Handbook of Epigenetics: The New Molecular and Medical Genetics [2 ed.] 0128053887, 9780128053881

Citation preview

HANDBOOK OF EPIGENETICS THE NEW MOLECULAR AND MEDICAL GENETICS SECOND EDITION

Edited by

Trygve O. Tollefsbol

University of Alabama at Birmingham, Birmingham, AL, United States

  

Academic Press is an imprint of Elsevier 125 London Wall, London EC2Y 5AS, United Kingdom 525 B Street, Suite 1800, San Diego, CA 92101-4495, United States 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom Copyright © 2017 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www.elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-805388-1 For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Mica Haley Acquisition Editor: Peter B. Linsley Editorial Project Manager: Lisa Eppich Production Project Manager: Chris Wortley Designer: Greg Harris Typeset by Thomson Digital

Contributors

Nobuyoshi Akimitsu  The University of Tokyo, Tokyo, Japan

Kimberly E. Hawkins  University of Florida, Gainesville, FL, United States

Juliana Almeida  University of São Paulo, São Paulo, Brazil; Royal Holloway, University of London, Egham, United Kingdom

Rita K. Hayford  Center for Integrated Biological & Environmental Research (CIBER), Delaware State University, Dover, DE, United States

Anthony Au  Institute of Bioproduct Development, University of Technology, Malaysia, Skudai, Johor, Malaysia

Zdenko Herceg  International Agency for Research on Cancer (IARC), Lyon, France

Vasudevan Ayyappan  Center for Integrated Biological & Environmental Research (CIBER), Delaware State University, Dover, DE, United States

Line Hjort  University of Copenhagen, Copenhagen, Denmark Xiaotong Hu  Sir Run Run Shaw Hospital, Zhejiang University, Hangzhou, Zhejiang, China

J. Armando Casas-Mollano  Yachay Tech University, Urcuquí, Ecuador

Eveline M. Ibeagha-Awemu  Agriculture and Agri-Food Canada, Sherbrooke Research and Development Centre, Sherbrooke, QC, Canada

Frances A. Champagne  Columbia University, New York, NY, United States Taiping Chen  The University of Texas MD Anderson Cancer Center, Smithville; Graduate School of Biomedical Sciences at Houston, Houston, TX, United States

Ali Jawaid  Brain Research Institute, Zürich, Switzerland

Ravindresh Chhabra  Panjab University, Chandigarh, Punjab, India

Venu (Kal) Kalavacharla  Center for Integrated Biological & Environmental Research (CIBER), Delaware State University, Dover, DE, United States

Astrid Jüngel  University Hospital Zurich, Gloriastrasse, Zurich, Switzerland

Tian Chi  ShanghaiTech University, Shanghai, China; Yale University Medical School, New Haven, CT, United States

Ashley M. Karnay  Drexel University College of Medicine, Philadelphia, PA, United States

James P. Curley  Columbia University, New York, NY, United States

Loo Keat Wei  Universiti Tunku Abdul Rahman, Kampar, Perak, Malaysia

Gary L. Dunbar  Central Michigan University, Mt. Pleasant; Field Neurosciences Institute, Saginaw, MI, United States

Hasan Khatib  University of Wisconsin, Madison, WI, United States

Mollee C. Dworkin  Center for Integrated Biological & Environmental Research (CIBER), Delaware State University, Dover, DE, United States

Kerstin Klein  University Hospital Zurich, Gloriastrasse, Zurich, Switzerland

Thomas Eggermann  Institute of Human Genetics, RWTH Aachen University, Aachen, Germany

Eloïse A. Kremer  Brain Research Institute, Zürich, Switzerland Ilkka Kronholm  Centre of Excellence in Biological Interactions, University of Jyväskylä, Jyväskylä, Finland

Felice Elefant  Drexel University, Philadelphia, PA, United States

Hervé Lalucque  Paris Diderot University, Paris, France

Hodaka Fujii  Research Institute for Microbial Diseases, Osaka University, Suita, Japan

Ho-Sun Lee  Seoul National University, Seoul, South Korea; International Agency for Research on Cancer (IARC), Lyon, France

Toshitsugu Fujita  Research Institute for Microbial Diseases, Osaka University, Suita, Japan

Yongqin Li  ShanghaiTech University, Shanghai, China

Krutika S. Gaonkar  Mayo Clinic, Rochester, MN, United States

Bo Liu  ShanghaiTech University, Shanghai, China Shuiping Liu  Sir Run Run Shaw Hospital, Zhejiang University, Hangzhou, Zhejiang, China

Steffen Gay  University Hospital Zurich, Gloriastrasse, Zurich, Switzerland

Chiao-Ling Lo  Indiana University School of Medicine, Indianapolis, IN, United States

Linn Gillberg  University of Copenhagen, Copenhagen, Denmark

Hanna Maciejewska-Rodrigues  University Hospital Zurich, Gloriastrasse, Zurich, Switzerland

Naoko Hattori  National Cancer Center Research Institute, Tokyo, Japan  

xi

xii

CONTRIBUTORS

Frédérique Magdinier  Aix Marseille University, French National Institute of Health and Medical Research (INSERM), Medical Genetics and Functional Genomics (GMGF), Marseille, France

Richard Saffery  Murdoch Children’s Research Institute, Parkville; University of Melbourne, Parkville, VIC, Australia Axel Schumacher  Munich, Bavaria, Germany Michael I. Shifman  Shriners Hospitals Pediatric Research Center (Center for Neural Repair and Rehabilitation), Philadelphia; Temple University School of Medicine, Philadelphia, PA, United States

Panchanan Maiti  Central Michigan University, Mt. Pleasant; Field Neurosciences Institute, Saginaw, MI, United States Fabienne Malagnac  University of Paris-Saclay, Paris, France Isabelle M. Mansuy  Brain Research Institute, Zürich, Switzerland

Philippe Silar  Paris Diderot University, Paris, France Athena Sklias  International Agency for Research on Cancer (IARC), Lyon, France

Shaoshuai Mao  ShanghaiTech University, Shanghai, China J. Alfredo Martinez  Centre for Nutrition Research, University of Navarra, Pamplona; Institute of Health Carlos III, Madrid, Spain

Susan L. Slager  Mayo Clinic, Rochester, MN, United States

Rahia Mashoodh  Columbia University, New York, NY, United States

Bhairavi Srinageshwar  Central Michigan University, Mt. Pleasant, MI, United States

Fermin I. Milagro  Centre for Nutrition Research, University of Navarra, Pamplona; Institute of Health Carlos III, Madrid, Spain

Mayavan Subramani  Center for Integrated Biological & Environmental Research (CIBER), Delaware State University, Dover, DE, United States

Rena Mizutani  The University of Tokyo, Tokyo, Japan

J. David Sweatt  Vanderbilt University, Nashville, TN, United States

Kjetil Søreide  Stavanger University Hospital, Stavanger; University of Bergen, Bergen, Norway

Shiraz Mujtaba  City University of New York, Brooklyn, NY, United States

Moshe Szyf  McGill University, Montreal, QC, Canada

Pamela N. Munster  University of California, San Francisco, CA, United States

Manuela Terranova-Barberio  University of California, San Francisco, CA, United States

Rabih Murr  Institute for Genetics and Genomics in Geneva (iGE3), University of Geneva, Geneva, Switzerland

Scott Thomas  University of California, San Francisco, CA, United States

Ru¯ ta Navakauskienė  Vilnius University, Vilnius, Lithuania

Shulan Tian  Mayo Clinic, Rochester, MN, United States

Tinh-Suong Nguyen  Paris Diderot University, Paris, France

Trygve O. Tollefsbol  Comprehensive Cancer Center, University of Alabama at Birmingham, Birmingham; Comprehensive Center for Healthy Aging, University of Alabama at Birmingham, Birmingham; Nutrition Obesity Research Center, University of Alabama at Birmingham, Birmingham; Comprehensive Diabetes Center, University of Alabama at Birmingham, Birmingham, AL, United States

Chris O’Neill  Kolling Institute for Medical Research,Royal North Shore Hospital, NSW, Australia Ifeanyi Okpala  City University of New York, Brooklyn, NY, United States Martin-Joseph Okpala  City University of New York, Brooklyn, NY, United States

Toshikazu Ushijima  National Cancer Center Research Institute, Tokyo, Japan

Zimuzoh Orakwue  City University of New York, Brooklyn, NY, United States

Ludovica Vanzan  Institute for Genetics and Genomics in Geneva (iGE3), University of Geneva, Geneva, Switzerland

Jeenah Park  University of California, San Francisco, CA, United States

Nicolas Veland  The University of Texas MD Anderson Cancer Center, Smithville; Graduate School of Biomedical Sciences at Houston, Houston, TX, United States

Jacob Peedicayil  Christian Medical College, Vellore, Tamil Nadu, India Gerd P. Pfeifer  Center for Epigenetics, Van Andel Research Institute, Grand Rapids, MI, United States

Günter Vogt  University of Heidelberg, Heidelberg, Germany

Vincenzo Pirrotta  Rutgers University, Piscataway, NJ, United States

Regan Vryer  Murdoch Children’s Research Institute, Parkville; University of Melbourne, Parkville, VIC, Australia

Tibor A. Rauch  Rush University Medical Center, Chicago, IL, United States

Martin M. Watson  Stavanger University Hospital, Stavanger, Norway

Marisol Resendiz  Indiana University School of Medicine, Indianapolis, IN, United States

Toshimichi Yamada  The University of Tokyo, Tokyo, Japan Huihuang Yan  Mayo Clinic, Rochester, MN, United States

Jérôme D. Robin  Institute for Research on Cancer and Aging, Nice (ICRAN), The National Center for Scientific Research (CNRS), French National Institute of Health and Medical Research (INSERM), Nice, France

Ericka Zacarias  Yachay Tech University, Urcuquí, Ecuador Allison Y. Zhong  University of California, San Francisco, CA, United States

Julien Rossignol  Central Michigan University, Mt. Pleasant, MI, United States

Feng C. Zhou  Indiana University School of Medicine, Indianapolis, IN, United States

  

Preface

Initiation of the second edition of Handbook of Epigenetics was undertaken for a number of reasons. First, the field of epigenetics has rapidly expanded since the first edition of this book appeared over half a decade ago. Although the first edition was very comprehensive, the current volume includes additional topics to maintain this comprehensive nature and to keep up with the expansion that the area of epigenetics has incurred since the release of the first edition of Handbook of Epigenetics. Second, the original topics that comprised the first edition have been updated and therefore contain the latest advances in each of the areas of epigenetics that are covered. Third, some topics from the first edition have been replaced with chapters that encompass more contemporary aspects of epigenetics, as well as cutting-edge techniques, to provide coverage of the latest developments in epigenetics. Fourth, many of the prior authors of the first edition of Handbook of Epigenetics were willing to contribute to a more updated volume, which serves to maintain the high standards that facilitated the success of the first edition of the book. The topics covered in this updated volume consist of an opening section that introduces the reader to the leading basic mechanisms of epigenetics, including DNA methylation, histone modifications, noncoding RNA, prions, higher-order chromatin organization, and polycomb mechanisms. As many of the breakthroughs in epigenetics have depended upon technological innovation, a number of chapters are also provided that cover numerous cutting-edge methodologies of epigenetics, which are revolutionizing this field of study. Likewise, the development of model organisms also serves as a driver of innovation, and the area of epigenetics has benefited from a number of developments in newer model organisms that are well covered in this updated volume. One of the more notable facets of epigenetics is that its biological impact is exceptionally vast. Many topics on this theme are covered and include the role of epigenetics in development, memory processes, transgenerational effects, aging, regeneration, stem cells, metabolic processes, and biomarker formation for diagnosis and

 

prognosis of epigenetic-based diseases. The evolution of the epigenetic machinery in plants and animals, as well as the role of epigenetics in adaptive evolution, is also included in this volume. In addition, a relatively newer topic of epigenetics is its role in epidemiology, and chapters are presented on the epigenetics of livestock breeding, nutriepigenomics, the impact of the environment on population epigenetics, and the influence of the gut microbiota on epigenetics, as well as population pharmacoepigenomics. The increasing role of epigenetics in human diseases and potential epigenetic therapies are highly important topics. This updated and expanded volume of Handbook of Epigenetics covers cancer epigenetics, as well as the role of epigenetics in autoimmune and brain disorders, metabolic diseases, and imprinting disorders. Fortunately, with the increasing knowledge of aberrant epigenetic mechanisms as an important factor in disease processes, the field has also experienced a renaissance in the development of agents that may prevent or treat many epigenetic disorders. Leading the way are the innovations in development and applications of DNA demethylating agents, as well as histone deacetylase inhibitors. Further, many advances with respect to novel combinations of epigenetic-based therapy are occurring, which opens a wealth of possibilities for treating diseases initiated or exacerbated by epigenetic aberrations. The future of epigenetics is certainly bright and the closing chapter of this book provides an illuminating vision of the many directions that this field will likely undergo over the ensuing years, perhaps most notably, the long sought-after techniques that allow direct epigenetic modifications to specific loci through epigenetic-editing methods. It is intended that the second edition of Handbook of Epigenetics will interest advanced students and researchers, as well as healthcare personnel in academics, industry, and clinical settings, and enhance the understanding and applications of the most recent advances in epigenetics in a comprehensive, practical, and fascinating manner.

xiii

Trygve O. Tollefsbol

C H A P T E R

1 An Overview of Epigenetics Trygve O. Tollefsbol Comprehensive Cancer Center, University of Alabama at Birmingham, Birmingham, AL, United States; Comprehensive Center for Healthy Aging, University of Alabama at Birmingham, Birmingham, AL, United States; Nutrition Obesity Research Center, University of Alabama at Birmingham, Birmingham, AL, United States; Comprehensive Diabetes Center, University of Alabama at Birmingham, Birmingham, AL, United States

O U T L I N E Introduction

3

Epigenetic Epidemiology

6

Molecular Mechanisms of Epigenetics

3

Epigenetics and Human Disease

6

Methods in Epigenetics

4

Epigenetic Therapy

7

Model Organisms of Epigenetics

5

The Future of Epigenetics

7

Factors Influencing Epigenetic Changes

5

Conclusions

7

Evolutionary Epigenetics

6

References

8

INTRODUCTION

MOLECULAR MECHANISMS OF EPIGENETICS

The term epigenetics was first introduced in 1942 by Conrad Waddinton and was defined as the causal interactions between genes and their products that allow for the phenotypic expression [1]. This term has now been somewhat redefined, and although there are many variants of the definition of this term today, a consensus definition is that epigenetics is the collective heritable changes in phenotype due to processes that arise independent of the primary DNA sequence. This heritability of epigenetic information was for many years thought to be limited to mitotic cellular divisions. However, it is now apparent that epigenetic processes can be transferred meiotically in organisms from one generation to another [2–3]. This phenomenon was first described in plants [4] and has been expanded to include yeast, Drosophila, mouse, and possibly humans [5–7].

Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00001-8 Copyright © 2017 Elsevier Inc. All rights reserved.

In most eukaryotes DNA methylation, the most studied of epigenetic processes, consists of transfer of a methyl moiety from S-adenosylmethionine (SAM) to the 5-position of cytosines in certain CpG dinucleotides. This important transfer reaction is catalyzed by the DNA methyltransferases (DNMTs). The three major DNMTs are DNMT1, 3A, and 3B, and DNMT1 catalyzes what is referred to as maintenance methylation that occurs during each cellular replication as the DNA is duplicated. The other major DNMTs, 3A and 3B, are characterized by their relatively higher de novo methylation activity where new 5-methylcytosines (5mCs) are introduced in the genome at sites that were not previously methylated. The most significant aspect of DNA methylation, which can also influence processes, such as X-chromosome

3

4

1.  An Overview of Epigenetics

inactivation and cellular differentiation, is its effects on gene expression. In general, the more methylated a gene regulatory region, the more likely it is that the gene activity will become downregulated and vice versa, although there are some notable exceptions to this dogma [8]. Chapter 2 of this book reviews the mechanisms of DNA methylation and demethylation during mammalian development. Recent advances have highlighted important roles of the ten-eleven translocation (TET) family of dioxygenases. These enzymes convert 5mC to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC), and appear to play important roles in the dramatic DNA demethylation that occurs during early development. The 5mC-oxidized derivatives may also serve as epigenetic marks that modulate chromatin regulation. Chromatin changes are another central epigenetic process that have an impact not only on gene expression, but also many other biological processes. Posttranslational modifications of histones, such as acetylation and methylation, occur in a site-specific manner and often influence the binding and activities of other proteins that influence gene regulation. The histone acetyltransferases (HATs) catalyze histone acetylation, and the histone deacetylases (HDACs) result in removal of acetyl groups from key histones that comprise the chromatin. These modifications can occur at numerous sties in the histones and are most common in the amino terminal regions of these proteins, as reviewed in Chapter 3. In general, increased histone acetylation is associated with greater gene activity and vice versa. Methylation of histones has variable effects on gene activity where lysine 4 (K4) methylation of histone H3 is often associated with increasing gene activity, whereas methylation of lysine 9 (K9) of histone H3 may lead to transcriptional repression. There is also considerable crosstalk between DNA methylation and histone modifications [9], such that cytosine methylation may increase the likelihood of H3K9 methylation and H3K9 methylation may promote cytosine methylation. The most exciting advances in epigenetics have been the discoveries that indicate that many other processes besides DNA methylation and histone modifications impact the epigenetic behavior of cells. For instance, noncoding RNA (Chapter 4), including both short (200 nucleotides) forms, often share protein and RNA components with the RNA interference (RNAi) pathway and they may also influence more traditional aspects of epigenetics, such as DNA methylation and chromatin marking. These effects appear to be widespread and occur in organisms ranging from protists to humans and the noncoding RNAs may serve as therapeutic targets for a number of human diseases. Prions are fascinating in that they can influence epigenetic processes independent of DNA and chroma-

tin. In Chapter 5 it is shown that structural heredity is also important in epigenetic expression where alternative states of macromolecular complexes or regulatory networks can have a major effect on phenotypic expression independent of changes in DNA sequences. The prion proteins are able to switch their structure in an autocatalytic manner that can not only influence epigenetic expression, but also lead to human disease. The position of a gene in a given chromosome can also greatly influence its expression (Chapter 6). Upon rearrangement, a gene may be relocated to a heterochromatic region of the genome, leading to gene silencing. Moreover, a change in position of a regulatory element may affect the maintenance of chromatin architecture and subsequently cellular functioning. Polycomb mechanisms are another relatively new aspect of epigenetics that control all of the major cellular differentiation pathways and are also involved in cell fate. Polycomb repression is very dynamic and can be easily reversed by activators, and they also raise the threshold of the signals or activators required for transcriptional activation, which places these fascinating proteins within the realm of epigenetic processes. Polycomb complexes can generate H2A ubiquitylation and H3K27 methylation that often mediate their repressive functions, and H3K27 may serve as an epigenetic memory for Polycomb repression as described in Chapter 7. Therefore, although DNA methylation and histone modifications are mainstays of epigenetics, recent advances have greatly expanded the epigenetic world to include many other processes, such as noncoding RNA, prions, chromosome position effects, and Polycomb mechanisms.

METHODS IN EPIGENETICS Many of the advances in epigenetics that have driven this field for the past 2 decades can be traced back to the technological breakthroughs that have made the many discoveries possible. We now have a wealth of information about key gene-specific epigenetic changes that occur in a myriad of biological processes. In Chapter 8, gene-specific techniques for determining DNA methylation are reviewed. These methods include bisulfite sequencing, methylation-specific PCR (MSP), real-time MSP, MethyLight, pyrosequencing, and high-resolution melting analyses. These techniques can be applied not only to mechanisms of epigenetic gene control, but also to diagnostic processes as well. In addition, there have been important breakthroughs in analyses of the methylome at a high resolution. Microarray platforms and high-throughput sequencing, which have largely replaced microarray platforms, have made possible new techniques to analyze genomewide features of epigenetics that are based on uses of methylation-sensitive

I.  Overview



Factors Influencing Epigenetic Changes

restriction enzymes, sodium bisulfite conversion, and affinity capture with antibodies or proteins that select methylated DNA sequences. Tiling arrays that may be employed for analysis of chromosomal segments or the whole genome and whole genome high-throughput sequencing are now a staple for reliable and specific methylomic analyses as described in Chapter 9. Chromatin immunoprecipitation and sequencing (ChIP-seq) is now frequently used in epigenetic analyses and can also be applied to single cells, as well as used for mapping two modifications simultaneously as described in Chapter 10. Likewise, genomewide expression analyses of RNAs have been made possible with RNA-seq that allows detection of a myriad of noncoding RNAs (Chapter 11). As there has been much information derived from epigenomic approaches, methods to analyze data from ChIP-seq and RNA-seq, for example, are becoming increasingly important and are delineated in Chapter 12. There is no question that developments in the tools for assessing epigenetic information have been and will continue to be an important driving force in advancing epigenetics.

MODEL ORGANISMS OF EPIGENETICS Epigenetic processes are widespread and much of our extant knowledge about epigenetics has been derived from model systems, both typical and unique. The ease of manipulation of prokaryotic organisms has facilitated discoveries in the molecular mechanisms of basic epigenetic processes. Although prokaryotic organisms do not contain chromatin, some viruses and bacteria are capable of encoding factors, which lead to posttranslational modifications of an epigenetic nature as described in Chapter 13. Drosophila is a mainstay model in biology in general and the epigenetics field is not an exception in this regard. For example, Chapter 14 discusses how the use of Drosophila as a model organism has significantly increased knowledge of chromatin organization and may have powerful potential in facilitating understanding of the epigenetics of neurological disorders. Probably the most useful model system in epigenetics to date is the mouse model (Chapter 15). Numerous different mouse models have been developed that are important in many different epigenetic processes, such as transgenerational epigenetics and imprinting, and these models have potential in illuminating human diseases, such as diabetes, neurological disorders, and cancer. Plant models (Chapter 16) are of great importance in epigenetics in part due to their plasticity and their ability to silence transposable elements. RNAi silencing in plants has been at the forefront of epigenetics and plant models will likely lead the way in several other epigenetic processes in the future. Use of plant models has also facilitated the development

5

of targeted epigenome editing in plants that may have great significance in enhancing crop performance. Thus, model development, such as the advances in techniques, have made many of the most exciting discoveries in epigenetics possible for a number of years.

FACTORS INFLUENCING EPIGENETIC CHANGES The functions of epigenetics are indeed numerous and it would be next to impossible to do complete justice in one book to this ever-expanding field. However, Chapters 17–25 illustrate a few of the many different functions that epigenetics mediates. For example, Chapter 17 reviews the role of epigenetic in development and illustrates that both gene-specific and global changes in DNA methylation are central to embryonic development. In addition, perturbations mediated by processes, such as stress during development, may lead to disease formation later in life. Chapter 18 reviews epigenetic biomarkers that are also very important and serve the key function of informing the staging and classification of disease, as well as guiding clinical management. Epigenetics is intricately linked to changes in the metabolism of organisms and these two processes cannot be fully understood separately. SAM is a universal methyl donor and drives many epigenetic processes (Chapter 19), and the importance of SAM in epigenetic mechanisms is great. Metabolic functions can also influence the chromatin, which is a major mediator of epigenetic processes (Chapter 20). It is now apparent that various environmental influences and metabolic compounds can regulate many enzymes that modify histones in mammals. Thus, metabolic processes impact both DNA methylation and chromatin remodeling, the two major epigenetic mediators, and it is likely that this field will continue to advance in an exponential manner. Stem cells rely in part on signals from the environment and epigenetic mechanisms have central roles in how stem cells respond to environmental influences (Chapter 21) and how therapeutic roles of stem cells may contribute greatly to management of many diseases, such as multiple sclerosis, Huntington’s disease, and Parkinson’s disease. Regenerative medicine is dependent upon stem cells and regulation of the neuron’s regenerative abilities after spinal cord injury involves key changes in the epigenome that regulate gene expression (Chapter 22). Profound new discoveries have occurred in the area of the epigenetics of memory processes. Recent exciting discoveries have shown that gene regulation through epigenetic mechanisms is necessary for changes in adult brain function and behavior based on life experiences (Chapter 23). Moreover, new drugs that impact epigenetic mechanisms may have future uses

I.  Overview

6

1.  An Overview of Epigenetics

in treating or alleviating cognitive dysfunction. Transgenerational inheritance (Chapter 24) is also a form of memory based in part on epigenetics in that early life experiences that impact epigenetic markers can greatly influence adult health and risk for diseases. In addition, the aging process is a form of epigenetic memory and experience in that our genes are epigenetically modified from our parents, as well as during our entire life spans, that can significantly impact the longevity of humans, as well as our risk for the many age-related diseases, many of which are also epigenetically based (Chapter 25). It is, therefore, apparent that epigenetics influences many different functions, and it is highly likely that many additional functions of epigenetics will be discovered in the future.

EVOLUTIONARY EPIGENETICS Although many think of epigenetic processes as being inherent and static to a specific organism, it is apparent that epigenetics has been a major force behind the evolutionary creation of new species. Chapter 26 reveals that epigenetic mediators, such as H3K4 and H3K27 methyltransferases, have significantly impacted plant evolution. The expansion of gene families of these enzymes may have allowed for changes in chromatin regulation of new genes and pathways that have modified plant evolution. The epigenetic machinery has also played a major role in the evolution of animals. As described in Chapter 27, DNA methylation may serve as a driver of evolution through its effects on gene regulation. Vertebrates have evolved genomewide methylation, while invertebrates are characterized by mosaic DNA methylation patterns and epigenotype diversification, followed by genetic fixation could be a major mechanism of evolution that is currently understudied and may receive increasing attention in future analyses of epigenetics in evolutionary processes. Adaptive evolution by natural selection requires heritability and spontaneous epigenetic changes may have been important in adaptive evolution; although, there are important differences between genetic variation and epigenetic variation with respect to supply and stability of epigenetic mutations (Chapter 28).

EPIGENETIC EPIDEMIOLOGY A very new and exciting area within the field of epigenetics is the impact of epigenetic mechanisms on livestock breeding. Chapter 29 indicates that the breeding programs of livestock that are currently in place account for only part of the phenotypic variance in traits and that epigenetic factors of variance need further consider-

ation. In fact, epigenetic modifications may have a major impact on the growth and development of livestock and profoundly affect phenotypic outcomes. Dietary factors are highly variable not only between individuals, but also among human populations and various nonhuman species. Many studies have shown that the diet has a profound effect on the epigenetic expression of the genome and, therefore, on the phenotype. An important example of this phenomenon is the link between the nutritional epigenome and the metabolic syndrome (Chapter 30) in that early epigenetic changes through nutritional means may exacerbate the risk of development of the metabolic syndrome later in life. Environmental agents other than diet also impact the epigenome. For example, Chapter 31 reviews the many environmental agents that can lead to alterations in the epigenome, thereby inducing toxicity or carcinogenesis. The environment also affects the gut microbiome (Chapter 32), which is known to impact epigenetic processes, and increased understanding of the dynamic interactions between the gut microbiome and epigenetic modifications may have considerable translational potential. Drugs also reshape the epigenome, which has opened the new field of pharmacoepigenomics. It is clear that certain populations respond differently to drugs and much of this variation may be explained by epigenetic factors (Chapter 33). Thus, epidemiological factors have great importance in epigenetics, and this is influenced by diet, environmental agents, the gut microbiome, drugs, and likely many other factors as well.

EPIGENETICS AND HUMAN DISEASE For the medical community, a major interest in epigenetics stems from the role of epigenetic changes in the etiology, progression, and diagnosis of human diseases. Cancer has long been associated with epigenetic alterations, and DNA methylation, chromatin modifications, and RNA-dependent regulation have all been shown to affect the incidence and severity of cancer (Chapter 34). Many immune disorders, such as systemic lupus erythematosus (SLE) and rheumatoid arthritis, as well as multiple sclerosis, have been associated with epigenetic aberrations (Chapter 35), and epigenetic processes have also been linked to brain disorders (Chapter 36). In the latter case, the Rett syndrome, Alzheimer’s diseases, Huntington’s disease, and even autism, to name a few, have been associated in at least some way with epigenetic alterations. Even schizophrenia and depression may have an epigenetic basis in their expression. However, system metabolic disorders may also be related to epigenetic aberrations. For example, obesity, gestational diabetes, and hypertension can influence the fetal chromatin and lead to an increased incidence in adult disease later in life (Chapter 37). As genomic imprinting is based

I.  Overview



7

Conclusions

on epigenetic mechanisms, it may come as no surprise that defects in imprinting can lead to a number of human diseases (Chapter 38). The Prader–Willi syndrome, Angelman syndrome, Silver Russell syndrome, and many other imprinting disorders, such as transient neonatal diabetes mellitus, are due to imprinting disorders that are based on epigenetic defects. Therefore, the number of diseases impacted by epigenetic processes is large and advances in the treatment of these disorders will likely depend in part on breakthroughs in epigenetic therapy.

EPIGENETIC THERAPY The development of DNA demethylation agents in clinical medicine as reviewed by Moshe Szyf in Chapter 39 has been a topic of high interest in epigenetics for decades. Although many studies have focused on use of demethylating agents in treating various hematological cancers, newer studies have also found that neuroepigenetic aberrations may be important targets for treating mental health disease. The use of combinatorial approaches using demethylating agents with classical agents may also have increasing potential in therapy. Although there are many epigenetic therapies that are in use and on the horizon, histone-modifying drugs have probably received the most attention in the clinics. Chief among these are the HDAC inhibitors. Vorinostat (Zolinza), for example, has been approved by the Food and Drug Administration for use in the treatment of patients with cutaneous T-cell lymphoma (Chapter 40). Many different HDAC inhibitors have been developed, and it is likely that significant improvements will occur for HDAC inhibitors, as well as many other drugs that can normalize aberrations in not only histone modifications, but also DNA methylation and perhaps some of the many other epigenetic processes that have been discovered. On the horizon is an increasing interest in combinatorial approaches using epigenetic-modifying agents with other forms of therapy as aforementioned. Moreover, as described in Chapter 41, the combined use of epigenetic-modifying drugs, such as DNA methylation and histone methyltransferase, as well as histone acetylase inhibitors may have significant potential.

THE FUTURE OF EPIGENETICS For many decades the field of epigenetics has grown exponentially in part because its many effects on the epigenome, which has led to a vast number of biological manifestations, involving both natural, as well as diseases processes. However, a significant impediment to progress in this field has been an outgrowth of the same phenomenon, that is, that the vast effects epigenetics has

on the genome has made locus-specific studies of epigenetics processes somewhat limited. However, recent developments may overcome this persistent impediment. Exciting new advances in epigenetic editing may contribute significantly to the long sought-after targeting of epigenetic modifications. The development of the clustered regularly interspaced short palindromic repeats (CRISPR) and CRISPR-associated protein (Cas) (CRISPR/Cas) system now enables epigenetics researchers to target loci of interest and to determine the modifiers of interest in a locus-specific manner as described in Chapter 42. This capacity for locus-specific epigenetic editing may spark a new revolution in epigenetic research and finally allow us to treat many diseases that are manifested by epigenetic aberrations with a significantly reduced risk to the patient due to the more targeted approach it manifests.

CONCLUSIONS Advances in understanding the basic machinery that drives the epigenetic processes of DNA methylation and histone modifications have raised the field of epigenetics well beyond original expectations. This area of research has also significantly expanded horizontally in that additional epigenetic processes, such as noncoding RNA, prion changes, and Polycomb mechanisms have now been established, and it is likely that more epigenetic processes will be discovered in the not too distant future. A major driving force in epigenetics has been the outstanding development of new technology that has not only served to stimulate new discoveries, but has also expanded the field by allowing for novel discoveries possible only through the use of these new tools. Advances in new model organisms for understanding epigenetic processes have also greatly stimulated this field of study. We now know that epigenetics is not only intricately associated with metabolism, but also functions in stem cell behavior, tissue regeneration, the transfer of information through generations, neurological memory processes, and even the aging of organisms. Epigenetics has also played roles in evolution, and has served as a molecular driver of mutations. Moreover, the changing environment is currently reshaping the evolution of many organisms through plastic epigenetic processes. Epidemiological factors, such as diet, environmental exposure, the gut microbiome, and drugs are also influencing our daily lives through epigenetics. Diseases that have been associated with epigenetic processes range from schizophrenia to cancer; the list of these diseases is ever-expanding. Fortunately, the field of epigenetic therapy is also expanding and the hope is that the future will see many novel treatments for the numerous diseases that are derived from epigenetic defects. Advances

I.  Overview

8

1.  An Overview of Epigenetics

in locus-specific targeting of epigenetics through epigenetic editing and CRISPR/Cas offers but one example, albeit an important, exciting, and perhaps revolutionizing example of changes in the future directions of epigenetic research.

References [1] Waddington CH. The epigenotype. Endeavour 1942;1:18–20. [2] Li Y, Saldanha SN, Tollefsbol TO. Impact of epigenetic dietary compounds on transgenerational prevention of human diseases. AAPS J. 2014;16:27–36. [3] Tollefsbol TO, (editor). Transgenerational epigenetics: evidence and debate. London: Academic Press; 2014.

[4] Brink RA, Styles ED, Axtell JD. Paramutation: directed genetic change. Paramutation occurs in somatic cells and heritably alters the functional state of a locus. Science 1968;159:161–70. [5] Cavalli G, Paro R. Epigenetic inheritance of active chromatin after removal of the main transactivator. Science 1999;286:955–8. [6] Grewal SI, Klar AJ. Chromosomal inheritance of epigenetic states in fission yeast during mitosis and meiosis. Cell 1996;86:95–101. [7] Rakyan VK, Blewitt ME, Druker R, Preis JI, Whitelaw E. Metastable epialleles in mammals. Trends Genet 2002;18:348–51. [8] Lai SR, Phipps SM, Liu L, Andrews LG, Tollefsbol TO. Epigenetic control of telomerase and modes of telomere maintenance in aging and abnormal systems. Front Biosci 2005;10:1779–96. [9] Fuks F, Burgers WA, Brehm A, Hughes-Davies L, Kouzarides T. DNA methyltransferase Dnmt1 associates with histone deacetylase activity. Nat Genet 2000;24:88–91.

I.  Overview

C H A P T E R

2 Mechanisms of DNA Methylation and Demethylation During Mammalian Development Nicolas Veland, Taiping Chen The University of Texas MD Anderson Cancer Center, Smithville, TX, United States; Graduate School of Biomedical Sciences at Houston, Houston, TX, United States

O U T L I N E Introduction

11

Conclusions

20

DNA Methylation Maintenance DNA Methylation De novo DNA Methylation

12 12 14

References

22

DNA Demethylation Passive DNA Demethylation Active DNA Demethylation

16 17 17

INTRODUCTION

methylation patterns are altered in pathological conditions. For example, cancer cells usually exhibit global hypomethylation and local hypermethylation, which contribute to genomic instability and tumor suppressor silencing, respectively. 5mC is not randomly distributed in the genome. In general, repetitive DNA sequences, including transposable elements and centromeric and pericentric satellite DNA, are heavily methylated. In contrast, CpG islands (CGIs, 1–2 kilobases of GC-rich regions) present in gene promoters are usually depleted of DNA methylation, with some exceptions. For example, CpGs on the inactive X chromosome in female cells are hypermethylated, and CpGs in imprinting control regions (ICRs) exhibit allele-specific methylation. On the other hand, gene bodies, especially exons, are often highly methylated. Unlike promoter methylation, which correlates with gene silencing, gene body methylation is often associated with transcriptional activity [3].

DNA methylation, the covalent addition of a methyl group to the fifth position of cytosine (5-methylcytosine, 5mC), is the most common form of DNA modification, which plays important roles in the regulation of chromatin structure and gene expression. DNA methylation is present in various organisms including many animals, plants, and fungi. In mammals, DNA methylation mostly occurs in the context of CpG dinucleotides, with 60%–80% of all CpGs in the genome being methylated, although non-CpG (i.e., CpA, CpT, or CpC) methylation is abundant in specific tissues and cell types, including embryonic stem cells (ESCs), oocytes, and brain tissue [1]. DNA methylation is essential for mammalian development and plays key roles in multiple biological processes, such as gene expression, genomic imprinting, X-chromosome inactivation, and transposon silencing [2]. Consistent with its pleiotropic roles, DNA Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00002-X Copyright © 2017 Elsevier Inc. All rights reserved.

11

12

2.  Mechanisms of DNA Methylation and Demethylation During Mammalian Development

DNA methylation patterns and levels are determined by the opposing actions of the methylation and demethylation machineries. The methylation machinery includes DNA methyltransferases (DNMTs), which catalyze the transfer of a methyl group from the methyl donor S-adenosyl-l-methionine (AdoMet or SAM) to the C-5 position of cytosines. There are three active DNMTs in mammals; DNMT1 is mainly responsible for maintaining DNA methylation patterns during DNA replication, whereas DNMT3A and DNMT3B function primarily as de novo methyltransferases that establish DNA methylation patterns [4]. DNMT2, a protein with conserved DNMT catalytic motifs, turned out to be an aspartic acid tRNA methyltransferase and has been renamed tRNA aspartic acid (D) methyltransferase 1 (TRDM1) [5]. From a chemical perspective, DNA methylation is considered a relatively stable modification. However, global demethylation occurs in preimplantation embryos and primordial germ cells (PGCs), and locus-specific demethylation takes place during cellular differentiation. DNA demethylation can be achieved by replicationdependent “passive” dilution of 5mC and replicationindependent “active” processes [6,7]. Great progress has been made in understanding the mechanisms of demethylation over the past several years, thanks to the discovery that the TET family of proteins—TET1, TET2, and TET3—function as 5mC dioxygenases that convert 5mC to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxylcytosine (5caC) [8–11]. There is

compelling evidence that 5hmC, 5fC, and 5caC serve as intermediates for DNA demethylation [6,7].

DNA METHYLATION In 1975, Holliday, Pugh and Riggs proposed that DNA methylation could be important for cellular memory by serving as a heritable epigenetic mark through cell division. Based on the complementarity of CpG/ CpG dyads, they reasoned that methylated CpG sites could be replicated semiconservatively during DNA replication [12,13]. The theory would predict the existence of two DNA methyltransferase activities; de novo methyltransferase(s) methylate unmodified DNA to establish DNA methylation patterns, and maintenance methyltransferase(s) methylate newly formed hemimethylated CpG sites during DNA replication to maintain the patterns (Fig. 2.1A). The hypothesis was subsequently validated, to a large extent, by the identification of DNA methyltransferases with distinct expression patterns, biochemical properties, and biological functions [4].

Maintenance DNA Methylation Once DNA methylation patterns are established during early embryogenesis, they are maintained in somatic cells in a cell type-specific manner. During each round

FIGURE 2.1  Overview of mechanisms of DNA methylation and demethylation. (A) During early development, de novo methylation adds methyl groups to specific CpG sites on both DNA strands (resulting in symmetric CpG methylation) and establishes DNA methylation patterns. After each round of DNA replication, the methylated CpG sites become hemimethylated, as the newly replicated daughter DNA strand is unmethylated. Maintenance methylation recognizes hemimethylated CpG sites and “copies” the DNA methylation pattern of the parental strand onto the daughter strand. (B) Failure in maintenance methylation results in replication-dependent loss of DNA methylation, a process known as passive demethylation. DNA demethylation can also be achieved by enzyme-mediated removal of the methyl group or replacement of methylated cytosines (or their derivatives) with unmodified cytosines, a process known as active demethylation, which is usually independently of DNA replication. Open and filled circles indicate unmethylated and methylated CpG dinucleotides, respectively.

II.  Molecular Mechanisms of Epigenetics



DNA Methylation

of DNA replication, DNA becomes hemimethylated, as only CpGs on the parental strand remain methylated while CpGs on the newly replicated daughter strand are unmethylated. To reestablish the symmetry of CpG methylation and keep the specificity, the maintenance DNA methyltransferase activity recognizes hemimethylated CpGs and methylates the corresponding CpGs on the daughter strand. Biochemical, cellular, and genetic evidence suggests that DNMT1 is the major maintenance methyltransferase [4]. In addition, a multidomain protein, UHRF1 (ubiquitin-like with PHD and RING finger domains 1), is essential for directing DNMT1 to DNA replication sites [14,15]. DNMT1 Mouse DNMT1, the first mammalian DNA methyltransferase gene identified [16], has several transcription start sites and produces two major protein products. Transcripts initiated within a somatic cell-specific exon encode the full-length DNMT1 protein that is expressed in somatic cell types, whereas transcripts initiated within an oocyte-specific exon utilize a downstream AUG as the translation initiation codon, resulting in the DNMT1o isoform that lacks the N-terminal 118 amino acids of fulllength DNMT1 (Fig. 2.2) [4]. Both isoforms are equally

13

functional in maintaining DNA methylation, although DNMT1o is more stable than DNMT1 [17]. DNMT1 has an N-terminal regulatory domain and a C-terminal catalytic domain (Fig. 2.2). While all DNMTs contain highly conserved DNA methyltransferase motifs in their catalytic domains, the DNMT1 and DNMT3 families show little sequence similarity in their regulatory domains. There are several unique domains in the N-terminal region of DNMT1 that confer its functional difference from DNMT3A/3B. A region at the very N terminus mediates the interaction between DNMT1 and DNA methyltransferase associated protein 1 (DMAP1), a protein implicated in histone acetylation and ATM signaling. The DMAP1-interaction domain is absent in the more stable DNMT1o isoform [17], suggesting that this domain or the interaction between DNMT1 and DMAP1 may be involved in regulating DNMT1 stability. The DMAP1-interaction domain is followed by a proliferating cell nuclear antigen (PCNA) binding domain (PBD), which is required for the interaction with the DNA replication machinery, and a nuclear localization signal (NLS) [4]. DNMT1 also contains a motif originally named as the replication foci-targeting sequence (RFTS). Recent evidence indicates that RFTS contains a ubiquitin-interacting motif (UIM) that recognizes ubiquitinated histone

FIGURE 2.2  Schematic diagram of major proteins involved in DNA methylation. (A) Protein domain architecture of mouse DNA methyltransferases (DNMTs). The DNMT1 and DNMT3 families of proteins share conserved catalytic motifs (I–X) in their C-terminal catalytic domains. DNMT3L lacks catalytic activity because some essential motifs are missing or mutated. The N-terminal regions of DNMT1 and DNMT3 proteins have little sequence similarity, with distinct domains that contribute to their functional specificities. DMAP1, DNA methyltransferase associated protein 1; PBD, PCNA-binding domain; NLS, nuclear localization sequence; RFTS, replication foci targeting sequence; UIM, ubiquitin interacting motif; CXXC, cysteine-rich motif; BAH, bromo-adjacent homology; GK, glycine/lysine-rich linker; I–X, DNA methyltransferase conserved catalytic motifs; PWWP, proline-tryptophan-tryptophan-proline; ADD, ATRX-DNMT3-DNMT3L. The start sites of the DNMT1o and DNMT3A2 isoforms are indicated. Locations of the most common alternatively spliced exons (exons 10, 11, 20, 21, 22) in DNMT3B are also indicated. (B) Protein domain architecture of mouse UHRF1. UBL, ubiquitin-like domain; TTD, tandem tudor domain; PHD, plant homeodomain; SRA, SET and RING associated; RING, really interesting new gene.

II.  Molecular Mechanisms of Epigenetics

14

2.  Mechanisms of DNA Methylation and Demethylation During Mammalian Development

H3 at lysine 18 (H3K18ub), a histone modification that serves as a docking site for DNMT1 targeting to replication foci [18]. Structural data obtained recently suggest that the RFTS domain also plays an autoinhibitory role in the regulation of DNMT1 activity by binding to the catalytic domain and blocks the catalytic center [19]. Additionally, DNMT1 contains a CXXC domain, a cysteinerich motif that binds unmethylated CpGs, and a pair of bromo-adjacent homology (BAH) domains, BAH1 and BAH2, with their role remaining unknown [4]. Biochemical assays reveal that DNMT1 preferentially methylates hemimethylated CpG dinucleotides, although it is capable of methylating unmethylated substrates as well [20]. DNMT1 is constitutively expressed in proliferating cells. During S phase, DNMT1 is upregulated and associates with the replication foci, suggesting that DNMT1-mediated methylation is coupled to DNA replication [21]. Genetic inactivation of DNMT1 in mouse ESCs results in global loss of DNA methylation, but does not affect de novo methylation of integrated provirus DNA [22,23]. Taken together, these results suggest that DNMT1 functions primarily as a maintenance methyltransferase. DNA methylation is dispensable in undifferentiated mouse ESCs, as DNMT1 knockout (KO), DNMT3A/3B double KO (DKO), and DNMT1/3A/3B triple KO (TKO) ESCs show no defects in viability and proliferation, but these cells die when induced to differentiate [23–27]. A recent study shows that human ESCs require DNMT1, but not DNMT3A and DNMT3B, for survival [28]. Mouse and human ESCs represent different pluripotent states, with human ESCs resembling the more mature epiblast state, which may explain the sensitivity of human ESCs to hypomethylation. DNMT1 is also required for the survival of mouse embryonic fibroblasts (MEFs) and the human colorectal cancer cell HCT116 [29,30]. These findings suggest crucial roles for DNMT1 in cellular differentiation and in the viability of differentiated cells. Consistent with this notion, complete inactivation of DNMT1 results in the arrest of mouse embryonic development around E9.5, when the embryo is in the process of differentiating into the three germ layers [23]. In human, missense DNMT1 mutations in the RFTS of the N-terminal regulatory domain, which likely results in hypomorphic alleles, have been identified in two related neurodegenerative disorders, hereditary sensory, and autonomic neuropathy with dementia and hearing loss type IE (HSAN IE) and autosomal dominant cerebellar ataxia, deafness, and narcolepsy (ADCA-DN) [31]. While the mechanisms by which these mutations lead to the disease phenotypes are unknown, changes in DNA methylation and gene expression likely play an important role. UHRF1 UHRF1, also known as nuclear protein of ∼95-kDa (NP95) in mouse and inverted CCAAT box-binding

protein (ICBP90) in human, is a multidomain protein. Genetic studies demonstrate that UHRF1 is essential for maintaining DNA methylation. Similar to the phenotype of DNMT1 deficiency, disruption of UHRF1 leads to embryonic lethality and global DNA hypomethylation [14,15,32]. Cellular studies also suggest functional interactions between DNMT1 and UHRF1. Both proteins are enriched at DNA replication foci during the S phase in normal cells, but DNMT1 fails to localize to these foci in UHRF1 KO cells [14,15]. These findings indicate that UHRF1 is a critical regulator of maintenance methylation by directing DNMT1 to hemimethylated CpG sites. UHRF1 has five conserved domains; a ubiquitin-like domain (UBL), a tandem tudor domain (TTD), a plant homeodomain (PHD), a SET and RING associated (SRA) domain, and a Really Interesting New Gene (RING) domain (Fig. 2.2). The mechanisms by which UHRF1 controls DNMT1 localization and DNA methylation are complex, and all the conserved domains, with the exception of UBL, have been implicated. UHRF1 interacts with DNMT1 and likely mediates its recruitment to chromatin, as the TTD and PHD of UHRF1 act in combination to recognize the histone H3 trimethylated at lysine 9 (H3K9me3) mark, and additionally, the PHD binds to histone H3 with unmethylated arginine 2 (H3R2me0) [33,34]. The SRA domain preferentially binds hemimethylated DNA and likely plays an important role in loading DNMT1 onto newly synthesized DNA substrates [14,15]. The RING finger domain has E3 ubiquitin ligase activity. UHRF1 has been shown to ubiquitinate histone H3 at lysine 18 (H3K18ub) in mammalian cells [18] and at lysine 23 (H3K23ub) using the cell-free system of Xenopus egg extracts [35]. Both studies show that UHRF1-dependent histone H3 ubiquitination is required for maintenance DNA methylation, suggesting that the H3 ubiquitination mark(s) provides a binding site(s) for DNMT1 [18,35].

De Novo DNA Methylation After fertilization, both the maternal and paternal genomes undergo global DNA demethylation during preimplantation development. As a result, DNA methylation marks inherited from gametes are largely erased by the blastocyst stage, with the exception of those in ICRs and some retrotransposons. After implantation, a wave of de novo methylation occurs in the epiblast to establish the initial pattern of DNA methylation. DNA methylation shows further changes during cellular differentiation, and the patterns are then stably maintained in a lineage-specific manner after successive cell divisions. Similar epigenetic reprograming events also take place during gametogenesis, including global demethylation in PGCs. Subsequently, a new round of de novo DNA methylation takes place in germ cells, which have already passed sex-determination, resulting in different

II.  Molecular Mechanisms of Epigenetics



DNA Methylation

methylation patterns in male and female gametes (sperm and egg) [2]. De novo methylation is mediated by DNMT3A and DNMT3B. A third member of the DNMT3 family, DNMT3-like (DNMT3L), has no catalytic activity, but is an important regulator of de novo methylation (Fig. 2.2) [4]. DNMT3A and DNMT3B The DNMT3A and DNMT3B genes were initially identified by searching an expressed sequence tag (EST) database using bacterial type II cytosine-5 methyltransferase sequences as queries. Their protein products have a similar structural organization, including a C-terminal catalytic domain that contains sequence motifs characteristic of all DNA methyltransferases (prokaryotic and eukaryotic) and an N-terminal regulatory domain that is distinct from that of DNMT1 (Fig. 2.2) [36]. The N-terminal regions of DNMT3A/3B contain a variable region (not conserved between DNMT3A and DNMT3B), followed by two conserved domains implicated in chromatin binding. The PWWP domain, an ∼150-residue domain with a conserved proline-tryptophan-tryptophan-proline (PWWP) motif, is necessary for heterochromatin targeting and mediates binding to histone H3 trimethylated at lysine 36 (H3K36me3) mark [37–39]. The ADD (ATRX-DNMT3-DNMT3L) domain, composed of two C4-type zinc fingers (GATA binding protein 1 (GATA1) and PHD-type), interacts with the N-terminal tail of histone H3 with unmodified lysine 4 (H3K4me0) [40,41]. Recent structural studies revealed that the ADD domain of DNMT3A interacts with its catalytic domain and blocks its DNA-binding affinity, resulting in autoinhibition. Unmodified histone H3 tail (but not H3 tail with H3K4me3) can disrupt the interaction between the ADD and the catalytic domains, leading to DNMT3A activation [42]. These findings indicate that DNMT3A and DNMT3B act as both “writers” and “readers” of epigenetic marks and that their activities and specificities are regulated by specific histone modifications. The conclusion that DNMT3A and DNMT3B function primarily as de novo methyltransferases is based on several lines of evidence. First, DNMT3A and DNMT3B expression correlates with de novo methylation during development. Specifically, DNMT3A and DNMT3B are highly expressed in early embryos (as well as ESCs) and developing germ cells, and their expression is significantly downregulated in somatic cells and when ESCs are differentiated [36]. In ESCs, DNMT3A transcription is mostly driven by an internal promoter, resulting in a shorter isoform known as DNMT3A2, which lacks the N-terminal 219 (mouse) or 223 (human) amino acids of full-length DNMT3A1 (Fig. 2.2). DNMT3A2 expression decreases with ESC differentiation and is replaced by the DNMT3A1 isoform, which is ubiquitously expressed at low levels in most somatic tissues [43]. DNMT3B

15

produces multiple alternatively spliced isoforms (∼30 reported to date), many of which lack catalytic activity but may play regulatory roles in DNA methylation [44]. In ESCs, the full-length isoform DNMT3B1, a catalytically active form, is the predominant product, and other isoforms, including the inactive DNMT3B6, are also expressed. In most somatic cells, DNMT3B expression is low, usually with both active (e.g., DNMT3B2) and inactive (e.g., DNMT3B3) isoforms [36,43]. Second, biochemical studies indicate that DNMT3A and DNMT3B behave as de novo methyltransferases do. In particular, recombinant DNMT3A and DNMT3B proteins show no preference for hemimethylated DNA over unmethylated DNA in vitro, unlike DNMT1, which preferentially methylates hemimethylated DNA [20,36]. Furthermore, in vitro and in vivo target analyses indicate that DNMT3A and DNMT3B could methylate cytosines at non-CpG sites, such as CpA and CpT, albeit with lower efficiency compared to CpG methylation [45]. Non-CpG methylation, which cannot be maintained during DNA replication, is mediated by de novo methyltransferases. Third, genetic studies provide definitive evidence for the involvement of DNMT3A and DNMT3B in de novo DNA methylation. Targeted disruption of both DNMT3A and DNMT3B blocks de novo methylation in mouse ESCs and early embryos, however, has no effect on the maintenance of methylation at imprinted loci [25]. Moreover, DNMT3A and DNMT3B cause de novo methylation when overexpressed in mammalian cells or ectopically expressed in transgenic flies or budding yeast [26,46–48]. While DNMT3A and DNMT3B methylate many genomic loci redundantly, they also have preferred and distinct targets. For example, DNMT3A is more efficient than DNMT3B in methylating major satellite repeats at pericentric heterochromatin, whereas DNMT3B preferentially methylates minor satellite repeats in centromeric regions [26]. Characterization of DNMT3A and DNMT3B KO mice suggests that these enzymes play distinct roles in developmental processes. DNMT3A KO mice develop to term and show no overt defects at birth, but die at ∼4 weeks of age. In contrast, disruption of DNMT3B leads to embryonic lethality at ∼E12.5, with multiple developmental defects. DNMT3A/3B DKO embryos exhibit more severe defects and die earlier (before E11.5) than DNMT3B KO embryos [25]. DNA methylation analysis of E9.5 embryos indicate that DNMT3B is largely responsible for methylation of germline-specific genes, pluripotency genes, and many developmental genes and DNMT3A and DNMT3B redundantly methylate some specific genes, such as Brdt, Dpep3, Cytip, and Crygd [49]. Conditional gene KO studies indicate that DNMT3A, but not DNMT3B, is essential for de novo methylation during gametogenesis, including the establishment of DNA methylation imprints [50]. Indeed, immunofluorescence

II.  Molecular Mechanisms of Epigenetics

16

2.  Mechanisms of DNA Methylation and Demethylation During Mammalian Development

experiments show abundant expression of DNMT3A, but not DNMT3B, in fully grown oocytes [51]. Consistent with their important roles in developmental processes in mice, DNMT3A and DNMT3B mutations are associated with human diseases [31]. Somatic DNMT3A mutations occur frequently in acute myeloid leukemia (AML) and other hematologic malignancies [31]. Many DNMT3A mutations have been identified, with the majority (>50%) of cases affecting Arg882 (R882) in the catalytic domain. Although almost all reported DNMT3A mutations in leukemia occur in only one allele, there is evidence that DNMT3A R882 mutant proteins have dominant-negative effects by interacting with wild-type DNMT3A to form functionally deficient complexes [52]. DNMT3A mutations are also associated with an overgrowth disorder [31]. Hypomorphic DNMT3B mutations account for ∼50% of cases with ICF (Immunodeficiency, Centromeric instability, and Facial anomalies) syndrome, a rare recessive autosomal disorder characterized by hypomethylation of special genomic regions (most notably classical satellites 2 and 3 repeats) and chromosomal defects in lymphocytes and antibody deficiency, as well as facial dysmorphism, failure to thrive, and mental retardation [31]. Several other ICF genes—ZBTB24 (zinc-fingerand BTB domain-containing 24), CDCA7 (cell division cycle associated 7), and HELLS (helicase, lymphoid-specific)—have also been identified [53,54]. HELLS (also known as LSH), a DNA helicase involved in chromatin remodeling, has been shown to regulate DNA methylation by affecting DNMT3B targeting to chromatin [55]. However, little is known about the biological functions of ZBTB24 and CDCA7 and, in particular, their links to DNA methylation. DNMT3L DNMT3L was originally identified by sequence database analysis. The protein product contains an ADD domain, but not a PWWP domain, in the N-terminal region. Its C-terminal region shares sequence homology with the catalytic domains of DNMT3A and DNMT3B, but lacks some sequence motifs essential for catalytic activity, including those required for AdoMet (methyl donor) binding (Fig. 2.2) [4]. Although DNMT3L has no methyltransferase activity, biochemical, and genetic evidence suggests that it is an important regulator of de novo methylation. DNMT3L interacts with DNMT3A and DNMT3B and significantly enhances their catalytic activity [56]. Crystallographic studies reveal that the C-terminal region of DNMT3L directly interacts with the catalytic domain of DNMT3A, and the DNMT3A/3L dimer further dimerizes through DNMT3A–DNMT3A interaction, forming a tetramer with two active sites [57]. Biochemical and structural data also indicate that the ADD domain of DNMT3L

binds the N-terminal tail of histone H3 with H3K4me0, suggesting that DNMT3L plays a role in determining the specificity of de novo methylation [40]. The expression pattern of DNMT3L during development is similar to that of DNMT3A and DNMT3B, with high expression in developing germ cells, early embryos, and ESCs and low expression in most somatic tissues [56]. DNMT3L KO mice are viable and grossly normal, suggesting that zygotic DNMT3L is not essential for development. However, both male and female KO mice fail to reproduce [56,58]. DNMT3L KO males are sterile, as they are unable to produce mature sperm (azoospermia). The spermatogenesis defect is due to loss of DNA methylation in germ cells, which results in reactivation of retrotransposons, inducing genomic instability, meiotic catastrophe, and ultimately apoptosis [59]. On the other hand, DNMT3L KO females are able to conceive, but embryos die by midgestation due to failure to establish DNA methylation imprints in oocytes [56,58]. These studies suggest that DNMT3L is a critical accessory factor for DNMT3A for de novo methylation during gametogenesis. Indeed, the phenotype of DNMT3L KO mice is almost identical to that of mice with conditional DNMT3A deletion in germ cells [50].

DNA DEMETHYLATION DNA methylation is generally stable in somatic cells. However, two waves of global demethylation take place during development. The first wave occurs in preimplantation embryos, which results in genome-wide erasure of DNA methylation marks inherited from both the paternal and maternal gametes, with the exception of methylation in special regions, such as ICRs and some retrotransposons (e.g., intracisternal-A particles) [2,6]. Reprogramming of the parental genomes in early embryos is important for the establishment of totipotency. The paternal and maternal genomes undergo demethylation through distinct mechanisms. Shortly after fertilization, the male pronucleus rapidly loses its 5mC signal before the onset of DNA replication, suggesting DNA replication-independent “active” demethylation. In contrast, the maternal genome is gradually demethylated during cleavage divisions, mainly through “passive” dilution of 5mC due to insufficient maintenance methylation [2,6] (Fig. 2.1B). The second wave of global demethylation occurs in PGCs, which is a critical reprogramming event that generates an epigenome for the development of germ cells. DNA methylation imprints at ICRs, which are protected in preimplantation embryos, are erased in PGCs. In mice, PGCs are specified around E7.25 in the epiblast of the developing embryo and then migrate along the embryonic–extraembryonic interface, eventually arriving

II.  Molecular Mechanisms of Epigenetics



DNA Demethylation

at the genital ridge. Demethylation takes place during PGC migration. Genome-wide methylation profiling of PGCs at different time points indicates that demethylation occurs in two phases. The first phase, beginning at ∼E8.5, results in genome-wide hypomethylation affecting almost all genomic regions, mainly through passive demethylation. The second phase, occurring from E9.5 to E13.5, affects special loci, including ICRs, germlinespecific genes, and CGIs on the X chromosome, and involves both active and passive mechanisms [6].

Passive DNA Demethylation DNA methylation patterns are maintained during DNA replication by the maintenance methylation machinery that “copies” the methylation pattern from the parental strand onto the daughter strand. Defects in maintenance methylation can therefore lead to replication-dependent dilution of 5mC, known as passive DNA demethylation. Passive demethylation is biologically important for the erasure of DNA methylation marks in both preimplantation embryos and PGCs. It is generally believed that deficiency in DNMT1 function is the major mechanism for passive demethylation of the maternal genome during early embryogenesis. In preimplantation embryos, the oocyte-specific DNMT1o variant, transmitted from oocytes, is the predominant DNMT1 isoform, although zygotic DNMT1 is also expressed at a low level [51]. In contrast to nuclear localization of DNMT1 in somatic cells, DNMT1o is localized in the cytoplasm of oocytes and preimplantation embryos [60]. The presence or absence of the N-terminal 118 amino acids in different DNMT1 isoforms does not account for the difference in subcellular localization, as DNMT1o has several functional NLSs and, when ectopically expressed in somatic cells, localizes in the nuclei [60]. The mechanism of DNMT1o cytoplasmic retention in oocytes and preimplantation embryos is still unknown. It is worth mentioning that demethylation of the paternal genome, which is initiated by an active process involving TET-mediated 5mC oxidation, also involves passive demethylation, as the oxidative products (i.e., 5hmC, 5fC, and 5caC) become diluted in a replicationdependent manner [61,62]. Despite extensive demethylation during preimplantation development, parental allele-specific DNA methylation in ICRs is faithfully maintained. Genetic studies indicate that DNMT1, but not DNMT3A and DNMT3B, is responsible for maintaining methylation imprints [51,63]. Early studies detected transient nuclear localization of DNMT1o at 8-cell stage embryos [60,63]. However, subsequent work using different DNMT1 antibodies failed to confirm the observation [51]. Nevertheless, the fact that imprinted loci are protected from this wave of demethylation suggests that DNMT1 is not exclusively

17

retained in the cytoplasm during preimplantation development. While it is poorly understood what confers the specificity of DNMT1, such that methylation is maintained at imprinted genes, but not at other sequences, genetic and epigenetic features may distinguish imprinted loci from other regions. In addition to DNMT1, several genes have been identified that are involved in the maintenance of methylation imprints [64]. For example, the Krüppel associated box (KRAB)-containing zincfinger protein ZFP57 is essential for the maintenance of genomic imprints in mouse [65], and human Zfp57 mutations are associated with hypomethylation at multiple imprinted loci in patients with transient neonatal diabetes [66]. Molecular experiments show that ZFP57 specifically binds the methylated allele of ICRs, recognizing a hexanucleotide sequence (TGCCGC) shared by ICRs, and interacts with KRAB-associated protein 1 (KAP1), a scaffold protein for multiple proteins including DNMTs and UHRF1 [67,68]. Thus, it is likely that nuclear DNMT1 protein in preimplantation embryos, albeit at low levels, is concentrated in special genomic regions, such as ICRs. In PGCs, passive demethylation is mainly responsible for the first phase of demethylation that erases methylation marks genome-wide. There is evidence that migrating PGCs, which are rapidly cycling, have little de novo or maintenance DNA methylation potential. Specifically, immunofluorescence analysis shows that DNMT3A, DNMT3B, and UHRF1 are repressed in PGCs [69]. The second phase of demethylation in PGCs, which affects special genomic regions including ICRs, involves TETmediated 5mC oxidation, followed by DNA replicationcoupled dilution of oxidized derivatives [6].

Active DNA Demethylation In contrast to DNA replication-dependent passive dilution of 5mC, active DNA demethylation refers to enzyme-mediated removal or modifications of 5mC with the regeneration of unmodified cytosine, which is generally independent of DNA replication. It has been known for a long time that active demethylation takes place during development. However, the mechanisms involved had been elusive until recently, when TET proteins were shown to oxidize 5mC into 5hmC, 5fC, and 5caC [8,9]. Studies in the last several years have demonstrated that these oxidized derivatives serve as intermediates for demethylation, as they can be recognized by specific DNA repair enzymes, eventually leading to their removal and replacement by unmodified cytosines [6]. TET Proteins TET1, the founding member of the TET family, was named for its involvement in the ten-eleven translocation [t(10;11)(q22;q23)] in rare cases of leukemia, which

II.  Molecular Mechanisms of Epigenetics

18

2.  Mechanisms of DNA Methylation and Demethylation During Mammalian Development

FIGURE 2.3  Schematic diagram of TET family of proteins. Domain architecture of the three mouse TET proteins (TET1, TET2, and TET3), and IDAX. All three TET proteins share a similar C-terminal catalytic domain with two characteristic features, a cysteine-rich region (Cys) and a double-stranded beta-helix (DSBH) fold. Their N-terminal regions are less conserved, with TET1 and TET3 containing a CXXC zinc finger domain. The start site of a shorter TET3 isoform is indicated.

results in the fusion of the TET1 gene, located on 10q22, with the mixed-lineage leukemia (MLL) gene, located on 11q23 [70]. TET2 and TET3 were identified on the basis of sequence homology. The TET family belongs to the 2-oxoglutarate (2OG)- and Fe(II)-dependent dioxygenase (2OGFeDO) superfamily [8]. All three TET proteins are capable of oxidizing 5mC into 5hmC, 5fC, and 5caC [6]. They share a similar structural organization, with a C-terminal catalytic domain and an N-terminal regulatory domain (Fig. 2.3). The catalytic domain consists of a cysteine-rich region and a double-stranded β-helix (DSBH) domain characteristic of the 2OGFeDO superfamily. Structural studies reveal that the catalytic core region specifically recognizes CpG dinucleotides, showing preference for 5mC as a substrate and also binding to 5mC derivatives for subsequent oxidation [71,72]. The Nterminal regions of TET proteins are less conserved. TET1 and TET3 contain a CXXC zinc finger domain. The CXXC domain of TET1 binds unmodified, 5mC-modified, and 5hmC-modified CpG-rich DNA, whereas the CXXC domain of TET3 binds unmodified cytosine regardless of whether it is in the CpG context [73,74]. It appears that TET2 has lost its CXXC domain during evolution, as a result of a chromosomal inversion event that has converted a portion of the ancestral TET2 gene into a separate gene, IDAX, which encodes a protein containing the original CXXC domain (Fig. 2.3). IDAX binds unmethylated CpG sequences via its CXXC domain, localizes at CGIs, and physically interacts with TET2, suggesting that it may play a role in recruiting TET2 to its genomic targets [75]. TET-Dependent Demethylation Pathways Since direct removal of the methyl group from 5mC by breaking the carbon─carbon bond is a thermodynamically unfavorable reaction, DNA demethylation is most likely achieved by a step-wise process involving intermediates. The finding that TET proteins convert 5mC

to 5hmC immediately raised the possibility that 5mC oxidation could be involved in DNA demethylation [8,9]. This notion was strengthened by subsequent work showing that 5hmC can be further oxidized by TET proteins to produce 5fC and 5caC, which can be recognized and excised from DNA by thymine glycosylase (TDG) [10,11]. Several mechanisms of DNA demethylation involving TET-mediated 5mC oxidation have been proposed (Fig. 2.4). First, as mentioned previously, 5hmC, 5fC, and 5caC can be passively diluted due to the lack of maintenance during DNA replication. Indeed, in vitro assays indicate that DNMT1 is much less efficient in utilizing hemihydroxymethylated CpG substrates compared to hemimethylated CpG substrates [76]. Second, excision of 5fC and 5caC by TDG generates abasic sites, which can be repaired by the base excision repair (BER) pathway to restore unmodified cytosines. In support of this mechanism, coexpression of TDG and TET proteins leads to depletion of 5fC and 5caC in HEK293 cells, whereas TDG deficiency results in substantial increases of 5fC and 5caC in mouse ESCs, as well as defects in demethylation and embryonic development in mice [11,77–79]. Third, it has been reported that 5hmC can be deaminated by the AID/APOBEC family of deaminases to produce 5-hydroxymethyluracil (5hmU), which can be excised by TDG and SMUG1, another DNA glycosylase, and replaced by cytosine through BER [77,80]. However, biochemical studies suggest that AID/APOBEC deaminases have no detectable deamination activity on 5hmC [81]. Deamination of 5mC to thymine (T) by AID/APOBEC enzymes, followed by replacement of the T:G mismatch via BER, has also been implicated in DNA demethylation, although the physiological relevance of this mechanism remains to be determined [6]. Finally, 5hmC, 5fC, and 5caC could be directly removed by DNA dehydroxymethylases, deformylases, and decarboxylases, respectively. One study showed that DNMT3A and

II.  Molecular Mechanisms of Epigenetics



DNA Demethylation

19

FIGURE 2.4  Proposed DNA demethylation pathways involving TET proteins. TET proteins initiate DNA demethylation by oxidizing 5-methylcytosine (5mC) to generate 5-hydroxymethylcytosine (5hmC), which can be further oxidized to 5-formylcytosine (5fC) and 5-carboxylcytosine (5caC). 5mC can also be deaminated by the AID/APOBEC family of deaminases to generate thymine (T). In vitro evidence suggests that 5hmC can also be deaminated by AID/APOBEC to generate 5hmU, however, the efficiency is low compared to other substrates and this pathway needs to be further confirmed. The 5mC derivatives 5fC, 5caC, T, and 5hmU can then be removed by DNA glycosylases (TDG and SMUG1) and replaced by unmodified cytosine (C) by the Base Excision Repair (BER) pathway. Alternatively, because 5hmC, 5fC, and 5caC are poor substrates for maintenance methylation by DNMT1, their removal can be achieved by DNA replication-dependent passive dilution. Theoretically, 5hmC, 5fC, and 5caC could be directly converted to unmodified cytosine by dehydroxymethylases, deformylases, and decarboxylases, respectively. The existence and identities of these enzymes remain to be determined, although in vitro experiments indicate that DNMT3A and DNMT3B could function as dehydroxymethylases under special experimental conditions. Black solid lines represent pathways with relatively strong experimental evidence, and dashed lines represent pathways that need to be further confirmed.

DNMT3B are able to convert 5hmC directly to C in vitro under oxidized redox conditions [82]. Another study showed that mouse ESC nuclear extract can convert 5caC to C [83]. However, these observations need to be confirmed. TET Proteins in Development Although all three TET proteins have similar activity, they show different expression patterns. TET1 and TET2 are expressed in migrating PGCs, whereas TET3 is highly expressed in oocytes and zygotes (fertilized eggs) [84–88]. Consistent with the expression patterns, genetic studies have demonstrated that TET1, TET2, and TET3

have distinct roles in DNA demethylation during development. In the zygote, the rapid loss of 5mC signal in the male pronucleus is accompanied with dramatic increases in 5hmC, 5fC, and 5caC, suggesting TET-mediated 5mC oxidation [62,84,85]. Recent methylome analysis reveals that 5mC oxidation also occurs in the female pronucleus, albeit to a lesser extent [89,90]. Since zygotic gene expression does not occur until the 2-cell stage, the TET protein(s) responsible for 5mC oxidation in the zygote must come from the oocyte. Indeed, depletion of maternal TET3, the predominant TET protein in oocytes, blocks 5mC oxidation in both the male and female pronuclei of

II.  Molecular Mechanisms of Epigenetics

20

2.  Mechanisms of DNA Methylation and Demethylation During Mammalian Development

zygote and inhibits demethylation during preimplantation development [84,86,89]. These results suggest that TET3-mediated 5mC oxidation is a key step that initiates demethylation in preimplantation embryos. Subsequent conversion of 5mC oxidized derivatives to unmodified cytosines likely involves both active mechanisms and passive dilution, although their relative significance is somewhat controversial [61,62,89,90]. While it is not fully understood why the efficiencies of 5mC oxidation differ significantly in the male and female pronuclei, even though they are exposed to the same environment, distinct epigenetic states may be responsible. For example, histone H3 dimethylated at lysine 9 (H3K9me2) is abundant in the female pronucleus but absent in the male pronucleus. There is evidence that the maternal factor PGC7 (also known as Stella and DPPA3), an H3K9me2binding protein, inhibits TET3 binding to the maternal chromatin [91]. TET-mediated 5mC oxidation is also involved in demethylation in PGCs. TET1 and TET2 are expressed between E9.25 and E11.5, whereas TET3 is undetectable [87,88]. The timing of TET1 and TET2 expression, which correlates with increases in 5hmC [88,92], corresponds to the second phase of demethylation that affects special regions, such as ICRs and germline-specific genes. Indeed, genetic ablation of TET1 or both TET1 and TET2 has no effect on global demethylation in PGCs (first phase), but leads to defective demethylation and altered expression of specific genes including imprinted genes and meiosis genes [87,93,94]. In addition to preimplantation embryos and PGCs, TET proteins are dynamically expressed during development and are present in multiple adult tissues. For example, TET1 and TET2, but not TET3, are highly expressed in mouse ESCs, and differentiation of ESCs leads to downregulation of TET1 and TET2 and upregulation of TET3 [6]. In addition to ESCs, neuronal cells have high levels of 5hmC [95]. Given that postmitotic neurons do not replicate, it is believed that demethylation in response to stimuli is achieved through active mechanisms. TET enzymes may play important roles in the regulation of DNA methylation and gene expression in neuronal cells. Genetic studies in KO mice suggest that TET proteins have both distinct and redundant functions. TET1-null mice are viable (although some are slightly smaller at birth) but show reduced fertility and impaired adult neurogenesis [87,96,97]. TET2-null mice are viable and fertile, however, adult animals develop hematologic phenotypes, characterized by inhibition of hematopoietic stem cell differentiation and myeloid malignancies [98–101]. This phenotype is consistent with the observation that TET2 is frequently mutated in various myeloid malignancies [102]. Zygotic deletion of TET3 leads to neonatal lethality, and maternal deletion impairs demethylation in preimplantation embryos and

high frequencies of embryonic defects, with only ∼20% surviving to term [86]. TET1/2 DKO mice show variable phenotypes: most of them exhibit embryonic defects and die perinatally, and small numbers are viable [94]. A recent report shows that TET1/2/3 TKO mice exhibit gastrulation defects, partly due to increased Lefty–Nodal signaling [103]. Consistent with the developmental phenotypes, TKO ESCs show differentiation defects and contribute poorly to chimeric embryos [104]. TKO ESCs also show telomere changes, however, the results reported by different research groups are inconsistent and need to be clarified [105,106]. It is worth mentioning that, most likely, not all developmental phenotypes of TET KO mice are attributable to defects in DNA demethylation. TET proteins, which are large molecules, also have noncatalytic functions, including regulating histone modifications and forming chromatin regulatory complexes [107–109].

CONCLUSIONS It is generally believed that DNA methylation functions cooperatively and collaboratively with other epigenetic mechanisms, such as posttranslational modifications of histones, to stably maintain gene silencing and chromatin structure. Genetic manipulations of key components of the DNA methylation and demethylation machineries, including DNMTs and TETs, in mice have greatly facilitated our understanding of the roles of DNA methylation and demethylation in developmental processes in mammals. One area that is still poorly understood is how DNA methylation patterns are specified during development and how DNA methylation is dysregulated in human diseases. Studies in recent years have identified mutations in major components of DNA methylation “writers” (DNMTs), “erasers” (e.g., TETs), and “readers” (e.g., MeCP2) in various diseases [31]. In general, the mechanisms by which these mutations contribute to pathogenesis remain to be elucidated. The discovery of TET proteins as 5mC dioxygenases was a breakthrough in the field, and ever since, great progress has been made in understanding the mechanisms involved in DNA demethylation. It is well established that TET-mediated 5mC oxidation is a key initiating event of DNA demethylation during development and the oxidized derivatives 5hmC, 5fC, and 5caC serve as intermediates. Less clear are the relevance and significance of the multiple pathways that have been proposed to eventually achieve demethylation (regeneration of unmodified cytosine). Moreover, there is evidence that 5mC oxidized derivatives, in addition to being intermediates of DNA demethylation, act as epigenetic marks for chromatin regulation. Their relevance in biological processes is largely unknown. In the coming years, we are expected to see major advances in these areas.

II.  Molecular Mechanisms of Epigenetics



Conclusions

Glossary Base excision repair (BER)  DNA repair pathway in which a DNA base is removed by a glycosylase enzyme and ultimately replaced by a new base. Blastocyst  Early stage embryos that have undergone the first cell lineage specification, which results in two primary cell types: the inner cell mass (ICM) and the trophoblasts, which will differentiate into embryonic and extra-embryonic tissues, respectively. Centromeric chromatin  Type of heterochromatin that is a constituent in the formation of centromeres, which is the region of the chromosome that links sister chromatids to the spindle during cell division. It is flanked by pericentric heterochromatin. CpG islands (CGIs)  CG-rich genomic regions, ranging from several hundred bp to 2 kb, often located at gene promoters. CpG sequences  Cytosine followed by a guanine in the 5′–3′ directions on the same strand of DNA linked by phosphate. It is the major substrate for DNA methylation in mammals. Cytosine methylation  The covalent addition of a methyl group to the fifth position of cytosine (5-methylcytosine, 5mC), mostly occurring within a CpG dinucleotide sequence, is the most common form of DNA modification, which plays important roles in the regulation of chromatin structure and gene expression. Cytosine hydroxymethylation  DNA modification generated by oxidation of methylcytosine. This modification is mediated by the TET family proteins and is considered both stable and an intermediate in DNA demethylation. DNA demethylation  Replacement of the methylcytosine with an unmodified cytosine, either through direct or indirect mechanisms. Dioxygenases  Enzymes that catalyze the addition of the two oxygen atoms from its molecular free form (O2) to one or two organic substrates. Embryonic stem cells  Stem cells derived from the undifferentiated inner cell mass from a blastocyst. Epiblast  Outer layer arising from the inner cell mass in the blastocyst, capable of differentiating into the three primary germ layers (ectoderm, endoderm, and mesoderm) and into the extraembryonic mesoderm of the visceral yolk sac, the allantois, and the amnion. Hemimethylated sites  DNA sequences (mostly CpGs) that are methylated only on one of the two DNA strands. Genomic imprinting  Epigenetic phenomenon by which certain genes are expressed in a specific parent-of-origin fashion. Germline-specific genes  Genes of unique gametogenic function that are tightly regulated outside gametes or the early embryo by DNA methylation. They are specifically demethylated during PGC specification. Heterochromatin  Tightly packed chromatin that is associated with structural functions and gene silencing. Imprinted locus (imprinted gene)  Genomic region with a methylation mark that is present only on the maternally or paternally derived copy of an allele. Imprinting control regions (ICRs)  Genomic region that acts, in a methylation-sensitive manner, to determine whether imprinted genes are expressed or not according to the parent of origin. Imprints  Allele-specific DNA methylation marks at ICRs. Inactive X chromosome  In females, one of the two X chromosomes is inactivated to prevent overexpression of X-linked gene products, in order to equalize gene dose between females (with two X chromosomes) and males (with one X chromosome). X chromosome inactivation is achieved through formation of heterochromatin, and the highly condensed inactive X chromosome is called a Barr body. Intracisternal A-type particles (IAPs)  Mouse specific class II long terminal repeat (LTR)-containing retroelements that retain high DNA methylation levels throughout development. Implantation  Early developmental stage where the embryo (blastocyst) adheres to the wall of the uterus.

21

Major satellite  Tandem repeating DNA sequences of 234 bp that are primarily located in the pericentromeric regions of the mouse genome. Minor satellite  Tandem repeating AT-rich DNA sequences of 120 bp that are located in the centromeric regions of the mouse genome. Pericentric heterochromatin  Major repressive heterochromatin domain in the nucleus, which is mainly composed of major satellite repeats. Pluripotency  The potential ability of a cell to differentiate into all the three embryonic germ layers (endoderm, mesoderm and ectoderm) and to give rise to any embryonic tissue and any adult cell type. ESCs are pluripotent. Preimplantation embryos  Early embryos from 2-cell to morula stage or early blastocyst before the onset of implantation. PGCs  Cells that give rise to both mature germ cells (oocytes and spermatozoa). Repetitive element  Sequence that is found in multiple copies in the genome, such as transposable elements (transposons), as well as centromeric and telomeric satellite repeats. Retroelements (retrotransposons)  Endogenous transposable elements that move along the genome via the transcription of an RNA intermediate. Totipotency Potential ability of a cell to give rise to all cell types in embryonic and extra-embryonic tissues. For example, zygote is a totipotent cell. Zygote  Cell formed by the fertilization of the oocyte with a sperm cell.

Abbreviations 5fC 5-Formylcytosine 5caC 5-Carboxylcytosine 5hmC 5-Hydroxymethylcytosine 5mC 5-Methylcytosine ADCA-DN Autosomal dominant cerebellar ataxia, deafness and narcolepsy ADD ATRX-DNMT3-DNMT3L AID Activation-induced cytidine deaminase AML Acute myeloid leukemia APOBEC Apolipoprotein B mRNA editing enzyme, catalytic polypeptide-like BER Base excision repair CGI CpG island CpG 5′-Cytosine nucleotide-phosphate-guanine nucleotide-3’ CXXC Cysteine-rich motif DNA Deoxyribonucleic acid DNMT DNA methyltransferase ESC Embryonic stem cell EST Expressed sequence tag H3K4me0 Unmodified histone H3 at lysine 4 H3K4me3 Trimethylation of histone H3 at lysine 4 H3K9me2 Dimethylation of histone H3 at lysine 9 H3K9me3 Trimethylation of histone H3 at lysine 9 H3R2me0 Unmodified histone H3 at arginine 2 H3K36me3 Trimethylation of histone H3 at lysine 36 H3K18ub Ubiquitinated histone H3 at lysine 18 H3K23ub Ubiquitinated histone H3 at lysine 23 HSAN IE Hereditary sensory and autonomic neuropathy with dementia and hearing loss type IE ICF Immunodeficiency, centromeric instability, and facial anomalies ICR Imprinting control region KO Knockout MeCP2 Methyl-CpG-binding protein 2 MLL Mixed-lineage leukemia PCNA Proliferating cell nuclear antigen PGC Primordial germ cell

II.  Molecular Mechanisms of Epigenetics

22

2.  Mechanisms of DNA Methylation and Demethylation During Mammalian Development

PHD Plant homeodomain PWWP Proline-tryptophan-tryptophan-proline RFTS Replication foci-targeting sequence RING Really Interesting New Gene SAM S-adenosyl-l-methionine SRA SET and RING associated TET Ten-eleven translocation TDG Thymine DNA glycosylase TTD Tandem tudor domain UHRF1 Ubiquitin-like with PHD and RING finger domains 1 UIM Ubiquitin-interacting motif

Acknowledgments Work in our laboratory is supported by grants from the National Institutes of Health (NIH, 1R01DK106418-01 and 1R01AI12140301A1 to T.C.) and a Rising Star Award (R1108 to T.C.) from Cancer Prevention and Research Institute of Texas (CPRIT). N.V. is a recipient of scholarships from CPRIT ​(RP140106) and the Center for Cancer Epigenetics (CCE) at MD Anderson Cancer Center. T.C. is a CPRIT Scholar in Cancer Research.

References [1] Patil V, Ward RL, Hesson LB. The evidence for functional non-CpG methylation in mammalian cells. Epigenetics 2014;9(6):823–8. [2] Smith ZD, Meissner A. DNA methylation: roles in mammalian development. Nat Rev Genet 2013;14(3):204–20. [3] Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet 2012;13(7):484–92. [4] Chen T, Li E. Structure and function of eukaryotic DNA methyltransferases. Curr Top Dev Biol 2004;60:55–89. [5] Goll MG, Kirpekar F, Maggert KA, Yoder JA, Hsieh CL, Zhang X, et al. Methylation of tRNA(Asp) by the DNA methyltransferase homolog Dnmt2. Science 2006;311(5759):395–8. [6] Zhao H, Chen T. Tet family of 5-methylcytosine dioxygenases in mammalian development. J Hum Genet 2013;58(7):421–7. [7] Wu H, Zhang Y. Reversing DNA methylation: mechanisms, genomics, and biological functions. Cell 2014;156(1–2):45–68. [8] Tahiliani M, Koh KP, Shen Y, Pastor WA, Bandukwala H, Brudno Y, et al. Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 2009;324(5929):930–5. [9] Ito S, D’Alessio AC, Taranova OV, Hong K, Sowers LC, Zhang Y. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature 2010;466(7310):1129–33. [10] Ito S, Shen L, Dai Q, Wu SC, Collins LB, Swenberg JA, et al. Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine. Science 2011;333(6047):1300–3. [11] He YF, Li BZ, Li Z, Liu P, Wang Y, Tang Q, et al. Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA. Science 2011;333(6047):1303–7. [12] Holliday R, Pugh JE. DNA modification mechanisms and gene activity during development. Science 1975;187(4173):226–32. [13] Riggs AD. X inactivation, differentiation, and DNA methylation. Cytogenet Cell Genet 1975;14(1):9–25. [14] Bostick M, Kim JK, Esteve PO, Clark A, Pradhan S, Jacobsen SE. UHRF1 plays a role in maintaining DNA methylation in mammalian cells. Science 2007;317(5845):1760–4. [15] Sharif J, Muto M, Takebayashi S, Suetake I, Iwamatsu A, Endo TA, et al. The SRA protein Np95 mediates epigenetic inheritance by recruiting Dnmt1 to methylated DNA. Nature 2007;450(7171):908–12.

[16] Bestor T, Laudano A, Mattaliano R, Ingram V. Cloning and sequencing of a cDNA encoding DNA methyltransferase of mouse cells. The carboxyl-terminal domain of the mammalian enzymes is related to bacterial restriction methyltransferases. J Mol Biol 1988;203(4):971–83. [17] Ding F, Chaillet JR. In vivo stabilization of the Dnmt1 (cytosine-5)-methyltransferase protein. Proc Natl Acad Sci USA 2002;99(23):14861–6. [18] Qin W, Wolf P, Liu N, Link S, Smets M, La Mastra F, et al. DNA methylation requires a DNMT1 ubiquitin interacting motif (UIM) and histone ubiquitination. Cell Res 2015;25(8):911–29. [19] Takeshita K, Suetake I, Yamashita E, Suga M, Narita H, Nakagawa A, et al. Structural insight into maintenance methylation by mouse DNA methyltransferase 1 (Dnmt1). Proc Natl Acad Sci USA 2011;108(22):9055–9. [20] Pradhan S, Bacolla A, Wells RD, Roberts RJ. Recombinant human DNA (cytosine-5) methyltransferase. I. Expression, purification, and comparison of de novo and maintenance methylation. J Biol Chem 1999;274(46):33002–10. [21] Leonhardt H, Page AW, Weier HU, Bestor TH. A targeting sequence directs DNA methyltransferase to sites of DNA replication in mammalian nuclei. Cell 1992;71(5):865–73. [22] Li E, Bestor TH, Jaenisch R. Targeted mutation of the DNA methyltransferase gene results in embryonic lethality. Cell 1992;69(6):915–26. [23] Lei H, Oh SP, Okano M, Juttermann R, Goss KA, Jaenisch R, et al. De novo DNA cytosine methyltransferase activities in mouse embryonic stem cells. Development 1996;122(10):3195–205. [24] Tucker KL, Talbot D, Lee MA, Leonhardt H, Jaenisch R. Complementation of methylation deficiency in embryonic stem cells by a DNA methyltransferase minigene. Proc Natl Acad Sci USA 1996;93(23):12920–5. [25] Okano M, Bell DW, Haber DA, Li E. DNA methyltransferases Dnmt3a and Dnmt3b are essential for de novo methylation and mammalian development. Cell 1999;99(3):247–57. [26] Chen T, Ueda Y, Dodge JE, Wang Z, Li E. Establishment and maintenance of genomic methylation patterns in mouse embryonic stem cells by Dnmt3a and Dnmt3b. Mol Cell Biol 2003;23(16):5594–605. [27] Tsumura A, Hayakawa T, Kumaki Y, Takebayashi S, Sakaue M, Matsuoka C, et al. Maintenance of self-renewal ability of mouse embryonic stem cells in the absence of DNA methyltransferases Dnmt1, Dnmt3a and Dnmt3b. Genes Cells 2006;11(7):805–14. [28] Liao J, Karnik R, Gu H, Ziller MJ, Clement K, Tsankov AM, et al. Targeted disruption of DNMT1, DNMT3A and DNMT3B in human embryonic stem cells. Nat Genet 2015;47(5):469–78. [29] Jackson-Grusby L, Beard C, Possemato R, Tudor M, Fambrough D, Csankovszki G, et al. Loss of genomic methylation causes p53-dependent apoptosis and epigenetic deregulation. Nat Genet 2001;27(1):31–9. [30] Chen T, Hevi S, Gay F, Tsujimoto N, He T, Zhang B, et al. Complete inactivation of DNMT1 leads to mitotic catastrophe in human cancer cells. Nat Genet 2007;39(3):391–6. [31] Hamidi T, Singh AK, Chen T. Genetic alterations of DNA methylation machinery in human diseases. Epigenomics 2015;7(2):247– 65. [32] Muto M, Kanari Y, Kubo E, Takabe T, Kurihara T, Fujimori A, et al. Targeted disruption of Np95 gene renders murine embryonic stem cells hypersensitive to DNA damaging agents and DNA replication blocks. J Biol Chem 2002;277(37):34549–55. [33] Rajakumara E, Wang Z, Ma H, Hu L, Chen H, Lin Y, et al. PHD finger recognition of unmodified histone H3R2 links UHRF1 to regulation of euchromatic gene expression. Mol Cell 2011;43(2):275–84. [34] Rothbart SB, Dickson BM, Ong MS, Krajewski K, Houliston S, Kireev DB, et al. Multivalent histone engagement by the linked

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45] [46] [47]

[48]

[49]

[50]

[51]

[52]

[53]

tandem Tudor and PHD domains of UHRF1 is required for the epigenetic inheritance of DNA methylation. Genes Dev 2013;27(11):1288–98. Nishiyama A, Yamaguchi L, Sharif J, Johmura Y, Kawamura T, Nakanishi K, et al. Uhrf1-dependent H3K23 ubiquitylation couples maintenance DNA methylation and replication. Nature 2013;502(7470):249–53. Okano M, Xie S, Li E. Cloning and characterization of a family of novel mammalian DNA (cytosine-5) methyltransferases. Nat Genet 1998;19(3):219–20. Chen T, Tsujimoto N, Li E. The PWWP domain of Dnmt3a and Dnmt3b is required for directing DNA methylation to the major satellite repeats at pericentric heterochromatin. Mol Cell Biol 2004;24(20):9048–58. Dhayalan A, Rajavelu A, Rathert P, Tamas R, Jurkowska RZ, Ragozin S, et al. The Dnmt3a PWWP domain reads histone 3 lysine 36 trimethylation and guides DNA methylation. J Biol Chem 2010;285(34):26114–20. Baubec T, Colombo DF, Wirbelauer C, Schmidt J, Burger L, Krebs AR, et al. Genomic profiling of DNA methyltransferases reveals a role for DNMT3B in genic methylation. Nature 2015;520(7546):243–7. Ooi SK, Qiu C, Bernstein E, Li K, Jia D, Yang Z, et al. DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature 2007;448(7154):714–7. Otani J, Nankumo T, Arita K, Inamoto S, Ariyoshi M, Shirakawa M. Structural basis for recognition of H3K4 methylation status by the DNA methyltransferase 3A ATRX-DNMT3-DNMT3L domain. EMBO Rep 2009;10(11):1235–41. Guo X, Wang L, Li J, Ding Z, Xiao J, Yin X, et al. Structural insight into autoinhibition and histone H3-induced activation of DNMT3A. Nature 2015;517(7536):640–4. Chen T, Ueda Y, Xie S, Li E. A novel Dnmt3a isoform produced from an alternative promoter localizes to euchromatin and its expression correlates with active de novo methylation. J Biol Chem 2002;277(41):38746–54. Duymich CE, Charlet J, Yang X, Jones PA, Liang G. DNMT3B isoforms without catalytic activity stimulate gene body methylation as accessory proteins in somatic cells. Nat Commun 2016;7:11453. Chédin F. The DNMT3 family of mammalian de novo DNA methyltransferases. Prog Mol Biol Transl Sci 2011;101:255–85. Hsieh CL. In vivo activity of murine de novo methyltransferases, Dnmt3a and Dnmt3b. Mol Cell Biol 1999;19(12):8211–8. Lyko F, Ramsahoye BH, Kashevsky H, Tudor M, Mastrangelo MA, Orr-Weaver TL, et al. Mammalian (cytosine-5) methyltransferases cause genomic DNA methylation and lethality in Drosophila. Nat Genet 1999;23(3):363–6. Hu JL, Zhou BO, Zhang RR, Zhang KL, Zhou JQ, Xu GL. The Nterminus of histone H3 is required for de novo DNA methylation in chromatin. Proc Natl Acad Sci USA 2009;106(52):22187–92. Borgel J, Guibert S, Li Y, Chiba H, Schubeler D, Sasaki H, et al. Targets and dynamics of promoter DNA methylation during early mouse development. Nat Genet 2010;42(12):1093–100. Kaneda M, Okano M, Hata K, Sado T, Tsujimoto N, Li E, et al. Essential role for de novo DNA methyltransferase Dnmt3a in paternal and maternal imprinting. Nature 2004;429(6994):900–3. Hirasawa R, Chiba H, Kaneda M, Tajima S, Li E, Jaenisch R, et al. Maternal and zygotic Dnmt1 are necessary and sufficient for the maintenance of DNA methylation imprints during preimplantation development. Genes Dev 2008;22(12):1607–16. Kim SJ, Zhao H, Hardikar S, Singh AK, Goodell MA, Chen T. A DNMT3A mutation common in AML exhibits dominant-negative effects in murine ES cells. Blood 2013;122(25):4086–9. de Greef JC, Wang J, Balog J, den Dunnen JT, Frants RR, Straasheijm KR, et al. Mutations in ZBTB24 are associated with im-

23

munodeficiency, centromeric instability, and facial anomalies syndrome type 2. Am J Hum Genet 2011;88(6):796–804. [54] Thijssen PE, Ito Y, Grillo G, Wang J, Velasco G, Nitta H, et al. Mutations in CDCA7 and HELLS cause immunodeficiency-centromeric instability-facial anomalies syndrome. Nat Commun 2015;6:7870. [55] Zhu H, Geiman TM, Xi S, Jiang Q, Schmidtmann A, Chen T, et al. Lsh is involved in de novo methylation of DNA. EMBO J 2006;25(2):335–45. [56] Hata K, Okano M, Lei H, Li E. Dnmt3L cooperates with the Dnmt3 family of de novo DNA methyltransferases to establish maternal imprints in mice. Development 2002;129(8): 1983–93. [57] Jia D, Jurkowska RZ, Zhang X, Jeltsch A, Cheng X. Structure of Dnmt3a bound to Dnmt3L suggests a model for de novo DNA methylation. Nature 2007;449(7159):248–51. [58] Bourc’his D, Xu GL, Lin CS, Bollman B, Bestor TH. Dnmt3L and the establishment of maternal genomic imprints. Science 2001;294(5551):2536–9. [59] Bourc’his D, Bestor TH. Meiotic catastrophe and retrotransposon reactivation in male germ cells lacking Dnmt3L. Nature 2004;431(7004):96–9. [60] Cardoso MC, Leonhardt H. DNA methyltransferase is actively retained in the cytoplasm during early development. J Cell Biol 1999;147(1):25–32. [61] Inoue A, Zhang Y. Replication-dependent loss of 5-hydroxymethylcytosine in mouse preimplantation embryos. Science 2011;334(6053):194. [62] Inoue A, Shen L, Dai Q, He C, Zhang Y. Generation and replication-dependent dilution of 5fC and 5caC during mouse preimplantation development. Cell Res 2011;21(12):1670–6. [63] Howell CY, Bestor TH, Ding F, Latham KE, Mertineit C, Trasler JM, et al. Genomic imprinting disrupted by a maternal effect mutation in the Dnmt1 gene. Cell 2001;104(6):829–38. [64] Ciccone DN, Chen T. Histone lysine methylation in genomic imprinting. Epigenetics 2009;4(4):216–20. [65] Li X, Ito M, Zhou F, Youngson N, Zuo X, Leder P, et al. A maternal-zygotic effect gene, Zfp57, maintains both maternal and paternal imprints. Dev Cell 2008;15(4):547–57. [66] Mackay DJ, Callaway JL, Marks SM, White HE, Acerini CL, Boonen SE, et al. Hypomethylation of multiple imprinted loci in individuals with transient neonatal diabetes is associated with mutations in ZFP57. Nat Genet 2008;40(8):949–51. [67] Quenneville S, Verde G, Corsinotti A, Kapopoulou A, Jakobsson J, Offner S, et al. In embryonic stem cells, ZFP57/KAP1 recognize a methylated hexanucleotide to affect chromatin and DNA methylation of imprinting control regions. Mol Cell 2011;44(3):361–72. [68] Zuo X, Sheng J, Lau HT, McDonald CM, Andrade M, Cullen DE, et al. Zinc finger protein ZFP57 requires its co-factor to recruit DNA methyltransferases and maintains DNA methylation imprint in embryonic stem cells via its transcriptional repression domain. J Biol Chem 2012;287(3):2107–18. [69] Kagiwada S, Kurimoto K, Hirota T, Yamaji M, Saitou M. Replication-coupled passive DNA demethylation for the erasure of genome imprints in mice. EMBO J 2013;32(3):340–53. [70] Lorsbach RB, Moore J, Mathew S, Raimondi SC, Mukatira ST, Downing JR. TET1, a member of a novel protein family, is fused to MLL in acute myeloid leukemia containing the t(10;11) (q22;q23). Leukemia 2003;17(3):637–41. [71] Hu L, Li Z, Cheng J, Rao Q, Gong W, Liu M, et al. Crystal structure of TET2-DNA complex: insight into TET-mediated 5mC oxidation. Cell 2013;155(7):1545–55. [72] Hu L, Lu J, Cheng J, Rao Q, Li Z, Hou H, et al. Structural insight into substrate preference for TET-mediated oxidation. Nature 2015;527(7576):118–22.

II.  Molecular Mechanisms of Epigenetics

24

2.  Mechanisms of DNA Methylation and Demethylation During Mammalian Development

[73] Xu Y, Wu F, Tan L, Kong L, Xiong L, Deng J, et al. Genome-wide regulation of 5hmC, 5mC, and gene expression by Tet1 hydroxylase in mouse embryonic stem cells. Mol Cell 2011;42(4):451–64. [74] Xu Y, Xu C, Kato A, Tempel W, Abreu JG, Bian C, et al. Tet3 CXXC domain and dioxygenase activity cooperatively regulate key genes for Xenopus eye and neural development. Cell 2012;151(6):1200–13. [75] Ko M, An J, Bandukwala HS, Chavez L, Aijo T, Pastor WA, et al. Modulation of TET2 expression and 5-methylcytosine oxidation by the CXXC domain protein IDAX. Nature 2013;497(7447):122–6. [76] Hashimoto H, Liu Y, Upadhyay AK, Chang Y, Howerton SB, Vertino PM, et al. Recognition and potential mechanisms for replication and erasure of cytosine hydroxymethylation. Nucleic Acids Res 2012;40(11):4841–9. [77] Cortellino S, Xu J, Sannai M, Moore R, Caretti E, Cigliano A, et al. Thymine DNA glycosylase is essential for active DNA demethylation by linked deamination-base excision repair. Cell 2011;146(1):67–79. [78] Shen L, Wu H, Diep D, Yamaguchi S, D’Alessio AC, Fung HL, et al. Genome-wide analysis reveals TET- and TDG-dependent 5-methylcytosine oxidation dynamics. Cell 2013;153(3):692–706. [79] Song CX, Szulwach KE, Dai Q, Fu Y, Mao SQ, Lin L, et al. Genome-wide profiling of 5-formylcytosine reveals its roles in epigenetic priming. Cell 2013;153(3):678–91. [80] Guo JU, Su Y, Zhong C, Ming GL, Song H. Hydroxylation of 5-methylcytosine by TET1 promotes active DNA demethylation in the adult brain. Cell 2011;145(3):423–34. [81] Nabel CS, Jia H, Ye Y, Shen L, Goldschmidt HL, Stivers JT, et al. AID/APOBEC deaminases disfavor modified cytosines implicated in DNA demethylation. Nat Chem Biol 2012;8(9):751–8. [82] Chen CC, Wang KY, Shen CK. The mammalian de novo DNA methyltransferases DNMT3A and DNMT3B are also DNA 5-hydroxymethylcytosine dehydroxymethylases. J Biol Chem 2012;287(40):33116–21. [83] Schiesser S, Hackner B, Pfaffeneder T, Muller M, Hagemeier C, Truss M, et al. Mechanism and stem-cell activity of 5-carboxycytosine decarboxylation determined by isotope tracing. Angew Chem Int Ed Engl 2012;51(26):6516–20. [84] Wossidlo M, Nakamura T, Lepikhov K, Marques CJ, Zakhartchenko V, Boiani M, et al. 5-Hydroxymethylcytosine in the mammalian zygote is linked with epigenetic reprogramming. Nat Commun 2011;2:241. [85] Iqbal K, Jin SG, Pfeifer GP, Szabo PE. Reprogramming of the paternal genome upon fertilization involves genomewide oxidation of 5-methylcytosine. Proc Natl Acad Sci USA 2011;108(9):3642–7. [86] Gu TP, Guo F, Yang H, Wu HP, Xu GF, Liu W, et al. The role of Tet3 DNA dioxygenase in epigenetic reprogramming by oocytes. Nature 2011;477(7366):606–10. [87] Yamaguchi S, Hong K, Liu R, Shen L, Inoue A, Diep D, et al. Tet1 controls meiosis by regulating meiotic gene expression. Nature 2012;492(7429):443–7. [88] Hackett JA, Sengupta R, Zylicz JJ, Murakami K, Lee C, Down TA, et al. Germline DNA demethylation dynamics and imprint erasure through 5-hydroxymethylcytosine. Science 2013;339(6118):448–52. [89] Guo F, Li X, Liang D, Li T, Zhu P, Guo H, et al. Active and passive demethylation of male and female pronuclear DNA in the mammalian zygote. Cell Stem Cell 2014;15(4):447–58. [90] Wang L, Zhang J, Duan J, Gao X, Zhu W, Lu X, et al. Programming and inheritance of parental DNA methylomes in mammals. Cell 2014;157(4):979–91. [91] Nakamura T, Liu YJ, Nakashima H, Umehara H, Inoue K, Matoba S, et al. PGC7 binds histone H3K9me2 to protect

[92]

[93] [94]

[95]

[96]

[97]

[98]

[99]

[100]

[101]

[102] [103]

[104]

[105]

[106]

[107]

[108]

[109]

against conversion of 5mC to 5hmC in early embryos. Nature 2012;486(7403):415–9. Yamaguchi S, Hong K, Liu R, Inoue A, Shen L, Zhang K, et al. Dynamics of 5-methylcytosine and 5-hydroxymethylcytosine during germ cell reprogramming. Cell Res 2013;23(3):329–39. Yamaguchi S, Shen L, Liu Y, Sendler D, Zhang Y. Role of Tet1 in erasure of genomic imprinting. Nature 2013;504(7480):460–4. Dawlaty MM, Breiling A, Le T, Raddatz G, Barrasa MI, Cheng AW, et al. Combined deficiency of Tet1 and Tet2 causes epigenetic abnormalities but is compatible with postnatal development. Dev Cell 2013;24(3):310–23. Kriaucionis S, Heintz N. The nuclear DNA base 5-hydroxymethylcytosine is present in Purkinje neurons and the brain. Science 2009;324(5929):929–30. Dawlaty MM, Ganz K, Powell BE, Hu YC, Markoulaki S, Cheng AW, et al. Tet1 is dispensable for maintaining pluripotency and its loss is compatible with embryonic and postnatal development. Cell Stem Cell 2011;9(2):166–75. Zhang RR, Cui QY, Murai K, Lim YC, Smith ZD, Jin S, et al. Tet1 Regulates Adult Hippocampal Neurogenesis and Cognition. Cell Stem Cell 2013;13(2):237–45. Moran-Crusio K, Reavie L, Shih A, Abdel-Wahab O, Ndiaye-Lobry D, Lobry C, et al. Tet2 loss leads to increased hematopoietic stem cell self-renewal and myeloid transformation. Cancer Cell 2011;20(1):11–24. Ko M, Bandukwala HS, An J, Lamperti ED, Thompson EC, Hastie R, et al. Ten-Eleven-Translocation 2 (TET2) negatively regulates homeostasis and differentiation of hematopoietic stem cells in mice. Proc Natl Acad Sci USA 2011;108(35): 14566–71. Li Z, Cai X, Cai CL, Wang J, Zhang W, Petersen BE, et al. Deletion of Tet2 in mice leads to dysregulated hematopoietic stem cells and subsequent development of myeloid malignancies. Blood 2011;118(17):4509–18. Quivoron C, Couronne L, Della Valle V, Lopez CK, Plo I, Wagner-Ballon O, et al. TET2 inactivation results in pleiotropic hematopoietic abnormalities in mouse and is a recurrent event during human lymphomagenesis. Cancer Cell 2011;20(1):25–38. Abdel-Wahab O. Genetics of the myeloproliferative neoplasms. Curr Opin Hematol 2011;18(2):117–23. Dai HQ, Wang BA, Yang L, Chen JJ, Zhu GC, Sun ML, et al. TETmediated DNA demethylation controls gastrulation by regulating Lefty-Nodal signalling. Nature 2016;538(7626):528–32. Dawlaty MM, Breiling A, Le T, Barrasa MI, Raddatz G, Gao Q, et al. Loss of Tet enzymes compromises proper differentiation of embryonic stem cells. Dev Cell 2014;29(1):102–11. Lu F, Liu Y, Jiang L, Yamaguchi S, Zhang Y. Role of Tet proteins in enhancer activity and telomere elongation. Genes Dev 2014;28(19):2103–19. Yang J, Guo R, Wang H, Ye X, Zhou Z, Dan J, et al. Tet enzymes regulate telomere maintenance and chromosomal stability of mouse ESCs. Cell Rep 2016;15(8):1809–21. Chen Q, Chen Y, Bian C, Fujiki R, Yu X. TET2 promotes histone O-GlcNAcylation during gene transcription. Nature 2013;493(7433):561–4. Vella P, Scelfo A, Jammula S, Chiacchiera F, Williams K, Cuomo A, et al. Tet proteins connect the O-linked N-acetylglucosamine transferase Ogt to chromatin in embryonic stem cells. Mol Cell 2013;49(4):645–56. Deplus R, Delatte B, Schwinn MK, Defrance M, Mendez J, Murphy N, et al. TET2 and TET3 regulate GlcNAcylation and H3K4 methylation through OGT and SET1/COMPASS. EMBO J 2013;32(5):645–55.

II.  Molecular Mechanisms of Epigenetics

C H A P T E R

3 Mechanisms of Histone Modifications Ludovica Vanzan*, Athena Sklias**, Zdenko Herceg**, Rabih Murr* *Institute for Genetics and Genomics in Geneva (iGE3), University of Geneva, Geneva, Switzerland; **International Agency for Research on Cancer (IARC), Lyon, France

O U T L I N E Introduction

25

Histone Modifications Proline Isomerization Citrullination Ubiquitination Sumoylation ADP-Ribosylation

25 26 26 26 31 31

Phosphorylation Methylation Acetylation

INTRODUCTION

Conclusions

39

References

39

and nucleosomes), and result in conformational changes translated into alterations in processes that use DNA as a template. Epigenetic modifications include RNAmediated interference, noncoding RNAs, nucleosome remodeling, DNA methylation (and its derivatives), and histone modifications. In this chapter, we will focus on histone modifications and their role in transcription and DNA repair.

The term “epigenetic” was introduced by Waddington in 1942 to describe “The interactions of genes with their environment that bring the phenotype into being.” In other words, epigenetic events are induced by geneextrinsic factors and affect the way the genetic “alphabet” is translated into proteins, the building blocks of the phenotype. However, it is not until relatively recently that epigenetics has emerged from the shadow of genetics and has become mainstream in modern biology, and is currently resting on well-defined molecular foundations that we will try to elaborate. Nevertheless, the field is still being continuously reshaped by the fast pace of conceptual and technological advances. Among many recent advances in the field is the realization that epigenetic mechanisms are not only involved in transcription regulation, but affect all processes using DNA as a template, including DNA repair and replication. Therefore, we intend to propose a paradigm shift in defining “epigenetic modifications” as heritable nongenetic modifications that take place on the chromatin (DNA Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00003-1 Copyright © 2017 Elsevier Inc. All rights reserved.

31 33 35

HISTONE MODIFICATIONS Eukaryotic chromatin is a highly condensed structure that is instrumental for fundamental nuclear processes, such as transcription, replication, and DNA repair. Chromatin exists in at least two distinct functional forms: a condensed form that generally lacks DNA regulatory activity, called heterochromatin, and a looser form that provides the environment for DNA regulatory processes, called euchromatin. Nucleosomes are the building blocks of chromatin and are formed by two turns of DNA (147 bp) wrapped around an octamer of two subunits of

25

26

3.  Mechanisms of Histone Modifications

each of the core histones H2A, H2B, H3, and H4. The chromatin structure fulfills essential functions by condensing DNA, preserving genetic information, and controlling gene expression. However, given its compacted structure, chromatin hinders important cellular processes, including transcription, replication, and the detection/repair of DNA breaks. Therefore, chromatin must be amenable to conformational changes that regulate accessibility of molecules involved in these processes. This raises a fundamental question in biology: how is chromatin remodeled? A part of the answer resides in the fact that cells have evolved epigenetic mechanisms, notably covalent posttranslational histone modifications that can alter the structure of chromatin. Histone modifications come in different “flavors” and affect multiple residues, primarily those located at amino-terminal (Nterminal) tails of histones that protrude from the nucleosome core (Figs. 3.1 and 3.2). Currently, there are at least 80 different identified types of histone modifications affecting more than 60 residues. This number is likely to be an underestimation, and the remarkable advances in proteomics, genomics, and computational biology will undoubtedly help to identify new modifications and target residues. The most studied types of histone modifications are acetylation, methylation, and phosphorylation. Additional modifications include ubiquitination, sumoylation, ADP-ribosylation, citrullination, and proline isomerization (Figs. 3.1 and 3.2). Traditionally, two mechanisms are thought to govern the function of these modifications [1]. First, these marks could affect the nucleosome–nucleosome or DNA–nucleosome interactions through the addition of physical entities or the change in histone charges. Second, they could serve as docking sites for the recruitment or repulsion of specific regulatory proteins. Additionally, numerous reports raised the possibility that these modifications can be interdependent and act in a combinatorial fashion, thus forming the “histone code” [2,3]. The molecular mechanisms involving histone modifications, their role in chromatin-based processes, and their interdependence will be discussed (Table 3.1). It has to be noted that, although an official nomenclature for histone-modifying enzymes was suggested [4], we chose to utilize the most widely used nomenclature for the sake of clarity and to avoid confusion.

Proline Isomerization Isomerization is defined as the transformation of a molecule into a different isomer, and it can adopt two distinct conformations: cis or trans. Protein isomerization was first described in 1968 [5] to dramatically affect protein conformation by disrupting the secondary structure of polypeptides. Isomerization occurs spontaneously, but enzymes called proline isomerases have evolved to accelerate switching between different conformations.

The first evidence of histone isomerization was reported in 2006 [6] when Frp4 was identified as a histone isomerase of prolines 30 and 38 (P30 and P38) on the histone H3 tail (Fig. 3.2). The conformational status of histone H3 P38 (H3P38) is necessary for the induction of methylation of lysine 36 of histone H3 (H3K36) and its isomerization appears to inhibit the ability of Set2 to methylate H3K36, resulting in defective transcription elongation. Moreover, a recent paper from Jane Mellor’s group described how acetylation of H3K14 promotes trans-isomerization at H3P16. This leads to a reduction of H3K4me3 levels in vivo, suggesting that H3P16 trans-isomerization could have a role in transcription repression [7].

Citrullination Arginine (R) residues of histones can be citrullinated by replacing an imine group with a carbonyl that causes the neutralization of positive charges of target proteins. This deimination reaction is catalyzed by peptidyl arginine deiminases (PADIs). Though there are five PADIs in humans, only PADI2 and PADI4 (previously referred to as PADI5) were shown to have a nuclear activity. Both are able to citrullinate H3R2, H3R8, H3R17, and H3R26, while H2AR3 and H4R3 have been described to be exclusively citrullinated by PADI4 [8–10] (Fig. 3.2). PADI4 exerts a dual enzymatic activity, as it converts arginine and methylated arginine into citrulline via deamination and demethylimination, respectively [11]. Recently, PADI4 was shown to citrullinate the H1R54 in the globular domain of H1 variants in mouse embryonic stem (ES) cells, which induces a global chromatin decondensation and plays an important role in the pluripotency of ES cells [12]. Likewise, H3R8 citrullination plays a role in chromatin opening by weakening the binding of HP1α, a heterochromatic protein, to H3K9me3 residues [13]. Histone citrullination has also been extensively studied in the context of estrogen receptor (ER)–mediated pathways. For example, H3R26 citrullination at ER-α target genes was shown to promote their expression by opening the chromatin [10]. Surprisingly, citrullination of H3R17 at ER-α target promoters leads to transcriptional repression by preventing CARM1 from methylating the same residue [14]. On the other hand, H4R3 is citrullinated by PADI4 in response to DNA damage and this results in the induction of apoptosis. PADI4 also interacts with tumor suppressor protein p53 during this process, which led to the proposal that PADI4 can act as a tumor suppressor [15].

Ubiquitination Ubiquitin is a 76–amino acid protein highly conserved in eukaryotes. Ubiquitination refers to the covalent attachment of one (monoubiquitination) or more

II.  Molecular Mechanisms of Epigenetics



Histone Modifications

27

FIGURE 3.1  Nonexhaustive list of posttranslational chemical modifications occurring on histone N-terminal tails in humans and rodents. These modifications are added preferentially to amino acids, such as lysine (K), arginine (R), serine (S), and threonine (T), and impact the DNA compaction and accessibility to transcriptional and DNA repair machineries. The commonly studied histone modifications are highlighted in green. SUMO, Small ubiquitin-related modifier protein.

(polyubiquitination) ubiquitin monomers to the amino group of a lysine residue. Although polyubiquitination typically marks a protein to be degraded via the 26S proteasome, and monoubiquitination modifies protein

function, these ubiquitin modifications may fulfill different roles depending on their target residues on histones. Histone H2A was the first histone shown to be ubiquitinated [16]. Subsequently, histones H2B (K119, K120,

II.  Molecular Mechanisms of Epigenetics

28

3.  Mechanisms of Histone Modifications

FIGURE 3.2  Posttranslational modifications of N-terminal tails of core histone H2A, H2B, H3, and H4. Superscripts (1/2/3) indicate that specific lysines can be mono-, di-, or trimethylated. Major histone acetyltransferases (ovals), methyltransferases (rectangles), and their main target residues are indicated. Asterisks indicate if the modification is described exclusively in certain species.

and K143) [17], H3, H4, and the linker histone H1 were likewise reported to be ubiquitinated (Fig. 3.2 and Table 3.1). However, ubiquitinated H2A and H2B represent the large majority of ubiquitinated histones and the function of other ubiquitinated histones is poorly defined. Moreover, although H2A and H2B may be polyubiquitinated [18,19], the monoubiquitinated forms are by far the most abundant. We will therefore primarily discuss the mechanisms of establishment and roles of H2A and H2B monoubiquitination (H2Aub and H2Bub). Histone ubiquitination consists of formation of an isopeptide bond between the C-terminus of ubiquitin and a lysine side chain of histones by sequential actions of E1-activating, E2-conjugating, and E3-ligase enzymes. Histone E3-ligases mostly belong to Homologous to E6AP-C Terminus (HECT) or Really Interesting New Gene (RING) protein families. H2A ubiquitination is primarily catalyzed by RING domain–containing members of the Polycomb Repressive Complex 1 (PRC1). The process starts with PRC2-dependent H3K27 trimethylation (H3K27me3) that serves as a docking site for PRC1 that

in turn ubiquitinates H2A. RING1B, the first identified ubiquitin ligase and member of the PRC1, is the main H2Aub E3-ligase at lysine 119, whose activity is highly dependent on interaction with two other RING domain– containing proteins in the PRC1, RING1A, and BMI1 [20,21]. The essential role of PRC2 in the recruitment of PRC1 was challenged by a number of reports. Indeed, the replacement of the PRC1 member CBX7 by RYBP in a newly identified PRC1-like complex allows the recruitment of the latter and subsequent PRC2-independent H2A ubiquitination [22]. Another surprising finding was that a variant of PRC1 could induce the recruitment of PRC2 [23]. Other PRC1-like complexes containing orthologs of PRC1 and RING1b, such as the RING-associated factor (dRAF) complex in D. melanogaster and bCL6 corepressor (bCoR) complex in mammals, can also ubiquitinate H2A [21,24]. Finally, RING-independent proteins, such as the 2A-HUB/hRUL138 and BRCA1, were shown to be able to ubiquitinate H2A although their actual contribution to the bulk of ubiquitinated histones in vivo is minimal. H2BK123-specific E2 in yeast, Rad6, was the first

II.  Molecular Mechanisms of Epigenetics



29

Histone Modifications

TABLE 3.1 Most Common Histone Modifications and Their Functions Across Different Model Organisms Histones

Residues

Modifications

Major functions

Genomic contexts

H2A

K5, K9

ac

DDR

—

K126

sum

Transcriptional repression

—

K119

ub

Transcriptional repression

—

S139

p

DDR

—

S129

p

DDR

—

N-term

PAR

DDR

—

K120

ub

Transcriptional activation: elongation

—

K9

me1, me2

Transcriptional activation

Euchromatin

me3

Transcriptional repression

Pericentric heterochromatin, permanent mark

K4

me1, me2, me3

Transcriptional activation, poised to transcriptional activation

Promoters and enhancers

K27

me1

Transcriptional activation

Promoters

me2

Transcriptional activation, enhancer silencing

Promoters and enhancers

me3

Transcriptional repression

Promoters, development, cell differentiation, X-chromosome silencing

ac

Transcriptional activation

Enhancers

me1

Transcriptional activation

—

me2

DDR

—

me3

Transcriptional activation: elongation

Exon borders/gene body

K9, K4, K14, K36

ac

Transcriptional activation

Promoters

R17, R26

me2a

Transcriptional activation

—

R8

cit

Transcriptional activation

Permissive chromatin

R3

cit

DDR

Apoptosis

S10

p

Transcriptional activation and repression

Promoters

K16

ac

Transcriptional activation and repression for DNA repair

—

K20

me1

Transcriptional activation and repression

Different spatiotemporal distribution of active and repressive marks

me2

Transcriptional activation and DDR

—

me3

Transcriptional repression

—

K4, K8, K12

ac

Transcriptional activation

Euchromatin

R3

me2a

Transcriptional activation

Promoters

me2s

Transcriptional repression

—

H2AX

H2B

H3

K36

H4

Note: Not all histone modifications are present across all model species (Homo sapiens, Saccharomyces cerevisiae, and Drosophila melanogaster). ac, Acetylation; cit, citrullination; DDR, DNA damage response; me, methylation (1, mono; 2, bi; 3, tri; a, asymmetric; s, symmetric); p, phosphorylation; PAR, poly-ADP-ribosylation; ub, ubiquitination.

histone E2 to be identified [25]. Rad6 activity is linked to the RING finger E3-ligase, Bre1. Homologs of Rad6 and Bre1 in humans were also shown to be involved in H2B lysine 120 monoubiquitination [26,27].

The addition of the ubiquitin moiety to histones is reversible through the activity of deubiquitinating enzymes (DUBs) that consist of ubiquitin C-terminal hydrolases and ubiquitin-specific processing proteases

II.  Molecular Mechanisms of Epigenetics

30

3.  Mechanisms of Histone Modifications

(UBPs). Sixteen UBPs have been identified so far in yeast, differing by the length of their amino-terminal part that confers their specificity [28]. DUBs specific to H2Bub are UBP8 and UBP10. UBP8 belongs to the SAGA complex and its activity is required for the transcription of SAGA-regulated genes [29,30]. Orthologs of UBP8 were found in Drosophila (Nonstop) and in human (USP22). UBP10 activity is SAGA independent but SIR dependent [31]. The deubiquitinase activity of its ortholog in human, USP36, was shown to be largely conserved [32]. UBP8 and UBP10, likely through their presence in different complexes, are recruited to distinct genomic regions. UBP8 is recruited to regions marked by H3K4me3, such as active promoters, while UBP10 colocalizes with regions marked by H3K79me3, such as telomeres [33]. DUBs specific to H2Aub are USP16, 2A-DUB, USP21, and BRCA1-associated protein 1 (BAP1). Finally, some DUBs, such as USP3, USP12, and USP46, have a dual specificity toward H2Aub and H2Bub [34]. Role of Histone Ubiquitination in Transcription Regulation Although H2Aub and H2Bub were reported to simultaneously activate and repress gene transcription, current evidence points toward a preferential association of H2Aub with transcription repression and of H2Bub with activation [35–37]. ChIP-Chip and ChIP-Seq experiments showed an enrichment of H2Bub upstream of PolII binding at transcription start sites (TSS) of highly expressed genes. This finding was in agreement with the requirement of H2Bub for the COMPASS complex–mediated di- and trimethylation of H3K4 and subsequent gene activation in yeast [38]. In agreement, ubiquitination of H2BK120, specifically deposited in humans by the RNF20/40 complex and UbcH6, was shown to be important for the methylation of H3K4 and H3K79 and subsequently for Hox gene activation [39]. Further support for the correlation of H2Bub with active transcription came from an elegant biochemical study in which two traceless orthogonal expressed protein ligation (EPL) reactions to specifically ubiquitinate H2BK120 incorporated into chemically defined nucleosomes. The results showed a direct activation of H3K79 methylation by hDot1L, a mark related to gene activation [40]. As both H2Bub and H3K79me are not associated with distal gene promoters, but rather with TSS and further with bodies of active genes, it was suggested that H2Bub is linked to transcription elongation rather than initiation. Indeed, H2Bub associates with the elongating PolII and was shown to be involved in the opening and reconstitution of the chromatin structure upon the passage of PolII across gene bodies [41,42]. In contrast, the presence of H2A ubiquitin ligases in several transcription repressor complexes suggests that H2Aub can repress transcription. This role was

confirmed in vitro by showing that H2Aub can inhibit the methylation of H3K4 [43]. Furthermore, in vivo observations that H2A ubiquitination is essential for Polycomb-mediated gene repression and that H2A DUBs are required for gene activation reinforced this hypothesis. However, the role of USP16 (H2A DUB) in HOX gene and X-chromosome silencing, the presence of H2B DUBs in some activator complexes, and their positive effect on transcription initiation and elongation suggests the involvement of ubiquitination–deubiquitination cycles in gene activation and repression. It is noteworthy that histone ubiquitination may also play a role in transcription by affecting other histone modifications. For example, H2Bub-dependent trans-tail regulation of H3K4 and H3K79 methylation is an important mechanism involved in transcription regulation in yeast [38,40,44,45]. These marks facilitate transcription initiation and elongation [46,47] and prevent Sir proteins from association with active euchromatic regions, thus restricting them to heterochromatic regions to mediate silencing. The conservation of H2Bub-dependent H3K4/ H3K79 methylation in higher eukaryotes is still debated: while some reports indicate that near ablation of H2Bub in human cells did not affect the global level of H3K4me3 [48], others reported that modifying H2Bub levels led to a correlated change in H3K4/H3K79 methylation [49,50]. These findings demonstrate that histone ubiquitination is interconnected with other histone modifications in the control of gene transcription. Role of Histone Ubiquitination in DNA Damage Response H2A, its variant H2AX, and H2B were shown to be monoubiquitinated at DNA damage sites. These events are important for double-strand break (DSB) repair, as disruption of ubiquitination leads to defects in DSB repair protein recruitment [51–53]. It is thought that monoubiquitination of histones facilitates DNA damage response (DDR) by contributing to chromatin “decondensation.” Moreover, similar to the involvement of ubiquitination–deubiquitination in transcription regulation, it was suggested that such a cycle is also important for efficient DDR. This notion was supported by the important role of USP3 and BRCC36 DUBs in DDR. However, the most established role of histone ubiquitination in DDR is related to H2AX polyubiquitination: H2AX is first phosphorylated at DSB sites leading to the recruitment of RNF8 and RNF168 that polyubiquitinate H2AX. This in turn allows the recruitment of RAP80 and subsequently 53BP1 and BRCA1, two essential proteins for DSB repair [54,55]. Finally, a recent study demonstrated that the linker histone H1 may play an important role in early steps of DDR. Upon DSBs, RNF8 mediates ubiquitination of H1K63, providing an initial binding platform for RNF168, that in turn ubiquitynates H2A [56].

II.  Molecular Mechanisms of Epigenetics



Histone Modifications

Sumoylation Sumoylation consists in the addition of a “small ubiquitin-related modifier protein” (SUMO) of ∼100 amino acids to target proteins. Sumoylation shares many biochemical aspects with ubiquitination. Indeed, SUMO is covalently attached to other proteins through the activities of ligases (E1, E2, and E3), it can be cleaved by peptidases, and in vertebrates, subtypes SUMO2/3 can sequentially be conjugated [57]. Contrary to ubiquitination in mammals, fewer ligase homologs can conjugate SUMO moieties to other proteins (1 E1, 1 E2, and at least 10 E3s) and in parallel just a few SUMO/sentrin-specific proteases (SENP1–3 and SENP5–7) can remove them. Histone sumoylation was first reported in 2003, when Shiio and Eisenman found that H4 can be modified by SUMO and they suggested that this modification leads to the repression of transcriptional activity through the recruitment of HDACs and HP1 proteins [58]. Currently, it is demonstrated that all four core histones can be sumoylated in yeast (Fig. 3.2). The putative sumoylation sites were identified as K6/7 and to a lesser extent K16/17 of H2B, K126 of H2A, and all lysines in the N-terminal tail of H4. Traditionally, histone sumoylation was considered to have a role in transcription repression by opposing other active marks, such as acetylation and ubiquitination [59]. However, a recent paper reported evidence of acetylation-dependent sumoylation events on H3 and H4 [60], which indicates that histone sumoylation might have, in some cases, a positive effect on transcription.

ADP-Ribosylation ADP-ribosylation is defined by the addition of an ADP-ribose moiety onto a protein using nicotinamide adenine dinucleotide (NAD) as a substrate. When the transfer takes place on an amino acid acceptor, it is referred to as mono- or poly-(ADP-ribosyl)ation (MARand PARylation, respectively). Initially these two modifications were thought to be performed by different enzymes, that is, ADP-ribosyltransferases (ARTs) and poly(ADP-ribose) polymerases (PARPs), respectively. However, advanced technologies showed that some PARP enzymes do not have a polymerase activity, but only ADP-ribosyl synthase activity, independently of the group they belong to. Hence, Hottiger et al. suggested that mammalian ARTs and PARPs be gradually renamed to diphtheria and cholera toxin-like ADP-ribosyltransferases (ARTDs and ARTCs, respectively) [61]. In humans, at least 15 ARTs/PARPs have been described to act in the nucleus but only 2, PARP1 and PARP2, interact with chromatin [62]. All core histones and linker histone H1 are subjects to MARylation, either in response to genotoxic stress or in physiological conditions, depending on the cell cycle stage, proliferation activity, and degree of

31

terminal differentiation [63,64]. PARylation can also be detected on the majority of histone types. There seems to be some specificity in the activity of PARP proteins on histones: PARP1 seems to preferentially PARylate H1, whereas PARP2 acts more on core histones. PARylation of H1 is believed to be responsible for its eviction from the chromatin and the further unwinding of the chromatin structure. Moreover, PARP1 is enriched in promoters and enhancers associated with actively transcribed genes [65,66]. In response to single-strand break (SSB), PARP1 and PARP2 PARylate the C- and N-termini of H1 and H2B, leading to the decondensation of the chromatin structure, probably due to PARylation-dependent histone eviction. This is proposed to facilitate the access of single-strand break repair (SSBR)/base excision repair (BER) factors to the site of damage. Moreover, PAR is used for tagging regions affected by DNA damage, allowing adequate response of the cell according to the extent of damage signaled by the presence of PAR moieties. Surprisingly, recent studies indicated that ribosylation in DDR may induce local chromatin condensation rather than relaxation. Indeed, PAR moieties are recognized by the macrodomain of the histone variant macroH2A1.1. This may lead to a transient condensation, increased phosphorylation of H2AX at the sites of break, and reduced recruitment of Ku70/80, leading to altered DDR [67]. However, it is difficult to reconcile this last study with the rest of the literature on the role of PARP in DNA repair. One possible explanation for this discrepancy would be that PARylation leads to a quick and transient compaction of the chromatin that protects the DNA from additional damage, and that this is rapidly reversed to allow DNA repair to take place. This hypothesis is supported by the quick and transient nature of the PAR-macroH2A1.1 interaction-dependent chromatin condensation that is gradually lost after the reduction of PARylation levels. To summarize, it is currently established that PARylation is important in response to DNA damage, although its precise role in DNA repair needs further determination.

Phosphorylation Protein phosphorylation represents the addition of a phosphate group by various specific protein kinases. Histones may be subjected to phosphorylation on serine, threonine, and tyrosine residues (Table 3.1). An important site for histone phosphorylation is the serine 10 of histone H3 (H3S10P) mediated by the Aurora-B kinase (Fig. 3.2). This event is essential for mitosis and meiosis [68]. In interphase, H3K9me3 recruits HP1 proteins that contribute to heterochromatin formation. However, during mitosis and meiosis this interaction is lost, although H3K9 is still methylated. It was shown that this is due

II.  Molecular Mechanisms of Epigenetics

32

3.  Mechanisms of Histone Modifications

to the phosphorylation of H3S10 that leads to eviction of HP1α, HP1β, and HP1γ [69], allowing the access of factors needed for proper condensation and segregation of the chromosomes. Serine129 of H2AX can also be phosphorylated by ATM and ATR in early stages of cellular response to DSBs. H2AX could be additionally phosphorylated on tyrosine 142, a modification that has a role in the choice between apoptosis and DNA repair in response to genotoxic stress. H4 (S1) and linker histone H1 (S18, S173, S189, T11, T138, and T155) can also be phosphorylated by the CK2 and DNA-PK, respectively. Role of Histone Phosphorylation in Transcription Regulation Understanding the relationship between histone phosphorylation and gene expression is far from complete. The ability of H3S10P to condense chromatin favors a role in transcriptional repression; however, evidence has accumulated to support an important role in transcriptional activation in various organisms. First, it was shown that the induction of immediate-early genes in growth factor–treated mammalian cells was accompanied by serine phosphorylation at the N-terminal tail of H3 [70]. Later on, Corces’ laboratory indicated that activation of heat shock genes in Drosophila correlates with hyperphosphorylation of H3S10. Consistently, dephosphorylation of H3S10 carried out by a phosphatase called protein phosphatase 2A (PP2A) results in transcription inhibition [71]. Additionally, H3S10P was shown to be important for the activation of NFkB-regulated genes [72]. ChIP-Chip analysis in budding yeast indicated that several kinases are not only present in the cytoplasm, but also on the chromatin of specific genes [73], which suggests that the kinase signal transduction cascade could have a direct effect on gene expression by phosphorylation of the histones of specific gene promoters. In addition to its independent role in transcription regulation, histone phosphorylation has a complex association with acetylation events: several studies provided evidence of H3S10P-dependent promotion of H3K14 acetylation and transcription activation at Gcn5 promoter [74–76]. The kinetics of this interaction, however, is not clear. On one hand, studies on the Gal1 promoter revealed that the H3K14ac mark is present prior to the H3S10P and that both modifications are necessary for Gal1 transcriptional activation [77]. On the other hand, phosphorylation, but not acetylation, is needed for Gal4 transcriptional activation [78]. Nevertheless, the phosphobinding proteins, such as 14-3-3, showed a higher affinity for the dual phosphoacetylation mark. Recently, another study by the Corces’s lab explored the genomewide acetylation–phosphorylation relationship and demonstrated the presence of the H3K9acS10P and H3K27acS28P dual marks on most enhancers and promoter sequences in Drosophila, as well as the importance

of these marks for the recruitment of 14-3-3 proteins [79]. Phosphorylation of the H3T11 and H3S28 has also been associated with H3 acetylation and transcriptional activation at enhancers and promoters [79–81]. In particular, H3S28P is considered as a hallmark of transcriptional response to cellular stress. Genomewide studies revealed its presence at more than 50% of stress-induced genes, contributing to the reduction of HDAC occupancy and thus increasing histone acetylation [82]. Role of Histone Phosphorylation in DNA Damage Response Beside its role in transcription, phosphorylation of histones, in particular phosphorylation of H2AX, has a major role in DDR and DNA repair (Table 3.1). Rapid phosphorylation of H2AX, at serine 129 (γ-H2AX) by the PI3K kinases at DSB sites, is one of the first and most easily detectable DNA damage signaling posttranslational events. It anticorrelates with the constitutive phosphorylation of H2AX-Y142 (by WSTF) that is present under physiological conditions and removed in response to DNA damage via the activity of Eya tyrosine phosphatase. The kinase activity of WSTF and the phosphatase activity of Eya were shown to be important for the early recruitment of phospho-ATM and MDC1 to sites of DNA damage, thus privileging DNA repair over apoptosis [83,84]. γ-H2AX can be detected over kilobases (in yeast) or megabases (in mammalian cells) from sites of DSBs [85], and is required for the retention/accumulation of repair proteins [86]. γ-H2AX also plays a role in cohesion binding to a large region around DSB, an event thought to be important for sister chromatid cohesion in postreplicative repair [87,88]. Interestingly, γ-H2AX is required for the recruitment of the yeast NuA4 histone acetyltransferase (HAT) complex, via its subunit Arp4, to sites of DSBs induced by HO endonuclease. The recruitment leads to histone acetylation, thereby opening the chromatin and facilitating repair [86]. Arp4 is also a subunit of the ATP-dependent chromatin-remodeling complex INO80/SWR1 and contributes to the recruitment of the latter to γ-H2AX-marked regions. The resulting remodeling activity seems to be required for the DSB repair. Hence it would appear that cells can utilize the activities of both histone-modifying and remodeling complexes to facilitate DNA repair. The precise role of γ-H2AX in DSB repair is still under debate. Originally, it was suggested that it is essential for the recruitment of DNA repair enzymes through their BRCA1 COOH terminal (BRCT) domain [89]. However, a study by Celeste et al. challenged this hypothesis by demonstrating that DNA repair proteins, including Brca1 and Nbs1, could be recruited to DNA breaks in the absence of γ-H2AX. On the other hand, the presence of γ-H2AX is essential for the formation of irradiation-induced foci (IRIF) [90], indicating that the role of γ-H2AX

II.  Molecular Mechanisms of Epigenetics



Histone Modifications

33

phosphorylation may be dispensable for the original recruitment of DNA repair factors, but indispensable for the accumulation/retention of these factors. Histone phosphorylation can also occur on H4S1 through the activity of casein kinase II (CK2) in response to DSBs and it facilitates nonhomologous end joining (NHEJ) repair [91]. A study from Côté’s laboratory demonstrated that this phosphorylation coincides with a decrease in acetylation, suggesting that it occurs after H4 deacetylation and may regulate chromatin restoration after repair is completed [92]. Finally, linker histone H1 phosphorylation by DNA-PKcs is also required for efficient NHEJ [93]. These studies demonstrate that histone phosphorylation plays an important role in gene transcription and DDR mainly through chromatin modulation, and that phosphorylated histones can have distinct effects on chromatin structure depending on the context and the exact targeted residue.

subsequently discovered and classified into two families of histone lysine demethylases [107,108]: the Jumonji ATrich interactive domain 1 (JARID1) family that catalyzes a FAD-dependent reaction and the Jumonji C (JmjC) domain–containing proteins (JMD2) family characterized by a Jumonji domain for the removal of the methyl group via a 2OG-dependent oxidation reaction [107]. The JMD2 family includes UTX (H3K27 demethylase), JMJD2D and JMJD2B (H3K9 demethylases), JHDM1 (H3K36 demethylase), and JHDM3A and JMJD2C/GASC1 (H3K9 and H3K36 demethylases). JARID1 proteins include RBP2, PLU1, SMCY/Jarid1ds, and SMCX. Histone arginine demethylation can occur via the formation of the intermediate citrulline by deimination catalyzed by the PADI family of demethylases [11,14]. In addition, JMJD6, a JmjC-containing protein, was identified as a direct arginine demethylase [11]. However, other structural and functional studies confuted this proposal, suggesting that JMJD6 is instead a lysine hydroxylase [109,110].

Methylation

Role of Histone Methylation in Transcription Regulation The effect of histone methylation on transcription depends on the position of the methylated residue: H3K4 methylation is linked to transcription initiation and elongation [44,46,47,111,112], and H3K36 methylation correlates with transcription elongation [2,46,100,113,114], whereas H3K9 and H3K27 methylation are involved in gene silencing (Table 3.1). ChIP-Chip and ChIP-seq experiments generated a detailed picture on histone methylation distribution. While H3K4me2/3 are enriched at active gene promoters [115–117], H3K4me2 further extends to active gene bodies. H3K4me1 is also enriched at active gene bodies and peaks at 5′ end of genes. H3K36 progressively shifts from monomethylation at promoters to trimethylation at 3′ end of genes [100,116,118–123]. Finally, H3K79me2 and H3K79me3 are highly enriched at TSS of active genes and slowly decline through the first intron [115,124,125]. H3K79me2/3 are also present at several origins of replication in mammals, where their role is to prevent endoreplication [126]. Inactive gene promoters, on the other hand, are characterized by high levels of H3K27me3 or H3K9me3 and low levels of H3K4me1/2/3. Active enhancers are marked by H3K27ac, H3K4me1, and H3K79me2/3, while inactive enhancers harbor H3K27me3 [127–130]. The precise underlying mechanism of histone methylation–dependent transcription regulation is still not clear. One hypothesis postulates that histone-modifying complexes and chromatin-remodeling factors recognize and bind to the methylation mark through their plant homeo domain (PHD), double Tudor domain (DTD), or double chromodomain (DCD) [131–133], thus activating transcription. Another mechanism involves the crosstalk between histone methylation, other histone

Protein methylation represents the covalent addition of a methyl group from the donor S-adenosylmethionine (SAM) on carboxyl groups of glutamate, leucine, and isoprenylated cysteine, or on the side-chain nitrogen atoms of lysine, arginine, and histidine residues [94] (Table 3.1). Arginine residues on histones can be mono- or dimethylated and lysines can be mono-, di-, or trimethylated [1]. Most histone lysine methyltransferases (HKMTs) contain an evolutionary conserved domain called Su(var)3-9, Enhancer of zeste, Trithorax (SET) responsible for the deposition of the methyl group specifically at H3K4, H3K9, H3K27, and H3K36 [95–97] (Fig. 3.2). Methylation of H3K9 is carried out by SUV39H1/2, G9a, GLP, and SETDB1, whereas H3K27 is methylated by the EZH2, a subunit of PRC2. NSD1 can mono- and dimethylate H3K36 [98], whereas dimethylation and trimethylation are catalyzed by ASH1 and SETD2, respectively [99–101]. H3K4 methylation could be mediated by several SET domain–containing methyltransferases, including MLL1–5, SET1A/B, SET7/9, SMYD1, SETMAR, and PRDM9 [102]. H3K79, less accessible, as it is located in the globule histone core, is methylated by Dot1L, an enzyme lacking the SET domain [103]. Protein arginine methyltransferases (PRMTs) are classified as type I and II, for monomethylation and asymmetric dimethylation, and type III for monomethylation and symmetric dimethylation. The most targeted histone residues are H3R2, H3R17, H3R26, and H4R3 [104] (Fig. 3.2). The first identified histone demethylase is LSD1, a nuclear amine oxidase homologous that mainly demethylates H3K4, using flavin adenosine dinucleotide (FAD) as a cofactor. LSD1 could also demethylate H3K9, when it is present in a complex with the androgen receptor [105,106]. A number of other related enzymes were

II.  Molecular Mechanisms of Epigenetics

34

3.  Mechanisms of Histone Modifications

modifications, and DNA methylation. In the following paragraphs, we will summarize the role of several histone methylation residues in transcription regulation. Methylation of H3K4 invariably correlates with active transcription. In support of the notion of methylation-dependent recruitment of transcription regulators, a study by the Roeders’s group demonstrated that H3K4me3 is recognized by the TAF3 subunit of TFIID, favoring the recruitment of the whole TFIID complex at the promoter of p53-dependent genes [134]. Moreover, the presence of H3K4me1/2/3 is mechanistically linked to depletion of DNA methylation at promoters and enhancers [135–137]. The study from Hsieh and Okitsu indicated that H3K4 methylation may directly lead to loss of DNA methylation [138]. Recent studies further found that unmethylated CpG islands recruit CpG-binding CXXC domain–containing proteins, such as MLL1 H3K4 methyltransferase and CFP1, a cofactor of SET1 H3K4 methyltransferase, which results in trimethylation of H3K4 [139–142]. The presence of H3K4me3 blocks the activity of the H3K4me0-interacting ADD domain contained in DNMT3a and DNMT3L and therefore protects CpG ­islands from DNA methylation [143–145]. Surprisingly, a recent study showed that the presence of H3K4me1 at promoters can correlate with gene repression. However, the study did not confirm the causal relationship, but rather suggested a role of H3K4me1 in directing further methylation of H3K4 during development and therefore limiting the appearance of H3K4me3 to previously marked H3K4me1 promoters [146]. H3K36 methylation marks at active gene bodies are likewise recognized by the chromodomain of the Eaf3 subunit of The Rpd3S HDAC [120,147–150]. This leads to deacetylation of gene bodies, thus preventing transcription initiation at cryptic sites. This role is reinforced by the presence of DNA methylation at transcribed gene bodies [151]. The cooccurrence of the two marks seems to depend on the preferential interaction of the PWWP domain contained in DNMT3a/b with H3K36me3 [152]. The presence of the PWWP domain is essential for DNMT3a/b activity, as its mutation, either experimentally induced or found in diseases, such as the ICF syndrome, leads to a decreased DNA methyltransferase activity. However, the relationship between H3K36me3 and DNMT3a/b recruitment might be more complex, as the removal of H3K36me3, mediated by knockdown of Setd2, reduced gene body methylation by affecting DNMT3b, but not DNMT3a recruitment in ES cells [153], while it had no effect on DNA methylation in a cancer cell line [154]. H3K79 methylation was also implicated in transcription activation and elongation [155]; however, the evidence for such a role are limited to experiments showing its effect on the activation of HOXA9 by inhibition of Sir2/3-dependent heterochromatin spreading.

Moreover, while both H3K79me2 and H3K79me3 are enriched in gene bodies in yeast and Drosophila, only Drosophila H3K79me2 correlates with active transcription [122,156]. Three lysine methylation sites are related to transcriptional repression: H3K9, H3K27, and H4K20. H3K9me2/3 is essential for heterochromatin formation and subsequent gene silencing [95,157,158]. The marks are recognized by HP1 chromodomain, which then propagates the heterochromatic state. The initial trigger for H3K9me3 deposition is still unclear: in Schizosaccharomyces pombe, it has been shown that Argonaute and Dicer, two proteins of the RNAi machinery, are required for heterochromatin formation [159]. Other studies suggested the involvement of specific transcription factors, such as Atf1, PCR1, and Taz1. Finally, H3K9 and DNA methylation are tightly interconnected: mutation of the methyl DNA–binding domain of the SUVH5 enzyme affects its activity, thus reducing the levels of H3K9me2 in plant [160]. Reciprocally, DNMT3a/b are recruited by direct interaction with HP1, and SUV39H knockout in mouse cells leads to alterations in the DNA methylation status [161]. Surprisingly, one study showed that H3K9me3 could be located in the gene bodies of active genes along with HP1 [162]. This observation led to the proposed model where H3K9me within the coding regions correlates with transcription, whereas H3K9me on promoters and intergenic regions is repressive. Methylation of H3K27 is likewise traditionally associated with gene silencing, contributing notably in HOX gene cluster regulation, X-chromosome inactivation, and genomic imprinting. Interestingly, genomewide experiments showed that PRC2 and H3K27me3 are enriched at and repress developmental gene promoters in ES cells. These promoters are also occupied by “pluripotency factors” Oct4, Sox2, and Nanog, indicating that these factors could contribute to PRC2 recruitment [163,164]. Other gene promoters, although not active in ES cells, are concomitantly enriched with H3K4me3 and H3K27me3. These “bivalent promoters” are thought to be in a poised state, ready to be either transcribed or silenced, according to the differentiation pathway the cell will follow [117,165,166]. Despite that DNA methylation and H3K27 methylation are both repressive marks, they are mutually exclusive at regulatory elements [167,168]. This might result from the association of PRC2 with TET1, a putative DNA demethylase [169]. It is hypothesized that developmental genes are first inhibited by H3K27me3 in pluripotent cells and later permanently silenced by DNA methylation during development. This is supported by the observation that PRC2 may recruit DNMTs [170]. Studies on the role of H4K20me1 in transcription generated conflicting results: while some reported a correlation with histone hyperacetylation at transcribed genes and promoters, particularly those with high CpG content

II.  Molecular Mechanisms of Epigenetics



Histone Modifications

[115,116,171,172], others have associated H4K20me1 to gene repression, notably during mitosis and X-chromosome inactivation [173,174]. Reports on H4K20me3, on the other hand, invariably associated it with transcription repression [157,175,176]. Arginine methylation has a bivalent role in transcription regulation: while monomethylation and asymmetric dimethylation, deposited by PRMTs type I, such as CARM1, PRMT1, and PRMT2, are associated with transcription activation, symmetric dimethylation by PRMT5 is linked to repression [177–179]. The first arginine methyltransferase identified as a transcriptional regulator was the PRMT4/CARM1 that deposits the H3R17me2a and the H3R26me2a marks [180,181]. PRMT4/CARM1 works as coactivator of nuclear receptors, ER in particular, and other transcription factors, such as p53, NFkB, PPAR-γ, and c-Fos [182–186]. However, H4R3me2a represents the majority of arginine methylation events and is deposited by PRMT1. Both the mark and the enzyme are enriched at active gene promoters, which favors the recruitment of p300 and other acetyltransferases [187,188]. For example, upon the activation of ER-α, the promoter of a downstream target called pS2 undergoes repetitive and regular cycles of activation [189] that correlate with the recruitment of PRMT1. However, following studies found that these activation cycles were probably not due to arginine methylation, but cycles of DNA methylation/demethylation, although this cyclical DNA (de) methylation [190,191] is currently a subject of controversy. Finally, numerous studies reported the importance of PRMT5, responsible for the H4R3me2s modification, in transcription repression [192,193]. In particular, its presence on genes involved in cell proliferation (e.g. c-Myc target gene CAD and tumor suppressors ST7 and NM23) correlates with the recruitment of mSin3/HDAC and Brg1/hBrm and subsequent repression [194,195]. In summary, histone methylation is prominently involved in both transcriptional repression and activation, in which some histone methylation marks serve as signals for transcriptional regulation, whereas others provide memory of transcriptional activity. Role of Histone Methylation in DNA Repair The involvement of lysine methylation, in processes other than transcriptional regulation, has recently received considerable attention. Upon DNA damage, H4K20 methylation by Set9 serves as a docking site for the recruitment of Crb2 via its double Tudor domains to sites of damage. Crb2 acts as a DNA damage sensor and checkpoint protein in S. pombe, and its recruitment increases cellular survival following genotoxic stress [196,197]. Its mammalian homolog, 53BP1, binds methylated H3 and H4 at sites of DSBs [196,198]. Interestingly, H3K9 methylation was linked to activation of HAT complexes (such as TIP60) following DNA damage [199]. Upon DSB, a complex

35

containing Kap1, HP1-β, and SUV39H1 is recruited, leading to trimethylation of H3K9. H3K9me3 then recruits and activates TIP60 and enhances the recruitment of the Kap1/HP1-β complex. TIP60 acetylates ATM and H4K16, resulting in subsequent recruitment of BRCA1 and depletion of 53BP1. In a negative feedback loop, ATM phosphorylates Kap1, releasing the Kap1/HP1/SUV39H1 complex from the damaged site [199]. The presence of either the BRCA1- or 53BP1-related repair machineries on DSB sites is thought to generate a dynamic alternation between homologous recombination and NHEJ [200]. Although the importance of histone methylation in DDR and DNA repair has been established by several studies, the molecular mechanisms involving this mark appears to be complex and remain debated.

Acetylation Histone acetylation is the process by which an acetyl group from acetyl-CoA is transferred to the lysine amino groups on the N-terminal tails of histones. This enzymatic activity is catalyzed by enzymes called HATs. HAT enzymes are often part of multisubunit complexes that contain one HAT catalytic subunit, adapter proteins, and in many cases, a large scaffold protein called TRRAP. Acetylation can occur on specific lysines in all four histones (Fig. 3.2 and Table 3.1). Hyperacetylation of histones is considered as a hallmark of transcriptionally active regions. However, acetylation is not exclusively related to transcription, but can affect other DNA-based cellular processes, such as DNA repair. Two classifications can be used to separate HATs. In the first classification, HAT complexes can be divided into two classes based on their suspected cellular origin and functions: A-type HATs are nuclear enzymes that catalyze acetylation on already deposited histones in the context of the chromatin. B-type HATs are cytoplasmic enzymes that are thought to be responsible for the acetylation of newly synthesized histones, leading to their transport from the cytoplasm to the nucleus where they are deposited onto newly replicated DNA [201,202]. On the other hand, the modern classification of HATs uses structural criteria, such as the presence of chromodomains, bromodomains, and zinc finger domains. This classification separates the HATs into two major families: Gcn5-related acetyltransferases (GNATs) and MOZ, Ybf2/Sas3, Sas2, and Tip60 (MYST)–related HATs. In addition to these families, one can add p300/CBP (KAT3A/B) HATs, the general transcription factor HATs that include the TFIID subunit TAFII250 and the nuclear hormone–related HATs: SRC1 and ACTR (SRC3). GNAT Superfamily All the GNAT superfamily members share structural and sequence similarity to Gcn5 (KAT2A). This

II.  Molecular Mechanisms of Epigenetics

36

3.  Mechanisms of Histone Modifications

superfamily is characterized by four regions with different degrees of conservation (labeled A–D) spanning over 100 residues. These regions were first defined by the comparison between Gcn5 and B-type Hat1. Motif A, also called AT domain, contains an Arg/Gln-X-XGly-X-Gly/Ala sequence and is shared with other HAT families. This motif is the most highly conserved and is important for acetyl-CoA recognition and binding [203]. The C motif is found in most of the GNAT family members, but not in the majority of other HATs. The GNAT superfamily contains over 10,000 members distributed in all kingdoms of life, including HATs and nonhistone acetyltransferases. The most relevant members of this family are: Gcn5, PCAF (KAT2B), Hat1 (KAT1), Elp3 (KAT9), and Hpa2. MYST Superfamily The MYST family was named after its founding members: MOZ (KAT6A), Ybf2/Sas3 (KAT6), Sas2 (KAT8), and Tip60 (KAT5) [204]. These proteins are grouped together on the basis of their close sequence similarities and their possession of a particular acetyltransferase homology region (part of motif A of the GNAT superfamily). This region binds acetyl-CoA [205], in addition to a zinc finger domain called C2HC (C-X2-C-X13-H-X-C), and a “E-R” motif (Esa1-Rpd3), both needed for the enzymatic activity and for the substrate recognition [206]. Recently, additional members of this family were identified, including Esa1 (KAT5 family) in yeast, MOF (KAT8 family) in Drosophila, and HBO1 (KAT7) and MORF (KAT6B) in mammals. Despite their structural similarities, members of this superfamily have diverse functions. They resemble those of the GNAT family, as both have an AT domain [207], but differ in C- and N-terminal parts, exhibiting affinity for different substrates. In addition, MYST family members possess either a chromodomain or an additional zinc finger domain termed the PHD domain [208,209]. P300/CBP P300 and CBP are large proteins (∼300 kDa) containing more than 2400 residues. Four interaction domains have been characterized throughout their sequence. These include a bromodomain also found in several other HATs, such as Gcn5 and PCAF. P300/CBP have homologs in most metazoans, but not in inferior eukaryotes, including yeast. They were first identified as transcriptional adaptors for many different transcription factors that directly contact DNA-bound activators. In vitro studies indicate that p300/CBP preferentially acetylate H2BK12, H2BK15, H3K14, H3K18, H3K56, H4K5, and H4K8 [210]. p300 and CBP are often referred to as a single entity, owing to their structural and functional similarity. However, the two proteins diverge in several functional and structural properties. First, some studies

identified phosphorylation residues specific for each of the two proteins [211]. Second, in response to ionizing radiation, p300, but not CBP, is important for apoptosis induction (probably through the activation of p53) [212]. Third, while both proteins are necessary for apoptosis and G1 arrest, differentiation and induction of the cell cycle inhibitor p21/Cip1 critically depends on p300, whereas induction of p27/Kip1 requires CBP [213]. Finally, loss-of-function studies showed that individual knockouts of the two proteins resulted in different phenotypes [214,215]. The specificity of the acetyltransferase activity of p300 and CBP proteins may explain, at least in part, their different functions. For instance, CBP has a preference for acetylating H4K12, while p300 preferentially acetylates H4K8 [212]. However, both proteins were shown to be able to acetylate H3K56 in collaboration with ASF1A histone chaperone. Another histone chaperone, CAF1, is needed for the incorporation of H3K56Ac into the chromatin, notably in response to DNA damage [216]. Other studies showed that H3K56 can be acetylated in the cytoplasm by HatB3.1, a Gcn5containing HAT complex, prior to its transport to the nucleus [217,218], which indicates that Gcn5 has both nuclear and cytoplasmic localization. Another HAT-like complex in yeast, known as Rtt109p (KAT11), was also identified to be responsible for H3K56 acetylation in the cytoplasm [219]. Nuclear Hormone–Related HATs Establishing that human coactivators ACTR (KAT13B), SRC-1 (KAT13A), and TIF2 (KAT13C) that interact with nuclear hormone receptors, exhibit HAT activity confirms the involvement of acetylation in yet another system of transcriptional regulation brought about by hormone signals and defines a distinct family of HATs. The members of this family share several similarities, including HAT domain in the C-terminus and an N-terminal basic helix–loop–helix/PAS region, as well as receptor and coactivator interaction domains [220]. HAT Complexes Most HAT enzymes, alone, are not able to acetylate histones in the context of nucleosomes. However, when present in multisubunit complexes, these enzymes become active and stable. Furthermore, the substrate of HAT enzymes may change according to the HAT complex to which they belong. HAT complexes are divided into several families. Between all HAT complexes, GNAT or SAGA-like HAT complexes (SAGA, SLIK, PCAF, STAGA, and TFTC) are unique by the fact that they contain TAFs. The enzymatic subunit of these complexes can be represented by Gcn5 or PCAF. To date, two complexes belonging to this group have been discovered in yeast (SAGA and SALSA/ SLIK) and three in humans (PCAF, STAGA, and TFTC).

II.  Molecular Mechanisms of Epigenetics



Histone Modifications

Subunits of these complexes include Ada proteins, Spt proteins, TAFs, SAP130, and TRRAP. Grant et al. in a pioneering study, identified four Gcn5-containing HAT complexes in yeast: SAGA, nucleosomal acetyltransferase of histone H3 (NuA3), nucleosomal acetyltransferase of histone H4 (NuA4), and ADA [221]. NuA3is a 500-kDa complex that exclusively acetylates H3. Peptide sequencing of proteins from the purified NuA3 complex identified Sas3, a MYST protein involved in silencing, as the catalytic HAT subunit of the complex. NuA3 also contains the TBP-associated factor, yTAF(II)30, and Yng1, a PHD finger–containing protein recently found to interact with H3K4me3. The interaction between Yng1 and H3K4me3 seems to promote NuA3 HAT activity at H3K14 and transcription at a subset of targeted ORFs [222]. In vitro studies on NuA3 indicated that it failed to interact with activation domains or to have any specificity in transcription activation [223,224]. NuA4 and its human homolog, TIP60 [225], are able to acetylate free histones H3, H4, and H2A. However, when histones are folded into nucleosomes, the activity of these enzymes depends on the presence of other proteins in the TIP60/ NuA4 complex. The HAT activity of this complex seems to be restricted to H4 and H2A [225]. Further studies identified three new complexes that share several subunits of the TIP60/NuA4 complex: p400 complex and TRRAP-BAF53-TIP48-TIP49–containing complex in humans and “Piccolo NuA4” in yeast. The p400 complex has no HAT activity, but can hydrolyze ATP [226]. The second human complex has HAT activity, but the HAT enzyme responsible for this activity is yet to be identified. The Piccolo NuA4, as its name indicates, is a small complex as it contains only three subunits (Tip60p/NuA4–Ing3–Epc1). This complex is conserved in humans and seems to represent the catalytic core of TIP60 complexes [227]. Although we tried to be as exhaustive and accurate as possible, the list of proteins that harbor HAT activity is rather long and their functional characterization is somewhat hampered by the wide spectrum of their targets, including histones and nonhistone proteins. Therefore, we do not pretend to exclude different classification systems and functions of HATs than the ones presented earlier. Histone Deacetylases There are four distinct families of histone deacetylases: class I (Rpd3-like) and class II (Hda1-like) histone deacetylases, class III NAD-dependent enzymes of the Sir family, and class IV for the single-member HDAC11 [228]. They are involved in multiple signaling pathways, are present in numerous repressive chromatin complexes, and do not show high specificity for a particular acetyl group. However, yeast enzyme Hda1 seems to have higher specificity for H3 and H2B, whereas Hos2 is

37

specific for H3 and H4. The fission yeast class III deacetylase, Sir2, and its human homolog, SirT2, preferentially deacetylate H4K16ac and H3K56ac. Role of Histone Acetylation in Transcription Regulation The established role of histone acetylation is to regulate gene transcription (Table 3.1). The first evidence of the involvement of HATs in transcription dates back to 1964, when it was observed that chromatin regions of actively transcribed genes tend to have hyperacetylated histones [229]. The addition of acetyl groups to histone tails was proposed to neutralize the histone charge, thus weakening histone–DNA interactions, thereby opening the chromatin structure and facilitating the access of transcription machinery [230]. Work from Craig Peterson’s laboratory demonstrated that the incorporation of H4K16ac into nucleosomal arrays impedes the formation of compacted chromatin fibers and prevents the ATPmediated chromatin-remodeling factors from mediating nucleosome sliding [231,232]. Likewise, H3K56ac was implicated in transcriptional activation by opening the chromatin structure. Indeed, H3K56 residue is facing the major groove of the DNA within the nucleosome, so it is in a particularly suitable position to affect histone/ DNA interactions when acetylated [233–235]. In addition, two other mechanisms by which histone acetylation facilitates transcription have been proposed. First, there is evidence that histone acetylation may serve as a docking site for the recruitment of transcription regulators [236,237]. Second, histone acetylation could contribute to the “histone code” [238,239]. HAT complexes from GNAT and MYST families were shown to be recruited to activator-bound nucleosomes resulting in transcriptional activation [92,240,241]. The recruitment of SAGA leads to acetylation of promoter-proximal H3, whereas recruitment of NuA4 results in a broader domain of H4 acetylation (more than 3 kb) [241]. This hyperacetylation of histones was linked to transcription activation of several genes [242]. Moreover, several studies indicated that TRRAP (a subunit of many HAT complexes) and its yeast homolog (Tra1) are recruited by c-Myc and N-Myc to the promoters of transcribed genes. The recruitment of TRRAP leads to increased histone acetylation by Tip60 and Gcn5/PCAF, followed by transcription activation [243–246]. Genome-wide analysis indicated that both gene-proximal (promoters) and -distal (enhancers) regulatory regions of active genes are acetylated (H3K27ac at both regions and H3K9ac at promoters). Moreover, the combination of H3K27ac with other histone modifications faithfully distinguishes active, poised, and inactive enhancers and promoters. Inactive promoters and enhancers are marked by H3K27me3 or DNA methylation. Acquisition of H3K4me2/me3 at promoters and H3K4me1 at enhancers renders them poised for activation

II.  Molecular Mechanisms of Epigenetics

38

3.  Mechanisms of Histone Modifications

that occurs upon the loss of HK27me3 and the appearance of H3K27ac [127,128,130]. Most evidence from biochemical and genetic studies pointed toward the essential role and nonredundancy of different HATs, meaning that each HAT would acetylate specific residues on histones. This specificity is also implied by the fact that HAT enzymatic components interact with different partners in the context of HAT complexes. This in turn predicts that different types of acetylated histones should have essential and nonoverlapping functions. Mutation and deletion experiments further supported the high degree of specificity of different HATs, which is reflected in the significant variability of the phenotypes produced by the ablation of HATs. For example, Tip60 was shown to be essential for early stages of development in mice (embryonic day 3.5, ED 3.5) and for maintaining stem cell pluripotency, whereas GCN5 is dispensable for pluripotency, and GCN5-depleted embryos can develop until ED 10.5 [247–250]. However, recent genomewide mapping of histone acetylation and HATs seems to point toward redundancy between different HATs and different histone acetylation marks. Indeed, the results from these approaches indicated the following points. (1) Histone acetylation, regardless of histone tail residues, invariably correlates with active gene expression. (2) Different HATs and acetylated histone residues are simultaneously enriched primarily at TSS and promoters of active genes (H2AK9ac, H2BK5ac, H3K9ac, H3K18ac, H3K27ac, H3K36ac, and H4K91ac), while others can extend concurrently to the coding sequence (H2BK12ac, H2BK20ac, H2BK120ac, H3K4ac, H4K5ac, H4K8ac, H4K12ac, and H4K16ac) [251,252]. These data suggest that different histone acetylation forms are redundant as they overlap. This apparent contradiction could be explained by a two-stage model: where one HAT sets a specific acetylation mark needed for transcription “initiation.” This creates an open chromatin context that allows the recruitment of other HATs and the deposition of other acetylation marks that allow the “maintenance” of transcription [253]. Although these evidences strongly argue for a correlation between histone acetylation and transcription activation, they do not provide a causal link between these two events. Even genetic experiments involving HAT ablation do not represent a formal proof of the role of histone acetylation in transcription regulation, particularly considering that these do not account for the activity of HATs on nonhistone proteins. On the other hand, experiments involving direct mutations of acetylatable histone residues in yeast [254] showed that, with the exception of H4K16, all acetylatable lysines on H4 are not essential for transcriptional regulation when mutated individually. This points again to a redundant nonessential function of most histone acetylation marks individually. These and other observations led to the proposal of

a model postulating that the combination of different histone modifications and other epigenetic marks may affect transcription indirectly by affecting the nucleosome occupancy [255]. However, Kimura and coworkers recently developed a fluorescent microscopy–based approach to follow H3K27ac’s effect on PolII activity in vivo with high temporal resolution. Results showed that H3K27ac is essential for PolII recruitment, transcription initiation, and elongation [256]. In conclusion, despite the plethora of research on histone acetylation, the exact mechanism by which it contributes to gene regulation is far from being understood. It is hoped that the recent emergence of a new generation of genome-editing tools will facilitate the addressing of this long-standing issue. Role of Histone Acetylation in DNA Repair A large body of evidence comforts the notion that HATs and histone acetylation, besides their role in transcription regulation, belong to a broad repertoire of histone modifications involved in DDR and DNA repair (Table 3.1). Early work on S. cerevisiae showed that mutation of any of the four N-terminal lysines of H4, target substrates for acetylation by Esa1 (catalytic component of the NuA4 complex), abolished repair of both DSB and UV lesions [257]. Tip60 (mammalian homolog of Esa1) was also shown to be important in DSB repair following genotoxic stress [225]. In addition, mutations in Yng2, another component of the NuA4 complex, results in hypersensitivity to and inefficient repair of DNA damage caused by replication fork stall [258] Further work revealed that N-terminal deletion of H3, mutations in specific H3 lysine residues, or mutations in H3 acetyltransferase HAT1 result in defects in DSB repair [233,259]. It was also shown that GCN5 is recruited to DSBs and that GCN5 deletion mutants are incompatible with viability when subjected to induction of a single DSB [260]. TATA box–binding protein–free TAFII (TFTC), a complex containing Gcn5 HAT, was found to preferentially acetylate H3 in nucleosomes containing UV-damaged DNA in mammalian cells [261], whereas STAGA, another Gcn5containing HAT complex, associates with UV damage– binding factor [262]. Mechanistic data for the role of acetylation in DNA repair has arisen from several reports on yeast and mammalian models. Binding of the NuA4 HAT complex at sites of DNA damage and site-specific histone H4 acetylation were found to occur concomitantly with H2A phosphorylation after induction of DSBs [86,263]. Furthermore, localized histone H3 and H4 acetylation and deacetylation is triggered by homology directed repair of DSBs. Consistent with this finding, Gcn5 and Esa1 HATs are recruited to chromatin around a DSB induced by HO endonuclease in yeast [260]. Alongside histone modification on the amino-terminal tails of histones, histone

II.  Molecular Mechanisms of Epigenetics



39

REFERENCES

core modifications also play a role in DNA repair. For instance, in budding yeast, H3K56 is acetylated on newly synthesized histones during S phase and H3K56ac disappears in G2. However, in the presence of DNA damage the deacetylases for H3K56, Hst3, and Hst4 (two paralogs of Sir2) are downregulated and the modification persists [264,265]. Moreover, the Rtt109 enzyme that acetylates H3K56, has recently been implicated in genome stability and DNA replication [234,266,267]. Furthermore, studies in mammalian cells provided evidence that histone acetylation by TRRAP/TIP60 HAT is important for recruitment/loading of repair proteins to sites of DNA DSBs and homology-directed DNA repair [268]. These findings suggest a model in which induction of DSBs leads to the recruitment of the TIP60/NuA4 complex to DSBs and concomitant acetylation of H4 Nterminal tails. These events result in the opening of the chromatin structure, facilitating the recruitment of DNA repair proteins [269,270]. Surprisingly, it was shown that transient H4 deacetylation is an early response to DNA damage that facilitates DSB repair by NHEJ pathway, as well as repression of transcription [200,271]. This somewhat unexpected finding urges for a more precise determination of the order of events in response to DNA damage. Various studies identified crosstalks between histone acetylation and several other DNA damage-induced histone modifications and chromatin-remodeling events [272–274]. For instance, following DSB, HP1-β is displaced from the chromatin, making H3K9me3 accessible. TIP60 is then tyrosine phosphorylated, which promotes its binding, via its chromodomain, to H3K9me3 marks [275]. This triggers TIP60-mediated acetylation and activation of the ATM kinase [276], thus promoting cell cycle checkpoint activation and cell survival [277]. Besides its role in ATM activation, TIP60-mediated histone acetylation is involved in dephosphorylation of γH2AX. The work from Workman’s laboratory showed that Drosophila TIP60 may acetylate phosphorylated H2Av (the fly ortholog of H2AX), an event resulting in exchange of phosphorylated histone with an unmodified equivalent [226]. In human cells, TIP60 was found to be involved in DNA damage–dependent acetylation of γ-H2AX at K5 that promotes its ubiquitination. The sequential acetylation and ubiquitination then enhances histone dynamics during DDR [278]. Furthermore, Jha et al. showed that H4 acetylation by human TIP60 HAT is required to remodel phosphoH2AX-containing nucleosome at sites of DNA breaks [279]. Lee et al. showed that BRG1, the catalytic subunit of SWI/SNF chromatinremodeling complex, binds to γ-H2AX nucleosomes by interacting with acetylated H3, independent of H2AX serine-139 phosphorylation, through its bromodomain. The authors suggested that histone H3 acetylation and H2AX phosphorylation cooperatively act in a feedback activation loop involved in DNA repair [280]. Together,

data from different model systems (including mammalian cells and yeast) clearly show that histone acetylation by HAT complexes are required for efficient repair of DNA breaks. In addition, these findings demonstrate an important role of the crosstalk between histone acetylation and other histone modifications in regulation of the DSB response.

CONCLUSIONS Cellular processes, such as transcription, DNA replication, and DNA repair, are regulated by an intimate and self-reinforcing crosstalk and interdependence between histone-modifying complexes and other histone-modifying activities, such as acetylation, phosphorylation, and methylation. Consistent with the critical function of histone modifications in key cellular processes, a large body of evidence has suggested that aberrations in these modifications are intimately linked to human pathologies. Most notably, recent genetic and molecular studies have directly implicated histone modifications and histone-modifying complexes in human cancer [281,282]. The fact that epigenetic alterations are, in contrast to genetic changes, reversible has important implications for human cancer treatment, as aberrant histone modifications are potential molecular targets for therapeutic intervention in human malignancies.

References [1] Kouzarides T. Chromatin modifications and their function. Cell 2007;128:693–705. [2] Strahl BD, Allis CD. The language of covalent histone modifications. Nature 2000;403:41–5. [3] Turner BM. Cellular memory and the histone code. Cell 2002;111:285–91. [4] Allis CD, Berger SL, Cote J, Dent S, Jenuwien T, Kouzarides T, et al. New nomenclature for chromatin-modifying enzymes. Cell 2007;131:633–6. [5] Ramachandran GN, Venkatachalam CM. Stereochemical criteria for polypeptides and proteins. IV. Standard dimensions for the cis-peptide unit and conformation of cis-polypeptides. Biopolymers 1968;6:1255–62. [6] Nelson CJ, Santos-Rosa H, Kouzarides T. Proline isomerization of histone H3 regulates lysine methylation and gene expression. Cell 2006;126:905–16. [7] Howe FS, Boubriak I, Sale MJ, Nair A, Clynes D, Grijzenhout A, et al. Lysine acetylation controls local protein conformation by influencing proline isomerization. Mol Cell 2014;55:733–44. [8] Cherrington BD, Morency E, Struble AM, Coonrod SA, Wakshlag JJ. Potential role for peptidylarginine deiminase 2 (PAD2) in citrullination of canine mammary epithelial cell histones. PLoS One 2010;5:e11768. [9] Nakashima K, Hagiwara T, Yamada M. Nuclear localization of peptidylarginine deiminase V and histone deimination in granulocytes. J Biol Chem 2002;277:49562–8. [10] Zhang X, Bolt M, Guertin MJ, Chen W, Zhang S, Cherrington BD, et al. Peptidylarginine deiminase 2-catalyzed histone H3 arginine 26 citrullination facilitates estrogen receptor α target gene activation. Proc Natl Acad Sci USA 2012;109:13331–6.

II.  Molecular Mechanisms of Epigenetics

40

3.  Mechanisms of Histone Modifications

[11] Wang Y, Wysocka J, Sayegh J, Lee YH, Perlin JR, Leonelli L, et al. Human PAD4 regulates histone arginine methylation levels via demethylimination. Science 2004;306:279–83. [12] Christophorou MA, Castelo-Branco G, Halley-Stott RP, Oliveira CS, Loos R, Radzisheuskaya A, et al. Citrullination regulates pluripotency and histone H1 binding to chromatin. Nature 2014;507:104–8. [13] Sharma P, Azebi S, England P, Christensen T, Moller-Larsen A, Petersen T, et al. Citrullination of histone H3 interferes with HP1-mediated transcriptional repression. PLoS Genet 2012;8:e1002934. [14] Cuthbert GL, Daujat S, Snowden AW, Erdjument-Bromage H, Hagiwara T, Yamada M, et al. Histone deimination antagonizes arginine methylation. Cell 2004;118:545–53. [15] Tanikawa C, Espinosa M, Suzuki A, Masuda K, Yamamoto K, Tsuchiya E, et al. Regulation of histone modification and chromatin structure by the p53-PADI4 pathway. Nat Commun 2012;3:676. [16] Goldknopf IL, Taylor CW, Baum RM, Yeoman LC, Olson MO, Prestayko AW, et al. Isolation and characterization of protein A24, a “histone-like” non-histone chromosomal protein. J Biol Chem 1975;250:7182–7. [17] West MH, Bonner WM. Histone 2B can be modified by the attachment of ubiquitin. Nucleic Acids Res 1980;8:4671–80. [18] Geng F, Tansey WP. Polyubiquitylation of histone H2B. Mol Biol Cell 2008;19:3616–24. [19] Zhou W, Wang X, Rosenfeld MG. Histone H2A ubiquitination in transcriptional regulation and DNA damage repair. Int J Biochem Cell Biol 2009;41:12–5. [20] Buchwald G, van der Stoop P, Weichenrieder O, Perrakis A, van Lohuizen M, Sixma TK. Structure and E3-ligase activity of the Ring-Ring complex of polycomb proteins Bmi1 and Ring1b. EMBO J 2006;25:2465–74. [21] Cao R, Tsukada Y, Zhang Y. Role of Bmi-1 and Ring1A in H2A ubiquitylation and Hox gene silencing. Mol Cell 2005;20:845–54. [22] Tavares L, Dimitrova E, Oxley D, Webster J, Poot R, Demmers J, et al. RYBP-PRC1 complexes mediate H2A ubiquitylation at polycomb target sites independently of PRC2 and H3K27me3. Cell 2012;148:664–78. [23] Blackledge NP, Farcas AM, Kondo T, King HW, McGouran JF, Hanssen LL, et al. Variant PRC1 complex-dependent H2A ubiquitylation drives PRC2 recruitment and polycomb domain formation. Cell 2014;157:1445–59. [24] Lagarou A, Mohd-Sarip A, Moshkin YM, Chalkley GE, Bezstarosti K, Demmers JA, et al. dKDM2 couples histone H2A ubiquitylation to histone H3 demethylation during Polycomb group silencing. Genes Dev 2008;22:2799–810. [25] Robzyk K, Recht J, Osley MA. Rad6-dependent ubiquitination of histone H2B in yeast. Science 2000;287:501–4. [26] Kim J, Hake SB, Roeder RG. The human homolog of yeast BRE1 functions as a transcriptional coactivator through direct activator interactions. Mol Cell 2005;20:759–70. [27] Koken MH, Reynolds P, Jaspers-Dekker I, Prakash L, Prakash S, Bootsma D, et al. Structural and functional conservation of two human homologs of the yeast DNA repair gene RAD6. Proc Natl Acad Sci USA 1991;88:8865–9. [28] Zhang Y. Transcriptional regulation by histone ubiquitination and deubiquitination. Genes Dev 2003;17:2733–40. [29] Daniel JA, Torok MS, Sun ZW, Schieltz D, Allis CD, Yates JR 3rd, et al. Deubiquitination of histone H2B by a yeast acetyltransferase complex regulates transcription. J Biol Chem 2004;279:1867–71. [30] Henry KW, Wyce A, Lo WS, Duggan LJ, Emre NC, Kao CF, et al. Transcriptional activation via sequential histone H2B ubiquitylation and deubiquitylation, mediated by SAGA-associated Ubp8. Genes Dev 2003;17:2648–63. [31] Emre NC, Ingvarsdottir K, Wyce A, Wood A, Krogan NJ, Henry KW, et al. Maintenance of low histone ubiquitylation by Ubp10

[32]

[33]

[34] [35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

[47]

[48]

[49]

correlates with telomere-proximal Sir2 association and gene silencing. Mol Cell 2005;17:585–94. Richardson LA, Reed BJ, Charette JM, Freed EF, Fredrickson EK, Locke MN, et al. A conserved deubiquitinating enzyme controls cell growth by regulating RNA polymerase I stability. Cell Rep 2012;2:372–85. Schulze JM, Hentrich T, Nakanishi S, Gupta A, Emberly E, Shilatifard A, et al. Splitting the task: Ubp8 and Ubp10 deubiquitinate different cellular pools of H2BK123. Genes Dev 2011;25:2242–7. Cao J, Yan Q. Histone ubiquitination and deubiquitination in transcription, DNA damage response, and cancer. Front Oncol 2012;2:26. Dawson BA, Herman T, Haas AL, Lough J. Affinity isolation of active murine erythroleukemia cell chromatin: uniform distribution of ubiquitinated histone H2A between active and inactive fractions. J Cell Biochem 1991;46:166–73. Nickel BE, Allis CD, Davie JR. Ubiquitinated histone H2B is preferentially located in transcriptionally active chromatin. Biochemistry 1989;28:958–63. Parlow MH, Haas AL, Lough J. Enrichment of ubiquitinated histone H2A in a low salt extract of micrococcal nuclease-digested myotube nuclei. J Biol Chem 1990;265:7507–12. Dover J, Schneider J, Tawiah-Boateng MA, Wood A, Dean K, Johnston M, et al. Methylation of histone H3 by COMPASS requires ubiquitination of histone H2B by Rad6. J Biol Chem 2002;277:28368–71. Zhu B, Zheng Y, Pham AD, Mandal SS, Erdjument-Bromage H, Tempst P, et al. Monoubiquitination of human histone H2B: the factors involved and their roles in HOX gene regulation. Mol Cell 2005;20:601–11. McGinty RK, Kim J, Chatterjee C, Roeder RG, Muir TW. Chemically ubiquitylated histone H2B stimulates hDot1L-mediated intranucleosomal methylation. Nature 2008;453:812–6. Fleming AB, Kao CF, Hillyer C, Pikaart M, Osley MA. H2B ubiquitylation plays a role in nucleosome dynamics during transcription elongation. Mol Cell 2008;31:57–66. Pavri R, Zhu B, Li G, Trojer P, Mandal S, Shilatifard A, et al. Histone H2B monoubiquitination functions cooperatively with FACT to regulate elongation by RNA polymerase II. Cell 2006;125:703–17. Nakagawa T, Kajitani T, Togo S, Masuko N, Ohdan H, Hishikawa Y, et al. Deubiquitylation of histone H2A activates transcriptional initiation via trans-histone cross-talk with H3K4 di- and trimethylation. Genes Dev 2008;22:37–49. Briggs SD, Bryk M, Strahl BD, Cheung WL, Davie JK, Dent SY, et al. Histone H3 lysine 4 methylation is mediated by Set1 and required for cell growth and rDNA silencing in Saccharomyces cerevisiae. Genes Dev 2001;15:3286–95. Ng HH, Xu RM, Zhang Y, Struhl K. Ubiquitination of histone H2B by Rad6 is required for efficient Dot1-mediated methylation of histone H3 lysine 79. J Biol Chem 2002;277:34655–7. Krogan NJ, Kim M, Tong A, Golshani A, Cagney G, Canadien V, et al. Methylation of histone H3 by Set2 in Saccharomyces cerevisiae is linked to transcriptional elongation by RNA polymerase II. Mol Cell Biol 2003;23:4207–18. Ng HH, Robert F, Young RA, Struhl K. Targeted recruitment of Set1 histone methylase by elongating Pol II provides a localized mark and memory of recent transcriptional activity. Mol Cell 2003;11:709–19. Shema E, Tirosh I, Aylon Y, Huang J, Ye C, Moskovits N, et al. The histone H2B-specific ubiquitin ligase RNF20/hBRE1 acts as a putative tumor suppressor through selective regulation of gene expression. Genes Dev 2008;22:2664–76. Jung I, Kim SK, Kim M, Han YM, Kim YS, Kim D, et al. H2B monoubiquitylation is a 5′-enriched active transcription mark and correlates with exon-intron structure in human cells. Genome Res 2012;22:1026–35.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[50] Kim J, Guermah M, McGinty RK, Lee JS, Tang Z, Milne TA, et al. RAD6-Mediated transcription-coupled H2B ubiquitylation directly stimulates H3K4 methylation in human cells. Cell 2009;137:459–71. [51] Ginjala V, Nacerddine K, Kulkarni A, Oza J, Hill SJ, Yao M, et al. BMI1 is recruited to DNA breaks and contributes to DNA damage-induced H2A ubiquitination and repair. Mol Cell Biol 2011;31:1972–82. [52] Moyal L, Lerenthal Y, Gana-Weisz M, Mass G, So S, Wang SY, et al. Requirement of ATM-dependent monoubiquitylation of histone H2B for timely repair of DNA double-strand breaks. Mol Cell 2011;41:529–42. [53] Pan MR, Peng G, Hung WC, Lin SY. Monoubiquitination of H2AX protein regulates DNA damage response signaling. J Biol Chem 2011;286:28599–607. [54] Kim H, Chen J, Yu X. Ubiquitin-binding protein RAP80 mediates BRCA1-dependent DNA damage response. Science 2007;316:1202–5. [55] Sobhian B, Shao G, Lilli DR, Culhane AC, Moreau LA, Xia B, et al. RAP80 targets BRCA1 to specific ubiquitin structures at DNA damage sites. Science 2007;316:1198–202. [56] Thorslund T, Ripplinger A, Hoffmann S, Wild T, Uckelmann M, Villumsen B, et al. Histone H1 couples initiation and amplification of ubiquitin signalling after DNA damage. Nature 2015;527:389–93. [57] Eisenhardt N, Chaugule VK, Koidl S, Droescher M, Dogan E, Rettich J, et al. A new vertebrate SUMO enzyme family reveals insights into SUMO-chain assembly. Nat Struct Mol Biol 2015;22:959–67. [58] Shiio Y, Eisenman RN. Histone sumoylation is associated with transcriptional repression. Proc Natl Acad Sci USA 2003;100:13225–30. [59] Nathan D, Ingvarsdottir K, Sterner DE, Bylebyl GR, Dokmanovic M, Dorsey JA, et al. Histone sumoylation is a negative regulator in Saccharomyces cerevisiae and shows dynamic interplay with positive-acting histone modifications. Genes Dev 2006;20:966–76. [60] Hendriks IA, D’Souza RC, Yang B, Verlaan-de Vries M, Mann M, Vertegaal AC. Uncovering global SUMOylation signaling networks in a site-specific manner. Nat Struct Mol Biol 2014;21:927–36. [61] Hottiger MO, Hassa PO, Luscher B, Schuler H, Koch-Nolte F. Toward a unified nomenclature for mammalian ADP-ribosyltransferases. Trends Biochem Sci 2010;35:208–19. [62] Hottiger MO. Nuclear ADP-ribosylation and its role in chromatin plasticity, cell differentiation, and epigenetics. Annu Rev Biochem 2015;84:227–63. [63] Kreimeyer A, Adamietz P, Hilz H. Alkylation-induced mono(ADP-ribosyl)-histones H1 and H2B. Hydroxylamineresistant linkage in hepatoma cells. Biol Chem Hoppe Seyler 1985;366:537–44. [64] Kreimeyer A, Wielckens K, Adamietz P, Hilz H. DNA repairassociated ADP-ribosylation in vivo. Modification of histone H1 differs from that of the principal acceptor proteins. J Biol Chem 1984;259:890–6. [65] Kraus WL. Transcriptional control by PARP-1: chromatin modulation, enhancer-binding, coregulation, and insulation. Curr Opin Cell Biol 2008;20:294–302. [66] Kraus WL, Hottiger MO. PARP-1 and gene regulation: progress and puzzles. Mol Aspects Med 2013;34:1109–23. [67] Timinszky G, Till S, Hassa PO, Hothorn M, Kustatscher G, Nijmeijer B, et al. A macrodomain-containing histone rearranges chromatin upon sensing PARP1 activation. Nat Struct Mol Biol 2009;16:923–9. [68] Wei Y, Mizzen CA, Cook RG, Gorovsky MA, Allis CD. Phosphorylation of histone H3 at serine 10 is correlated with chromosome condensation during mitosis and meiosis in Tetrahymena. Proc Natl Acad Sci USA 1998;95:7480–4.

41

[69] Fischle W, Tseng BS, Dormann HL, Ueberheide BM, Garcia BA, Shabanowitz J, et al. Regulation of HP1-chromatin binding by histone H3 methylation and phosphorylation. Nature 2005;438:1116–22. [70] Mahadevan LC, Willis AC, Barratt MJ. Rapid histone H3 phosphorylation in response to growth factors, phorbol esters, okadaic acid, and protein synthesis inhibitors. Cell 1991;65:775–83. [71] Nowak SJ, Corces VG. Phosphorylation of histone H3: a balancing act between chromosome condensation and transcriptional activation. Trends Genet 2004;20:214–20. [72] Macdonald N, Welburn JP, Noble ME, Nguyen A, Yaffe MB, Clynes D, et al. Molecular basis for the recognition of phosphorylated and phosphoacetylated histone h3 by 14-3-3. Mol Cell 2005;20:199–211. [73] Pokholok DK, Zeitlinger J, Hannett NM, Reynolds DB, Young RA. Activated signal transduction kinases frequently occupy target genes. Science 2006;313:533–6. [74] Lo WS, Trievel RC, Rojas JR, Duggan L, Hsu JY, Allis CD, et al. Phosphorylation of serine 10 in histone H3 is functionally linked in vitro and in vivo to Gcn5-mediated acetylation at lysine 14. Mol Cell 2000;5:917–26. [75] Lo WS, Duggan L, Emre NC, Belotserkovskya R, Lane WS, Shiekhattar R, et al. Snf1--a histone kinase that works in concert with the histone acetyltransferase Gcn5 to regulate transcription. Science 2001;293:1142–6. [76] Cheung P, Tanner KG, Cheung WL, Sassone-Corsi P, Denu JM, Allis CD. Synergistic coupling of histone H3 phosphorylation and acetylation in response to epidermal growth factor stimulation. Mol Cell 2000;5:905–15. [77] Walter W, Clynes D, Tang Y, Marmorstein R, Mellor J, Berger SL. 14-3-3 interaction with histone H3 involves a dual modification pattern of phosphoacetylation. Mol Cell Biol 2008;28:2840–9. [78] Labrador M, Corces VG. Phosphorylation of histone H3 during transcriptional activation depends on promoter structure. Genes Dev 2003;17:43–8. [79] Kellner WA, Ramos E, Van Bortle K, Takenaka N, Corces VG. Genome-wide phosphoacetylation of histone H3 at Drosophila enhancers and promoters. Genome Res 2012;22:1081–8. [80] Kim JY, Banerjee T, Vinckevicius A, Luo Q, Parker JB, Baker MR, et al. A role for WDR5 in integrating threonine 11 phosphorylation to lysine 4 methylation on histone H3 during androgen signaling and in prostate cancer. Mol Cell 2014;54:613–25. [81] Metzger E, Yin N, Wissmann M, Kunowska N, Fischer K, Friedrichs N, et al. Phosphorylation of histone H3 at threonine 11 establishes a novel chromatin mark for transcriptional regulation. Nat Cell Biol 2008;10:53–60. [82] Sawicka A, Hartl D, Goiser M, Pusch O, Stocsits RR, Tamir IM, et al. H3S28 phosphorylation is a hallmark of the transcriptional response to cellular stress. Genome Res 2014;24:1808–20. [83] Cook PJ, Ju BG, Telese F, Wang X, Glass CK, Rosenfeld MG. Tyrosine dephosphorylation of H2AX modulates apoptosis and survival decisions. Nature 2009;458:591–6. [84] Xiao A, Li H, Shechter D, Ahn SH, Fabrizio LA, Erdjument-Bromage H, et al. WSTF regulates the H2A.X. DNA damage response via a novel tyrosine kinase activity. Nature 2009;457:57–62. [85] Shroff R, Arbel-Eden A, Pilch D, Ira G, Bonner WM, Petrini JH, et al. Distribution and dynamics of chromatin modification induced by a defined DNA double-strand break. Curr Biol 2004;14:1703–11. [86] Downs JA, Allard S, Jobin-Robitaille O, Javaheri A, Auger A, Bouchard N, et al. Binding of chromatin-modifying activities to phosphorylated histone H2A at DNA damage sites. Mol Cell 2004;16:979–90. [87] Strom L, Lindroos HB, Shirahige K, Sjogren C. Postreplicative recruitment of cohesin to double-strand breaks is required for DNA repair. Mol Cell 2004;16:1003–15.

II.  Molecular Mechanisms of Epigenetics

42

3.  Mechanisms of Histone Modifications

[88] Unal E, Arbel-Eden A, Sattler U, Shroff R, Lichten M, Haber JE, et al. DNA damage response pathway uses histone modification to assemble a double-strand break-specific cohesin domain. Mol Cell 2004;16:991–1002. [89] Kobayashi J, Tauchi H, Sakamoto S, Nakamura A, Morishima K, Matsuura S, et al. NBS1 localizes to gamma-H2AX foci through interaction with the FHA/BRCT domain. Curr Biol 2002;12:1846–51. [90] Celeste A, Fernandez-Capetillo O, Kruhlak MJ, Pilch DR, Staudt DW, Lee A, et al. Histone H2AX phosphorylation is dispensable for the initial recognition of DNA breaks. Nat Cell Biol 2003;5:675–9. [91] Cheung WL, Turner FB, Krishnamoorthy T, Wolner B, Ahn SH, Foley M, et al. Phosphorylation of histone H4 serine 1 during DNA damage requires casein kinase II in S. cerevisiae. Curr Biol 2005;15:656–60. [92] Utley RT, Lacoste N, Jobin-Robitaille O, Allard S, Cote J. Regulation of NuA4 histone acetyltransferase activity in transcription and DNA repair by phosphorylation of histone H4. Mol Cell Biol 2005;25:8179–90. [93] Kysela B, Chovanec M, Jeggo PA. Phosphorylation of linker histones by DNA-dependent protein kinase is required for DNA ligase IV-dependent ligation in the presence of histone H1. Proc Natl Acad Sci USA 2005;102:1877–82. [94] Clarke S. Protein methylation. Curr Opin Cell Biol 1993;5:977–83. [95] Tschiersch B, Hofmann A, Krauss V, Dorn R, Korge G, Reuter G. The protein encoded by the Drosophila position-effect variegation suppressor gene Su(var)3-9 combines domains of antagonistic regulators of homeotic gene complexes. EMBO J 1994;13:3822–31. [96] Spellmon N, Holcomb J, Trescott L, Sirinupong N, Yang Z. Structure and function of SET and MYND domain-containing proteins. Int J Mol Sci 2015;16:1406–28. [97] Schuhmacher MK, Kudithipudi S, Kusevic D, Weirich S, Jeltsch A. Activity and specificity of the human SUV39H2 protein lysine methyltransferase. Biochim Biophys Acta 2015;1849:55–63. [98] Kudithipudi S, Lungu C, Rathert P, Happel N, Jeltsch A. Substrate specificity analysis and novel substrates of the protein lysine methyltransferase NSD1. Chem Biol 2014;21:226–37. [99] Dorighi KM, Tamkun JW. The trithorax group proteins Kismet and ASH1 promote H3K36 dimethylation to counteract Polycomb group repression in Drosophila. Development 2013;140:4182–92. [100] Kizer KO, Phatnani HP, Shibata Y, Hall H, Greenleaf AL, Strahl BD. A novel domain in Set2 mediates RNA polymerase II interaction and couples histone H3 K36 methylation with transcript elongation. Mol Cell Biol 2005;25:3305–16. [101] Zhang Y, Xie S, Zhou Y, Xie Y, Liu P, Sun M, et al. H3K36 histone methyltransferase Setd2 is required for murine embryonic stem cell differentiation toward endoderm. Cell Rep 2014;8:1989– 2002. [102] Gu B, Lee MG. Histone H3 lysine 4 methyltransferases and demethylases in self-renewal and differentiation of stem cells. Cell Biosci 2013;3:39. [103] Feng Q, Wang H, Ng HH, Erdjument-Bromage H, Tempst P, Struhl K, et al. Methylation of H3-lysine 79 is mediated by a new family of HMTases without a SET domain. Curr Biol 2002;12:1052–8. [104] Di Lorenzo A, Bedford MT. Histone arginine methylation. FEBS Lett 2011;585:2024–31. [105] Metzger E, Wissmann M, Yin N, Muller JM, Schneider R, Peters AH, et al. LSD1 demethylates repressive histone marks to promote androgen-receptor-dependent transcription. Nature 2005;437:436–9. [106] Shi Y, Lan F, Matson C, Mulligan P, Whetstine JR, Cole PA, et al. Histone demethylation mediated by the nuclear amine oxidase homolog LSD1. Cell 2004;119:941–53.

[107] Tsukada Y, Fang J, Erdjument-Bromage H, Warren ME, Borchers CH, Tempst P, et al. Histone demethylation by a family of JmjC domain-containing proteins. Nature 2006;439:811–6. [108] Whetstine JR, Nottke A, Lan F, Huarte M, Smolikov S, Chen Z, et al. Reversal of histone lysine trimethylation by the JMJD2 family of histone demethylases. Cell 2006;125:467–81. [109] Mantri M, Krojer T, Bagg EA, Webby CA, Butler DS, Kochan G, et al. Crystal structure of the 2-oxoglutarate- and Fe(II)-dependent lysyl hydroxylase JMJD6. J Mol Biol 2010;401:211–22. [110] Webby CJ, Wolf A, Gromak N, Dreger M, Kramer H, Kessler B, et al. Jmjd6 catalyses lysyl-hydroxylation of U2AF65, a protein associated with RNA splicing. Science 2009;325:90–3. [111] Bernstein BE, Humphrey EL, Erlich RL, Schneider R, Bouman P, Liu JS, et al. Methylation of histone H3 Lys 4 in coding regions of active genes. Proc Natl Acad Sci USA 2002;99:8695–700. [112] Santos-Rosa H, Schneider R, Bannister AJ, Sherriff J, Bernstein BE, Emre NC, et al. Active genes are tri-methylated at K4 of histone H3. Nature 2002;419:407–11. [113] Li B, Howe L, Anderson S, Yates JR 3rd, Workman JL. The Set2 histone methyltransferase functions through the phosphorylated carboxyl-terminal domain of RNA polymerase II. J Biol Chem 2003;278:8897–903. [114] Xiao B, Wilson JR, Gamblin SJ. SET domains and histone methylation. Curr Opin Struct Biol 2003;13:699–705. [115] Vakoc CR, Sachdeva MM, Wang H, Blobel GA. Profile of histone lysine methylation across transcribed mammalian chromatin. Mol Cell Biol 2006;26:9185–95. [116] Barski A, Cuddapah S, Cui K, Roh TY, Schones DE, Wang Z, et al. High-resolution profiling of histone methylations in the human genome. Cell 2007;129:823–37. [117] Bernstein BE, Mikkelsen TS, Xie X, Kamal M, Huebert DJ, Cuff J, et al. A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 2006;125:315–26. [118] Bannister AJ, Schneider R, Myers FA, Thorne AW, Crane-Robinson C, Kouzarides T. Spatial distribution of di- and tri-methyl lysine 36 of histone H3 at active genes. J Biol Chem 2005;280:17732– 6. [119] Bell O, Wirbelauer C, Hild M, Scharf AN, Schwaiger M, MacAlpine DM, et al. Localized H3K36 methylation states define histone H4K16 acetylation during transcriptional elongation in Drosophila. EMBO J 2007;26:4974–84. [120] Lee JS, Shilatifard A. A site to remember: H3K36 methylation a mark for histone deacetylation. Mutat Res 2007;618:130–4. [121] Mikkelsen TS, Ku M, Jaffe DB, Issac B, Lieberman E, Giannoukos G, et al. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature 2007;448:553–60. [122] Pokholok DK, Harbison CT, Levine S, Cole M, Hannett NM, Lee TI, et al. Genome-wide map of nucleosome acetylation and methylation in yeast. Cell 2005;122:517–27. [123] Rao B, Shibata Y, Strahl BD, Lieb JD. Dimethylation of histone H3 at lysine 36 demarcates regulatory and nonregulatory chromatin genome-wide. Mol Cell Biol 2005;25:9447–59. [124] Djebali S, Davis CA, Merkel A, Dobin A, Lassmann T, Mortazavi A, et al. Landscape of transcription in human cells. Nature 2012;489:101–8. [125] Huff JT, Plocik AM, Guthrie C, Yamamoto KR. Reciprocal intronic and exonic histone modification regions in humans. Nat Struct Mol Biol 2010;17:1495–9. [126] Fu H, Maunakea AK, Martin MM, Huang L, Zhang Y, Ryan M, et al. Methylation of histone H3 on lysine 79 associates with a group of replication origins and helps limit DNA replication once per cell cycle. PLoS Genet 2013;9:e1003542. [127] Creyghton MP, Cheng AW, Welstead GG, Kooistra T, Carey BW, Steine EJ, et al. Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proc Natl Acad Sci USA 2010;107:21931–6.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[128] Heintzman ND, Hon GC, Hawkins RD, Kheradpour P, Stark A, Harp LF, et al. Histone modifications at human enhancers reflect global cell-type-specific gene expression. Nature 2009;459:108–12. [129] Heintzman ND, Stuart RK, Hon G, Fu Y, Ching CW, Hawkins RD, et al. Distinct and predictive chromatin signatures of transcriptional promoters and enhancers in the human genome. Nature Genet 2007;39:311–8. [130] Rada-Iglesias A, Bajpai R, Swigut T, Brugmann SA, Flynn RA, Wysocka J. A unique chromatin signature uncovers early developmental enhancers in humans. Nature 2011;470:279–83. [131] Champagne KS, Kutateladze TG. Structural insight into histone recognition by the ING PHD fingers. Curr Drug Targets 2009;10:432–41. [132] Maurer-Stroh S, Dickens NJ, Hughes-Davies L, Kouzarides T, Eisenhaber F, Ponting CP. The Tudor domain ‘Royal Family’: Tudor, plant Agenet, Chromo, PWWP and MBT domains. Trends Biochem Sci 2003;28:69–74. [133] Kim J, Daniel J, Espejo A, Lake A, Krishna M, Xia L, et al. Tudor, MBT and chromo domains gauge the degree of lysine methylation. EMBO Rep 2006;7:397–403. [134] Lauberth SM, Nakayama T, Wu X, Ferris AL, Tang Z, Hughes SH, et al. H3K4me3 interactions with TAF3 regulate preinitiation complex assembly and selective gene activation. Cell 2013;152:1021–36. [135] Balasubramanian D, Akhtar-Zaidi B, Song L, Bartels CF, Veigl M, Beard L, et al. H3K4me3 inversely correlates with DNA methylation at a large class of non-CpG-island-containing start sites. Genome Med 2012;4:47. [136] Fernandez AF, Bayon GF, Urdinguio RG, Torano EG, Garcia MG, Carella A, et al. H3K4me1 marks DNA regions hypomethylated during aging in human stem and differentiated cells. Genome Res 2015;25:27–40. [137] Stadler MB, Murr R, Burger L, Ivanek R, Lienert F, Scholer A, et al. DNA-binding factors shape the mouse methylome at distal regulatory regions. Nature 2011;480:490–5. [138] Okitsu CY, Hsieh CL. DNA methylation dictates histone H3K4 methylation. Mol Cell Biol 2007;27:2746–57. [139] Clouaire T, Webb S, Bird A. Cfp1 is required for gene expressiondependent H3K4 trimethylation and H3K9 acetylation in embryonic stem cells. Genome Biol 2014;15:451. [140] Clouaire T, Webb S, Skene P, Illingworth R, Kerr A, Andrews R, et al. Cfp1 integrates both CpG content and gene activity for accurate H3K4me3 deposition in embryonic stem cells. Genes Dev 2012;26:1714–28. [141] Thomson JP, Skene PJ, Selfridge J, Clouaire T, Guy J, Webb S, et al. CpG islands influence chromatin structure via the CpGbinding protein Cfp1. Nature 2010;464:1082–6. [142] Xu C, Bian C, Lam R, Dong A, Min J. The structural basis for selective binding of non-methylated CpG islands by the CFP1 CXXC domain. Nat Commun 2011;2:227. [143] Hashimoto H, Vertino PM, Cheng X. Molecular coupling of DNA methylation and histone methylation. Epigenomics 2010;2:657– 69. [144] Long HK, King HW, Patient RK, Odom DT, Klose RJ. Protection of CpG islands from DNA methylation is DNA-encoded and evolutionarily conserved. Nucleic Acids Res 2016;gkw258. [145] Ooi SK, Qiu C, Bernstein E, Li K, Jia D, Yang Z, et al. DNMT3L connects unmethylated lysine 4 of histone H3 to de novo methylation of DNA. Nature 2007;448:714–7. [146] Cheng J, Blum R, Bowman C, Hu D, Shilatifard A, Shen S, et al. A role for H3K4 monomethylation in gene repression and partitioning of chromatin readers. Mol Cell 2014;53:979–92. [147] Carrozza MJ, Li B, Florens L, Suganuma T, Swanson SK, Lee KK, et al. Histone H3 methylation by Set2 directs deacetylation of coding regions by Rpd3S to suppress spurious intragenic transcription. Cell 2005;123:581–92.

43

[148] Joshi AA, Struhl K. Eaf3 chromodomain interaction with methylated H3-K36 links histone deacetylation to Pol II elongation. Mol Cell 2005;20:971–8. [149] Keogh MC, Kurdistani SK, Morris SA, Ahn SH, Podolny V, Collins SR, et al. Cotranscriptional set2 methylation of histone H3 lysine 36 recruits a repressive Rpd3 complex. Cell 2005;123:593–605. [150] Li B, Gogol M, Carey M, Lee D, Seidel C, Workman JL. Combined action of PHD and chromo domains directs the Rpd3S HDAC to transcribed chromatin. Science 2007;316:1050–4. [151] Lorincz MC, Dickerson DR, Schmitt M, Groudine M. Intragenic DNA methylation alters chromatin structure and elongation efficiency in mammalian cells. Nat Struct Mol Biol 2004;11:1068–75. [152] Dhayalan A, Rajavelu A, Rathert P, Tamas R, Jurkowska RZ, Ragozin S, et al. The Dnmt3a PWWP domain reads histone 3 lysine 36 trimethylation and guides DNA methylation. J Biol Chem 2010;285:26114–20. [153] Baubec T, Colombo DF, Wirbelauer C, Schmidt J, Burger L, Krebs AR, et al. Genomic profiling of DNA methyltransferases reveals a role for DNMT3B in genic methylation. Nature 2015;520:243–7. [154] Hahn MA, Wu X, Li AX, Hahn T, Pfeifer GP. Relationship between gene body DNA methylation and intragenic H3K9me3 and H3K36me3 chromatin marks. PLoS One 2011;6:e18844. [155] Steger DJ, Lefterova MI, Ying L, Stonestrom AJ, Schupp M, Zhuo D, et al. DOT1L/KMT4 recruitment and H3K79 methylation are ubiquitously coupled with gene transcription in mammalian cells. Mol Cell Biol 2008;28:2825–39. [156] Schubeler D, MacAlpine DM, Scalzo D, Wirbelauer C, Kooperberg C, van Leeuwen F, et al. The histone modification pattern of active genes revealed through genome-wide chromatin analysis of a higher eukaryote. Genes Dev 2004;18:1263–71. [157] McEwen KR, Ferguson-Smith AC. Distinguishing epigenetic marks of developmental and imprinting regulation. Epigenetics Chromatin 2010;3:2. [158] Richards EJ, Elgin SC. Epigenetic codes for heterochromatin formation and silencing: rounding up the usual suspects. Cell 2002;108:489–500. [159] Saksouk N, Simboeck E, Dejardin J. Constitutive heterochromatin formation and transcription in mammals. Epigenetics Chromatin 2015;8:3. [160] Rajakumara E, Law JA, Simanshu DK, Voigt P, Johnson LM, Reinberg D, et al. A dual flip-out mechanism for 5mC recognition by the Arabidopsis SUVH5 SRA domain and its impact on DNA methylation and H3K9 dimethylation in vivo. Genes Dev 2011;25:137–52. [161] Lehnertz B, Ueda Y, Derijck AA, Braunschweig U, Perez-Burgos L, Kubicek S, et al. Suv39h-mediated histone H3 lysine 9 methylation directs DNA methylation to major satellite repeats at pericentric heterochromatin. Curr Biol 2003;13:1192–200. [162] Vakoc CR, Mandat SA, Olenchock BA, Blobel GA. Histone H3 lysine 9 methylation and HP1gamma are associated with transcription elongation through mammalian chromatin. Mol Cell 2005;19:381–91. [163] Boyer LA, Plath K, Zeitlinger J, Brambrink T, Medeiros LA, Lee TI, et al. Polycomb complexes repress developmental regulators in murine embryonic stem cells. Nature 2006;441:349–53. [164] Lee TI, Jenner RG, Boyer LA, Guenther MG, Levine SS, Kumar RM, et al. Control of developmental regulators by Polycomb in human embryonic stem cells. Cell 2006;125:301–13. [165] Harikumar A, Meshorer E. Chromatin remodeling and bivalent histone modifications in embryonic stem cells. EMBO Rep 2015;16:1609–19. [166] Mantsoki A, Devailly G, Joshi A. CpG island erosion, polycomb occupancy and sequence motif enrichment at bivalent promoters in mammalian embryonic stem cells. Sci Rep 2015;5:16791. [167] Brinkman AB, Gu H, Bartels SJ, Zhang Y, Matarese F, Simmer F, et al. Sequential ChIP-bisulfite sequencing enables direct ge-

II.  Molecular Mechanisms of Epigenetics

44

[168]

[169]

[170]

[171]

[172]

[173]

[174]

[175]

[176]

[177]

[178]

[179]

[180]

[181]

[182]

[183]

[184]

[185]

3.  Mechanisms of Histone Modifications

nome-scale investigation of chromatin and DNA methylation cross-talk. Genome Res 2012;22:1128–38. Jermann P, Hoerner L, Burger L, Schubeler D. Short sequences can efficiently recruit histone H3 lysine 27 trimethylation in the absence of enhancer activity and DNA methylation. Proc Natl Acad Sci USA 2014;111:E3415–21. Neri F, Incarnato D, Krepelova A, Rapelli S, Pagnani A, Zecchina R, et al. Genome-wide analysis identifies a functional association of Tet1 and Polycomb repressive complex 2 in mouse embryonic stem cells. Genome Biol 2013;14:R91. Vire E, Brenner C, Deplus R, Blanchon L, Fraga M, Didelot C, et al. The Polycomb group protein EZH2 directly controls DNA methylation. Nature 2006;439:871–4. Karlic R, Chung HR, Lasserre J, Vlahovicek K, Vingron M. Histone modification levels are predictive for gene expression. Proc Natl Acad Sci USA 2010;107:2926–31. Talasz H, Lindner HH, Sarg B, Helliger W. Histone H4-lysine 20 monomethylation is increased in promoter and coding regions of active genes and correlates with hyperacetylation. J Biol Chem 2005;280:38814–22. Karachentsev D, Sarma K, Reinberg D, Steward R. PR-Set7-dependent methylation of histone H4 Lys 20 functions in repression of gene expression and is essential for mitosis. Genes Dev 2005;19:431–5. Kohlmaier A, Savarese F, Lachner M, Martens J, Jenuwein T, Wutz A. A chromosomal memory triggered by Xist regulates histone methylation in X inactivation. PLoS Biol 2004;2:E171. Henckel A, Nakabayashi K, Sanz LA, Feil R, Hata K, Arnaud P. Histone methylation is mechanistically linked to DNA methylation at imprinting control regions in mammals. Hum Mol Genet 2009;18:3375–83. Yokoyama Y, Matsumoto A, Hieda M, Shinchi Y, Ogihara E, Hamada M, et al. Loss of histone H4K20 trimethylation predicts poor prognosis in breast cancer and is associated with invasive activity. Breast Cancer Res 2014;16:R66. Hansen KR, Hazan I, Shanker S, Watt S, Verhein-Hansen J, Bahler J, et al. H3K9me-independent gene silencing in fission yeast heterochromatin by Clr5 and histone deacetylases. PLoS Genet 2011;7:e1001268. Tae S, Karkhanis V, Velasco K, Yaneva M, Erdjument-Bromage H, Tempst P, et al. Bromodomain protein 7 interacts with PRMT5 and PRC2, and is involved in transcriptional repression of their target genes. Nucleic Acids Res 2011;39:5424–38. Wang L, Pal S, Sif S. Protein arginine methyltransferase 5 suppresses the transcription of the RB family of tumor suppressors in leukemia and lymphoma cells. Mol Cell Biol 2008;28:6262–77. Chen D, Ma H, Hong H, Koh SS, Huang SM, Schurter BT, et al. Regulation of transcription by a protein methyltransferase. Science 1999;284:2174–7. Jacques SL, Aquino KP, Gureasko J, Boriack-Sjodin PA, Porter Scott M, Copeland RA, et al. CARM1 preferentially methylates H3R17 over H3R26 through a random kinetic mechanism. Biochemistry 2016;55:1635–44. Fuhrmann J, Thompson PR. Protein arginine methylation and citrullination in epigenetic regulation. ACS Chem Biol 2016;11:654–68. Wu J, Cui N, Wang R, Li J, Wong J. A role for CARM1-mediated histone H3 arginine methylation in protecting histone acetylation by releasing corepressors from chromatin. PLoS One 2012;7:e34692. An W, Kim J, Roeder RG. Ordered cooperative functions of PRMT1, p300, and CARM1 in transcriptional activation by p53. Cell 2004;117:735–48. Gao WW, Xiao RQ, Peng BL, Xu HT, Shen HF, Huang MF, et al. Arginine methylation of HSP70 regulates retinoid acidmediated RARbeta2 gene activation. Proc Natl Acad Sci USA 2015;112:E3327–36.

[186] Kim JK, Lim Y, Lee JO, Lee YS, Won NH, Kim H, et al. PRMT4 is involved in insulin secretion via the methylation of histone H3 in pancreatic beta cells. J Mol Endocrinol 2015;54:315–24. [187] Li J, Zhou F, Zhan D, Gao Q, Cui N, Li J, et al. A novel histone H4 arginine 3 methylation-sensitive histone H4 binding activity and transcriptional regulatory function for signal recognition particle subunits SRP68 and SRP72. J Biol Chem 2012;287:40641–51. [188] Chittka A. Dynamic distribution of histone H4 arginine 3 methylation marks in the developing murine cortex. PLoS One 2010;5:e13807. [189] Metivier R, Penot G, Hubner MR, Reid G, Brand H, Kos M, et al. Estrogen receptor-alpha directs ordered, cyclical, and combinatorial recruitment of cofactors on a natural target promoter. Cell 2003;115:751–63. [190] Kangaspeska S, Stride B, Metivier R, Polycarpou-Schwarz M, Ibberson D, Carmouche RP, et al. Transient cyclical methylation of promoter DNA. Nature 2008;452:112–5. [191] Metivier R, Gallais R, Tiffoche C, Le Peron C, Jurkowska RZ, Carmouche RP, et al. Cyclical DNA methylation of a transcriptionally active promoter. Nature 2008;452:45–50. [192] Tarighat SS, Santhanam R, Frankhouser D, Radomska HS, Lai H, Anghelina M, et al. The dual epigenetic role of PRMT5 in acute myeloid leukemia: gene activation and repression via histone arginine methylation. Leukemia 2016;30:789–99. [193] Tsutsui T, Fukasawa R, Shinmyouzu K, Nakagawa R, Tobe K, Tanaka A, et al. Mediator complex recruits epigenetic regulators via its two cyclin-dependent kinase subunits to repress transcription of immune response genes. J Biol Chem 2013;288:20955–65. [194] Cosgrove MS, Boeke JD, Wolberger C. Regulated nucleosome mobility and the histone code. Nat Struct Mol Biol 2004;11:1037–43. [195] Pal S, Vishwanath SN, Erdjument-Bromage H, Tempst P, Sif S. Human SWI/SNF-associated PRMT5 methylates histone H3 arginine 8 and negatively regulates expression of ST7 and NM23 tumor suppressor genes. Mol Cell Biol 2004;24:9630–45. [196] Botuyan MV, Lee J, Ward IM, Kim JE, Thompson JR, Chen J, et al. Structural basis for the methylation state-specific recognition of histone H4-K20 by 53BP1 and Crb2 in DNA repair. Cell 2006;127:1361–73. [197] Sanders SL, Portoso M, Mata J, Bahler J, Allshire RC, Kouzarides T. Methylation of histone H4 lysine 20 controls recruitment of Crb2 to sites of DNA damage. Cell 2004;119:603–14. [198] Huyen Y, Zgheib O, Ditullio RA Jr, Gorgoulis VG, Zacharatos P, Petty TJ, et al. Methylated lysine 79 of histone H3 targets 53BP1 to DNA double-strand breaks. Nature 2004;432:406–11. [199] Ayrapetov MK, Gursoy-Yuzugullu O, Xu C, Xu Y, Price BD. DNA double-strand breaks promote methylation of histone H3 on lysine 9 and transient formation of repressive chromatin. Proc Natl Acad Sci USA 2014;111:9169–74. [200] Tang J, Cho NW, Cui G, Manion EM, Shanbhag NM, Botuyan MV, et al. Acetylation limits 53BP1 association with damaged chromatin to promote homologous recombination. Nat Struct Mol Biol 2013;20:317–25. [201] Allis CD, Chicoine LG, Richman R, Schulman IG. Depositionrelated histone acetylation in micronuclei of conjugating Tetrahymena. Proc Natl Acad Sci USA 1985;82:8048–52. [202] Ruiz-Carrillo A, Wangh LJ, Allfrey VG. Processing of newly synthesized histone molecules. Science 1975;190:117–28. [203] Vetting MW, LP SdC, Yu M, Hegde SS, Magnet S, Roderick SL, et al. Structure and functions of the GNAT superfamily of acetyltransferases. Arch Biochem Biophys 2005;433:212–26. [204] Borrow J, Stanton VP Jr, Andresen JM, Becher R, Behm FG, Chaganti RS, et al. The translocation t(8;16)(p11;p13) of acute myeloid leukaemia fuses a putative acetyltransferase to the CREBbinding protein. Nat Genet 1996;14:33–41. [205] Neuwald AF, Landsman D. GCN5-related histone N-acetyltransferases belong to a diverse superfamily that includes the yeast SPT10 protein. Trends Biochem Sci 1997;22:154–5.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[206] Akhtar A, Becker PB. The histone H4 acetyltransferase MOF uses a C2HC zinc finger for substrate recognition. EMBO Rep 2001;2:113–8. [207] Yan Y, Barlev NA, Haley RH, Berger SL, Marmorstein R. Crystal structure of yeast Esa1 suggests a unified mechanism for catalysis and substrate binding by histone acetyltransferases. Mol Cell 2000;6:1195–205. [208] Doyon Y, Cote J. The highly conserved and multifunctional NuA4 HAT complex. Curr Opin Genet Dev 2004;14:147–54. [209] Doyon Y, Selleck W, Lane WS, Tan S, Cote J. Structural and functional conservation of the NuA4 histone acetyltransferase complex from yeast to humans. Mol Cell Biol 2004;24:1884–96. [210] Schiltz RL, Mizzen CA, Vassilev A, Cook RG, Allis CD, Nakatani Y. Overlapping but distinct patterns of histone acetylation by the human coactivators p300 and PCAF within nucleosomal substrates. J Biol Chem 1999;274:1189–92. [211] Kalkhoven E. CBP and p300: HATs for different occasions. Biochem Pharmacol 2004;68:1145–55. [212] McManus KJ, Hendzel MJ. Quantitative analysis of CBP- and P300-induced histone acetylations in vivo using native chromatin. Mol Cell Biol 2003;23:7611–27. [213] Kawasaki H, Eckner R, Yao TP, Taira K, Chiu R, Livingston DM, et al. Distinct roles of the co-activators p300 and CBP in retinoicacid-induced F9-cell differentiation. Nature 1998;393:284–9. [214] Kung AL, Rebel VI, Bronson RT, Ch’ng LE, Sieff CA, Livingston DM, et al. Gene dose-dependent control of hematopoiesis and hematologic tumor suppression by CBP. Genes Dev 2000;14:272–7. [215] Yao TP, Oh SP, Fuchs M, Zhou ND, Ch’ng LE, Newsome D, et al. Gene dosage-dependent embryonic development and proliferation defects in mice lacking the transcriptional integrator p300. Cell 1998;93:361–72. [216] Das C, Lucia MS, Hansen KC, Tyler JK. CBP/p300-mediated acetylation of histone H3 on lysine 56. Nature 2009;459:113–7. [217] Adkins MW, Carson JJ, English CM, Ramey CJ, Tyler JK. The histone chaperone anti-silencing function 1 stimulates the acetylation of newly synthesized histone H3 in S-phase. J Biol Chem 2007;282:1334–40. [218] Sklenar AR, Parthun MR. Characterization of yeast histone H3specific type B histone acetyltransferases identifies an ADA2independent Gcn5p activity. BMC Biochem 2004;5:11. [219] Tsubota T, Berndsen CE, Erkmann JA, Smith CL, Yang L, Freitas MA, et al. Histone H3-K56 acetylation is catalyzed by histone chaperone-dependent complexes. Mol Cell 2007;25:703–12. [220] Taylor BL, Zhulin IB. PAS domains: internal sensors of oxygen, redox potential, and light. Microbiol Mol Biol Rev 1999;63:479– 506. [221] Grant PA, Duggan L, Cote J, Roberts SM, Brownell JE, Candau R, et al. Yeast Gcn5 functions in two multisubunit complexes to acetylate nucleosomal histones: characterization of an Ada complex and the SAGA (Spt/Ada) complex. Genes Dev 1997;11:1640–50. [222] Taverna SD, Ilin S, Rogers RS, Tanny JC, Lavender H, Li H, et al. Yng1 PHD finger binding to H3 trimethylated at K4 promotes NuA3 HAT activity at K14 of H3 and transcription at a subset of targeted ORFs. Mol Cell 2006;24:785–96. [223] Utley RT, Ikeda K, Grant PA, Cote J, Steger DJ, Eberharter A, et al. Transcriptional activators direct histone acetyltransferase complexes to nucleosomes. Nature 1998;394:498–502. [224] Wallberg AE, Neely KE, Gustafsson JA, Workman JL, Wright AP, Grant PA. Histone acetyltransferase complexes can mediate transcriptional activation by the major glucocorticoid receptor activation domain. Mol Cell Biol 1999;19:5952–9. [225] Ikura T, Ogryzko VV, Grigoriev M, Groisman R, Wang J, Horikoshi M, et al. Involvement of the TIP60 histone acetylase complex in DNA repair and apoptosis. Cell 2000;102:463–73. [226] Kusch T, Florens L, Macdonald WH, Swanson SK, Glaser RL, Yates JR 3rd, et al. Acetylation by Tip60 is required for

45

selective histone variant exchange at DNA lesions. Science 2004;306:2084–7. [227] Boudreault AA, Cronier D, Selleck W, Lacoste N, Utley RT, Allard S, et al. Yeast enhancer of polycomb defines global Esa1dependent acetylation of chromatin. Genes Dev 2003;17:1415–28. [228] Gregoretti IV, Lee YM, Goodson HV. Molecular evolution of the histone deacetylase family: functional implications of phylogenetic analysis. J Mol Biol 2004;338:17–31. [229] Allfrey VG, Faulkner R, Mirsky AE. Acetylation and methylation of histones and their possible role in the regulation of RNA synthesis. Proc Natl Acad Sci USA 1964;51:786–94. [230] Workman JL, Kingston RE. Alteration of nucleosome structure as a mechanism of transcriptional regulation. Annu Rev Biochem 1998;67:545–79. [231] Shogren-Knaak M, Ishii H, Sun JM, Pazin MJ, Davie JR, Peterson CL. Histone H4-K16 acetylation controls chromatin structure and protein interactions. Science 2006;311:844–7. [232] Shogren-Knaak M, Peterson CL. Switching on chromatin: mechanistic role of histone H4-K16 acetylation. Cell Cycle 2006;5:1361–5. [233] Masumoto H, Hawke D, Kobayashi R, Verreault A. A role for cell-cycle-regulated histone H3 lysine 56 acetylation in the DNA damage response. Nature 2005;436:294–8. [234] Schneider J, Bajwa P, Johnson FC, Bhaumik SR, Shilatifard A. Rtt109 is required for proper H3K56 acetylation: a chromatin mark associated with the elongating RNA polymerase II. J Biol Chem 2006;281:37270–4. [235] Xu F, Zhang K, Grunstein M. Acetylation in histone H3 globular domain regulates gene expression in yeast. Cell 2005;121:375–85. [236] de la Cruz X, Lois S, Sanchez-Molina S, Martinez-Balbas MA. Do protein motifs read the histone code? Bioessays 2005;27:164–75. [237] Owen DJ, Ornaghi P, Yang JC, Lowe N, Evans PR, Ballario P, et al. The structural basis for the recognition of acetylated histone H4 by the bromodomain of histone acetyltransferase gcn5p. EMBO J 2000;19:6141–9. [238] Jenuwein T, Allis CD. Translating the histone code. Science 2001;293:1074–80. [239] Rea S, Eisenhaber F, O’Carroll D, Strahl BD, Sun ZW, Schmid M, et al. Regulation of chromatin structure by site-specific histone H3 methyltransferases. Nature 2000;406:593–9. [240] Brownell JE, Allis CD. Special HATs for special occasions: linking histone acetylation to chromatin assembly and gene activation. Curr Opin Genet Dev 1996;6:176–84. [241] Vignali M, Steger DJ, Neely KE, Workman JL. Distribution of acetylated histones resulting from Gal4-VP16 recruitment of SAGA and NuA4 complexes. EMBO J 2000;19:2629–40. [242] Nourani A, Doyon Y, Utley RT, Allard S, Lane WS, Cote J. Role of an ING1 growth regulator in transcriptional activation and targeted histone acetylation by the NuA4 complex. Mol Cell Biol 2001;21:7629–40. [243] Arabi A, Wu S, Ridderstrale K, Bierhoff H, Shiue C, Fatyol K, et al. c-Myc associates with ribosomal DNA and activates RNA polymerase I transcription. Nat Cell Biol 2005;7:303–10. [244] Fernandez PC, Frank SR, Wang L, Schroeder M, Liu S, Greene J, et al. Genomic targets of the human c-Myc protein. Genes Dev 2003;17:1115–29. [245] Herceg Z, Li H, Cuenin C, Shukla V, Radolf M, Steinlein P, et al. Genome-wide analysis of gene expression regulated by the HAT cofactor Trrap in conditional knockout cells. Nucleic Acids Res 2003;31:7011–23. [246] Knoepfler PS, Zhang XY, Cheng PF, Gafken PR, McMahon SB, Eisenman RN. Myc influences global chromatin structure. EMBO J 2006;25:2723–34. [247] Cai Y, Jin J, Swanson SK, Cole MD, Choi SH, Florens L, et al. Subunit composition and substrate specificity of a MOF-containing

II.  Molecular Mechanisms of Epigenetics

46

3.  Mechanisms of Histone Modifications

histone acetyltransferase distinct from the male-specific lethal (MSL) complex. J Biol Chem 2010;285:4268–72. [248] Ding L, Paszkowski-Rogacz M, Nitzsche A, Slabicki MM, Heninger AK, de Vries I, et al. A genome-scale RNAi screen for Oct4 modulators defines a role of the Paf1 complex for embryonic stem cell identity. Cell Stem Cell 2009;4:403–15. [249] Fazzio TG, Huff JT, Panning B. An RNAi screen of chromatin proteins identifies Tip60-p400 as a regulator of embryonic stem cell identity. Cell 2008;134:162–74. [250] Xu W, Edmondson DG, Evrard YA, Wakamiya M, Behringer RR, Roth SY. Loss of Gcn5l2 leads to increased apoptosis and mesodermal defects during mouse development. Nat Genet 2000;26:229–32. [251] Wang Z, Zang C, Cui K, Schones DE, Barski A, Peng W, et al. Genome-wide mapping of HATs and HDACs reveals distinct functions in active and inactive genes. Cell 2009;138:1019–31. [252] Wang Z, Zang C, Rosenfeld JA, Schones DE, Barski A, Cuddapah S, et al. Combinatorial patterns of histone acetylations and methylations in the human genome. Nat Genet 2008;40:897–903. [253] Anamika K, Krebs AR, Thompson J, Poch O, Devys D, Tora L. Lessons from genome-wide studies: an integrated definition of the coactivator function of histone acetyl transferases. Epigenetics Chromatin 2010;3:18. [254] Dion MF, Altschuler SJ, Wu LF, Rando OJ. Genomic characterization reveals a simple histone H4 acetylation code. Proc Natl Acad Sci USA 2005;102:5501–6. [255] Henikoff S, Shilatifard A. Histone modification: cause or cog? Trends Genet 2011;27:389–96. [256] Stasevich TJ, Hayashi-Takanaka Y, Sato Y, Maehara K, Ohkawa Y, Sakata-Sogawa K, et al. Regulation of RNA polymerase II activation by histone acetylation in single living cells. Nature 2014;516:272–5. [257] Bird AW, Yu DY, Pray-Grant MG, Qiu Q, Harmon KE, Megee PC, et al. Acetylation of histone H4 by Esa1 is required for DNA double-strand break repair. Nature 2002;419:411–5. [258] Choy JS, Kron SJ. NuA4 subunit Yng2 function in intra-S-phase DNA damage response. Mol Cell Biol 2002;22:8215–25. [259] Qin S, Parthun MR. Histone H3 and the histone acetyltransferase Hat1p contribute to DNA double-strand break repair. Mol Cell Biol 2002;22:8353–65. [260] Tamburini BA, Tyler JK. Localized histone acetylation and deacetylation triggered by the homologous recombination pathway of double-strand DNA repair. Mol Cell Biol 2005;25:4903–13. [261] Brand M, Moggs JG, Oulad-Abdelghani M, Lejeune F, Dilworth FJ, Stevenin J, et al. UV-damaged DNA-binding protein in the TFTC complex links DNA damage recognition to nucleosome acetylation. EMBO J 2001;20:3187–96. [262] Martinez E, Palhan VB, Tjernberg A, Lymar ES, Gamper AM, Kundu TK, et al. Human STAGA complex is a chromatin-acetylating transcription coactivator that interacts with pre-mRNA splicing and DNA damage-binding factors in vivo. Mol Cell Biol 2001;21:6782–95. [263] Downs JA, Cote J. Dynamics of chromatin during the repair of DNA double-strand breaks. Cell Cycle 2005;4:1373–6. [264] Celic I, Masumoto H, Griffith WP, Meluh P, Cotter RJ, Boeke JD, et al. The sirtuins hst3 and Hst4p preserve genome integrity by controlling histone H3 lysine 56 deacetylation. Curr Biol 2006;16:1280–9.

[265] Maas NL, Miller KM, DeFazio LG, Toczyski DP. Cell cycle and checkpoint regulation of histone H3 K56 acetylation by Hst3 and Hst4. Mol Cell 2006;23:109–19. [266] Driscoll R, Hudson A, Jackson SP. Yeast Rtt109 promotes genome stability by acetylating histone H3 on lysine 56. Science 2007;315:649–52. [267] Han J, Zhou H, Horazdovsky B, Zhang K, Xu RM, Zhang Z. Rtt109 acetylates histone H3 lysine 56 and functions in DNA replication. Science 2007;315:653–5. [268] Murr R, Loizou JI, Yang YG, Cuenin C, Li H, Wang ZQ, et al. Histone acetylation by Trrap-Tip60 modulates loading of repair proteins and repair of DNA double-strand breaks. Nat Cell Biol 2006;8:91–9. [269] Loizou JI, Murr R, Finkbeiner MG, Sawan C, Wang ZQ, Herceg Z. Epigenetic information in chromatin: the code of entry for DNA repair. Cell Cycle 2006;5:696–701. [270] Murr R, Vaissiere T, Sawan C, Shukla V, Herceg Z. Orchestration of chromatin-based processes: mind the TRRAP. Oncogene 2007;26:5358–72. [271] Hsiao KY, Mizzen CA. Histone H4 deacetylation facilitates 53BP1 DNA damage signaling and double-strand break repair. J Mol Cell Biol 2013;5:157–65. [272] Dantuma NP, van Attikum H. Spatiotemporal regulation of posttranslational modifications in the DNA damage response. EMBO J 2016;35:6–23. [273] Gospodinov A, Herceg Z. Chromatin structure in double strand break repair. DNA Repair 2013;12:800–10. [274] Gospodinov A, Herceg Z. Shaping chromatin for repair. Mutat Res 2013;752:45–60. [275] Sun Y, Jiang X, Xu Y, Ayrapetov MK, Moreau LA, Whetstine JR, et al. Histone H3 methylation links DNA damage detection to activation of the tumour suppressor Tip60. Nat Cell Biol 2009;11:1376–82. [276] Sun Y, Jiang X, Chen S, Fernandes N, Price BD. A role for the Tip60 histone acetyltransferase in the acetylation and activation of ATM. Proc Natl Acad Sci USA 2005;102:13182–7. [277] Kaidi A, Jackson SP. KAT5 tyrosine phosphorylation couples chromatin sensing to ATM signalling. Nature 2013;498:70–4. [278] Ikura T, Tashiro S, Kakino A, Shima H, Jacob N, Amunugama R, et al. DNA damage-dependent acetylation and ubiquitination of H2AX enhances chromatin dynamics. Mol Cell Biol 2007;27:7028–40. [279] Jha S, Shibata E, Dutta A. Human Rvb1/Tip49 is required for the histone acetyltransferase activity of Tip60/NuA4 and for the downregulation of phosphorylation on H2AX after DNA damage. Mol Cell Biol 2008;28:2690–700. [280] Lee HS, Park JH, Kim SJ, Kwon SJ, Kwon J. A cooperative activation loop among SWI/SNF, gamma-H2AX and H3 acetylation for DNA double-strand break repair. EMBO J 2010;29:1434–45. [281] Plass C, Pfister SM, Lindroth AM, Bogatyrova O, Claus R, Lichter P. Mutations in regulators of the epigenome and their connections to global chromatin patterns in cancer. Nat Rev Genet 2013;14:765–80. [282] Ryan RJ, Bernstein BE. Molecular biology. Genetic events that shape the cancer epigenome. Science 2012;336:1513–4.

II.  Molecular Mechanisms of Epigenetics

C H A P T E R

4 The Epigenetics of Noncoding RNA Ravindresh Chhabra Panjab University, Chandigarh, Punjab, India

O U T L I N E Introduction The Noncoding RNA Genetics and Epigenetics Epigenetic Regulation of miRNA Expression Epitranscriptomics of miRNAs Regulation of Epigenetics by miRNAs Plausible Interdependence Between miRNA Targeting and mRNA Methylation

Epigenetic Regulation of lncRNAs Epitranscriptomics of lncRNA Regulation of Epigenetics by lncRNAs

47 47 49 51 52 53

54 55 55

Conclusions

56

References

57

54

INTRODUCTION 1%–2% of these transcripts code for proteins, rest all are ncRNAs [2–4]. The ncRNAs less than 200 nucleotides are called small noncoding RNAs and greater than 200 nucleotides are called long noncoding RNAs (lncRNAs). The small ncRNAs can be further classified into miRNAs, siRNAs, piRNAs, and snoRNAs. The properties of different classes of ncRNAs are discussed in brief in Fig. 4.1A. The ncRNAs, both small and long, have been shown to regulate most of the biological processes including apoptosis [5], Wnt signalling [6], epigenetics (Tables 4.1 and 4.2), embryonic development [7], and tissue morphogenesis [8]. The biogenesis of miRNAs has been extensively reported in literature [9]. The miRNA genes are transcribed by Pol II promoters to produce primary transcripts (pri-transcript) of miRNA, which are several thousand nucleotides in length. The pri-transcript is cleaved by Drosha/DGCR8 complex to form about hundred nucleotides long precursor miRNA (pre-miRNA). The premiRNA is then transported out of the nucleus by RanGTP/Exportin-5 complex where it is further cleaved by Dicer/TRBP complex to produce mature miRNA of ∼22

The Noncoding RNA The central dogma of life had clearly established the importance of RNA molecule in the flow of genetic information. The understanding of transcription and translation processes further elucidated three distinct classes of RNA—mRNA, tRNA, and rRNA. mRNA carries the information from DNA and gets translated to structural or functional proteins, hence they are referred to as the coding RNA (RNA which codes for proteins). tRNA and rRNA helps in the process of translation among other functions. A major part of the DNA does not code for proteins and was earlier referred to as junk DNA. The scientists started realizing the role of the junk DNA in late 1990s and the ENCODE project initiated in 2003 proved the significance of junk DNA beyond any doubt [1]. Many RNA types are now known to be transcribed from DNA in the same way as mRNA, however, unlike mRNA they do not get translated into any protein, hence they are collectively referred to as noncoding RNA (ncRNA). The studies have revealed that upto 90% of the eukaryotic genome is transcribed but only

Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00004-3 Copyright © 2017 Elsevier Inc. All rights reserved.

47

48

4.  The Epigenetics of Noncoding RNA

FIGURE 4.1  (A) Different classes of ncRNAs associated with epigenetic phenomenon. The number of RNAs, belonging to each class is as per their listing in the referenced databases [31,105–107]. These numbers are bound to increase in future. The numbers are not available for endogenous siRNAs. #The size of snoRNAs is highly variable with certain snoRNAs running into a length of more than 500 nucleotides [107]. (B) Schematic representation of miRNA and lncRNA biogenesis and their mechanism of action in regulating epigenetic modifications. miRNA biogenesis begins in the nucleus with the formation of primary miRNA transcript (pri-miRNA), which is a few thousand kilobases in length. The transcription is mediated by either RNA polymerase II (Pol II) or RNA polymerase III (Pol III). The pri-miRNAs are cleaved by Drosha–DGCR8 complex to generate precursor miRNA (pre-miRNA), which is few hundred bases in length. The pre-miRNA is then exported to the cytoplasm by Exportin-5/Ran-GTP and further cleaved by Dicer/TRBP complex to give rise to ∼21 nucleotides long mature miRNA. The mature miRNA is capable of mRNA degradation or translational repression of its target mRNAs. Most of the lncRNAs are nucleus bound and do not have such an elaborate process of biogenesis. The intergenic lncRNAs are transcribed by Pol II and the intronic lncRNAs are transcribed by Pol III. To regulate epigenetic mechanisms, the miRNAs target the expression of proteins like DNMTs while the lncRNAs cause chromatin remodeling by recruiting proteins like PRC2 or DNMTs to preferential chromosome locations. The gray blocks represent different genes and the straight lines connecting these blocks represent intergenic regions in DNA. Me, Methylation.

nucleotides in length. The mature miRNA along with Argonaute 2 proteins is loaded onto the RISC complex, binds the target mRNA and causes mRNA degradation or translational repression. In contrast, the lncRNAs do not have an elaborate process of biogenesis and resemble to a large extent the transcription of protein coding genes. The lncRNAs are further classified based on their genomic

loci and the direction of their transcription [10,11]. The various types of lncRNAs are long intergenic ncRNAs (lincRNAs, originate from the region between proteincoding genes), intronic lncRNAs (originate from the intronic regions of protein-coding genes), sense lncRNAs (originate from exonic regions of protein-coding genes), antisense lncRNAs (transcribed in the direction opposite

II.  Molecular Mechanisms of Epigenetics



49

Introduction

TABLE 4.1 A Brief List of the miRNAs Associated with Epigenetics Phenomenon and the Corresponding Pathological Condition miRNA

Epigenetic modification/DNMT target

Pathological condition

References

miR-155

Hypomethylated

B-cell lymphoma

[34]

miR-191

Hypomethylated

HCC

[35]

let-7a-3

Hypomethylated

Lung adenocarcinoma

[36]

miR-375

Hypomethylated

Er-α positive breast cancer

[37]

miR-200a/b

Hypomethylated

Pancreatic cancer

[38]

miR-126

Hypomethylated

SLE

[80]

miR-124-1, -2, -3

Hypermethylated

Pancreatic cancer and ALL

[40,41]

miR-124-1, -2,

Hypermethylated

Cervical cancer

[42]

miR-124-3

Hypermethylated

HCC

[43]

miR-34b

Hypermethylated

AML, NSCLC

[44,45]

miR-34a

Hypermethylated

Prostate cancer

[46,47]

miR-34a, -b, -c

Hypermethylated

Colorectal cancer

[48,49]

miR-127

Histone deacetylation and Hypermethylation

Bladder carcinoma

[52]

miR-29a, -b, -c

DNMT3A

Lung cancer, AML, cerebral Ischemia

[71,73,108]

miR-29a, -b

DNMT3B

Lung cancer, AML

[71,73]

miR-34b

DNMT3B

Prostate cancer

[70]

miR-152

DNMT1

Cholangiocarcinoma, HCC, atherosclerosis, Ovarian cancer, Endometrial cancer

[67,69,74,75,77]

DNMT1

HCC, ovarian cancer

[75,109]

miR-29b

DNMT1

AML

[73]

miR-148

DNMT1

Cholangiocarcinoma

[67]

miR-342

DNMT1

Colorectal cancer

[76]

DNMT1

Prostate cancer

[70]

DNMT1

SLE

[72,80]

a

miR-185 a

miR-34b a

miR-126, -148, -21 a

Indirect inhibitor.

to the protein-coding genes), pseudogene lncRNAs (originate from pseudogenes), enhancer lncRNAs (originate from enhancer sequences) and promoter lncRNAs (originate from promoter sequences). Most of the lncRNAs are transcribed by Pol II complex; however, some of them are also transcribed by Pol III complex [11]. The lncRNAs regulate gene expression in a variety of ways including chromatin reprogramming. The biogenesis of miRNAs and lncRNAs is depicted in Fig. 4.1B. siRNAs originate from transposon, sense-antisense, and stem-loop transcripts. The dsRNA/stem-loop transcripts, formed in the nucleus, are cleaved by Dicer in the cytoplasm to give rise to functional siRNAs. siRNAs, like miRNAs interact with argonaute 2 protein and cause degradation of their target mRNA [12]. piRNAs are transcribed from intergenic repetitive elements or transposons. ssRNA transcripts from these regions act as precursors to piRNA. Unlike miRNAs and siRNAs, they are

not dependent of Drosha/Dicer for their biogenesis and interact with PIWI proteins [12]. The detailed pathway, however, remains elusive for both siRNAs and piRNAs. Most of the snoRNAs are encoded in the intronic regions with the exception of a few, which originate from nonprotein coding genes. The introns excised during mRNA processing are cleaved by exonucleases to give rise to mature snoRNAs [13]. The snoRNAs are majorly involved in the modification of rRNA but additional functions have also been elucidated including a role similar to miRNAs or as the precursor to miRNAs [14,15].

Genetics and Epigenetics Genetics is a branch of science, which encompasses flow of information from one generation to the next in the form of DNA sequence. It is a science, which attributes a particular function or a phenotype of a cell to

II.  Molecular Mechanisms of Epigenetics

50

4.  The Epigenetics of Noncoding RNA

TABLE 4.2 Some of the lncRNAs Involved in Epigenetic Regulation lncRNA

Function

Biological role

Cell line/tissue

References

HOTTIP

Induces transcription by H3K4me3

Activation of HOXA genes

Foreskin and lung fibroblasts

[89]

ecCEBPA

Associates with DNMT1 and inhibits methylation

Induces transcription of CEBPA

HL-60 and U937 cell line

[96]

DACOR1

Recruits DNMT1 to specific genomic loci and induces methylation

Blocks transcription

HCT116 cell line

[97]

H19

Inhibits SAHH to alter DNMT3b mediated methylation

Developmental aberrations

Mouse muscle cells

[99]

Evf2

Inhibits methylation of Dlx5/6 enhancer elements and differentially regulates the Dlx5/6 expression by controlling the recruitment of activator and repressor proteins

Differential regulation of genes with shared regulatory elements

Mouse model system

[101]

Kcnq1ot1

Recruits DNMT1 to differentially methylated region

Regulates imprinting of genes within the Kcnq1 domain

Mouse embryonic stem cells

[110]

ANRIL

PRC2 recruitment to and H3K27me3 of p15INK4B

Induces cellular proliferation

WI-38 cell line

[111]

AIR

Targets H3K9 histone methyltransferase G9a to chromatin

Induces transcriptional repression

Mouse models

[112]

HOTAIR

Genome wide induction of H3K27me3 and targeting of PRC2 complex on selective genes

Induces cancer metastasis

MCF7, MDA-MB-231, SKBR3, breast cancer

[113]

a specific DNA sequence or sequences. However, there are instances where genotype to phenotype outcomes cannot be explained solely on the basis of DNA sequence. This is because there are additional regulatory factors, which determine the genotype to phenotype conversions. Epigenetics is a branch of science, which deals with the regulation of the genotype via the modification of nucleotide sequences (DNA and RNA) and/ or histone proteins to alter its phenotypic outcome. RD Hotchkiss was the first to report such a modification of DNA sequence [16]. In this report, DNA was found to be methylated at the C5 position of cytosine. Since then, about seven modifications have been reported in DNA, and more than 100 such modifications have been elucidated for RNAs [17,18]. Out of these, the most frequently studied and reported modification remains to be methylation both in case of DNA, as well as RNA (including mRNA, tRNA, rRNA, lncRNA, and miRNAs). DNA methylation occurs majorly in promoter regions and to a lesser degree in intragenic regions and is responsible for either inducing or inhibiting transcription [19]. Among the histone modifications, H3K4me3, H3K27me3, H3K9me3, and H3K36me3 are reported to be the key regulators of mRNA expression [20]. RNA methylation is comparatively more complex as it can affect the ribose sugar, as well as all the bases [17,21]. Among the RNA species, tRNA and rRNA methylations have been studied in detail because of their higher stability than mRNA. The recent innovations in NGS technology

have however, made it easier to study methylations in mRNA too. The most common methylations reported in mRNA include m6A (methylation of adenosine at its N6 position) and m5C (methylation of cytosine at its C5 position). m6A occurs mostly in 3′-UTRs, internal long exons and around stop codons [22,23] and m5C were reportedly found to be enriched in UTR regions and many of these regions in 3′-UTRs overlap with the binding region of Argonaute proteins [24]. While m6A modification clearly destabilizes the mRNA [23,25], m5C modification may be speculated to stabilize the mRNA as Karikó et al. in 2008 observed fourfold increase in the translation of mRNA modified with m5C [26]. Genomic imprinting is a special type of epigenetic phenomenon in which a small fixed set of genes is expressed in a parental-origin specific manner [27]. Genomic imprinting is essential for normal mammalian development and growth. The aberration in genomic imprinting results in a number of diseases including Angelmann syndrome, Prader–Willi syndrome, autism and so on [27]. The underlying mechanism of DNA and histone modification and the involvement of noncoding RNA are the same for imprinting at developmental stages and the epigenetic regulation at later stages [28]. This chapter will mostly discuss about the epigenetics in context to two major classes of noncoding RNA, miRNAs, and lncRNAs. This is because there is sufficient literature available for both these classes in context

II.  Molecular Mechanisms of Epigenetics



Introduction

to epigenetics to arrive at some conclusive inferences. However, that is in no way implying that epigenetic modifications are not known for other noncoding RNA. In fact, the epigenetics of siRNAs and piRNAs have also been reported [29,30].

Epigenetic Regulation of miRNA Expression There are around 1800 miRNA sequences spread across the entire human genome, which give rise to ∼2500 functional miRNAs [31]. They can be intronic or intergenic miRNAs depending on their genomic loci. The intergenic miRNAs are transcribed independently by their promoters. In contrast, the intronic miRNA may either share the promoter with the host gene or may be transcribed independently of the host gene promoter [32]. Epigenetic modification of miRNA promoter region has been implicated in both physiological and pathological conditions. In fact, an estimated 50% of miRNA genes have been associated with CpG islands, the hotspot for DNA methylation [33]. Hypermethylation/hypomethylation of miRNA promoters has been observed in a number of cancers. Baer et al., in their study on whole miRnome promoter methylation in chronic lymphocytic leukemia (CLL) revealed that 90 miRNA promoters were hypomethylated and 38 promoters were hypermethylated [34]. There are, however, more reports on hypermethylated miRNA promoters in the literature. This may be attributed to the availability of pharmacological demethylating agents, which makes it comparatively easier to identify the function of hypermethylated miRNA promoters. Nevertheless, hypomethylated miRNAs have also been implicated in a number of cancers. The increased expression of miR-155 in higher grade B-cell lymphoma [34], miR-191 in hepatocellular carcinoma (HCC) [35], let-7a-3 in lung adenocarcinoma [36], miR-375 in Er-α positive breast cancer [37], and miR-200a/b in pancreatic cancer [38] is because of the hypomethylation of their respective promoters. The hypomethylated miR-375 inhibits the expression of RASD1, an antiproliferative factor in Erα positive breast cancer [37] and the hypomethylated let-7a-3 causes increase in expression of CDK6, PCNA, PRDX1, and CXCL5 (which induce lung cancer progression) and reduces the expression of PPARG, TGFB2, and SFRP1 (which inhibit lung cancer proliferation) [36]. The direct functional target of hypomethylated miRNAs remains unknown in most of the studies. In addition to hypomethylation, histone modification is also responsible for the elevated levels of miR-375 in Er-α positive breast cancer [37]. The hypermethylation of miR-124 promoter is frequently seen in the tumors of lung, breast, colon, and leukemia and lymphoma [39]. This hypermethylation

51

is however, absent in neuroblastoma and sarcoma [39]. miR-124 originates from three different genomic loci and hence, are classified as miR-124-1, -124-2, and -124-3, which implies the sequence of their mature form is the same, while sequence of primary and precursor forms is different. The methylation pattern of all three miR-124 promoters varies in different cancers. While all three types are hypermethylated in pancreatic cancer [40] and acute lymphoblastic leukemia (ALL) [41], miR-124-1 and -2 are hypermethylated in cervical cancer [42] and only miR-124-3 is hypermethylated in HCC [43]. The silencing of miR-124 as a result of hypermethylation induces the expression of vimentin (EMT marker) in HCC [43], Rac1(oncogene) in pancreatic cancer [40] and CDK6 (cell cycle regulator) in colorectal cancer [39] and ALL [41]. The hypermethylation of miR-124 is correlated with increased mortality and relapse rate in ALL [41] and is responsible for enhanced proliferation and invasion in pancreatic cancer [40]. The promoter hypermethylation is also responsible for silencing of miR-34 family in acute myeloid leukemia (AML) [44], nonsmall cell lung cancer (NSCLC) [45], prostate cancer [46,47], and colorectal cancer [48,49]. This aberrant methylation of miR-34 family members induces proliferation in AML and metastasis in NSCLC, colorectal cancer and prostate cancer. Moreover, the reduced miR-34 induces cancer stem cell phenotype [50], which has been implicated in chemoresistance, metastasis, and relapse of a number of cancers. Since miR-34 silencing caused by hypermethylation could aid cancer progression in different ways, any drug which reverses the methylation pattern of miR-34 would make an excellent therapeutic. Difluorinatedcurcumin and BioResponse 3,3′-diindolylmethane are two such compounds, which have been shown to have demethylation effect on miR-34 in chemoresistant colon cancer cell line [49] and in prostate cancer cells [51], respectively. In addition to promoter methylation, histone modification can also regulate the expression of miRNAs. At times, both histone modification and DNA methylation coordinate to regulate miRNA expression. For instance, in bladder carcinoma, histone deacetylation, and DNA methylation of miRNA promoter was found to be responsible for downregulation of miR-127 and as a result, increase in expression of its target, BCL6 [52]. Similarly, in ALL, DNA methylation, and histone modifications (enhanced levels of H3K9me2 and/ or low levels of H3K4me3) in CpG islands around 13 miRNAs downregulated the expression of all 13 miRNAs [53]. An interesting liaison among epigenetics, imprinting, and miRNAs was described in mice where miR-127 and miR-136 mediate the function of imprinting to silence

II.  Molecular Mechanisms of Epigenetics

52

4.  The Epigenetics of Noncoding RNA

FIGURE 4.2  The imprinted gene, Rtl1 is silenced by miR-127 and miR-136. On the maternal chromosome, intergenic differentially methylated region (IG-DMR) downstream of Rtl1 gene is unmethylated, which allows the transcription of Rtl1, as well as antisense-Rtl1. The antisense Rtl1 is further processed to form mature miR-127 and miR-136 which bind to the Rtl1 transcript and silences its expression. On the paternal chromosome; however, IG-DMR is methylated which precludes the transcription of the antisense-Rtl1 transcript, thus allowing the formation of functional Rtl1 protein. Me, Methylation.

the expression of Rtl1, the imprinted gene responsible for placenta formation (Fig. 4.2). Rtl1 gene is maternally imprinted as the intergenic differentially methylated region (IG-DMR) downstream of Rtl1 gene is unmethylated while in the paternal chromosome, the methylated IG-DMR prevents the silencing of Rtl1. The unmethylated IG-DMR allows the transcription of antisense Rtl1 transcript, which is further processed to form miR127 and miR-136. Both these miRNAs bind to the Rtl1 transcript and causes its degradation [54]. It was also observed that it is not just the absence of methylation, which induces the expression of miR-127 and miR-136 however, it is the epigenetic imprinting which controls their expression [55]. The examples discussed here do not give a complete list of hypermethylated miRNAs and their associated functional significance however, provide a general idea about the significance of epigenetic regulation of miRNAs in the normal functioning of different cell types.

Epitranscriptomics of miRNAs In contrast to the effect of DNA modifications on the miRNA expression pattern, the direct modifications of miRNA and their associated function are not as extensively reported. Methylation of miRNAs is a fundamental part of miRNA biogenesis in plants however, has rarely been reported in case of animals. In plants, miRNA mediated degradation of its mRNA-target bound in miRNA-Ago complex is facilitated by uridylation of 3′ends of mRNA by HESO1 (HEN1 suppressor) enzyme. Plant miRNAs are protected from this uridylation as they are methylated

at 3′ ends by the enzyme, HEN1 during their biogenesis [56]. This results in efficient recycling of plant miRNAs [57]. miRNA mediated degradation of mRNA in animals also happens by uridylation [58] but herein, very little is known about the recycling of miRNAs. Since many reports in the literature suggest that uridylation reduces the activity, as well as stability of miRNAs [59,60], it may be assumed that miRNAs may also get degraded along with its target. Methylation of miRNAs in humans was first reported in 2014 when it was observed that different forms of miR125b (primary miRNA, precursor miRNA and mature miRNA) in HeLa cells are methylated at adenosine sites (m6A) by NSUN2 [61]. This is quite surprising, as traditionally, NSUN2 has been known to add methyl group to C5 of cytosine (m5C). The reason for changed substrate preference of NSUN2 remains unknown. The m6A methylation at any stage of miR-125b biogenesis causes inhibition of the subsequent step in biogenesis. This implies methylation of primary miRNA reduced its processing to precursor form, methylation of precursor miRNA reduced its cleavage to mature form and methylation of mature miRNA inhibited its function by attenuating the recruitment of RISC by miRNA. Since miR-125b originates from two different sites in the genome, they are of two subtypes, mir-125b-1 and mir-125b-2. Although, all the forms of both miR-125b-1 and -125b-2 were found to be methylated, the functional effect was surprisingly observed only in case of miR-125b-2. The effect of m6A on miRNAs is not just exclusive to miR-125b but extends to a majority of miRNAs methylated at m6A [62]. A rather fundamental role of methylation in miRNA biogenesis in humans was recently uncovered by Alarcón and coworkers [63]. They reported that methyltransferase-like

II.  Molecular Mechanisms of Epigenetics



Introduction

3 (METTL3) mediates m6A methylation of pri-miRNAs leading to their recognition and processing by DGCR8. They also proposed that RNA-binding protein HNRNPA2B1 could be the one which recognise the m6A mark on the pri-miRNA and promotes its processing by recruiting DGCR8 [64]. m6A has similar effects (destabilization and protein recruitment) on mRNA and lncRNAs, as well [25,65,66], thus implying a universal function of m6A across various RNA classes.

Regulation of Epigenetics by miRNAs DNA methylation is one of the most studied epigenetic modifications. In mammals, there are three enzymes known to add methyl group to the specific sites in DNA, collectively referred to DNA methyltransferases (DNMT1, DNMT3a, and DNMT3b). DNMT1 is referred to as the maintenance methyltransferase as it maintains the methylation pattern during DNA replication and in contrast, DNMT3a and DNMT3b are called de novo methyltransferases as they carry out de novo methylation during embryonic development. Since methylation patterns are crucial for maintaining the healthy state in individuals, the aberration in expression of the DNMTs results in a number of diseases including systemic lupus erythematosus (SLE) and cancers of liver, prostate, and ovary. In many of these diseases, the aberrant expression of DNMTs is because of miRNAs [67–72]. A few of these miRNA-DNMTs interactions are discussed here. miR-29b targets DNMT3a and -3b in lung cancer, as well as in AML [71,73]. In both these diseases, the expression of DNMT3a and -3b increases as a result of reduced miR-29b. This leads to enhanced methylation and subsequent silencing of tumor suppressor genes, FHIT and WWOX in lung cancer [71] and p15INK4b and ESR1 in AML [73]. Apart from directly targeting DNMT3a and -3b, miR-29b also targets DNMT1 indirectly by downregulating its transcriptional activator, Sp1. This makes miR-29b a potent hypomethylating agent, which can cause global DNA hypomethylation [73]. The loss of miR152 causes enhanced expression of its target, DNMT1 in atherosclerosis [74], cholangiocarcinoma [67], HCC [69], and ovarian cancer [75]. The enhanced DNMT1 leads to enhanced methylation and subsequent silencing of an antiatherosclerotic gene, ERα [74], and tumor-suppressor genes, CDH1 and GSTP1 in HCC [69]. The genes targeted for methylation by enhanced DNMT1 in cholangiocarcinoma and ovarian cancer remains unknown. miR-29b and miR-152 therefore act as tumor suppressor genes by keeping the expression of DNMTs in check. Similar functions have also been reported for miR-148 in cholangiocarcinoma [67], miR-185 in ovarian cancer [75], miR-34b in prostate cancer [70], miR-342 in colorectal cancer [76], and miR-152 in endometrial cancer [77]. A rather elaborate mechanism is involved in an indirect

53

activation of de novo methylation by miR-290 cluster during mouse development. Herein, mir-290 cluster prevents the expression of Rbl2, the inhibitor of DNMT3a and -3b. The absence of miR-290 would thus cause genomic hypomethylation resulting in elongated telomeres and increased telomere recombination (Fig. 4.3) [78]. Embryonic stem cell maintenance is also dependent on miR-290 cluster regulated de novo methylation. During differentiation of embryonic stem cells, miR-290 mediates indirect activation of DNMTs and represses the expression of Oct4 and Nanog (stem cell marker genes) via promoter methylation [79]. While cancer is associated with hypermethylation by DNMTs, an autoimmune disease SLE is associated with hypomethylation caused by reduced expression of DNMT1. Enhanced expression of miR-148 [72] and miR-126 [80] was found to be responsible for this hypomethylation as they both target DNMT1. The hypomethylation of genes, CD11a and CD70 and their subsequent activation are speculated to be cause SLE [80]. Interestingly, reduced expression of miR-148 causes hypermethylation in cholangiocarcinoma [67], thereby stressing the importance for tight regulation of miRNA expression.

FIGURE 4.3  The indirect regulation of DNMT3a/b by miR-290 cluster affects the cellular differentiation. The de novo methylation of Nanog and Oct4 by DNMT3a/b is essential for the stem cell differentiation during development but Rbl2 protein blocks DNMT3a/b. During the onset of cellular differentiation, miR-290 cluster inhibits the expression of Rbl2 and thus allows DNMT3a/b to methylate and silence Nanog and Oct4, the genes responsible for stem cell maintenance. The up arrow indicates enhanced expression and the down arrow indicates reduced expression during cellular differentiation. Me, Methylation.

II.  Molecular Mechanisms of Epigenetics

54

4.  The Epigenetics of Noncoding RNA

There are a few studies, which imply a potential feedback mechanism between the miRNAs and DNMTs. For instance, miR-126 responsible for downregulating DNMT1 in SLE is itself hypomethylated in its promoter region [80], reduced expression of miR-34b responsible for upregulating DNMT1 in prostate cancer could be attributed to its hypermethylated upstream region [70] and reduced expression of miR-152 responsible for upregulating DNMT1 in endometrial cancer could be because of its hypermethylated upstream region [77]. The mechanism, which helps the cell, decides whether to silence miRNAs or its target DNMTs remains unexplored. Some of the miRNAs associated with epigenetics phenomenon are listed in Table 4.1. miRNAs can be expected to similarly regulate the m5C methylation of mRNA by targeting the enzymes NSUN2 and TRDMT1 and m6A methylation of mRNA by targeting the enzymes METTL3, METTL14, WTAP, and FTO. There have been few reports which have tried to identify the miRNAs which target these genes but none of these reports show the functional effect of such miRNA mediated regulation of mRNA methylation [81–83].

Plausible Interdependence Between miRNA Targeting and mRNA Methylation Majority of the 144 types of modifications in RNA [18] were discovered in rRNA or tRNA owing to their abundance and stability in contrast to short life span of mRNAs. In the past few years, however, the advancement of technology has allowed for the study of mRNA modifications. The most commonly studied modifications in mRNA are methylation of Adenosine at N6 position (m6A) [23,84] and methylation of cytosine at C5 position (m5C) [24,85]. Intriguingly, both m6A and m5C sites are found to be enriched in 3′UTR region [22,24,84], the preferred region of miRNAs for binding its target mRNA and subsequently inhibiting its expression. For m5C sites in 3′UTR, the enrichment was even more in the binding region of argonaute protein [24]. Since miRNAs target the 3′UTR sites by forming a complex with argonaute protein, this suggests a potential role of m5C in the miRNA mediated target inhibition. In the 3′UTRs where both m6A and TargetScan predicted miRNA binding sites are present, the m6A precede miRNA-binding sites 62% of the time [22]. miRNAs were speculated to cause this m6A methylation as majority of the highly expressed miRNAs of brain were shown to have higher percentage of targets with m6A sites [22]. m6A sites have also been reported to alter the structure of RNA to aid/prevent its binding to specific proteins [64,86]. Specifically, binding of HUR protein (which blocks miRNA–mRNA interaction) is prevented by m6A methylation [87]. Additionally, while m6A has a negative effect on the stability of mRNA [25,65], m5C may

stabilize the mRNA [26]. It is thus worth speculating that presence of m6A in 3′UTR regions alters the mRNA structure in a way which makes it amenable for miRNA binding and subsequent degradation in contrast to m5C which somehow prevents the miRNA from inhibiting the expression of its target mRNA. These speculations are however, yet to be validated experimentally. This aspect of mRNA methylation in context to miRNA function had been reviewed previously [68].

Epigenetic Regulation of lncRNAs The mammalian genome encodes thousands of lncRNAs, which may be intragenic or intergenic [88]. They may or may not be polyadenylated. lncRNAs have critical roles in various biological processes including embryonic development [7] and tissue morphogenesis [8]. In spite of all these important functions, lncRNAs are poorly conserved in mammalian species [89]. The regulation of lncRNA expression is not yet fully understood but there are many reports, which highlight the epigenetic modulation of lncRNAs [20,90,91]. Modification of lysine residues of histone protein H3 at position 4 (H3K4), 9 (H3K9), 27 (H3K27), and 36 (H3K36) play a vital regulatory role in the gene expression. The most common modifications at the aforementioned residues are acetylation and methylation. While H3K4 methylation, H3K36 methylation and H3K9 acetylation are usually associated with euchromatin implying actively transcribed regions, H3K9 and H3K27 methylation are associated in heterochromatin regions, implying silencing of gene expression in those regions. Sati et al., in 2012 performed a comprehensive study to locate epigenetic marks across lncRNAs loci in humans [20]. The epigenetic marks discussed in their study included DNA methylation and histone modifications, H3K27me3, and H3K9me3 (associated with reduced expression) and H3K4me3 and H3K36me3 (associated with enhanced expression). They observed that in contrast to the protein-coding genes where hypomethylation around or at transcription start site (TSS) corresponds to higher expression and hypermethylation around or at TSS corresponds to reduced expression in lncRNAs, the methylation density in lncRNAs is high in the downstream region of TSS, irrespective of their expression levels. Thus, suggesting that other factors may be needed to regulate lncRNAs. CpG islands present in lncRNAs were marked by H3K4me3 [20], which generally corresponds to transcriptionally active chromatin [92]. Also, H3K4me3 and H3K36me3 modifications in the TSS and gene body, respectively, corresponded to increased expression of lncRNAs [20]. Between the repressive marks, while H3K9me3 was present, H3K27me3 was absent in TSS of the highly expressed lncRNA [20]. In short, many of the epigenetic marks known for the protein coding

II.  Molecular Mechanisms of Epigenetics



Introduction

genes like DNA methylation and H3K9me3 behave ambiguously or have no effect in case of lncRNAs but for other epigenetic marks, their effect is similar to that in protein coding genes. In embryonic stem cells, lncRNAs with high levels of H3K27me3 in their promoter region have reduced expression and inhibiting this modification by silencing H3K27me3 methyltransferase (Ezh2) induces their expression [90]. The expression of XIST, lncRNA responsible for X chromosome inactivation in females is regulated by Tsix (noncoding antisense transcript to XIST) via histone modifications [93]. In males, Tsix induces H3K9 trimethylation and DNA methylation [93] and prevent H3K27 trimethylation [93,94] making the embryonic stem cells unable to transcribe XIST, thereby preventing X-inactivation. In case of females, there is asymmetric regulation of Tsix where Tsix is repressed and XIST expression is induced only on the future inactive X [95]. There are, however, contradictory reports on the mechanism behind Tsix mediated induction of XIST (Fig. 4.4) [93,94]. On activated X-chromosome, Tsix maintains the repression on XIST expression by directing DNMT3a to methylate the promoter of XIST and thus allowing the genes of X chromosome to be expressed.

Epitranscriptomics of lncRNA Site-specific methylation of lncRNAs came to light in 2013 when Amort et al. showed site specific cytosine methylation of XIST and HOTAIR lncRNAs by analyzing a part of the sequences of these 2 lncRNAs [21]. HOTAIR was always found to be methylated at cytosine position at 1683 in HEK293, NT2, Hs578T, and BT-20 and the

55

HOC7 cell lines. The methylated site C1683 of HOTAIR is in the vicinity of the sequence responsible for interacting with histone demethylase complex LSD1/CoREST/ REST, thereby suggesting that this methylation could be important for HOTAIR-LSD1 interaction. While HOTAIR always showed consistent methylation at C1683 position, XIST showed variable methylation pattern at cytosine residues at position 701, 702, 703, 711, and 712 in HEK293 and that too only in about 1/5th of the number of tested clones. Moreover, the identical XIST sequence in mice did not show methylations at the corresponding sites. The methylated cytosines might prevent the interaction of XIST with polycomb-repressive complex 2 (PRC2), which aids in chromatin remodeling via histone modifications.

Regulation of Epigenetics by lncRNAs miRNA mediates epigenetic regulation usually by directly or indirectly controlling the expression of DNMTs. In contrast, the lncRNAs regulate epigenetic phenomenon by remodeling chromatin structure, which is usually aided by their structure and their ability to recruit proteins like DNMTs and PRC2 to preferential chromosomal locations (Fig. 4.1). Di Ruscio et al., in 2013 identified a novel lncRNA, ecCEBPA in HL-60 and U937 cell line. ecCEBPA is an intragenic lncRNA, which is concomitantly expressed along with CEBPA transcript. The ecCEBPA lncRNA is localized in nucleus and a small region of ecCEBPA is capable of physically associating with DNMT1. The association of ecCEBPA and DNMT1 prevents the methylation of promoter region of CEBPA and thereby

FIGURE 4.4  The proposed mechanisms for X-chromosome inactivation by Xist/Tsix. (A) The research findings of Sun et al. [94] state that prior to X-chromosome inactivation (pre-XCI), Tsix maintains euchromatin state by H3K4 dimethylation and precludes the transcription of Xist. During XCI, Tsix expression is repressed, which causes heterochromatinization indicated by loss of H3K4 dimethylation and gain of H3K27 trimethylation. This leads to transcription of Xist, which causes XCI. (B) A contradictory mechanism was proposed by Navarro et al. [93] where during pre-XCI, DNA methylation and H3K9 trimethylation maintains the heterochromatin state of the Xist promoter, which precludes the transcription of Xist. During XCI, Tsix expression is repressed which causes euchromatinization of the Xist promoter indicated by gain of H3K4 di/ trimethylation and H3K9 acetylation. This allows Xist to express and cause XCI. Ac, Acetylation; Me, methylation.

II.  Molecular Mechanisms of Epigenetics

56

4.  The Epigenetics of Noncoding RNA

induces the transcription of CEBPA [96]. This effect of ecCEBPA may not be limited to CEBPA locus but extend to other genomic loci too. Also, ecCEBPA is not the exclusive lncRNA to associate with DNMT1. In fact, many nonpolyadenylated lncRNAs have been shown to have a similar function as ecCEBPA. In 2015, a subset of lncRNAs was identified which interacts with DNMT1 in a colon cancer cell line, HCT116 [97]. One of these lncRNAs was found to be associated with the 31 genomic loci known to be differentially methylated regions in colon tumors. Additionally, this lncRNA had reduced expression in colon tumors prompting the authors to call it DACOR1 (DNMT1-associated Colon Cancer Repressed lncRNA 1). Just like ecCEBPA, DACOR1 affects DNA methylation levels at multiple sites in the human genome but unlike ecCEBPA, which suppressed DNA methylation, DACOR1 enhanced DNA methylation. It seems ecCEBPA competitively binds DNMT1 owing to its higher binding affinity over the genomic loci [96] while DACOR1 may help recruit DNMT1 to specific genomic loci [97]. Also, DNMT1–DACOR1 interaction indirectly elevates S-adenosyl methionine (SAM) levels, the preferred methyl donor for DNMT1 activity [97]. The exact mechanism behind this remains unexplored but the contradictory effect of two lncRNAs points to the complex paradigm of DNMT1-lncRNA interactions. Both these studies indicate that gene expression pattern dependent on DNA methylation can get significantly altered without any changes in histone complex or DNMT1 expression. It is highly probable that other DNMTs (DNMT3a and 3b) may also display such association with lncRNA and in turn be regulated by them [98]. An indirect regulation of DNMT3b by H19 lncRNA has already been observed [99]. H19 inhibits the enzyme S-adenosylhomocysteine hydrolase (SAHH), the only mammalian enzyme capable of hydrolyzing S-adenosylhomocysteine (SAH). SAH acts as a feedback inhibitor of S-adenosylmethionine-dependent methyltransferases. Hence, knockdown of H19 increases DNMT3b-mediated methylation of Nctc1 within the Igf2-H19-Nctc1 locus [99]. In a way, H19 lncRNA can affect the methylation by all the DNMTs since SAHH does not bring about this effect by physically interacting with DNMT rather by inhibiting SAM dependent methyltransferase reaction. Also, the methyltransferases are not equally sensitive to SAHH, hence H19 mediated SAHH knockdown may result in any of the three outcomes—hypermethylation, hypomethylation, or no change in methylation [99,100]. In addition to regulating CpG island methylation, the lncRNAs have also been shown to differentially regulate transcription of adjacent genes via their effect on DNA enhancer elements. One such study was published by Berghoff et al. in 2013, where it was shown that lncRNA Evf2 prevents the site specific methylation of Dlx5/6 ultraconserved enhancer at two sites [101]. Such an effect

of lncRNA is referred to as the trans effect in contrast to the localized effect of lncRNAs where it is referred to as cis effect. Some of the lncRNAs involved in regulation of epigenetics are listed in Table 4.2.

CONCLUSIONS The investigations into the role of noncoding RNAs in epigenetics have revealed a lot of insight into the complexity of gene regulation. In the last decade, the advent of NGS has made it easier to uncover the genome wide histone modifications and methylation patterns and understand the role of noncoding RNAs in the epigenetic phenomenon. The aberrant regulation of epigenetic pathways by miRNAs and lncRNAs has been documented in a number of diseases but it is still far before we could understand the general rules or principles which govern the interactions between noncoding RNA and epigenetics. In addition to miRNAs and lncRNAs, a few reports also describe the role of siRNAs and piRNAs in regulating the epigenetic phenomenon [102–104], however, there is comparatively little information to review them in detail. With continuous innovations in high throughput techniques, it can be safely assumed that our knowledge of this area would increase at a very rapid pace. Many of the newer links between noncoding RNA and epigenetics are already on the brink of discovery including the role of methylations (both m5C and m6A) present in the 3′UTR of mRNA on the binding capability of miRNAs. It could certainly be assumed that the ncRNA-epigenetic interactions could make for excellent therapeutic targets for pathological conditions resulting from aberrant epigenetic modifications.

Abbreviations BCL6 B-cell CLL/lymphoma 6 CDH1 Cadherin 1 CDK6 Cyclin dependent kinase 6, CXCL5 C-X-C motif chemokine ligand 5 DGCR8 DGCR8, microprocessor complex subunit (DiGeorge syndrome critical region gene 8) DNMT DNA methyltransferase EMT Epithelial–mesenchymal transition ERα (official name ESR1) Estrogen receptor 1 ESR1 Estrogen receptor 1 FHIT Fragile histidine triad FTO FTO, alpha-ketoglutarate dependent dioxygenase (fat mass and obesity associated) GSTP1 Glutathione S-transferase pi 1 HEN1 double-stranded RNA binding protein-related HESO1 Nucleotidyltransferase family protein HNRNPA2B1 heterogeneous nuclear ribonucleoprotein A2/B1 METTL methyltransferase-like NANOG Nanog homeobox NSUN2 NOP2/Sun RNA methyltransferase family member 2

II.  Molecular Mechanisms of Epigenetics



REFERENCES

OCT4 (official name POU5F1) POU class 5 homeobox 1 p15INK4b (official name CDKN2B) Cyclin-dependent kinase inhibitor 2B PCNA Proliferating cell nuclear antigen PPARG Peroxisome proliferator activated receptor gamma PRDX1 Peroxiredoxin Rac1 Ras-related C3 botulinum toxin substrate 1 RASD1 Ras related dexamethasone induced 1 Rbl2 RB transcriptional corepressor like 2 Rtl1 Retrotransposon-like 1 SFRP1 Secreted frizzled related protein 1 TGFB2 Transforming growth factor beta 2 TRDMT1 tRNA aspartic acid methyltransferase 1 WTAP Wilms tumor 1 associated protein WWOX WW domain containing oxidoreductase

Acknowledgments This work was supported by the INSPIRE Faculty fellowship awarded to RC (Code: IFA-LSBM-18) by Department of Science and Technology, India.

References [1] Dunham I, Kundaje A, Aldred SF, Collins PJ, Davis CA, Doyle F, et al. An integrated encyclopedia of DNA elements in the human genome. Nature 2012;489:57–74. [2] Kaikkonen MU, Lam MTY, Glass CK. Non-coding RNAs as regulators of gene expression and epigenetics. Cardiovasc Res 2011;90:430–40. [3] Frith MC, Pheasant M, Mattick JS. The amazing complexity of the human transcriptome. Eur J Hum Genet 2005;13:894–7. [4] Djebali S, Davis CA, Merkel A, Dobin A, Lassmann T, Mortazavi A, et al. Landscape of transcription in human cells. Nature 2012;489:101–8. [5] Ambros V. The functions of animal microRNAs. Nature 2004;431:350–5. [6] Kim NH, Kim HS, Kim NG, Lee I, Choi HS, Li XY, et al. p53 and MicroRNA-34 are suppressors of canonical Wnt signaling. Sci Signal 2011;4:ra71. [7] Sauvageau M, Goff LA, Lodato S, Bonev B, Groff AF, Gerhardinger C, et al. Multiple knockout mouse models reveal lincRNAs are required for life and brain development. Elife 2013;2013:e01749. [8] Grote P, Wittler L, Hendrix D, Koch F, Währisch S, Beisaw A, et al. The tissue-specific lncRNA Fendrr is an essential regulator of heart and body wall development in the mouse. Dev Cell 2013;24:206–14. [9] Winter J, Jung S, Keller S, Gregory RI, Diederichs S. Many roads to maturity: microRNA biogenesis pathways and their regulation. Nat Cell Biol 2009;11:228–34. [10] Kowalczyk MS, Higgs DR, Gingeras TR. Molecular biology: RNA discrimination. Nature 2012;482:310–1. [11] Khandelwal A, Bacolla A, Vasquez KM, Jain A. Long noncoding RNA: a new paradigm for lung cancer. Mol Carcinog 2015;54:1235–51. [12] Kim VN, Han J, Siomi MC. Biogenesis of small RNAs in animals. Nat Rev Mol Cell Biol 2009;10:126–39. [13] Filipowicz W, Pogacˇi V. Biogenesis of small nucleolar ribonucleoproteins. Curr Opin Cell Biol 2002;14:319–27. [14] Scott MS, Avolio F, Ono M, Lamond AI, Barton GJ. Human miRNA precursors with box H/ACA snoRNA features. PLoS Comput Biol 2009;5.

57

[15] Ender C, Krek A, Friedländer MR, Beitzinger M, Weinmann L, Chen W, et al. A human snoRNA with microRNA-like functions. Mol Cell 2008;32:519–28. [16] Hotchkiss RD. The quantitative separation of purines, pyrimidines, and nucleosides by paper. J Biol Chem 1948;175:315–32. [17] Cantara WA, Crain PF, Rozenski J, McCloskey JA, Harris KA, Zhang X, et al. The RNA modification database, RNAMDB: 2011 update. Nucleic Acids Res 2011;39:D195–201. [18] Machnicka MA, Milanowska K, Oglou OO, Purta E, Kurkowska M, Olchowik A, et al. MODOMICS: A database of RNA modification pathways—2013 update. Nucleic Acids Res 2013;41: D262–7. [19] Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat Rev Genet 2012;13:484–92. [20] Sati S, Ghosh S, Jain V, Scaria V, Sengupta S. Genome-wide analysis reveals distinct patterns of epigenetic features in long noncoding RNA loci. Nucleic Acids Res 2012;40:10018–31. [21] Amort T, Souliere MF, Wille A, Jia XY, Fiegl H, Worle H, et al. Long non-coding RNAs as targets for cytosine methylation. RNA Biol 2013;10:1003–8. [22] Meyer KD, Saletore Y, Zumbo P, Elemento O, Mason CE, Jaffrey SR. Comprehensive analysis of mRNA methylation reveals enrichment in 3′UTRs and near stop codons. Cell 2012;149:1635–46. [23] Fu Y, Dominissini D, Rechavi G, He C. Gene expression regulation mediated through reversible m6A RNA methylation. Nat Rev Genet 2014;15:293–306. [24] Squires JE, Patel HR, Nousch M, Sibbritt T, Humphreys DT, Parker BJ, et al. Widespread occurrence of 5-methylcytosine in human coding and non-coding RNA. Nucleic Acids Res 2012;40: 5023–33. [25] Wang Y, Li Y, Toth JI, Petroski MD, Zhang Z, Zhao JC. N6-methyladenosine modification destabilizes developmental regulators in embryonic stem cells. Nat Cell Biol 2014;16:191–8. [26] Karikó K, Muramatsu H, Welsh FA, Ludwig J, Kato H, Akira S, et al. Incorporation of pseudouridine into mRNA yields superior nonimmunogenic vector with increased translational capacity and biological stability. Mol Ther 2008;16:1833–40. [27] Ferguson-Smith AC. Genomic imprinting: the emergence of an epigenetic paradigm. Nat Rev Genet 2011;12:565–75. [28] Barlow DP. Genomic imprinting: a mammalian epigenetic discovery model. Annu Rev Genet 2011;45:379–403. [29] Ji L, Chen X. Regulation of small RNA stability: methylation and beyond. Cell Res 2012;22:624–36. [30] Kirino Y, Mourelatos Z. Mouse Piwi-interacting RNAs are 2′-Omethylated at their 3′ termini. Nat Struct Mol Biol 2007;14: 347–8. [31] Griffiths-Jones S. The microRNA registry. Nucleic Acids Res 2004;32:D109–11. [32] Monteys AM, Spengler RM, Wan J, Tecedor L, Lennox KA, Xing Y, et al. Structure and activity of putative intronic miRNA promoters. RNA 2010;16:495–505. [33] Wang Z, Yao H, Lin S, Zhu X, Shen Z, Lu G, et al. Transcriptional and epigenetic regulation of human microRNAs. Cancer Lett 2013;331:1–10. [34] Baer C, Claus R, Frenzel LP, Zucknick M, Park YJ, Gu L, et al. Extensive promoter DNA hypermethylation and hypomethylation is associated with aberrant microRNA expression in chronic lymphocytic leukemia. Cancer Res 2012;72:3775–85. [35] He Y, Cui Y, Wang W, Gu J, Guo S, Ma K, et al. Hypomethylation of the hsa-miR-191 locus causes high expression of hsa-mir-191 and promotes the epithelial-to-mesenchymal transition in hepatocellular carcinoma. Neoplasia 2011;13:841–53. [36] Brueckner B, Stresemann C, Kuner R, Mund C, Musch T, Meister M, et al. The human let-7a-3 locus contains an epigenetically regulated microRNA gene with oncogenic function. Cancer Res 2007;67:1419–23.

II.  Molecular Mechanisms of Epigenetics

58

4.  The Epigenetics of Noncoding RNA

[37] de Souza Rocha Simonini P, Breiling A, Gupta N, Malekpour M, Youns M, Omranipour R, et al. Epigenetically deregulated microRNA-375 is involved in a positive feedback loop with estrogen receptor alpha in breast cancer cells. Cancer Res 2010;70:9175–84. [38] Li A, Omura N, Hong S-M, Vincent A, Walter K, Griffith M, et al. Pancreatic cancers epigenetically silence SIP1 and hypomethylate and overexpress miR-200a/200b in association with elevated circulating miR-200a and miR-200b levels. Cancer Res 2010;70: 5226–37. [39] Lujambio A, Ropero S, Ballestar E, Fraga MF, Cerrato C, Setién F, et al. Genetic unmasking of an epigenetically silenced microRNA in human cancer cells. Cancer Res 2007;67:1424–9. [40] Wang P, Chen L, Zhang J, Chen H, Fan J, Wang K, et al. Methylation-mediated silencing of the miR-124 genes facilitates pancreatic cancer progression and metastasis by targeting Rac1. Oncogene 2014;33:514–24. [41] Agirre X, Vilas-Zornoza A, Jiménez-Velasco A, Martin-Subero JI, Cordeu L, Gárate L, et al. Epigenetic silencing of the tumor suppressor microRNA Hsa-miR-124a regulates CDK6 expression and confers a poor prognosis in acute lymphoblastic leukemia. Cancer Res 2009;69:4443–53. [42] Wilting SM, van Boerdonk RAA, Henken FE, Meijer CJLM, Diosdado B, Meijer GA, et al. Methylation-mediated silencing and tumour suppressive function of hsa-miR-124 in cervical cancer. Mol Cancer 2010;9:167. [43] Furuta M, Kozaki KI, Tanaka S, Arii S, Imoto I, Inazawa J. miR124 and miR-203 are epigenetically silenced tumor-suppressive microRNAs in hepatocellular carcinoma. Carcinogenesis 2009;31:766–76. [44] Pigazzi M, Manara E, Bresolin S, Tregnago C, Beghin A, Baron E, et al. MicroRNA-34b promoter hypermethylation induces CREB overexpression and contributes to myeloid transformation. Haematologica 2013;98:602–10. [45] Watanabe K, Emoto N, Hamano E, Sunohara M, Kawakami M, Kage H, et al. Genome structure-based screening identified epigenetically silenced microRNA associated with invasiveness in non-small-cell lung cancer. Int J Cancer 2012;130:2580–90. [46] Liu C, Kelnar K, Liu B, Chen X, Calhoun-Davis T, Li H, et al. The microRNA miR-34a inhibits prostate cancer stem cells and metastasis by directly repressing CD44. Nat Med 2011;17: 211–5. [47] Lodygin D, Tarasov V, Epanchintsev A, Berking C, Knyazeva T, Körner H, et al. Inactivation of miR-34a by aberrant CpG methylation in multiple types of cancer. Cell Cycle 2008;7:2591–600. [48] Toyota M, Suzuki H, Sasaki Y, Maruyama R, Imai K, Shinomura Y, et al. Epigenetic silencing of microRNA-34b/c and B-cell translocation gene 4 is associated with CpG island methylation in colorectal cancer. Cancer Res 2008;68:4123–32. [49] Roy S, Levi E, Majumdar AP, Sarkar FH. Expression of miR-34 is lost in colon cancer which can be re-expressed by a novel agent CDF. J Hematol Oncol 2012;5:58. [50] Chhabra R, Saini N. microRNAs in cancer stem cells: current status and future directions. Tumour Biol 2014;35:8395–405. [51] Kong D, Heath E, Chen W, Cher M, Powell I, Heilbrun L, et al. Epigenetic silencing of miR-34a in human prostate cancer cells and tumor tissue specimens can be reversed by BR-DIM treatment. Am J Transl Res 2012;4:14–23. [52] Saito Y, Liang G, Egger G, Friedman JM, Chuang JC, Coetzee GA, et al. Specific activation of microRNA-127 with downregulation of the proto-oncogene BCL6 by chromatin-modifying drugs in human cancer cells. Cancer Cell 2006;9:435–43. [53] Roman-Gomez J, Agirre X, Jiménez-Velasco A, Arqueros V, Vilas-Zornoza A, Rodriguez-Otero P, et al. Epigenetic regulation of MicroRNAs in acute lymphoblastic leukemia. J Clin Oncol 2009;27:1316–22.

[54] Seitz H, Youngson N, Lin SP, Dalbert S, Paulsen M, Bachellerie JP, et al. Imprinted microRNA genes transcribed antisense to a reciprocally imprinted retrotransposon-like gene. Nat Genet 2003;34:261–2. [55] Cui XS, Zhang DX, Ko YG, Kim NH. Aberrant epigenetic reprogramming of imprinted microRNA-127 and Rtl1 in cloned mouse embryos. Biochem Biophys Res Commun 2009;379:390–4. [56] Li J, Yang Z, Yu B, Liu J, Chen X. Methylation protects miRNAs and siRNAs from a 3′-end uridylation activity in Arabidopsis. Curr Biol 2005;15:1501–7. [57] Ren G, Xie M, Zhang S, Vinovskis C, Chen X, Yu B. Methylation protects microRNAs from an AGO1-associated activity that uridylates 5′ RNA fragments generated by AGO1 cleavage. Proc Natl Acad Sci USA 2014;111:6365–70. [58] Shen B, Goodman HM. Uridine addition after microRNA-directed cleavage. Science 2004;306:997. [59] Heo I, Joo C, Kim YK, Ha M, Yoon MJ, Cho J, et al. TUT4 in concert with Lin28 cuppresses cicroRNA biogenesis through pre-microRNA uridylation. Cell 2009;138:696–708. [60] Heo I, Ha M, Lim J, Yoon MJ, Park JE, Kwon SC, et al. Monouridylation of pre-microRNA as a key step in the biogenesis of group II let-7 microRNAs. Cell 2012;151:521–32. [61] Yuan S, Tang H, Xing J, Fan X, Cai X, Li Q, et al. Methylation by NSun2 represses the levels and function of miR-125b. Mol Cell Biol 2014;34:3630–41. [62] Berulava T, Rahmann S, Rademacher K, Klein-Hitpass L, Horsthemke B. N6-Adenosine Methylation in MiRNAs. PLoS One 2015;10:e0118438. [63] Alarcón CR, Lee H, Goodarzi H, Halberg N, Tavazoie SF. N6methyladenosine marks primary microRNAs for processing. Nature 2015;519:482–5. [64] Alarcón CR, Goodarzi H, Lee H, Liu X, Tavazoie S, Tavazoie SF. HNRNPA2B1 is a mediator of m(6)A-dependent nuclear RNA processing events. Cell 2015;162:1299–308. [65] Fu Y, Dominissini D, Rechavi G, He C. Gene expression regulation mediated through reversible m(6)A RNA methylation. Nat Rev Genet 2014;15:293–306. [66] Pan T. N6-methyl-adenosine modification in messenger and long non-coding RNA. Trends Biochem Sci 2013;38:204–9. [67] Braconi C, Huang N, Patel T. Microrna-dependent regulation of DNA methyltransferase-1 and tumor suppressor gene expression by interleukin-6 in human malignant cholangiocytes. Hepatology 2010;51:881–90. [68] Chhabra R. miRNA and methylation: a multifaceted liaison. ChemBioChem 2015;16:195–203. [69] Huang J, Wang Y, Guo Y, Sun S. Down-regulated microRNA-152 induces aberrant DNA methylation in hepatitis B virus-related hepatocellular carcinoma by targeting DNA methyltransferase 1. Hepatology 2010;52:60–70. [70] Majid S, Dar AA, Saini S, Shahryari V, Arora S, Zaman MS, et al. miRNA-34b inhibits prostate cancer through demethylation, active chromatin modifications, and AKT pathways. Clin Cancer Res 2013;19:73–84. [71] Fabbri M, Garzon R, Cimmino A, Liu Z, Zanesi N, Callegari E, et al. MicroRNA-29 family reverts aberrant methylation in lung cancer by targeting DNA methyltransferases 3A and 3B. Proc Natl Acad Sci USA 2007;104:15805–10. [72] Pan W, Zhu S, Yuan M, Cui H, Wang L, Luo X, et al. MicroRNA-21 and microRNA-148a contribute to DNA hypomethylation in lupus CD4+ T cells by directly and indirectly targeting DNA methyltransferase 1. J Immunol 2010;184:6773–81. [73] Garzon R, Liu S, Fabbri M, Liu Z, Heaphy CEA, Callegari E, et al. MicroRNA-29b induces global DNA hypomethylation and tumor suppressor gene reexpression in acute myeloid leukemia by targeting directly DNMT3A and 3B and indirectly DNMT1. Blood 2009;113:6411–8.

II.  Molecular Mechanisms of Epigenetics



59

REFERENCES

[74] Wang YS, Chou WW, Chen KC, Cheng HY, Lin RT, Juo SHH. Microrna-152 mediates dnmt1-regulated DNA methylation in the estrogen receptor α gene. PLoS One 2012;7:e30635. [75] Xiang Y, Ma N, Wang D, Zhang Y, Zhou J, Wu G, et al. MiR-152 and miR-185 co-contribute to ovarian cancer cells cisplatin sensitivity by targeting DNMT1 directly: a novel epigenetic therapy independent of decitabine. Oncogene 2013;33:378–86. [76] Wang H, Wu J, Meng X, Ying X, Zuo Y, Liu R, et al. Microrna-342 inhibits colorectal cancer cell proliferation and invasion by directly targeting DNA methyltransferase 1. Carcinogenesis 2011;32:1033–42. [77] Tsuruta T, Kozaki K-I, Uesugi A, Furuta M, Hirasawa A, Imoto I, et al. miR-152 is a tumor suppressor microRNA that is silenced by DNA hypermethylation in endometrial cancer. Cancer Res 2011;71:6450–62. [78] Blasco M, Benetti R, Gonzalo S, Jaco I, Muñoz P, Gonzalez S, et al. A mammalian microRNA cluster controls DNA methylation and telomere recombination via Rbl2-dependent regulation of DNA methyltransferases. Nat Struct Mol Biol 2008;15:268–79. [79] Sinkkonen L, Hugenschmidt T, Berninger P, Gaidatzis D, Mohn F, Artus-Revel CG, et al. MicroRNAs control de novo DNA methylation through regulation of transcriptional repressors in mouse embryonic stem cells. Nat Struct Mol Biol 2008;15:259–67. [80] Zhao S, Wang Y, Liang Y, Zhao M, Long H, Ding S, et al. MicroRNA-126 regulates DNA methylation in CD4+ T cells and contributes to systemic lupus erythematosus by targeting DNA methyltransferase 1. Arthritis Rheum 2011;63:1376–86. [81] Lal A, Thomas MP, Altschuler G, Navarro F, O’Day E, Li XL, et al. Capture of microRNA-bound mRNAs identifies the tumor suppressor miR-34a as a regulator of growth factor signaling. PLoS Genet 2011;7:e1002363. [82] Baek D, Villén J, Shin C, Camargo FD, Gygi SP, Bartel DP. The impact of microRNAs on protein output. Nature 2008;455:64–71. [83] Krishnan K, Steptoe AL, Martin HC, Wani S, Nones K, Waddell N, et al. MicroRNA-182-5p targets a network of genes involved in DNA repair. RNA 2013;19:230–42. [84] Dominissini D, Moshitch-Moshkovitz S, Schwartz S, SalmonDivon M, Ungar L, Osenberg S, et al. Topology of the human and mouse m6A RNA methylomes revealed by m6A-seq. Nature 2012;485:201–6. [85] Hussain S, Aleksic J, Blanco S, Dietmann S, Frye M. Characterizing 5-methylcytosine in the mammalian epitranscriptome. Genome Biol 2013;14:215. [86] Zhou KI, Parisien M, Dai Q, Liu N, Diatchenko L, Sachleben JR, et al. N6-methyladenosine modification in a long noncoding RNA hairpin predisposes its conformation to protein binding. J Mol Biol 2016;428:822–33. [87] Warren L, Manos PD, Ahfeldt T, Loh YH, Li H, Lau F, et al. Highly efficient reprogramming to pluripotency and directed differentiation of human cells with synthetic modified mRNA. Cell Stem Cell 2010;7:618–30. [88] Mercer TR, Dinger ME, Mattick JS. Long non-coding RNAs: insights into functions. Nat Rev Genet 2009;10:155–9. [89] Wang KC, Yang YW, Liu B, Sanyal A, Corces-Zimmerman R, Chen Y, et al. A long noncoding RNA maintains active chromatin to coordinate homeotic gene expression. Nature 2011;472:120–4. [90] Wu SC, Kallin EM, Zhang Y. Role of H3K27 methylation in the regulation of lncRNA expression. Cell Res 2010;20:1109–16. [91] Li Y, Zhang Y, Li S, Lu J, Chen J, Wang Y, et al. Genome-wide DNA methylome analysis reveals epigenetically dysregulated non-coding RNAs in human breast cancer. Sci Rep 2015;5:8790. [92] Hahn MA, Wu X, Li AX, Hahn T, Pfeifer GP. Relationship between gene body DNA methylation and intragenic H3K9ME3 and H3K36ME3 chromatin marks. PLoS One 2011;6:e18844. [93] Navarro P, Page DR, Avner P, Rougeulle C. Tsix-mediated epigenetic switch of a CTCF-flanked region of the Xist promoter

[94]

[95] [96]

[97]

[98] [99]

[100]

[101]

[102]

[103] [104]

[105]

[106]

[107]

[108]

[109]

[110]

[111]

[112]

[113]

determines the Xist transcription program. Genes Dev 2006;20:2787–92. Sun BK, Deaton AM, Lee JT. A transient heterochromatic state in Xist preempts X inactivation choice without RNA stabilization. Mol Cell 2006;21:617–28. Lee JT, Lu N. Targeted mutagenesis of Tsix leads to nonrandom X inactivation. Cell 1999;99:47–57. Di Ruscio A, Ebralidze AK, Benoukraf T, Amabile G, Goff LA, Terragni J, et al. DNMT1-interacting RNAs block gene-specific DNA methylation. Nature 2013;503:371–6. Merry CR, Forrest ME, Sabers JN, Beard L, Gao XH, Hatzoglou M, et al. DNMT1-associated long non-coding RNAs regulate global gene expression and DNA methylation in colon cancer. Hum Mol Genet 2015;24:6240–53. Lai F, Shiekhattar R. Where long noncoding RNAs meet DNA methylation. Cell Res 2014;24:263–4. Zhou J, Yang L, Zhong T, Mueller M, Men Y, Zhang N, et al. H19 lncRNA alters DNA methylation genome wide by regulating Sadenosylhomocysteine hydrolase. Nat Commun 2015;6:10221. Tehlivets O, Malanovic N, Visram M, Pavkov-Keller T, Keller W. S-adenosyl-l-homocysteine hydrolase and methylation disorders: yeast as a model system. Biochim Biophys Acta—Mol Basis Dis 2013;1832:204–15. Berghoff EG, Clark MF, Chen S, Cajigas I, Leib DE, Kohtz JD. Evf2 (Dlx6as) lncRNA regulates ultraconserved enhancer methylation and the differential transcriptional control of adjacent genes. Development 2013;140:4407–16. Brennecke J, Malone CD, Aravin AA, Sachidanandam R, Stark A, Hannon GJ. An epigenetic role for maternally inherited piRNAs in transposon silencing. Science 2008;322:1387–92. Henderson IR, Jacobsen SE. Epigenetic inheritance in plants. Nature 2007;447:418–24. Verdel A, Vavasseur A, Le Gorrec M, Touat-Todeschini L. Common themes in siRNA-mediated epigenetic silencing pathways. Int J Dev Biol 2009;53:245–57. Sai lakshmi S, Agrawal S. piRNABank: a web resource on classified and clustered Piwi-interacting RNAs. Nucleic Acids Res 2008;36. Volders PJ, Verheggen K, Menschaert G, Vandepoele K, Martens L, Vandesompele J, et al. An update on LNCipedia: a database for annotated human lncRNA sequences. Nucleic Acids Res 2015;43:D174–80. Lestrade L, Weber MJ. snoRNA-LBME-db, a comprehensive database of human H/ACA and C/D box snoRNAs. Nucleic Acids Res 2006;34:D158–62. Pandi G, Nakka VP, Dharap A, Roopra A, Vemuganti R. MicroRNA miR-29c down-regulation leading to de-repression of its target DNA methyltransferase 3a promotes ischemic brain damage. PLoS One 2013;8:e58039. Qadir XV, Han C, Lu D, Zhang J, Wu T. MicroRNA-185 inhibits hepatocellular carcinoma growth by targeting the DNMT1/ PTEN/Akt pathway. Am J Pathol 2014;184:2355–64. Mohammad F, Mondal T, Guseva N, Pandey GK, Kanduri C. Kcnq1ot1 noncoding RNA mediates transcriptional gene silencing by interacting with Dnmt1. Development 2010;137:2493–9. Kotake Y, Nakagawa K, Kitagawa K, Suzuki S, Liu N, Kitagawa M, et al. Long non-coding RNA ANRIL is required for the PRC2 recruitment to and silencing of p15(INK4B) tumor suppressor gene. Oncogene 2011;30:1956–62. Nagano T, Mitchell JA, Sanz LA, Pauler FM, Ferguson-Smith AC, Feil R, et al. The Air noncoding RNA epigenetically silences transcription by targeting G9a to chromatin. Science 2008;322: 1717–20. Gupta Ra, Shah N, Wang KC, Kim J, Horlings HM, Wong DJ, et al. Long non-coding RNA HOTAIR reprograms chromatin state to promote cancer metastasis. Nature 2010;464:1071–6.

II.  Molecular Mechanisms of Epigenetics

C H A P T E R

5 Prions and Prion-Like Phenomena in Epigenetic Inheritance Tinh-Suong Nguyen*, Hervé Lalucque*, Fabienne Malagnac**, Philippe Silar* *Paris Diderot University, Paris, France; **University of Paris-Saclay, Paris, France

O U T L I N E Structural Heredity Prions of Saccharomyces cerevisiae and Podospora anserina Self Driven Assembly of Hsp60 Mitochondrial Chaperonin Cytotaxis of Cilia and Other Complex Structures

Crippled Growth, a Self-Sustained and Mitotically Inheritable Signaling Pathway in the Filamentous Fungus Podospora anserina The White/Opaque Switch of Candida albicans, an Epigenetic Switch at the Transcription Level

62 62 65 65

Mixed Heredity: A Prion That Propagates by Covalent Autoactivation

65

Regulatory Inheritance The Lactose Operon and its Positive Feedback Loop

66 66

This discussion is based on the idea of two types of cellular regulatory systems, both capable of maintaining persistent cellular characteristics but achieving homeostasis by different means. The current concept of a primary genetic material (DNA), replicating by a template mechanism, is opposed to a homeostatic system operating by, perhaps, self-regulating metabolic patterns. —D.L. Nanney, 1958 [1]

67

Conclusions

68

References

69

acid-bearing entities in the cytoplasm. Theoretical considerations, first made by Max Delbrück [2], proposed that negative-interacting metabolic networks could generate alternative states, stable enough to be passed on during cell division. Similarly, as early as 1961, a model based on structure inheritance was proposed by Marcou and Rizet to account for a case of non-mendelian inheritance in the fungus Podospora anserina [3]. Early experiments with the lactose operon of Escherichia coli [4], proved that indeed metabolic networks could generate inheritable alternative metabolic states, and studies with paramecia showed that complex subcellular structures, that is; cilia, which direct their own assembly in a template-assisted fashion [5], could create alternative states that were inheritable during cell division and even sexual reproduction. This led to the definition of two kinds of epigenetic inheritance [6]: the structural inheritance, based on

In the early period of the 20th century, almost as soon as the laws of Mendel were rediscovered, characters that would not follow the rules of classical mendelian segregation were discovered. Most of these cases of non-mendelian heredity are presently accounted for by mutations in eukaryotic organelle genomes (plastids and mitochondria), by cytoplasmic symbionts or by viruses and virus-like particles. However, a subset of these phenomena cannot be explained by the presence of nucleic Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00005-5 Copyright © 2017 Elsevier Inc. All rights reserved.

67

61

62

5.  Prions and Prion-Like Phenomena in Epigenetic Inheritance

the transmission of alternative structures of macromolecules and macromolecular complexes, and the regulatory inheritance, based on the alternative states adopted by metabolic or regulatory pathways. As we shall see later, such a clear-cut difference may not be made in some instances, which are clearly a mix of the two types. It is interesting to note that the acceptance that DNA, and not proteins, is the genetic material took a long time. Purification of DNA associated with genetic transformation was the key experiment that permitted final recognition. The discovery in the mid-1960s of mitochondrial and plastid genomic DNAs [7,8] has then eclipsed for 3 decades the studies on many “genes” with unorthodox segregation, that could be due to structural or regulatory inheritance. Their analysis was reignited by the proposal of R. Wickner in 1994 [9] that two of them, the [PSI+] and [URE3] elements of the yeast Saccharomyces cerevisiae, could be due to inheritable changes in protein structure. At that time, the concern with the mad-cow disease, which affected a large part of the cattle in the United Kingdom and some other European countries, whose etiologic agent appears to be composed only of proteins, made the scientific community more receptive to unorthodox ideas regarding inheritance. As we shall see, we now can transform cells to alternative “states” with purified prions, a feat that would have postponed the recognition of DNA as the genetic material, and reinforced the hypothesis that proteins were the genetic material, would one of the unorthodox “gene” have been chosen in the pioneering transformation experiments. Here, we will review only a few cases of structural and regulatory inheritance due to space constraint. Indeed, an ever-increasing array of phenomena is now attributed to prions and prion-like elements. Only those prototypic are discussed in the following sections.

STRUCTURAL HEREDITY At the present time, two different kinds of structural heredity have been clearly demonstrated, that of prions, which is based on the structural changes in a single polypeptide, and cytotaxis in which a large macromolecular complex is concerned.

Prions of Saccharomyces cerevisiae and Podospora anserina The term “prion” for proteinaceous infectious particle was first proposed by Prusiner [10] to characterize the etiologic agents of some, at the time, bizarre diseases of mammals called transmissible spongiform encephalopathies (TSEs), including scrapie in sheep, as well as Kuru, Creutzfeldt–Jakob in human and later on the mad cow diseases. A basic definition of this term is the following: a

protein able to adopt two distinct conformations, one of which can convert the other one. Usually, prion proteins may adopt monomeric or oligomeric states. The protein can change spontaneously from the monomeric to the oligomeric forms with a low frequency. Importantly, oligomers trigger the switch of the monomers toward the oligomeric and infectious form. This autocatalytic process leads thus to the depletion of the monomers and to the accumulation of the oligomers. TSEs are probably caused by an aberrant folding of the PrPc protein into the infectious PrPSc form [11]. Afterward, this concept was successfully extended to explain the peculiar features of two non-mendelian elements of S. cerevisiae, [PSI+] and [URE3] [9]. Now, the term prion is no longer restricted to TSE agents, however, refers to any protein able to adopt an infectious conformation. The prion transition alters the function of the protein and consequently the phenotype of the cell. The aggregated state, as well as the associated phenotype, is infectious and stably transmitted from generations to generations both by mitosis and meiosis. Thus, yeast prions act as protein-based genetic element corresponding to an elegant epigenetic heredity. Several genetic criteria have been retained to suggest a prion behavior for a cellular protein [12]: (1) a prion can be cured, but it can reappear in the cured strain with a constant frequency because the protein able to change to an infectious form is still present, (2) overproduction of a protein capable of becoming a prion increases the frequency of the prion arising de novo, and (3) if the prion phenotype is due to the absence of the normal form of the protein, then the phenotype of null mutant of the gene for the protein is the same as that of the strain containing the aggregated prion. Also, this gene is required for the prion to propagate. The first two yeast proteins obeying the previously mentioned criteria and showed to be true prions in yeast were eRF3 and Ure2p. [PSI+], the prion of the release factor eRF3, also called Sup35p, affects the efficiency of translation termination. This may result in significant morphological or physiological switch when the transition is made [13]. [URE3], the prion of the protein Ure2p alters nitrogen catabolism. Although, the genetic, biochemical, and cell biological analysis of these two prions, especially [PSI+], have boosted the comprehension of prion properties, the definitive demonstration that a protein is infectious based on cell transformation was obtained in the ascomycete fungus P. anserina [14]. This organism contains a true prion that displays the expected properties, the HET-s protein involved in heterokaryon incompatibility [15]. Prion aggregates of HET-s obtained from recombinant protein made in E. coli were introduced by ballistic transformation into prion-free cells of P. anserina and were shown to induce a phenotypic conversion toward the prion containing cells [14]. This demonstration is formally equivalent to

II.  Molecular Mechanisms of Epigenetics



63

Structural Heredity

TABLE 5.1 Summary of the Prion Proteins in Different Organisms Protein

Organism

Prion

References

PrP

Mammals

TSE agent

[11]

HET-s

P. anserina

HET-s

[14]

+

Sup35

S. cerevisiae

[PSI ]

[17,18,30]

Ure2

S. cerevisiae

[URE3]

[31]

Rnq1 New1 Swi1 Cyc8 MOT3 SFP1 Mod5 Nup100

S. cerevisiae

+

[19]

+

[PIN ]

S. cerevisiae

[NU ]

[32]

S. cerevisiae

+

[SWI ]

[33]

S. cerevisiae

+

[OCT ]

[34] +

S. cerevisiae

[MOT3 ]

[35]

S. cerevisiae

+

[36]

S. cerevisiae S. cerevisiae

[ISP ] +

[MOD ]

[37] +

[NUP100 ] +

[38]

Pma1 (?)

S. cerevisiae

[GAR ]

[29]

CPEB

A. californica

—

[21]

LD

A. Thaliana

—

[24]

Two studies suggest that dozen additional yeast proteins, not indicated in the table, may also be prion [32,35].

transformation experiments conducted by Avery et al. [16] proving that DNA is the support of genetic information. To date, this kind of transformation demonstration was performed for only few additional examples in S. cerevisiae [17–20], but an ever-growing number of S. cerevisiae proteins are strongly suspected to be prions (Table 5.1). Moreover, proteins behaving as prions when expressed in S. cerevisiae have been found in animals [21–23] and more recently in plants [24]. The CPEB proteins of Aplysia and mouse have been shown to adopt prion conformation in neurons [23,25] and the Aplysia protein to undergo typical structural switch into β-sheetrich fibers [26]. Other prions may include the amyloid-β and tau proteins involved in amyloid fibrils in Alzheimer diseases, although this is still a subject of contention [27]. Conversely, some genetic traits behaving like prion phenomenon are still mysterious. These not only include the [NSI+] determinant that controls translation accuracy in S. cerevisiae in a fashion reminiscent of [PSI+] [28], but also of [GAR+] [29] that affects carbon source utilization. Domain analysis showed that the yeast and P. anserina prions contain a modular prion domain, dispensable for the cellular function but required and sufficient for the prion properties [39,40]. Most proteins fused to it behave like a prion. For example, in cells expressing a fusion protein composed of a prion domain and the GFP protein, two populations of cells may be observed, one with a homogenous cytoplasmic fluorescence corresponding to the monomeric form and the other displaying intense punctuated foci due to oligomerization of the fusion

protein [40]. Studies have focused on these domains to detect important features. In S. cerevisiae, an important feature is richness in glutamines (Q) and asparagines (N), since all known yeast prion have such a domain. The specific primary sequence is probably not critical, because a prion domain with a sequence randomly shuffled is still able to form prion [41]. A bioinformatic analysis of the yeast genome performed to select gene coding a protein with a Q/N rich domain [42] permitted the identification of the Rnq1 protein (rich in N and Q) that has prion properties, confirming the importance of this feature. However, it may not be universal, since the prion domains of Het and PrP are not Q/N rich and exhibit no obvious bias in their amino-acid composition. Interestingly, other parts of the protein, not included in the prion domain sensu stricto, may be required for the stability of prions in vivo by allowing efficient transmission of prion aggregates during cell divisions [43]. Despite their variation in primary sequences, all prions appear to adopt a similar conformation into fibrils (see Ref. [44] for a review). The monomer is soluble, rich in α-helix and protease sensitive. The infectious form is oligomeric, rich in β-sheet and partially protease resistant. The oligomers of all known prions appear to be able to form amyloid fibers. This fibrous aggregates, identified by a birefringence when stained with Congo Red, are characteristically composed of proteins rich in β− sheet structures. These kinds of fibers are detected in prion disease in mammals in the brain of affected individuals and also in vitro for recombinant yeast prion

II.  Molecular Mechanisms of Epigenetics

64

5.  Prions and Prion-Like Phenomena in Epigenetic Inheritance

proteins. Note that these fibers are not restricted to prion proteins; indeed some native amyloïds exist like the curlin protein in E. coli and even in the silk of some spiders. Progresses are made in the understanding of the structural basis of infectivity [45]. To date only the structure of the Het-s prion has been resolved [46,47] and key residues and inhibitory domains enabling or preventing the folding and polymerization have been identified [48,49]. As expected for a mechanism based on protein conformations, chaperones are implicated in the stability of prions. Hsp104, Hsp70, and Hsp40 families of chaperone but also Hsp70 or Hsp40 activating proteins are involved in this regulation. Hsp104 breaks aggregates into smaller ones: Hsp40 and Hsp70 help to refold the proteins into their native conformation. Hsp104p was the first chaperone identified to modulate prion [50]. Surprisingly both overexpression and depletion of Hsp104 led to the loss of [PSI+]. Hsp104 is required to break large aggregates of prion producing new oligomers efficient for polymerization or seeds. The normal level of Hsp104 permits the retention of seeds large enough to not be refolded by Hsp40/Hsp70, as well as, to have an adequate number of seeds for efficient segregation during cellular division. Among chaperones, Hsp104 appears to be the major player as it is required for all the yeast prions. Moreover, the first compound destabilizing [PSI+], guanidium hydrochlorid, is an inhibitor of Hsp104 [51]. However, Hsp104 overexpression affects only [PSI+], indicating a complex role for Hsp104 with a general effect on yeast prions and another activity specific to [PSI+]. The effect of overexpression/depletion of members of the families Hsp40 and Hsp70 on the prion stability is more complex than those of Hsp104 (see Ref. [52] for a review). The search for others players modulating prion stability provided one of the most astonishing results in prion studies: the appearance of [PSI+] is itself controlled by another prion [53]. In absence of this prion, called [PIN+] for PSI-inducible in yeast cells, PSI is unable to appear. This PIN element is usually due to the prion conformation of Rnq1, however, experimentally several other prions including [URE3] permit the formation of [PSI+] independently of Rnq1 [32]. The effect of [PIN+] is not limited to [PSI+], as it also influences the appearance of [URE3] [54]. This kind of interactions seems to be not restricted to positive ones as it is suggested that [PSI+] and [URE3] antagonize each other propagation and de novo appearance [55]. Destabilizing interactions between [PSI+] and [PIN+] have also been described [56]. Additionally, the New1 protein responsible for the [NU+] prion is able to cleave the [PSI+] amyloid fibers, suggesting another level of complex interaction between prion proteins in both polymerized and nonpolymerized forms [57]. These data exemplify interactions between prions in a manner similar to interactions between alleles of different genes, emphasizing the similarity between true genes and “prion genes”.

The similarity of prions with true genes is even more pronounced, since “allelic variants” called strains have been discovered for several prions [18,54,58–60]. Indeed, the observation that cells with the prion [PSI+] may present different stabilities of [PSI+] during cell division and different expressivities of the phenotype led to the proposal that strains of [PSI+] exist [58]. The same observation was also made with TSE diseases that would appear to be caused by various strains of infectious agents [60–62]. The question is how a protein able to switch to an inactive and infectious state may be connected to several phenotypes. The answer lies in the ability of a unique prion protein to adopt many distinct infectious conformations [17,18]. As the first aggregated proteins appear spontaneously, one conformation is adopted among many possibilities. Then, this conformation acts as a template for further aggregation and is accurately transmitted to the successive cycles of conversion. Each conformation gives birth to a strain or variant of prion presenting distinct properties: stability, structure of the amyloid fibers, number of seeds, but also proportion of the prion protein in the aggregated form versus the soluble form [17,18,63]. At the level of the organism, these different strains trigger phenotypes with a more or less pronounced severity or effects. Interestingly, some are able to act as template on proteins with a different primary structure. This enables for example, some of the conformers from one species, but not all of them, to transmit its conformation to the homologous prion protein from a different species, a process known as speciesbarrier crossing [64]. Prions are clearly endowed with the ability to transmit information from one cell generation to the next one and importantly the aggregated form can be purified and used in transformation experiments. Some are even responsible for transmitting diseases in mammals, clearly demonstrating their stability outside cells or organisms. They should thus be considered as true “hereditary units” in their ability to carry genetic information. At the present time, their actual role in cell physiology is unclear. Mammalian prions are clearly detrimental infectious agents. The P. anserina prion may be regarded as beneficial or detrimental [15,65,66], while, in yeast, prions are regarded either as enabling adaptation [13,37,67,68] or diseases [69–71]. Indeed, for most prions, only the monomeric form has a biological activity. Exceptions to this rule includes HET-s, which appears active only in the oligomeric and infectious conformation [15,72]. However, for [MOD+], polymerization inactivates an isopentenyl transferase, regulating sterol biosynthesis and thereby triggering resistance to antifungal compounds, suggesting that both the prion and nonprion form have some biological activity [36]. The most interesting example of a functional role for a prion protein is that of Aplysia CPEB [21], a neuronal mRNA translation regulator. Indeed,

II.  Molecular Mechanisms of Epigenetics



Mixed Heredity: A Prion That Propagates by Covalent Autoactivation

65

this protein is likely involved through its prion switch in stabilization of long-term memory [73].

in the human mHsp60 gene lead to the development of neurodegenerative disorders (see [76] for a review).

Self Driven Assembly of Hsp60 Mitochondrial Chaperonin

Cytotaxis of Cilia and Other Complex Structures

Unlike prions, which can be viewed as abnormal proteins, the S. cerevisiae Hsp60 chaperonin provides a clear example of a structure catalyzing its own folding. It demonstrates the necessity for correctly folded preexisting oligomers to ensure the correct folding of further monomers [74]. Some proteins, which are imported from the cytosol into mitochondria, cross the mitochondrial membranes in an unfolded conformation and then are folded in the matrix by Hsp60. Monomers of Hsp60 form a complex, arranged as two stacked 7-mer rings. Once assembled in the matrix of mitochondria, these 14-mer complexes bind unfolded proteins to catalyze their proper folding in an ATP-dependent manner. However, Hsp60 proteins are also encoded by a nuclear gene and translated in the cytosol as precursors, which are then translocated into the mitochondrial matrix. So, how could they assemble themselves without preexisting 14-mer complexes to fold them? To address this question, Cheng et al. took advantage of a temperature-sensitive lethal mutation in the hsp60 gene [74]. At 23°C, the hsp60ts mutant cells grew normally, however, when the temperature was shifted to 37°C, the mutant cells stopped to grow within one generation, because the impaired Hsp60 complex fails in folding and assembly of imported mitochondrial proteins. An attempt to rescue the growth deficient phenotype of hsp60ts mutant strains was set up with a high copy plasmid, containing the coding sequence of the wild-type Hsp60 precursor, driven by the inducible galactose promoter. Cultures were first shifted from 23 to 37°C and 2 h later; expression of wild-type hsp60 was induced by addition of galactose. In these conditions, the growth deficient phenotype of the mutant strain was not rescued. However, strikingly, when expression of wild-type Hsp60 subunits was induced by addition of galactose for 2 h before the temperature shift to 37°C, the mutant cells could grow. This means that wild-type hsp60 complexes, can rescue the mutant phenotype at restrictive temperature only when expressed at permissive temperature, indicating that Hsp60 is required for its own assembly. In another words, newly Hsp60 imported subunits can be assembled only if preexisting Hsp60 complexes are present in the matrix of mitochondria. More generally, this study strongly suggests that biogenesis of organelles, such as mitochondria is probably not a de novo process, but rather relies on preexisting structures, acting as a template. If this template is lost along the path, although the protein subunits are produced, no functional organelles would be made. In mammals, the Hsp60 chaperonin is also essential for viability [75] and some mutations

Prions and Hsp60 are homopolymers of a single protein. However, in the cells, most structures are built up from several different polypeptides and additional molecules, such as RNAs, cofactors, and so on. While in many cases it has been shown that these complex structures are able to correctly fold themselves spontaneously, often with the help of chaperones, in other cases the preexistence of some structural information is necessary to obtain a correct organization. This was first shown by Beisson and Sonneborn on the orientation of cilia in paramecium [5]. Ciliates, like paramecium, are large cells that display a complex organization. Especially, their cortex is endowed with cilia that are all oriented in the same direction, permitting efficient swimming. Beisson and Sonneborn “grafted” in an inverted orientation rows of cilia in Paramecium aurelia, producing “variants” with abnormal swimming behavior [5]. These variants could be maintained over 800 mitotic generations and maintained in sexual crosses. This heritability is due to the fact that preexisting cilia direct the correct insertion and orientation of newly formed cilia. This process whereby an old cellular structure orders a new one was called cytotaxis [77]. Cytotaxis of cilia/flagella have also been described in Tetrahymena [78] Chlamydomonas [79], and Trypanosoma [80]. Additional examples for other structures have been described in Paramecium, including handedness [81,82] and doublets (see Ref. [83] for a review). Even in S. cevevisiae, cortical inheritance has been described [84]. The mechanism involved in cilium insertion in Paramecium has been analyzed at the ultrastructural level [85] and mutant searches [83] have uncovered Tetrahymena, and Paramecium nuclear mutants with altered cortical elements, which should enable a full understanding at the molecular level how the old structure directs the construction of the new one.

MIXED HEREDITY: A PRION THAT PROPAGATES BY COVALENT AUTOACTIVATION In this section, we will present one process that stands at the border between structural and regulatory inheritance. It is based on the inheritable autocatalytic cleavage of a protease [86]. The yeast protease B, PrB, is a subtilisin/furin class serine protease derived from a larger, catalytically inactive proform encoded by the gene PRB1. The final steps in the maturation of the proenzyme PrB are sequential truncations occurring in the lysosome-like yeast vacuole, catalyzed by protease A, PrA, and finally PrB itself. Mature PrB

II.  Molecular Mechanisms of Epigenetics

66

5.  Prions and Prion-Like Phenomena in Epigenetic Inheritance

protease activates other vacuolar hydrolases, such as carboxipeptidase Y (CpY), whose activity can be easily assayed. Deletion of PEP4, encoding the PrA precursor leads to accumulation of the immature form of PrB and therefore loss of its activity, as seen by lack of CpY activity. However, the disappearance of mature PrB after deletion of PEP4 is progressive. CpY activity can be detected in pep4∆ strains for more than 20 mitotic generations. This hysteresis of PrB activity is referred to as “phenotypic lag” and it is believed to reflect dilution during growth of PEP4 mRNA and PrA protease. The phenotypic lag was initially observed during growth on dextrose medium because dextrose represses PRB1 transcription. Roberts and Wickner [86] tested if this lag might be prolonged after transfer onto glycerol medium, which does not repress PRB1 expression. The authors sporulated a diploid heterozygous for PEP4 deletion and germinated the meiotic products on glycerol medium. They found that CpY remained active indefinitely, even in the colonies derived from the pep4∆ spores. However, when transferred onto dextrose medium, which represses PRB1 transcription, CpY activity of these pep4∆ cells was progressively lost and not restored by a return to glycerol medium. The “PrB+” state expressing the CpY hydrolase is infectious during cytoduction experiments (i.e., cytoplasmic mixing without caryogamy), even when both donor and recipient pep4∆ strains are grown on glycerol. This was the demonstration that the PrB+ state is triggered by a cytoplasmic and infectious factor called [β]. They further showed that the cells mutated for PrB do not contain [β] and that the overexpression of PRB1 increases the frequency of [β] appearance in pep4∆ cells that have been previously cured by extended growth on dextrose medium. In this system, PrpA is only needed for the initial conversion of PrB in the absence of [β]. To the authors, the fairly unconventional behavior of the PrB+ state is reminiscent to that of structure-based prion, indicating that any enzyme could be a prion, provided that its activity depends on self-modification in trans and that there is a mechanism by which it can be transmitted from individual to individual [87]. As we will see later, an additional example of such behavior exhibited by kinases involved in signaling indicates that this is indeed the case. However, because these kinases are involved in regulation, we will discuss them in the next section dealing with regulatory inheritance.

REGULATORY INHERITANCE There is now a large body of literature dealing with the behavior of regulatory networks, especially their ability to generate emergent properties enabling cells to finely tune their response to various environmental changes (see Ref. [88–90] for reviews). In some cases, these properties result in the generation of bistable states that are inheritable in

a more or less faithful fashion [91–96]. Below we discuss three examples of such regulatory inheritance.

The Lactose Operon and its Positive Feedback Loop In their seminal studies of 1956 [4], Novick and Weiner, and later on Cohn and Horibata [97–99], showed that under defined conditions, it is possible to obtain an epigenetic inheritability of the activation status of the lactose operon in E. coli. Indeed, when grown at low concentrations of a gratuitous inducer, E. coli are either not induced for their lactose operon or fully induced and are never found in an intermediate state. Noninduced cells can change spontaneously toward the induced state with a constant probability. Thus, when transferred from a medium lacking the inducer toward a medium containing the inducer, the population accumulates more and more cells that have made the transition toward the active state. These cells do not invade the population since the induced cells grow more slowly than the noninduced ones, permitting a dynamic equilibrium in the population. That both the on and off states are inheritable was demonstrated by diluting an equilibrated population in new medium that contained the inducer at an inoculum of one cell per new culture. Two kinds of cultures where obtained, one composed of fully induced cells that originated from a bacterium that was already induced and the other which accumulated induced cells at the level of the parental culture before dilution. The cause of this behavior is the presence of a positive regulatory loop in the lactose regulation, whereby the entrance of enough inducer inside the cell activates the operon and especially the production of permease, which in turn allows more inducer to enter the cell. It was recently demonstrated that the stochastic complete dissociation of the lactose repressor, which binds as a tetramer, triggers the initial burst of production of the permease [100]. Another piece of information gained from the study of the lactose operon is that the behavior of the cells is strongly dependent upon the inducer and glucose concentrations [101]. Glucose represses the operon while the inducer activates it. Three main regulatory behaviors can be adopted by the cells, (1) monostable induced at low glucose and high inducer concentrations, (2) monostable uninduced at high glucose and low inducer lever, and (3) the bistable state described previously at intermediate concentrations. Importantly, cells that are placed in these intermediate concentrations will behave differently if they originate from the monostable induced or monostable uninduced conditions. This lactose operon is prototypic of systems with a positive autoregulatory loop and mathematic models describing their properties are available [101–103]. More, complex inheritable units could be envisioned. They are all based on positive autoregulatory loops or its

II.  Molecular Mechanisms of Epigenetics



Regulatory Inheritance

derivative, the reciprocal repression as the one present in the C1/Cro interaction of the lambda phage. That these transcription factors that negatively regulate the other are able to produce epigenetic inheritable switch is known for a long time [95,104]. Readers interested in the emergent properties of regulatory networks, including epigenetic inheritance can refer to a number of excellent reviews [88–90]. Most of the researches carried out today are performed on man-made regulatory networks. We discuss below two examples of regulatory inheritance encountered in wild organisms, the sectors of filamentous fungi and the yeasts phenotypic switches.

Crippled Growth, a Self-Sustained and Mitotically Inheritable Signaling Pathway in the Filamentous Fungus Podospora anserina P. anserina is a saprophytic filamentous fungus used as a model organism for decades. In the 1990s, Silar and coworkers noticed that sectors of altered growth could be seen on P. anserina growing thalli [105]. This cell degeneration phenomenon, called crippled growth (CG), was easily visible macroscopically, displaying highly pigmented, flat, and female-sterile mycelium as opposed to normal growth (NG). Curiously, the development of these sectors occurred only in special genetic or environmental conditions [105,106]. The switch is controlled in both directions by environmental stimuli [105]. It was rapidly demonstrated that no nucleic acid was involved in the genesis of these sectors and that the presence of C, a cytoplasmic and infectious factor was associated with CG. The mycelium of P. anserina can thus exhibit a bistability at the morphological level. Similar phenomena were previously described and reported to be very frequent in filamentous ascomycetes (see Ref. [107] for a review). They were generally due to the presence in the cell of cytoplasmic and infectious factors, whose properties appear strikingly similar to prions. Apart from CG, only the “secteur” phenomenon of Nectria haematococca has been studied [108], but in this instance, no clear model on how the infectious factor is generated is presently available. In the case of CG, a genetic analysis [106] permitted to retrieve numerous genes, which are required to produce C. Some of these “IDC” genes (impaired in the development of CG) were cloned and showed to encode a MAP kinase kinase kinase (MAPKKK) [109] and a MAP kinase kinase (MAPKK) [110]. These two proteins are members of a large family of kinases present in all eukaryotes. They act in a sequential manner, that is, the MAPKKK activates by phosphorylation the MAPKK, which in turn activates by phopshorylation a MAP kinase (or MAPK). These IDC mutants unable to produce C were null mutants of either the MAPKKK or MAPKK genes. Further genetic inactivation of the gene coding the downstream MAP kinase also showed it to be a key element in the genesis of C [110]. Moreover, overexpression of the MAPKKK

67

and MAPK was shown to facilitate the development of CG [109,110]. (1) Presence of a cytoplasmic and infectious factor, (2) necessity of a gene for its propagation, and (3) increased frequency of appearance of the infectious factor when the gene is overexpressed are properties exhibited by genes coding for prions. Here, the three genes coding the MAPKKK, MAPKK, and MAPK display these properties. A model related to that of prions but based on an autocatalytic activation loop in the MAPK cascade has thus been proposed to account for the C element [109,110]. Some element(s) downstream of the MAPK would be able to activate directly or indirectly the upstream MAPKKK. In this model, the C element corresponds to components of the cascade in the active state, which are able to activate in trans other molecules that are in the inactive state. This results in the complete conversion of the inactive factors to their active form. This strikingly resembles the ability of prions to promote their own aggregation or that of the [β] “prion” in S. cerevisiae to promote its own maturation. This model is supported by experiments in Xenopus eggs, in which it was demonstrated that the presence of a positive self-regulation in the p42 MAP kinase cascade entails the presence of only two states: one in which no active MAPK is present and the other in which all MAPK molecules are active, the intermediate states being transient [111]. It was also shown that transfer of cytoplasm from an activated egg to an inactivated one results in complete activation of the MAPK cascade in the recipient [111]. This property is conserved over three transfers, in conditions where no cytoplasm originating from the first egg is present. In essence, this is strikingly similar to the cytoplasmic and infectious factors detected in CG and related phenomena. This regulatory inheritance has many properties in common with prions. It however, displays several differences. First, it relies on many proteins (the whole signaling cascade or at least a subset of the cascade). This implies that the genetic basis is more complex than for prion [106]. In the case of CG, additional factors have been identified and were shown to be necessary for producing the C element, but likely not to be present in the regulatory loop [112–115]. Many genes restricting the spread of C have also been identified adding another level of complexity [106]. Second, the development and/or spreading of C are highly dependent upon the environmental conditions, a property also exhibited by the [β] factor. Because regulatory networks can adopt complex behavior, depending upon the level of expression of the key “flexibility loci” [116], it is not surprising that the determinism of CG is quite complex and depends upon numerous genetic and environmental factors.

The White/Opaque Switch of Candida albicans, an Epigenetic Switch at the Transcription Level Many species of yeast have the ability to switch at various frequencies between different states [13,117–122].

II.  Molecular Mechanisms of Epigenetics

68

5.  Prions and Prion-Like Phenomena in Epigenetic Inheritance

These switches may be caused by classical transcriptional gene silencing (see [123] for a review) or prions [13]. However, regulatory inheritance may also be involved. The most studied of the switches is the white/opaque transition exhibited by C. albicans [124]. This transition is present in this diploid fungus when homozygous for the mating type [125,126] or when grown in a condition resembling the environment within its host [127]. The cells may then adopt two morphologies: round cells forming white colonies and bigger more elongated cell forming colonies that are more translucent. In fact, the two types of cell differ by an impressive array of differences [128–130]. C. albicans causes mycosis in human and the switch likely enable the fungus to adapt to the various niches it will encounter in the human body [131,132]. To understand how the transition is controlled, genes downregulated by the a1-α2 heterodimers encoded at the C. albicans mating type were searched, since the white/ opaque switch is present in strains homozygous for their mating-type [125]. Transcriptomic or chromatin-immunoprecipitation approaches identified the WOR1/TOS9 transcription factor as specifically expressed in opaque cells [133–135]. Gene inactivation of WOR1 showed that the cells were locked in the white state. Ectopic expression of WOR1, even as a single pulse or in cells heterozygous for the mating type, was sufficient to convert the whole cell population to the opaque state. Finally, it was shown that WOR1 binds the promoter of its own gene and thereby activates its own expression [135]. Overall, these data permitted the formulation of a model for the white/opaque transition based on the self-activation of the WOR1 transcription factor [133–135]. WOR1 is absent in white cells. Random fluctuation in the transcription of the WOR1 locus permits the expression of a few molecules of WOR1, resulting in further transcription of the WOR1 gene, locking the cell in a state with a high concentration of WOR1. The other targets of WOR1 are then regulated promoting the physiological and morphological changes to the opaque state [136,137]. Although it has not yet been formally demonstrated that the opaque state is infectious toward the white state, the similarity of this bistable system with the ones created by classical, β-type and C-type infectious prions is evident. Studies showed that the WOR1 positive regulatory loop is embedded in a complex network of seven transcription factors (WOR1, WOR2, WOR3, WOR4, EFG1, CZF1, and AHR1) with positive and negative feedback loops [138–140]. The multiplicity of the feedbacks appears to ensure a faithful transmission of the white and opaque states through numerous cell generations and accounts for the previously known roles of various transcription factors and chromatin remodeling factors in the control of the transition [141–144]. It also enables in some clinical isolates the formation of a third state called “gray” that displays a distinct transcriptional profile and

physiology [145]. This state does not necessitate WOR1 to appear. However, a wor1 efg1 double mutant is locked in the gray state, suggesting a connection between this new state and the regulatory network controlling the white/opaque switch. Like for the CG of P. anserina, the environment controls the switch in both directions. While high temperature triggers the opaque to white switch [124,146], numerous factors are able to trigger the white to opaque transition [147,148], and a study showed that slowing the cell cycle by many means is sufficient to increase the white to opaque switch frequency [149]. The signalling pathways controlling the transition include the Hog1 MAP kinase pathway, which controls stress response in many fungi [150].

CONCLUSIONS The various examples aforementioned show that epigenetic states can be conferred in many ways (Fig. 5.1), provided that an autoregulatory loop (or a double repression) is present. The presence of this loop ensures that two mutually exclusive states may be exhibited by cells with an identical genome and grown in the same conditions. In the case of the structural inheritance that we have presented, the loops ensure the faithful reproduction of a structure made of proteins (prions, Hsp60, or cilia). There are, however, suggestions that another component of the cell, the membranes, may adopt alternative states [151]. In general, the influence of the environment on this kind of inheritance is moderate. On the contrary, the regulatory inheritance is usually greatly influenced by the environment, since any modification in the concentration of key factors [116] under the influence of external stimuli may drastically alter the behavior of the pathway. Importantly, in the case of regulatory inheritance the phenotype of the cells will depend upon their history, however, not in a directed fashion as in a Lamarckian inheritance. Often, this inheritance results from the emergent properties of complex networks or structural properties of domains that adopt particular structures. It is thus not easy to know whether the inheritable behavior is a by-product of these, or whether it participates in a process essential to the life cycle. If the white/opaque transition appears to confer a selective advantage to C. albicans, what of yeast prions and CG? Certainly in some cases, prions and prion-like element are involved in differentiation processes [152]. However, human prions clearly cause severe diseases and it has been proposed that regulatory inheritance may be involved in cancer formation [153]. This regulatory inheritance is presently known in both eukaryotes and prokaryotes, and occurs at all level of gene regulation (transcription, signal transduction, carbon metabolism...). As yet, the full scope of this kind of inheritance

II.  Molecular Mechanisms of Epigenetics



REFERENCES

69

FIGURE 5.1  Schematic diagram of the prions and prion-like elements of eukaryotes discussed in this chapter. See text for detailed explanation on each phenomenon.

based on prion and prion-like phenomena is unknown. We here have presented data obtained with a few model microorganisms, since they are more easily tractable than the multicellular animals and plants. However, we propose that the same epigenetic mechanisms are prevalent in all organisms. In a recent paper [154], it was demonstrated that over expression of nearly 50 S. cerevisiae proteins was able to create prion-like inheritable traits, expanding the repertoire of proteins endowed of such property in this yeast to those lacking Q or N rich regions.

References [1] Nanney DL. Epigenetic control systems. Proc Natl Acad Sci USA 1958;44:712–7. [2] Delbrück M. Comment of an article by Sonneborn, T.M. and Beale, G.H. In Unités biologiques douées de continuité, génétique. C.N.R.S., Paris, France, 1949. p. 33–34. [3] Marcou D. Notion de longévité et nature cytoplasmique du déterminant de la sénescence chez quelques champignons. Ann Sci Natur Bot Sér 1961;12(2):653–764. [4] Novick A, Weiner M. Enzyme induction as an all-or-none phenomenon. Proc Natl Acad Sci USA 1957;43:553–66. [5] Beisson J, Sonneborn TM. Cytoplasmic inheritance of the organization of the cell cortex in Paramecium aurelia. Proc Natl Acad Sci USA 1965;53:275–82. [6] Beisson J. Non-nucleic acid inheritance and epigenetic phenomena. Cell Biology, vol. 1. San Francisco USA: Academic Press Inc; 1977. p. 375–421. [7] Kirk JTO. Roots: the discovery of chloroplast DNA. BioEssays 1986;4:36–8. [8] Mounolou JC, Lacroute F. Mitochondrial DNA: an advance in eukaryotic cell biology in the 1960s. Biol Cell 2005;97:743–8. [9] Wickner RB. [URE3] as an altered URE2 protein: evidence for a prion analog in Saccharomyces cerevisiae. Science 1994;264:566–9. [10] Prusiner SB. Novel proteinaceous infectious particles cause scrapie. Science 1982;216:136–44.

[11] Legname G, Baskakov IV, Nguyen HO, Riesner D, Cohen FE, DeArmond SJ, et al. Synthetic mammalian prions. Science 2004;305:673–6. [12] Wickner RB, Shewmaker F, Kryndushkin D, Edskes HK. Protein inheritance (prions) based on parallel in-register beta-sheet amyloid structures. Bioessays 2008;30:955–64. [13] True HL, Berlin I, Lindquist SL. Epigenetic regulation of translation reveals hidden genetic variation to produce complex traits. Nature 2004;431:184–7. [14] Maddelein ML, Dos Reis S, Duvezin-Caubet S, Coulary-Salin B, Saupe SJ. Amyloid aggregates of the HET-s prion protein are infectious. Proc Natl Acad Sci USA 2002;99:7402–7. [15] Coustou V, Deleu C, Saupe S, Begueret J. The protein product of the het-s heterokaryon incompatibility gene of the fungus Podospora anserina behaves as a prion analog. Proc Natl Acad Sci USA 1997;94:9773–8. [16] Avery OT, McLeod CM, McCarty M. Studies on the chemical nature of the substance inducing transformation of pneumococcal types. J Experim Med 1944;79:137–58. [17] King CY, Diaz-Avalos R. Protein-only transmission of three yeast prion strains. Nature 2004;428:319–23. [18] Tanaka M, Chien P, Naber N, Cooke R, Weissman JS. Conformational variations in an infectious protein determine prion strain differences. Nature 2004;428:323–8. [19] Patel BK, Liebman SW. “Prion-proof” for [PIN+]: infection with in vitro-made amyloid aggregates of Rnq1p-(132–405) induces [PIN+]. J Mol Biol 2007;365:773–82. [20] Brachmann A, Baxa U, Wickner RB. Prion generation in vitro: amyloid of Ure2p is infectious. EMBO J 2005;24:3082–92. [21] Si K, Lindquist S, Kandel ER. A neuronal isoform of the aplysia CPEB has prion-like properties. Cell 2003;115:879–91. [22] Heinrich SU, Lindquist S. Protein-only mechanism induces self-perpetuating changes in the activity of neuronal Aplysia cytoplasmic polyadenylation element binding protein (CPEB). Proc Nat Acad Sci 2011;108:2999–3004. [23] Stephan JS, Joseph S, Fioriti L, Lamba N, Colnaghi L, Karl K, et al. The CPEB3 protein is a functional prion that interacts with the actin cytoskeleton. Cell Rep 2015;11:1772–85. [24] Chakrabortee S, Kayatekin C, Newby GA, Mendillo ML, Lancaster A, Lindquist S. Luminidependens (LD) is an Arabidopsis protein with prion behavior. Proc Natl Acad Sci 2016;113:6065–70.

II.  Molecular Mechanisms of Epigenetics

70

5.  Prions and Prion-Like Phenomena in Epigenetic Inheritance

[25] Si K, Choi Y-B, White-Grindley E, Majumdar A, Kandel ER. Aplysia CPEB can form prion-like multimers in sensory neurons that contribute to long-term facilitation. Cell 2010;140:421–35. [26] Raveendra BL, Siemer AB, Puthanveettil SV, Hendrickson WA, Kandel ER, McDermott AE. Characterization of prion-like conformational changes of the neuronal isoform of Aplysia CPEB. Nat Struct Mol Biol 2013;20:495–501. [27] Abbott A. The red-hot debate about transmissible Alzheimer’s. Nature 2016;531:294–7. [28] Kondrashkina AM, Antonets KS, Galkin AP, Nizhnikov AA. Prion-like determinant [NSI +] decreases the expression of the SUP45 gene in Saccharomyces cerevisiae. Mol Biol 2014;48:688–93. [29] Brown JC, Lindquist S. A heritable switch in carbon source utilization driven by an unusual yeast prion. Genes Dev 2009;23:2320–32. [30] Sparrer HE, Santoso A, Szoka FC Jr, Weissman JS. Evidence for the prion hypothesis: induction of the yeast [PSI+] factor by in vitro- converted Sup35 protein. Science 2000;289:595–9. [31] Brachmann A, Baxa U, Wickner RB. Prion generation in vitro: amyloid of Ure2p is infectious. Embo J 2005;24:3082–92. [32] Derkatch IL, Bradley ME, Hong JY, Liebman SW. Prions affect the appearance of other prions: the story of [PIN(+)]. Cell 2001;106:171–82. [33] Du Z, Park KW, Yu H, Fan Q, Li L. Newly identified prion linked to the chromatin-remodeling factor Swi1 in Saccharomyces cerevisiae. Nat Genet 2008;40:460–5. [34] Patel BK, Gavin-Smyth J, Liebman SW. The yeast global transcriptional co-repressor protein Cyc8 can propagate as a prion. Nat Cell Biol 2009;11:344–9. [35] Alberti S, Halfmann R, King O, Kapila A, Lindquist S. A systematic survey identifies prions and illuminates sequence features of prionogenic proteins. Cell 2009;137:146–58. [36] Rogoza T, Goginashvili A, Rodionova S, Ivanov M, Viktorovskaya O, Rubel A, et al. Non-Mendelian determinant [ISP+] in yeast is a nuclear-residing prion form of the global transcriptional regulator Sfp1. Proc Natl Acad Sci 2010;107:10573–7. [37] Suzuki G, Shimazu N, Tanaka M. A yeast prion, Mod5, promotes acquired drug resistance and cell survival under environmental stress. Science 2012;336:355–9. [38] Halfmann R, Wright JR, Alberti S, Lindquist S, Rexach M. Prion formation by a yeast GLFG nucleoporin. Prion 2012;6:391–9. [39] Li L, Lindquist S. Creating a protein-based element of inheritance. Science 2000;287:661–4. [40] Patino MM, Liu JJ, Glover JR, Lindquist S. Support for the prion hypothesis for inheritance of a phenotypic trait in yeast. Science 1996;273:622–6. [41] Ross ED, Baxa U, Wickner RB. Scrambled prion domains form prions and amyloid. Mol Cell Biol 2004;24:7206–13. [42] Sondheimer N, Lindquist S. Rnq1: an epigenetic modifier of protein function in yeast. Mol Cell 2000;5:163–72. [43] Verges KJ, Smith MH, Toyama BH, Weissman JS. Strain conformation, primary structure and the propagation of the yeast prion [PSI+]. Nat Struct Mol Biol 2011;18:493–9. [44] Baxa U, Cassese T, Kajava AV, Steven AC. Structure, function, and amyloidogenesis of fungal prions: filament polymorphism and prion variants. Adv Protein Chem 2006;73:125–80. [45] Tessier PM, Lindquist S. Unraveling infectious structures, strain variants and species barriers for the yeast prion [PSI+]. Nat Struct Mol Biol 2009;16:598–605. [46] Ritter C, Maddelein ML, Siemer AB, Luhrs T, Ernst M, Meier BH, et al. Correlation of structural elements and infectivity of the HET-s prion. Nature 2005;435:844–8. [47] Wasmer C, Schutz A, Loquet A, Buhtz C, Greenwald J, Riek R, et al. The molecular organization of the fungal prion HET-s in its amyloid form. J Mol Biol 2009;394:119–27.

[48] Daskalov A, Gantner M, Walti MA, Schmidlin T, Chi CN, ­Wasmer C, et al. Contribution of specific residues of the betasolenoid fold to HET-s prion function, amyloid structure and stability. PLoS Pathog 2014;10:e1004158. [49] Greenwald J, Buhtz C, Ritter C, Kwiatkowski W, Choe S, Maddelein ML, et al. The mechanism of prion inhibition by HET-S. Mol Cell 2010;38:889–99. [50] Chernoff YO, Lindquist SL, Ono B, Inge-Vechtomov SG, Liebman SW. Role of the chaperone protein Hsp104 in propagation of the yeast prion-like factor [psi+]. Science 1995;268:880–4. [51] Jung G, Masison DC. Guanidine hydrochloride inhibits Hsp104 activity in vivo: a possible explanation for its effect in curing yeast prions. Curr Microbiol 2001;43:7–10. [52] Rikhvanov EG, Romanova NV, Chernoff YO. Chaperone effects on prion and nonprion aggregates. Prion 2007;1:217–22. [53] Derkatch IL, Bradley ME, Zhou P, Chernoff YO, Liebman SW. Genetic and environmental factors affecting the de novo appearance of the [PSI+] prion in Saccharomyces cerevisiae. Genetics 1997;147:507–19. [54] Bradley ME, Edskes HK, Hong JY, Wickner RB, Liebman SW. Interactions among prions and prion “strains” in yeast. Proc Natl Acad Sci USA 2002;99(Suppl 4):16392–9. [55] Schwimmer C, Masison DC. Antagonistic interactions between yeast [PSI(+)] and [URE3] prions and curing of [URE3] by Hsp70 protein chaperone Ssa1p but not by Ssa2p. Mol Cell Biol 2002;22:3590–8. [56] Bradley ME, Liebman SW. Destabilizing interactions among [PSI(+)] and [PIN(+)] yeast prion variants. Genetics 2003;165:1675–85. [57] Inoue Y, Kawai-Noma S, Koike-Takeshita A, Taguchi H, Yoshida M. Yeast prion protein New1 can break Sup35 amyloid fibrils into fragments in an ATP-dependent manner. Gene Cell 2011;16:545–56. [58] Derkatch IL, Chernoff YO, Kushnirov VV, Inge-Vechtomov SG, Liebman SW. Genesis and variability of [PSI] prion factors in Saccharomyces cerevisiae. Genetics 1996;144:1375–86. [59] Schlumpberger M, Wille H, Baldwin MA, Butler DA, Herskowitz I, Prusiner SB. The prion domain of yeast Ure2p induces autocatalytic formation of amyloid fibers by a recombinant fusion protein. Protein Sci 2000;9:440–51. [60] Morales R, Abid K, Soto C. The prion strain phenomenon: molecular basis and unprecedented features. Biochim Biophys Acta 2007;1772:681–91. [61] Pattison IH, Millson GC. Further experimental observations on scrapie. J Comp Pathol 1961;71:350–9. [62] Pattison IH, Millson GC. Scrapie produced experimentally in goats with special reference to the clinical syndrome. J Comp Pathol 1961;71:101–9. [63] Tanaka M, Collins SR, Toyama BH, Weissman JS. The physical basis of how prion conformations determine strain phenotypes. Nature 2006;442:585–9. [64] Tanaka M, Chien P, Yonekura K, Weissman JS. Mechanism of cross-species prion transmission: an infectious conformation compatible with two highly divergent yeast prion proteins. Cell 2005;121:49–62. [65] Debets AJM, Dalstra HJP, Slakhorst M, Koopmanschap B, Hoekstra RF, Saupe SJ. High natural prevalence of a fungal prion. Proc Natl Acad Sci 2012;109:10432–7. [66] Dalstra HJP, Swart K, Debets AJM, Saupe SJ, Hoekstra RF. Sexual transmission of the [Het-s] prion leads to meiotic drive in Podospora anserina. Proc Natl Acad Sci USA 2003;100:6616–21. [67] Jarosz Daniel F, Lancaster Alex K, Brown Jessica CS, Lindquist S. An evolutionarily conserved prion-like element converts wild fungi from metabolic specialists to generalists. Cell 2014;158:1072–82.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[68] Jarosz DF, Brown JCS, Walker GA, Datta MS, Ung WL, Lancaster AK, et al. Cross-kingdom chemical communication drives a heritable, mutually beneficial prion-based transformation of metabolism. Cell 2014;158:1083–93. [69] Nakayashiki T, Kurtzman CP, Edskes HK, Wickner RB. Yeast prions [URE3] and [PSI+] are diseases. Proc Natl Acad Sci USA 2005;102:10575–80. [70] Wickner RB, Edskes HK, Gorkovskiy A, Bezsonov EE, Stroobant EE. Chapter four—yeast and fungal prions: amyloid-handling systems, amyloid structure, and prion biology. In: Theodore Friedmann JCD, Stephen FG, editors. Advances in genetics, vol. 93. Academic Press; 2016. p. 191–236. [71] Kelly AC, Shewmaker FP, Kryndushkin D, Wickner RB. Sex, prions, and plasmids in yeast. Proc Natl Acad Sci 2012;109:E2683–90. [72] Daskalov A, Habenstein B, Martinez D, Debets AJM, Sabaté R, Loquet A, et al. Signal transduction by a fungal NOD-like receptor based on propagation of a prion amyloid fold. PLoS Biol 2015;13:e1002059. [73] Si K, Kandel ER. The role of functional prion-like proteins in the persistence of memory. Cold Spring Harb Perspect Biol 2016;8(4):8. [74] Cheng MY, Hartl FU, Horwich AL. The mitochondrial chaperonin hsp60 is required for its own assembly. Nature 1990;348:455–8. [75] Christensen JH, Nielsen MN, Hansen J, Fuchtbauer A, Fuchtbauer EM, West M, et al. Inactivation of the hereditary spastic ­paraplegia-associated Hspd1 gene encoding the Hsp60 chaperone results in early embryonic lethality in mice. Cell Stress Chaperones 2010;15:851–63. [76] Bross P, Magnoni R, Bie AS. Molecular chaperone disorders: defective Hsp60 in neurodegeneration. Curr Top Med Chem 2012;12:2491–503. [77] Sonneborn TM. The determinants and evolution of life. the differentiation of cells. Proc Natl Acad Sci USA 1964;51:915–29. [78] Ng SF, Frankel J. 180 degrees rotation of ciliary rows and its morphogenetic implications in Tetrahymena pyriformis. Proc Natl Acad Sci USA 1977;74:1115–9. [79] Feldman JL, Geimer S, Marshall WF. The mother centriole plays an instructive role in defining cell geometry. PLoS Biol 2007;5:e149. [80] Moreira-Leite FF, Sherwin T, Kohl L, Gull K. A trypanosome structure involved in transmitting cytoplasmic information during cell division. Science 2001;294:610–2. [81] Nelsen EM, Frankel J. Maintenance and regulation of cellular handedness in Tetrahymena. Development 1989;105:457–71. [82] Nelsen EM, Frankel J, Jenkins LM. Non-genic inheritance of cellular handedness. Development 1989;105:447–56. [83] Frankel J. What do genic mutations tell us about the structural patterning of a complex single-celled organism? Eukaryot Cell 2008;7:1617–39. [84] Chen T, Hiroko T, Chaudhuri A, Inose F, Lord M, Tanaka S, et al. Multigenerational cortical inheritance of the Rax2 protein in orienting polarity and division in yeast. Science 2000;290:1975–8. [85] Iftode F, Fleury-Aubusson A. Structural inheritance in Paramecium: ultrastructural evidence for basal body and associated rootlets polarity transmission through binary fission. Biol Cell 2003;95:39–51. [86] Roberts BT, Wickner RB. Heritable activity: a prion that propagates by covalent autoactivation. Genes Dev 2003;17:2083–7. [87] Roberts BT, Wickner RB. A new kind of prion: a modified protein necessary for its own modification. Cell Cycle 2004;3:100–3. [88] Brandman O, Meyer T. Feedback loops shape cellular signals in space and time. Science 2008;322:390–5. [89] Drubin DA, Way JC, Silver PA. Designing biological systems. Genes Dev 2007;21:242–54. [90] Sprinzak D, Elowitz MB. Reconstruction of genetic circuits. Nature 2005;438:443–8.

71

[91] Becskei A, Seraphin B, Serrano L. Positive feedback in eukaryotic gene networks: cell differentiation by graded to binary response conversion. Embo J 2001;20:2528–35. [92] Gardner TS, Cantor CR, Collins JJ. Construction of a genetic toggle switch in Escherichia coli. Nature 2000;403:339–42. [93] Kobayashi H, Kaern M, Araki M, Chung K, Gardner TS, Cantor CR, Collins JJ. Programmable cells: interfacing natural and engineered gene networks. Proc Natl Acad Sci USA 2004;101:8414–9. [94] Tchuraev RN, Stupak IV, Tropynina TS, Stupak EE. Epigenes: design and construction of new hereditary units. FEBS Lett 2000;486:200–2. [95] Toman Z, Dambly C, Radman M. Induction of a stable, heritable epigenetic change by mutagenic carcinogens: a new test system. IARC Sci Publ 1980;27:243–55. [96] Ajo-Franklin CM, Drubin DA, Eskin JA, Gee EP, Landgraf D, Phillips I, et al. Rational design of memory in eukaryotic cells. Genes Dev 2007;21:2271–6. [97] Cohn M, Horibata K. Analysis of the differentiation and of the heterogeneity within a population of Escherichia coli undergoing induced beta-galactosidase synthesis. J Bacteriol 1959;78:613–23. [98] Cohn M, Horibata K. Inhibition by glucose of the induced synthesis of the beta-galactoside-enzyme system of Escherichia coli. Analysis of maintenance. J Bacteriol 1959;78:601–12. [99] Cohn M, Horibata K. Physiology of the inhibition by glucose of the induced synthesis of the beta-galactoside enzyme system of Escherichia coli. J Bacteriol 1959;78:624–35. [100] Choi PJ, Cai L, Frieda K, Xie XS. A stochastic single-molecule event triggers phenotype switching of a bacterial cell. Science 2008;322:442–6. [101] Ozbudak EM, Thattai M, Lim HN, Shraiman BI, Van Oudenaarden A. Multistability in the lactose utilization network of Escherichia coli. Nature 2004;427:737–40. [102] Laurent M, Kellershohn N. Multistability: a major means of differentiation and evolution in biological systems. Trends Biochem Sci 1999;24:418–22. [103] Santillan M, Mackey MC. Quantitative approaches to the study of bistability in the lac operon of Escherichia coli. J R Soc Interface 2008;5(Suppl 1):S29–39. [104] Toman Z, Dambly-Chaudiere C, Tenenbaum L, Radman M. A system for detection of genetic and epigenetic alterations in Escherichia coli induced by DNA-damaging agents. J Mol Biol 1985;186:97–105. [105] Silar P, Haedens V, Rossignol M, Lalucque H. Propagation of a novel cytoplasmic, infectious and deleterious determinant is controlled by translational accuracy in Podospora anserina. Genetics 1999;151:87–95. [106] Haedens V, Malagnac F, Silar P. Genetic control of an epigenetic cell degeneration syndrome in Podospora anserina. Fungal Genet Biol 2005;42:564–77. [107] Silar P, Daboussi MJ. Non-conventional infectious elements in filamentous fungi. Trends Genet 1999;15:141–5. [108] Graziani S, Silar P, Daboussi MJ. Bistability and hysteresis of the ‘Secteur’ differentiation are controlled by a two-gene locus in Nectria haematococca. BMC Biol 2004;2:18. [109] Kicka S, Silar P. PaASK1, a mitogen-activated protein kinase kinase kinase that controls cell degeneration and cell differentiation in Podospora anserina. Genetics 2004;166:1241–52. [110] Kicka S, Bonnet C, Sobering AK, Ganesan LP, Silar P. A mitotically inheritable unit containing a MAP kinase module. Proc Natl Acad Sci USA 2006;103:13445–50. [111] Bagowski CP, Ferrell JE Jr. Bistability in the JNK cascade. Curr Biol 2001;11:1176–82. [112] Jamet-Vierny C, Debuchy R, Prigent M, Silar P. IDC1, a Pezizomycotina-specific gene that belongs to the PaMpk1 MAP kinase transduction cascade of the filamentous fungus Podospora anserina. Fungal Genet Biol 2007;44:1219–30.

II.  Molecular Mechanisms of Epigenetics

72

5.  Prions and Prion-Like Phenomena in Epigenetic Inheritance

[113] Lacaze I, Lalucque H, Siegmund U, Silar P, Brun S. Identification of NoxD/Pro41 as the homologue of the p22phoxNADPH oxidase subunit in fungi. Mol Microbiol 2015;95:1006–24. [114] Chan Ho Tong L, Silar P, Lalucque H. Genetic control of anastomosis in Podospora anserina. Fungal Genet Biol 2014;70C:94–103. [115] Lalucque H, Malagnac F, Brun S, Kicka S, Silar P. A non-mendelian MAPK-generated hereditary unit controlled by a second MAPK pathway in Podospora anserina. Genetics 2012;191:419–33. [116] Bhalla US, Ram PT, Iyengar R. MAP kinase phosphatase as a locus of flexibility in a mitogen-activated protein kinase signaling network. Science 2002;297:1018–23. [117] Clemons KV, Hanson LC, Stevens DA. Colony phenotype switching in clinical and non-clinical isolates of Saccharomyces cerevisiae. J Med Vet Mycol 1996;34:259–64. [118] Lachke SA, Joly S, Daniels K, Soll DR. Phenotypic switching and filamentation in Candida glabrata. Microbiology 2002;148: 2661–74. [119] Lachke SA, Srikantha T, Tsai LK, Daniels K, Soll DR. Phenotypic switching in Candida glabrata involves phase-specific regulation of the metallothionein gene MT-II and the newly discovered hemolysin gene HLP. Infect Immun 2000;68:884–95. [120] Fries BC, Goldman DL, Casadevall A. Phenotypic switching in Cryptococcus neoformans. Microbes Infect 2002;4:1345–52. [121] Enger L, Joly S, Pujol C, Simonson P, Pfaller M, Soll DR. Cloning and characterization of a complex DNA fingerprinting probe for Candida parapsilosis. J Clin Microbiol 2001;39:658–69. [122] Joly S, Pujol C, Schroppel K, Soll DR. Development of two species-specific fingerprinting probes for broad computer-assisted epidemiological studies of Candida tropicalis. J Clin Microbiol 1996;34:3063–71. [123] Malagnac F, Silar P. Epigenetics of Eukaryotic Microbes. In: Tollefsbol T, editor. Handbook of epigenetics. San Diego: Academic Press; 2011. p. 185–201. Chapter 13. [124] Slutsky B, Staebell M, Anderson J, Risen L, Pfaller M, Soll DR. “White-opaque transition”: a second high-frequency switching system in Candida albicans. J Bacteriol 1987;169:189–97. [125] Miller MG, Johnson AD. White-opaque switching in Candida albicans is controlled by mating-type locus homeodomain proteins and allows efficient mating. Cell 2002;110:293–302. [126] Lockhart SR, Pujol C, Daniels KJ, Miller MG, Johnson AD, Pfaller MA, et al. In Candida albicans, white-opaque switchers are homozygous for mating type. Genetics 2002;162:737–45. [127] Xie J, Tao L, Nobile CJ, Tong Y, Guan G, Sun Y, et al. Whiteopaque switching in natural MTLa/alpha isolates of Candida albicans: evolutionary implications for roles in host adaptation, pathogenesis, and sex. PLoS Biol 2013;11:e1001525. [128] Soll DR. High-frequency switching in Candida albicans. Clin Microbiol Rev 1992;5:183–203. [129] Soll DR. Gene regulation during high-frequency switching in Candida albicans. Microbiology 1997;143(Pt 2):279–88. [130] Lan CY, Newport G, Murillo LA, Jones T, Scherer S, Davis RW, et al. Metabolic specialization associated with phenotypic switching in Candida albicans. Proc Natl Acad Sci USA 2002;99:14907–12. [131] Vargas K, Messer SA, Pfaller M, Lockhart SR, Stapleton JT, Hellstein J, et al. Elevated phenotypic switching and drug resistance of Candida albicans from human immunodeficiency virus-positive individuals prior to first thrush episode. J Clin Microbiol 2000;38:3595–607. [132] Lohse MB, Johnson AD. Differential phagocytosis of white versus opaque Candida albicans by Drosophila and mouse phagocytes. PLoS One 2008;3:e1473. [133] Huang G, Wang H, Chou S, Nie X, Chen J, Liu H. Bistable expression of WOR1, a master regulator of white-opaque switching in Candida albicans. Proc Natl Acad Sci USA 2006;103:12813–8.

[134] Srikantha T, Borneman AR, Daniels KJ, Pujol C, Wu W, Seringhaus MR, et al. TOS9 regulates white-opaque switching in Candida albicans. Eukaryot Cell 2006;5:1674–87. [135] Zordan RE, Galgoczy DJ, Johnson AD. Epigenetic properties of white-opaque switching in Candida albicans are based on a selfsustaining transcriptional feedback loop. Proc Natl Acad Sci USA 2006;103:12807–12. [136] Tuch BB, Mitrovich QM, Homann OR, Hernday AD, Monighetti CK, De La Vega FM, et al. The transcriptomes of two heritable cell types illuminate the circuit governing their differentiation. PLoS Genet 2010;6:e1001070. [137] Si H, Hernday AD, Hirakawa MP, Johnson AD, Bennett RJ. Candida albicans white and opaque cells undergo distinct programs of filamentous growth. PLoS Pathog 2013;9:e1003210. [138] Zordan RE, Miller MG, Galgoczy DJ, Tuch BB, Johnson AD. Interlocking transcriptional feedback loops control white-opaque switching in Candida albicans. PLoS Biol 2007;5:e256. [139] Lohse MB, Johnson AD. Identification and characterization of Wor4, a new transcriptional regulator of white-opaque switching. G3 Genes Genomes Genetics 2016;6:721–9. [140] Hernday AD, Lohse MB, Fordyce PM, Nobile CJ, DeRisi JL, Johnson AD. Structure of the transcriptional network controlling white-opaque switching in Candida albicans. Mol Microbiol 2013;90:22–35. [141] Sonneborn A, Tebarth B, Ernst JF. Control of white-opaque phenotypic switching in Candida albicans by the Efg1p morphogenetic regulator. Infect Immun 1999;67:4655–60. [142] Perez-Martin J, Uria JA, Johnson AD. Phenotypic switching in Candida albicans is controlled by a SIR2 gene. Embo J 1999;18:2580–92. [143] Klar AJ, Srikantha T, Soll DR. A histone deacetylation inhibitor and mutant promote colony-type switching of the human pathogen Candida albicans. Genetics 2001;158:919–24. [144] Srikantha T, Tsai L, Daniels K, Klar AJ, Soll DR. The histone deacetylase genes HDA1 and RPD3 play distinct roles in regulation of high-frequency phenotypic switching in Candida albicans. J Bacteriol 2001;183:4614–25. [145] Tao L, Du H, Guan G, Dai Y, Nobile CJ, Liang W, et al. Discovery of a white-gray-opaque tristable phenotypic switching system in Candida albicans: roles of non-genetic diversity in host adaptation. PLoS Biol 2014;12:e1001830. [146] Rikkerink EH, Magee BB, Magee PT. Opaque-white phenotype transition: a programmed morphological transition in Candida albicans. J Bacteriol 1988;170:895–9. [147] Huang G, Srikantha T, Sahni N, Yi S, Soll DR. CO(2) regulates white-to-opaque switching in Candida albicans. Curr Biol 2009;19:330–4. [148] Ramirez-Zavala B, Reuss O, Park YN, Ohlsen K, Morschhauser J. Environmental induction of white-opaque switching in Candida albicans. PLoS Pathog 2008;4:e1000089. [149] Alby K, Bennett R. Stress-induced phenotypic switching in Candida albicans. Mol Biol Cell 2009;20:3178–91. [150] Liang S-H, Cheng J-H, Deng F-S, Tsai P-A, Lin C-H. A novel function for hog1 stress-activated protein kinase in controlling white-opaque switching and mating in Candida albicans. Eukaryotic Cell 2014;13:1557–66. [151] Lockshon D. A heritable structural alteration of the yeast mitochondrion. Genetics 2002;161:1425–35. [152] Malagnac F, Silar P. Regulation, cell differentiation and proteinbased inheritance. Cell Cycle 2006;5:2584–7. [153] Blagosklonny MV. Molecular theory of cancer. Cancer Biol Ther 2005;4:621–7. [154] Chakrabortee S, Byers JS, Jones S, Garcia DM, Bhullar B, Chang A, She R, Fremin B, Lindquist S, Jarosz DF. Intrinsically disordered proteins drive emergence and inheritance of biological traits. Cell 2016;167:369–81.

II.  Molecular Mechanisms of Epigenetics

C H A P T E R

6 Higher-Order Chromatin Organization in Diseases: From Chromosomal Position Effect to Phenotype Variegation Jérôme D. Robin*, Frédérique Magdinier** *Institute for Research on Cancer and Aging, Nice (ICRAN), The National Center for Scientific Research (CNRS), French National Institute of Health and Medical Research (INSERM), Nice, France; **Aix Marseille University, French National Institute of Health and Medical Research (INSERM), Medical Genetics and Functional Genomics (GMGF), Marseille, France

O U T L I N E Introduction

73

CPE in Model Organisms

74

Telomeric Position Effect

76

Setting the Frontiers of Chromatin Domains

77

Genome Topology and Scaffolding of Chromosomal Domains

77

INTRODUCTION

78

Chromosomal Position Effect in Human Pathologies

79

Telomeric Position Effect in Human Pathologies

82

Conclusions

86

References

87

modifications of specific amino acids of the core histones (H2A, H2B, H3, and H4) are translated into distinct information; the presence of histone variants (i.e., H2.AX, H2.AZ, MacroH2A, CENPA, and H3.3); the covalent modification of the underlying DNA; its composition in nonhistone-binding factors; the spatial localization within the nucleoplasm and its dynamics during the cell cycle. In particular, mitosis is associated with profound changes in the organization of chromatin, the transient disappearance of the nuclear envelope and most subnuclear organelles, eviction of transcription factors, and gene silencing. Nevertheless, at the end of mitosis, the cell fate-specific transcription program and epigenetic profile is faithfully recapitulated requiring a maintenance mechanism named mitotic bookmarking [1–4].

The genome of eukaryotes is composed of thousands of genes and even more interspersed noncoding sequences. Constraining up to tens of billions of bases within a nucleus of a few microns in diameter requires a high level of DNA compaction that must also exhibit high plasticity to allow efficient realization of cellular functions. Structure of the chromatin fiber regulates the accessibility of DNA to the plethora of factors involved in the regulation of gene expression, DNA replication, or response to DNA damage. Each chromatin state can be defined in four dimensions by its level of compaction; its topological state; the positioning and the spacing of nucleosomes; its histone code, predicting how the posttranslational Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00006-7 Copyright © 2017 Elsevier Inc. All rights reserved.

Boundary Elements are Involved in Functional Genome Partitioning

73

74

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

Generally, open chromatin, where most of the transcription occurs, is referred to as “euchromatin,” whereas condensed chromatin, where transcription is generally inhibited, is referred to as “heterochromatin”; although various types of chromatin structure are evoked under these denominations. In eukaryotes, heterochromatin is hypoacetylated and lacks methylation of H3K4, but is enriched in methylated DNA, histone H3K9 di- and trimethylation, H3K27 and H4K20 methylation, HP1 binding, and can spread over flanking regions, thereby inducing transcriptional silencing [5,6]. However, a number of genes escape this strict dichotomy and show a bivalent chromatin signature with both repressive (H3K27me3) and active transcription (H3K4me1–3) marks, allowing poised expression during development and cell lineage differentiation [7–10]. Overall, it has to be kept in mind that chromatin is highly plastic and that transition from one state to another occurs during development or cell differentiation, in particular during the transition from chromatin-relaxed pluripotent cells to cells committed in a specific lineage [7,10]. Genes, regulatory elements, and repetitive DNA are interspersed, at the chromosome level, in a mosaic of condensed and open regions. The proximity of different types of chromatin can influence gene expression either positively (enhancer proximity) or negatively (silencer proximity) [11–13]. Structural variations, such as insertion, deletion, duplication, translocation, or inversion of DNA segments, are often observed in human diseases. Variants, especially at noncoding regions, can interrupt chromatin topology, resulting in ectopic interactions between distant regulatory elements. Upon rearrangement, a gene relocated in the vicinity of heterochromatin can become silent in a subset of cells, leading to a characteristic variegated pattern of expression, as a consequence of a position effect (position effect variegation or PEV). Thus, chromosomal position effect (CPE or chromosomal position effect) refers to the differences in gene expression when a gene is positioned at different positions and in different chromatin environments in the genome. Moreover, telomere proximity can trigger gene silencing by telomeric position effect (or TPE) [14]. To an extent, the identity of chromatin domains is maintained by different factors, such as cis-regulators, “fuzzy boundaries,” or insulators that limit the influence of one region on the adjacent one [15–17]. The goal of this chapter is not to provide a detailed review of all the experimental work that has been published on CPE, PEV, or TPE in different cellular or animal models, but rather to describe the main features of CPE and TPE and their implication in human pathologies.

CPE IN MODEL ORGANISMS CPE was originally discovered in flies in the 1920s by Sturtevant who first described facet changes in the eyes of flies linked to duplication of the Bar locus that influences expression of the Bar gene, possibly due to a complex mechanism that was named “position effect” [18]. Later on, Muller showed that an inversion of the X-chromosome and relocalization of the white gene close to pericentromeric heterochromatin was associated with a “mottled” phenotype with each eye having some white (mutant) and red (wildtype) regions with variation from eye to eye. This phenomenon was dubbed PEV [11,18–20]. These early observations led to the description of two types of mechanisms defined as stable position effect and PEV, respectively (Fig. 6.1). Stable position effect is associated with gene silencing upon transgene multimerization (i.e., repeat-induced silencing) [21,22], while PEV is produced by the translocation of euchromatic genes into heterochromatin or insertion of heterochromatin-prone sequences into euchromatin domains and subsequent spreading of heterochromatic features into euchromatin [23–25]. This spreading can be variable from cell to cell and leads to the characteristic mosaic appearance of variegated expression. As a consequence, a transgene located in constitutive heterochromatin adopts the compact nucleosomal structure of the insertion site. Moreover, if some transgenes do not variegate when repeated in tandem, certain ones are more prone to silencing, challenging their use in transgenesis. In flies, the genetic dissection of this process has been performed by means of dominant suppressor [Su(var), suppressor of variegation] and enhancer [E(var), enhancer of variegation] mutations. Around 140 suppressors and 230 enhancers of variegation have been identified and approximately 30, fully characterized [11,26,27]. Most of these modifiers are components of heterochromatin, enzymes that modify histones, nonhistone proteins, or nuclear architectural proteins, such as the heterochromatin protein 1, HP1, Su(var)3–9, the histone H3K9 methyltransferase (HMTase), the Suv4-20 HMTase, histone acetyl transferases (HATs), the Drosophila homolog of the mammalian LSD1 amine oxidase that demethylates H3K4me2 and H3K4me1, components of the TIP60 complex or remodeling and spacing factor (Rsf) (reviewed in Ref. [27]). Recent advances have shown the existence of different subgroups of PEV modifiers, depending on their ability to modulate silencing and the chromatin context [28]. Different screens were also performed in mice to identify loci or genes that either decrease or increase the degree of epigenetic silencing [29–31]. Most of the genes identified encode epigenetic modifiers, including

II.  Molecular Mechanisms of Epigenetics



CPE in Model Organisms

75

FIGURE 6.1  Chromatin configuration modulates gene expression and phenotype variegation. (A) Gene expression can be modulated by the chromatin context. Condensed chromatin (constitutive heterochromatin, repetitive DNA, and telomeric DNA) can induce spreading of heterochromatin marks, nucleosomes repositioning, or recruitment of chromatin-remodeling complexes, leading to gene repression. The extent of gene silencing might be different from cell to cell in a given tissue, leading to a variegated expression pattern and mosaic phenotype. This phenomenon is called position effect variegation (PEV) and has been observed in all eukaryotic cells. The presence of “fuzzy boundaries” maintains a dynamic equilibrium between the two types of chromatin and limits heterochromatin spreading. (B) The presence of boundary elements (or insulators) defines strict borders between chromatin regions and separate functional domains by recruiting specific factors. (C) Insulators can also regulate cis-regulating elements, such as silencer, (S) or enhancers (E).

the DNA methyltransferases Dnmt1 and Dnmt3b; the Hdac1 histone deacetylase; the Suvar39h1 and Setdb1 HMTases; the Smarca5, Smarcc1, Pbrm1, or Baz1b chromatin-remodeling factors; the Trim28 transcriptional regulator; and epigenetic modifiers, such as Smchd1 or D14Abb1e. Interestingly, screens performed in the mouse led to the identification of additional factors that do not have orthologs in Drosophila, especially among genes involved in the regulation of DNA methylation. So far, screening for modifiers of PEV has been challenging in human cells until the recent advances of

genome-editing approaches. Nonlethal haploid genetic screen in human KBM7 chronic myeloid leukemia cells carrying a single copy of each chromosome, except chromosome 8 and sex chromosome [32], uncovered the existence of the Human Silencing Hub (HUSH) complex composed of four factors essential for gene silencing: the SETDB1 HMTase, M-phase phosphoprotein (MPP8), Periphilin, and FAM208A (TASOR, transgene activation repressor) [33]. The last three factors are associated in the same complex that in turn recruits SETDB1, through interaction with the MPP8 chromodomain.

II.  Molecular Mechanisms of Epigenetics

76

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

Silenced transgenes probably do not recruit the HUSH complex directly, but are rather modified by recruitment of HUSH, SETDB1, and H3K9me3 that are present at the heterochromatic sites of insertion, thereby reinforcing silencing in preexisting heterochromatin domains. HUSH repression does not require HP1 or tripartite motif containing 28 (TRIM28). Orthologs of HUSH complex subunits are not found in flies, but highly conserved from fish to humans [33].

TELOMERIC POSITION EFFECT Apart from PEV, silencing also occurs in the vicinity of telomeres from Saccharomyces cerevisiae to Homo sapiens (Fig. 6.2). TPE in yeast was first demonstrated by insertion of a construct containing a URA3 auxotrophic marker 1.1 kb from a newly formed telomere. Expression of the URA3 gene allows cell growth in the absence of uracil. However, on plates containing 5-fluoro-orotic acid, a drug toxic for cells expressing URA3, 20%–60% of cells were able to grow, suggesting that the subtelomeric URA3 was silenced by telomere proximity [14]. TPE is characterized by reversibility, promoter independence, and variegated expression.

Interestingly, despite structural differences at telomeres formed by transposition of the HeTA and TART retrotransposons to chromosome ends and arrays of telomere-associated sequences (TAS) [34,35], Drosophila melanogaster also exhibits telomeric silencing [36]. However, genetic modifiers of PEV display little or no effect on TPE, suggesting the existence of specialized mechanisms. In S. cerevisiae more than 50 proteins that modulate TPE have been identified [14,37,38], but only a few exhibit a specific and complete suppression of telomeric silencing [39]. Among them, silent information regulators (Sir)–complex proteins (Sir2p, Sir3p, and Sir4p) [40], Ku heterodimer components (yKu70p and yKu80p) [41,42], and Rap1p [43] are absolutely required [12,39]. Sir proteins are in limited amount and concentrated at telomere clusters located at the nuclear periphery [44,45] and establishment of silencing requires the localization of the cis-acting silencer at these Sir-rich regions [44]. In other organisms, some of the factors mediating TPE are functional orthologs of S. cerevisiae proteins. However, other factors, such as trimethylated H3K9 or HP1 required for TPE in Drosophila and fission yeast, are absent in S. cerevisiae. [46–49]. In eukaryotes, both telomeres and subtelomeres contain nucleosomes [50–52] enriched in chromatin marks

FIGURE 6.2  Telomere length modulates expression of subtelomeric genes by continuous cis-spreading or long-distance looping. (A) In eukaryotes, telomeres silence proximal gene (light green) through a mechanism named telomeric position effect (TPE) involving the nucleation of heterochromatin and spreading of chromatin modifications from the telomere to the subtelomeric region. TPE involves component of the shelterin complex. Composition of subtelomeric regions likely modulates TPE through unknown mechanisms. Some genes are not modulated by TPE (blue), while expression of other genes located up to 10 Mb from the chromosome end (dark green) can be modulated by telomere length, suggesting a discontinuous phenomenon. This mechanism involves looping of telomeres toward its target genes and modulates gene expression through telomere position effect over long distance (TPE-OLD). By comparison to PEV, TPE might thus be an alternative and specialized silencing process acting through the interaction between components of the shelterin complex and chromatin-remodeling factors. (B) Upon telomere shortening, telomeres and subtelomeric regions are less compacted and loose heterochromatin marks. This euchromatinization alleviates TPE and leads to the induction of subtelomeric gene expression. Upon telomere shortening, telomere-associated looping is impaired and expression of TPE-OLD genes increased.

II.  Molecular Mechanisms of Epigenetics



Genome Topology and Scaffolding of Chromosomal Domains

found at constitutive heterochromatin regions [53–57] (Fig. 6.2A). The capacity of mammalian telomeres to induce position effect has been controversial for many years, mostly due to technical challenges. The first evidence in vivo came from the analysis of replication timing of human chromosome 22 carrying a chromosomal abnormality [58]. Evidence for transcriptional silencing in the vicinity of human telomeres [59,60] and identification of modulators of TPE [61–63] were finally obtained experimentally by a telomere-seeding procedure where natural telomeric regions have been replaced by artificial ones using a construct containing a reporter gene. Among others, SIRT6 (Sir2p homolog), HP1, CBX complex, and Suv39h were identified as important factors dictating TPE and the chromatin landscape both at telomeres and subtelomeres, highlighting similarities with other organisms. In human, the first endogenous TPE gene described was ISG15 (located at the 1p36 locus) [64]. However, unlike ISG15, the expression of genes located closer to the 1p telomeres are not influenced by telomere length and TPE factors, such as SIRT6 [63]. This suggested an additional mechanism more complex than a continuous silencing phenomenon (Fig. 6.2A). As described yeast, where a telomere loop represses distal gene while not affecting intervening gene, a recent study uncovered an equivalent mechanism in humans [65]. This extended TPE called TPE-over long distances (TPEOLD) explains through formation of a looping structure the peculiar case of ISG15 and showed the potential to influence genes located as far as 10 Mb away from telomeres, allowing the identification of new genes regulated by telomeres [65]. Altogether, those experimental systems revealed that telomere length and architecture along with chromatin-remodeling factors are involved in TPE and TPE-OLD (Fig. 6.2).

SETTING THE FRONTIERS OF CHROMATIN DOMAINS Active and inactive chromatin domains are often juxtaposed, but their respective identities are maintained by specialized elements [66–68] (Fig. 6.1A). The first insulator elements, scs and scs′, were described in Drosophila at the heat shock locus [69]. Since, many more insulators have been described and defined by two nonexclusive properties: enhancer-blocking and barrier activities [67] (Fig. 6.1B–C). Barrier insulators protect genes or regions from the spreading of heterochromatin (Fig. 6.1B). Enhancer-blocking insulators can disrupt the communication between a promoter and a cis-regulatory element when placed in-between, without preventing their interactions with other genetic elements (Fig. 6.1C). Insulator elements bind specific proteins. In Drosophila, five proteins with enhancer-blocking activity have

77

been identified: Zw5, BEAF-32 [70,71], the GAGA factor [72,73], Su(Hw), and an ortholog of the CCCTC-binding factor, CTCF [74,75]. CTCF is the only conserved protein displaying such activity in vertebrates [76]. There are two nonexclusive mechanistic models explaining how these sequences and associated factors block the spreading of repressive marks. The first one involves a dynamic model based on a competition between euchromatin- and heterochromatin-remodeling complexes (Fig. 6.1A). Another strategy that stops the spreading of heterochromatin relies on the properties of various sequences to induce local nucleosome exclusion or replacement of core histones by histone variants, considering barrier activity as a physical block to the spreading of chromatin modifications (Fig. 6.1A).

GENOME TOPOLOGY AND SCAFFOLDING OF CHROMOSOMAL DOMAINS Recent experimental procedure developments, such as the chromatin conformation capture (3C) [77,78] and derivatives (4C, 5C, and HiC), ChIA-PET [79], RNA tagging and recovery of associated proteins (RNA TRAP) or DamID [80], provided new insights into the spatial organization of chromatin during interphase and the description of subdomains, such as lamin-associated domains (LADs) or nucleolar-associated domains (NADs) formed by association of chromatin with subnuclear structures, the nuclear lamina and the nucleolus, respectively. Individual chromosomes occupy specific nuclear subdomains defined as chromosome territories and are organized in active and open domains or inactive and condensed structures named topologically associated domains (TADs) [81–84] (Fig. 6.3A). These domains exhibit frequent intrachromatin domain interactions but rarely interdomain interactions, and maintain enhancer– promoter interactions or control spatiotemporal gene expression pattern (Fig. 6.3A). However, independent studies in ESCs or ESCs-derived neurons also described well-defined “insulated neighborhoods” of interactions within TADs [84–86]. TADs are separated by boundary regions that contain CTCF-binding sites or housekeeping genes able to block interactions between adjacent TADs. TADs are conserved between cell types but contain smaller domains, or sub-TADs, which are differentially regulated depending on the cell type or differentiation stage, and TADs from related species are highly similar [84,87]. Deletion or inversion of boundaries disrupts TAD organization and gene expression in cultured cells and animals [83,88] (Fig. 6.3B). Notably, almost half of the epigenetic marks retained on chromosomes at mitosis correspond to boundaries of topological domains underlining the

II.  Molecular Mechanisms of Epigenetics

78

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

FIGURE 6.3  Chromosome rearrangements cause disruption of topologically associated chromatin domains and modification in gene expression. (A) Chromatin can be organized in topologically associated domains (TADs) flanked and separated from adjacent TADs by insulator elements bound by CTCF. TADs exhibit interdomain interactions. Within individual TADs, genes can be regulated by specific enhancer elements ensuring correct spatiotemporal expression. (B) Constitutive pathologies involving chromosomal rearrangements (deletions, duplications, and inversions) of large chromosomal domains can modify TAD boundaries and formation of chromatin loops, leading to the ectopic interactions with distant enhancers and modifications in gene expression.

importance of mitotic bookmarking for the restoration of long-range interaction and TADs after cell division [2].

BOUNDARY ELEMENTS ARE INVOLVED IN FUNCTIONAL GENOME PARTITIONING Eukaryotic genomes are partitioned in functional domains often separated by boundary elements. Boundaries are often associated with housekeeping genes, tRNAs, and short interspersed elements (SINE). Insulator elements are also involved in the delimitation of functional domains and genome partitioning. In Drosophila, for instance, the Su(Hw) protein is present at band/ interband transitions on polytene chromosomes [89–91]. In human cells, recent genome-wide studies in different cell types identified up to 39,000 putative sites for the architectural CTCF protein that binds to a wide range of sequences. Distribution of these sites along chromosomes follows the distribution of genes; CTCF-depleted regions often correspond to clusters of coexpressed genes, while CTCF-enriched regions contain genes with multiple alternative promoters [92–94]. These sites are involved in long-range interactions, act together with cohesins and mediator complexes, [95–97] and often correspond to boundaries between internal and peripheral sequences based on their interaction with LADs [80]. The main functions of CTCF in boundary activity, chromatin organization, and looping have been

described recently in several reviews [98,99], but very little is known about the dysregulation of insulator elements in human diseases. Recent evidence suggests a role for CTCF in pathologies. For instance, microdeletion or microduplications of the CTCF sites at the IGF2/H19 locus are associated with some cases of nonsyndromic Wilms’ tumors [100]. CTCF also flanks CTG/CAG trinucleotide repeats at several disease-associated loci, such as the DM1 locus implicated in myotonic dystrophy (or Steinert disease, OMIM #160900), an autosomal dominant multisystemic disorder characterized by myotonia, muscular dystrophy, cataracts, hypogonadism, cardiac conduction anomaly, and diabetes mellitus [101–103], or other diseases linked to triplet expansion. The presence of CTCF-binding sites upstream and downstream of the expanded CTG tract restricts the extent of antisense transcription and constrains the spreading of heterochromatin from the trinucleotide repeats and DNA methylation [101,102]. CTCF may also be associated with another muscular dystrophy named facio–scapulo–humeral dystrophy (FSHD, OMIM #158900–158901). This autosomal dominant pathology is linked to the shortening of an array of repeated macrosatellite elements at the distal end of chromosome 4q arm [104,105]. Normal 4q35 ends carry from 11–150 copies of this element, while this number is reduced to 1–10 repeats in most patients. D4Z4 acts as a potent CTCF- and A-type lamins–dependent insulator in cells from patients, but not in control cells (see further for details) [106].

II.  Molecular Mechanisms of Epigenetics



Chromosomal Position Effect in Human Pathologies

In addition, one can speculate that hypomorphic mutations in factors associated with CTCF chromatin barriers, such as cohesins mutated in the Cornelia de Lange syndrome [107], or lamins that give rise to at least 10 distinct and heterogeneous genetic diseases called laminopathies [108] might also alter boundary activity and induce CPE in human cells. Overall, disrupting the integrity of the complexes mediating the partitioning of functional chromatin domains may cause a wide range of defects by affecting insulators and chromatin boundaries that could lead or contribute to various syndromes.

CHROMOSOMAL POSITION EFFECT IN HUMAN PATHOLOGIES Mechanisms through which CPE can cause human diseases are diverse: separation of the transcription unit from an essential distant regulatory element, juxtaposition of the gene with the enhancer of another gene, competition for the same regulatory element, or classical PEV in which a gene is moved to a new chromatin environment [109]. CPE-associated pathogenesis has been described in cancers, as well as in constitutional pathologies, essentially in the context of chromosomal rearrangements (translocations, deletions, and inversions). For instance, malignant hemopathies are characterized by acquired chromosomal rearrangements, mainly translocations, which are clonal, nonrandom, recurrent, and often tumor-type specific. These translocations have two main consequences: (1) the formation of a chimerical gene encoding a new fusion protein or (2) the combination between the coding region of a gene and the promoter/enhancer region of another one, leading to inappropriate overexpression of the former gene [110]. A well-known example is the t(8;14)(q24;q32) translocation in Burkitt’s lymphoma, an aggressive B-cell neoplasm that places the c-MYC gene (chr. 8) near the enhancer of the heavy chain of immunoglobulin on chromosome 14, resulting in overexpression of c-MYC in B cells [111]. Other translocations of genes, such as BCL2 (follicular lymphoma) [112], BCL6 (diffuse large B-cell lymphoma) [113], and the TMPRSS2-ETS fusion gene involved in prostate cancer [114] can also be cited. More recently, group 3 and 4 medulloblastomas, a leading cause of cancer-related mortality in children, have been associated with somatic genomic rearrangements associated with GFI1/GFI1B activation. This “enhancer hijacking” induces redistribution of the two genes from regions of transcriptionally silent chromatin to regions enriched in acetylated H3K9 and K27, usually associated with superenhancers [115]. The involvement of CPE in constitutional pathologies is in most cases more difficult to prove even if this mechanism is often proposed to explain disease-associated

79

phenotype in the absence of gene disruption [116–118]. However, a firm demonstration is rarely found in the literature for several reasons. First, in developmental syndromes, the causative gene often shows a specific spatiotemporal pattern of expression and access to the tissue from the patient is, in most cases, impossible. Second, regulatory elements are often located far from the gene of interest, sometimes included in other neighboring genes, and can regulate more than one gene. Finally, if the position effect seems evident when a chromosomal breakpoint occurs in the vicinity of a gene for which the associated phenotype is well known, it is less obvious when the phenotype is not related to the gene or when the gene function is unknown. Examples of pathologies in which CPE has been suspected or proved are summarized in Table 6.1. Most of these examples correspond to disruption, but in few cases, complex rearrangements, such as paracentric inversion or other constitutional complex chromosomal rearrangements, might also lead to position effects. In addition, DNA copy number variation (CNV) associated with human genetic diversity might account for some of the gene variation existing in the human genome. The majority of CNVs are biallelic polymorphisms (deletions, insertions, or duplications larger than 1 kb and up to several megabases), but different regions with segmental duplications have been observed, likely due to nonallelic homologous recombination [166]. With regard to CPE, not all gene CNVs result in changes in gene expression levels, but a negative correlation between copy number and expression level was found for 5%–15% of CNVs [167–170]. CPE has been involved in a number of developmental syndromes. In particular, two regions involved in congenital defects have been thoroughly characterized, and chromosomal aberrations allowed the identification of specific subregions associated with long-range position effect [159,160,171–174]. The first region is the SOX9 locus on chromosome 17 [175]. Its haploinsufficiency is responsible for campomelic dysplasia (OMIM #114290), a rare autosomal dominant disorder involving shortening and bowing of long bones, skeletal malformations, Pierre–Robin sequence (PRS, associating micrognathia, glossoptosis, and cleft palate), hypoplastic lung, and male-to-female sex-reversal. Translocation or inversion breakpoints sparing the SOX9 sequence allowed the identification of three different clusters associated with either: (1) campomelic dysplasia, (2) Pierre–Robin sequence, or (3) an intermediate acampomelic/campomelic phenotype. The second region is associated with split-hand/foot malformation (SHFM1, OMIM #183600), an autosomal dominant congenital limb defect characterized by a median cleft of the hand or foot (ectrodactyly) resulting in an aspect of “lobster claw” and caused by chromosomal

II.  Molecular Mechanisms of Epigenetics

80

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

TABLE 6.1 Constitutional Syndromes Involving Position Effect Mechanisms Genes

Loci

Phenotypes

Modifications

References

APC

5q22.2

Adenomatous polyposis

Disruption

[119]

DLX5/DLX6/DSS 1

7q22

Split-hand/foot malformation type I

Disruption

[120]

DSC3/2/DSG1/4/2/ TTR/DTNA

18q12.1

ASD

Duplication/CNV

[121]

FGFR4/MSXS

5q35

Hunter–McAlpine syndrome

Duplication

[122]

FOXC1

6p25

Primary congenital glaucoma

Disruption

[123]

FOXC2

16q24.3

Lymphedema distichiasis

Disruption

[124]

Rett syndrome

Deletion

[125]

FOXG1 FOXL2

3q23

Blepharophimosis ptosis epicanthus inversus syndrome

Disruption/deletion

[126,127]

FOXP2

7q31

Speech and language disorder

Disruption

[128]

GLI3

7p13

Greig cephalopolysyndactyly syndrome

Disruption

[129]

HBA HBB

16p13.3 11p15.5

Alpha-thalassemia Gamma/Beta-thalassemia

Deletion

[130–132]

HOXB

17q21.3

Intellectual impairment, hexadactyly

Disruption

[133]

HOXD

2q31

Limb malformations

Disruption

[134,135]

LCT

2q21

Adult-type hypolactasia

Mutation

[136]

LMNB1

5q23.2

Adult-onset demyelinating leukodystrophy

Deletion

[137,138]

MAF

16q23

Cataract, anterior segment dysgenesis, microphthalmia

Disruption

[139]

PAX6

11p13

Aniridia

Disruption

[140]

PITX1

5q31.1

Liebendberg syndrome

Deletion

PITX2

4q25

Rieger syndrome

Disruption

[141]

PLP1

Xq22

Pelizaeus–Merzbacher syndrome/spastic paraplegia, neuropathy

Disruption/duplication

[142,143]

POU3F4

Xq21.1

X-linked deafness

Deletion

[144]

REEP3

10q21.3

Autism

Disruption

[145]

RUNX2

6p21

Cleidocranial dysplasia

Disruption

[146]

SALL1

16q12.1

Townes–Brocks syndrome

Disruption

[147]

SDC2

8q22

Autism, multiple exostoses

Disruption

[148]

SHH

7q36

Holoprosencephaly

Disruption

[149]

SHH SHOX

7q36 Xp22.3

Pre-axial polydactyly Leri–Weill dyschondrosteosis

Disruption Deletion/mutation

[150–152]

SIX3

2p21

Holoprosencephaly

Disruption

[153]

SOST

17q21

Van Buchem disease

Deletion

[154]

SOX3

Xq27.1

Hypoparathyroidism

Disruption

[155]

SOX4/ID4

6p22.3

Mesomelic dysplasia: hypoplastic tibiae/fistulae

deletion

[156]

SOX9

17q24.3

Campomelic dysplasia/Pierre–Robin sequence

Disruption Deletion/mutation

[157,158] [159,160]

SRY

Yp11.3

Sex reversal

Deletion

[161]

TGFB2

1q41

Peters anomaly

Disruption

[162]

TRPS1

8q23.3

Ambras syndrome

Disruption/deletion

[163]

TWIST

7p21.1

Saethre–Chotzen syndrome

Disruption

[164]

WNT6

2q35–36.1

F syndrome, syndactyly

Inversion

[165]

II.  Molecular Mechanisms of Epigenetics



Chromosomal Position Effect in Human Pathologies

rearrangements of the 7q21q22 region (deletions, translocations, and inversions) [120,176–178]. Three distinct subregions within the SHFM1 locus containing tissuespecific enhancers controlling DLX5, DLX6, and DSS1 genes have been defined [179–182]. Aniridia (OMIM #106210) is an autosomal dominant panocular disorder characterized by complete or partial iris and foveal hypoplasia resulting in reduced visual acuity and nystagmus presenting in early infancy together with ocular abnormalities caused by haploinsufficiency of the PAX6 gene. Some patients carry 11p13 rearrangements (deletions or translocations) downstream of the PAX6 gene [183] in the final intron of the ELP4 gene containing a tissue-specific enhancer, leading to PAX6 haploinsufficiency [118,183]. An interesting familial case of PLP1-related diseases also highlighted the fact that position effect can modulate disease penetrance and even rescue the phenotype [184]. These diseases are disorders of myelin formation in the central nervous system and include Pelizaeus– Merzbacher disease (PMD, OMIM #312080) and type 2 spastic paraplegia (OMIM #312920). PMD is an X-linked recessive disease that begins in infancy and manifests by nystagmus, hypotonia, and cognitive impairment progressing to severe spasticity and ataxia caused by duplications, point mutations, and deletions of the PLP1 gene. Some cases of PMD resulting from position effect linked to small rearrangements in the region surrounding the coding sequence have been described [142,143]. Among which, one particular family composed of a mother, with late-onset PMD, carried a balanced insertion of a small segment of chromosome X including the PLP1 gene into the terminal region of the long arm of chromosome 19 [ins(19;X) (13.4;q22.2q22.2)]. One of her sons, with a mild PMD form, inherited the derivative X and carried no PLP1 copy. The second son, who is healthy, inherited the derivative chromosome 19 and two copies of PLP1. However, the PLP1 copy inserted on chromosome 19 might be silenced by the chromatin context and telomere proximity, rescuing the phenotype usually associated with abnormal PLP1 dosage. Repeat expansion diseases are caused by an increase in the number of trinucleotide repeats, which are usually polymorphic in the general population, but become unstable beyond a certain number of repeats. These trinucleotide repeats are either found in coding regions resulting in the production of a protein with altered function (i.e., Huntington disease), or in noncoding regions resulting in an altered transcription [fragile X syndrome (FXS), Friedreich’s ataxia (FDRA), and myotonic dystrophy] caused by epigenetic changes and heterochromatin formation [185]. Among them, FXS (OMIM #300624), the most common form of inherited intellectual impairment is caused by expansion of a CGG trinucleotide in the 5′-UTR of

81

the FMR1 gene. In the normal population, the number of repeats varies from 6 to 54. Between 55 and 200 CGG, the repeats become meiotically unstable (premutation), while the full mutation corresponds to more than 200 repeats, and is accompanied by hypermethylation of the repeats and neighboring sequences [186], histone H3 and H4 lysine hypoacetylation, H3K9 methylation, and FMR1 silencing [187–189]. Another interesting disease linked to triplet expansion is FRDA (OMIM #2293000), an autosomal recessive neurodegenerative disease characterized by difficulties to coordinate movements, dysarthria, loss of reflexes, pes cavus, scoliosis, cardiomyopathy, and diabetes mellitus. FRDA is caused by the expansion of a GAA trinucleotide in the first intron of the FXN gene. Repeats range from 6 to 34 in the general population and there are over 66 in patients. Expansion is associated with a decrease in FXN transcription level [190] and heterochromatin formation from the expanded GAA triplet-repeat sequence in intron 1 to the FXN promoter region [191–193]. As observed in FXS, specific CpG sites are hypermethylated in the FXN intron 1 compared to control. Hypoacetylation of H3 and H4 histones and hypermethylation of H3K9 modulate FXN promoter activity [191–193]. In addition, emerging evidence highlights the key role of TADs and TAD boundaries in congenital diseases [194]. For instance, deletions, inversions, or duplications altering the structure of the TAD spanning the WNT6/IHH/EPHA4/PAX3 locus has been implicated in at least three related genetic disorders associated with limb developmental defects and altered digits. The rearrangements disrupt the normal topology of proteincoding genes without affecting the coding sequences themselves, but enhancers and TAD boundaries result in inappropriate long-distance interactions and misexpression of the different genes. In three families affected with a dominantly inherited brachydactyly characterized by short digits, highresolution CGH revealed a heterozygous mutation of 1.75–1.9 Mb on chromosome 2q35–36, including the EPHA4 gene, a large region of its surrounding TAD, and the noncoding part of the PAX3 TAD removing the PAX3/EPHA4 boundary [165]. In a second family, affected with the F syndrome characterized by severe complex syndactyly of the hand and feet polydactyly, affected individuals carry a 1.1-Mb heterozygous inversion or a 1.4-Mb heterozygous duplication arranged in direct tandem orientation. In both cases, the breakpoint was located in the vicinity of the WNT6 gene [165]. Finally, in a family affected with a severe polysyndactyly and craniofacial abnormality, a 900-kb duplication brings the IHH gene in proximity to the centromeric potion of the EPHA4-containing TAD. The pathogenic effect of the different rearrangements was further

II.  Molecular Mechanisms of Epigenetics

82

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

confirmed in vivo by the 4C technique and in engineered mouse models that recapitulate the human phenotype. Wildtype animals show minimal interactions of Pax3, Wnt6, and Ihh with noncoding sequences in the Epha4 TAD, while all three genes showed increased interaction in strains carrying engineered mutations mimicking variants found in the patients, with novel interactions within the Epha4 TAD and a fusion of adjacent TADs. Overall, the different rearrangements described in patients led to deletion of a boundary element that prevents inappropriate cross-TAD chromatin interactions with a cluster of regulatory elements driving limb expression [165]. Genomic rearrangements associated with disease have also been described at the SOX4 locus in unrelated patients with mesomelic dysplasia where microdeletions involve four genes, three TADs, and two boundaries [156]. Deletion of the TAD boundary was also proposed in the Liebenberg syndrome (OMIM #186550), an autosomal dominant disease in which the arms acquire morphological features of the leg. In this case, the deletion removes the H2AFY gene in a region 300-kb upstream of the PITX1 gene, which determines the hind limb identity. The H2AFY gene, which encodes the H2A histone family member Y, might act as a barrier element that separates the PITX1 TAD from neighboring regulators [195]. In a case of autosomal dominant adult-onset demyelinating leukodystrophy (ADLD; OMIM #169500), causing central nervous system demyelination, a deletion upstream of the LMNB1 gene eliminates the TAD and TAD boundary, causing ectopic interactions between nearby TADs, at least three forebrain enhancers, and the LMNB1 promoter, leading to lamin B1 overexpression, as observed in other typical ADLD cases [137,138]. Overall, these examples indicate that phenotypes can be associated with disruption of interactions between regulatory sequences located in rearranged regions and distant genes (Fig. 6.3B). Nevertheless, it remains difficult to provide a conclusive computational answer on the effect of a given chromosomal rearrangement on long-distance interactions, TAD formation, and enhancer–promoter crosstalks without further experimental validation. In addition, CNVs challenge the deciphering of these long-distance interactions, and it was estimated that approximately 11.8% of CNVs might be related to the disruption of interactions between regulatory sequences and gene promoters and misregulation of phenotypically relevant genes, underlining the importance of topological domain analysis in deleted regions. Such a phenomenon could explain the Williams– Beuren syndrome (OMIM #194050), an autosomal dominant genomic disorder caused by 7q11.2 microdeletions and CNVs. The classical Williams–Beuren deletion includes 28 genes. Expression level of genes in the deleted region is half that of the control, but nonhemizygous genes located several megabases away from the

deletion also harbor significantly reduced expression levels, suggesting that long-range interactions between cisregulatory elements might modulate the expression of genes not included in the deletion and participate in the disease phenotype [196]. A similar phenomenon might also occur in the Di George syndrome (OMIM #188400) characterized by a wide range of defects and high clinical variability and caused by a 1.5–3 Mb hemizygous deletion of the 22q11.2 locus. TBX1 haploinsufficiency is responsible for most of the physical malformations [197]. However, as described earlier variable expression of genes outside the deleted region might also contribute to the phenotype. In addition, although still speculative, one might also expect that disruption of chromatin boundaries or chromatin domains might also impact mitotic bookmarking and precise reactivation of tissue-specific gene expression upon mitosis exit.

TELOMERIC POSITION EFFECT IN HUMAN PATHOLOGIES The most distal unique regions of chromosomes and telomeres are separated by different types of subtelomeric repeats varying in size from 10 to ∼100 kb in human cells. Subtelomeres are highly polymorphic gene-rich regions, and RNAs produced by these regions include transcripts from multicopy protein-encoding gene families, single genes, and a large variety of noncoding RNAs (reviewed in Ref. [34]). Large variations of subtelomeric DNA were among the first examples of CNVs in human cells. The recombination rate at chromosome ends is higher than in the rest of the genome. Telomeres influence the maintenance and recombination of the adjacent subtelomeric regions and vice versa. Subtelomeres may buffer or facilitate the spreading of silencing that emanates from the telomere, as observed in other species. Due to their heterochromatic signature, telomeres and subtelomeres have been considered as transcriptionally silent. Nevertheless, telomeric repeats RNA transcripts (TERRA), transcribed from subtelomeric promoters recently changed the paradigm of this “silent” region [198]. Contradictory reports have presented cases where TERRAs are either increased when telomeres are short or downregulated when telomeres are long [198] and vice versa [199–201]. Further studies are thus required to fully appreciate the potential beneath TERRAs production and their implication in disease or potential role in TPE, as TERRAs can act as chromatin modifiers. Subtelomeric regions are not only associated with genome evolution, but also human disorders, and the existence of TPE has been evoked in many cases for patients carrying truncated chromosome ends that have been repaired by the process of telomeric healing

II.  Molecular Mechanisms of Epigenetics



Telomeric Position Effect in Human Pathologies

(de novo addition of telomeric repeats by telomerase at breaks devoid of exact telomeric repeats), telomeric capture (resulting from a break-induced replication event between a truncated chromosome and the distal arm of another chromosome), or formation of ring chromosomes (formed by fusion between the p and q arms within the same chromosome). TPE is influenced by telomere length, a variable that is partially inherited, tissue specific, and influenced by environment-driven insults, notwithstanding the telomere length heterogeneity of each chromosome end. At present, the molecular mechanisms associated with these interchromosomal variations remain poorly investigated. Moreover, differential sensitivity, inherited, and insult-driven telomere shortening have been proposed as an explanation for generational differences in some human disorders, including in congenital syndromes associated with telomere deficiency (or telomeropathies). Subtelomeric imbalances that include deletions, duplications, unbalanced translocations, and complex rearrangements [202] are terminal, as well as interstitial and extremely variable in size [203]. Some telomeric polymorphisms and transmitted subtelomeric imbalances are benign and not associated with any phenotypical manifestation [204,205]. However, the importance of subtelomeric rearrangements affecting all chromosomes, with the exception of the short arms of acrocentric chromosomes, is well established in 5%–10% of idiopathic mental retardation [157,206,207]. The refinement of diagnostic techniques allowed the identification of a number of genes, but many cases remain poorly characterized [158,208,209], and rather a few deleterious subtelomeric imbalances are associated with a distinct, recognizable phenotype. For example, in the cri-du-chat syndrome (5pter deletion), which is characterized by microcephaly, facial dysmorphism, high-pitched cat-like cry, severe intellectual impairment, and speech delay [210,211], critical regions corresponding to cry (5p15.31), speech delay (5p15.32– 15.33), and facial dysmorphism (5p15.31–15.2) have been delineated [164]. However, no gene has been identified for each specific feature yet. In other cases, candidate genes remain to be identified. The 1p36 monosomy syndrome is the most common subtelomeric microdeletion syndrome with a frequency of 1/5000. It is characterized by intellectual impairment, developmental delay, hearing loss, seizures, growth impairment, hypotonia, heart defect, and a distinctive dysmorphism [212]. In the absence of common breakpoint or common deletion interval, no correlation between the deletion size and the number of clinical features could be made [212], leading to the hypothesis that the 1p36 monosomy syndrome might be due to a positional effect rather than haploinsufficiency of contiguous genes.

83

Apart from subtelomeric rearrangements, hundreds of patients carrying ring chromosomes have been reported with various combinations of malformations, minor abnormalities, growth retardation, and intellectual impairment [213,214]. Ring chromosomes are circularized chromosomes with either a nonsupernumerary ring, which replaces one of the two alleles or the presence of a supernumerary ring chromosome. Supernumerary ring chromosomes are usually associated with an abnormal phenotype. Intact rings putatively causing TPEs have been reported for the different autosomes [215–217]. They can be formed by deletion near the end(s) of chromosomes followed by fusion at breakage points or fusion without loss of genetic material. The resulting phenotypes vary greatly depending on the size and nature of the deleted segments. In rings formed by telomere–telomere fusion with little or no loss of chromosomal material, the “ring syndrome” might be associated with the silencing of genes in the vicinity of a longer telomere. Moreover, analysis of telomere length in patients with ring chromosomes and their parents suggest that some chromosomes might be prone to ring formation when reaching a critical telomere length [218], suggesting that besides influencing the phenotypic spectrum, telomere length might also influence the familial transmission of the mosaic ring. To date, the implication of TPE and TPE-OLD in diseases is solely illustrated in FSHD [106,219,220]. As mentioned, this disease is a skeletal muscle dystrophy with a late age of onset linked, for most cases, to contraction of the D4Z4 repetitive array on the 4q35 locus, In addition to a CTCF-associated insulator activity and its capacity to tether telomeres to the nuclear envelope [106], each D4Z4 repeats contains the DUX4 open reading frame encoding a highly conserved double homeobox protein. In FSHD cells, DUX4 might be transcribed in 1/1,000 cells, as opposed to 1/10,000 in non-FSHD cells, making DUX4 a candidate gene for FSHD pathogenesis. It was recently shown that TPE alleviates DUX4 expression, as telomere length reduction correlates with the increased expression seen in the fraction of DUX4-positive cells expected in FSHD [219]. Besides TPE, TPE-OLD can reshape the chromatin landscape of the 4q35 locus, enhancing the expression of genes (i.e., SORBS2) located as far as 5-Mb upstream of the D4Z4 repeats, reinforcing the notion of a cooperative effect between telomere length and size of the D4Z4 array [106,219–221]. TPE and TPE-OLD could thus elegantly explain the late-stage onset of FSHD, as well as the wide heterogeneity of its clinical manifestation. Based on these observations, one can speculate that accelerated myoblast proliferation triggers telomere shortening, DUX4 activation, and accumulation of toxic factors in foci, leading in turn to cycles of degeneration and regeneration and potential exhaustion in the pool of cells able to regenerate the damaged muscle. However,

II.  Molecular Mechanisms of Epigenetics

84

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

other candidate genes (FRG1, FAT1, SORBS2, etc.) and epigenetic changes might also contribute to the disease. Hence to the best of our knowledge, TPE can extend as far as 80 kb (DUX4) and TPE-OLD as far as 10 Mb in human cells, unraveling potential an unthought of implication of telomeres on gene regulation or pathologies through position effect mechanisms. According to the work on TPE-OLD, all telomeres are not equal and disparities can be found (i.e., tissue specificity), with chromosome ends highly affected by telomere length changes (i.e., numerous genes affected; 3p, 6p, 12p, and 20p) or not (no change in gene transcription; 7q, 11q, and 13q). This suggests previously unthought of telomere length implications in pathologies linked to genes located within the “extended” reach of telomeres (e.g., 10 Mb). Hence, one can argue that telomere dynamics and role in chromatin structure can influence the onset

of pathologies and variability in clinical manifestations (due to selective telomere shortening), as seen in FSHD. A list of genes linked to pathologies and located in subtelomeric regions at an arbitrary distance of 10 Mb from the telomere is provided in Table 6.2. Altogether, this suggests that a wide spectrum of diseases might be affected by telomere length through long-distance regulation of their associated genes. This cut-off at 10 Mb and the tissue-specific diversity of telomere length might underestimate all the possible consequences of telomere length on pathologies, interindividual variability, or age of onset. Due to the paucity of cases and the variability in the size of the terminal deletion, genotype–phenotype correlations are not established for most syndromes involving either subtelomeric imbalance or ring formation. Epigenetic modifications, chromatin condensation, or loop formation

TABLE 6.2 Diseases Linked to Gene Located in Subtelomeric Regions

Chromosomes

Telomere distances (Mb)

Genes associated

Diseases associated

OMIM #

1q

7.6

FH

Fumarase deficiency/leiomyomatosis and renal cell cancer

606812/150800

1q

5.8

SDCCAG8

Senior–Loken syndrome 7

613615

2p

3

RPS7

Diamond–Blackfan anemia 8

612563

3p

4.5

ITPR1

Spinocerebellar ataxia 15 and 29

606658/117360

3p

8.8

CAV3

Cardiomyopathy, familial hypertrophic/creatine phosphokinase, elevated serum/long QT syndrome 9/muscular dystrophy, limb-girdle, type IC/myopathy, distal, Tateyama type/rippling muscle disease

192600/123320/611818/ 607801/614321/ 606072

3q

8.4

TP63

ADULT syndrome/ectrodactyly, ectodermal dysplasia, and cleft lip/palate syndrome 3/Hay–Wells syndrome/limb-mammary syndrome/orofacial cleft 8/Rapp–Hodgkin syndrome/splithand/foot malformation 4

103285/604292/10626 0/603543/129400/6 05289

3q

4.5

OPA1

Optic atrophy 1/optic atrophy with or without deafness, ophthalmoplegia, myopathy, ataxia, neuropathy

165500/125250

3q

0.1

RPL35A

Diamond–Blackfan anemia 5

612528

4p

0.6

PDE6B

Night blindness, congenital stationary, autosomal dominant 2/ retinitis pigmentosa 40

163500/613801

4p

0.9

IDUA

Mucopolysaccharidosis Ih/s/Is

607014/607015/607016

4p

1.8

FGFR3

Achondroplasia/bladder cancer, somatic/CATSHL syndrome/ cervical cancer, somatic/colorectal cancer, somatic/Crouzon syndrome with acanthosis nigricans/hypochondroplasia/ LADD syndrome/Muenke syndrome/nevus, epidermal, somatic/spermatocytic seminoma, somatic/thanatophoric dysplasia, type I and II

100800/109800/610474 /603956/114500/612 247/146000/149730/ 602849/162900/2733 00/187600/187601

4p

2.8

SH3BP2

Cherubism

118400

4p

3.2

HTT

Huntington disease

613004

4p

3.5

DOK7

Fetal akinesia deformation sequence/myasthenia limb-girdle familial

208150/254300

4p

6.3

WFS1

Deafness, autosomal dominant 6/14/38/Wolfram syndrome

600965/222300/614296

II.  Molecular Mechanisms of Epigenetics



85

Telomeric Position Effect in Human Pathologies

TABLE 6.2 Diseases Linked to Gene Located in Subtelomeric Regions (cont.)

Chromosomes

Telomere distances (Mb)

Genes associated

Diseases associated

OMIM #

4q

4

CYP4V2

Bietti crystalline corneoretinal dystrophy

608614

5p

0.2

SDHA

Cardiomyopathy dilated 1GG/Leigh syndrome/mitochondrial respiratory chain complex II deficiency/paragangliomas 5

613642/256000/ 252011/614165

5p

1.2

TERT

Related to: bone marrow failure, telomere-related, 1/coronary artery disease/dyskeratosis congenita, autosomal dominant 2 and 4/leukemia, acute myeloid/melanoma, cutaneous malignant, 9/pulmonary fibrosis, telomere related, 1

614742/613989/601626 /615134/614742

5p

7.8

MTRR

Homocystinuria–megaloblastic anemia cbl E type/susceptibility to homocystinuria–megaloblastic anemia, cbl E type

236270/601634

5q

6.6

MSX2

Craniosynostosis, type 2/parietal foramina 1/parietal foramina with cleidocranial dysplasia

604757/168500/168550

5q

4.2

NSD1

Beckwith–Wiedemann syndrome/Sotos syndrome 1

130650/601626/117550

5q

3.3

PROP1

Pituitary hormone deficiency, combined, 2

262600

5q

3.1

NHP2

Dyskeratosis congenita, autosomal recessive 2

613987

5q

0.7

FLT4

Hemangioma, capillary infantile, somatic/lymphedema, hereditary I

602089/153100

6p

7.5

DSP

Arrhythmogenic right ventricular dysplasia 8/dilated cardiomyopathy with woolly hair and keratoderma/ epidermolysis bullosa, lethal acantholytic/keratosis palmoplantaris striata II/skin fragility–woolly hair syndrome

607450/605676/609638 /612908/607655

6q

9.3

PARK2

Adenocarcinoma of lung, somatic/adenocarcinoma of lung, somatic/Parkinson disease, juvenile, type 2

211980/167000/600116

6q

0.2

TBP

Spinocerebellar ataxia 17/suceptibility to Parkinson

607136/168600

9p

2.6

VLDLR

Cerebellar hypoplasia and intellectual impairment with or without quadrupedal locomotion 1

224050

9p

6.5

GLDC

Glycine encephalopathy

238300

12p

0.9

WNK1

Neuropathy, hereditary sensory and autonomic, type II/ pseudohypoaldosteronism, type IIC

201300/614492

12p

2

CACNA1C

Brugada syndrome 3/Timothy syndrome

611875/601005

12p

2.2

CACNA1C

Retinal cone dystrophy 4/susceptibility to major affective disorder-9

610478/612372

12p

5

KCNA1

Episodic ataxia/myokymia syndrome/atrial fibrillation familial 7

160120/612240

12p

6

VWF

von Willebrand disease, type 1/types 2A, 2B, 2M, and 2N/type 3

193400/613554/277480

12p

7

ATN1

Dentatorubral pallidoluysian atrophy

125370

12p

7.3

PEX5

Peroxisome biogenesis disorder 2

214110/202370

12q

9.7

EIF2B1

Leukoencephalopathy with vanishing white matter

603896

12q

9.7

TCTN2

Meckel syndrome 8

613885

12q

9.6

ATP6V0A2

Cutis laxa, autosomal recessive, type IIA/wrinkly skin syndrome

219200/278250

13q

4.2

COL4A1

Angiopathy, hereditary, with nephropathy, aneurysms, and muscle/brain small vessel disease with Axenfeld–Rieger anomaly/brain small vessel disease with hemorrhage/ porencephaly 1

611773/607595/175780

14q

4.9

DYNC1H1

Charcot–Marie–Tooth disease, axonal, type 20/intellectual impairment, autosomal dominant 13/spinal muscular atrophy, lower extremity, autosomal dominant

158600/614228/614563

(Continued) II.  Molecular Mechanisms of Epigenetics

86

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

TABLE 6.2 Diseases Linked to Gene Located in Subtelomeric Regions (cont.)

Chromosomes

Telomere distances (Mb)

Genes associated

14q

2.1

18q

Diseases associated

OMIM #

AKT1

Breast cancer, somatic/colorectal cancer, somatic/Cowden syndrome 6/ovarian cancer, somatic/Proteus syndrome, somatic

114480/114500/61510 9/167000/176920/1 81500

0.6

CTDP1

Congenital cataracts, facial dysmorphism, and neuropathy

604168

20p

2.6

IDH3B

Retinitis pigmentosa 46

612572

20p

3.9

PANK2

HARP syndrome/neurodegeneration with brain iron accumulation 1

607236/234200

20p

4.7

PRNP

Creutzfeldt–Jakob disease/Gerstmann–Sträussler–Scheinker disease/Huntington disease–like 1/insomnia, fatal familial/ prion disease with protracted course

123400/137440/603218 /600072/606688

20p

5.3

PROKR2

Hypogonadotropic hypogonadism 3 with or without anosmia

244200

Xp

0.6

SHOX

Langer mesomelic dysplasia/Leri–Weill dyschondrosteosis/short stature, idiopathic familial

249700/127300/300582

Xp

2.8

ARSE

Chondrodysplasia punctata, X-linked recessive

302950

Xp

0.6

NLGN4X

Intellectual impairment, X-linked/susceptibility to Asperger and autism

300495/300497

Xp

8.5

KAL1

Hypogonadotropic hypogonadism 1 with or without anosmia (Kallmann syndrome 1)

308700

Xp

9.7

GPR143

Nystagmus 6, congenital, X-linked/ocular albinism, type I, Nettleship–Falls type

300814/300500

Yp

0.5

SHOX

Short stature, idiopathic familial

400020

Yp

2.6

SRY

46XX sex reversal 1/46XY sex reversal 1

400045/400044

might alter the expression of a number of genes in the vicinity, rearranged region, or at distance by modulating the formation or higher-order structures. However, characterization of the rearrangement’s effect on gene expression is still needed. Moreover, a number of genes implicated diseases are found in these subtelomeric regions (Table 6.2), suggesting that telomere length might also participate in the disease phenotype of a wide range of pathologies.

CONCLUSIONS The extent by which epigenetic changes and position effects contribute to the diversity of human phenotypes is increasingly being recognized, especially when genome-wide association fails to establish a clear genotype/­ phenotype correlation. Through this review, our aim was to emphasize that besides genetic changes linked to inherited pathologies, alterations of chromatin marks and chromatin topology and disruption of the equilibrium maintaining epigenetic information might be associated with disease onset, susceptibility, or penetrance. Unraveling the complexity of human CPE and TPE in the context of health and diseases is contingent upon

our knowledge of epigenetic regulations implicated in pathogenic pathways, and decades of extensive research on model organisms should help to validate the hypothetical mechanisms involved. In the future, a better understanding of human PEV, the identification of proteins and pathways involved in its regulation, and the regional specificities (telomere proximity, euchromatic, or heterochromatic regions) might thus be considered for a better understanding of the clinical variability and pathomechanisms associated with complex diseases. Together with epigenetic regulation in general, deciphering the mechanisms involved in gene variegation is thus a major challenge of the postgenomic era for the understanding and cure of a wide range of human diseases. Altogether, this also underlines the necessity to revisit old findings and observations in light of recent advances in whole genome and large-scale tridimensional chromatin analyses, as structural variants might perturb TAD structure and result in aberrant long-distance interactions. In agreement with this hypothesis, this might also explain how different types of large-scale structural changes might give rise to the same phenotype.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

References [1] Egli D, Birkhoff G, Eggan K. Mediators of reprogramming: transcription factors and transitions through mitosis. Nat Rev Mol Cell Biol 2008;9(7):505–16. [2] Kadauke S, Blobel GA. Mitotic bookmarking by transcription factors. Epigenetics Chromatin 2013;6(1):6. [3] Zaidi SK, Grandy RA, Lopez-Camacho C, Montecino M, van Wijnen AJ, Lian JB, et al. Bookmarking target genes in mitosis: a shared epigenetic trait of phenotypic transcription factors and oncogenes? Cancer Res 2014;74(2):420–5. [4] Caravaca JM, Donahue G, Becker JS, He X, Vinson C, Zaret KS. Bookmarking by specific and nonspecific binding of FoxA1 pioneer factor to mitotic chromosomes. Genes Dev 2013;27(3):251–60. [5] Peng JC, Karpen GH. Epigenetic regulation of heterochromatic DNA stability. Curr Opin Genet Dev 2008;18(2):204–11. [6] Trojer P, Reinberg D. Facultative heterochromatin: is there a distinctive molecular signature? Mol Cell 2007;28(1):1–13. [7] Bernstein BE, Mikkelsen TS, Xie X, Kamal M, Huebert DJ, Cuff J, et al. A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 2006;125(2):315–26. [8] Mikkelsen TS, Ku M, Jaffe DB, Issac B, Lieberman E, Giannoukos G, et al. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature 2007;448(7153):553–60. [9] Margueron R, Trojer P, Reinberg D. The key to development: interpreting the histone code? Curr Opin Genet Dev 2005;15(2):163–76. [10] Azuara V, Perry P, Sauer S, Spivakov M, Jorgensen HF, John RM, et al. Chromatin signatures of pluripotent cell lines. Nat Cell Biol 2006;8(5):532–8. [11] Girton JR, Johansen KM. Chromatin structure and the regulation of gene expression: the lessons of PEV in Drosophila. Adv Genet 2008;61:1–43. [12] Ottaviani A, Gilson E, Magdinier F. Telomeric position effect: from the yeast paradigm to human pathologies? Biochimie 2008;90(1):93–107. [13] Rabbitts TH, Forster A, Baer R, Hamlyn PH. Transcription enhancer identified near the human C mu immunoglobulin heavy chain gene is unavailable to the translocated c-myc gene in a Burkitt lymphoma. Nature 1983;306(5945):806–9. [14] Gottschling DE, Aparicio OM, Billington BL, Zakian VA. Position effect at S. cerevisiae telomeres: reversible repression of Pol II transcription. Cell 1990;63(4):751–62. [15] Phillips-Cremins JE, Corces VG. Chromatin insulators: linking genome organization to cellular function. Mol Cell 2013;50(4):461–74. [16] Chetverina D, Aoki T, Erokhin M, Georgiev P, Schedl P. Making connections: insulators organize eukaryotic chromosomes into independent cis-regulatory networks. BioEssays 2014;36(2):163–72. [17] Wang J, Lawry ST, Cohen AL, Jia S. Chromosome boundary elements and regulation of heterochromatin spreading. Cell Mol Life Sci 2014;71(24):4841–52. [18] Strurtevant A. The effect of unequal crossing over the Bar locus in Drosophila. Genetics 1925;10:117–47. [19] Muller HJ. Types of visible variations induced by X-rays in Drosophila. J Genet 1930;22:299–304. [20] Muller HJ. Further studies on the nature and causes of gene mutations. Proc Sixth Int Congr Genet 1932;1:213–55. [21] Dorer DR, Henikoff S. Transgene repeat arrays interact with distant heterochromatin and cause silencing in cis and trans. Genetics 1997;147(3):1181–90. [22] Garrick D, Fiering S, Martin DI, Whitelaw E. Repeat-induced gene silencing in mammals. Nat Genet 1998;18(1):56–9. [23] Eissenberg JC, James TC, Foster-Hartnett DM, Hartnett T, Ngan V, Elgin SC. Mutation in a heterochromatin-specific chromosomal protein is associated with suppression of position-effect variegation in Drosophila melanogaster. Proc Natl Acad Sci USA 1990;87(24):9923–7.

87

[24] Festenstein R, Sharghi-Namini S, Fox M, Roderick K, Tolaini M, Norton T, et al. Heterochromatin protein 1 modifies mammalian PEV in a dose- and chromosomal-context-dependent manner. Nat Genet 1999;23(4):457–61. [25] James TC, Elgin SC. Identification of a nonhistone chromosomal protein associated with heterochromatin in Drosophila melanogaster and its gene. Mol Cell Biol 1986;6(11):3862–72. [26] Schulze SR, Wallrath LL. Gene regulation by chromatin structure: paradigms established in Drosophila melanogaster. Annu Rev Entomol 2007;52:171–92. [27] Elgin SC, Reuter G. Position-effect variegation, heterochromatin formation, and gene silencing in Drosophila. Cold Spring Harb Perspect Biol 2013;5(8):a017780. [28] Phalke S, Nickel O, Walluscheck D, Hortig F, Onorati MC, Reuter G. Retrotransposon silencing and telomere integrity in somatic cells of Drosophila depends on the cytosine-5 methyltransferase DNMT2. Nat Genet 2009;41(6):696–702. [29] Ashe A, Morgan DK, Whitelaw NC, Bruxner TJ, Vickaryous NK, Cox LL, et al. A genome-wide screen for modifiers of transgene variegation identifies genes with critical roles in development. Genome Biol 2008;9(12):R182. [30] Blewitt ME, Vickaryous NK, Hemley SJ, Ashe A, Bruxner TJ, Preis JI, et al. An N-ethyl-N-nitrosourea screen for genes involved in variegation in the mouse. Proc Natl Acad Sci USA 2005;102(21):7629–34. [31] Akhtar W, de Jong J, Pindyurin AV, Pagie L, Meuleman W, de Ridder J, et al. Chromatin position effects assayed by thousands of reporters integrated in parallel. Cell 2013;154(4):914–27. [32] Kotecki M, Reddy PS, Cochran BH. Isolation and characterization of a near-haploid human cell line. Exp Cell Res 1999;252(2):273–80. [33] Tchasovnikarova IA, Timms RT, Matheson NJ, Wals K, Antrobus R, Gottgens B, et al. Gene silencing. Epigenetic silencing by the HUSH complex mediates position-effect variegation in human cells. Science 2015;348(6242):1481–5. [34] Mefford HC, Trask BJ. The complex structure and dynamic evolution of human subtelomeres. Nat Rev Genet 2002;3(2):91–102. [35] Mason JM, Biessmann H. The unusual telomeres of Drosophila. Trends Genet 1995;11(2):58–62. [36] Gehring WJ, Klemenz R, Weber U, Kloter U. Functional analysis of the white gene of Drosophila by P-factor-mediated transformation. EMBO J 1984;3(9):2077–85. [37] Craven RJ, Petes TD. Involvement of the checkpoint protein Mec1p in silencing of gene expression at telomeres in Saccharomyces cerevisiae. Mol Cell Biol 2000;20(7):2378–84. [38] Pryde FE, Louis EJ. Limitations of silencing at native yeast telomeres. EMBO J 1999;18(9):2538–50. [39] Mondoux MA, Zakian VA. Telomere position effect: silencing near the end. In: De Lange T, Lundblad V, Blackburn EH, editors. Telomeres. 2nd ed. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 2006. p. 261–316. [40] Aparicio OM, Billington BL, Gottschling DE. Modifiers of position effect are shared between telomeric and silent mating-type loci in S. cerevisiae. Cell 1991;66(6):1279–87. [41] Boulton SJ, Jackson SP. Components of the Ku-dependent non-homologous end-joining pathway are involved in telomeric length maintenance and telomeric silencing. EMBO J 1998;17(6):1819–28. [42] Laroche T, Martin SG, Gotta M, Gorham HC, Pryde FE, Louis EJ, et al. Mutation of yeast Ku genes disrupts the subnuclear organization of telomeres. Curr Biol 1998;8(11):653–6. [43] Kyrion G, Boakye KA, Lustig AJ. C-terminal truncation of RAP1 results in the deregulation of telomere size, stability, and function in Saccharomyces cerevisiae. Mol Cell Biol 1992;12(11):5159–73. [44] Maillet L, Boscheron C, Gotta M, Marcand S, Gilson E, Gasser SM. Evidence for silencing compartments within the yeast nucleus: a role for telomere proximity and Sir protein concentration in silencer-mediated repression. Genes Dev 1996;10(14):1796–811.

II.  Molecular Mechanisms of Epigenetics

88

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

[45] Gotta M, Laroche T, Formenton A, Maillet L, Scherthan H, Gasser SM. The clustering of telomeres and colocalization with Rap1, Sir3, and Sir4 proteins in wild-type Saccharomyces cerevisiae. J Cell Biol 1996;134(6):1349–63. [46] Cooper JP, Nimmo ER, Allshire RC, Cech TR. Regulation of telomere length and function by a Myb-domain protein in fission yeast. Nature 1997;385(6618):744–7. [47] Kanoh J, Ishikawa F. spRap1 and spRif1, recruited to telomeres by Taz1, are essential for telomere function in fission yeast. Curr Biol 2001;11(20):1624–30. [48] Wallrath LL, Elgin SC. Position effect variegation in Drosophila is associated with an altered chromatin structure. Genes Dev 1995;9(10):1263–77. [49] Ekwall K, Javerzat JP, Lorentz A, Schmidt H, Cranston G, Allshire R. The chromodomain protein Swi6: a key component at fission yeast centromeres. Science 1995;269(5229):1429–31. [50] de Lange T, Shiue L, Myers RM, Cox DR, Naylor SL, Killery AM, et al. Structure and variability of human chromosome ends. Mol Cell Biol 1990;10(2):518–27. [51] Makarov VL, Lejnine S, Bedoyan J, Langmore JP. Nucleosomal organization of telomere-specific chromatin in rat. Cell 1993;73(4):775–87. [52] Pisano S, Galati A, Cacchione S. Telomeric nucleosomes: forgotten players at chromosome ends. Cell Mol Life Sci 2008;66(22):3553–63. [53] Benetti R, Garcia-Cao M, Blasco MA. Telomere length regulates the epigenetic status of mammalian telomeres and subtelomeres. Nat Genet 2007;39(2):243–50. [54] Benetti R, Gonzalo S, Jaco I, Schotta G, Klatt P, Jenuwein T, et al. Suv4-20h deficiency results in telomere elongation and derepression of telomere recombination. J Cell Biol 2007;178(6):925–36. [55] Garcia-Cao M, O’Sullivan R, Peters AH, Jenuwein T, Blasco MA. Epigenetic regulation of telomere length in mammalian cells by the Suv39h1 and Suv39h2 histone methyltransferases. Nat Genet 2004;36(1):94–9. [56] Gonzalo S, Garcia-Cao M, Fraga MF, Schotta G, Peters AH, Cotter SE, et al. Role of the RB1 family in stabilizing histone methylation at constitutive heterochromatin. Nat Cell Biol 2005;7(4):420–8. [57] Gonzalo S, Jaco I, Fraga MF, Chen T, Li E, Esteller M, et al. DNA methyltransferases control telomere length and telomere recombination in mammalian cells. Nat Cell Biol 2006;8(4):416–24. [58] Ofir R, Wong AC, McDermid HE, Skorecki KL, Selig S. Position effect of human telomeric repeats on replication timing. Proc Natl Acad Sci USA 1999;96(20):11434–9. [59] Baur JA, Zou Y, Shay JW, Wright WE. Telomere position effect in human cells. Science 2001;292(5524):2075–7. [60] Koering CE, Pollice A, Zibella MP, Bauwens S, Puisieux A, Brunori M, et al. Human telomeric position effect is determined by chromosomal context and telomeric chromatin integrity. EMBO Rep 2002;3(11):1055–61. [61] Uhlirova R, Horakova AH, Galiova G, Legartova S, Matula P, Fojtova M, et al. SUV39h- and A-type lamin-dependent telomere nuclear rearrangement. J Cell Biochem 2010;109(5):915–26. [62] Tennen RI, Chua KF. Chromatin regulation and genome maintenance by mammalian SIRT6. Trends Biochem Sci 2011;36(1):39–46. [63] Tennen RI, Bua DJ, Wright WE, Chua KF. SIRT6 is required for maintenance of telomere position effect in human cells. Nat Commun 2011;2:433. [64] Lou Z, Wei J, Riethman H, Baur JA, Voglauer R, Shay JW, et al. Telomere length regulates ISG15 expression in human cells. Aging 2009;1(7):608–21. [65] Robin JD, Ludlow AT, Batten K, Magdinier F, Stadler G, Wagner KR, et al. Telomere position effect: regulation of gene expression with progressive telomere shortening over long distances. Genes Dev 2014;28(22):2464–76. [66] Valenzuela L, Kamakaka RT. Chromatin insulators. Annu Rev Genet 2006;40:107–38.

[67] Gaszner M, Felsenfeld G. Insulators: exploiting transcriptional and epigenetic mechanisms. Nat Rev Genet 2006;7(9):703–13. [68] Fourel G, Magdinier F, Gilson E. Insulator dynamics and the setting of chromatin domains. Bioessays 2004;26(5):523–32. [69] Udvardy A, Maine E, Schedl P. The 87A7 chromomere. Identification of novel chromatin structures flanking the heat shock locus that may define the boundaries of higher order domains. J Mol Biol 1985;185(2):341–58. [70] Gaszner M, Vazquez J, Schedl P. The Zw5 protein, a component of the scs chromatin domain boundary, is able to block enhancerpromoter interaction. Genes Dev 1999;13(16):2098–107. [71] Zhao K, Hart CM, Laemmli UK. Visualization of chromosomal domains with boundary element-associated factor BEAF-32. Cell 1995;81(6):879–89. [72] Ohtsuki S, Levine M. GAGA mediates the enhancer blocking activity of the eve promoter in the Drosophila embryo. Genes Dev 1998;12(21):3325–30. [73] Belozerov VE, Majumder P, Shen P, Cai HN. A novel boundary element may facilitate independent gene regulation in the Antennapedia complex of Drosophila. EMBO J 2003;22(12):3113–21. [74] Moon H, Filippova G, Loukinov D, Pugacheva E, Chen Q, Smith ST, et al. CTCF is conserved from Drosophila to humans and confers enhancer blocking of the Fab-8 insulator. EMBO Rep 2005;6(2):165–70. [75] Gerasimova TI, Lei EP, Bushey AM, Corces VG. Coordinated control of dCTCF and gypsy chromatin insulators in Drosophila. Mol Cell 2007;28(5):761–72. [76] Bell AC, West AG, Felsenfeld G. The protein CTCF is required for the enhancer blocking activity of vertebrate insulators. Cell 1999;98(3):387–96. [77] Dekker J, Rippe K, Dekker M, Kleckner N. Capturing chromosome conformation. Science 2002;295(5558):1306–11. [78] Dekker J, Marti-Renom MA, Mirny LA. Exploring the three-dimensional organization of genomes: interpreting chromatin interaction data. Nat Rev Genet 2013;14(6):390–403. [79] de Wit E, de Laat W. A decade of 3C technologies: insights into nuclear organization. Genes Dev 2012;26(1):11–24. [80] Guelen L, Pagie L, Brasset E, Meuleman W, Faza MB, Talhout W, et al. Domain organization of human chromosomes revealed by mapping of nuclear lamina interactions. Nature 2008;453(7197):948–51. [81] Sexton T, Yaffe E, Kenigsberg E, Bantignies F, Leblanc B, Hoichman M, et al. Three-dimensional folding and functional organization principles of the Drosophila genome. Cell 2012;148(3):458–72. [82] Dixon JR, Selvaraj S, Yue F, Kim A, Li Y, Shen Y, et al. Topological domains in mammalian genomes identified by analysis of chromatin interactions. Nature 2012;485(7398):376–80. [83] Nora EP, Lajoie BR, Schulz EG, Giorgetti L, Okamoto I, Servant N, et al. Spatial partitioning of the regulatory landscape of the Xinactivation centre. Nature 2012;485(7398):381–5. [84] Phillips-Cremins JE, Sauria ME, Sanyal A, Gerasimova TI, Lajoie BR, Bell JS, et al. Architectural protein subclasses shape 3D organization of genomes during lineage commitment. Cell 2013;153(6):1281–95. [85] Dowen JM, Bilodeau S, Orlando DA, Hubner MR, Abraham BJ, Spector DL, et al. Multiple structural maintenance of chromosome complexes at transcriptional regulatory elements. Stem Cell Rep 2013;1(5):371–8. [86] Ji X, Dadon DB, Powell BE, Fan ZP, Borges-Rivera D, Shachar S, et al. 3D Chromosome regulatory landscape of human pluripotent cells. Cell Stem Cell 2016;18(2):262–75. [87] Rao SS, Huntley MH, Durand NC, Stamenova EK, Bochkov ID, Robinson JT, et al. A 3D map of the human genome at kilobase resolution reveals principles of chromatin looping. Cell 2014;159(7):1665–80.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[88] Andrey G, Montavon T, Mascrez B, Gonzalez F, Noordermeer D, Leleu M, et al. A switch between topological domains underlies HoxD genes collinearity in mouse limbs. Science 2013;340(6137):1234167. [89] Gerasimova TI, Corces VG. Polycomb and trithorax group proteins mediate the function of a chromatin insulator. Cell 1998;92(4):511–21. [90] Ramos E, Ghosh D, Baxter E, Corces VG. Genomic organization of gypsy chromatin insulators in Drosophila melanogaster. Genetics 2006;172(4):2337–49. [91] Parnell TJ, Kuhn EJ, Gilmore BL, Helou C, Wold MS, Geyer PK. Identification of genomic sites that bind the Drosophila suppressor of Hairy-wing insulator protein. Mol Cell Biol 2006;26(16):5983–93. [92] Xie X, Mikkelsen TS, Gnirke A, Lindblad-Toh K, Kellis M, Lander ES. Systematic discovery of regulatory motifs in conserved regions of the human genome, including thousands of CTCF insulator sites. Proc Natl Acad Sci USA 2007;104(17):7145–50. [93] Kim TH, Abdullaev ZK, Smith AD, Ching KA, Loukinov DI, Green RD, et al. Analysis of the vertebrate insulator protein CTCFbinding sites in the human genome. Cell 2007;128(6):1231–45. [94] Barski A, Cuddapah S, Cui K, Roh TY, Schones DE, Wang Z, et al. High-resolution profiling of histone methylations in the human genome. Cell 2007;129(4):823–37. [95] Wendt KS, Yoshida K, Itoh T, Bando M, Koch B, Schirghuber E, et al. Cohesin mediates transcriptional insulation by CCCTCbinding factor. Nature 2008;451(7180):796–801. [96] Stedman W, Kang H, Lin S, Kissil JL, Bartolomei MS, Lieberman PM. Cohesins localize with CTCF at the KSHV latency control region and at cellular c-myc and H19/Igf2 insulators. EMBO J 2008;27(4):654–66. [97] Parelho V, Hadjur S, Spivakov M, Leleu M, Sauer S, Gregson HC, et al. Cohesins functionally associate with CTCF on mammalian chromosome arms. Cell 2008;132(3):422–33. [98] Ghirlando R, Felsenfeld G. CTCF: making the right connections. Genes Dev 2016;30(8):881–91. [99] Ong CT, Corces VG. CTCF: an architectural protein bridging genome topology and function. Nat Rev Genet 2014;15(4):234–46. [100] Scott RH, Douglas J, Baskcomb L, Huxter N, Barker K, Hanks S, et al. Constitutional 11p15 abnormalities, including heritable imprinting center mutations, cause nonsyndromic Wilms tumor. Nat Genet 2008;40(11):1329–34. [101] Filippova GN, Thienes CP, Penn BH, Cho DH, Hu YJ, Moore JM, et al. CTCF-binding sites flank CTG/CAG repeats and form a methylation-sensitive insulator at the DM1 locus. Nat Genet 2001;28(4):335–43. [102] Cho DH, Thienes CP, Mahoney SE, Analau E, Filippova GN, Tapscott SJ. Antisense transcription and heterochromatin at the DM1 CTG repeats are constrained by CTCF. Mol Cell 2005;20(3):483–9. [103] Lopez Castel A, Nakamori M, Tome S, Chitayat D, Gourdon G, Thornton CA, et al. Expanded CTG repeat demarcates a boundary for abnormal CpG methylation in myotonic dystrophy patient tissues. Hum Mol Genet 2011;20(1):1–15. [104] Wijmenga C, Hewitt JE, Sandkuijl LA, Clark LN, Wright TJ, Dauwerse HG, et al. Chromosome 4q DNA rearrangements associated with facioscapulohumeral muscular dystrophy. Nat Genet 1992;2(1):26–30. [105] Sarfarazi M, Wijmenga C, Upadhyaya M, Weiffenbach B, Hyser C, Mathews K, et al. Regional mapping of facioscapulohumeral muscular dystrophy gene on 4q35: combined analysis of an international consortium. Am J Hum Genet 1992;51(2):396–403. [106] Ottaviani A, Schluth-Bolard C, Rival-Gervier S, Boussouar A, Rondier D, Foerster AM, et al. Identification of a perinuclear positioning element in human subtelomeres that requires A-type lamins and CTCF. EMBO J 2009;28(16):2428–36.

89

[107] Watrin E, Kaiser FJ, Wendt KS. Gene regulation and chromatin organization: relevance of cohesin mutations to human disease. Curr Opin Genet Dev 2016;37:59–66. [108] Cau P, Navarro C, Harhouri K, Roll P, Sigaudy S, Kaspi E, et al. Nuclear matrix, nuclear envelope and premature aging syndromes in a translational research perspective. Semin Cell Dev Biol 2014;29:125–47. [109] Kleinjan DJ, van Heyningen V. Position effect in human genetic disease. Hum Mol Genet 1998;7(10):1611–8. [110] Rabbitts TH. Chromosomal translocations in human cancer. Nature 1994;372(6502):143–9. [111] Rabbitts TH, Translocations. master genes, and differences between the origins of acute and chronic leukemias. Cell 1991;67(4):641–4. [112] Korsmeyer SJ. Chromosomal translocations in lymphoid malignancies reveal novel proto-oncogenes. Annu Rev Immunol 1992;10:785–807. [113] Lossos IS, Akasaka T, Martinez-Climent JA, Siebert R, Levy R. The BCL6 gene in B-cell lymphomas with 3q27 translocations is expressed mainly from the rearranged allele irrespective of the partner gene. Leukemia 2003;17(7):1390–7. [114] Kumar-Sinha C, Tomlins SA, Chinnaiyan AM. Recurrent gene fusions in prostate cancer. Nat Rev Cancer 2008;8(7):497–511. [115] Northcott PA, Lee C, Zichner T, Stutz AM, Erkek S, Kawauchi D, et al. Enhancer hijacking activates GFI1 family oncogenes in medulloblastoma. Nature 2014;511(7510):428–34. [116] State MW, Greally JM, Cuker A, Bowers PN, Henegariu O, Morgan TM, et al. Epigenetic abnormalities associated with a chromosome 18(q21-q22) inversion and a Gilles de la Tourette syndrome phenotype. Proc Natl Acad Sci USA 2003;100(8):4684–9. [117] Wieacker P, Apeshiotis N, Jakubiczka S, Volleth M, Wieland I. Familial translocation t(1;9) associated with macromastia: molecular cloning of the breakpoints. Sex Dev 2007;1(1):35–41. [118] Kleinjan DA, Seawright A, Schedl A, Quinlan RA, Danes S, van Heyningen V. Aniridia-associated translocations, DNase hypersensitivity, sequence comparison and transgenic analysis redefine the functional domain of PAX6. Hum Mol Genet 2001;10(19):2049–59. [119] de Chadarevian JP, Dunn S, Malatack JJ, Ganguly A, Blecker U, Punnett HH. Chromosome rearrangement with no apparent gene mutation in familial adenomatous polyposis and hepatocellular neoplasia. Pediatr Dev Pathol 2002;5(1):69–75. [120] Crackower MA, Scherer SW, Rommens JM, Hui CC, Poorkaj P, Soder S, et al. Characterization of the split hand/split foot malformation locus SHFM1 at 7q21.3-q22.1 and analysis of a candidate gene for its expression during limb development. Hum Mol Genet 1996;5(5):571–9. [121] Wang P, Carrion P, Qiao Y, Tyson C, Hrynchak M, Calli K, et al. Genotype-phenotype analysis of 18q12.1-q12.2 copy number variation in autism. Eur J Med Genet 2013;56(8):420–5. [122] Jamsheer A, Sowinska A, Simon D, Jamsheer-Bratkowska M, Trzeciak T, Latos-Bielenska A. Bilateral radial agenesis with absent thumbs, complex heart defect, short stature, and facial dysmorphism in a patient with pure distal microduplication of 5q35.2-5q35.3. BMC Med Genet 2013;14:13. [123] Akasaka K, Nishimura A, Takata K, Mitsunaga K, Mibuka F, Ueda H, et al. Upstream element of the sea urchin arylsulfatase gene serves as an insulator. Cell Mol Biol 1999;45(5):555–65. [124] Fang J, Dagenais SL, Erickson RP, Arlt MF, Glynn MW, Gorski JL, et al. Mutations in FOXC2 (MFH-1), a forkhead family transcription factor, are responsible for the hereditary lymphedemadistichiasis syndrome. Am J Hum Genet 2000;67(6):1382–8. [125] Allou L, Lambert L, Amsallem D, Bieth E, Edery P, Destree A, et al. 14q12 and severe Rett-like phenotypes: new clinical insights and physical mapping of FOXG1-regulatory elements. Eur J Hum Genet 2012;20(12):1216–23.

II.  Molecular Mechanisms of Epigenetics

90

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

[126] Crisponi L, Deiana M, Loi A, Chiappe F, Uda M, Amati P, et al. The putative forkhead transcription factor FOXL2 is mutated in blepharophimosis/ptosis/epicanthus inversus syndrome. Nat Genet 2001;27(2):159–66. [127] Beysen D, Raes J, Leroy BP, Lucassen A, Yates JR, Clayton-Smith J, et al. Deletions involving long-range conserved nongenic sequences upstream and downstream of FOXL2 as a novel diseasecausing mechanism in blepharophimosis syndrome. Am J Hum Genet 2005;77(2):205–18. [128] Kosho T, Sakazume S, Kawame H, Wakui K, Wada T, Okoshi Y, et al. De-novo balanced translocation between 7q31 and 10p14 in a girl with central precocious puberty, moderate mental retardation, and severe speech impairment. Clin Dysmorphol 2008;17(1):31–4. [129] Vortkamp A, Gessler M, Grzeschik KH. GLI3 zinc-finger gene interrupted by translocations in Greig syndrome families. Nature 1991;352(6335):539–40. [130] Barbour VM, Tufarelli C, Sharpe JA, Smith ZE, Ayyub H, Heinlein CA, et al. alpha-thalassemia resulting from a negative chromosomal position effect. Blood 2000;96(3):800–7. [131] Kioussis D, Vanin E, deLange T, Flavell RA, Grosveld FG. Betaglobin gene inactivation by DNA translocation in gamma betathalassaemia. Nature 1983;306(5944):662–6. [132] Driscoll MC, Dobkin CS, Alter BP. Gamma delta beta-thalassemia due to a de novo mutation deleting the 5′ beta-globin gene activation-region hypersensitive sites. Proc Natl Acad Sci USA 1989;86(19):7470–4. [133] Yue Y, Farcas R, Thiel G, Bommer C, Grossmann B, Galetzka D, et al. De novo t(12;17)(p13.3;q21.3) translocation with a breakpoint near the 5′ end of the HOXB gene cluster in a patient with developmental delay and skeletal malformations. Eur J Hum Genet 2007;15(5):570–7. [134] Spitz F, Montavon T, Monso-Hinard C, Morris M, Ventruto ML, Antonarakis S, et al. A t(2;8) balanced translocation with breakpoints near the human HOXD complex causes mesomelic dysplasia and vertebral defects. Genomics 2002;79(4):493–8. [135] Dlugaszewska B, Silahtaroglu A, Menzel C, Kubart S, Cohen M, Mundlos S, et al. Breakpoints around the HOXD cluster result in various limb malformations. J Med Genet 2006;43(2):111–8. [136] Enattah NS, Sahi T, Savilahti E, Terwilliger JD, Peltonen L, Jarvela I. Identification of a variant associated with adult-type hypolactasia. Nat Genet 2002;30(2):233–7. [137] Giorgio E, Robyr D, Spielmann M, Ferrero E, Di Gregorio E, Imperiale D, et al. A large genomic deletion leads to enhancer adoption by the lamin B1 gene: a second path to autosomal dominant adult-onset demyelinating leukodystrophy (ADLD). Hum Mol Genet 2015;24(11):3143–54. [138] Brussino A, Vaula G, Cagnoli C, Panza E, Seri M, Di Gregorio E, et al. A family with autosomal dominant leukodystrophy linked to 5q23.2-q23.3 without lamin B1 mutations. Eur J Neurol 2010;17(4):541–9. [139] Jamieson RV, Perveen R, Kerr B, Carette M, Yardley J, Heon E, et al. Domain disruption and mutation of the bZIP transcription factor, MAF, associated with cataract, ocular anterior segment dysgenesis and coloboma. Hum Mol Genet 2002;11(1):33–42. [140] Fantes J, Redeker B, Breen M, Boyle S, Brown J, Fletcher J, et al. Aniridia-associated cytogenetic rearrangements suggest that a position effect may cause the mutant phenotype. Hum Mol Genet 1995;4(3):415–22. [141] Flomen RH, Gorman PA, Vatcheva R, Groet J, Barisic I, Ligutic I, et al. Rieger syndrome locus: a new reciprocal translocation t(4;12)(q25;q15) and a deletion del(4)(q25q27) both break between markers D4S2945 and D4S193. J Med Genet 1997;34(3):191–5. [142] Lee JA, Madrid RE, Sperle K, Ritterson CM, Hobson GM, Garbern J, et al. Spastic paraplegia type 2 associated with axonal neuropathy and apparent PLP1 position effect. Ann Neurol 2006;59(2):398–403.

[143] Muncke N, Wogatzky BS, Breuning M, Sistermans EA, Endris V, Ross M, et al. Position effect on PLP1 may cause a subset of Pelizaeus-Merzbacher disease symptoms. J Med Genet 2004;41(12):e121. [144] de Kok YJ, Vossenaar ER, Cremers CW, Dahl N, Laporte J, Hu LJ, et al. Identification of a hot spot for microdeletions in patients with X-linked deafness type 3 (DFN3) 900 kb proximal to the DFN3 gene POU3F4. Hum Mol Genet 1996;5(9):1229–35. [145] Castermans D, Vermeesch JR, Fryns JP, Steyaert JG, Van de Ven WJ, Creemers JW, et al. Identification and characterization of the TRIP8 and REEP3 genes on chromosome 10q21.3 as novel candidate genes for autism. Eur J Hum Genet 2007;15(4):422–31. [146] Fernandez BA, Siegel-Bartelt J, Herbrick JA, Teshima I, Scherer SW. Holoprosencephaly and cleidocranial dysplasia in a patient due to two position-effect mutations: case report and review of the literature. Clin Genet 2005;68(4):349–59. [147] Marlin S, Blanchard S, Slim R, Lacombe D, Denoyelle F, Alessandri JL, et al. Townes-Brocks syndrome: detection of a SALL1 mutation hot spot and evidence for a position effect in one patient. Hum Mutat 1999;14(5):377–86. [148] Ishikawa-Brush Y, Powell JF, Bolton P, Miller AP, Francis F, Willard HF, et al. Autism and multiple exostoses associated with an X;8 translocation occurring within the GRPR gene and 3’ to the SDC2 gene. Hum Mol Genet 1997;6(8):1241–50. [149] Roessler E, Ward DE, Gaudenz K, Belloni E, Scherer SW, Donnai D, et al. Cytogenetic rearrangements involving the loss of the Sonic Hedgehog gene at 7q36 cause holoprosencephaly. Hum Genet 1997;100(2):172–81. [150] Lettice LA, Heaney SJ, Purdie LA, Li L, de Beer P, Oostra BA, et al. A long-range Shh enhancer regulates expression in the developing limb and fin and is associated with preaxial polydactyly. Hum Mol Genet 2003;12(14):1725–35. [151] Benito-Sanz S, Thomas NS, Huber C, Gorbenko del Blanco D, Aza-Carmona M, Crolla JA, et al. A novel class of Pseudoautosomal region 1 deletions downstream of SHOX is associated with Leri–Weill dyschondrosteosis. Am J Hum Genet 2005;77(4):533–44. [152] Sabherwal N, Bangs F, Roth R, Weiss B, Jantz K, Tiecke E, et al. Long-range conserved non-coding SHOX sequences regulate expression in developing chicken limb and are associated with short stature phenotypes in human patients. Hum Mol Genet 2007;16(2):210–22. [153] Wallis DE, Roessler E, Hehr U, Nanni L, Wiltshire T, RichieriCosta A, et al. Mutations in the homeodomain of the human SIX3 gene cause holoprosencephaly. Nat Genet 1999;22(2):196–8. [154] Balemans W, Patel N, Ebeling M, Van Hul E, Wuyts W, Lacza C, et al. Identification of a 52 kb deletion downstream of the SOST gene in patients with van Buchem disease. J Med Genet 2002;39(2):91–7. [155] Bowl MR, Nesbit MA, Harding B, Levy E, Jefferson A, Volpi E, et al. An interstitial deletion-insertion involving chromosomes 2p25.3 and Xq27.1, near SOX3, causes X-linked recessive hypoparathyroidism. J Clin Invest 2005;115(10):2822–31. [156] Flottmann R, Wagner J, Kobus K, Curry CJ, Savarirayan R, Nishimura G, et al. Microdeletions on 6p22.3 are associated with mesomelic dysplasia Savarirayan type. J Med Genet 2015;52(7):476–83. [157] Giraudeau F, Aubert D, Young I, Horsley S, Knight S, Kearney L, et al. Molecular-cytogenetic detection of a deletion of 1p36.3. J Med Genet 1997;34(4):314–7. [158] Kleefstra T, Brunner HG, Amiel J, Oudakker AR, Nillesen WM, Magee A, et al. Loss-of-function mutations in euchromatin histone methyl transferase 1 (EHMT1) cause the 9q34 subtelomeric deletion syndrome. Am J Hum Genet 2006;79(2):370–7. [159] Velagaleti GV, Bien-Willner GA, Northup JK, Lockhart LH, Hawkins JC, Jalal SM, et al. Position effects due to chromosome

II.  Molecular Mechanisms of Epigenetics



[160]

[161]

[162]

[163]

[164]

[165]

[166]

[167]

[168]

[169]

[170]

[171]

[172]

[173]

[174]

[175]

91

REFERENCES

breakpoints that map approximately 900 Kb upstream and approximately 1.3 Mb downstream of SOX9 in two patients with campomelic dysplasia. Am J Hum Genet 2005;76(4):652–62. Benko S, Fantes JA, Amiel J, Kleinjan DJ, Thomas S, Ramsay J, et al. Highly conserved non-coding elements on either side of SOX9 associated with Pierre Robin sequence. Nat Genet 2009;41(3):359–64. McElreavey K, Vilain E, Barbaux S, Fuqua JS, Fechner PY, Souleyreau N, et al. Loss of sequences 3′ to the testis-determining gene, SRY, including the Y pseudoautosomal boundary associated with partial testicular determination. Proc Natl Acad Sci USA 1996;93(16):8590–4. David D, Cardoso J, Marques B, Marques R, Silva ED, Santos H, et al. Molecular characterization of a familial translocation implicates disruption of HDAC9 and possible position effect on TGFbeta2 in the pathogenesis of Peters anomaly. Genomics 2003;81(5):489–503. Fantauzzo KA, Tadin-Strapps M, You Y, Mentzer SE, Baumeister FA, Cianfarani S, et al. A position effect on TRPS1 is associated with Ambras syndrome in humans and the Koala phenotype in mice. Hum Mol Genet 2008;17(22):3539–51. Zhang X, Snijders A, Segraves R, Niebuhr A, Albertson D, Yang H, et al. High-resolution mapping of genotype-phenotype relationships in cri du chat syndrome using array comparative genomic hybridization. Am J Hum Genet 2005;76(2):312–26. Lupianez DG, Kraft K, Heinrich V, Krawitz P, Brancati F, Klopocki E, et al. Disruptions of topological chromatin domains cause pathogenic rewiring of gene–enhancer interactions. Cell 2015;161(5):1012–25. Goidts V, Cooper DN, Armengol L, Schempp W, Conroy J, Estivill X, et al. Complex patterns of copy number variation at sites of segmental duplications: an important category of structural variation in the human genome. Hum Genet 2006;120(2):270–84. Aldred PM, Hollox EJ, Armour JA. Copy number polymorphism and expression level variation of the human alpha-defensin genes DEFA1 and DEFA3. Hum Mol Genet 2005;14(14):2045–52. Stranger BE, Forrest MS, Dunning M, Ingle CE, Beazley C, Thorne N, et al. Relative impact of nucleotide and copy number variation on gene expression phenotypes. Science 2007;315(5813):848–53. Botuyan MV, Lee J, Ward IM, Kim JE, Thompson JR, Chen J, et al. Structural basis for the methylation state-specific recognition of histone H4-K20 by 53BP1 and Crb2 in DNA repair. Cell 2006;127(7):1361–73. McCarroll SA, Hadnott TN, Perry GH, Sabeti PC, Zody MC, Barrett JC, et al. Common deletion polymorphisms in the human genome. Nat Genet 2006;38(1):86–92. Foster JW, Dominguez-Steglich MA, Guioli S, Kwok C, Weller PA, Stevanovic M, et al. Campomelic dysplasia and autosomal sex reversal caused by mutations in an SRY-related gene. Nature 1994;372(6506):525–30. Pfeifer D, Kist R, Dewar K, Devon K, Lander ES, Birren B, et al. Campomelic dysplasia translocation breakpoints are scattered over 1 Mb proximal to SOX9: evidence for an extended control region. Am J Hum Genet 1999;65(1):111–24. Pop R, Conz C, Lindenberg KS, Blesson S, Schmalenberger B, Briault S, et al. Screening of the 1 Mb SOX9 5′ control region by array CGH identifies a large deletion in a case of campomelic dysplasia with XY sex reversal. J Med Genet 2004;41(4):e47. Wagner T, Wirth J, Meyer J, Zabel B, Held M, Zimmer J, et al. Autosomal sex reversal and campomelic dysplasia are caused by mutations in and around the SRY-related gene SOX9. Cell 1994;79(6):1111–20. Fonseca AC, Bonaldi A, Bertola DR, Kim CA, Otto PA, ViannaMorgante AM. The clinical impact of chromosomal rearrangements with breakpoints upstream of the SOX9 gene: two novel

[176]

[177]

[178]

[179]

[180]

[181]

[182]

[183]

[184]

[185] [186]

[187]

[188]

[189]

[190]

[191]

[192]

de novo balanced translocations associated with acampomelic campomelic dysplasia. BMC Med Genet 2013;14:50. Scherer SW, Poorkaj P, Allen T, Kim J, Geshuri D, Nunes M, et al. Fine mapping of the autosomal dominant split hand/split foot locus on chromosome 7, band q21.3-q22.1. Am J Hum Genet 1994;55(1):12–20. Ignatius J, Knuutila S, Scherer SW, Trask B, Kere J. Split hand/ split foot malformation, deafness, and mental retardation with a complex cytogenetic rearrangement involving 7q21.3. J Med Genet 1996;33(6):507–10. van Silfhout AT, van den Akker PC, Dijkhuizen T, Verheij JB, Olderode-Berends MJ, Kok K, et al. Split hand/foot malformation due to chromosome 7q aberrations (SHFM1): additional support for functional haploinsufficiency as the causative mechanism. Eur J Hum Genet 2009;17(11):1432–8. Scherer SW, Cheung J, MacDonald JR, Osborne LR, Nakabayashi K, Herbrick JA, et al. Human chromosome 7: DNA sequence and biology. Science 2003;300(5620):767–72. Birnbaum RY, Everman DB, Murphy KK, Gurrieri F, Schwartz CE, Ahituv N. Functional characterization of tissue-specific enhancers in the DLX5/6 locus. Hum Mol Genet 2012;21(22):4930–98. Birnbaum RY, Clowney EJ, Agamy O, Kim MJ, Zhao J, Yamanaka T, et al. Coding exons function as tissue-specific enhancers of nearby genes. Genome Res 2012;22(6):1059–68. Kouwenhoven EN, van Heeringen SJ, Tena JJ, Oti M, Dutilh BE, Alonso ME, et al. Genome-wide profiling of p63 DNA-binding sites identifies an element that regulates gene expression during limb development in the 7q21 SHFM1 locus. PLoS Genet 2010;6(8):e1001065. Lauderdale JD, Wilensky JS, Oliver ER, Walton DS, Glaser T. 3′ Deletions cause aniridia by preventing PAX6 gene expression. Proc Natl Acad Sci USA 2000;97(25):13755–9. Inoue K, Osaka H, Thurston VC, Clarke JT, Yoneyama A, Rosenbarker L, et al. Genomic rearrangements resulting in PLP1 deletion occur by nonhomologous end joining and cause different dysmyelinating phenotypes in males and females. Am J Hum Genet 2002;71(4):838–53. Nageshwaran S, Festenstein R. Epigenetics and triplet-repeat neurological diseases. Front Neurol 2015;6:262. Oberle I, Rousseau F, Heitz D, Kretz C, Devys D, Hanauer A, et al. Instability of a 550-base pair DNA segment and abnormal methylation in fragile X syndrome. Science 1991;252(5010):1097–102. Coffee B, Zhang F, Ceman S, Warren ST, Reines D. Histone modifications depict an aberrantly heterochromatinized FMR1 gene in fragile X syndrome. Am J Hum Genet 2002;71(4):923–32. Coffee B, Zhang F, Warren ST, Reines D. Acetylated histones are associated with FMR1 in normal but not fragile X-syndrome cells. Nat Genet 1999;22(1):98–101. Chutake YK, Lam CC, Costello WN, Anderson MP, Bidichandani SI. Reversal of epigenetic promoter silencing in Friedreich ataxia by a class I histone deacetylase inhibitor. Nucleic Acids Res 2016;gkw107. Sakamoto N, Ohshima K, Montermini L, Pandolfo M, Wells RD. Sticky DNA, a self-associated complex formed at long GAA*TTC repeats in intron 1 of the frataxin gene, inhibits transcription. J Biol Chem 2001;276(29):27171–7. Al-Mahdawi S, Pinto RM, Ismail O, Varshney D, Lymperi S, Sandi C, et al. The Friedreich ataxia GAA repeat expansion mutation induces comparable epigenetic changes in human and transgenic mouse brain and heart tissues. Hum Mol Genet 2008;17(5):735–46. Greene E, Mahishi L, Entezam A, Kumari D, Usdin K. Repeatinduced epigenetic changes in intron 1 of the frataxin gene and its consequences in Friedreich ataxia. Nucleic Acids Res 2007;35(10):3383–90.

II.  Molecular Mechanisms of Epigenetics

92

6.  HIGHER-ORDER CHROMATIN ORGANIZATION IN DISEASES

[193] Saveliev A, Everett C, Sharpe T, Webster Z, Festenstein R. DNA triplet repeats mediate heterochromatin-protein-1-sensitive variegated gene silencing. Nature 2003;422(6934):909–13. [194] Ibn-Salem J, Kohler S, Love MI, Chung HR, Huang N, Hurles ME, et al. Deletions of chromosomal regulatory boundaries are associated with congenital disease. Genome Biol 2014;15(9):423. [195] Spielmann M, Brancati F, Krawitz PM, Robinson PN, Ibrahim DM, Franke M, et al. Homeotic arm-to-leg transformation associated with genomic rearrangements at the PITX1 locus. Am J Hum Genet 2012;91(4):629–35. [196] Merla G, Howald C, Henrichsen CN, Lyle R, Wyss C, Zabot MT, et al. Submicroscopic deletion in patients with Williams-Beuren syndrome influences expression levels of the nonhemizygous flanking genes. Am J Hum Genet 2006;79(2):332–41. [197] Chieffo C, Garvey N, Gong W, Roe B, Zhang G, Silver L, et al. Isolation and characterization of a gene from the DiGeorge chromosomal region homologous to the mouse Tbx1 gene. Genomics 1997;43(3):267–77. [198] Arnoult N, Van Beneden A, Decottignies A. Telomere length regulates TERRA levels through increased trimethylation of telomeric H3K9 and HP1alpha. Nat Struct Mol Biol 2012;19(9):948–56. [199] Azzalin CM, Reichenbach P, Khoriauli L, Giulotto E, Lingner J. Telomeric repeat containing RNA and RNA surveillance factors at mammalian chromosome ends. Science 2007;318(5851):798–801. [200] Smirnova A, Gamba R, Khoriauli L, Vitelli V, Nergadze SG, Giulotto E. TERRA expression levels do not correlate with telomere length and radiation sensitivity in human cancer cell lines. Front Oncol 2013;3:115. [201] Schoeftner S, Blasco MA. Developmentally regulated transcription of mammalian telomeres by DNA-dependent RNA polymerase II. Nat Cell Biol 2008;10(2):228–36. [202] Shao L, Shaw CA, Lu XY, Sahoo T, Bacino CA, Lalani SR, et al. Identification of chromosome abnormalities in subtelomeric regions by microarray analysis: a study of 5,380 cases. Am J Med Genet A 2008;146A(17):2242–51. [203] Ballif BC, Sulpizio SG, Lloyd RM, Minier SL, Theisen A, Bejjani BA, et al. The clinical utility of enhanced subtelomeric coverage in array CGH. Am J Med Genet A 2007;143A(16):1850–7. [204] Ledbetter DH, Martin CL. Cryptic telomere imbalance: a 15-year update. Am J Med Genet C 2007;145C(4):327–34. [205] Barber JC. Terminal 3p deletions: phenotypic variability, chromosomal non-penetrance, or gene modification? Am J Med Genet A 2008;146A(14):1899–901. [206] Flint J, Wilkie AO, Buckle VJ, Winter RM, Holland AJ, McDermid HE. The detection of subtelomeric chromosomal rearrangements in idiopathic mental retardation. Nat Genet 1995;9(2):132–40. [207] Horsley SW, Knight SJ, Nixon J, Huson S, Fitchett M, Boone RA, et al. Del(18p) shown to be a cryptic translocation using a multiprobe FISH assay for subtelomeric chromosome rearrangements. J Med Genet 1998;35(9):722–6. [208] Lamb J, Harris PC, Wilkie AO, Wood WG, Dauwerse JG, Higgs DR. De novo truncation of chromosome 16p and healing with (TTAGGG)n in the alpha-thalassemia/mental retardation syndrome (ATR-16). Am J Hum Genet 1993;52(4):668–76.

[209] Walter S, Sandig K, Hinkel GK, Mitulla B, Ounap K, Sims G, et al. Subtelomere FISH in 50 children with mental retardation and minor anomalies, identified by a checklist, detects 10 rearrangements including a de novo balanced translocation of chromosomes 17p13.3 and 20q13.33. Am J Med Genet A 2004;128A(4):364–73. [210] Bonaglia MC, Giorda R, Mani E, Aceti G, Anderlid BM, Baroncini A, et al. Identification of a recurrent breakpoint within the SHANK3 gene in the 22q13.3 deletion syndrome. J Med Genet 2006;43(10):822–8. [211] Wilson HL, Wong AC, Shaw SR, Tse WY, Stapleton GA, Phelan MC, et al. Molecular characterisation of the 22q13 deletion syndrome supports the role of haploinsufficiency of SHANK3/ PROSAP2 in the major neurological symptoms. J Med Genet 2003;40(8):575–84. [212] Gajecka M, Mackay KL, Shaffer LG. Monosomy 1p36 deletion syndrome. Am J Med Genet C 2007;145C(4):346–56. [213] Kosztolanyi G. Does “ring syndrome” exist? An analysis of 207 case reports on patients with a ring autosome. Hum Genet 1987;75(2):174–9. [214] Cote GB, Katsantoni A, Deligeorgis D. The cytogenetic and clinical implications of a ring chromosome 2. Ann Genet 1981;24(4):231–5. [215] Pezzolo A, Gimelli G, Cohen A, Lavaggetto A, Romano C, Fogu G, et al. Presence of telomeric and subtelomeric sequences at the fusion points of ring chromosomes indicates that the ring syndrome is caused by ring instability. Hum Genet 1993;92(1):23–7. [216] Sigurdardottir S, Goodman BK, Rutberg J, Thomas GH, Jabs EW, Geraghty MT. Clinical, cytogenetic, and fluorescence in situ hybridization findings in two cases of “complete ring” syndrome. Am J Med Genet 1999;87(5):384–90. [217] Vermeesch JR, Baten E, Fryns JP, Devriendt K. Ring syndrome caused by ring chromosome 7 without loss of subtelomeric sequences. Clin Genet 2002;62(5):415–7. [218] Surace C, Berardinelli F, Masotti A, Roberti MC, Da Sacco L, D’Elia G, et al. Telomere shortening and telomere position effect in mild ring 17 syndrome. Epigenetics Chromatin 2014;7(1):1. [219] Stadler G, Rahimov F, King OD, Chen JC, Robin JD, Wagner KR, et al. Telomere position effect regulates DUX4 in human facioscapulohumeral muscular dystrophy. Nat Struct Mol Biol 2013;20(6):671–8. [220] Robin JD, Ludlow AT, Batten K, Gaillard MC, Stadler G, Magdinier F, et al. SORBS2 transcription is activated by telomere position effect-over long distance upon telomere shortening in muscle cells from patients with facioscapulohumeral dystrophy. Genome Res 2015;25(12):1781–90. [221] Arnoult N, Schluth-Bolard C, Letessier A, Drascovic I, BouarichBourimi R, Campisi J, et al. Replication timing of human telomeres is chromosome arm-specific, influenced by subtelomeric structures and connected to nuclear localization. PLoS Genet 2010;6(4):e1000920.

II.  Molecular Mechanisms of Epigenetics

C H A P T E R

7 Polycomb Mechanisms and Epigenetic Control of Gene Activity Vincenzo Pirrotta Rutgers University, Piscataway, NJ, United States

O U T L I N E Introduction

93

Higher Order Interactions

100

The Hardware PRC1 PRC2 Polycomb Targets

94 94 94 95

Epigenetic Maintenance or Cellular Memory

102

Maintenance of the Nonrepressed State

102

The Bivalent State

103

Association of PRC2 and PRC1 Complexes

98

Differentiation

104

Mechanisms of PcG Repression

99

Conclusions

104

PRC2 Global Activities

99

References

106

PRC2 Modulation and Product Feedback Effects

100

INTRODUCTION

PcG complexes are found in the nuclei of most cells but the mechanisms by which they are targeted, which genes they affect, how they repress expression, and how the repressed state is remembered from one cell cycle to the next remain important research questions. The molecular analysis of Polycomb mechanisms can be said to have begun with the isolation of the Drosophila Polycomb gene in 1991 [1]. It is sobering to realize that, after more than 25 years the answers to some fundamental questions remain unclear: how are PcG complexes recruited to target genes? How do they repress? In this article, I have not tried to cover all aspects of PcG mechanisms, rather, I have focused on presenting the basics and giving an overview of the current understanding of the role played by these mechanisms, what advances have been made and what partial answers have been achieved, as well as their limitations. Many of these advances have been made using Drosophila as a model system but many important insights have been reached by comparing mammalian systems

The Polycomb Group (PcG) proteins were first discovered in the study of Drosophila homeotic genes and their regulation. They derive their name from the fact that the first signs of a decrease in PcG function is often a homeotic transformation of posterior legs toward anterior legs, which in male flies, have a characteristic comb-like set of bristles. Studies in flies, mammals, and other metazoans have now shown that they are part of a genome-wide mechanism for controlling genomic priorities that governs the ability of genes to respond to signals, channels differentiation pathways, and ultimately determines the developmental fate and identities of cells. PcG mechanisms do this by targeting most if not all of the key genes in the genome that govern cellular decisions from cell cycle progression to apoptosis and including in particular the effectors controlling the variety of differentiation programs. Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00007-9 Copyright © 2017 Elsevier Inc. All rights reserved.

93

94

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

with flies and other organisms. PcG mechanisms have been the subject of a number of recent reviews that treat many aspects in greater detail and to which the reader is referred for additional information [2–5].

THE HARDWARE The essential PcG machinery consists of two multiprotein complexes, Polycomb Repressive Complex 1 and 2, or PRC1 and PRC2. In recent years, however, the picture has been complicated by the discovery of a large number of variant complexes, particularly in the case of PRC1.

PRC1 The essential core of this complex is a heterodimer of a RING protein with a structurally related partner, a Polycomb Group RING Finger (PCGF) protein. This core has catalytic activity, acting as an E3 ubiquityl transferase that targets histone H2A lysine 119 (H2AK119), producing a monoubiquitylated derivative [6]. In Drosophila, there is one RING protein, the product of the Sce gene, and two PCGF alternatives, Psc and Su(z)2. The canonical PRC1 complex utilizes either Psc or Su(z)2 as partners of RING/Sce. It includes the eponymous Polycomb protein Pc, which contains a chromodomain that interacts specifically with histone H3 trimethylated at lysine 27 (H3K27me3) [7], and one of two closely related Polyhomeotic proteins, Ph-p and Ph-d. The Ph proteins contain a protein–protein interaction domain called sterile alpha motif (SAM) also known as SPM, which, in vitro, mediates the formation of Ph oligomers [8] and which interacts with a similar domain in another PcG protein, Sex Comb On Midleg (Scm), which appears to be more loosely associated with PRC1 [9] This combination of components constitutes a canonical PRC1 complex, whose key ability to recognize H3K27me3 links it to the action of the PRC2 complex. A noncanonical Drosophila PRC1-like complex called dRAF was reported to include RING/Sce and Psc, as well as the H3K36 demethylase KDM2 and to be responsible for most of the ubiquitylation of histone H2AK118 (corresponding to mammalian H2AK119) [10]. Although probably incorrect in detail, this was the first report of a noncanonical PRC1 complex. More recent work shows that KDM2 does not seem to be required for most of the H2A ubiquitylation or for homeotic gene repression in Drosophila. Instead a noncanonical PRC1 complex containing a third PCGF protein, L(3)73Ah as a RING partner is responsible for much of the H2AK118ub [11]. In mammals there are two RING homologs, RING1 and RING2 (also known as RING1A and RING1B) of which RING2 is the more abundant. Mammals also have six PCGF proteins, each of which specifies additional

combinatorial properties of the complex. PCGF2 and PCGF4, also known as MEL18 and BMI1, respectively, allow the further binding of a chromodomain-containing homolog (one of six CBX proteins: CBX2, CBX4, CBX6, CBX7, and CBX8) and a Polyhomeotic homolog (one of three: PHC1–3) [12]. These complexes represent varieties of a canonical PRC1 complex, so called because it corresponds structurally to the major Drosophila PRC1, which includes the chromodomain-containing Polycomb protein. In fact, however, the RING heterodimer core can alternatively bind the ubiquitin-binding protein RYBP or its close relative YAF2. The RYBP/YAF2 binding is mutually exclusive with the CBX and PHC binding. Various reports have argued that most of the ubiquityl transferase activity is confined to the RYBP/YAF2-­containing complexes and that the canonical PRC1 type of complex has little H2A ubiquitylating activity [12–14]. Other PCGF factors specify the involvement of other components [12,13]. The best-studied combination is the PRC1.1 complex, involving PCGF1, which recruits the binding of BCOR and KDM2B [15,16]. A schematic summary of PRC1 complexes is shown in Figs. 7.1 and 7.2.

PRC2 The catalytic component of this complex is Enhancer of zeste [E(z)], a SET domain histone methyltransferase [17–21]. Alone, however, the catalytic domain is not appropriately folded for enzymatic activity. ­Methyltransferase function requires the other core components, Suppressor of zeste-12 [Su(z)12], Extra Sex Combs (Esc) or its close homolog Escl, and, to a lesser extent, the histone chaperone RbAp48 (also known as RBBP4/7). In vivo, the enzymatic

FIGURE 7.1  Canonical PRC1 complexes. Canonical PRC1 complexes preserve the structure of the Drosophila PRC1 complex shown on the left and include a chromodomain component: Pc in the Drosophila PRC1 and a CBX component in the human PRC1. The components are color-coded to identify homologs and alternative components are indicated. The looser association of the Scm component is indicated by the double-headed red arrow.

II.  Molecular Mechanisms of Epigenetics



The Hardware

FIGURE 7.2  Variant PRC1 complexes. The human PRC1.1 complex is shown in schematic form. In other variant PRC1 complexes, the replacement of PCGF1 by PCGF3, 5, or 6 directs the recruitments of other components, in addition to RYBP/YAF2. The two Drosophila variant PRC1s, dRAF and L(3) are shown on the right. It is not known what role KDM2 plays in dRAF and the composition of the L(3) complex has not been determined.

target of this complex is histone H3K27 and PRC2 is responsible for most mono- and all di- and trimethylation at this position. The mammalian PRC2 is entirely homologous, including the proteins SUZ12, EED (embryonic ectoderm development), and the E(z) homolog EZH2 (or its more specialized homolog EZH1). In both flies and mammals, variants of this complex have been found, containing additional factors or, in mammals, different isoforms of EED that modulate its specificity [22,23]. The additional factors are generally substoichiometric and not essential for catalytic activity, though they may modulate it. The major optional components that have been reported are

95

Adipocyte Enhancer Binding Protein 2 AEBP2 (Drosophila Jing) [24,25], Jumonji AT-Rich Interactive Domain JARID2 [26–29], Drosophila Polycomblike Pcl [30] or its mammalian homologs PHF1, MTF2, PHF19 (also known as PCL1,2,3) [31–34], C17orf96 [35], and histone deacetylase HDAC1 (Rpd3) [17,19]. Notably, the presence of AEBP2 is ­mutually exclusive with the presence of the PCL or C17orf96 cofactors [36,37]. A summary of PRC2 structure is shown in Fig. 7.3. The two mammalian enhancer of zeste genes: Ezh1 and Ezh2 differ widely in their abundance in different tissues and in the catalytic activities of the PRC2 complexes they form. In general, EZH2 is both more abundant and more active than EZH1 but, while its expression depends on the cell cycle-associated factor E2F and declines in senescent or terminally differentiated cells, EZH1 becomes predominant as EZH2 decreases and continues to be expressed in postmitotic cells. Although much less active than EZH2, EZH1 can partly substitute for it, at least in some tissues. PRC2 complexes containing EZH1 bind more strongly to nucleosomes and to chromatin and may, in some cases, help EZH2-PRC2 to be recruited to some target sites, possibly by forming PRC2 dimers [38].

Polycomb Targets A characteristic feature of classical Polycomb repression is that it requires both PRC1 and PRC2 complexes at specific target genes. In fact, however, there are cases known in which one but not the other complex appears to be required. A few loci in Drosophila bind PRC1 but not PRC2 and are not marked by H3K27me3 [39–42], while a few genomic regions have prominent H3K27me3 enrichment but no PRC1 binding. In mammalian cells,

FIGURE 7.3  PRC2 complex. The figure shows the core components of the Drosophila and human PRC2 complex. The additional components indicated further increase stability, modulate enzymatic activity, and may confer target specificity.

II.  Molecular Mechanisms of Epigenetics

96

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

knockdown experiments have shown that, while most PcG target genes require both PRC1 and PRC2 for repression, some require only PRC1 and some only PRC2 [43]. In general, however, genes specifically controlled by PcG mechanisms require a recruiting element that binds PRC1 and PRC2 complexes. Thus, what is needed is a coordinated recruitment of two types of complex, neither of which has sequence-specific components. Furthermore, while PcG proteins are present in most cells, a given target gene may be repressed in one type of cell but not in another type. Therefore, some additional features of the target gene chromatin must specify whether it is to bind PcG complexes or not. To understand how this may be done, we will examine first what appear to be the major modes of PcG recruitment in flies and in mammals. Drosophila Polycomb Response Elements Sequence elements able to autonomously recruit PcG complexes and establish repressive states were first identified in Drosophila by genetic and functional analyses [40,44,45]. These Polycomb Response Elements (PREs) may be found close to the promoter of the target gene, but are often ten or more kilobases distant, upstream, within introns, or downstream of the polyadenylation site. Genes with complex regulatory regions may possess more than one PRE. An extreme case is the Hox gene Abd-A, which is regulated by at least six PREs separated by insulator elements and enhancers [42,46]. PREs are of variable length ranging from a few tens to a few hundred nucleotides whose sequence is not conserved but generally contains numerous consensus binding motifs for certain DNAbinding proteins (Fig. 7.4). Most commonly, PRE-binding proteins include GAGA binding factor GAF/Trl [47], Pleiohomeotic Pho [48], Dorsal Switch Protein 1 Dsp1 [49], Sp1-like Factor for Pairing-sensitive Silencing Spps [50], ADH Transcription Factor 1 Adf1 [51], Combgap Cg [52]. However, none of these proteins are specific for PREs: they are found at many other sites in the genome

FIGURE 7.4  Polycomb response elements. The figure represents the bxd PRE of the Drosophila Ubx gene. PREs are generally nucleosomedepleted, contain DNase hypersensitive sites and consensus sequences for several DNA-binding factors, two of which, GAF and the PhoRC complex, are shown. These factors are thought to act cooperatively to recruit PRC1 and PRC2.

with no relationship to PcG proteins. In addition, no one of these proteins is obligatorily found at all PREs. Thus, the distribution suggests that these proteins cooperate to recruit PcG complexes and that different combinations may be effective. Of these, the Pleiohomeotic (Pho) protein or its close homolog Phol, is particularly interesting. Pho was originally identified as a PcG protein, whose absence causes derepression of homeotic genes, the only PcG protein with sequence-specific DNA-binding ability [53,54]. Pho/Phol are homologs of the mammalian YinYang1(YY1) protein, a transcription factor that has both repressive and activating features. In Drosophila, Pho is found in a Pho repressive complex PhoRC that includes SFMBT (Scm-related containing Four MBT domains), a chromatin protein whose MBT domains are able to bind to mono- and dimethylated lysines [55,56]. More significant, however, is that SFMBT contains a zinc finger motif that binds to Scm, whose SAM (Sterile Alpha Motif) domain in turn interacts with the SAM domain of the Ph component of PRC1 [57]. Consistent with this, the binding of PRC1 and of PhoRC is interdependent and each stabilizes the other [58,59]. In mammals, despite the conservation of Scm and related MBT-containing proteins, and YY1, the interrelationship of YY1 and PRC1 binding is not conserved and there is little overlap between their genomic binding sites. Paradoxically, perhaps, but consistent with the need to allow access to multiple DNAbinding proteins, PREs are generally nucleosome-depleted [60–62] and, like enhancers and promoters, are also prominent sites for the binding of the histone acetylase CREB-binding protein CBP [63]. How PRC2 is recruited to PREs is less clear. A hierarchical recruitment model was proposed by Wang et al. [64], according to which PhoRC initially recruited PRC2, which subsequently produced H3K27me3 in the surrounding region, which in turn recruited ­chromodomain-containing canonical PRC1 complexes. This model was appealing in its simplicity and very influential, despite the fact that PRC1 is generally localized at the PRE, which is usually nucleosome-depleted, hence contains little H3K27me3, while the H3K27me3 domain can extend for tens or even hundreds of kilobases. Instead, recent results show that, in the absence of PRC2, PREs retain most of their ability to recruit PRC1, while in the absence of the canonical PRC1 complex, many but not all PREs lose most of their ability to bind PRC2, thus arguing that PRC2 binding needs help from PRC1 [65]. About one third of the PREs bind PRC2 independently of PRC1, suggesting that somewhat different recruitment mechanisms may be utilized at different sites. Mammalian PcG Recruitment In mammals PRE-like recruiting elements do not seem to be widely utilized. While a few sites in the mouse or human genome have been identified that resemble

II.  Molecular Mechanisms of Epigenetics



The Hardware

­ rosophila PREs in function [66,67], it has become clear D that the majority of PcG binding sites correspond to CpG islands [68–71]. These are short sequence tracts around 1 kb long, highly enriched for the CpG dinucleotide, which is otherwise underrepresented in vertebrate genomes. CpG islands are targeted by many transcription factors, often through a modified zinc finger motif called the CXXC domain, and correspond, in fact, to the promoter regions of most regulated genes [72,73]. A connection between CpG islands and PcG complexes was made by the discovery that KDM2B, a H3K36 histone demethylase that contains a CXXC domain, binds to CpG islands and recruits the PRC1.1 variant complex [16,74–76]. In principle, therefore all CpG-rich promoters are potential targets of PRC1.1 unless its binding is discouraged by specific modifications, such as DNA methylation of the CpG island or by its occupancy by transcription factors. It is not clear to what extent binding by PRC1.1 might be intrinsically repressive but it certainly becomes so when it mediates the recruitment of PRC2. Using a variety of tethering experiments, it was shown that binding of PRC1.1 produces a domain of histone H2A monoubiquitylation at lysine 119 and that this results in recruitment of PRC2 [74,77]. In fact, tethering the catalytic core of the complex suffices to recruit PRC2, while a mutation of the catalytic activity prevents PRC2 recruitment. The implication that H2AK119ub recruits PRC2 was supported by the finding that, in Drosophila and in mammals, the PRC2-associated factors AEBP2 and JARID2 bind to H2AK119ub [78]. The chain of events is then completed when PRC2 generates a domain of H3K27me3, which in turn recruits the canonical, chromodomain-containing PRC1 complex (Fig. 7.5A–C). This attractive model has quickly run into difficulties. Many sites in the mouse genome contain PRC1.1 and H2AK119ub but not H3K27me3 [79]. In mouse intestinal stem cells, knockout of RING1,2 eliminates H2AK119ub but does not affect H3K27me3 [80] and in human HEK293T cells, a human differentiated cell line, targeting of KDM2B produces H2AK119ub but does not recruit PRC2 [75]. In Drosophila, H2AK118 ubiquitylation is widely distributed in the genome and is enriched at many PcG binding sites, where it coincides with H3K27me3 but it is absent at the major homeotic genes of the bithorax and antennapedia complexes [11]. In fact, H2A ubiquitylation is removed at these sites by the action of the PRDUB complex composed of Additional Sex Combs Asx and the deubiquitylase Calypso, h ­ omologous to mammalian ASXL1 and BAP1, respectively [81]. The deubiquitylating action is required for effective repression of these homeotic genes. H2AK118ub is widespread in the Drosophila genome, without a corresponding presence of PRC2 or H3K27me3 [11]. Most conclusive are the reports that mutations of the RING catalytic domain abolishing H2A ubiquitylation in flies and mammals do not prevent

97

FIGURE 7.5  Model for the recruitment of PcG complexes to CpG islands. (A) According to this model [74], the variant complex PRC1.1 binds to CpG islands through the CXXC zinc finger of KDM2B. PRC1.1 has a strong catalytic activity that monoubiquitylates flanking nucleosomes at H2K119. (B) PRC2 is recruited through the affinity of AEBP2 and JARID2 for H2AK119ub. Then PRC2 trimethylates flanking nucleosomes at H3K27. (C) The canonical PRC1 complex, which includes a CBX component, binds to the H3K27me3 through its chromodomain. The model is presented here as a possible example of recruitment in mammals. Conflicting evidence raises questions about how broadly it might be applicable. A definitive general model for recruitment of PRC complexes to mammalian promoters has not been established at this time.

repression of target genes and have only mild effects on H3K27me3 [79,82]. In addition, in Drosophila, clonal patches lacking ubiquitylatable H2A maintain repression of homeotic genes [79]. The significance of this last result is weakened by the fact that the major homeotic genes in Drosophila are specifically lacking in H2A ubiquitylation through the action of the PR-DUB [11,81] and can be normally repressed without H2A ubiquitylation. It remains possible that H2AK119ub-mediated recruitment of PRC2 operates in some circumstances, perhaps in embryonic stem cells but not later in development, or at some target sites but not others.

II.  Molecular Mechanisms of Epigenetics

98

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

It is important to note that PcG proteins are present in most cells but they do not obligatorily bind and repress all potential target genes. Drosophila PREs do not constitutively recruit PcG complexes. In the Drosophila embryo, homeotic genes are activated or repressed in patterns ultimately depending on short-lived maternal cues. PcG complexes bind and repress only those target genes that are not active [83]. In mammalian embryonic stem cells, transcriptionally active CpG islands do not recruit KDM2B or PcG complexes. It appears then that the transcriptional status is an important cue determining the recruitment of PcG complexes. This was dramatically demonstrated by the finding that, in mouse embryonic stem cells, direct block of RNA Pol II is sufficient to recruit PRC2 to CpG island promoters [84]. The conclusion indicated by these results is that PcG mechanisms are targeted to genes that are not transcriptionally ­active. It follows then that, in the general case, genes must be turned off by specific repressors before they can be subject to PcG repression or, in other words, that PcG mechanisms maintain repression that has been initiated by other means. Other Mechanisms of Recruitment PcG complexes can be recruited by different mechanisms in different circumstances or at different potential targets. Various cases have been described in which the recruitment of PcG complexes is instigated or facilitated by protein complexes, such as the inhibitor of neuronal differentiation REST (Fig. 7.6A) or repressors of the SNAIL family [85,86]. In a number of cases, long noncoding RNAs (lncRNAs) have been implicated in the recruitment or targeting of Polycomb repression. For example, lncRNAs have been reported to bind PRC2 and direct its action to HOX genes [87–89]. PcG regulation of the INK4A/ARF locus depends on the lncRNA ANRIL (Fig. 7.6B) transcribed from the locus itself and required for the recruitment of PRC2 to the locus [90]. Recruitment of PRC2 by Xist RNA has been reported to be involved in mammalian X chromosome inactivation [91] although the specificity and mechanism have been debated [92]. Recently, the binding of PRC2 to Xist RNA was reported to be mediated by the JARID2 component of PRC2 [93], which has been shown to contain an RNA-binding domain [94]. It remains unclear, however, how this binding and subsequent targeting of the complex would be regulated to account for the specificity of PcG repression. Finally, PRC2 has differential affinity for certain histone modifications, which can help to target it preferentially to certain chromatin regions. For example, the core PRC2 constituent Esc/EED binds directly to nucleosomes containing H3K27me3 (Fig. 7.6C). Although this affinity is unlikely to result in stable binding, it is essential for preferential targeting of PRC2 activity to previously repressed sites (see further). Other ­chromatin features can

FIGURE 7.6  Alternative recruitment mechanisms. Specific recruitment mechanisms may be used for specific targets. These have mostly been studied for mammalian PRC2 recruitment but similar alternatives are likely to exist for PRC1 recruitment and in Drosophila, as well as in mammals. (A) DNA-binding proteins may interact directly with PRC2 and participate in its recruitment to specific targets: an example is the REST repressor, which is said to both silence target neural genes and bring to them the PRC2 complex. (B) Long noncoding RNAs have been reported to recruit PcG complexes to target genes. In the case shown here, the ANRIL lncRNA binds PRC2 and targets it to the INK4A/ARF locus. (C) Chromatin features, such as H3K27me3 can help to bring PRC1 and/or PRC2 to regions that had previously been PcG-repressed.

prevent or displace PRC2 binding. The presence of H3K36me3, a modification associated with transcriptional elongation, antagonizes PRC2 binding to chromatin. Phosphorylation of H3S28, adjacent to the methylation target H3K27, displaces PRC2 and PRC1 and provides a means of rapidly reactivating a PcG-repressed gene without the need to remove the H3K27 methylation [95]. Remarkably, PRC2 containing EZH1 is not affected by this so-called “phospho-switch” and its continued binding is important to regulate gene expression in myogenic differentiation [96].

ASSOCIATION OF PRC2 AND PRC1 COMPLEXES While PRC1 and PRC2 components bind primarily at the sites to which they are recruited, the domain of H3K27 trimethylation is often much broader in somatic cells. This can be explained by a local chromatin looping that is well known to produce frequent contacts between adjacent chromatin regions [60,97,98], bringing the associated PRC2 complex to bear on a broader chromatin region. Once this region becomes trimethylated, the ­contacts are further enhanced by the interaction of the chromodomain-containing PRC1 with the H3K27me3. Interestingly, when a PcG target gene is in the active state, it generally loses PRC2 binding and H3K27me3 but retains a significant presence of PRC1 at PREs [99].

II.  Molecular Mechanisms of Epigenetics



99

PRC2 Global Activities

It is clear that, at most target genes, the repressed state requires both PRC1 and PRC2 and that loss of either complex derepresses the target genes. A direct interaction between the two complexes has often been suspected. In the early Drosophila embryo, tethering a PRC1 component to a reporter gene can establish repression by recruiting PRC2 and tethering the Esc component of PRC2 can similarly recruit PRC1 [100]. At this stage, the two complexes seem able to coprecipitate but lose this ability at later stages. More recently, additional links between the two complexes have been identified. One report found that the N-terminal region of the PRC2 core component EED but not other PRC2 proteins interacted directly with the RING component of a canonical PRC1 complex in extracts of mouse embryonic stem cells [101]. This region of EED is also the domain that interacts with EZH2, so that binding to RING is incompatible with participation in PRC2. It is not clear what the biological role of this interaction might be or why it has not been detected previously. Another report using rapid affinity purification found that, in Drosophila, Scm, previously known to interact with PRC1, copurified with both PRC1 and PRC2 complexes [102]. Conversely, tagged Scm was able to recover both PRC1 and PRC2 components. Scm is an important contributor to PRC1 recruitment because it binds independently to PREs [103] and helps stabilize interactions between the Ph component and the PhoRC complex, which also binds weakly to PREs [58,59]. Scm might therefore serve as the strong man that helps to hold together PRC1, PhoRC, and PRC2 at the PRE.

MECHANISMS OF PcG REPRESSION Initial views of PcG repression envisioned the effective agent to be PRC1 recruited by the H3K27me3 mark deposited by PRC2. Repression was thought to be due to the aggregation of PRC1 complexes on the target chromatin, which compacted it and rendered the underlying DNA inaccessible to transcription factors and RNA polymerase. This model was originally inspired by then-current views of heterochromatic silencing and was further fueled by reports of in vitro clumping of a chromatin fiber induced by purified PRC1 complexes, as visualized by electron microscopy [104]. Under similar in vitro conditions, purified PRC1 was shown be a powerful inhibitor of the SWI/SNF nucleosome remodeling complex [105]. In vivo, a FISH imaging study of a mammalian Hox cluster showed that it became more extended, looping out of the chromosome territory as the Hox genes became transcriptionally active [106,107]. This was interpreted as a physical decondensation that occurred upon derepression. These conclusions, depending on a chromatin-­ condensing ability of PcG complexes, became increas-

ingly problematic as chromatin immunoprecipitation assays in Drosophila showed that PRC1 binding is preferentially localized not at the promoter but at the PRE [42]. In addition, Drosophila PREs are not inaccessible but rather nucleosome-depleted and DNaseI hypersensitive [108], PcG repression does not prevent access to transcription factors and RNA polymerase though it prevents transcription initiation [109], suggesting that the repressive mechanism interferes with the preinitiation complex. Furthermore, far from being compacted and impervious, PRC1 binding sites are in fact regions of very active nucleosome turnover [110]. A variety of imaging studies found no correlation between Polycomb proteins and condensed chromatin in the nucleus [111–113]. The looping out of mammalian Hox genes upon derepression might be better understood as occurring at the level of looped chromatin domains rather than as a result of decompaction at the level of nucleosomes. Attention then turned to the histone H2A ubiquityl transferase activity or PRC1. It was reported that H2A monoubiquitylation interfered with transcription elongation [114,115]. However, it is clear from genetic knockout experiments that both PRC2 and canonical PRC1 are needed for effective repression at most target genes. This implies that noncanonical PRC1 complexes, while they efficiently ubiquitylate H2A, are not by themselves able to repress CpG island promoters in mammals or PRE-associated genes in Drosophila. More recently, furthermore, it has become clear that H2A ubiquitylation is not required for PcG repression in flies or in mammals [79,82]. A different line was suggested recently by the finding that Pc and its mammalian CBX counterparts interact with the same region of the histone acetylase CBP that is involved in the autoacetylation required to promote acetylase activity [116]. Thus Pc binding and autoacetylation are mutually exclusive. The authors argue that PRC1 would therefore inhibit CBP associated with promoters or enhancers and inhibit their activation. They propose a much more global role for PRC1, or at least for Pc, in the genome in which it would regulate both global and promoterassociated CBP acetylation activity. A global role for PRC1 at both active and silent promoters was also proposed in another study [117]. It is by no means excluded, therefore that PRC1 makes important contributions to transcriptional silencing but the focus is now shifting to PRC2 and H2K27 methylation as the key effectors.

PRC2 GLOBAL ACTIVITIES A clue to the repressive activity of H3K27 methylation might come from the effect of H3K27me2 rather than H3K27me3. Seen globally, in fact, the major activity of PRC2 is the dimethylation of H3K27, which is highly abundant in the genome, representing 50%–70%

II.  Molecular Mechanisms of Epigenetics

100

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

of all histone H3, compared to 5%–10% for H3K27me3 [118–120]. H3K27me2 is essentially enriched in all regions that are weakly transcribed or entirely transcriptionally inactive (except for those that contain H3K27me3) [11,120]. ­Regions enriched for H3K27me2 correspond in fact to the inactive “black chromatin” states described by Filion et al. [121]. In Drosophila, inactivation of PRC2 using a conditional allele of E(z) results in a global increase in transcriptional activity in proportion to the H3K27me2 levels [11]. Thus, transcripts from intergenic regions or previously silent genes increase 50–100-fold but even those from very active genes increase 2–4-fold. This global transcriptional derepression is dependent on the H3K27 demethylase UTX, which is also a component of the CBP acetylase [122] and of the Trr H3K4 monomethylase [123]. PRC2 inactivation results in fact in a global increase in H3K27ac and H3K4me1 [11,120]. These r­esults show that PRC2, as well as CBP and Trr and their accompanying UTX have untargeted, roaming activities and suggest that they act on the entire genome with antagonistic results for the access of DNA-binding factors to the underlying DNA [124,125]. H3K27ac and H3K4me1 are chromatin marks associated with enhancer and promoter activation [126–128] and, in fact, loss of PRC2 function results in inappropriate activation of enhancers in mouse embryonic stem cells [120]. In sum, it appears that H3K27 dimethylation is a global activity that ­suppresses adventitious and pervasive chromatinopening and transcriptional activities and that this ­suppression might simply result from preventing the acetylation of H3K27.

PRC2 MODULATION AND PRODUCT FEEDBACK EFFECTS During chromatin replication, old nucleosomes are disrupted but their core histone H3/H4 tetramer is randomly redistributed to the two daughter DNA molecules (reviewed in ref. [129]). The newly replicated DNA therefore has half the complement of nucleosomes and new nucleosomes must be deposited to fill the gaps. As a result, the histone methyl marks present on the old chromatin become diluted twofold. To maintain the mark, therefore it is essential to restore full H3K27 trimethylation rapidly and efficiently. PRC2 operates stepwise rather than processively and the trimethylation step is known to be much slower than the mono- or dimethylation [130,131]. Several feedback features modulate PRC2 activity and help to steer it to specific chromatin targets. The catalytic activity of PRC2 can be decreased by interactions of the PRC2 component Su(z)12 with chromatin features, such as H3K4 methylation or H3K36 methylation, both marks associated with transcriptional activity [132,133]. This helps to focus PRC2 function on genomic regions

that are not transcriptionally active. On the positive side, methylation activity is stimulated by interaction of the Su(z)12 component of PRC2 with unmethylated H3K36 in the vicinity, with the result that transcriptionally inactive chromatin, which has a higher nucleosome density, is a favored substrate [134]. This study also showed that the higher nucleosome density precedes PcG recruitment, rather than being caused by it. A particularly important effect is a positive product feedback mediated by the Esc/EED subunit of PRC2. This component contains a hydrophobic pocket that binds methylated lysine, with a preference for H3K27me2 and H3K27me3 [135]. This interaction has a powerful stimulatory effect on the catalytic activity so that chromatin that already contains H3K27me2 or H3K27me3 is a preferred methylation target. PRC2-dependent gene repression is lost when the Esc/EED residues involved in this interaction are mutated, showing that this feature is essential for effective maintenance of the repressed chromatin state, probably by accelerating the trimethylation step at sites that were previously trimethylated.

HIGHER ORDER INTERACTIONS The foregoing discussion suggests the possibility that the repressive action might be primarily a function of PRC2 H3K27 dimethylation and that the trimethylation that is characteristic of targeted PcG silencing might simply make the repression stronger, stable, and more effective. It is clear that, at most target genes, the repressed state requires both PRC1 and PRC2 and that loss of either complex derepresses the target genes. There are several possible ways in which the canonical PRC1 complex might make an important contribution to repression. One is by helping to spread the domain of H3K27 trimethylation [60,97]. Local looping is a normal feature of the chromatin fiber that can bring the PRC2 complex bound to the recruiting site in contact with flanking regions. This contact, however, would be too brief to allow the addition of a third methyl group, a slow and inefficient process. An important role of the canonical PRC1 complex would be to make such contacts less transient by the interaction of the Pc chromodomain with the surrounding H3K27me3. This could greatly facilitate the spread of the H3K27me3 domain and accelerate its full reconstitution after the dilution resulting from a round of replication. The repressive action of PcG complexes might be enhanced by interactions between distant PcG target sites in the nucleus. This possibility was first suggested by the fact that, in Drosophila, transgenes containing a PRE often display a pairing effect: the repressive activity of the PRE becomes much stronger when the transgene is made homozygous [44,45]. This is attributed to the homologous pairing of chromosomes that, in Drosophila,

II.  Molecular Mechanisms of Epigenetics



Higher Order Interactions

brings the two copies of the transgene in close proximity during interphase. This pairing effect suggests that the physical proximity of two or more PREs enhances or stabilizes repression. The ability of PREs to act in trans is strikingly illustrated by the fact that if the PRE is deleted from one copy of the transgene, the remaining PRE on the homologous chromosome can repress both copies [136]. The idea that endogenous Polycomb targets might tend to cluster spatially in the nucleus was suggested by the fact that PcG foci visualized in the interphase diploid nucleus of normal cells are variable in size and intensity and fewer than expected from the number of PcG-repressed sites in the genome [137,138]. Although other explanations can be found for this, for example, the fact that PcG targets tend to be clustered in the primary ­genomic sequence, further evidence confirms that PcG target sites in the nucleus can, in fact, contact one another [139–141]. Such clustering of PcG binding sites could result in stronger and stable repression by creating local high concentrations of PcG proteins. Three factors have been reported to modulate the clustering of PcG targets: oligomerization of PRC1 through the Ph component, SUMOylation of the Pc component, and interactions mediated by insulator or architectural proteins. The Ph subunit, which is present in canonical but not in noncanonical PRC1 complexes, contains a SAM domain which has been shown to polymerize head-totail [8]. Mutations that specifically abolish the polymerization of the Ph SAM domain also cause loss of stable binding to and derepression of PcG target genes and loss of the ability to cluster in the nucleus [142–145]. It is important to note, however, that the Ph SAM domain is also known to be involved in interaction with the SAM domain of the Scm protein, which is essential for PRC1 function [9]. Thus, the essential role of the Ph SAM domain might be not to mediate PRC1 oligomerization but to interact with Scm, which is a precondition for stable PRC1 recruitment. Furthermore, Ph oligomerization is held in check by the glycosyl transferase Ogt in flies and mammals [146–148]. Ogt is responsible for the O-linked acetylglucosamine acylation (GlcNAcylation) of a serine/threonine-rich region in the Ph protein, a modification that interferes with the SAM domain-mediated polymerization. Ogt function is essential for effective repression and deletion of the serine/threonine-rich region has the same effect as loss of Ogt, implying that unimpeded polymerization prevents correct function of PRC1. Thus, GlcNAcylation might discourage Ph oligomerization in favor of interaction with Scm. While it is likely that the Scm interaction is the key function of the Ph SAM domain, it cannot be excluded at this point that a limited degree of Ph-mediated oligomerization contributes to PRC1 function. An unbiased search for mutations that alter the distribution of PcG proteins in the nucleus revealed an-

101

other protein modification that affects the clustering of PRC1. Among the most important candidates revealed by this screen were smt3, the gene encoding the Small ­Ubiquitin-like MOdifier SUMO, and velo, encoding an endopeptidase that removes the SUMO modification [149]. Significantly, when the SUMO gene was knocked down, PRC1 formed fewer and larger aggregates in the nucleus. Depletion of Velo caused a decrease in the number and intensity of Pc binding sites determined by ChIP, a more diffused Pc localization in the nucleus and a derepression of homeotic genes. The human Pc2 protein, now known as CBX4, is known to be a target of SUMOylation, as well as itself an E3 SUMO ligase that SUMOylates other proteins [150]. The PcG protein Scm is also known to be a target of SUMOylation, which reduces its binding to chromatin [151,152]. The effective target for the nuclear localization of Pc is the protein itself, as shown by the fact that mutating the lysines affected by SUMOylation mimicks the effect of SUMO loss of function [149]. In conclusion, Pc SUMOylation appears to modulate aggregation of PRC1 in the nucleus; its absence results in excessive aggregation while uncontrolled Pc SUMOylation weakens interaction with chromatin and results in excessive dispersion in the nucleus. A third type of activity that promotes the interaction between PcG target sites in the genome is that of architectural proteins, such as insulator binding fac­ tors and CTCF in particular. Once again, the story starts with observations made with Drosophila transgenes. As mentioned previously, transgenes containing a PRE show much more powerful silencing when homozygous, an effect attributed to the pairing of homologous chromosomes in Drosophila. In tests to determine possible ­ interactions between transgene copies inserted at nonhomologous sites, pairing-like effects were observed in certain combinations but only if the transgenes included an insulator binding site [136]. In some cases, the interacting transgenes were inserted on two different chromosomes and mutation knocking out the insulator binding protein abolished the interaction. Attempts to visualize such interactions in the nucleus using transgenes tagged with tandem arrays of Lac operator sequences and GFP-tagged Lac repressor showed that transgenes containing PREs and insulator elements could indeed colocalize physically [153]. Further work confirmed this colocalization and its dependence on the insulator element [154] rather than on PcG proteins and showed that even stronger insulator-dependent colocalization could result from the ability of PREs to recruit Trithorax when the transgene was in the active state [155]. These interactions are part of the complex networks [156] underlying the architectural organization of the genome into Topologically Associated Domains (TADs) and involving intra- and inter-TAD long-range interactions (for ­review, see ref. [157]).

II.  Molecular Mechanisms of Epigenetics

102

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

EPIGENETIC MAINTENANCE OR CELLULAR MEMORY In any one cell, some of the potential PcG targets bind PcG proteins and are repressed while others are active. What determines which genes will be repressed? The early answer for the Drosophila Hox genes, the prototypical PcG targets, was that if the gene had been active in the early embryo, it remained potentially unrepressed by PcG mechanisms and could function at later stages, provided the appropriate activators were present [83]. If the gene was not active in the early embryo, it bound PcG proteins and became repressed. Once the PcG proteins established repression, this state was remembered in succeeding cell cycles and was reestablished in the cellular progeny. Conversely, if the gene had been active in the early embryo, this activity somehow set a protective barrier that prevented PcG repression at later stages. Genetic evidence showed that the functions responsible for this kind of memory of the active state were associated with the trithorax and ash1 genes (see further). This degree of memory that makes the consequences of early decisions persist throughout development is likely to be specific for Drosophila homeotic genes. Many other developmental genes in flies and mammals need to be turned on later in development and often in a more dynamic fashion. It remains true, however, that genes repressed by PcG mechanisms tend to remain repressed from one cell cycle to the next unless a surge of activators succeeds in overcoming repression and switch the gene to an active mode. The close association of H3K27 trimethylation with PcG repression immediately suggests that this methyl mark might be responsible for the epigenetic memory of the repressed state and this has been widely assumed. In order for such a histone mark to be responsible for the epigenetic memory, it must be able to perform two functions. One is to facilitate or stabilize the binding of PcG complexes and the other is to trigger the self-renewal of the mark every cell cycle. As discussed previously, the methyl mark does appear to stabilize the binding of PRC1 to target genes and, although the mechanism is not entirely clear; it provides an advantage over transcriptional reactivation. It has been argued that the PRC2 complex associates with the replication fork and might recognize and restore the H3K27me3 mark as the ­replication fork transits [158]. This does not agree well with studies using SILAC (Stable Isotope Labeling with Amino acids in Cell culture) and mass spectrometry, which show that restoration of H3K27 trimethylation after DNA replication is very slow and gradual during the cell cycle [131]. Several mechanisms of context regulation of PRC2 activity (see section “PRC2 Modulation and Product Feedback Effects”), as well as the binding of PRC2 to recruiting sites and chromodomain-mediat-

ed local looping would make PcG target genes and regions previously trimethylated preferential targets for remethylation. These mechanisms ensure that domains of H3K27me3 tend to be maintained from one cell cycle to the next and provide a memory of the repressed state. Preexisting H3K27me3 is not sufficient, however, to maintain itself indefinitely in the absence of continued recruitment of PRC2 to a stable binding site. Deletion of a PRE in the course of development, results in eventual derepression [159].

MAINTENANCE OF THE NONREPRESSED STATE What prevents PcG complexes from acting on target genes that should remain in an active mode? Overexpression of an activator trumps PcG repression and can reactivate a repressed gene [160–162]. However, the genetic evidence in Drosophila shows that the maintenance of a program of expression of PcG target genes, requires the trithorax (trx) and ash1 genes. These functions do not activate or derepress target genes but they cooperate with activators to stimulate transcription, antagonize PcG repression, and preserve an epigenetic memory of the derepressed state [163,164]. Interestingly, however, these are necessary only in the presence of PcG complexes. That is, the genetic evidence in Drosophila shows that when the PcG repressive mechanisms are disabled, trx and ash1 functions can be dispensed with. This strongly implies that their main function is to antagonize PcG repression and, in fact, loss of trx causes the PcG repression of homeotic genes in domains where they are normally active. Trx and Ash1 are both methyltransferases producing respectively H3K4me3 and H3K36me2 [165–167]. Trx binds to PREs both in the active and in the PcG-repressed state [60,168] and is in some way necessary for the recruitment of Ash1 when the target gene is in the active mode. A comparison of the active versus the repressed state of PcG target genes shows that the active state is accompanied by the loss of PRC2 and H3K27me3 and of much but not necessarily all PRC1 binding at the PRE. In addition, Trx and Ash1 bind to the entire transcription unit and are accompanied by a high level of H3K27ac and H3K36me1 [169]. The implication is that these histone marks stimulate transcription and antagonize the return of PcG silencing [124]. The mammalian homologs of Trx are MLL1 and 2 (Mixed Lineage Leukemia), which, like TRX, are needed for the effective expression of Hox genes. Unlike Trx, MLL proteins contain a CXXC zinc finger that targets them to CpG islands so they are potentially available at genes subject to PcG repression. Trx/MLL complexes belong to a set of related assemblies sometimes referred to as COMPASS complexes after the founding member,

II.  Molecular Mechanisms of Epigenetics



The Bivalent State

the yeast Set1/COMPASS complex [123,170,171]. They include the fly and mammalian SET1 complexes, methylases responsible for promoter-associated H3K4me3 necessary for effective transcriptional elongation [172], the Drosophila Trx complex and its mammalian homologs MLL1/2, which also produce H3K4me3, and the Drosophila Trr complex and mammalian counterparts MLL3/4, responsible for the H3K4me1 associated with enhancer and promoter activation [173]. Unlike Set1 mutations, which cause depletion of H3K4me3, loss of Trx/ MLL has little effect on total H3K4me3, indicating that they are specialized functions that affect a specific subset of genes [174]. How Trx/MLL complexes are more effective in antagonizing PcG repression is far from clear. A chromatin context containing H3K4me3 is known to inhibit PRC2 [132] but this is also produced by the SET1 methyltransferase. Another inhibitory histone mark is H3K36 methylation [132,133], normally produced towards the 3′ end of a transcription unit by the SET2 methyltransferase. The high level of H3K36 methylation produced by Ash1 may be a more effective antagonist of PRC2 activity. Perhaps most definitively antagonistic is H3K27 acetylation. Both Trx and Ash1 have been reported to act in conjunction with the CBP histone acetylase [175–177], which is responsible for the H3K27 acetylation found along with Trx and Ash1 at derepressed PcG target genes [169].

THE BIVALENT STATE Embryonic stem (ES) cells are pluripotent: they are able to enter any differentiation pathway and to contribute to any tissue if introduced into an early embryo. Therefore, the genes controlling all pathways are potentially available for expression. However, these genes must not be expressed if the cells are to remain pluripotent. The critical differentiation genes are targets of a complex regulatory circuitry involving the pluripotency factors, Oct4, Sox2, and Nanog (reviewed in ref. [178]). These genes show the presence of the PRC2 and the associated H3K27me3 mark and, at the same time, of H3K4me3, a mark usually associated with transcriptional activity and, in these genes, attributed to MLL2 [179]. This condition has been called “bivalent” and thought to correspond to a poised state ready to switch into a fully active mode or into a fully repressed mode [180]. Careful re-ChIP experiments show that these domains correspond to nucleosomes that lack DNA methylation and containing one histone H3 with the K4me3 mark and one with the K27me3 mark but no H3 containing both modifications [119,181]. Upon differentiation, the bivalent state is resolved toward active transcription or long-term repression and, correspondingly, either the active mark or the repressive mark is exclusively retained.

103

More recent studies have shown that many genes remain in a bivalent state not only in various types of multipotent tissue stem cells but in some cases also in differentiated cells. These may represent genes poised to respond to signals, not necessarily for differentiation [69,182]. A detailed study of genes in such a bivalent state showed that they are not in fact transcriptionally repressed in the usual sense [183]. Their promoter regions are acetylated and bind RNA pol II whose CTD is phosphorylated at Serine-5 but not Serine-2. These genes are not only poised but weakly active, producing low levels of transcripts relative to the amount of polymerase present at the promoter. Loss of both RING1 and RING2 causes the progressive loss of H2A ubiquitylation without affecting the distributions of PRC2 or H3K27me3. In this study, the loss of ubiquitylation was accompanied by the transcriptional upregulation of the bivalent genes and onset of differentiation. The interpretation that H2A ubiquitylation interferes with some stage of the transcription process was consistent with the report that H2A ubiquitylation prevents recruitment of the FACT complex and therefore interferes with effective transcriptional elongation [115]. This widely accepted interpretation was seriously challenged by recent findings that H2A ubiquitylation is neither necessary nor sufficient for PcG repression in flies and mammals [11,79,82]. The idea that bivalency is a part of a specific mechanism to “poise” genes in undifferentiated cells for activation or silencing is probably incorrect. An alternative view considers it a default state [184], the result of a progression of events that shape chromatin states in the early embryo. CpG islands tend to recruit H3K4 methylases, such as Set1 and MLL2 that contain a CXXC CpG-­ binding motif. The resulting H3K4me3 inhibits DNA methyltransferases, thus preventing long-term shutdown. If genes so marked are not transcriptionally activated by specific factors in embryonic stem cells, they are then targets for PRC2 and thus become bivalent. Perhaps a good way to envision this state is suggested by the role of Trx and Ash1 seen in Drosophila. According to this, the bivalent state would correspond to a dynamic equilibrium between Trx/MLL and ASH1L (the mammalian Ash1 homologue) on one hand, and PcG mechanisms on the other, presumably driven by subliminal amounts of activators and resulting in low levels of transcription, possibly frequently aborted, or incompletely spliced. In the absence of any activators, PcG target genes would be simply repressed. When subliminal levels of activators are available (in the pluripotent state), the bivalent state would be reached with extensive binding of Trx/MLL and ASH1L. The role of PcG complexes would thus be to raise the threshold for activators that, abetted by Trx/ MLL and Ash1L, would be required for productive transcriptional activation. Decreased levels of the pluripotency factors and increased levels of specific activators

II.  Molecular Mechanisms of Epigenetics

104

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

CONCLUSIONS

or repressors would tilt the equilibrium towards either full activation or full repression.

DIFFERENTIATION The elevated levels of PcG proteins in ES cells and the binding to a multitude of target genes controlling differentiation pathways suggested that PcG repression might be essential for the maintenance of the pluripotent state. Surprisingly, it has been possible to generate mouse ES cells entirely lacking PRC2 function [185–187]. Recent studies [84] confirm that in these cells, there is little change in the expression of PcG target genes and the cells remain pluripotent, indicating that other features of the pluripotency circuitry are sufficient maintain the repression. What is defective in these cells is the ability to differentiate correctly. If differentiation is induced, the pluripotency genes and genes specific for other differentiation pathways are not properly turned off, the appropriate differentiation genes are not correctly activated and the differentiation program fails to be carried out. A recent work analyzing the role of PcG mechanisms in epidermal differentiation used a conditional knockout of EZH2 to show that, once differentiation is under way [188], PcG repression is no longer needed to keep other differentiation pathways turned off. Other mechanisms, probably initiated by G9A methylation of H3K9 and recruitment of DNA methyltransferases, have already established long-term silencing [189,190]. The lack of PRC2 function, however, disrupts the normal process of differentiation, resulting in the premature expression of the differentiated functions and exit from the cell cycle before the appropriate precursor cells have been produced. PcG complexes regulate Ink4A and B, two important inhibitors of cyclin-dependent kinases. Ink4A and B are not directly involved in differentiation but their derepression results in cell cycle arrest in tissue stem cells or precursor cells [191,192]. In normal epidermal differentiation, for example, the expression of PRC2 components gradually winds down, allowing the gradual derepression of Ink4A/B and eventual exit from the cell cycle as the epidermal differentiation program unfolds. Ebbing levels of PcG proteins are not the only things that determine the reactivation of repressed genes, including Ink4A/B. As intimated previously, balance between repression and expression depends on the relative levels of PcG proteins, TRX/MLL, the relevant transcriptional activators and the chromatin remodeling complexes that refashion the chromatin landscape to allow transcriptional activity. In certain cell types, the levels of the SWI/SNF chromatin remodeler are a determining factor in the decision whether to derepress Ink4A/B, inducing exit from the cell cycle and cellular senescence [193]. The SWI/SNF complex facilitates eviction of the PcG complexes and the action of TRX/MLL.

If PcG repression is so plastic, dynamic, and easily reversed by the presence of activators, what is epigenetic about it and how does it differ from any that of any other negative regulator? A way to view the role of PcG mechanisms is that they are the valves that regulate the flow of transcriptional activity in the nucleus: they raise the threshold of the signals or activators required to turn on a gene. Furthermore, this higher threshold is transmitted to the cellular progeny. As a result, these mechanisms can differentiate the ability of cells to respond to transcriptional signals depending on the history of the cell lineage. This, as developmental biologists have known for a long time, is essential for pattern formation, morphogenesis, organogenesis. As differentiation proceeds, alternative pathways have to be turned off more or less permanently. In mammals, this does not seem to be done by PcG mechanisms but by DNA methylation. Although phylogenetically ancient and present in many other insects, DNA methylation is not used in Drosophila development and PcG mechanisms have adapted to take over that function in ways that are adequate at least for the short lifespan and rapid developmental strategies of the fruitfly. How PcG mechanisms are modulated in response to metabolic states, intercellular signaling, stress response mechanisms and many other factors in development, aging, homeostasis, and disease, are questions that are now being actively studied with potential for many exciting advances. In this article, I have used PcG mechanisms in the plural to emphasize the realization that there is no single account that describes all the ways in which these versatile proteins are used in the genome or in different cells and different developmental contexts. Only some of these are discussed here and, doubtless, many more are yet to be discovered.

Abbreviations Abd-A Abdominal-A, Drosophila homeotic gene Adf1 ADH Transcription Factor 1, Drosophila DNAbinding factor AEBP2 Adipocyte Enhancer Binding Protein 2, optional component of PRC2 ANRIL Antisense Noncoding RNA in the INK4 Locus, lncRNA encoded by INK4A/ARF and required to recruit PcG complexes there Ash1 Absent, Small, or Homeotic discs, Drosophila histone H3K36 methyltransferase BCOR BCL6 Corepressor, component of noncanonical PRC1 BMI1 B cell-specific Moloney murine leukemia virus integration site-1, PCGF4 C17orf96 Optional component of mammalian PRC2 CBP CREB-Binding Protein, a powerful acetyltransferase CBX2,4,6,7,8 Chromobox protein homologs, mammalian Pc homologs

II.  Molecular Mechanisms of Epigenetics



CONCLUSIONS

Cg Combgap, Drosophila DNA-binding factor ChIP Chromatin Immunoprecipitation CpG islands characteristic genomic elements rich in CpG dinucleotides, often transcription initiation sites CTCF CCCTC-binding Factor, chromatin insulatorbinding protein CTD C-terminal Domain of RNA pol II CXXC motif a type of zinc-finger domain that binds to CpGrich regions dRAF Drosophila RING-Associated Factors, a noncanonical PRC1 Dsp1 Dorsal switch protein 1, Drosophila DNA-binding factor E(z) Enhancer of zeste, catalytic component of Drosophila PRC2 EED Embryonic Ectoderm Development, mammalian homolog of Drosophila Esc, component of PRC2 ES cells Embryonic Stem cells, pluripotent mammalian embryonic cells Esc, Escl Extra sex combs, Extra sex combs-like, component of Drosophila PRC2 EZH1, EZH2 Enhancer of zeste Homologs, catalytic component of mammalian PRC2 FACT Facilitates Chromatin Transcription, a complex that promotes transcriptional elongation FISH Fluorescent In Situ Hybridization GAF GAGA binding Facttor, Drosophila DNA-binding factor GFP Green Fluorescent Protein G9A mammalian histone H3K9 methyltransferase HOX genes genes encoding transcription factors homologous to Drosophila homeotic genes INK4A/ARF locus CDKN2A, mammalian complex locus encoding peptide inhibitors of cyclin kinases JARID2 Jumonji AT-rich Interactive Domain, optional component of PRC2 KDM2, KDM2B Histone H3K36 demethylase L(3)73Ah Lethal on chromosome 3 at 73Ah, a Drosophila PCGF3 homolog lncRNAs lonc noncoding RNAs MEL18 Mouse melanoma protein-18, PCGF2 MLL1,2 Mixed Lineage Leukemia 1,2, mammalian H3K4 methyltransferases homologs of Drosophila Trx MLL3,4 Mixed Lineage Leukemia 3,4, mammalian H3K4 methyltransferases homologs of Drosophila Trr MTF2 Metal-response element-binding Transcription Factor 2, also known as PCL2, optional component of mammalian PRC2 Nanog named from the Celtic Tir na nÓg legend, mammalian pluripotency factor OCT4 Octamer-binding transcription factor, mammalian pluripotency factor Ogt O-linked acetylglucosamine (GlcNAc) transferase PcG Polycomb Group PCGF Polycomb Group RING Finger protein, forms heterodimer with RING Pcl Polycomb-like, optional component of Drosophila PRC2 Ph-p, Ph-d Polyhomeotic proximal and distal, Drosophila PRC1 component PHC1-3 mammalian Ph homologs PHF1 PHD Finger protein 1, also known as PCL1, optional component of mammalian PRC2 PHF17 PHD Finger protein 17, also known as PCL3, optional component of mammalian PRC2

105

Pleiohomeotic, DNA-binding protein required at many PREs PhoRC Pho Repressive Complex, contains Pho and SFMBT PRC1, PRC2 Polycomb Repressive Complex 1, 2 PRE Polycomb Response Element, DNA element that recruits PcG proteins Psc Posterior sex comb, Drosophila PCGF protein RbAp48 Retinoblastoma Associated Protein48, also known as RBBP4/7, Cap1, component of Drosophila PRC2 Re-ChIP sequential ChIP to identify association of two immunological markers REST RE1-Silencing Transcription Factor, inhibitor of neuronal differentiation RING so-called Really Interesting New Gene, PRC1 component RNA pol II RNA polymerase II RYBP RING and YY1 Binding Protein SAM Sterile Alpha Motif, structural protein motif that can polymerize head to tail Sce Sex comb extra, Drosophila PcG gene, encodes RING Scm Sex comb on midleg, PcG component, associates with PRC1, PRC2, PhoRC Set1/COMPASS Complex Proteins Associated with Set1, yeast histone H3K4 methyltransferase complex and related complexes in flies and mammals SFMBT Scm-related containing Four MBT domains, component of PhoRC SILAC Stable Isotope Labeling with Amino acids in Cell culture SNAIL a family of mammalian repressors related to Drosophila Snail SOX2 Sex-determining region Y-box 2; mammalian pluripotency factor SPM see SAM Spps Sp1-like Factor for pairing-Sensitive Silencing, Drosophila DNA-binding factor Su(z)12 Suppressor of zeste 12, component of Drosophila PRC2 Su(z)2 Suppressor of zeste-2, Drosophila PCGFprotein SUMO Small Ubiquitin-like Modifier SUZ112 mammalian Su(z)12 homolog, component of PRC2 SWI/SNF nucleosome remodeling complex related to yeast SWItch/Sucrose Nonfermentable TAD Topologically Asociated Domain Trithorax Drosophila histone H3K4 methyltranferase Trr Trithorax-related, Drosophila histone H3K4 methyltransferase Ubx Ultrabithorax, Drosophila homeotic gene UTX Ubiquitously Transcribed tetratricopeptide repeat X chromosome, histone H3K27 demethylase Xist X inactive specific transcript, lncRNA required for mammalian X chromosome inactivation YAF2 YY1-Associated Factor 2, RYBP homolog Pho

Note: In an attempt to reduce confusion, I have written the abbreviated Drosophila protein names with initial capital only; mammalian protein names are abbreviated with all capitals. Naturally, there are exceptions, particularly when the same acronym is used for both Drosophila and mammals.

II.  Molecular Mechanisms of Epigenetics

106

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

References [1] Paro R, Hogness DS. The Polycomb protein shares a homologous domain with a heterochromatin-associated protein of Drosophila. Proc Natl Acad Sci USA 1991;88:263–7. [2] Simon JA, Kingston RE. Occupying chromatin: Polycomb mechanisms for getting to genomic targets, stopping transcriptional traffic, and staying put. Mol Cell 2013;49:808–24. [3] Schwartz YB, Pirrotta V. A new world of Polycombs: unexpected partnerships and emerging functions. Nat Rev Genet 2013;14:853–64. [4] Piunti A, Shilatifard A. Epigenetic balance of gene expression by Polycomb and COMPASS families. Science 2016;352:aad9780. [5] Geisler SJ, Paro R. Trithorax and Polycomb group-­dependent regulation: a tale of opposing activities. Development 2015;142: 2876–87. [6] de Napoles M, Mermoud JE, Wakao R, Tang YA, Endoh M, ­Appanah R, et al. Polycomb group proteins Ring1A/B link ubiquitylation of histone H2A to heritable gene silencing and X inactivation. Dev Cell 2004;7:663–76. [7] Fischle W, Wang Y, Jacobs SA, Kim Y, Allis CD, Khorasanizadeh S. Molecular basis for the discrimination of repressive methyllysine marks in histone H3 by Polycomb and HP1 chromodomains. Genes Dev 2003;17:1870–81. [8] Kim CA, Gingery M, Pilpa RM, Bowie JU, The SAM. domain of polyhomeotic forms a helical polymer. Nat Struct Biol 2002;9:453–6. [9] Peterson AJ, Mallin DR, Francis NJ, Ketel CS, Stamm J, Voeller RK, et al. Requirement for sex comb on midleg protein interactions in Drosophila polycomb group repression. Genetics 2004;167:1225–39. [10] Lagarou A, Mohd-Sarip A, Moshkin YM, Chalkley GE, ­Bezstarosti K, Demmers JAA, et al. dKDM2 couples histone H2A ubiquitylation to histone H3 demethylation during Polycomb group silencing. Genes Dev 2008;22:2799–810. [11] Lee H-G, Kahn TG, Simcox A, Schwartz YB, Pirrotta V. Genomewide activities of Polycomb complexes control pervasive ­transcription. Genome Res 2015;25:1170–81. [12] Gao Z, Zhang J, Bonasio R, Strino F, Sawai A, Parisi F, et al. PCGF homologs, CBX proteins, and RYBP define functionally distinct PRC1 family complexes. Mol Cell 2012;45:344–56. [13] Tavares L, Dimitrova E, Oxley D, Webster J, Poot R, Demmers J, et al. RYBP-PRC1 complexes mediate H2A ubiquitylation at polycomb target sites independently of PRC2 and H3K27me3. Cell 2012;148:664–78. [14] Taherbhoy AM, Huang OW, Cochran AG. BMI1-RING1B is an autoinhibited RING E3 ubiquitin ligase. Nat Commun 2015;6:7621. [15] Junco SE, Wang R, Gaipa JC, Taylor AB, Schirf V, Gearhart MD, et al. Structure of the polycomb group protein PCGF1 in complex with BCOR reveals basis for binding selectivity of PCGF homologs. Structure 2013;21:665–71. [16] Farcas AM, Blackledge NP, Sudbery I, Long HK, McGouran JF, Rose NR, et al. KDM2B links the polycomb repressive complex 1 (PRC1) to recognition of CpG islands. eLife Sci. 2012;1:e00205. [17] Tie F, Furuyama T, Prasad-Sinha J, Jane E, Harte PJ. The Drosophila polycomb group proteins ESC and E(Z) are present in a complex containing the histone-binding protein p55 and the histone deacetylase RPD3. Development 2001;128:275–86. [18] Cao R, Wang L, Wang H, Xin L, Erdjument-Bromage H, Tempst P, et al. Role of histone H3 lysine 27 methylation in polycombgroup silencing. Science 2002;298:1039–43. [19] Czermin B, Melfi R, McCabe D, Seitz V, Imhof A, Pirrotta V. Drosophila Enhancer of Zeste/ESC complexes have a histone H3 methyltransferase activity that marks chromosomal Polycomb sites. Cell 2002;111:185–96.

[20] Müller J, Hart CM, Francis NJ, Vargas ML, Sengupta A, Wild B, et al. Histone methyltransferase activity of a Drosophila Polycomb group repressor complex. Cell 2002;111:197–208. [21] Kuzmichev A, Nishioka K, Erdjument-Bromage H, Tempst P, Reinberg D. Histone methyltransferase activity associated with a human multiprotein complex containing the Enhancer of Zeste protein. Genes Dev 2002;22:2893–905. [22] Tie F, Prasad-Sinha J, Birve A, Rasmuson-Lestander Å, Harte PJ. A 1-megadalton ESC/E(Z) complex from Drosophila that contains Polycomb like and RPD3. Mol Cell Biol 2003;23:3352–62. [23] Kuzmichev A, Jenuwein T, Tempst P, Reinberg D. Different Ezh2-containing complexes target methylation of histone H1 or nucleosomal histone H3. Mol Cell 2004;14:183–93. [24] Kim H, Kang K, Kim J. AEBP2 as a potential targeting protein for Polycomb Repression Complex PRC2. Nucl Acids Res 2009;37:2940–50. [25] Ciferri C, Lander GC, Maiolica A, Herzog F, Aebersold R, ­Nogales E. Molecular architecture of human polycomb repressive complex 2. eLife 2012;1:e00005. [26] Li G, Margueron R, Ku M, Chambon P, Bernstein BE, Reinberg D. Jarid2 and PRC2, partners in regulating gene expression. Genes Dev 2010;24(4):368–80. [27] Peng JC, Valouev A, Swigut T, Zhang J, Zhao Y, Sidow A, et al. ­Jarid2/jumonji coordinates control of PRC2 enzymatic activity and target gene occupancy in pluripotent cells. Cell 2009;139: 1290–302. [28] Shen X, Kim W, Fujiwara Y, Simon MD, Liu Y, Mysliwiec MR, et al. Jumonji modulates polycomb activity and self-renewal versus differentiation of stem cells. Cell 2009;139:1303–14. [29] Pasini D, Cloos PAC, Walfridsson J, Olsson L, Bukowski J­-P, Johansen JV, et al. JARID2 regulates binding of the Polycomb repressive complex 2 to target genes in ES cells. Nature 2010;464:306–10. [30] Nekrasov M, Klymenko T, Fraterman S, Papp B, Oktaba K, ­Köcher T, et al. Pcl-PRC2 is needed to generate high levels of H3-K27 trimethylation at Polycomb target genes. EMBO J 2007;26:4078–88. [31] Sarma K, Margueron R, Ivanov A, Pirrotta V, Reinberg D. Ezh2 requires PHF1 to efficiently catalyze H3 lysine 27 trimethylation In Vivo. Mol Cell Biol 2008;28:2718–31. [32] Ballaré C, Lange M, Lapinaite A, Martin GM, Morey L, Pascual G, et al. Phf19 links methylated Lys36 of histone H3 to regulation of Polycomb activity. Nat Struct Mol Biol 2012;19:1257–65. [33] Casanova M, Preissner T, Cerase A, Poot R, Yamada D, Li X, et al. Polycomblike 2 facilitates the recruitment of PRC2 Polycomb group complexes to the inactive X chromosome and to target loci in embryonic stem cells. Development 2011;138:1471–82. [34] Hunkapiller J, Shen Y, Diaz A, Cagney G, McCleary D, RamalhoSantos M, et al. Polycomb-like 3 promotes polycomb repressive complex 2 binding to CpG islands and embryonic stem cell selfrenewal. PLoS Genet 2012;8:e1002576. [35] Smits AH, Jansen PWTC, Poser I, Hyman AA, Vermeulen M. Stoichiometry of chromatin-associated protein complexes revealed by label-free quantitative mass spectrometry-based proteomics. Nucleic Acids Res 2013;41:e28. [36] Grijzenhout A, Godwin J, Koseki H, Gdula M, Szumska D, McGouran JF, et al. Functional analysis of AEBP2, a PRC2 ­ Polycomb protein, reveals a Trithorax phenotype in embryonic development and in ES cells. Development 2016;143:2716–23. [37] Alekseyenko AA, Gorchakov AA, Kharchenko PV, Kuroda MI. Reciprocal interactions of human C10orf12 and C17orf96 with PRC2 revealed by BioTAP-XL cross-linking and affinity purification. Proc Natl Acad Sci USA 2014;111:2488–93. [38] Son J, Shen SS, Margueron R, Reinberg D. Nucleosome-binding activities within JARID2 and EZH1 regulate the function of PRC2 on chromatin. Genes Dev 2013;27:2663–77.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[39] Rastelli L, Chan CS, Pirrotta V. Related chromosome binding sites for zeste, suppressors of zeste and Polycomb group proteins in Drosophila and their dependence on Enhancer of zeste function. EMBO J 1993;12:1513–22. [40] Fauvarque M-O, Dura J-M. polyhomeotic regulatory sequences induce developmental regulator-dependent variegation and targeted P-element insertions in Drosophila. Genes Dev 1993;7:1508–20. [41] Bloyer S, Cavalli G, Brock HW, Dura J-M. Identification and characterization of polyhomeotic PREs and TREs. Dev Biol 2003;261:426–42. [42] Schwartz YB, Kahn TG, Nix DA, Li X-Y, Bourgon R, Biggin M, et al. Genome-wide analysis of Polycomb targets in Drosophila melanogaster. Nat Genet 2006;38:700–5. [43] Leeb M, Pasini D, Novatchkova M, Jaritz M, Helin K, Wutz A. Polycomb complexes act redundantly to repress genomic repeats and genes. Genes Dev 2010;24:265–76. [44] Chan C-, Rastelli ., Pirrotta V. A Polycomb response element in the Ubx gene that determines an epigenetically inherited state of repression. EMBO J 1994;13:2553–64. [45] Kassis JA. Unusual properties of regulatory DNA from the Drosophila engrailed gene: three “pairing-sensitive” sites within a 1.6 kb region. Genetics 1994;136:1025–38. [46] Mihaly J, Barges S, Sipos L, Maeda R, Cleard F, Hogga I, et al. Dissecting the regulatory landscape of the Abd-B gene of the ­bithorax complex. Development 2006;133:2983–93. [47] Horard B, Tatout C, Poux S, Pirrotta V. Structure of a Polycomb Response element and in vitro binding of Polycomb group complexes containing GAGA factor. Mol Cell Biol 2000;20:3187–97. [48] Americo G, Whiteley M, Brown JL, Fujioka M, Jaynes JB, Kassis JA. A complex array of DNA-binding complexes required for pairing-sensitive silencing by a Polycomb Group response element from the Drosophila engrailed gene. Genetics 2002;160:1561–71. [49] Déjardin J, Rappailles A, Cuvier O, Grimaud C, Decoville M, Locker D, et al. Recruitment of Drosophila Polycomb group proteins to chromatin by DSP1. Nature 2005;434:533–8. [50] Brown JL, Kassis JA. Spps, a Drosophila Sp1/KLF family member, binds to PREs and is required for PRE activity late in development. Development 2010;137:2597–602. [51] Orsi GA, Kasinathan S, Hughes KT, Saminadin-Peter S, Henikoff S, Ahmad K. High-resolution mapping defines the cooperative architecture of Polycomb response elements. Genome Res 2014;25:809–20. [52] Ray P, De S, Mitra A, Bezstarosti K, Demmers JAA, Pfeifer K, Kassis JA. Combgap contributes to recruitment of Polycomb group proteins in Drosophila. Proc Natl Acad Sci USA 2016;113:3826–31. [53] Fritsch C, Brown JL, Kassis JA, Müller J. The DNA-binding Polycomb group protein Pleiohomeotic mediates silencing of a Drosophila homeotic gene. Development 1999;126:3905–13. [54] Brown JL, Fritsch C, Mueller J, Kassis JA. The Drosophila pholike gene encodes a YY1-related DNA binding protein that is redundant with pleiohomeotic in homeotic gene silencing. Development 2003;130:285–94. [55] Klymenko T, Papp B, Fischle W, Kocher T, Schelder M, Fritsch C, Wild B, Wilm M, Muller J. A Polycomb group protein complex with sequence-specific DNA-binding and selective methyllysine-binding activities. Genes Dev 2006;20:1110–22. [56] Grimm C, Matos R, Ly-Hartig N, Steuerwald U, Lindner D, Rybin V, et al. Molecular recognition of histone lysine methylation by the Polycomb group repressor dSfmbt. EMBO J 2009;28:1965–77. [57] Peterson AJ, Kyba M, Bornemann D, Morgan K, Brock HW, Simon J. A domain shared by the Polycomb group proteins Scm and ph mediates heterotypic and homotypic interactions. Mol Cell Biol 1997;17:6683–92.

107

[58] Kahn TG, Stenberg P, Pirrotta V, Schwartz YB. Combinatorial interactions are required for the efficient recruitment of Pho repressive complex (PhoRC) to polycomb response elements. PLoS Genet 2014;10:e1004495. [59] Frey F, Sheahan T, Finkl K, Stoehr G, Mann M, Benda C, et al. Molecular basis of PRC1 targeting to Polycomb response elements by PhoRC. Genes Dev 2016;30:1116–27. [60] Kahn TG, Schwartz YB, Dellino GI, Pirrotta V. Polycomb complexes and the propagation of the methylation mark at the Drosophila Ubx gene. J Biol Chem 2006;281:29064–75. [61] Papp B, Muller J. Histone trimethylation and the maintenance of transcriptional ON and OFF states by trxG and PcG proteins. Genes Dev 2006;20:2041–54. [62] Mohd-Sarip A, van der Knaap JA, Wyman C, Kanaar R, Schedl P, Verrijzer CP. Architecture of a Polycomb nucleoprotein complex. Mol Cell 2006;24:91–100. [63] Philip P, Boija A, Vaid R, Churcher A, Meyers D, Cole P, et al. CBP binding outside of promoters and enhancers in Drosophila melanogaster. Epigen Chromatin 2015;8:48. [64] Wang L, Brown JL, Cao R, Zhang Y, Kassis JA, Jones RS. Hierarchical recruitment of Polycomb Group silencing complexes. Mol Cell 2004;14:637–46. [65] Kahn TG, Dorafshan E, Schultheis D, Zare A, Stenberg P, Reim I, et al. Interdependence of PRC1 and PRC2 for recruitment to Polycomb Response Elements. Nucleic Acids Res 2016;44(21): 10132–49. [66] Sing A, Pannell D, Karaiskakis A, Sturgeon K, Djabali M, Ellis J, et al. A vertebrate polycomb response element governs segmentation of the posterior hindbrain. Cell 2009;138:885–97. [67] Woo CJ, Kharchenko PV, Daheron L, Park PJ, Kingston RE. A region of the human HOXD cluster that confers polycomb-group responsiveness. Cell 2010;140:99–110. [68] Mikkelsen TS, Ku M, Jaffe DB, Issac B, Lieberman E, Giannoukos G, et al. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature 2007;448:553–60. [69] Ku M, Koche RP, Rheinbay E, Mendenhall EM, Endoh M, Mikkelsen TS, et al. Genomewide analysis of PRC1 and PRC2 occupancy identifies two classes of bivalent domains. PLoS Genet 2008;4:e1000242. [70] Mendenhall EM, Koche RP, Truong T, Zhou VW, Issac B, Chi AS, et al. GC-rich sequence elements recruit PRC2 in mammalian ES cells. PLoS Genet 2010;6:e1001244. [71] Lynch MD, Smith AJH, De Gobbi M, Flenley M, Hughes JR, Vernimmen D, et al. An interspecies analysis reveals a key role for unmethylated CpG dinucleotides in vertebrate Polycomb complex recruitment. EMBO J 2012;31:317–29. [72] Thomson JP, Skene PJ, Selfridge J, Clouaire T, Guy J, Webb S, et al. CpG islands influence chromatin structure via the CpGbinding protein Cfp1. Nature 2010;464:1082–6. [73] Deaton AM, Bird A. CpG islands and the regulation of transcription. Genes Dev 2011;25:1010–22. [74] Blackledge NP, Farcas AM, Kondo T, King HW, McGouran JF, Hanssen LLP, et al. Variant PRC1 complex-dependent H2A ubiquitylation drives PRC2 recruitment and Polycomb domain formation. Cell 2014;157:1445–59. [75] Wu X, Johansen JV, Helin K. Fbxl10/Kdm2b recruits polycomb repressive complex 1 to CpG islands and regulates H2A ubiquitylation. Mol Cell 2013;49:1134–46. [76] He J, Shen L, Wan M, Taranova O, Wu H, Zhang Y. Kdm2b maintains murine embryonic stem cell status by recruiting PRC1 complex to CpG islands of developmental genes. Nat Cell Biol 2013;15:373–84. [77] Cooper S, Dienstbier M, Hassan R, Schermelleh L, Sharif J, Blackledge NP, et al. Targeting polycomb to pericentric heterochromatin in embryonic stem cells reveals a role for H2AK119u1 in PRC2 recruitment. Cell Rep 2014;7:1456–70.

II.  Molecular Mechanisms of Epigenetics

108

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

[78] Kalb R, Latwiel S, Baymaz HI, Jansen PWTC, Müller CW, Vermeulen M, et al. Histone H2A monoubiquitination promotes histone H3 methylation in Polycomb repression. Nat Struct Mol Biol 2014;21:569–71. [79] Pengelly AR, Kalb R, Finkl K, Müller J. Transcriptional repression by PRC1 in the absence of H2A monoubiquitylation. Genes Dev 2015;29:1487–92. [80] Chiacchiera F, Rossi A, Jammula S, Piunti A, Scelfo A, Ordóñdez-Morán P, et al. Polycomb complex PRC1 preserves intestinal stem cell identity by sustaining Wnt/β-Catenin Transcriptional Activity. Cell Stem Cell 2016;18:91–103. [81] Scheuermann JC, de Ayala Alonso AG, Oktaba K, Ly-Hartig N, McGinty RK, Fraterman S, et al. Histone H2A deubiquitinase activity of the Polycomb repressive complex PR-DUB. Nature 2010;465:243–7. [82] Illingworth RS, Moffat M, Mann AR, Read D, Hunter CJ, Pradeepa MM, et al. The E3 ubiquitin ligase activity of RING1B is not essential for early mouse development. Genes Dev 2015;29:1897–902. [83] Poux S, Kostic C, Pirrotta V. Hunchback-independent silencing of late Ubx enhancers by a Polycomb group response element. EMBO J 1996;15:4713–22. [84] Riising EM, Comet I, Leblanc B, Wu X, Johansen JV, Helin K. Gene silencing triggers polycomb repressive complex 2 recruitment to CpG islands genome wide. Mol Cell 2014;55:347–60. [85] Dietrich N, Lerdrup M, Landt E, Agrawal-Singh S, Bak M, Tommerup N, et al. REST-mediated recruitment of polycomb repressor complexes in mammalian cells. PLoS Genet 2012;8:e1002494. [86] Arnold P, Schöler A, Pachkov M, Balwierz PJ, Jørgensen H, Stadler MB, et al. Modeling of epigenome dynamics identifies transcription factors that mediate Polycomb targeting. Genome Res 2013;23:60–73. [87] Rinn JL, Kertesz M, Wang JK, Squazzo SL, Xu X, Brugmann SA, et al. Functional demarcation of active and silent chromatin domains in human HOX loci by noncoding RNAs. Cell 2007;129:1311–23. [88] Tsai M-C, Manor O, Wan Y, Mosammaparast N, Wang JK, Lan F, et al. Long noncoding RNA as modular scaffold of histone modification complexes. Science 2010;329:689–93. [89] Liu G-Y, Zhao G-N, Chen X-F, Hao D-L, Zhao X, Lv X, et al. The long noncoding RNA Gm15055 represses Hoxa gene expression by recruiting PRC2 to the gene cluster. Nucleic Acids Res 2016;44:2613–27. [90] Kotake Y, Nakagawa T, Kitagawa K, Suzuki S, Liu N, Kitagawa M, et al. Long non-coding RNA ANRIL is required for the PRC2 recruitment to and silencing of p15INK4B tumor suppressor gene. Oncogene 2011;30:1956–62. [91] Zhao J, Sun BK, Erwin JA, Song J-J, Lee JT. Polycomb proteins targeted by a short repeat RNA to the mouse X chromosome. Science 2008;322:750–6. [92] Davidovich C, Zheng L, Goodrich KJ, Cech TR. Promiscuous RNA binding by Polycomb repressive complex 2. Nat Struct Mol Biol 2013;20:1250–7. [93] da Rocha ST, Boeva V, Escamilla-Del-Arenal M, Ancelin K, Granier C, Matias NR, et al. Jarid2 is implicated in the initial Xist-induced targeting of PRC2 to the inactive X chromosome. Mol Cell 2016;53:301–16. [94] Kaneko S, Bonasio R, Saldaña-Meyer R, Yoshida T, Son J, Nishino K, et al. Interactions between JARID2 and noncoding RNAs regulate PRC2 recruitment to chromatin. Mol Cell 2014;53:290–300. [95] Gehani SS, Agrawal-Singh S, Dietrich N, Christophersen NS, Helin K, Hansen K. Polycomb group protein displacement and gene activation through MSK-dependent H3K27me3S28 phosphorylation. Mol Cell 2010;39:886–900. [96] Stojic L, Jasencakova Z, Prezioso C, Stutzer A, Bodega B, Pasini, et al. Chromatin regulated interhange between Polycomb repres-

[97]

[98] [99]

[100]

[101]

[102]

[103]

[104] [105]

[106]

[107]

[108]

[109]

[110] [111]

[112]

[113]

[114]

[115]

sive complex 2 (PRC2)-Ezh2 and PRC2-Ezh1 complexes controls myogenin activation in skeletal muscle cells. Epigen Chromatin 2011;4:16. Comet I, Schuettengruber B, Sexton T, Cavalli G. A chromatin insulator driving three-dimensional Polycomb response element (PRE) contacts and Polycomb association with the chromatin fiber. Proc Natl Acad Sci USA 2011;108:2294–9. Dekker J. The three ’C’ s of chromosome conformation capture: controls, controls, controls. Nat Meth 2006;3:17–21. Bowman SK, Deaton AM, Domingues H, Wang PI, Sadreyev RI, Kingston RE, et al. H3K27 modifications define segmental regulatory domains in the Drosophila bithorax complex. eLife 2014;3:e02833. Poux S, Melfi R, Pirrotta V. Establishment of Polycomb silencing requires a transient interaction between PC and ESC. Genes Dev 2001;15:2509–14. Cao Q, Wang X, Zhao M, Yang R, Malik R, Qiao Y, et al. The central role of EED in the orchestration of polycomb group complexes. Nat Commun 2014;5:3127. Kang H, McElroy KA, Jung YL, Alekseyenko AA, Zee BM, Park PJ, et al. Sex comb on midleg (Scm) is a functional link between PcG-repressive complexes in Drosophila. Genes Dev 2015;29:1136–50. Wang L, Jahren N, Miller EL, Ketel CS, Mallin DR, Simon JA. Comparative analysis of chromatin binding by sex comb on midleg (SCM) and other polycomb group repressors at a drosophila hox gene. Mol Cell Biol 2010;30:2584–93. Francis NJ, Kingston RE, Woodcock CL. Chromatin compaction by a polycomb group protein complex. Science 2004;306:1574–7. Shao Z, Raible F, Mollaaghababa R, Guyon JR, Wu C-t, Bender W, et al. Stabilization of chromatin structure by PRC1, a Polycomb complex. Cell 1999;98:37–46. Chambeyron S, Bickmore WA. Chromatin decondensation and nuclear reorganization of the HoxB locus upon induction of transcription. Genes Dev 2004;18:1119–30. Morey C, Da Silva NR, Perry P, Bickmore WA. Nuclear reorganisation and chromatin decondensation are conserved, but distinct, mechanisms linked to Hox gene activation. Development 2007;134:909–19. Mishra RK, Mihaly J, Barges S, Spierer A, Karch F, Hagstrom K, et al. The iab-7 polycomb response element maps to a nucleosome-free region of chromatin and requires both GAGA and pleiohomeotic for silencing activity. Mol Cell Biol 2001;21:1311–8. Dellino GI, Schwartz YB, Farkas G, McCabe D, Elgin SCR, Pirrotta V. Polycomb silencing blocks transcription initiation. Mol Cell 2004;13:887–93. Mito Y, Henikoff JG, Henikoff S. Histone replacement marks the boundaries of cis-regulatory domains. Science 2007;315:1408–11. Buchenau P, Hodgson J, Strutt H, Arndt-Jovin DJ. The distribution of polycomb-group proteins during cell division and development in Drosophila embryos: impact on models for silencing. J Cell Biol 1998;141:469–81. Ficz G, Heintzmann R, Arndt-Jovin DJ. Polycomb group protein complexes exchange rapidly in living Drosophila. Development 2005;132:3963–76. Cheutin T, Cavalli G. Progressive polycomb assembly on H3K27me3 compartments generates polycomb bodies with developmentally regulated motion. PLoS Genet 2012;8:e1002465. Stock JK, Giadrossi S, Casanova M, Brookes E, Vidal M, Koseki H, et al. Ring1-mediated ubiquitination of H2A restrains poised RNA polymerase II at bivalent genes in mouse ES cells. Nat Cell Biol 2007;9:1428–35. Zhou W, Zhu P, Wang JK, Pascual G, Ohgi KA, Lozach J, et al. Histone H2A monoubiquitination represses transcription by inhibiting RNA polymerase II transcriptional elongation. Mol Cell 2008;29:69–80.

II.  Molecular Mechanisms of Epigenetics



REFERENCES

[116] Tie F, Banerjee R, Fu C, Stratton CA, Fang M, Harte PJ. Polycomb inhibits histone acetylation by CBP by binding directly to its catalytic domain. Proc Natl Acad Sci USA 2016;113:E744–53. [117] Schaaf CA, Misulovin Z, Gause M, Koenig A, Gohara DW, Watson A, et al. Cohesin and polycomb proteins functionally interact to control transcription at silenced and active genes. PLoS Genet 2013;9:e1003560. [118] Jung HR, Pasini D, Helin K, Jensen ON. Quantitative mass spectrometry of histones H3.2 and H3.3 in Suz12-deficient mouse embryonic stem cells reveals distinct dynamic post-translational modifications at Lys-27 and Lys-36. Mol Cell Proteomics 2010;9:838–50. [119] Voigt P, LeRoy G, Drury WJI, Zee BM, Son J, Beck DB, et al. Asymmetrically modified nucleosomes. Cell 2012;151:181–93. [120] Ferrari KJ, Scelfo A, Jammula S, Cuomo A, Barozzi I, Stützer A, et al. Polycomb-dependent H3K27me1 and H3K27me2 regulate active transcription and enhancer fidelity. Mol Cell 2014;53:49–62. [121] Filion GJ, van Bemmel JG, Braunschweig U, Talhout W, Kind J, Ward LD, et al. Systematic protein location mapping reveals five principal chromatin types in drosophila cells. Cell 2010;143:212–24. [122] Tie F, Banerjee R, Conrad PA, Scacheri PC, Harte PJ. Histone demethylase utx and chromatin remodeler BRM bind directly to CBP and modulate acetylation of histone H3 lysine 27. Mol Cell Biol 2012;32:2323–34. [123] Mohan M, Herz H-M, Smith ER, Zhang Y, Jackson J, Washburn MP, et al. The COMPASS family of H3K4 methylases in drosophila. Mol Cell Biol 2011;31:4310–8. [124] Tie F, Banerjee R, Stratton CA, Prasad-Sinha J, Stepanik V, Zlobin A, et al. CBP-mediated acetylation of histone H3 lysine 27 antagonizes Drosophila Polycomb silencing. Development 2009;136:3131–41. [125] Pasini D, Malatesta M, Jung HR, Walfridsson J, Willer A, Olsson L, et al. Characterization of an antagonistic switch between histone H3 lysine 27 methylation and acetylation in the transcriptional regulation of Polycomb group target genes. Nucl Acids Res 2010;38:4958–69. [126] Heintzman ND, Stuart RK, Hon G, Fu YT, Ching CW, Hawkins RD, et al. Distinctive and predictive chromain signatures of transcriptional promoters and enhancers inthe human genome. Nat Genet 2007;39:311–8. [127] Creyghton MP, Cheng AW, Welstead GG, Kooistra T, Carey BW, Steine EJ, et al. Histone H3K27ac separates active from poised enhancers and predicts developmental state. Proc Natl Acad Sci 2010;107:21931–6. [128] Rada-Iglesias A, Bajpai R, Swigut T, Brugmann SA, Flynn RA, Wysocka J. A unique chromatin signature uncovers early developmental enhancers in humans. Nature 2011;470:279–83. [129] Groth A, Rocha W, Verreault A, Almouzni G. Chromatin challenges during DNA replication and repair. Cell 2007;128:721–33. [130] McCabe MT, Graves AP, Ganji G, Diaz E, Halsey WS, Jiang Y, et al. Mutation of A677 in histone methyltransferase EZH2 in human B-cell lymphoma promotes hypertrimethylation of histone H3 on lysine 27 (H3K27). Proc Natl Acad Sci USA 2012;109:2989–94. [131] Alabert C, Barth TK, Reverón-Gómez N, Sidoli S, Schmidt A, Jensen ON, et al. Two distinct modes for propagation of histone PTMs across the cell cycle. Genes Dev 2015;29:585–90. [132] Schmitges FW, Prusty AB, Faty M, Stützer A, Lingaraju GM, Aiwazian J, et al. Histone Methylation by PRC2 is inhibited by active chromatin marks. Mol Cell 2011;42:330–41. [133] Yuan W, Xu M, Huang C, Liu N, Chen S, Zhu B. H3K36 methylation antagonizes PRC2-mediated H3K27 methylation. J Biol Chem 2011;286:7983–9. [134] Yuan W, Wu T, Fu H, Dai C, Wu H, Liu N, et al. Dense chromatin activates polycomb repressive complex 2 to regulate H3 Lysine 27 Methylation. Sci 2012;337:971–5.

109

[135] Margueron R, Justin N, Ohno K, Sharpe ML, Son J, Drury WJ III, et al. Role of the polycomb protein EED in the propagation of repressive histone marks. Nature 2009;461:762–7. [136] Sigrist CJA, Pirrotta V. Chromatin insulator elements block the silencing of a target gene by the Drosophila Polycomb Response Element (PRE) but allow trans interactions between PREs on different chromosomes. Genetics 1997;147:209–21. [137] Grimaud C, Bantignies F, Pal-Bhadra M, Ghana P, Bhadra U, Cavalli G. RNAi components are required for nuclear clustering of Polycomb Group Response Elements. Cell 2006;124:957–71. [138] Ren X, Vincenz C, Kerppola TK. Changes in the distributions and dynamics of polycomb repressive complexes during embryonic stem cell differentiation. Mol Cell Biol 2008;28:2884–95. [139] Bantignies F, Roure V, Comet I, Leblanc B, Schuettengruber B, Bonnet J, et al. Polycomb-dependent regulatory contacts between distant hox loci in drosophila. Cell 2011;144:214–26. [140] Tolhuis B, Blom M, Kerkhoven RM, Pagie L, Teunissen H, Nieuwland M, et al. Interactions among polycomb domains are guided by chromosome architecture. PLoS Genet 2011;7:e1001343. [141] Sexton T, Yaffe E, Kenigsberg E, Bantignies F, Leblanc B, Hoichman M, et al. Three-dimensional folding and functional organization principles of the drosophila genome. Cell 2012;148:458–72. [142] Robinson AK, Leal BZ, Chadwell LV, Wang R, Ilangovan U, Kaur Y, et al. The growth-suppressive function of the polycomb group protein polyhomeotic is mediated by polymerization of its sterile alpha motif (SAM) domain. J Biol Chem 2012;287:8702–13. [143] Isono K, Endo TA, Ku M, Yamada D, Suzuki R, Sharif J, et al. SAM domain polymerization links subnuclear clustering of PRC1 to gene silencing. Dev Cell 2013;26:565–77. [144] Nanyes DR, Junco SE, Taylor AB, Robinson AK, Patterson NL, Shivarajpur A, et al. Multiple polymer architectures of human polyhomeotic homolog 3 sterile alpha motif. Proteins 2014;82:2823–30. [145] Wani AH, Boettiger AN, Schorderet P, Ergun A, Munger C, Sadreyev RI, et al. Chromatin topology is coupled to Polycomb group protein subnuclear organization. Nat Commun 2016;7. [146] Gambetta MC, Oktaba K, Muller J. Essential role of the glycosyltransferase Sxc/Ogt in polycomb repression. Science 2009;325:93–6. [147] Chalkley RJ, Thalhammer A, Schoepfer R, Burlingame AL. Identification of protein O-GlcNAcylation sites using electron transfer dissociation mass spectrometry on native peptides. Proc Natl Acad Sci USA 2009;106:8894–9. [148] Gambetta MC, Müller J. O-GlcNAcylation prevents aggregation of the polycomb group repressor polyhomeotic. Dev Cell 2014;31:629–39. [149] Gonzalez I, Mateos-Langerak J, Thomas A, Cheutin T, Cavalli G. Identification of regulators of the three-dimensional polycomb organization by a microscopy-based genome-wide RNAi screen. Mol Cell 2014;54:485–99. [150] Kagey MH, Melhuish TA, Wotton D. The Polycomb protein Pc2 is a SUMO E3. Cell 2003;113:127–37. [151] Kang X, Qi Y, Zuo Y, Wang Q, Zou Y, Schwartz RJ, et al. SUMOspecific protease 2 is essential for suppression of polycomb group protein-mediated gene silencing during embryonic development. Mol Cell 2010;38:191–201. [152] Smith M, Mallin DR, Simon JA, Courey AJ. Small ubiquitin-like modifier (SUMO) conjugation impedes transcriptional silencing by the polycomb group repressor sex comb on midleg. J Biol Chem 2011;286:11391–400. [153] Vazquez J, Müller M, Pirrotta V, Sedat JW. The Mcp element mediates stable long-range chromosome-chromosome interactions in Drosophila. Mol Biol Cell 2006;17:2158–65. [154] Li H-B, Muller M, Bahechar IA, Kyrchanova O, Ohno K, Georgiev P, Pirrotta V. Insulators, not polycomb response elements are required for long-range interactions between polycomb targets in drosophila melanogaster. Mol Cell Biol 2011;31:616–25.

II.  Molecular Mechanisms of Epigenetics

110

7.  Polycomb Mechanisms and Epigenetic Control of Gene Activity

[155] Li H-B, Ohno K, Gui H, Pirrotta V. Insulators target active genes to transcription factories and polycomb-repressed genes to polycomb bodies. PLoS Genet 2013;9:e1003436. [156] Li L, Lyu X, Hou C, Takenaka N, Nguyen HQ, Ong C-T, et al. Widespread rearrangement of 3D chromatin organization underlies polycomb-mediated stress-induced silencing. Mol Cell 2015;58:216–31. [157] Dekker J, Marti-Renom MA, Mirny LA. Exploring the threedimensional organization of genomes: interpreting chromatin interaction data. Nat Rev Genet 2013;14:390–403. [158] Hansen KH, Bracken AP, Pasini D, Dietrich N, Gehani SS, Monrad A. A model for transmission of the H3K27me3 epigenetic mark. Nat Cell Biol 2008;10:1291–300. [159] Busturia A, Wightman CD, Sakonju S. A silencer is required for maintenance of transcriptional repression throughout Drosophila development. Development 1997;124:4343–50. [160] Zink D, Paro R. Drosophila Polycomb-group regulated chromatin inhibits the accessibility of a trans-activator to its target DNA. EMBO J 1995;14:5660–71. [161] Cavalli G, Paro R. The Drosophila Fab-7 chromosomal element conveys epigenetic inheritance during mitosis and meiosis. Cell 1998;93:505–18. [162] Cavalli G, Paro R. Epigenetic inheritance of active chromatin after removal of the main transactivator. Science 1999;286:955–8. [163] Poux S, Horard B, Sigrist CJA, Pirrotta V. The Drosophila Trithorax protein is a coactivator required to prevent re-establishment of Polycomb silencing. Development 2002;129:2483–93. [164] Klymenko T, Müller J. The histone methyltransferases Trithorax and Ash1 prevent transcriptional silencing by Polycomb group proteins. EMBO Rep 2004;5:373–7. [165] Milne TA, Briggs SD, Brock HW, Martin ME, Gibbs D, llis CD, et al. MLL targets SET domain methyltransferase activity to Hox gene promoters. Mol Cell 2002;10:1107–17. [166] Smith ST, Petruk S, Sedkov Y, Cho E, Tillib S, Canaani E, et al. Modulation of heat shock gene expression by the TAC1 chromatin-modifying complex. Nat Cell Biol 2004;6:162–7. [167] Tanaka Y, Katagiri Z, Kawahashi K, Kioussis D, Kitajima S. Trithorax-group protein ASH1 methylates histone H3 lysine 36. Gene 2007;397:161–8. [168] Orlando V, Jane EP, Chinwalla V, Harte PJ, Paro R. Binding of Trithorax and Polycomb proteins to the bithorax complex: dynamic changes during early Drosophila embryogenesis. EMBO J 1998;17:5141–50. [169] Schwartz YB, Kahn TG, Stenberg P, Ohno K, Bourgon R, Pirrotta V. Alternative epigenetic chromatin states of polycomb target genes. PLoS Genet 2010;6:e1000805. [170] Miller T, Krogan NJ, Dover J, Erdjument-Bromage H, Tempst P, Johnston M, et al. COMPASS: a complex of proteins associated with a trithorax-related SET domain protein. Proc Natl Acad Sci USA 2001;98:12902–7. [171] Smith E, Lin C, Shilatifard A. The super elongation complex (SEC) and MLL in development and disease. Genes Dev 2011;25:661–72. [172] Lee J-H, Tate CM, You J-S, Skalnik DG. Identification and characterization of the human Set1B Histone H3-Lys4 methyltransferase complex. J Biol Chem 2007;282:13419–28. [173] Herz H-M, Mohan M, Garruss AS, Liang K, Takahashi Y-h, Mickey K, et al. Enhancer-associated H3K4 monomethylation by Trithorax-related, the Drosophila homolog of mammalian Mll3/ Mll4. Genes Dev 2012;26:2604–20. [174] Wu M, Wang PF, Lee JS, Martin-Brown S, Florens L, Washburn M, et al. Molecular Regulation of H3K4 Trimethylation by Wdr82, a Component of Human Set1/COMPASS. Mol Cel. Biol 2008;28:7337–44.

[175] Petruk S, Sedkov Y, Smith S, Tillib S, Kraevski V, Nakamura T, et al. Trithorax and dCBP acting in a complex to maintain expression of a homeotic gene. Science 2001;294:1331–4. [176] Bantignies F, Goodman RH, Smolik SM. Functional interaction between the coactivator Drosophila CREB-binding protein and ASH1, a member of the trithorax group of chromatin modifiers. Mol Cell Biol 2000;20:9317–30. [177] Ernst P, Wang J, Huang M, Goodman RH, Korsmayer SJ. MLL and CREB bind cooperatively to the nuclear coactivator CREBbinding protein. Mol Cell Biol 2001;21:2249–58. [178] Young RA. Control of the embryonic stem cell state. Cell 2011;144:940–54. [179] Hu D, Garruss AS, Gao X, Morgan MA, Cook M, Smith ER, et al. The Mll2 branch of the COMPASS family regulates bivalent promoters in mouse embryonic stem cells. Nat Struct Mol Biol 2013;20:1093–7. [180] Bernstein BE, Mikkelsen TS, Xie X, Kamal M, Huebert DJ, Cuff J, et al. A bivalent chromatin structure marks key developmental genes in embryonic stem cells. Cell 2006;125:315–26. [181] Kinkley S, Helmuth J, Polansky JK, Dunkel I, Gasparoni G, Fröhler, et al. reChIP-seq reveals widespread bivalency of H3K4me3 and H3K27me3 in CD4+ memory T cells. Nat Comm 2016;7:12514. [182] Voigt P, Tee W-W, Reinberg D. A double take on bivalent promoters. Genes Dev 2013;27:1318–38. [183] Stock JK, Giadrossi S, Casanova M, Brookes E, Vidal M, Koseki H, Brockdorff N, Fisher AG, Pombo A. Ring1-mediated ubiquitination of H2A restrains poised RNA polymerase II at bivalent genes in mouse ES cells. Nat Cell Biol 2007;9:1428–35. [184] Wachter E, Quante T, Merusi C, Arczewska A, Stewart F, Webb S, et al. Synthetic CpG islands reveal DNA sequence determinants of chromatin structure. eLife 2015;3:e03397. [185] Montgomery ND, Yee D, Chen A, Kalantry S, Chamberlain SJ, Otte AP, Magnuson T. The murine Polycomb Group protein Eed is required for global histone H3 lysine-27 methylation. Curr Biol 2005;15:942–7. [186] Pasini D, Bracken AP, Hansen JB, Capillo M, Helin K. The Polycomb Group Protein Suz12 Is Required for Embryonic Stem Cell Differentiation. Mol Cell Biol 2007;27:3769–79. [187] Chamberlain SJ, Yee D, Magnuson T. Polycomb repressive complex 2 is dispensable for maintenance of embryonic stem cell pluripotency. Stem Cells 2008;26:1496–505. [188] Ezhkova E, Pasolli HA, Parker JS, Stokes N, Su Ih, Hannon G, et al. Ezh2 orchestrates gene expression for the stepwise differentiation of tissue-specific stem cells. Cell 2009;136:1122–35. [189] Feldman N, Gerson A, Fang J, Li E, Zhang Y, Sihinkai Y, et al. G9a-mediated irreversible epigenetic inactivation of Oct3/4 during early embryogenesis. Nat Cell Biol 2006;8:188–94. [190] Epsztejn-Litman S, Feldman N, Abu-Remaileh M, Shufaro Y, Gerson A, Ueda J, et al. De novo DNA methylation promoted by G9a prevents reprogramming of embryonically silenced genes. Nat Struct Mol Biol 2008;15:1176–83. [191] Jacobs JJL, Kleboom K, Marino S, DePinho RA, van Lohuizen M. The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation and senescence through the ink4a locus. Nature 1999;397:164–8. [192] Bracken AP, Kleine-Kohlbrecher D, Dietrich N, Pasini D, Gargiulo G, Beekman C, et al. The Polycomb group proteins bind throughout the INK4A-ARF locus and are disassociated in senescent cells. Genes Dev 2007;21:525–30. [193] Kheradmand Kia S, Gorski MM, Giannakopoulos S, Verrijzer CP. SWI/SNF mediates polycomb eviction and epigenetic reprogramming of the INK4b-ARF-INK4a locus. Mol Cell Biol 2008;28:3457–64.

II.  Molecular Mechanisms of Epigenetics

C H A P T E R

8 Analysis of Gene-Specific DNA Methylation Naoko Hattori, Toshikazu Ushijima National Cancer Center Research Institute, Tokyo Japan

O U T L INE Introduction

113

Principles of DNA Methylation Analysis

113

Characteristics of Individual Techniques Methylation-Sensitive Restriction Enzyme–Based Analyses Bisulfite Sequencing Combined Bisulfite Restriction Analysis Methylation-Specific PCR Real-Time MSP MethyLight and Digital MethyLight Methylation-Sensitive High-Resolution Melting Analysis

115 115 115 116 116 116 117 119

INTRODUCTION Gene- or region-specific DNA methylation analysis is necessary in various situations, and a variety of methods are available. It is important to become familiar with the characteristics of each method, including the amount of DNA required, flexibility in selection of CpG sites to analyze, degree of quantitation of the technique, technical complexity, and the cost (Table 8.1). For example, if one wants to analyze DNA methylation as a cause of gene silencing, a specific region that controls gene expression should be analyzed [1], and a method with flexibility in selecting a region to analyze should be used. If one aims for diagnostic applications, a method that is highly accurate should be adopted [2]. In this chapter, we first introduce principles of DNA methylation analysis, and then summarize characteristics Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00008-0 Copyright © 2017 Elsevier Inc. All rights reserved.

Pyrosequencing MassARRAY

119 119

Tips for Individual Methods Tips for Bisulfite-Mediated Conversion Preparation of Fully Methylated and Fully Unmethylated DNA Preparation of Standard DNA Tips for Bisulfite Sequencing Tips for MSP and Quantitative MSP

120 120

Epilogue

123

References

123

120 120 121 121

of individual methods. Finally, we provide tips necessary to perform the analyses based on bisulfite-mediated conversion, bisulfite sequencing (BS), methylation-specific PCR (MSP), and quantitative MSP.

PRINCIPLES OF DNA METHYLATION ANALYSIS DNA methylation can be analyzed based on several principles that differentially recognize 5-methylcytosine (Cm) from cytosine (C). The first principle depends upon methylation-sensitive restriction enzymes, whose activity is affected by the presence of a methyl group on a cytosine at a CpG site(s) within restriction sites (Fig. 8.1A). The vast majority of methylation-sensitive restriction enzymes, such as HpaII and SmaI, are inactive on methylated

113

114

8.  Analysis of Gene-Specific DNA Methylation

TABLE 8.1 Characteristics of Methods for Gene-Specific Methylation Analysis Amount of DNA required

Flexibility in selection of a region analyzed

Quantification

Ease of use

Cost

Applications

Southern blot hybridization

Large

Low

No

Intermediate

Low

Detection of methylated/ unmethylated status at specific CpG sites

BS

Small

High

No

Intermediate

Low

Analysis of methylation patterns on individual DNA molecules

DigitalBS

Small

High

Yes

Intermediate

Intermediate

Analysis of methylation patterns on individual DNA molecules, and measurement of actual copy number in a sample

AmpliconBS

Small

High

Yes

Intermediate

High

Analysis of methylation patterns on individual DNA molecules, and measurement of methylation levels

COBRA

Small

Low

Yes

Easy

Low

Detection of DNA molecules methylated/unmethylated at a specific CpG site

MSP

Small

High

No

Easy

Low

Detection of DNA molecules methylated/unmethylated at a specific region

Real-time MSP using SYBR Green I

Small

High

Yes

Easy

Low

Quantitative analysis of DNA molecules methylated/ unmethylated at a specific region

MethyLight

Small

Intermediate

Yes

Intermediate

Intermediate

Quantitative analysis of DNA molecules methylated/ unmethylated at a specific region

Digital MethyLight

Small

Intermediate

Yes

Intermediate

High

Quantitative analysis of DNA molecules methylated/ unmethylated at a specific region, and measurement of methylation levels

Pyrosequencing

Small

High

Yes

Intermediate

High

Quantitative methylation analysis of multiple CpG sites

MassARRAY

Small

High

Yes

Easy

High

Quantitative methylation analysis of multiple CpG sites

AmpliconBS, Amplicon bisulfite sequencing; BS, bisulfite sequencing; COBRA, combined bisulfite restriction analysis; DigitalBS, digital bisulfite sequencing; MSP, methylation-specific PCR.

CpG sites, but a unique methylation-sensitive restriction enzyme, McrBC is inactive on unmethylated CpG sites. Differential cleavage can be detected by Southern blot hybridization, gel electrophoresis, digital electrophoresis, and real-time PCR. The second principle depends on bisulfite-mediated DNA conversion. This treatment converts unmethylated C into uracil (U) very rapidly, whereas it converts methylated C extremely slowly [3]. Under optimized conditions, a difference in the methylation status of a CpG site can be converted into a difference in the sequence, UpG or CpG. Once a difference of the methylation status is converted into a difference of DNA se-

quence, it can be detected by various methods, such as BS, MSP, real-time MSP, combined bisulfite restriction analysis (COBRA), pyrosequencing, and MassARRAY analysis (Table 8.1). Third, methylated cytosines can be specifically recognized by an antimethylcytidine antibody or methylated DNA–binding (MBD) protein. After appropriate shearing of DNA, methylated DNA can be collected using these affinity methods. This principle is mainly used for genomewide screening techniques [4]. Fourth, the fraction of methylcytosine in the entire genomic DNA can be measured by HPLC or mass spectrometry [5]. As this method does not contain sequence

III.  Methods in Epigenetics



Characteristics of Individual Techniques

115

FIGURE 8.1  Principles of DNA methylation detection. (A) Detection by a methylation-sensitive restriction enzyme. Genomic DNA is digested with a methylation-sensitive restriction enzyme (HpaII) when its restriction site (CCGG) is unmethylated, but not digested when the site is methylated. Whether genomic DNA is digested or not represents the methylation status in the original DNA. (B) Detection by bisulfite-mediated DNA conversion. Unmethylated cytosines are converted very rapidly into uracil by deamination, whereas methylated cytosines are converted extremely slowly. Therefore, a difference in the methylation status of a CpG site can be converted into a difference in the sequence, UpG or CpG. After bisulfite-mediated DNA conversion, the upper and lower strands are no longer complementary. Cm, Methylated cytosine.

information, this can be used solely to measure global methylation levels.

CHARACTERISTICS OF INDIVIDUAL TECHNIQUES Methylation-Sensitive Restriction Enzyme–Based Analyses Southern blot hybridization for DNA methylation analysis is based on DNA digestion by a pair of methylation-sensitive and -insensitive restriction enzymes. Subsequently, DNA methylation levels are detected by hybridization using a probe for a specific genomic region (Southern blot analysis) [6] by electrophoresis followed by PCR (methylation-sensitive restriction fingerprinting) [7] or by real-time PCR. The methylation status of a restriction recognition site can be detected by monitoring the band positions of DNA fragments flanking the restriction sites. The advantage of this technique is that its quantitative results reflect the amounts of digested and undigested DNA molecules. Southern blot analysis is especially useful for analysis of repetitive sequences because multiple similar sequences in the genome can be analyzed by a single probe. However, methylation-sensitive restriction enzyme–based techniques analyze a limited number of CpG sites located within restriction recognition sites, and Southern blot analysis requires a large amount of high-quality DNA.

Although these techniques used to be used frequently before bisulfite conversion–based techniques became popular, they are used only occasionally now.

Bisulfite Sequencing Bisulfite-converted DNA is amplified by PCR using primers located in genomic regions lacking CpG sites. In conventional BS, the PCR product is sequenced, usually after cloning of the PCR product, and CpG sites within the amplified region are interrogated (Fig. 8.2) [8]. Cytosine (C) and thymine (T) at a CpG site in the converted DNA show methylated and unmethylated C, respectively, in the original DNA. BS has advantages: its capacity of investigating the methylation status of every single CpG site between the primers, and determining how multiple CpG sites in a single DNA molecule are methylated. However, this method suffers from technical limitations similarly to other bisulfite-mediated analyses, which will be described further, and the cumbersomeness of sequencing. Two novel ways of BS have been developed by the combination with next-generation sequencing technology and with digital PCR (dPCR) (Fig. 8.2) [9]. Combination of BS and next-generation sequencing, namely amplicon bisulfite sequencing (AmpliconBS), allows analysis of a large number of molecules and increases the accuracy of methylation levels [2,9]. In the combination with the dPCR technique, that is, digital bisulfite sequencing (DigitalBS), dPCR is performed under the

III.  Methods in Epigenetics

116

8.  Analysis of Gene-Specific DNA Methylation

FIGURE 8.2  Principles of BS, AmpliconBS, and DigitalBS. Bisulfite-converted DNA is amplified by PCR with primers covering no CpG sites (universal primers; uni-primers). For conventional BS, the PCR product is cloned, and individual clones are sequenced. For AmpliconBS, a DNA library is prepared from the amplicon, barcoding and emulsion PCR are conducted, and the produce is deeply sequenced by a next-generation sequencer. Deep sequencing enables accurate measurement of methylation levels. For DigitalBS, the PCR reaction is processed in a multiwell plate with one or no copy of template DNA under the presence of an intercalating dye, and amplified PCR products are sequenced. DigitalBS enables unbiased amplification of DNA molecules, regardless of the number of methylated CpG sites within a molecule. These methods provide information on the methylation patterns of individual DNA molecules at single-CpG resolution. Methylated and unmethylated CpG sites are shown by closed and open circles, respectively.

presence of an intercalating dye, and amplified PCR product is sequenced. DigitalBS allows unbiased amplification of any DNA molecule, regardless of methylation patterns of a molecule.

Combined Bisulfite Restriction Analysis COBRA is based on the appearance or disappearance of a restriction enzyme recognition site after bisulfite conversion (Fig. 8.3) [10]. By quantifying the ratio of digested and undigested PCR products, the ratio of methylated and unmethylated DNA molecules can be quantified. This technique has the advantages of quantitative measurement of methylation levels and ease of procedure. However, this method is limited to the analysis of only one CpG site. However, as multiple CpG sites within a small genomic region tend to be coordinately methylated or unmethylated [5,11], analysis of a single CpG site can predict the methylation status of the surrounding region. A modified protocol for COBRA, Bio-COBRA, was therefore developed [12]. Bio-COBRA incorporates an

FIGURE 8.3  Principles of COBRA. Bisulfite-converted DNA is amplified by PCR with primers covering no CpG sites, and the PCR product is digested with a restriction enzyme (TaqI). In the COBRA assay shown here, if the cytosine in the CpG site is methylated, the restriction site will remain. On the other hand, if the site is unmethylated, the restriction site will be lost. Quantitative analysis of methylation levels is achieved by subsequent gel electrophoresis and measurement of cleaved and uncleaved bands.

electrophoresis step of the digested PCR product in a microfluidics chip, such as Bioanalyzer or TapeStation (Agilent), and provides a rapid and quantitative assessment of DNA methylation statuses in a large sample set.

Methylation-Specific PCR This method interrogates methylation statuses of several CpGs at primer sites by performing PCR with primers specific to methylated or unmethylated sequences and observing the presence or absence of a PCR product (Fig. 8.4A) [13]. If both forward and reverse primer regions are methylated, intervening CpG sites are also likely to be methylated. DNA molecules with mosaic methylation patterns at primer sites are not amplified. This method has advantages of high flexibility in selecting a genomic region to analyze because PCR primers can be designed at arbitrary positions, even if the region to be analyzed is CpG-rich, and technical simplicity. However, MSP can easily produce false-positive and -negative results. Therefore, it is critically important to use the optimal number of PCR cycles and annealing temperatures with appropriate negative controls, which will be described in the section “Tips for MSP and Quantitative MSP.”

Real-Time MSP Real-time MSP is performed by real-time detection of a MSP product (Fig. 8.4B). Real-time MSP has shared advantages with MSP, such as flexibility in selecting a genomic

III.  Methods in Epigenetics



Characteristics of Individual Techniques

117

FIGURE 8.4  Principles of MSP, quantitative MSP, and MethyLight. (A) Methylation statuses at several CpGs within primer sequences are interrogated by performing PCR with primers specific to methylated or unmethylated templates (M-primer and U-primer) and monitoring the presence or absence of a PCR product. PCR conditions needs to be optimized using fully methylated DNA and fully unmethylated DNA. Methylated and unmethylated CpG sites are shown by closed and open circles, respectively. (B) The number of methylated and unmethylated DNA molecules by MSP can be quantified using intercalating dye (qMSP) or a TaqMan probe (MethyLight). Full M, Fully methylated DNA; Full U, fully unmethylated DNA.

region to analyze and suitability for analysis of a large number of samples as MSP, and its unique advantage of accurate and sensitive assessment of DNA methylation levels. On the other hand, PCR products are detected by an intercalating dye, such as SYBR Green I, and it can even detect nonspecifically amplified DNA and primer dimers. Therefore, confirmation of specific amplification by melting analysis of the PCR product is essential. It is reported that the use of a new fluorescent dye, such as SYTO-82, can produce more accurate melting results [14].

MethyLight and Digital MethyLight MSP products can also be detected using a TaqMan probe (MethyLight) [15]. Advantage of MethyLight is its higher specificity than real-time MSP (Fig. 8.5B), as a TaqMan probe anneals only to a specific sequence (methylated or unmethylated sequence). Potential disadvantage is

the high cost of a TaqMan probe. In real-time MSP and MethyLight, DNA methylation level can be quantified in two manners. Absolute methylation level can be quantified by comparing the amplification curve of a test sample with those of standard samples containing known numbers of DNA molecules (Fig. 8.5A). Alternatively, the relative methylation level can be estimated by analyzing sample DNA and fully methylated DNA (control) for target and reference loci (Fig. 8.5B). Using universal primers (and a probe), the amounts of sample DNA and control DNA are normalized. Under this condition, methylationspecific primers (and a probe) can reveal the fraction of methylated DNA molecules in a sample. MethyLight can be used to detect a PCR product by dPCR, and the combined method is referred to as digital MethyLight [9,16,17]. In the dPCR, bisulfite-treated DNA is diluted and dispensed over multiple wells so that each well will contain one or no template DNA molecule, and

III.  Methods in Epigenetics

118

8.  Analysis of Gene-Specific DNA Methylation

FIGURE 8.5  Strategies for quantifying methylation levels. (A) Absolute quantification using standard DNA samples. Sample DNA, control DNA (fully methylated DNA or fully unmethylated DNA), and standard DNA are amplified with M-primer or with U-primer. An accurate number of methylated or unmethylated DNA molecules can be quantified by comparing the amplification curve of sample DNA with those of standard DNA containing known numbers of DNA molecules. Methylation level is calculated using the following formula; methylation level (%) = 100 × (number of methylated DNA molecules)/(number of methylated DNA molecules + number of unmethylated DNA molecules). (B) Relative quantification using a reference locus and control DNA. The amplification curves show representative MethyLight data. The amounts of sample DNA and control DNA (fully methylated DNA) are normalized by universal primers (uni-primer) and a probe. Under this condition, the amounts of methylated DNA are compared between the sample and control using methylation-specific primers (M-primer) and a probe. A ∆∆Ct value is obtained as shown and the relative methylation level can be calculated. Methylated and unmethylated CpG sites are shown as closed and open circles, respectively. III.  Methods in Epigenetics



Characteristics of Individual Techniques

119

the product is detected by MethyLight. The copy numbers of methylated and unmethylated DNA can be calculated based upon the dilution and the final template DNA molecules detected. Advantage of digital MethyLight is absolute quantification without calibration using a standard DNA samples. Digital MethyLight is resistant to background noise and PCR contaminants, and is considered promising for the analysis of biological fluid samples, such as blood and urine.

Methylation-Sensitive High-Resolution Melting Analysis High-resolution melting (HRM) analysis detects the GC content of a DNA fragment by monitoring its melting curve over gradually increasing temperatures [18]. For methylation-sensitive high-resolution melting (MS-HRM) analysis, bisulfite-modified DNA is amplified using primers common to the methylated and unmethylated DNA (primers not containing a CpG site), and the difference in the numbers of methylated CpG sites is converted into the difference in GC contents of the PCR product, specifically that in the melting curves. The methylation level of a sample can be quantified by comparing the melting profile of its PCR product with those of standard DNA samples consisting of 0, 20, 40, 60, 80, and 100% methylated DNA. The advantages of this method include the fast and simple procedure with high reproducibility and capability of analyzing multiple CpG sites within a region. However, careful optimization of the PCR conditions is required for methylation-sensitive high-resolution melting analysis because the sensitivity is strongly affected by the primer sequences and annealing temperature [19].

Pyrosequencing Pyrosequencing detects methylation levels of individual CpG sites in a PCR product obtained by primers common to methylated and unmethylated sequences after bisulfite conversion. The amounts of C and T at individual sites are converted into the amounts of pyrophosphates released using the primer-extension method, and their amounts are accurately quantified bioluminometrically using the Pyrosequencer system (Qiagen) (Fig. 8.6). The advantages of pyrosequencing are its accurate quantitative results and ease of daily procedure. However, the design of suitable primers that do not include CpG sites and do not have PCR bias is difficult depending upon the local sequence, and an instrument specifically designed for this analysis is required.

MassARRAY MassARRAY detects methylation levels of multiple CpG sites in a PCR product of bisulfite-treated DNA

FIGURE 8.6  The principle of pyrosequencing. C/T polymorphisms in the PCR product are investigated by measuring pyrophosphate released at individual sites. The amount of pyrophosphate is converted into a light signal, and then shown as a pyrogram.

by detecting the difference in the masses of cytosine and thymine produced as follows. In this technique, bisulfite-modified DNA is amplified by PCR using a forward primer and T7 promoter-coupled reverse primer, both of which are common to methylated and unmethylated sequences. The PCR product is transcribed in vitro using the T7 promoter, and the in vitro transcript is cleaved into short fragments, regardless of the methylation statuses, using RNase A (Fig. 8.7). Although RNase A cleaves at pyrimidine bases (uracil and cytosine), uracil-specific cleavage is achieved by substituting rCTP with dCTP during the in vitro transcription. The cleaved fragments are analyzed by matrix-assisted laser desorption/ionization time-of-flight (MALDITOF) mass spectrometry, and a difference in the mass (16 Da) of product with C (methylated in the original DNA) and that with T (unmethylated) is detected. The advantages of MassARRAY include its quantitative measurement of DNA methylation statuses of multiple CpG sites and its capacity to analyze a large number of samples. However, MassARRAY requires costly specialized instrument.

III.  Methods in Epigenetics

120

8.  Analysis of Gene-Specific DNA Methylation

be selected by assessing both the maintenance of DNA integrity (number of DNA molecules remaining as a PCR template) and conversion efficiency (rate of conversion of unmethylated cytosines to uracils) after the reaction. To estimate the conversion ratio, BS is recommended. We can sequence a number of clones and count the number of unconverted cytosines at non-CpG sites. We can analyze 500 cytosines by sequencing 10 clones of a PCR product with 50 cytosines at non-CpG sites (typically 200–300 bp) and assess conversion rate at the level of 0.2%. To estimate the degree of DNA degradation, analysis of the number of DNA molecules amplified by real-time PCR is recommended. If we use 10 ng of DNA (corresponding to 2800 molecules of template DNA) for real-time PCR, we ideally detect 280 molecules (10%) after bisulfite-mediated conversion. FIGURE 8.7  The Principle of MassARRAY. The PCR product amplified from bisulfite-converted DNA is transcribed in vitro, and cleaved by RNase A. A difference of C and T is detected as a difference in the mass of 16 Da by matrix-assisted laser desorption/ionization time-of-flight mass spectrometry, and differences of multiple CpG sites in a fragment can be detected.

Preparation of Fully Methylated and Fully Unmethylated DNA

We will first introduce tips for bisulfite-mediated conversion, preparation of control DNA for this method, and preparation of standard DNA because these procedures are common and important for all the analyses based on bisulfite-mediated DNA conversion. Next, specific tips for BS and MSP are summarized, as they are the most popular methods because of their simplicity and cost effectiveness.

PCR conditions, including an optimal primer set and an optimal annealing temperature, should be determined using fully methylated and fully unmethylated DNA controls [22]. The fully unmethylated DNA can be prepared by amplifying normal cell DNA, such as blood DNA from a healthy individual, with a GenomiPhi DNA amplification kit (GenomiPhi). The fully methylated DNA can be prepared by treatment of normal cell DNA with SssI methylase. However, genomic regions amplified by GenomiPhi are biased. If copy numbers of various genomic regions need to be matched in fully unmethylated and fully methylated DNAs, the fully methylated DNA should be prepared by methylating DNA amplified by GenomiPhi with SssI methylase.

Tips for Bisulfite-Mediated Conversion

Preparation of Standard DNA

During bisulfite-mediated conversion, DNA degradation is induced and the number of DNA molecules that can serve as a PCR template decreases to 5%–10% of the DNA before treatment [20]. Therefore, the most important issues in bisulfite-mediated conversion are: (1) to avoid DNA degradation as much as possible and (2) to achieve complete conversion of unmethylated cytosine to uracil. Although bisulfite conversion reactions were conducted at 55°C for 16 h more than 10 years ago [21], it is now recognized that reactions at a high temperature for a shorter period are more effective for complete conversion and alleviation of DNA degradation. Multiple companies provide kits for DNA denaturation, bisulfite conversion, DNA purification, and desulfonation, and emphasize efficiency in conversion. Nevertheless, the most suitable kit should

To estimate the degree of DNA degradation and to quantify DNA methylation levels by real-time MSP, standard DNA with known numbers of DNA molecules is necessary (Fig. 8.5A), and this can be prepared in two ways. First, the PCR product can be purified by a gel-filtration column to remove unused nucleotides and primers. Second, the PCR product is cloned into a plasmid, and the plasmid is linearized by a restriction enzyme. As the molecular weight of the PCR product or the plasmid with the insert can be calculated, the number of DNA molecules in a sample can be calculated from its DNA concentration. Preparation of standard DNA by cloning a PCR product has the advantage of accuracy and availability of a large amount of standard DNA, but has the disadvantage of requiring time for cloning and confirmation of an insert.

TIPS FOR INDIVIDUAL METHODS

III.  Methods in Epigenetics



Tips for Individual Methods

121

Tips for Bisulfite Sequencing

Tips for MSP and Quantitative MSP

BS is capable of analyzing detailed DNA methylation patterns of individual DNA molecules in a given region of the genome. It also provides quantitative information on the ratio of methylated and unmethylated DNA molecules. Especially, AmpliconBS has enabled this with ease. At the same time, caution must be exercised to obtain unbiased amplification of both unmethylated and methylated DNA molecules.

MSP is flexible in selecting regions for analysis and can be performed with ease and at a low cost. Quantitative MSP provides accurate, sensitive, and quantitative assessment of DNA methylation levels. Under good conditions, DNA methylation levels obtained by real-time MSP have a variation ≤20% of the mean methylation level. To maximize these advantages, there are some tips for conducting MSP and quantitative MSP.

PCR Conditions for Unbiased Amplification It is well known that, depending upon PCR conditions, there can be a PCR bias that leads to preferential amplification of either unmethylated or methylated DNA [22,23]. In most cases, unmethylated DNA is preferentially amplified, but methylated DNA can be preferentially amplified with specific primers [23]. To avoid this PCR bias, a PCR condition that equally amplifies fully methylated and fully unmethylated DNA controls should be established by selecting an optimal primer set and an optimal annealing temperature (Fig. 8.8A) [22]. When accurate estimation of the ratio of methylated and unmethylated DNA is necessary, control DNA containing an equal number of fully methylated and unmethylated DNA molecules should be prepared by mixing such DNA, and simultaneously analyzed to obtain a ratio of 50% (Fig. 8.8B).

Primer Design A genomic region to be analyzed need to be carefully selected, as in other analyses, and primers specific to methylated or unmethylated DNA should be designed within the region. The 3′ end of a primer should be located at a polymorphic C/T site, and multiple CpG sites should be located near the 3′ end (Fig. 8.9A). Difficulty in designing primers specific to unmethylated DNA is frequently encountered, and use of the other DNA strand (bottom strand) is often helpful. Freely available software or websites assist the designing of MSP primers, for example, Methyl Primer Express (Applied Biosystems), BiSearch (http://bisearch.enzim.hu), and MethPrimer (http:// www.urogene.org/cgi-bin/methprimer/methprimer.cgi).

PCR Cycles to Avoid Artifacts Even if optimal PCR conditions are used, PCR cycles should be minimized as long as a sufficient amount of a PCR product for cloning is obtained. Excessive PCR cycles cause denaturation of the PCR product in the absence of Taq polymerase activity, and produce the amplification of chimeric products and even PCR products that were not present in the template DNA. Excessive PCR cycles also exaggerate the difference of PCR efficiency between methylated and unmethylated DNA.

PCR Conditions for Specific Amplification The annealing temperature and magnesium concentration should be optimized using fully methylated and fully unmethylated DNA controls. A good condition for primers specific to methylated DNA amplifies fully methylated DNA efficiently, but not fully unmethylated DNA (Fig. 8.9B). A good condition for primers specific to unmethylated DNA amplifies fully unmethylated DNA efficiently, but not fully methylated DNA. In the case of real-time MSP, an optimal condition can be determined by the amplification curve and the melting curve (Fig. 8.9C). The amplification curve under

FIGURE 8.8  Optimization of BS. (A) Optimization of the primer sets for BS. The influence of primers on PCR efficiency was examined using fully methylated DNA and fully unmethylated DNA by real-time PCR. (B) Confirmation of unbiased amplification. Methylated and unmethylated cytosines are shown by closed and open circles, respectively. The proportion of methylated clones was 40%, indicating that appropriate PCR conditions and unbiased amplification were achieved.

III.  Methods in Epigenetics

122

8.  Analysis of Gene-Specific DNA Methylation

FIGURE 8.9  Optimization of real-time PCR conditions. (A) Primer design for MSP and real-time MSP. Primers specific to methylated and unmethylated DNA (M- and U-primers, respectively) should contain multiple CpG sites near and at their 3′ ends, and are desirably located in the same locations. (B) Optimization of the annealing temperature for MSP. For the M-primer, annealing temperatures of 57 and 60°C was considered to be optimal because fully methylated DNA was amplified with good efficiency, but fully unmethylated DNA was not amplified at all. For the U-primer, only an annealing temperature of 57°C was considered to be optimal because fully methylated DNA was efficiently amplified at 54 and 57°C, but fully methylated DNA was also amplified at 54°C. (C) Optimization of the annealing temperature for real-time MSP. The real-time PCR amplification curve showed high PCR efficiency under annealing temperatures of 54, 57, and 60°C. The melting curve showed a single peak, thus specific amplification, under annealing temperatures of 54, 57, and 60°C. If multiple optimal annealing temperatures are available, a higher temperature is preferred for specificity. Optimal conditions in real-time MSP are often different from those in MSP due to the presence of an intercalating agent, even if the same primer set is used.

good conditions shows a steep rise at an early PCR cycle, and a flat plateau of target methylated (or unmethylated) DNA. The melting curve of the target DNA should show a single sharp peak. Nontarget DNA should be not be amplified in early cycles, and PCR products produced in late cycles, if any, should not overlap with the melting curve of the target DNA.

Quantity of Template DNA Both MSP and real-time MSP can achieve high sensitivity, such as detecting one methylated DNA molecule among 1000 molecules. However, as mentioned earlier, the number of DNA molecules that can serve as a PCR template decreases to 5%–10% during bisulfite-mediated conversion, caution must be exercised so that sufficient

III.  Methods in Epigenetics



REFERENCES

copies of template DNA are present in a PCR solution. Supposing that one human haploid genome weighs 3.6 pg and that 10% of the DNA molecules are recovered as a template for PCR after bisulfite-mediated conversion, only 28 molecules are available for PCR in 1 ng of genomic DNA before bisulfite treatment. To achieve a sensitivity of 1%, 1000 molecules (10 methylated molecules) in a PCR solution is necessary, and this corresponds to 36 ng DNA before bisulfite treatment.

EPILOGUE Regional DNA methylation analysis is useful not only for research, but also for diagnostic purposes. Selecting an appropriate technique and conducting experiments under optimal conditions are required to obtain reliable data. We hope that this chapter will help investigators select appropriate techniques.

References [1] Ushijima T. Detection and interpretation of altered methylation patterns in cancer cells. Nat Rev Cancer 2005;5:223–31. [2] BLUEPRINT Consortium. Quantitative comparison of DNA methylation assays for biomarker development and clinical applications. Nat Biotechnol 2016;34(7):726–37. [3] Hayatsu H, Wataya Y, Kazushige K. The addition of sodium bisulfite to uracil and to cytosine. J Am Chem Soc 1970;92:724–6. [4] Rauch T, Pfeifer GP. Methylated-CpG island recovery assay: a new technique for the rapid detection of methylated-CpG islands in cancer. Lab Invest 2005;85:1172–80. [5] Kaneda A, Tsukamoto T, Takamura-Enya T, Watanabe N, Kaminishi M, Sugimura T, et al. Frequent hypomethylation in multiple promoter CpG islands is associated with global hypomethylation, but not with frequent promoter hypermethylation. Cancer Sci 2004;95:58–64. [6] Bird AP, Southern EM. Use of restriction enzymes to study eukaryotic DNA methylation: I. The methylation pattern in ribosomal DNA from Xenopus laevis. J Mol Biol 1978;118:27–47. [7] Huang TH, Laux DE, Hamlin BC, Tran P, Tran H, Lubahn DB. Identification of DNA methylation markers for human breast carcinomas using the methylation-sensitive restriction fingerprinting technique. Cancer Res 1997;57:1030–4. [8] Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res 1994;22:2990–7.

123

[9] Weisenberger DJ, Trinh BN, Campan M, Sharma S, Long TI, Ananthnarayan S, et al. DNA methylation analysis by digital bisulfite genomic sequencing and digital MethyLight. Nucleic Acids Res 2008;36:4689–98. [10] Xiong Z, Laird PW. COBRA: a sensitive and quantitative DNA methylation assay. Nucleic Acids Res 1997;25:2532–4. [11] Kaneda A, Kaminishi M, Yanagihara K, Sugimura T, Ushijima T. Identification of silencing of nine genes in human gastric cancers. Cancer Res 2002;62:6645–50. [12] Brena RM, Auer H, Kornacker K, Plass C. Quantification of DNA methylation in electrofluidics chips (Bio-COBRA). Nat Protoc 2006;1:52–8. [13] Herman JG, Graff JR, Myohanen S, Nelkin BD, Baylin SB. Methylation-specific PCR: a novel PCR assay for methylation status of CpG islands. Proc Natl Acad Sci USA 1996;93:9821–6. [14] Gudnason H, Dufva M, Bang DD, Wolff A. Comparison of multiple DNA dyes for real-time PCR: effects of dye concentration and sequence composition on DNA amplification and melting temperature. Nucleic Acids Res 2007;35:e127. [15] Eads CA, Danenberg KD, Kawakami K, Saltz LB, Blake C, Shibata D, et al. MethyLight: a high-throughput assay to measure DNA methylation. Nucleic Acids Res 2000;28:E32. [16] Redshaw N, Huggett JF, Taylor MS, Foy CA, Devonshire AS. Quantification of epigenetic biomarkers: an evaluation of established and emerging methods for DNA methylation analysis. BMC Genomics 2014;15:1174. [17] Wiencke JK, Bracci PM, Hsuang G, Zheng S, Hansen H, Wrensch MR, et al. A comparison of DNA methylation specific droplet digital PCR (ddPCR) and real time qPCR with flow cytometry in characterizing human T cells in peripheral blood. Epigenetics 2014;9:1360–5. [18] Wojdacz TK, Hansen LL. Reversal of PCR bias for improved sensitivity of the DNA methylation melting curve assay. Biotechniques 2006;41:274–8. [19] Wojdacz TK, Dobrovic A. Methylation-sensitive high resolution melting (MS-HRM): a new approach for sensitive and high-throughput assessment of methylation. Nucleic Acids Res 2007;35:e41. [20] Munson K, Clark J, Lamparska-Kupsik K, Smith SS. Recovery of bisulfite-converted genomic sequences in the methylationsensitive QPCR. Nucleic Acids Res 2007;35:2893–903. [21] Grunau C, Clark SJ, Rosenthal A. Bisulfite genomic sequencing: systematic investigation of critical experimental parameters. Nucleic Acids Res 2001;29:E65. [22] Warnecke PM, Stirzaker C, Song J, Grunau C, Melki JR, Clark SJ. Identification and resolution of artifacts in bisulfite sequencing. Methods 2002;27:101–7. [23] Warnecke PM, Stirzaker C, Melki JR, Millar DS, Paul CL, Clark SJ. Detection and measurement of PCR bias in quantitative methylation analysis of bisulphite-treated DNA. Nucleic Acids Res 1997;25:4422–6.

III.  Methods in Epigenetics

C H A P T E R

9 Methods for Assessing DNA Cytosine Modifications Genome-Wide Tibor A. Rauch*, Gerd P. Pfeifer** *Rush University Medical Center, Chicago, IL, United States; **Center for Epigenetics, Van Andel Research Institute, Grand Rapids, MI, United States

O U T L I N E Introduction Techniques Based on Methylation-Sensitive Restriction Endonucleases and PCR Targeted and Whole Genome Bisulfite Sequencing Infinium Methylation 450 Bead Chip Other Sodium Bisulfite-Based Approaches Methylated DNA Immunoprecipitation

126 127 128 128 128

INTRODUCTION Historically, DNA methylation has been the first epigenetic modification to be discovered [1]. The processes governing enzymatic DNA methylation, as well as the biological properties of this modification are quite well understood. A functional role of DNA methylation at CpG sequences in epigenetic control, gene regulation, X chromosome inactivation and cell differentiation was first proposed in 1975 [2,3]. Today, the connections between DNA methylation, gene activity and other epigenetic marks, such as histone modifications are being studied in many laboratories. Aberrations in DNA methylation patterns as a cause or consequence of diseases, such as cancer, are analyzed and described in detail. To explore DNA methylation profiles at the genome scale, a wide range of approaches have been developed. Most of the methods were originally used for detecting methylation changes at the single gene level but by coupling them with microarray or high throughput sequencing (HTS) platforms, genome-wide analysis tools have become available. On microarray platforms, proHandbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00009-2 Copyright © 2017 Elsevier Inc. All rights reserved.

MBD Protein-Based Affinity Pulldown Methylated-CpG Island Recovery Assay 5-Hydroxymethylcytosine Mapping Methodologies TET-Assisted Bisulfite Sequencing Oxidative Bisulfite Sequencing Future Directions and Challenges

125

References

129 129 130 131 131 131 132

moter, or CpG island arrays have often been used to analyze important regulatory regions. Tiling arrays can be used to investigate segments of specific chromosomes or the entire genome. Whole genome HTS is now a widely used approach for exploring DNA methylation profiles. Most DNA methylation analysis methods can be categorized into well-characterized groups on the basis of their principles (Fig. 9.1). Earlier methods are based on restriction endonucleases that possess altered sensitivity toward methylated cytosine residues present at the cleavage site. In this way the restriction endonuclease digestion pattern depends on the methylation status of the cleavage sites and ultimately reflects methylation profiles of the given chromosomal region. Other techniques are based on antibodies or proteins that bind to methylated DNA. Resolution at the single nucleotide level often requires bisulfite-based DNA sequencing approaches for which high throughput techniques have been developed. In this review, we will describe several of the more commonly used genomescale methods for DNA methylation analysis in some detail.

125

126

9.  Methods for Assessing DNA Cytosine Modifications Genome-Wide

FIGURE 9.1  Methods for genome-scale analysis of DNA methylation and DNA hydroxymethylation profiles. The major methods discussed in this chapter are subdivided into restriction enzyme-based, sodium bisulfite-dependent and affinity-based approaches.

The first edition of this handbook provided a detailed overview of the development of genome-scale DNA methylation analysis methods, but new technical advances have occurred in this field, and a number of earlier methods including restriction landmark genomic scanning (RLGS), methylation-sensitive restriction fingerprinting (MSRF), methylation-sensitive representational difference analysis (MS-RDA), and methylation-specific digital karyotyping (MSDK) are not commonly used anymore. Newer methodologies including Infinium CpG arrays, reduced representation bisulfite sequencing (RRBS) and methods for detection of 5-hydroxymethylcytosine have been added.

Techniques Based on Methylation-Sensitive Restriction Endonucleases and PCR One of the earliest restriction enzyme-based approaches developed is differential methylation hybridization (DMH), which allows the simultaneous determination of the methylation status of a large number of CpG-island loci [4]. CpG island-containing DNA fragments are gridded on high-density arrays, genomic DNA from the tissues of interest is digested with methylation-sensitive enzymes, and digestion products are used as templates for PCR after ligation of linkers. The resulting fragments are used as probes to screen for hypermethylated sequences, for example in cancer tissues. A more recently developed related method is the HELP assay (HpaII tiny fragment enrichment by ligation-mediated PCR), which involves cohybridization of the DNA samples to a genomic DNA microarray after cutting with a methylation-sensitive restriction enzyme or its methylation-insensitive isoschizomer [5]. A variation of this approach can be used in combination with highthroughput DNA sequencing [6]. Methylated CpG island amplification coupled microarray (MCAM) also is based on methylation-sensitive restriction enzymes and PCR using flanking primers

followed by sequence analysis or microarray probing. MCAM is a useful approach for simultaneous identification of differentially methylated genomic regions [7,8] (Fig. 9.2). Initially, the genomic DNA is cleaved with a methylation- sensitive enzyme (e.g., SmaI: 5′-CCCGGG-3′) that creates blunt ends and in this way contributes to the elimination of unmethylated sequences from the subsequent adaptor ligation step. The second enzyme XmaI recognizes the same sequence independently of methylation status and creates sticky ends that makes the appropriate adaptor ligation possible. The adaptor-ligated methylated fraction is PCR-amplified, labelled with fluorescent dyes and hybridized onto a microarray platform. Performing MCAM on normal and disease-derived tissue can easily reveal a number of disease-associated DNA methylation changes [9,10] (Fig. 9.2).

FIGURE 9.2  The main steps of methylated CpG island amplification coupled microarray (MCAM). Details are in the text. Arrows point to SmaI and XmaI cleavage sites.

III.  Methods in Epigenetics



Introduction

Variations (e.g., application of different enzymes) of the outlined restriction enzyme-based approaches have been used but all of them are based on similar principles. These methods can be combined with high-throughput sequencing [11]. The main drawback of the methylationsensitive cleavage-dependent methods is that they can provide only limited methylation profile analysis since there is no restriction enzyme that cleaves appropriately within all genomic loci.

Targeted and Whole Genome Bisulfite Sequencing The set of techniques to be described next makes use of the differential sensitivity of cytosine and 5-methylcytosine (5mC) towards chemical modification. This methodology is challenging when applied to the whole genome mostly at the level of bioinformatics analysis and cost per sample. The basis of sodium bisulfite sequencing is the fact that cytosine is deaminated to uracil by sodium bisulfite but 5-methylcytosine is resistant to bisulfite-induced deamination [12–14]. Bisulfite sequencing provides single base resolution for analysis of DNA methylation patterns [15]. It is based on sequencing of PCR products of bisulfite-treated DNA to profile DNA methylation at specific loci or along entire segments of chromosomes as reported in the first larger scale application of this technique [16]. The latter project analyzed DNA methylation patterns across human chromosomes 6, 20, and 22 in 12 human tissues, covering 22% of the CpGs on these chromosomes. Ideally, one would like to apply this approach to whole genome methylation profiling. The current approaches for high throughput sequence analysis, for example, on the Roche 454, Illumina and ABI/SOLiD platforms, have provided technology for unprecedented large-scale analysis on a rapid time scale and for a reasonable and still declining cost. Whole genome bisulfite sequencing (WGBS, also referred to as “Methyl-Seq”) has been initially accomplished for the relatively small genome of the plant Arabidopsis thaliana [17] and more recently, for the genome of various mouse and human cell types [18–20]. The DNA is fragmented with sonication to 150–200 bp and cytosine-methylated adapters are ligated before the bisulfite-conversion. The bisulfitetreated DNA molecules are enriched by PCR initiated from the adapters and then subjected to sequencing. The sequencing reads are then processed and aligned to “computationally bisulfite-converted” genomes. Due to the nature of the initial cytosine deamination step, both strands of each locus are no longer complementary and are sequenced separately. To get quantitative information on the percent methylation at individual CpG sites on each strand, a read coverage of at least 20 or 30× is generally recommended. The current cost of sequencing

127

one sample by WGBS is approximately US $4000, still putting larger scale projects using this technology out of reach for most laboratories. Researchers have developed the RRBS method that allows one to analyze methylation profiles at the single nucleotide level in a more limited way, mostly focusing on CpG islands [21] (Fig. 9.3). The original RRBS protocol was combined with Sanger sequencing [22], which was later replaced by HTS. As a first step, genomic DNA is digested with an enzyme (e.g., MspI, 5′-CCGG-3′) that mainly targets CpG-rich regions, which is followed by adapter ligation and fragment size selection. Fragments in the 150–225 bp size range are isolated and subjected to bisulfite conversion. Bisulfite-treated samples are PCR amplified and sequenced [23]. Although MethylSeq/WGBS has a much greater genome-wide coverage of CpGs compared to RRBS, RRBS provides a good coverage of CpG islands and is substantially more costeffective [11]. Other more gene-specifically targeted approaches are based on the enrichment of specific genomic regions using gene-specific probes. One method specifically captures a subset of genomic targets for single-molecule

FIGURE 9.3  Reduced representation bisulfite sequencing (RRBS). MspI cleavage enriches genomic fragments from CpG islands and other CpG-rich regions. See text for details.

III.  Methods in Epigenetics

128

9.  Methods for Assessing DNA Cytosine Modifications Genome-Wide

­bisulfite sequencing [24,25]. For example, a set of ∼30,000 padlock probes was designed to assess methylation of ∼66,000 CpG sites within 2,020 CpG islands on human chromosomes 12, 20, and 34 [25]. Other genomic enrichment approaches for targeted bisulfite sequencing have been developed [26]. Locus-specific HTS-based bisulfite deep sequencing analysis is becoming increasingly useful for detection of rare methylated alleles in a population of otherwise unmethylated sequences [27,28].

Infinium Methylation 450 Bead Chip The Illumina Infinium array is a company-developed array platform (Fig. 9.4) [29]. Using the Infinium 450k methylation assay, one can assess the methylation status of ∼480,000 cytosines at CpG sequences, which are selected primarily from promoter regions, first exons, gene bodies, and 3′UTRs. A new generation of this platform, the Epic array, incorporates additional sequences of the genome that are of particular interest, including a large number of enhancer sequences to bring the total number of analyzed CpG sites to ∼850,000 [30]. Genomic DNA is bisulfite-treated before hybridizing onto the microarray chip. The design of oligonucleotides is based on the “all-or-none” concept, surmising that the methylation status is correlated in a 50 bp stretch of DNA, that is, most neighboring CpGs are thought to be in either a methylated or unmethylated state. The

method is based on gene-specific and methylationdependent single nucleotide primer extension on the bisulfite-converted DNA using primer extension that distinguishes between unmethylated and methylated CpGs after bisulfite conversion. Single-base extension of the probes incorporates a labeled ddNTP, which is subsequently detected by array scanning. The level of methylation for the interrogated locus can be determined by calculating the ratio of the fluorescent signals from the methylated versus the unmethylated state. This technology is now used very commonly and there are large numbers of publications in the literature. Even though the genomic CpG coverage is still only ∼3% of all CpGs (28 million on each strand), the design of the probes permits a focused analysis of many genomic regions of interest.

Other Sodium Bisulfite-Based Approaches After conversion of genomic DNA with sodium bisulfite, medium-throughput approaches have been developed to analyze methylation profiles of a number of genes simultaneously. In the mass spectrometry-based MALDI-TOF MS (Sequenom approach), bisulfite-pretreated genomic DNA is used as a template for PCR amplification of specific genes [31]. One of the PCR primers is designed in such a way that a T7 RNA polymerase promoter is attached to the 5′ end. By using this promoter the amplicon is transcribed in vitro into single stranded RNA and the RNA is subsequently digested with ribonuclease A at uracil residues. MALDI-TOF mass spectrometry of the cleavage products can provide quantitative data regarding the methylation status of most CpG dinucleotides in the tested region because (1) the endonuclease cleavage occurs quantitatively so that the potential fragment variants are predictable, (2) the methylation state-related sequence alteration from C to T (or G to A on the opposite strand) yields a 16-Da mass difference (Fig. 9.5). MALDI-TOF MS is a mediumthroughput approach that provides direct quantitative methylation data for many genes simultaneously [32]. The analysis is usually conducted in multi-well format and is limited by the number of gene targets that are amplified and analyzed in each run.

Methylated DNA Immunoprecipitation

FIGURE 9.4  Infinium bead arrays. Gene-specific primers are extended to reveal the methylation status of each locus after bisulfite conversion.

Affinity purification is the principle of the third group of techniques applied to genome-wide DNA methylation mapping. Methylated DNA fragments can be affinity purified by using either a 5mC-specific antibody or proteins that specifically bind to methylated DNA sequences [33–36]. Conversely, DNA fragments containing unmethylated CpG sites can be enriched by using specific protein domains, for exmaple, the CXXC domain

III.  Methods in Epigenetics



Introduction

129

FIGURE 9.6  Methylated DNA immunoprecipitation (MeDIPseq). Delineation of the MeDIP method is in the text. Filled and open lollipops mark methylated and unmethylated CpG sites, respectively.

MBD Protein-Based Affinity Pulldown FIGURE 9.5  MALDI-TOF MS (“Sequenom” approach). The principle of MALDI-TOF MS methylation detection, which is based on bisulfite conversion, is outlined in the text. Filled and open lollipops mark methylated and unmethylated CpG sites, respectively. Boldfaced and colored letters help in following the nucleotide changes through the procedure.

that have a selective affinity for unmethylated CpGcontaining DNA [37]. A 5-methylcytosine-specific antibody has been raised and used for immunoprecipitation of densely methylated DNA sequences [35] (Fig. 9.6). In this method, the fragments generated by sonication of genomic DNA are incubated with anti-5mC antibody and the captured methylated fraction is extracted from the reaction by using Protein A/G beads. Isolated DNA fragments are deproteinized, fluorescently labeled and hybridized onto different microarray platforms or analyzed by massively parallel sequencing [38]. The main limitation of the Methylated DNA Immunoprecipitation (MeDIP) method is the quality of 5mC antibodies used in the procedure and the requirement of single-stranded (denatured) DNA for analysis. Methylation profiles can be generated with a resolution of approximately 100 base pairs. Like all affinity-based methods, MeDIP cannot measure absolute methylation levels, i.e. the percent methylation at CpG sites (0–100) within a particular region. It can only provide relative levels of methylation as indicated by different peak heights between different genomic regions.

The proteins MBD1, MBD2, MBD3, MBD4, and MeCP2 comprise a small family of nuclear proteins that share a common methyl-CpG binding domain (MBD). Each of these proteins, probably with the exception of human MBD3, is capable of binding specifically to CpGmethylated DNA. An affinity column-based method was initially developed that employs the methyl-CpG binding domain of MeCP2 [39,40]. Recombinant MBD fragments can be endowed with an affinity tag, such as a His-tag or GST-tag. Incubating fragmented genomic DNA with the tagged MBDs provides a convenient way for enrichment of CpG-methylated DNA fragments. These fragments can then be analyzed, as done earlier using microarrays [36] or currently by using high throughput sequencing (MBD-seq) [11,41–43].

Methylated-CpG Island Recovery Assay Among the MBD proteins, MBD2b has the strongest affinity to methylated DNA and can form heterodimers with other MBD proteins via its C-terminal coiled-coil domain. MBD3L1, a related protein and member of the MBD2/3 sub-family, has no DNA binding domain itself but it is a binding partner of MBD2b via heterodimer formation [44]. The MBD2b/MBD3L1 protein complex has higher affinity to methylated DNA than MBD2b alone [33,34]. In the MIRA procedure, the fragmented (restriction enzyme-cut or sonicated) genomic DNA is incubated with the bacterially expressed and purified GST-MBD2b and His-MBD3L1 proteins. The high

III.  Methods in Epigenetics

130

9.  Methods for Assessing DNA Cytosine Modifications Genome-Wide

FIGURE 9.7  Methylated-CpG island recovery assay (MIRA-seq). The steps of the MIRA procedure are described in the text. Filled and open lollipops mark methylated and unmethylated CpG sites, respectively.

affinity MBD2b/MBD3L1 complex specifically binds to the methylated genomic DNA fragments and, since MBD2b is GST-tagged; it is easy to purify the complex containing methylated DNA by applying glutathionecoated magnetic beads [33,45–47] (Fig. 9.7). MIRA does not depend on restriction enzyme recognition sites or sodium bisulfite conversion of the DNA and it works on double-stranded DNA. There is no dependence on DNA sequences other than that it requires a minimum of two methylated CpGs in the captured fragment. MIRA can reliably be performed with a few hundred nanograms of genomic DNA. The MIRA technique has been used to profile DNA methylation patterns at a resolution of 100 base pairs in the entire genome of normal human B lymphocytes providing one of the first mammalian “methylome” data sets [48]. MIRA analysis is compatible with different microarray platforms [45,46] and with high-throughput DNA sequencing (MIRA-seq) [49–53].

5-Hydroxymethylcytosine Mapping Methodologies A recent technological challenge affecting almost all types of DNA methylation analysis has emerged from the discovery of 5-hydroxymethylcytosine (5hmC) in mammalian DNA [54,55]. This base modification, hitherto thought to be a product of oxidative DNA damage, is in fact produced by an enzymatic pathway involving the TET family 5mC oxidases. 5hmC appears to be quite abundant in certain mammalian tissues and may have regulatory roles distinct from that of 5mC [56]. It should be noted that

5hmC prevents cleavage by certain (and probably most) methylation-sensitive restriction endonucleases making it indistinguishable from 5-methylcytosine by these approaches [57,58]. Moreover, 5-hydroxymethylcytosine in DNA is resistant to bisulfite-induced deamination similar as is 5-methylcytosine and scores identically in bisulfite sequencing [58,59]. It has been shown that 5hmC reacts with bisulfite and, instead of leading to deamination, this reaction gives rise to cytosine 5-methylenesulfonate as the product [60]. Cytosine 5-methylenesulfonate is only very slowly deaminated by treatment with bisulfite. However, 5hmC is not recognized by anti-5mC antibodies and by MBD family proteins that bind to methylated CpGs [58]. Thus, using existing 5-methylcytosine-related technology, one needs to be careful if the goal is to distinguish between the two modified cytosine bases. This is particularly important in tissues where high levels of 5hmC are present, such as the brain. Affinity or pulldown approaches have been used to map the distribution of 5hmC in the genome. One approach is based on anti-5hmC antibodies and is called hMeDIP, a technique analogous to MeDIP [61–63]. Another very specific approach for mapping of 5hmC is based on selective chemical labeling of 5hmC. This method employs the T4 bacteriophage beta-glucosyltransferase to transfer a modified glucose containing an azide group onto the hydroxyl group of 5hmC. The azide group is then used for coupling to biotin, affinity enrichment, and DNA sequencing [64] (Fig. 9.1). Although bisulfite-treatment cannot be used directly for identifying 5hmC at individual nucleotide positions,

III.  Methods in Epigenetics



Introduction

131

by employing some modifications—before the actual bisulfite-treatment—the bisulfite approach can produce base resolution sequence information that unambiguously identifies positions of these modified cytosines in genomic DNA. Currently there are two methods that can be used for identification of 5hmC at single-base resolution.

due to the rarity of the 5hmC base, which requires sequence coverage 5–50 times greater than when 5mC is analyzed (depending on the tissue). A less expensive but more restricted methodology is the combination of TAB-seq and RRBS, which limits the analysis to mostly CpG-rich genomic regions [68].

TET-Assisted Bisulfite Sequencing

Oxidative Bisulfite Sequencing

TET-assisted bisulfite sequencing (TAB-seq) relies on the application of two enzymes; it first employs βglucosyltransferase-mediated glucosylation of 5-hydroxymethylcytosine and then TET-catalyzed oxidation of 5-methylcytosine to 5-carboxylcytosine [65] (Fig. 9.8). Thus, genomic DNA samples are treated with β-glucosyltransferase and the glycosylated 5-hydroxymethylcytosines are thus protected from subsequent TET-mediated oxidization. Next, bisulfite conversion of the resulting pretreated DNA samples are conducted, which converts unmodified cytosine and 5-carboxylcytosine residues to uracils, while leaving the glycosylationprotected 5hmC bases intact. Next, BS-treated samples are used for sequencing library generation, and any Cs in the sequencing reads can be interpreted as 5hmC. This method has been used [66,67] but is relatively expensive

In Oxidative bisulfite sequencing (OxBS-Seq), genomic DNA is treated with a strong oxidizing agent (KRuO4) that induces the conversion of 5-hydroxymethylcytosines to 5-formylcytosine but leaves 5-methylcytosine and unmodified cytosine residues intact [69] (Fig. 9.9). Subsequent bisulfite treatment converts 5-formylcytosine and unmodified cytosine residues to uracils. Next, the BS-treated DNA sample is used for sequencing library synthesis and is subjected to HTS. Of note, Ts in the sequencing reads can be either 5-hydroxymethylcytosines or unmodified cytosines; therefore OxBS-Seq and conventional BS-Seq data must be compared to each other for revealing the positions of 5hmC residues, which is a disadvantage relative to TAB-seq. The same limitations regarding high sequence coverage as for TAB-seq apply here.

FIGURE 9.8  TET-assisted bisulfite sequencing (TAB-seq). Whereas regular bisulfite sequencing cannot distinguish between the two modified bases and provides the sum of 5mC and 5hmC at each position, TAB-seq provides a direct display of 5hmC only at base resolution. caC, 5-Carboxylcytosine; CMS, cytosine-5-methylsulfonate; gmC, glucosylated 5- hydroxymethylcytosine; hmC, 5-hydroxymethylcytosine; mC, 5-methylcytosine.

FIGURE 9.9  Oxidative bisulfite sequencing (OxBS-seq). OxBS-seq displays 5mC only and requires a comparison to standard BS-seq to infer the sequence positions and abundance of 5hmC and 5mC. CMS, Cytosine-5-methylsulfonate; fC, 5-formylcytosine; hmC, 5-hydroxymethylcytosine; mC, 5-methylcytosine.

III.  Methods in Epigenetics

132

9.  Methods for Assessing DNA Cytosine Modifications Genome-Wide

Future Directions and Challenges The many different DNA methylation profiling techniques described in this chapter will continue to be applied in various settings of molecular analysis and probably will at some point move into the clinical arena for disease diagnosis and treatment stratification. The particular method of choice depends on many parameters including the scope of analysis that is desired, the exact questions to be pursued, and sample size, as well as cost. It is expected that the technological development of highthroughput approaches will continue at an accelerated pace allowing cost effective analysis of large sample series. The current challenges and cost surrounding whole genome bisulfite sequencing will probably be solved in the future. However, it may not always be desirable or even necessary to have information on the methylation status of every single one of the ∼28 million CpG sites in the haploid human genome. For determining the identity and number of methylated genes, in cancer for example, a lower resolution, affinity-based approach will be fully sufficient, and for disease diagnosis, only a subset of specific genes will most likely be relevant. Entirely new technologies are on the horizon. Single molecule DNA sequencing includes a set of promising new technologies in which 5-methylcytosine or its oxidation products may be distinguished from the other four standard DNA bases [70]. Until now, this method has not found widespread application yet but further developments in this area are likely. Another topic of current interest is the cellular heterogeneity inherent to tissues, in particular tumors. Single cell analysis of DNA methylation patterns by whole genome bisulfite analysis is challenging but is now on the horizon [71,72]. Finally, very recent studies have shown that the DNA base N6-methyladenine (6mA), previously though to be chiefly an epigenetic modification in prokaryotes, also occurs in higher organisms including worms, flies, algae, and even mammals [73,74]. The detection of 6mA provides its unique challenges. Bisulfite conversion-based approaches cannot be used to sequence 6mA. Instead, this base can be mapped using restriction enzymes that are inhibited by 6mA within the cleavage site, by affinity-based approaches using, for example, antibodies and by single molecule sequencing.

References [1] Wyatt GR. Occurrence of 5-methylcytosine in nucleic acids. Nature 1950;166(4214):237–8. [2] Holliday R, Pugh JE. DNA modification mechanisms and gene activity during development. Science 1975;187(4173):226–32. [3] Riggs AD. X inactivation, differentiation, and DNA methylation. Cytogenet Cell Genet 1975;14(1):9–25. [4] Huang TH, Perry MR, Laux DE. Methylation profiling of CpG islands in human breast cancer cells. Hum Mol Genet 1999;8(3): 459–70.

[5] Khulan B, Thompson RF, Ye K, Fazzari MJ, Suzuki M, Stasiek E, et al. Comparative isoschizomer profiling of cytosine methylation: the HELP assay. Genome Res 2006;16(8):1046–55. [6] Oda M, Glass JL, Thompson RF, Mo Y, Olivier EN, Figueroa ME, et al. High-resolution genome-wide cytosine methylation profiling with simultaneous copy number analysis and optimization for limited cell numbers. Nucleic Acids Res 2009;37(12):3829–39. [7] Estecio MR, Yan PS, Ibrahim AE, Tellez CS, Shen L, Huang TH, et al. High-throughput methylation profiling by MCA coupled to CpG island microarray. Genome Res 2007;17(10):1529–36. [8] Toyota M, Ho C, Ahuja N, Jair KW, Li Q, Ohe-Toyota M, et al. Identification of differentially methylated sequences in colorectal cancer by methylated CpG island amplification. Cancer Res 1999;59(10):2307–12. [9] Kleb B, Estecio MR, Zhang J, Tzelepi V, Chung W, Jelinek J, et al. Differentially methylated genes and androgen receptor re-expression in small cell prostate carcinomas. Epigenetics 2016;11(3): 184–93. [10] Okamoto Y, Shinjo K, Shimizu Y, Sano T, Yamao K, Gao W, et al. Hepatitis virus infection affects DNA methylation in mice with humanized livers. Gastroenterology 2014;146(2):562–72. [11] Harris RA, Wang T, Coarfa C, Nagarajan RP, Hong C, Downey SL, et al. Comparison of sequencing-based methods to profile DNA methylation and identification of monoallelic epigenetic modifications. Nat Biotechnol 2010;28(10):1097–105. [12] Hayatsu H. Discovery of bisulfite-mediated cytosine conversion to uracil, the key reaction for DNA methylation analysis—a personal account. Proc Jpn Acad Ser B Phys Biol Sci 2008;84(8): 321–30. [13] Hayatsu H, Wataya Y, Kazushige K. The addition of sodium bisulfite to uracil and to cytosine. J Am Chem Soc 1970;92(3):724–6. [14] Clark SJ, Harrison J, Paul CL, Frommer M. High sensitivity mapping of methylated cytosines. Nucleic Acids Res 1994;22(15): 2990–7. [15] Frommer M, McDonald LE, Millar DS, Collis CM, Watt F, Grigg GW, et al. A genomic sequencing protocol that yields a positive display of 5-methylcytosine residues in individual DNA strands. Proc Natl Acad Sci USA 1992;89(5):1827–31. [16] Eckhardt F, Lewin J, Cortese R, Rakyan VK, Attwood J, Burger M, et al. DNA methylation profiling of human chromosomes 6 20 and 22. Nat Genet 2006;38(12):1378–85. [17] Cokus SJ, Feng S, Zhang X, Chen Z, Merriman B, Haudenschild CD, et al. Shotgun bisulphite sequencing of the Arabidopsis genome reveals DNA methylation patterning. Nature 2008;452(7184):215–9. [18] Lister R, Pelizzola M, Dowen RH, Hawkins RD, Hon G, TontiFilippini J, et al. Human DNA methylomes at base resolution show widespread epigenomic differences. Nature. 2009;462: 315–322. [19] Mo A, Luo C, Davis FP, Mukamel EA, Henry GL, Nery JR, et al. Epigenomic landscapes of retinal rods and cones. Elife 2016;5:e11613. [20] Schultz MD, He Y, Whitaker JW, Hariharan M, Mukamel EA, Leung D, et al. Human body epigenome maps reveal noncanonical DNA methylation variation. Nature 2015;523(7559):212–6. [21] Meissner A, Mikkelsen TS, Gu H, Wernig M, Hanna J, Sivachenko A, et al. Genome-scale DNA methylation maps of pluripotent and differentiated cells. Nature 2008;454(7205):766–70. [22] Meissner A, Gnirke A, Bell GW, Ramsahoye B, Lander ES, Jaenisch R. Reduced representation bisulfite sequencing for comparative high-resolution DNA methylation analysis. Nucleic Acids Res 2005;33(18):5868–77. [23] Gu H, Smith ZD, Bock C, Boyle P, Gnirke A, Meissner A. Preparation of reduced representation bisulfite sequencing libraries for genome-scale DNA methylation profiling. Nat Protoc 2011;6(4): 468–81.

III.  Methods in Epigenetics



REFERENCES

[24] Ball MP, Li JB, Gao Y, Lee JH, LeProust EM, Park IH, et al. Targeted and genome-scale strategies reveal gene-body methylation signatures in human cells. Nat Biotechnol 2009;27(4):361–8. [25] Deng J, Shoemaker R, Xie B, Gore A, LeProust EM, AntosiewiczBourget J, et al. Targeted bisulfite sequencing reveals changes in DNA methylation associated with nuclear reprogramming. Nat Biotechnol 2009;27(4):353–60. [26] Lister R, Ecker JR. Finding the fifth base: genome-wide sequencing of cytosine methylation. Genome Res 2009;19(6):959–66. [27] Du Y, Li M, Chen J, Duan Y, Wang X, Qiu Y, et al. Promoter targeted bisulfite sequencing reveals DNA methylation profiles associated with low sperm motility in asthenozoospermia. Hum Reprod 2016;31(1):24–33. [28] Pabinger S, Ernst K, Pulverer W, Kallmeyer R, Valdes AM, Metrustry S, et al. Analysis and visualization tool for targeted amplicon bisulfite sequencing on ion torrent sequencers. PLoS One 2016;11(7):e0160227. [29] Dedeurwaerder S, Defrance M, Calonne E, Denis H, Sotiriou C, Fuks F. Evaluation of the infinium methylation 450K technology. Epigenomics 2011;3(6):771–84. [30] Moran S, Arribas C, Esteller M. Validation of a DNA methylation microarray for 850,000 CpG sites of the human genome enriched in enhancer sequences. Epigenomics 2016;8(3):389–99. [31] Ehrich M, Nelson MR, Stanssens P, Zabeau M, Liloglou T, Xinarianos G, et al. Quantitative high-throughput analysis of DNA methylation patterns by base-specific cleavage and mass spectrometry. Proc Natl Acad Sci USA 2005;102(44):15785–90. [32] Olk-Batz C, Poetsch AR, Nollke P, Claus R, Zucknick M, Sandrock I, et al. Aberrant DNA methylation characterizes juvenile myelomonocytic leukemia with poor outcome. Blood 2011;117(18): 4871–80. [33] Rauch T, Li H, Wu X, Pfeifer GP. MIRA-assisted microarray analysis, a new technology for the determination of DNA methylation patterns, identifies frequent methylation of homeodomain-containing genes in lung cancer cells. Cancer Res 2006;66(16):7939–47. [34] Rauch T, Pfeifer GP. Methylated-CpG island recovery assay: a new technique for the rapid detection of methylated-CpG islands in cancer. Lab Invest 2005;85(9):1172–80. [35] Weber M, Davies JJ, Wittig D, Oakeley EJ, Haase M, Lam WL, et al. Chromosome-wide and promoter-specific analyses identify sites of differential DNA methylation in normal and transformed human cells. Nat Genet 2005;37(8):853–62. [36] Gebhard C, Schwarzfischer L, Pham TH, Schilling E, Klug M, Andreesen R, et al. Genome-wide profiling of CpG methylation identifies novel targets of aberrant hypermethylation in myeloid leukemia. Cancer Res 2006;66(12):6118–28. [37] Illingworth R, Kerr A, Desousa D, Jorgensen H, Ellis P, Stalker J, et al. A novel CpG island set identifies tissue-specific methylation at developmental gene loci. PLoS Biol 2008;6(1):e22. [38] Li D, Zhang B, Xing X, Wang T. Combining MeDIP-seq and MREseq to investigate genome-wide CpG methylation. Methods 2015;72:29–40. [39] Cross SH, Charlton JA, Nan X, Bird AP. Purification of CpG islands using a methylated DNA binding column. Nat Genet 1994;6(3):236–44. [40] Zhang X, Yazaki J, Sundaresan A, Cokus S, Chan SW, Chen H, et al. Genome-wide high-resolution mapping and functional analysis of DNA methylation in arabidopsis. Cell 2006;126(6): 1189–201. [41] Serre D, Lee BH, Ting AH. MBD-isolated Genome Sequencing provides a high-throughput and comprehensive survey of DNA methylation in the human genome. Nucleic Acids Res 2010;38(2):391–9. [42] Brinkman AB, Simmer F, Ma K, Kaan A, Zhu J, Stunnenberg HG. Whole-genome DNA methylation profiling using MethylCapseq. Methods 2010;52(3):232–6.

133

[43] Subhash S, Andersson PO, Kosalai ST, Kanduri C, Kanduri M. Global DNA methylation profiling reveals new insights into epigenetically deregulated protein coding and long noncoding RNAs in CLL. Clin Epigenetics 2016;8:106. [44] Jiang CL, Jin SG, Pfeifer GP. MBD3L1 is a transcriptional repressor that interacts with methyl-CpG-binding protein 2 (MBD2) and components of the NuRD complex. J Biol Chem 2004;279(50):52456–64. [45] Rauch T, Wang Z, Zhang X, Zhong X, Wu X, Lau SK, et al. Homeobox gene methylation in lung cancer studied by genome-wide analysis with a microarray-based methylated CpG island recovery assay. Proc Natl Acad Sci USA 2007;104(13):5527–32. [46] Rauch TA, Zhong X, Wu X, Wang M, Kernstine KH, Wang Z, et al. High-resolution mapping of DNA hypermethylation and hypomethylation in lung cancer. Proc Natl Acad Sci USA 2008;105(1):252–7. [47] Rauch TA, Pfeifer GP. The MIRA method for DNA methylation analysis. Methods Mol Biol 2009;507:65–75. [48] Rauch TA, Wu X, Zhong X, Riggs AD, Pfeifer GP. A human B cell methylome at 100-base pair resolution. Proc Natl Acad Sci USA 2009;106(3):671–8. [49] Jung M, Kadam S, Xiong W, Rauch TA, Jin SG, Pfeifer GP. MIRAseq for DNA methylation analysis of CpG islands. Epigenomics 2015;7(5):695–706. [50] Jin SG, Xiong W, Wu X, Yang L, Pfeifer GP. The DNA methylation landscape of human melanoma. Genomics 2015;106(6):322–30. [51] Almamun M, Levinson BT, Gater ST, Schnabel RD, Arthur GL, Davis JW, et al. Genome-wide DNA methylation analysis in precursor B-cells. Epigenetics 2014;9(12):1588–95. [52] Lee EJ, Luo J, Wilson JM, Shi H. Analyzing the cancer methylome through targeted bisulfite sequencing. Cancer Lett 2013;340(2):171–8. [53] Benjamin AL, Green BB, Crooker BA, McKay SD, Kerr DE. Differential responsiveness of Holstein and Angus dermal fibroblasts to LPS challenge occurs without major differences in the methylome. BMC Genomics 2016;17:258. [54] Kriaucionis S, Heintz N. The nuclear DNA base 5-hydroxymethylcytosine is present in Purkinje neurons and the brain. Science 2009;324(5929):929–30. [55] Tahiliani M, Koh KP, Shen Y, Pastor WA, Bandukwala H, Brudno Y, et al. Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 2009;324(5929):930–5. [56] Hahn MA, Szabo PE, Pfeifer GP. 5-Hydroxymethylcytosine: a stable or transient DNA modification? Genomics 2014;104: 314–23. [57] Tardy-Planechaud S, Fujimoto J, Lin SS, Sowers LC. Solid phase synthesis and restriction endonuclease cleavage of oligodeoxynucleotides containing 5-(hydroxymethyl)-cytosine. Nucleic Acids Res 1997;25(3):553–9. [58] Jin SG, Kadam S, Pfeifer GP. Examination of the specificity of DNA methylation profiling techniques towards 5-methylcytosine and 5-hydroxymethylcytosine. Nucleic Acids Res 2010;38(11):e125. [59] Huang Y, Pastor WA, Shen Y, Tahiliani M, Liu DR, Rao A. The behaviour of 5-hydroxymethylcytosine in bisulfite sequencing. PLoS One 2010;5(1):e8888. [60] Hayatsu H, Shiragami M. Reaction of bisulfite with the 5-hydroxymethyl group in pyrimidines and in phage DNAs. Biochemistry 1979;18(4):632–7. [61] Jin SG, Wu X, Li AX, Pfeifer GP. Genomic mapping of 5-hydroxymethylcytosine in the human brain. Nucleic Acids Res 2011;39:5015–24. [62] Stroud H, Feng S, Morey Kinney S, Pradhan S, Jacobsen SE. 5-Hydroxymethylcytosine is associated with enhancers and gene bodies in human embryonic stem cells. Genome Biol 2011; 12(6):R54.

III.  Methods in Epigenetics

134

9.  Methods for Assessing DNA Cytosine Modifications Genome-Wide

[63] Williams K, Christensen J, Pedersen MT, Johansen JV, Cloos PA, Rappsilber J, et al. TET1 and hydroxymethylcytosine in transcription and DNA methylation fidelity. Nature 2011;473(7347):343–8. [64] Song CX, Szulwach KE, Fu Y, Dai Q, Yi C, Li X, et al. Selective chemical labeling reveals the genome-wide distribution of 5-hydroxymethylcytosine. Nat Biotechnol 2011;29(1):68–72. [65] Yu M, Hon GC, Szulwach KE, Song CX, Jin P, Ren B, et al. Tetassisted bisulfite sequencing of 5-hydroxymethylcytosine. Nat Protoc 2012;7(12):2159–70. [66] Chen K, Zhang J, Guo Z, Ma Q, Xu Z, Zhou Y, et al. Loss of 5-hydroxymethylcytosine is linked to gene body hypermethylation in kidney cancer. Cell Res 2016;26(1):103–18. [67] Hahn MA, Qiu R, Wu X, Li AX, Zhang H, Wang J, et al. Dynamics of 5-hydroxymethylcytosine and chromatin marks in Mammalian neurogenesis. Cell Rep 2013;3(2):291–300. [68] Hahn MA, Li AX, Wu X, Pfeifer GP. Single base resolution analysis of 5-methylcytosine and 5-hydroxymethylcytosine by RRBS and TAB-RRBS. Methods Mol Biol 2015;1238:273–87.

[69] Booth MJ, Ost TW, Beraldi D, Bell NM, Branco MR, Reik W, et al. Oxidative bisulfite sequencing of 5-methylcytosine and 5-hydroxymethylcytosine. Nat Protoc 2013;8(10):1841–51. [70] Korlach J, Turner SW. Going beyond five bases in DNA sequencing. Curr Opin Struct Biol 2012;22(3):251–61. [71] Farlik M, Sheffield NC, Nuzzo A, Datlinger P, Schonegger A, Klughammer J, et al. Single-cell DNA methylome sequencing and bioinformatic inference of epigenomic cell-state dynamics. Cell Rep 2015;10(8):1386–97. [72] Gravina S, Dong X, Yu B, Vijg J. Single-cell genome-wide bisulfite sequencing uncovers extensive heterogeneity in the mouse liver methylome. Genome Biol 2016;17(1):150. [73] Pfeifer GP, Epigenetics:. An elusive DNA base in mammals. Nature 2016;532(7599):319–20. [74] Sun Q, Huang S, Wang X, Zhu Y, Chen Z, Chen D. N6-methyladenine functions as a potential epigenetic mark in eukaryotes. Bioessays 2015;37(11):1155–62.

III.  Methods in Epigenetics

C H A P T E R

10 Analyses of Genome-Wide Histone Modifications in the Mammalian Genome Shulan Tian, Susan L. Slager, Krutika S. Gaonkar, Huihuang Yan Mayo Clinic, Rochester, MN, United States

O U T L I N E Introduction

135

High-Throughput Technologies to Study Histone Modifications ChIP-Seq in Studying Histone Modifications Analyses of Genome-Wide Histone Modification Data

136 136 138

Genome-Wide Profiles and Functions of Histone Modifications Histone Methylation Histone Acetylation Histone Ubiquitination Histone Phosphorylation Histone Variants and Linker Histone H1 Chromatin Regulators

139 139 142 143 143 143 144

INTRODUCTION In eukaryote cells, DNA is tightly packed into nucleosomes together with a histone octamer consisted of core histones H2A, H2B, H3, and H4 [1,2]. The nucleosomes are connected by linker DNA and histone H1 to form chromatin fibers, which are folded into three-dimensional structure [3]. Both core and linker histones can not only be posttranslationally modified, predominantly on their tail domains, but can also be within the globular domains [4–6]. The level and distribution of a histone modification reflect the interplay between chromatin regulators (CRs) that deposit (writer), remove (eraser), or bind (reader) the modification [5,6]. Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00010-9 Copyright © 2017 Elsevier Inc. All rights reserved.

Histone Modifications in Key Genomic Regions Histone Modifications in Promoters Histone Modifications in Enhancers Histone Modifications in Gene Bodies

145 145 146 147

Alteration of Histone Modifications in Disease

147

Conclusions and Perspectives

148

Acknowledgments

149

Glossary

149

Abbreviations

149

References

149

Histone modifications play key roles in establishing chromatin structure, accessibility, and function [7–9] by recruiting effector proteins that contain chromatin reader domains [4,6,10], thus further influencing the modifications at nearby residues and DNA–histone interactions [11]. Histone modifications influence multiple biological processes, from transcriptional regulation [3,11,12], DNA replication [3], and DNA repair [11,12] to chromosomal maintenance [3,13]. Despite nearly identical DNA sequence, cells from different tissues could have distinct gene expression profiles [14]. The cellular gene expression dynamics in response to environmental factors, developmental signals, or disease progression is regulated through histone modifications, along with

135

136

10.  Analyses of Genome-Wide Histone Modifications in the Mammalian Genome

other epigenetic mechanisms involving DNA methylation, histone variants, CRs, nucleosome occupancy and positioning, chromatin remodeling, high-order chromatin configuration, and noncoding RNAs [1,2,15,16]. There are at least 18 groups of histone modifications [6,7,17], totaling over 100 residue-specific modifications [1,12]. The four most common histone modifications are methylation (lysine and arginine), acetylation (lysine), phosphorylation (lysine, serine, and threonine) and ubiquitination (lysine) [5,7,17]. Chromatin immunoprecipitation and sequencing (ChIP-seq) has been widely used for genome-wide mapping of histone modifications [18,19], histone variants [20], and CRs [21] in both cell lines and primary tissues. These studies revealed distinct histone modification patterns in key genomic regions and the alterations in human disease. New protocols can even generate profiles from single cells [9,22] and map two modifications simultaneously [23–25]. In this chapter, we describe progress in genome-wide mapping of common histone modifications, histone variants, and CRs, mostly in the human genome through ChIP-seq. We summarize chromatin states of genomic regions, such as promoters, enhancers, and gene bodies. We present key findings illustrating the cell-type specificity of histone modifications, their roles in transcriptional regulation, and alteration of occupancy linked to cancer and noncancerous diseases. Finally, we highlight major challenges and areas for future studies in this field.

HIGH-THROUGHPUT TECHNOLOGIES TO STUDY HISTONE MODIFICATIONS Mass spectrometry (MS)–based analyses can be used to identify the type, abundance, and coexistence of histone modifications [6]; nevertheless, they have low resolution and fail to link histone modifications to genomic locations [9,26]. ChIP-chip was first used to map the binding sites of transcriptional activators and to identify the associated target genes in yeast [27], but it has limitations in genome coverage, sensitivity, and throughput when applied to much larger genomes, such as human and mouse. ChIP-seq can achieve much greater genome coverage and single-nucleosome or subnucleosomal resolution [26,28,29]. In high-resolution ChIP-seq, chromatin is digested by micrococcal nuclease, which can generate mononucleosome fragments and subnucleosomal DNAs of T single-nucleotide polymorphisms (SNPs) cannot be distinguished from the deamination of unmethylated Cs in the bisulfite-converted DNA. This shortcoming can be overcome by Bis-SNP that uses Bayesian inference to improve DNA methylation calling accuracy, as well as to determine monoallelic DNA methylation and SNPs in cis-regulatory sequences [57]. Upon obtaining the processed DNA methylation data, these data can be compiled into the CpG methylation table. This table can be converted into various file formats, such as bigbed and bigwig [58], in order to be imported to UCSC Genome Browser, Ensembl, WashU Epigenome Browser, MethylomeDB Browser, Integrated Genome Browser, Integrative Genomics Viewer, and Epiviz [58–61] for region-specific visualization. Furthermore, several types of diagrams (e.g., box plots, violin plots, Hilbert curve, Tree-like diagrams, and scatter plots) can be used to visualize the global changes and distribution patterns in DNA methylation data [62]. A number of software- and web-based tools for bisulfite sequencing are listed in Table 12.2.

PROCESSING MICROARRAY-BASED DNA METHYLATION DATA A number of microarray-based technologies are also commercially available for high-throughput methylation analysis, such as Infinium Methylation Arrays (Illumina), Human CpG Island Microarray Kit (Agilent), GeneChip Human Promoter 1.0R Array, and GeneChip Human Tiling 2.0R Array Set (Affymetrix). Illumina BeadChip platform provide a comprehensive genomewide DNA methylation profiling of >450,000 (HumanMethylation450) and >850,000 (MethylationEPIC) CpG sites located throughout the human genome. Each array comprises a combination of Infinium I and Infinium II assay chemistries, for the quantitative genotyping of C/T polymorphism in the bisulfite-converted DNA [83]. Similar to bisulfite sequencing, this method detects DNA methylation at specific regions across the promoter, enhancer, 5′-UTR, first exon, gene body, and 3′-UTR. Data generated from HumanMethylation450 BeadChip are processed via image processing and data normalization. Image processing can be carried out with vendor-provided software, such as Illumina Bead Scan, and raw data is subsequently normalized with GenomeStudio Methylation Module software by obtaining methylation β value. The enumerated β value is equivalent to the absolute DNA methylation levels that are derived from bisulfite-sequencing data. In which, β = M/ (M + U + α) where α, M, and U represent arbitrary offset (usually set at 100), methylated-, and unmethylated signal intensities, respectively [84]. Alternatively, the absolute DNA methylation levels can be derived from the M value, which is equivalent to a logit transformation of the β value, by which M = log2[(M + α)/(U + α)] [84]. Precision and accuracy of absolute DNA methylation levels can also be improved by subset quantile normalization of the β value using subset-quantile within array normalization (SWAN) method [85]. A recent study suggested that the combination of Noob + BMIQ + RUVm batch-correction method further improves the sensitivity and reproducibility of data processing and analysis [86]. Differential DNA methylation analysis can be performed in intuitive graphics (e.g., heat maps, scatter plots, and line plots) and integrated with gene expression data.

Methylated DNA Enrichment Alternatively, DNA methylation can be assayed using enrichment-based methods. MeDIP uses 5mC antibodies against methylcytosines to enrich methylated CpG regions, which are further analyze with NGS (MeDIP-seq). Another method (MBD-seq) utilizes the specificity of MBD protein-binding sites to capture methylated CpG. On the other hand, unmethylated DNA canbe enriched

III.  Methods in Epigenetics



TABLE 12.2 Software and Web-Based Tools for Bisulfite Sequencing Analysis Input formats

Output formats

Main features

Web links

References

BAM-ABS

Perl

bed, sam

text, sam

BAM-ABS built from Bayesian model is using posterior probability to assign ambiguously mapped reads in each candidate genomic location generated from Bismark output

https://github.com/ zhanglabvt/BAM_ABS

[63]

BatMeth

C++

fastq

text, pdf

BatMeth, an algorithm that combines mismatch counting, list filtering, mismatch stage filtering, and fast mapping onto two indexes to map the bisulfite sequencing (base and color) reads generated from Illumina and SOLiD, respectively

http://code.google. com/p/batmeth/

[64]

BatMis

C++

fasta, fastq

text

BatMis is a Burrows–Wheeler transformation based short read aligner for k mismatches. It allows up to 10 mismatches in the whole read

http://code.google. com/p/batmis/

[65]

BDPC

Php

abi, text, fasta

png

BDPC is an online tool that uses BISMA to analyze bisulfite sequencing reads. It compiles, clusters, and compares DNA methylation data according to their sources (tissues), and allows visualization of DNA methylation data in the form of heat maps

http://services.ibc.unistuttgart.de/BDPC/

[66]

BiQ Analyzer HiMod

Java

fasta, fastq, bam

tab-delimited text, bedgraph, svg

BiQ Analyzer HiMod, a cross-platform Java application that identifies locus-specific DNA methylation of cytosine and its oxidative residues. It is an updated version of BiQ Analyzer HT

http://biq-analyzerhimod.bioinf.mpi-inf. mpg.de/

[67]

BISMA

Perl

abi, tab-delimited text, multifasta

HTML, png

BISMA is a web-based tool that utilizes PHP coding language to analyze methylation levels of repetitive sequence. It automatically detects sequence direction and the presence of insertions or deletions. It generates graphical outputs of CpG site distribution and condensed methylation pattern. It also calculates the average methylation at each CpG site and the overall methylation levels

http://services.ibc.unistuttgart.de/BDPC/ BISMA/

[68]

Bismark

Perl

fastq, fasta, multi-fasta

bam, tab-delimited text, bedgraph

Bismark allows bisulfite sequence read mapping and methylation calling in a single step. It handles bisulfite sequence data from directional and nondirectional libraries. The output can discriminate cytosines in CpG, CHG, and CHH contexts

www.bioinformatics. bbsrc.ac.uk/projects/ bismark/

[52]

Bis-SNP

Java script, Perl

bam

bed, wig, vcf

Bis-SNP, a nonbisulfite SNP caller that is designed based on GATK framework. It takes advantage on Map-Reduce computation strategy, while applying Bayesian inference to evaluate the base calls and base quality scores of SNPs and detects the heterozygous SNPs for monoallelic DNA methylation

http://epigenome.usc. edu/publicationdata/ bissnp2011

[57]

Processing Microarray-Based DNA Methylation Data

III.  Methods in Epigenetics

Programming languages

173

(Continued)

174

TABLE 12.2  Software and Web-Based Tools for Bisulfite Sequencing Analysis (cont.) Input formats

Output formats

Main features

Web links

References

BRAT-nova

C++

fastq

sam

BRAT-nova is an improved version of BRAT and BRAT-BW. It is a memory efficient tool that maps single- and pair-end bisulfite sequencing short reads (>50 nucleotides) using Farragina–Manzini index. It trims low-quality read ends for enumerating the methylation levels accurately

http://compbio.cs.ucr. edu/brat/

[55]

BSMAP

C++

fasta, fastq, sam

sam, tab-delimited text

BSMAP combines genome hashing and bitwise masking to align gapped and pair-end sequencing reads

http://code.google. com/p/bsmap/

[48]

BSPAT 2.3.2-beta

Java, Jsp

fasta, fastq

text, png, eps

BSPAT is a web-based tool that utilizes sequence mapping strategy to provide fast sequencing alignment for regions with extreme high read depths. It provides visual summary for simple allele-specific methylation patterns, while allowing various integrative analyses

http://cbc.case.edu/ BSPAT

[69]

BSseeker

Python

fasta, fastq, qseq

sam, tab-delimited text

BSseeker is the mapping program that can tag bisulfite sequencing reads

http://pellegrini.mcdb. ucla.edu/BS_Seeker/ BS_Seeker.html

[50]

Bycom

Perl

fastq, sam, bam

text

Bycom built from Bayesian inference model can identify methylcytosines and SNPs from bisulfite-sequencing mapped and unmapped reads. It recognizes sequencing errors, nonconversion rate and cell heterozygosis of the samples

https://sourceforge.net/ projects/bycom/

[70]

DMEAS

C++, Perl

bam, tabdelimited text

tab-delimited text, jpg, bmp

DMEAS uses the output file from Bismark to identify DNA methylation patterns, enumerate DNA methylation levels, and perform methylation entropy analysis via a modified version of Shannon entropy equation

https://sourceforge.net/ projects/dmeas/files/

[71]

EpiGRAPH

Java, Phyton

xml

text

EpiGRAPH is a web service that integrates vertebrate genome and epigenome datasets based on 10 attributes. The 10 attributes are: DNA sequence, DNA structure, repetitive DNA, chromosomal organization, evolutionary history, population variation, regulatory regions, genes, transcriptome, and epigenome and chromatin structures

http://epigraph.mpi-inf. mpg.de/

[72]

GBSA

Python

bsp, sam, bam

bedgraph, tabdelimited text

GBSA uses the output files from BSMAP, RRBSmap, and BSseeker to score methylation levels at gene-centric or gene-independent loci

http://ctrad-csi.nus.edu. sg/gbsa

[73]

GSNAP

C++, Perl

fasta, fastq

tab-delimited text, sam

GSNAP aligns single- and paired-end reads from 14 nucleotides onward. It allows SNP-tolerant alignment and detects interchromosomal-, short-, and longdistance splicing, while calculating DNA methylation levels

http://research-pub.gene. com/gmap/

[74]

12.  Computational Epigenetics

III.  Methods in Epigenetics

Programming languages



C++

fastq

text

HPG-Methyl aligns as short as 75 bisulfite sequencing reads using Burrows–Wheeler transform method and Smith–Waterman algorithm

anonymous@clariano. uv.es (password: anonymous)

[75]

MethylCoder

C++, Python

fasta, fastq

tab-delimited text, sam, text

MethylCoder corrects soft-masked- and overlapping paired-end bisulfite sequencing reads to distinguishe DNA methylation levels at a single-base resolution

https://github.com/ brentp/methylcode

[76]

MethGo

Python

bam, fasta, cgmap, gtf

text, png

MethGo is a postalignment tool that reveals the coverage distribution of each cytosine, global, and gene-centric methylation levels, as well as methylation levels at transcription factor–binding sites

http://paoyangchenlaboratory.github.io/ methgo/

[77]

MethylMapper

Perl, NCBIBlast

fasta

text, xls

MethylMapper combines a (any) primer design program and BLASTN to generate the cluster plots for DNA methylation status

http://genome.ufl.edu/ methyl

[78]

MethTools 2.0

HTML, PerlCgi script

fasta

fasta, gif, png, svg, pdf

MethTools 2.0 is a second-generation web-based tool of MethTools. It compares bisulfite-PCR targeted sequence data with the unconverted genomic sequence. It generates graphical displays of methylation patterns, methylation densities, and methylated consensus sequences, where all of these graphical displays can be edited using Adobe Illustrator. It calculates methylated and unmethylated cytosines for each PCR clone and their average methylation levels. An error report for sequence quality evaluation can be generated upon request

http://methdb.igh.cnrs. fr/methtools/

[79]

QUMA

HTML, Java script, PerlCgi script

fasta, multi-fasta, genbank, plain sequence

text, csv, png

QUMA is an interactive web-based tool that accepts two types of input, namely a PCR target of an unconverted genomic sequence and raw bisulfite sequences. It is an all-in-one data-processing tool with variable output styles that provides consistent quality control on the aligned sequences. Fisher’s exact test and Mann–Whitney U-test are performed if the two groups of raw bisulfite sequences are fed into the optional fields as input

http://quma.cdb.riken. jp/

[80]

RRBSmap

C++

fasta, fastq, sam, bam

tab-delimited text, bed, sam, bam

RRBSMAP is a short-read alignment tool for singleended bisulfite sequencing reads that mapped on enzyme restriction sites. It reads up to 144 nucleotides, while allowing 15 mismatches and a gap size up to 3 bp

http://rrbsmap. computationalepigenetics.org/; http://code.google. com/p/bsmap/

[47]

RMAP

Python

fastq, fasta, fasta + prb

bed, fasta

RMAP maps paired-end bisulfite sequencing short reads in the presences or absences of error probability information. Read length or the number of mismatches is not a limiting factor for the mapping

http://www.cmb. usc.edu/people/ andrewds/rmap

[53]

Processing Microarray-Based DNA Methylation Data

III.  Methods in Epigenetics

HPG-Methyl

175

(Continued)

176

TABLE 12.2  Software and Web-Based Tools for Bisulfite Sequencing Analysis (cont.) Programming languages

Input formats

Output formats

Main features

Web links

References

TAMeBS

C++

fastq, fasta

tab-delimited text

TAMeBS implements bidirectional FMindex data structure, seed-and-extend sequence alignment scheme and likelihood-ratio scoring matrix to align bisulfite sequencing reads. It calculates the likelihoodratio and analyzes cytosine methylation from the mapped reads

https://sourceforge.net/ projects/tamebs/

[81]

WBSA

Jsp, Java, Perl

fastq

fastq, png

WBSA is a web-based tool that executes read mapping from WGBS and RRBS, determines their methylation levels, and annotates the correlation between differentially methylated regions and genes

http://wbsa.big.ac.cn

[82]

12.  Computational Epigenetics

III.  Methods in Epigenetics

bam, Binary alignment/map; BatMet, Basic Alignment Tool for Methylation; BatMis, Basic Alignment Tool for Mismatches; BDPC, Bisulfite Sequencing Data Presentation and Compilation; BISMA, Bisulfite Sequencing DNA Methylation Analysis; bmp, bitmap, BRAT, Bisulfite-Treated Reads Analysis Tool; bsp, binary space partition, BSPAT, Bisulfite Sequencing Pattern Analysis Tool; csv, comma-seperated value; DMEAS, DNA Methylation Entropy Analysis Software; eps, encapsulated postscript; GBSA, Genome Bisulfite Sequencing Analyser; gif, graphics interchange format; GSNAP, Genomic Short-Read Nucleotide Alignment Program; gtf, general transfer format; jsp, javaserver pages; png, portable network graphics; sam, sequence alignment/map; SNP, single-nucleotide polymorphism; svg, scalable vector graphics; QUMA, Quantification Tool for Methylation Analysis; vcf, variant calling format; WBSA, Web Service for Bisulfite Sequencing Data Analysis; wig, wiggle track format; xls, Microsoft Excel file format; xml, extensible markup language.



Processing Microarray-Based DNA Methylation Data

using restriction enzymes that cleave at unmethylated CpG sites (MRE-seq). However, data generated from these enrichment techniques should undergo quality control testing with saturation and coverage analysis, sequence pattern coverage analysis, and calculation of the overall CpG enrichment [87]. For MeDIP-seq data processing, the methylated DNA enriched with methylation-specific antibodies is aligned to the reference genome using alignment software, such as Bowtie, Burrow–Wheeler Transform algorithm (BWA), Mapping and Assembly with Quality (Maq), and others. Upon alignment, the relative enrichment score, which indicates the local DNA methylation levels, is enumerated by comparing the frequencies of specific DNA fragments in the genomic region to a control experiment. This localized DNA methylation levels is subsequently transformed to the absolute DNA methylation levels via Bayesian Tool for Methylation Analysis (Batman) [88] or MEDIPS algorithm [87]. However, both Batman and MEDIPS algorithms do not include quality control on MeDIP-seq data. This shortcoming is improved with MeQA pipeline that provides quality control on the MeDIP-seq data, which also distributes the sequence reads, while enumerating DNA methylation levels [89]. Recently, Li et al. [90] proposed a combination of MeDIP-seq and MRE-seq methods for the accurate identification of intermediate methylation regions. Enrichment-based data generated from this integrated analysis is subsequently analyzed with M&M and methylCRF algorithms. M&M algorithm determines differentially methylated regions (DMRs) by modeling a probability score for different DNA methylation levels observed on MeDIP-seq and MRE-seq data [91]. The Conditional Random Fields–based methylCRF algorithm combines MeDIP-seq and MRE-seq data to predict genomewide or monoallelic DNA methylation levels using the conditional probability model. By conditioning the DNA methylation states, methylCRF algorithm can determine the potential complex relationships between methylated CpGs and related model variables, in which this complex remodeling is absent in hidden Markov model, that uses joint probability for all model variables. Therefore, methylCRF algorithm provides sufficient freedom in data selection for estimating the absolute methylation levels [92].

ChIP-Chip and ChIP Sequencing ChIP is a type of immunoprecipitation method that used to capture and examine whether proteins of interest are associated with specific genomic regions in vivo [93]. In ChIP steps, cells are first treated with formaldehyde to induce protein–DNA cross-links, and then fragmented by sonication [94,95]. Cross-linked DNA fragments are selectively immunoprecipitated using specific antibod-

177

ies, and their cross-links can be reversed by heat [94,95]. As a result, the DNA fragments isolated can be further mapped with ChIP-chip array (Affymetrix, NimbleGen, and Agilent) or ChIP sequencing. ChIP-chip is a microarray-based platform that allows the identification of DNA protein–binding sites on a genomewide scale. The computational analysis for ChIPchip is depending mainly on the type of data (e.g., histone modifications, open chromatin, DNA methylation, or transcription factor–binding sites) that have been produced by ChIP experiment. At first, the collected raw probe intensities must be quantile normalized and standardized to a common median intensity [96]. Alternatively, Model-Based Analysis of Tiling Arrays (MAT) and TileProbe are commonly used for tilling array normalization and background correction [97,98]. RMAT is an R/Bioconductor package that is built in MAT program for normalizing and analyzing ChIP-chip data [99]. Software for the Analysis and Normalisation of Data from ChIP-chip Assays of Two or more Linked Experiments (Sandcastle) is an another R package that implements a newly developed normalization process for comparative analysis of data obtained from different arrays that normalizes all datasets to a common background [100]. Furthermore, several peak detection methods are available for facilitating routine processing of ChIP-chip datasets, such as ChIPOTle, ChIPpeakAnno, Ringo, TileMap, Tilescope, and T-quantile [96,101–104]. ChIP-seq is a variant of ChIP-chip that uses highthroughput DNA sequencing for the mapping of transcription factor binding and modified histones. The initial steps for ChIP data processing are the image analysis and base calling phases. This step is usually carried out using specific software provided by NGS platforms (e.g., Illumina Genome Analyzer, Applied Biosystems’ SOLiD, and Helicos platform). Several open-source software are available for genome alignment, similar to those used for bisulfite sequencing, but not limited to BWA, Bowtie ELAND, GSNAP, and Mapping and Assembly with Qualities (MAQ) [105–107]. To identify DNA protein–binding sites from ChIP-seq data, a variety of peak calling methods have been developed. The peak calling methods are subdivided according to three types of peaks: (1) sharp [e.g., Model-Based Analysis of Chip-Seq (MACS), FindPeaks, PeakSeq, PICS (Probabilistic Inference for ChIP-Seq), and Site Identification from Short Sequence Reads (SISSRs)], (2) broad [e.g., BroadPeak, Control based ChIP-Seq Analysis Tools (CCAT), Spatial Clustering Approach for the Identification of ChIP-Enriched Regions (SICER), and RSEG], and (3) mixed [e.g., SPP, Hypergeometric Optimization of Motif EnRichment (HOMER), MACS2, and Zero-Inflated Negative Binomial Algorithm (ZINBA)] [108–117]. Of which, these peak callers are used to identify significantly enriched regions, such as histone modifications

III.  Methods in Epigenetics

178

12.  Computational Epigenetics

and transcription factor–binding sites, respectively, where mostly relying on a Poisson distribution. MACS predicts the transcription factor–binding sites or histone modification regions by removing redundant reads, performs read-shifting via a sliding window approach, estimates fragment length, and enumerates the false discovery rate empirically for reducing the bias for ChIP-seq signals [104]. Meanwhile, FindPeaks simply enumerates biologically significance regions without comparing the ChIP-seq signals of the experimental sample against the control sample [118]. SICER clusters an entire genome into eligible windows, scores each of the eligible windows based on the Poisson distribution, and ranks the enriched windows by assuming which windows that secure a higher score have a lower chance to be enriched, while the remaining regions within a predefined distance are combined as an “island” [113]. Alternatively, differential biological significant regions can also be detected via ChIPComp, which corrects background noise and normalizes ChIP-seq signals based on a Poisson distribution [119]. Differential biologically significant regions identified by ChIPComp have to be merged with the significantly differential enriched regions detected by other peak callers, such as diffReps, to be named as consensus candidate regions [120]. DBChIP is purposely designed to analyze the differential transcription factor–binding sites in ChIP-seq data across multiple experimental conditions [121]. RSEG, which is based on a hidden Markov model, focused on the identifying domain boundaries and finding genomic regions that were marked with differential histone modification markers [115]. HOMER contains a set of tools for Motif Discovery and NGS analysis [122]. HOMER was initially written as a de novo motif discovery algorithm in Perl and C++ languages and has been applied for finding 8–20 bp motifs in large-scale epigenomics data. Differential peak calling methods are available for identifying significant differences in two ChIP-seq signals across multiple biological conditions. One-stage DIffereNtial peak caller (ODIN) built from hidden Markov model, filters both broad and sharp peak signals, performs peak calling, determines differential peaks in two sets of ChIP-seq experiments, and calculates P value in an integrated framework via a Binomial or a Poisson distribution and a combination of both distributions [123]. ODIN is meant for determining differential peak signals in two ChIP-seq experiments without replicates. Meanwhile, THOR analyzes differential peaks in two sets of ChIP-seq experiments with replicates [124]. THOR also normalizes peak signals from replicates to housekeeping genes to determine the signal-to-noise ratios via hidden Markov model. MAnorm is designed for comparative analysis of differential binding regions between ChIP-Seq datasets, based on a Bayesian model [125]. MACS2, an extended version of MACS, is used

for calling differential peak regions, by applying a log10 likelihood ratio cutoff [126]. HOMER uses zero or one occurrence per sequence together with binomial hypergeometric enrichment calculations to identify DNA motifs from peak files [122]. Other important databases and software tools for ChIP-seq and ChIP-chip data are presented in Table 12.3.

DIFFERENTIALLY METHYLATED REGIONS The most common goal of DNA methylation is the identification of DMRs between cases and controls. Methods for DMR detection are designed based on statistic tests, such as t-test, Wilcoxon rank-sum, bump hunting, FET, ANOVA, Wald test, logistic regression, likelihood-ratio, mixture models, Shannon entropy, logistic M values, feature selection, stratification of t-tests, and linear regression with batch effect removal and peak detection [60,62]. Notably, a multiple testing correction procedure is required for DMR detection at epigenomic levels (e.g., multiple methylated loci and regions), by applying false discovery rate–adjusted P value (q value). New emerging tools are available for the identification of DMRs, such as Quantitative Differentially Methylated Regions (QDMR), BIMMER, BiSeq, Bisulfighter, bsseq, BSmooth, City of Hope CpG Island Analysis Pipeline (COHCAP), dmrFinder [135], CpG_MPs, Differential Methylation Analysis Package (DMAP), DMR-cate, FastDMA, MethPipe, methylKit, MOdel-based Analysis of Bisulfite Sequencing data (MOABS), Specific Methylation Analysis and Report Tool (SMART), swDMR, and others [136–151]. Table 12.4 lists R/Bioconductor packages that are useful for DMR analysis. QDMR is an online tool that applies Shannon entropy for the identification and quantification DMRs from genomewide methylation profiles [136]. BSmooth implements Bowtie 2 or Merman aligner for the alignment of WGBS data, performs quality control, and identifies DMRs by taking biological variability into account [138]. BiSeq, a R/Bioconductor package, implements an algorithm to detect DMRs in the targeted bisulfite sequencing data obtained from RRBS [139]. MethyPipe directly align whole-genome bisulfite sequencing data via BWA, and apply sliding window approach for the identification of DMRs [169]. MOABS uses Beta Binomial hierarchical model to identify DMRs and differential methylated cytosines with a 10-fold coverage at single-base resolution [147]. FastDMA and DMRcate have been applied to identify DMRs from Infinium 450K Beadchip microarray data [142–149]. Bisulfighter integrates LAST alignment tool for methylated cytosine calling and comparative methylomics (ComMet) algorithm for DMR determination [145]. ComMet framework built

III.  Methods in Epigenetics

Programming languages

Input formats

Output formats

ChIPDiff

C++

text

ChIPnorm

C++

dbHiMo



TABLE 12.3 Databases and Software Tools for Histone Modification Analysis References

xls

ChIPDiff applies hidden Markov model to predict differential histone modification sites at genomewide levels by comparingtwo ChIP-seq libraries

http://cmb.gis.a-star. edu.sg/ChIPSeq/ paperChIPDiff.htm

[127]

text

gif, text

ChIPnorm is a two-stage statistical method that normalizes and removes background noise and local genomic bias in ChIP-seq data, while identifying differential histone modification regions in cells

http://lcbb.epfl.ch/ software.html

[128]

HTML, Php, Css3 Java script, Perl

fasta

jpeg, fasta, csv, png

dbHiMo built from hidden Markov model is a web-based epigenomics platform that facilitates functional and comparative analysis of histone modifying enzymes

http://hme.riceblast. snu.ac.kr/

[129]

Epidaurus

Python and C

bigwig bed

xls, pdf, png

Epidaurus is a standalone software that shows interdataset differences and relevance via data aggregation, integration, and visualization. Datasets that can be analyzed by Epidaurus includes DNase-seq, FAIRE-seq, and histone and transcription factor ChIP-seq. It also allows the users to analyze their customized datasets

http:// bioinformaticstools. mayo.edu:8080/ Epidaurus/

[130]

EpiRegNet

Perl, Php

Roadmap Epigenomics Project, ENCODE

png, text

EpiRegNet, a web server that identifies epigenetic factors for gene regulation, performs colocalization analysis between epigenetic factors and transcription factors, and constructs epigenetic regulatory network of histone marks

http://jjwanglab.org/ EpiRegNet/

[131]

HistoneHits

AJAX, Google Web Toolkit version 1.4, Perl-CGI script

Published systematic screens

text, HTML, xls

HistoneHits is a database of histone mutants and their phenotypes. Their phenotypes include single and multiple base substitution(s), posttranslational modifications, and evolutionary conservation. It also allows users to submit their new histone modification data

http://54.235.254.95/ histonehits/

[40]

HHMD

Jsp, Java

UCSC, ENCODE Canada’s Michael Smith Genome Sciences Centre, Laboratory of Molecular Immunology, National Heart Lung and Blood Institute, NIH

bed, png

HHMD is a cross-linked database that stores and integrates histone modification datasets obtained from UCSC, ENCODE Canada’s Michael Smith Genome Sciences Centre, and Laboratory of Molecular Immunology, National Heart Lung and Blood Institute, NIH. It contains manually curated histone modifications information for nine human cancers, such as colon, lymphoma, stomach, prostate, breast, lung, ovary, and pancreas cancers, and leukemia

http://bioinfo.hrbmu. edu.cn/hhmd; http://www.hhmd. org

[7]

HIstome

Python, Php, XHTML, Java script

PubMed, UniprotK, Swiss-Prot

xls, png

HIstome is a manually curated database that displays human histone proteins, their variants, and their sites of posttranslational modifications

http://www.iiserpune. ac.in/∼coee/ histome/; http:// www.actrec.gov.in/ histome/

[38]

179

Web links

Differentially Methylated Regions

III.  Methods in Epigenetics

Key points

(Continued)

Input formats

Output formats

HistoneDB 2.0

Python

NCBI, fasta

ModMine

Perl, Python, Ruby, Java, Java script

MUSIC

SysPTM 2.0

Key points

Web links

References

fasta

HistoneDB 2.0 with variants stores the manually and automatically curated histone protein sequences, and the annotated histone variants for subsequent Hidden Markov Models training. It performs multiple sequence alignments on various histone protein sequences, displays phylogenetic trees of histone variants, and allows a comparative analysis of histone protein sequences

http://www.ncbi.nlm. nih.gov/projects/ HistoneDB2.0

[39]

bir-tab, modENCODE, PubMed, xml, fasta, gff

xgmml, svg, xml, sif, tabdelimited text, commadelimited text, json

ModMine is a data warehouse for modENCODE project. It stores modENCODE data, which can be categorized as gene sequences, mRNA, and noncoding RNA expression profiles, transcription factor–binding sites, chromatin structures, DNA replication, copy number variations, and histone modification, and replacement data of Caenorhabditis elegans and Drosophila melanogaster. It allows a fine-grained analysis of modENCODE and nonmodENCODE datasets

http://intermine. modencode.org/

[132]

C++

bam

bed

MUSIC algorithm filters systematic noise, uses multiscale decomposition, and mappability correction methods to determine enriched regions in ChIP-Seq

https://github.com/ gersteinlab/MUSIC

[133]

Perl, Java

PubMed

png

SysPTM integrates 10 external resources of posttranslational modifications and contains various analysis tools, such as PTMBlast, PTMPathway, PTMPhylog, PTMCluster, SysPTM-H, and PTMGO. It annotates posttranslational modifications of proteins, while allowing systematic investigation of their function.

http://lifecenter.sgst. cn/SysPTM/

[134]

bam, Binary alignment/map; bed, browser extensible data; csv, comma-seperated value; dbHiMo, Database for Histone-Modifying Enzymes; EpiRegNet, Epigenetic Regulatory Network; gif, graphics interchange format; gff, general feature format; HHMD, Human Histone Modification Database; HIstome, Histone Infobase; jsp, javaserver pages; modENCODE, Model Organism Encyclopaedia of DNA Elements; png, portable network graphics; sif, simple interaction format; svg, scalable vector graphics; xgmml, extensible graph markup and modeling language; xls, Microsoft Excel file format; xml, extensible markup language.

12.  Computational Epigenetics

III.  Methods in Epigenetics

Programming languages

180

TABLE 12.3  Databases and Software Tools for Histone Modification Analysis (cont.)



TABLE 12.4 R/Bioconductor Package for Methylation Analysis Web links

References

Aclust

Aclust uses A-clustering algorithm to detect neighboring CpG sites that are correlated with each other. It identifies coregulated methylation regions that are associated with environmental exposure via a generalized estimating equation

http://www.hsph.harvard.edu/tamar-sofer/packages/

[152]

BayMeth

BayMeth for MeDIP-seq and MBD-seq data analyses, uses an empirical Bayesian approach to transform the observed read densities into regional methylation levels. It also distinguishes low methylation levels for inefficient capture

http://bioconductor.org/packages/release/bioc/ html/Repitools.html

[153]

BiSeq

BiSeq identifies DMRs from targeted bisulfite sequencing data

https://www.bioconductor.org/packages/release/ bioc/html/BiSeq.html

[139]

Bump hunting

Bump hunting is a generic method that integrates surrogate variable analysis with regression modeling, smoothing method, and multiple comparison method to detect DMRs from Infinium 450K BeadChip data

http://bioconductor.org/packages/release/bioc/ html/bumphunter.html

[154]

COHCAP

COHCAP integrates DNA methylation data with gene expression profiles without normalizing the raw data

https://bioconductor.org/packages/release/bioc/ html/COHCAP.html

[141]

CompEpiTools CompEpiTools generates read counts metrics in genomic regions, annotates genomic regions, as well as functional enhancers and lncRNAs, and integrates visualization of heterogeneous data types

http://bioconductor.org/packages/release/bioc/ html/compEpiTools.html

[148]

DSS-single

DSS-single implements Wald statistics for count-based sequencing data. It identifies DMRs from whole-genome bisulfite sequencing data without replicates. It also performs spatial correlation and Bayes shrinkage estimating procedure between methylation levels, biological variations, and sequence depth

https://www.bioconductor.org/packages/release/ bioc/html/DSS.html

[155]

EDMR

EDMR is built upon EM algorithm to detect and annotate regional methylation

https://github.com/ShengLi/edmr

[156]

ELMER

ELMER uses DNA methylation data for the identification of enhancers. It relates DNA methylation and gene expression data to determine transcription factor network and regulatory element landscape in primary tissue

https://bioconductor.org/packages/release/bioc/ html/ELMER.html

[157]

ICDMR

ICDMR is an unsupervised method that clusters and determines DMRs from individuals and between subgroup in a population

http://www.csjfann.ibms.sinica.edu.tw/eag/programlist/ICDMR/ICDMR.html

[158]

IMA

IMA enumerates methylation index via mean, median and Tukey’s Biweight robust average approaches. It also performs differential methylation analysis based on Bayes statistics(Wilcoxon rank-sum test, Student’s t-test, and general linear models), and multiple testing correction algorithms(Bonferroni correction). It summarizes site- and region-levels methylation change from Illumina 450K DNA methylation data

https://rforge.net/IMA/

[159]

Minfi

Minfi identifies DMRs from kilobase to megabase scale of Infinium microarrays data via bump hunting and block-finding approaches

http://bioconductor.org/packages/release/bioc/ html/minfi.html

[160]

Repitools

Repitools is a toolbox that visualizes, analyzes, and integrates enrichment-based ChIP-seq and MeDIP-seq epigenomic data in the context of differential methylation and gene expression

http://bioconductor.org/packages/release/bioc/ html/Repitools.html

[161]

M3D

M3D is a nonparametric kernel-based test that identifies spatial DNA methylation profiles via maximum mean discrepancy method. It also determines a methylation profile that is not influenced by interreplicate variations

http://watson.nci.nih.gov/bioc_mirror/packages/ release/bioc/html/M3D.html

[162]

181

(Continued)

Differentially Methylated Regions

III.  Methods in Epigenetics

Key points

Key points

References

MethyAnalysis MethyAnalysis analyzes DNA methylation profiles based on chromosomal location. It visualizes DNA methylation data and annotates differentially methylation regions via sliding window approach

https://www.bioconductor.org/packages/release/ bioc/html/methyAnalysis.html

[2]

MethylKit

MethylKit clusters and annotates methylation and hydroxymethylation sequencing data. It allows quality data visualization and differential methylation analysis

https://github.com/al2na/methylKit; http://code. google.com/p/methylkit

[137]

MethylMix

MethylMix implements beta mixture model that can determine the hyper- and hypomethylated genes in a disease and calculates the differential methylation value

https://bioconductor.org/packages/release/bioc/ html/MethylMix.html

[163]

MethyPipe

MethyPipe aligns bisulfite sequencing read via Burrow–Wheeler Transform algorithm. It also annotates methylation data and identifies DMRs

http://sunlab.lihs.cuhk.edu.hk/methy-pipe/

[164]

MethySig

MethySig uses a beta-binomial approach to determine genomewide methylation differences across biological samples with regard to different treatments or disease groups

http://sartorlab.ccmb.med.umich.edu/software

[165]

MethVisual

MethVisual imports and aligns DNA sequences. It performs quality control on the aligned sequence using Needleman–Wunsch algorithm. It utilizes lollipop diagram to indicate neighboring DNA methylation and distant CpG sites

http://methvisual.molgen.mpg.de/

[166]

M&M

M&M algorithm determines DMRs by generating a probability score for the observed differences in DNA methylation levels that are obtained from MeDIP-seq and MRE-seq data. It also calculates false discovery rate via Benjamini–Hochberg method

https://www.bioconductor.org/packages/devel/ bioc/html/methylMnM.html

[91]

Rnbeads

RnBeads is applicable for large-scale DNA methylation analysis, such as epigenomewide association studies. It generates detailed annotated hypertext reports after analyzing preprocessed data of MeDIP-seq, MBD-seq, bisulfite sequencing, and Infinium microarray

http://rnbeads.mpi-inf.mpg.de/

[167]

WFMM

WFMM is build from a Bayesian approach which can model the complex irregular methylation datasets within a linear mixed models

https://biostatistics.mdanderson.org/SoftwareDownload/

[168]

A-clustering, Adjacent site clustering; COHCAP, City of Hope CpG Island Analysis Pipeline; DMR, differentially methylated regions; EDMR, Empirically Differentially Methylated Regions; ELMER, Enhancer Linking by Methylation/Expression Relationships; EM, expectation maximization; ICDMR, Identification of Consistently Differentially Methylated Regions; IMA, Illumina Methylation Analyzer; WFMM, WaveletBased Functional Mixed Models.

12.  Computational Epigenetics

III.  Methods in Epigenetics

Web links

182

TABLE 12.4  R/Bioconductor Package for Methylation Analysis (cont.)



Prediction of Epigenetic Status

from expectation-maximization algorithm for hidden Markov model is used to adjust the parameters for the dataset. BIMMER applies hidden Markov model and expectation-maximization algorithm to detect DMRs for MBDCap-seq data [144]. DMAP uses fragment-based approach to produce coverage-filtered reference methylomes, and determines DMRs from RRBS and WGBS data via statistical analysis of Chi-square test, Fisher exact tests, and ANOVA [146]. SwDMR combines statistical tools, such as t-test, Wilcoxon test, Chi-square test, Fisher-exact test, Kruskal–Wallis test, and ANOVA, to determine DMRs from WGBS data based on sliding window method [150]. SMART is an entropy-based framework that integrates the WGBS methylomes and identifies DMRs across multiple cell types [151].

PREDICTION OF EPIGENETIC STATUS Computational modeling and algorithms, to detect the presence or absence of DNA methylation in a given genomic region, can be performed with DNA methylation prediction. DNA methylation prediction algorithms can accurately predict epigenome based on the properties of genome, such as DNA sequence, DNA structure, repetitive DNA elements, transcription factor–binding sites, number of SNPs, guanine to cytosine (GC) content, Alu elements, evolutionary conservation, histone modification marks, and methylation state of neighboring regions [166,170–174]. Indeed, our previous work suggested that DNA methylation prediction might be useful to study the epigenetic information of disease-related genes [175], before performing the laboratory- and clinical-based experiments [176]. This cost-saving and timesaving approach is applicable for epigenetic studies in human diseases [177,178]. A large number of computational tools and databases have been developed to predict DNA methylation state of CpG dinucleotides, such as CpGProD, CpGcluster, DeepMethyl, iDNA-Methyl, Methylator, MethDB, MethFinder, Epigram, EpiGRAPH, EMBOSS genome browsers, and so on [19,29,72,179–185]. Most of these methods are based on DNA sequence characteristics combined with a machine learning algorithms (Support Vector Machine and Artificial Neural Network Models) [182,186]. The Epigram pipeline can capture the cis-elements to predict histone modification and DNA methylation profiles from DNA motifs [184]. For instance, a GC-rich DNA motif tends to interact with long noncoding RNAs to recruit Polycomb repressive complex 2 to the binding region of histone 3 lysine 27 trimethylation (H3K27me3) for gene repression [184]. Moreover, an epigenome prediction pipeline (EpiGRAPH), developed to predict the overall strength of CpG islands and the maps of bona fide CpG islands, by integrating the predicted

183

unmethylated score, predicted promoter activity score and open chromatin score, such as histone H3K4me2, H3K4me3, H3K9/14 acetylation, DNase I hypersensitivity, and SP1 binding [185]. Besides, the epigenomic data are extremely useful in predicting chromatin interaction [187], their association with diseases [188,189] and phenotypes, such as hypertriglyceridemic waist [190]. Furthermore, histone modification ChIP-chip and ChIP-seq data can be used to predict transcriptional factor–binding sites via Chromia [191,192]. Chromia, built from the hidden Markov model, is an integrated tool that identifies the functional target loci for transcription factors at genomewide levels [193]. In contrast, histone modifications can be predicted with remarkable accuracy from transcription factor– binding profiles through logistic regression classifiers [194]. Differential transcription factor–binding patterns have been associated with differential gene expression levels, of which, a particular set of interacting transcriptional factors may increase or decrease the gene expression activity [175]. Gene expression profiles can be predicted from DNA methylation data via several algorithms [195,196]. A machine learning–based algorithm that uses Relief-based feature selection and random forest classification method is deriving differential gene expression from DNA methylation and histone H3 methylation modification data [195]. This algorithm is particularly important for formalin-fixed paraffin-embedded tumor samples, in which the results of gene expression might be impacted by RNA degradation during sample handling and storage steps [195]. Following the success in deriving gene expression profiles from DNA methylation data, Baur and Bozdag [197] developed a sequential feature selection algorithm using K-Nearest Neighbors to select the probes for gene expression experiment. Additionally, Binding and Expression Target Analysis (BETA) may be used for the analysis of ChIP-seq data in combination with differential gene expression data, to predict whether a transcription factor is enhancing or repressing gene activity, and to locate transcription factors–binding sites, that is, DNA motifs and target genes. It consists of three subprotocols, that is, BETA-basic, -plus, and -minus [197]. BETA-basic and -plus predict if a transcription factor activates or represses the gene, BETA-plus also identifies direct target genes and analyzes DNA motifs in the targeted regions. If the binding data of a transcription factor is available, a user can opt for BETA-minus [197]. Interestingly, it has been proposed that aging can be predicted by DNA methylation [198]. DNA Methylation Age Calculator estimates the DNA methylation age of most tissues and cell types, by using the methylation data measured from Illumina DNA Infinium platform (e.g., 27K or 450K) that consists of an identifier for CpG probes. The prediction depends on several

III.  Methods in Epigenetics

184

12.  Computational Epigenetics

characteristics: (1) value of DNA methylation: the value of DNA methylation age is close to zero for embryonic and induced pluripotent stem cells; (2) cell passage number: a significant correlation between DNA methylation age and cell passage number exists, (3) heritability of age acceleration: it is highly heritable, and (4) applicablity to different tissues: it is applicable to heart, liver, and kidney tissues in chimpanzee model [198].

INTEGRATIVE EPIGENOMICS Integration of epigenomics data (both DNA methylation and histone modification), along with other omics data, provides new insights into the complex regulation of biological systems. Clustering- and machine learningbased methods can be used to classify the genome into different chromatin states and to explore the interplay between epigenetic signatures and gene expression. Clustering-based methods, such as hierarchical clustering and k-means clustering, have been widely applied in the integration of epigenomics data. Chromatin-Profile Alignment followed by Tree-Clustering Algorithm (ChAT) is an unsupervised hierarchical clustering method that discovers andenomic regions and characterizes chromatin signatures at genomewide levels [199]. ChromaSig can be used to find, cluster, align, and orient common chromatin signatures and to discover motifs related to transcriptional enhancers and promoters from ChIPseq and ChIP-chip data, via hierarchical clustering with a Euclidean distance measure [200]. Chromatin Analysis and Exploration tool (ChAsE) is built from k-means clustering approach that reveals the correlations between DNA methylation, histone modifications, and transcription factors by using interactive heat maps and plot interfaces [201]. Hidden Markov model was proposed to annotate and characterize distinct chromatin states, including promoter, transcription, and intergenic regions [202]. ChIPmeta is used for the joint analysis of ChIP-chip and ChIP-seq data, by applying hierarchical hidden Markov model [94]. ChromHMM was then developed based on multivariate hidden Markov model, which learns and characterizes chromatin states, as well as integrates multiple chromatin datasets and visualizes the genomewide maps of chromatin state annotations in genome browsers [202]. Later, a sparsely correlated Hidden Markov Model (scHMM) was proposed for better mapping of epigenomic states, by performing simultaneous hidden Markov model inference for multiple genomic datasets [203]. Genomic Annotation using Temporal Epigenomic data (GATE) model is a combination of finite mixture model and hidden Markov model that can be used to annotate the epigenomic states based on the dynamic changes of epigenetic modifications [204]. Tree-hmm

uses a Bayesian tree hidden Markov model to determine chromatin markers at different chromatin states and the chromatin states at different disease progression stages [205]. Spectral Learning for Annotating Chromatin Labels and Epigenomes (Spectacle) implements spectral learning algorithm for hidden Markov model, instead of expectation-maximization algorithm, to annotate functional chromatin state and to identify enhancer subtypes enriched in GWAS SNPs [206]. Bayesian network has been proposed to infer causal and/or combinatorial relationships among different histone modifications and gene expression [207]. Segway, a dynamic Bayesian network model can identify the joint patterns across histone modification, transcription factor, and open chromatin datasets [208]. SeqSpider is a Bayesian network structure inference algorithm that predicts regulatory relationships and/or interactions between different biological factors, by performing the integrated analysis of heterogeneous epigenomic datasets (e.g., ChIP-seq data, BS-seq data, and RNA-seq data) [209]. Sparse partial correlation network determines the association among histone modifications, by deriving undirected networks depending on partial correlations between histone modifications [210]. Hierarchically linked infinite hidden Markov model (hiHMM), a new Bayesian nonparametric method, infers chromatin state maps in multiple genomes across different species, cell types, and developmental stages [41]. Epigenomics data can be further integrated with genomics, transcriptomics, proteomics, and metabolomics data, toward a more complete picture of the regulation and function of complex biological systems. The ENCODE project has released comprehensive resources of 3D chromatin structure, DNA methylation, RNA expression, transcription factors, DNA accessibility, DNA– and RNA–protein interactions, across different cell and tissue types in the model organisms of Homo sapiens, Mus musculus, D. melanogaster, and C. elegans [44]. Multiple datasets include raw sequencing data, processed data, reference data, experimental metadata, software tools, and other related documentation that are accessible at ENCODE Portal (www.encodeproject.org) or through application program interface (www.encodeproject. org/help/rest-api/) [211]. These data can be used for the integrative analysis of diverse omics datasets. When clubbed with integrative bioinformatics methods, it can help to elucidate the molecular mechanisms underlying diseases, and identify novel biomarkers and therapeutic targets. For instance, SIGMA2 is designed for an integrative analysis of genomics (e.g., DNA copy number and allelic imbalance), epigenomics (e.g., DNA methylation and histone modification), and transcriptomics (e.g., gene and miRNA expression) in the context of cancer [212]. Cistrome is an integrative pipeline analysis that built upon the Galaxy open source framework, and has

III.  Methods in Epigenetics



REFERENCES

29 ChIP-chip– and ChIP-seq–specific tools focused on peak calling, gene expression analysis, and motif discovery via SeqPos algorithm [213]. InCroMAP is another tool that specifically integrates omics data from different microarray platforms and extends to the visualization of metabolic and signaling pathways [214]. ChroGPS has been proposed to integrate the genomic and epigenomic data in a single map, based on multidimensional scaling techniques [215]. Integration analysis of cancer genomics, epigenomics, and transcriptomics data may also be achieved by the use of LASSO and ENET penalized regression approaches in combination with a permutation-based MaxT method [216]. Integrated Clustering of Multidimensional Biomedical Data (ICM) is a newly developed web server for integrated clustering of clinical data and multiomics data, by applying three different types of algorithms, namely Similarity Network Fusion (SNF), integrative clustering (iCluster), and Concatenation. The obtained clustering data can be presented in the form of interactive heat map, network, and table [217].

CONCLUDING REMARKS The emerging field of computational epigenetics is moving from hypothesis-driven, reductionist, and bottom-up approach toward a more data-driven, holistic, and top-down approach. There is a need for a paradigm shift toward the field of computational epigenetics. Over last five years, the availability of efficient computational tools for DNA methylation, histone modifications, transcription factor binding, nucleosome positioning, and chromosomal organization has become increasingly important for relevant researchers. Integrative data analysis of multiomics data,by combining bioinformatics, statistics, and mathematical systems biology approaches, could contribute greatly to our understanding of epigenetic modification and transcriptional regulation at the systemic level. However, it is important to appreciate that the development of next-generation sequencing and microarray platforms pose major challenges and opportunities for computational epigenetics, with regards to data analysis, management, storage, processing, and interpretation. With growing volumes and increased complexity of epigenomic data generated by these high-throughput experimental methods, data analysis typically requires a large storage capacity and high-performance computation engine. Even though many tools have been developed for analyzing multiomics data, the efforts of data integration and interpretation still provide a room for improvement. Therefore, improving prediction accuracy and maintaining computational efficiency of the available bioinformatic tools can lead to a fruitful computational epigenome research project.

185

Acknowledgments We acknowledge the support from Ministry of Higher Education Fundamental Research Grant Scheme (FRGS/1/2015/SKK08/ UTAR/02/3) and Universiti Tunku Abdul Rahman under the UTAR Research Fund 6200/LG7. We appreciate Shu Chai Ching for her help in references preparation.

References [1] Allis CD, Jenuwein T. The molecular hallmarks of epigenetic control. Nat Rev Genet 2016;1–14. [2] Du J, Johnson LM, Jacobsen SE, Patel DJ. DNA methylation pathways and their crosstalk with histone methylation. Nat Rev Mol Cell Biol 2015;16(9):519–32. [3] Laird PW. Principles and challenges of genome-wide DNA methylation analysis. Nat Rev Genet 2010;11(3):191–203. [4] Sant KE, Nahar MS, Dolinoy DC. DNA methylation screening and analysis. Developmental toxicology: methods and protocols. New York, NY: Springer; 2012. [pp. 385–406]. [5] Plongthongkum N, Diep DH, Zhang K. Advances in the profiling of DNA modifications: cytosine methylation and beyond. Nat Rev Genet 2014;15(10):647–61. [6] Moran S, Martínez-Cardús A, Sayols S, Musulén E, Balañá C, Estival-Gonzalez A, et al. Epigenetic profiling to classify cancer of unknown primary: a multicentre, retrospective analysis. Lancet Oncol 2016;17:1386–95. [7] Dong X, Weng Z. The correlation between histone modifications and gene expression. Epigenomics 2013;5:113–6. [8] Bannister AJ, Kouzarides T. Regulation of chromatin by histone modifications. Cell Res 2011;21(3):381–95. [9] Tessarz P, Kouzarides T. Histone core modifications regulating nucleosome structure and dynamics. Nat Rev Mol Cell Biol 2014;15(11):703–8. [10] Kimura H. Histone modifications for human epigenome analysis. J Human Genet 2013;58(7):439–45. [11] Telese F, Ma Q, Perez PM, Notani D, Oh S, Li W, et al. LRP8-Reelin-regulated neuronal enhancer signature underlying learning and memory formation. Neuron 2015;86(3):696–710. [12] Hubbard T, Barker D, Birney E, Cameron G, Chen Y, Clark L, et al. The Ensembl genome database project. Nucleic Acids Res 2002;30(1):38–41. [13] Karolchik D, Hinrichs AS, Kent WJ. The UCSC genome browser. Nucleic Acids Res 2004;32:D493–6. [14] Zhou X, Wang T. Using the wash U Epigenome browser to examine genome-wide sequencing data. Curr Protoc Bioinformatics 2012;(Suppl. 40):1–14. [15] Karnik R, Meissner A. Browsing (epi)genomes: a guide to data resources and epigenome browsers for stem cell researchers. Cell Stem Cell 2013;13(1):14–21. [16] Gu F, Doderer MS, Huang YW, Roa JC, Goodfellow PJ, Kizer EL, et al. CMS: a web-based system for visualization and analysis of genome-wide methylation data of human cancers. PLoS One 2013;8(4):e60980. [17] Kuo H-C, Lin P-Y, Chung T-C, Chao C-M, Lai L-C, Tsai M-H, et al. DBCAT: database of CpG islands and analytical tools for identifying comprehensive methylation profiles in cancer cells. J Comput Biol 2011;18(8):1013–7. [18] Raies AB, Mansour H, Incitti R, Bajic VB. DDMGD: the database of text-mined associations between genes methylated in diseases from different species. Nucleic Acids Res 2015;43: D879–86. [19] Wang Y, Liu T, Xu D, Shi H, Zhang C, Mo Y-Y, et al. Predicting DNA methylation state of CpG dinucleotide using genome topological features and deep networks. Sci Rep 2016;6:19598.

III.  Methods in Epigenetics

186

12.  Computational Epigenetics

[20] Lv J, Liu H, Su J, Wu X, Liu H, Li B, et al. DiseaseMeth: a human disease methylation database. Nucleic Acids Res 2012;40:1030–5. [21] Halachev K, Bast H, Albrecht F, Lengauer T, Bock C. EpiExplorer: live exploration and global analysis of large epigenomic datasets. Genome Biol 2012;13(10):R96. [22] Medvedeva YA, Lennartsson A, Ehsani R, Kulakovskiy IV, Vorontsov IE, Panahandeh P, et al. EpiFactors: a comprehensive database of human epigenetic factors and complexes. Database 2015;2015:1–10. [23] Cho SY, Chai JC, Park SJ, Seo H, Sohn CB, Lee YS. EPITRANS: a database that integrates epigenome and transcriptome data. Mol Cells 2013;36(5):472–5. [24] Bai W, Yang W, Wang W, Wang Y, Liu C, Jiang Q, et al. GED: a manually curated comprehensive resource for epigenetic modification of gametogenesis. Brief Bioinform 2017;18(1):53–9. [25] Baek SJ, Yang S, Kang TW, Park SM, Kim YS, Kim SY. MENT: methylation and expression database of normal and tumor tissues. Gene 2013;518(1):194–200. [26] Geisen S, Barturen G, Alganza ÁM, Hackenberg M, Oliver JL, NGSmeth DB. An updated genome resource for high quality, single-cytosine resolution methylomes. Nucleic Acids Res 2014;42:53–9. [27] Zou D, Sun S, Li R, Liu J, Zhang J, Zhang Z. MethBank: database integrating next-generation sequencing single-base-resolution DNA methylation programming data. Nucleic Acids Res 2015;43:D54–8. [28] Song Q, Decato B, Hong EE, Zhou M, Fang F, Qu J, et al. A reference methylome database and analysis pipeline to facilitate integrative and comparative epigenomics. PLoS One 2013;8(12):e81148. [29] Grunau C, Renault E, Rosenthal A, Roizes G. MethDB—a public database for DNA methylation data. Nucleic Acids Res 2001;29(1):270–4. [30] Huang W-Y, Hsu S-D, Huang H-Y, Sun Y-M, Chou C-H, Weng S-L, et al. MethHC: a database of DNA methylation and gene expression in human cancer. Nucleic Acids Res 2015;43:D856–61. [31] Xin Y, Chanrion B, O’Donnell AH, Milekic M, Costa R, Ge Y, et al. MethylomeDB: a database of DNA methylation profiles of the brain. Nucleic Acids Res 2012;40:1245–9. [32] He X, Chang S, Zhang J, Zhao Q, Xiang H, Kusonmano K, et al. MethyCancer: the database of human DNA methylation and cancer. Nucleic Acids Res 2008;36(Suppl. 1):836–41. [33] Shi J, Hu J, Zhou Q, Du Y, Jiang C. PEpiD: a prostate epigenetic database in mammals. PLoS One 2013;8(5):e64289. [34] Ongenaert M, Van Neste L, De Meyer T, Menschaert G, Bekaert S, Van Criekinge W. PubMeth: a cancer methylation database combining text-mining and expert annotation. Nucleic Acids Res 2008;36(Suppl. 1):842–6. [35] Díez-Villanueva A, Mallona I, Peinado MA. Wanderer, an interactive viewer to explore DNA methylation and gene expression data in human cancer. Epigenetics Chromatin 2015;8(1):22. [36] Fingerman IM, McDaniel L, Zhang X, Ratzat W, Hassan T, Jiang Z, et al. NCBI epigenomics: a new public resource for exploring epigenomic data sets. Nucleic Acids Res 2011;39(Suppl. 1):D908–12. [37] Marzi MJ, Ghini F, Cerruti B, De Pretis S, Bonetti P, Giacomelli C, et al. Degradation dynamics of microRNAs revealed by a novel pulse-chase approach. Genome Res 2016;26:554–65. [38] Khare SP, Habib F, Sharma R, Gadewal N, Gupta S, Galande S. HIstome—a relational knowledgebase of human histone proteins and histone modifying enzymes. Nucleic Acids Res 2012;40:1–6. [39] Draizen EJ, Shaytan AK, Mariño-Ramírez L, Talbert PB, Landsman D, Panchenko AR. HistoneDB 2.0: a histone database with variants-an integrated resource to explore histones and their variants. Database 2016;2016:baw014. [40] Huang H, Maertens AM, Hyland EM, Dai J, Norris A, Boeke JD, et al. HistoneHits: a database for histone mutations and their phenotypes. Genome Res 2009;19(4):674–81.

[41] Sohn K-A, Ho JWK, Djordjevic D, Jeong H-h, Park PJ, Kim JH. hiHMM: Bayesian non-parametric joint inference of chromatin state maps. Bioinformatics 2015;btv117. [42] Adams D, Altucci L, Antonarakis SE, Ballesteros J, Beck S, Bird A, et al. BLUEPRINT to decode the epigenetic signature written in blood. Nat Biotechnol 2012;30(3):224–6. [43] Kundaje A, Meuleman W, Ernst J, Bilenky M, Yen A, HeraviMoussavi A, et al. Integrative analysis of 111 reference human epigenomes. Nature 2015;518(7539):317–30. [44] Consortium EP. An integrated encyclopedia of DNA elements in the human genome. Nature 2012;489(7414):57–74. [45] Li Y, Tollefsbol TO. DNA methylation detection: bisulfite genomic sequencing analysis. Epigenet Protoc 2011;11–21. [46] Li H, Ruan J, Durbin R. Mapping short DNA sequencing reads and calling variants using mapping quality scores. Genome Res 2008;18(11):1851–8. [47] Smith AD, Chung W-Y, Hodges E, Kendall J, Hannon G, Hicks J, et al. Updates to the RMAP short-read mapping software. Bioinformatics 2009;25(21):2841–2. [48] Xi Y, Bock C, Muller F, Sun D, Meissner A, Li W. RRBSMAP: a fast, accurate and user-friendly alignment tool for reduced representation bisulfite sequencing. Bioinformatics 2012;28(3):430–2. [49] Coarfa C, Yu F, Miller CA, Chen Z, Harris RA, Milosavljevic A. Pash 3.0: a versatile software package for read mapping and integrative analysis of genomic and epigenomic variation using massively parallel DNA sequencing. BMC Bioinform 2010;11(1): 572. [50] Chen P-Y, Cokus S, Pellegrini M. BS Seeker: precise mapping for bisulfite sequencing. BMC Bioinform 2010;11(1):203. [51] Li H, Homer N. A survey of sequence alignment algorithms for next-generation sequencing. Brief Bioinform 2010;11(5):473–83. [52] Krueger F, Andrews SR. Bismark: a flexible aligner and methylation caller for Bisulfite-Seq applications. Bioinformatics 2011;27(11):1571–2. [53] Sun R, Tian Y, Chen X. TAMeBS: a sensitive bisulfite-sequencing read mapping tool for DNA methylation analysis. Bioinformatics and biomedicine (BIBM), 2014 IEEE international conference. IEEE; 2014. pp. 176–181. [54] Chen H, Smith AD, Chen T. WALT: fast and accurate read mapping for bisulfite sequencing. Bioinformatics 2016;22:3507–9. [55] Harris EY, Ounit R, Lonardi S. BRAT-nova: fast and accurate mapping of bi-sulfite-treated reads. Bioinformatics 2016;btw226. [56] Kunde-Ramamoorthy G, Coarfa C, Laritsky E, Kessler NJ, Harris RA, Xu M, et al. Comparison and quantitative verification of mapping algorithms for whole-genome bisulfite sequencing. Nucleic Acids Res 2014;42(6):e43. [57] Liu Y, Siegmund KD, Laird PW, Berman BP. Bis-SNP: combined DNA methylation and SNP calling for Bisulfite-seq data. Genome Biol 2012;13(7):1. [58] Kent WJ, Zweig AS, Barber G, Hinrichs AS, Karolchik D. BigWig and BigBed: enabling browsing of large distributed datasets. Bioinformatics 2010;26(17):2204–7. [59] Nicol JW, Helt GA, Blanchard SG, Raja A, Loraine AE. The Integrated Genome Browser: free software for distribution and exploration of genome-scale datasets. Bioinformatics 2009;25(20):2730–1. [60] Robinson JT, Thorvaldsdóttir H, Winckler W, Guttman M, Lander ES, Getz G, et al. Integrative genomics viewer. Nat Biotechnol 2011;29(1):24–6. [61] Chelaru F, Smith L, Goldstein N, Bravo HC. Epiviz interactive visual analytics for functional genomics data. Nat Methods 2014;11(9):938–40. [62] Bock C. Analysing interpreting DNA methylation data. Nat Rev Genet 2012;13(10):705–19. [63] Tran H, Wu X, Tithi S, Sun M-A, Xie H, Zhang L. A Bayesian assignment method for ambiguous bisulfite short reads. PLoS One 2016;11(3):e0151826.

III.  Methods in Epigenetics



REFERENCES

[64] Lim J-Q, Tennakoon C, Li G, Wong E, Ruan Y, Wei C-L, et al. BatMeth: improved mapper for bisulfite sequencing reads on DNA methylation. Genome Biol 2012;13:R82. [65] Tennakoon C, Purbojati RW, Sung WK. BatMis: a fast algorithm for k-mismatch mapping. Bioinformatics 2012;28(16):2122–8. [66] Rohde C, Zhang Y, Stamerjohanns H, Hecher K, Reinhardt R, Jeltsch A. New clustering module in BDPC bisulfite sequencing data presentation and compilation web application for DNA methylation analyses. BioTechniques 2009;47(3):781–3. [67] Becker D, Lutsik P, Ebert P, Bock C, Lengauer T, Walter Jr. BiQ Analyzer HiMod: an interactive software tool for high-throughput locus-specific analysis of 5-methylcytosine and its oxidized derivatives. Nucleic Acids Res 2014;42(W1):501–7. [68] Rohde C, Zhang Y, Reinhardt R, Jeltsch A. BISMA—fast and accurate bisulfite sequencing data analysis of individual clones from unique and repetitive sequences. BMC Bioinform 2010;11:230. [69] Hu K, Ting AH, Li J. BSPAT: a fast online tool for DNA methylation co-occurrence pattern analysis based on high-throughput bisulfite sequencing data. BMC Bioinform 2015;16(1):1–9. [70] Xie Q, Liu Q, Mao F, Cai W, Wu H, You M, et al. A Bayesian framework to identify methylcytosines from high-throughput bisulfite sequencing data. PLoS Comput Biol 2014;10(9):e1003853. [71] He J, Sun X, Shao X, Liang L, Xie H. DMEAS: DNA methylation entropy analysis software. Bioinformatics. 2013;29(16):2044–5. [72] Bock C, Halachev K, Büch J, Lengauer T. EpiGRAPH: user-friendly software for statistical analysis and prediction of (epi-) genomic data. DNA Seq 2009;10(2):R14. [73] Benoukraf T, Wongphayak S, Hakim L, Hadi A, Wu M, Soong R. GBSA: a comprehensive software for analysing whole genome bisulfite sequencing data. Nucleic Acids Res 2013;41(4):e55. [74] Wu TD, Nacu S. Fast and SNP-tolerant detection of complex variants and splicing in short reads. Bioinformatics 2010;26(7): 873–81. [75] Tarraga J, Perez M, Orduna JM, Duato J, Medina I, Dopazo J. A parallel and sensitive software tool for methylation analysis on multicore platforms. Bioinformatics 2015;31(19):3130–8. [76] Pedersen B, Hsieh TF, Ibarra C, Fischer RL. MethylCoder: software pipeline for bisulte-treated sequences. Bioinformatics 2011;27(17):2435–6. [77] Liao W-W, Yen M-R, Ju E, Hsu F-M, Lam L, Chen P-Y. MethGo: a comprehensive tool for analyzing whole-genome bisulfite sequencing data. BMC Genomics 2015;16(Suppl. 12):S11. [78] Ordway JM, Bedell JA, Citek RW, Nunberg AN, Jeddeloh JA. MethylMapper: a method for high-throughput, multilocus bisulfite sequence analysis and reporting. BioTechniques 2005;39(4):464–7. [79] Maspimby J-L, Grunau C. MethTools 2.0-a webtool for the generation and analysis of DNA methylation patterns from bisulfite sequencing raw data. Available from: https://www.researchgate. net/publication/255645896_MethTools_20_-_a_webtool_for_ the_generation_and_analysis_of_DNA_methylation_patterns_ from_bisulfite_sequencing_raw_data [80] Kumaki Y, Oda M, Okano M. QUMA: quantification tool for methylation analysis. Nucleic Acids Res 2008;36:170–5. [81] Sun R, Tian Y, Chen X. TAMeBS: a sensitive bisulfite-sequencing read mapping tool for DNA methylation analysis. InBioinformatics and Biomedicine (BIBM), 2014 IEEE International Conference on 2014 Nov 2 (pp. 176-181). IEEE. [82] Liang F, Tang B, Wang Y, Wang J, Yu C, Chen X, et al. WBSA: web service for bisulfite sequencing data analysis. PLoS One 2014;9(1):1–9. [83] Dedeurwaerder S, Defrance M, Bizet M, Calonne E, Bontempi G, Fuks F. A comprehensive overview of Infinium HumanMethylation450 data processing. Brief Bioinform 2013;bbt054. [84] Pidsley R, Wong CCY, Volta M, Lunnon K, Mill J, Schalkwyk LC. A data-driven approach to preprocessing Illumina 450K methylation array data. BMC Genomics 2013;14(1):293.

187

[85] Maksimovic J, Gordon L, Oshlack A. SWAN: subset-quantile within array normalization for Illumina Infinium HumanMethylation450 BeadChips. Genome Biol 2012;13(6):R44. [86] Liu J, Siegmund KD. An evaluation of processing methods for HumanMethylation450 BeadChip data. BMC Genomics 2016;17(469):1–11. [87] Lienhard M, Grimm C, Morkel M, Herwig R, Chavez L. MEDIPS: genome-wide differential coverage analysis of sequencing data derived from DNA enrichment experiments. Bioinformatics 2014;30(2):284–6. [88] Down TA, Rakyan VK, Turner DJ, Flicek P, Li H, Kulesha E, et al. A Bayesian deconvolution strategy for immunoprecipitation-based DNA methylome analysis. Nat Biotechnol 2008;26(7):779–85. [89] Huang J, Renault V, Sengenes J, Touleimat N, Michel S, Lathrop M, et al. MeQA: a pipeline for MeDIP-seq data quality assessment and analysis. Bioinformatics 2012;28(4):587–8. [90] Li D, Zhang B, Xing X, Wang T. Combining MeDIP-seq and MRE-seq to investigate genome-wide CpG methylation. Methods 2015;72:29–40. [91] Zhang B, Zhou Y, Lin N, Lowdon RF, Hong C, Nagarajan RP, et al. Functional DNA methylation differences between tissues, cell types, and across individuals discovered using the M&M algorithm. Genome Res 2013;23(9):1522–40. [92] Stevens M, Cheng JB, Li D, Xie M, Hong C, Maire CL, et al. Estimating absolute methylation levels at single-CpG resolution from methylation enrichment and restriction enzyme sequencing methods. Genome Res 2013;23(9):1541–53. [93] Aparicio O, Geisberg JV, Sekinger E, Yang A, Moqtaderi Z, Struhl K. Chromatin immunoprecipitation for determining the association of proteins with specific genomic sequences in vivo. Curr Protoc Mol Biol 2005;21–3. [94] Choi H, Nesvizhskii AI, Ghosh D, Qin ZS. Hierarchical hidden Markov model with application to joint analysis of ChIP-chip and ChIP-seq data. Bioinformatics 2009;25(14):1715–21. [95] Collas P. The current state of chromatin immunoprecipitation. Mol Biotech 2010;45(1):87–100. [96] Bolstad BM, Irizarry RA, Åstrand M, Speed TP. A comparison of normalization methods for high density oligonucleotide array data based on variance and bias. Bioinformatics 2003;19(2):185–93. [97] Johnson WE, Li W, Meyer CA, Gottardo R, Carroll JS, Brown M, et al. Model-based analysis of tiling-arrays for ChIP-chip. Proc Natl Acad Sci USA 2006;103(33):12457–62. [98] Judy JT, Ji H. TileProbe: modeling tiling array probe effects using publicly available data. Bioinformatics 2009;25(18):2369–75. [99] Droit A, Cheung C, Gottardo R. rMAT-an R/Bioconductor package for analyzing ChIP-chip experiments. Bioinformatics 2010;26(5):678–9. [100] Bennett M, Evans KE, Yu S, Teng Y, Webster RM, Powell J, et al. Sandcastle: software for revealing latent information in multiple experimental ChIP-chip datasets via a novel normalisation procedure. Sci Rep 2015;5. [101] Buck MJ, Nobel AB, Lieb JD. ChIPOTle: a user-friendly tool for the analysis of ChIP-chip data. Genome Biol 2005;6(11):R97. [102] Toedling J, Sklyar O, Huber W. Ringo—an R/Bioconductor package for analyzing ChIP-chip readouts. BMC Bioinform 2007;8(1):221. [103] Zhang ZD, Rozowsky J, Lam HYK, Du J, Snyder M, Gerstein M. Tilescope: online analysis pipeline for high-density tiling microarray data. Genome Biol 2007;8(5):R81. [104] Zhu LJ, Gazin C, Lawson ND, Pagès H, Lin SM, Lapointe DS, et al. ChIPpeakAnno: a bioconductor package to annotate ChIPseq and ChIP-chip data. BMC Bioinform 2010;11(1):237. [105] Park PJ. ChIP-Seq: advantages and challenges of a maturing technology. Nat Rev Genet 2009;10(10):669–80. [106] Pepke S, Wold B, Mortazavi A. Computation for ChIP-seq and RNA-seq studies. Nat Methods 2009;6(110):1–25.

III.  Methods in Epigenetics

188

12.  Computational Epigenetics

[107] Furey TS. ChIP-seq and beyond: new and improved methodologies to detect and characterize protein-DNA interactions. Nat Rev Genet 2012;13(12):840–52. [108] Zhang Y, Liu T, Meyer CA, Eeckhoute J, Johnson DS, Bernstein BE, et al. Model-based analysis of ChIP-Seq (MACS). Genome Biol 2008;9(9):R137. [109] Rozowsky J, Euskirchen G, Auerbach RK, Zhang ZD, Gibson T, Bjornson R, et al. PeakSeq enables systematic scoring of ChIP-seq experiments relative to controls. Nat Biotechnol 2009;27(1):66–75. [110] Zhang X, Robertson G, Krzywinski M, Ning K, Droit A, Jones S, et al. PICS: probabilistic Inference for ChIP-seq. Biometrics 2011;67(1):151–63. [111] Narlikar L, Jothi R. ChIP-Seq data analysis: identification of protein–DNA binding sites with SISSRs peak-finder. Next generation microarray bioinformatics: methods and protocols. New York, NY: Springer; 2012. [pp. 305–322]. [112] Wang J, Lunyak VV, Jordan IK. BroadPeak: a novel algorithm for identifying broad peaks in diffuse ChIP-seq datasets. Bioinformatics 2013;bts722. [113] Zang C, Schones DE, Zeng C, Cui K, Zhao K, Peng W. A clustering approach for identification of enriched domains from histone modification ChIP-Seq data. Bioinformatics 2009;25(15):1952–8. [114] Xu H, Handoko L, Wei X, Ye C, Sheng J, Wei C-L, et al. A signal– noise model for significance analysis of ChIP-seq with negative control. Bioinformatics 2010;26(9):1199–204. [115] Song Q, Smith AD. Identifying dispersed epigenomic domains from ChIP-Seq data. Bioinformatics 2011;27(6):870–1. [116] Kharchenko PV, Tolstorukov MY, Park PJ. Design and analysis of ChIP-seq experiments for DNA-binding proteins. Nat Biotechnol 2008;26(12):1351–9. [117] Rashid NU, Giresi PG, Ibrahim JG, Sun W, Lieb JD. ZINBA integrates local covariates with DNA-seq data to identify broad and narrow regions of enrichment, even within amplified genomic regions. Genome Biol 2011;12(7):R67. [118] Fejes AP, Robertson G, Bilenky M, Varhol R, Bainbridge M, Jones SJM. FindPeaks 3.1: a tool for identifying areas of enrichment from massively parallel short-read sequencing technology. Bioinformatics 2008;24(15):1729–30. [119] Chen L, Wang C, Qin ZS, Wu H. A novel statistical method for quantitative comparison of multiple ChIP-seq datasets. Bioinformatics 2015;btv094. [120] Shen L, Shao N-Y, Liu X, Maze I, Feng J, Nestler EJ. diffReps: detecting differential chromatin modification sites from ChIP-seq data with biological replicates. PLoS One 2013;8(6):e65598. [121] Liang K, Keles S. Detecting ¸ differential binding of transcription factors with ChIP-seq. Bioinformatics 2012;28(1):121–2. [122] Heinz S, Benner C, Spann N, Bertolino E, Lin YC, Laslo P, et al. Simple combinations of lineage-determining transcription factors prime cis-regulatory elements required for macrophage and B cell identities. Mol Cell 2010;38(4):576–89. [123] Allhoff M, Seré K, Chauvistré H, Lin Q, Zenke M, Costa IG. Detecting differential peaks in ChIP-seq signals with ODIN. Bioinformatics 2014;30(24):3467–75. [124] Allhoff M, Seré K, Pires JF, Zenke M, Costa IG. Differential peak calling of ChIP-seq signals with replicates with THOR. Nucleic Acids Res 2016;44(20):e153. [125] Shao Z, Zhang Y, Yuan G-C, Orkin SH, Waxman DJ. MAnorm: a robust model for quantitative comparison of ChIP-Seq data sets. Genome Biol 2012;13(3):R16. [126] Feng J, Liu T, Qin B, Zhang Y, Liu XS. Identifying ChIP-seq enrichment using MACS. Nat Protoc 2012;7(9):1728–40. [127] Xu H, Wei CL, Lin F, Sung WK. An HMM approach to genomewide identification of differential histone modification sites from ChIP-seq data. Bioinformatics 2008;24(20):2344–9. [128] Nair NU, Sahu AD, Bucher P, Moret BME. ChIPnorm: a statistical method for normalizing and identifying differential

[129]

[130]

[131]

[132]

[133]

[134]

[135]

[136]

[137]

[138]

[139]

[140]

[141]

[142] [143]

[144]

[145]

[146]

[147]

[148]

regions in histone modification ChIP-seq libraries. PLoS One 2012;7(8):e39573. Choi J, Kim KT, Huh A, Kwon S, Hong C, Asiegbu FO, et al. DbHiMo: a web-based epigenomics platform for histone-modifying enzymes. Database 2015;2015:bav052. Wang L, Huang H, Dougherty G, Zhao Y, Hossain A, Kocher JPA. Epidaurus: aggregation and integration analysis of prostate cancer epigenome. Nucleic Acids Res 2015;43(2):e7. Wang LY, Wang P, Li MJ, Qin J, Wang X, Zhang MQ, et al. EpiRegNet: constructing epigenetic regulatory network from high throughput gene expression data for humans. Epigenetics 2011;6(12):1505–12. Contrino S, Smith RN, Butano D, Carr A, Hu F, Lyne R, et al. modMine: flexible access to modENCODE data. Nucleic Acids Res 2012;40:1082–8. Harmanci A, Rozowsky J, Gerstein M. MUSIC: identification of enriched regions in ChIP-Seq experiments using a mappabilitycorrected multiscale signal processing framework. Genome Biol 2014;15(10):474. Li J, Jia J, Li H, Yu J, Sun H, He Y, et al. SysPTM 2.0: an updated systematic resource for post-translational modification. Database 2014;2014:bau025. Klein H-U, Hebestreit K. An evaluation of methods to test predefined genomic regions for differential methylation in bisulfite sequencing data. Brief Bioinform 2015;bbv095. Zhang Y, Liu H, Lv J, Xiao X, Zhu J, Liu X, et al. QDMR: a quantitative method for identification of differentially methylated regions by entropy. Nucleic Acids Res 2011;gkr053. Akalin A, Kormaksson M, Li S, Garrett-Bakelman FE, Figueroa ME, Melnick A, et al. methylKit: a comprehensive R package for the analysis of genome-wide DNA methylation profiles. Genome Biol 2012;13(10):R87. Hansen KD, Langmead B, Irizarry RA. BSmooth: from whole genome bisulfite sequencing reads to differentially methylated regions. Genome Biol 2012;13(10):R83. Hebestreit K, Dugas M, Klein H-U. Detection of significantly differentially methylated regions in targeted bisulfite sequencing data. Bioinformatics 2013;29(13):1647–53. Su J, Yan H, Wei Y, Liu H, Liu H, Wang F, et al. CpG_MPs: identification of CpG methylation patterns of genomic regions from high-throughput bisulfite sequencing data. Nucleic Acids Res 2012;gks829. Warden CD, Lee H, Tompkins JD, Li X, Wang C, Riggs AD, et al. COHCAP: an integrative genomic pipeline for single-nucleotide resolution DNA methylation analysis. Nucleic Acids Res 2013;41(11):1–11. Wu D, Gu J, Zhang MQ. FastDMA: an Infinium Human Methylation450 BeadChip analyzer. PLoS One 2013;8(9):e74275. Henry VJ, Bandrowski AE, Pepin A-S, Gonzalez BJ, Desfeux A. OMICtools: an informative directory for multi-omic data analysis. Database 2014;2014:bau069. Mao Z, Ma C, Huang THM, Chen Y, Huang Y. BIMMER: a novel algorithm f or detecting differential DNA methylation regions from MBDCap-seq data. BMC Bioinform 2014;15(Suppl. 12):S6. Saito Y, Tsuji J, Mituyama T. Bisulfighter: accurate detection of methylated cytosines and differentially methylated regions. Nucleic Acids Res 2014;gkt1373. Stockwell PA, Chatterjee A, Rodger EJ, Morison IM. DMAP: differential methylation analysis package for RRBS and WGBS data. Bioinformatics 2014;btu126. Sun D, Xi Y, Rodriguez B, Park HJ, Tong P, Meong M, et al. MOABS: model based analysis of bisulfite sequencing data. Genome Biol 2014;15(2):R38. Kishore K, de Pretis S, Lister R, Morelli MJ, Bianchi V, Amati B, et al. MethylPipe and compEpiTools: a suite of R packages for the integrative analysis of epigenomics data. BMC Bioinform 2015;16(1):313.

III.  Methods in Epigenetics



REFERENCES

[149] Peters TJ, Buckley MJ, Statham AL, Pidsley R, Samaras K, Lord RV, et al. De novo identification of differentially methylated regions in the human genome. Epigenetics Chromatin 2015;8(1):6. [150] Wang Z, Li X, Jiang Y, Shao Q, Liu Q, Chen B, et al. swDMR: a sliding window approach to identify differentially methylated regions based on whole genome bisulfite sequencing. PLoS One 2015;10(7):e0132866. [151] Liu H, Liu X, Zhang S, Lv J, Li S, Shang S, et al. Systematic identification and annotation of human methylation marks based on bisulfite sequencing methylomes reveals distinct roles of cell type-specific hypomethylation in the regulation of cell identity genes. Nucleic Acids Res 2016;44(1):75–94. [152] Sofer T, Schifano ED, Hoppin JA, Hou L, Baccarelli AA. A-clustering: a novel method for the detection of co-regulated methylation regions, and regions associated with exposure. Bioinformatics 2013;29(22):2884–91. [153] Riebler A, Menigatti M, Song JZ, Statham AL, Stirzaker C, Mahmud N, et al. BayMeth: improved DNA methylation quantification for affinity capture sequencing data using a flexible Bayesian approach. Genome Biol 2014;15(2):R35. [154] Jaffe AE, Murakami P, Lee H, Leek JT, Fallin MD, Feinberg AP, et al. Bump hunting to identify differentially methylated regions in epigenetic epidemiology studies. Int J Epidemiol 2012;41(1):200–9. [155] Wu H, Xu T, Feng H, Chen L, Li B, Yao B, et al. Detection of differentially methylated regions from whole-genome bisulfite sequencing data without replicates. Nucleic Acids Res 2015;43(21):e141. [156] Li S, Garrett-Bakelman FE, Akalin A, Zumbo P, Levine R, To BL, et al. An optimized algorithm for detecting and annotating regional differential methylation. BMC Bioinform 2013;14(Suppl. 5):S10. [157] Yao L, Shen H, Laird PW, Farnham PJ, Berman BP. Inferring regulatory element landscapes and transcription factor networks from cancer methylomes. Genome Biol 2015;16(1):105. [158] Hsiao CL, Hsieh AR, Lian IB, Lin YC, Wang HM, Fann CSJ. A novel method for identification and quantification of consistently differentially methylated regions. PLoS One 2014;9(5):1–11. [159] Wang D, Yan L, Hu Q, Sucheston LE, Higgins MJ, Ambrosone CB, et al. IMA: an R package for high-throughput analysis of Illumina’s 450K Infinium methylation data. Bioinformatics 2012;28(5):729–30. [160] Aryee MJ, Jaffe AE, Corrada-Bravo H, Ladd-Acosta C, Feinberg AP, Hansen KD, et al. Minfi: a flexible and comprehensive Bioconductor package for the analysis of Infinium DNA methylation microarrays. Bioinformatics 2014;30(10):1363–9. [161] Statham AL, Strbenac D, Coolen MW, Stirzaker C, Clark SJ, Robinson MD. Repitools: an R package for the analysis of enrichment-based epigenomic data. Bioinformatics 2010;26(13):1662–3. [162] Mayo T, Schweikert G, Sanguinetti G. M3D: a kernel-based test for shape changes in methylation profiles. arXiv 2014;1410:6677. [163] Gevaert O. MethylMix: an R package for identifying DNA methylation-driven genes. Bioinformatics 2015;31(11):1839–41. [164] Park Y, Figueroa ME, Rozek LS, Sartor MA. MethylSig: a whole genome DNA methylation analysis pipeline. Bioinformatics 2014;2014:btu339. [165] Zackay A, Steinhoff C. MethVisual—visualization and exploratory statistical analysis of DNA methylation profiles from bisulfite sequencing. BMC Res Notes 2010;3(1):337. [166] Zheng H, Wu H, Li J, Jiang S-W. CpGIMethPred: computational model for predicting methylation status of CpG islands in human genome. BMC Med Genomics 2013;6(1):S13. [167] Assenov Y, Müller F, Lutsik P, Walter J, Lengauer T, Bock C. Comprehensive analysis of DNA methylation data with RnBeads. Nat Methods 2014;11(11):1138–40.

189

[168] Lee W, Morris JS. Identification of differentially methylated loci using wavelet-based functional mixed models. Bioinformatics 2015;32(5):664–72. [169] Jiang P, Sun K, Lun FMF, Guo AM, Wang H, Chan KCA, et al. Methy-Pipe: an integrated bioinformatics pipeline for whole genome bisulfite sequencing data analysis. PLoS One 2014;9(6):e100360. [170] Bock C, Paulsen M, Tierling S, Mikeska T, Lengauer T, Walter J. CpG island methylation in human lymphocytes is highly correlated with DNA sequence, repeats, and predicted DNA structure. PLoS Genet 2006;2(3):e26. [171] Das R, Dimitrova N, Xuan Z, Rollins RA, Haghighi F, Edwards JR, et al. Computational prediction of methylation status in human genomic sequences. Proc Natl Acad Sci USA 2006;103(28): 10713–6. [172] Fang F, Fan S, Zhang X, Zhang MQ. Predicting methylation status of CpG islands in the human brain. Bioinformatics 2006;22(18):2204–9. [173] Fan S, Zhang MQ, Zhang X. Histone methylation marks play important roles in predicting the methylation status of CpG islands. Biochem Biophy Res Commun 2008;374(3):559–64. [174] Kim S, Li M, Paik H, Nephew K, Shi H, Kramer R, et al. Predicting DNA methylation susceptibility using CpG flanking sequences. Pac Symp Biocomput 2008;13:315–26. [175] Wei K, Sutherland H, Camilleri E, Haupt LM, Griffiths LR, Gan SH. Computational epigenetic profiling of CpG islets in MTHFR. Mol Biol Rep 2014;41(12):8285–92. [176] Wei LK, Sutherland H, Au A, Camilleri E, Haupt LM, Gan SH, Griffiths LR. Methylenetetrahydrofolate Reductase CpG Islands: Epigenotyping. J Clin Lab Anal 2016;30(4):335–44. [177] Wei LK, Sutherland H, Au A, Camilleri E, Haupt LM, Gan SH, et al. A potential epigenetic marker mediating serum folate and vitamin B12 levels contributes to the risk of ischemic stroke. Biomed Res Int 2015;2015:167976. [178] Bradshaw G, Sutherland HG, Camilleri ET, Lea RA, Haupt LM, Griffiths LR. Genetic and epigenetic variants in the MTHFR gene are not associated with non-Hodgkin lymphoma. Meta Gene 2015;6:91–5. [179] Ponger L, Mouchiroud D. CpGProD: identifying CpG islands associated with transcription start sites in large genomic mammalian sequences. Bioinformatics 2002;18(4):631–3. [180] Hackenberg M, Previti C, Luque-Escamilla PL, Carpena P, Martínez-Aroza J, Oliver JL. CpGcluster: a distance-based algorithm for CpG-island detection. BMC Bioinformatics 2006;7(1):446. [181] Liu Z, Xiao X, Qiu W-R, Chou K-C. iDNA-Methyl: identifying DNA methylation sites via pseudo trinucleotide composition. Anal Biochem 2015;474:69–77. [182] Bhasin M, Garg A, Raghava GPS. PSLpred: prediction of subcellular localization of bacterial proteins. Bioinformatics 2005;21(10):2522–4. [183] He S, He H, Xu W, Huang X, Jiang S, Li F, et al. ICM: a web server for integrated clustering of multidimensional biomedical data. Nucleic Acids Res 2016;44:W154–9. [184] James P, Girijadevi R, Charles S, Pillai MR. MethFinder—a software package for prediction of human tissue-specific methylation status of CpG islands. Bioinformation 2013;9(1):61–4. [185] Whitaker JW, Chen Z, Wang W. Predicting the human epigenome from DNA motifs. Nat Methods 2015;12(3):265–72. [186] Rice P, Longden I, Bleasby A. EMBOSS: the European Molecular Biology Open Software Suite. Trends Genet 2000;16(6):276–7. [187] Zhou X, Li Z, Dai Z, Zou X. Prediction of methylation CpGs and their methylation degrees in human DNA sequences. Comput Biol Med 2012;42(4):408–13. [188] He B, Chen C, Teng L, Tan K. Global view of enhancer–promoter interactome in human cells. Proc Natl Acad Sci USA 2014;111(21):E2191–9.

III.  Methods in Epigenetics

190

12.  Computational Epigenetics

[189] Birney E, Smith GD, Greally JM. Epigenome-wide association studies and the interpretation of disease-omics. PLoS Genet 2016;12(6):e1006105. [190] Montano C, Taub MA, Jaffe A, Briem E, Feinberg JI, Trygvadottir R, et al. Association of DNA methylation differences with schizophrenia in an epigenome-wide association study. JAMA Psychiatry 2016;73(5):506–14. [191] Mamtani M, Kulkarni H, Dyer TD, Göring HHH, Neary JL, Cole SA, et al. Genome-and epigenome-wide association study of hypertriglyceridemic waist in Mexican American families. Clin Epigenet 2016;8(1):6. [192] Ernst J, Kellis M. Discovery and characterization of chromatin states for systematic annotation of the human genome. Nat Biotechnol 2010;28(8):817–25. [193] Won K-J, Ren B, Wang W. Genome-wide prediction of transcription factor binding sites using an integrated model. Genome Biol 2010;11(1):R7. [194] Won K-J, Xu Z, Zhang X, Whitaker JW, Shoemaker R, Ren B, et al. Global identification of transcriptional regulators of pluripotency and differentiation in embryonic stem cells. Nucleic Acids Res 2012;40(17):8199–209. [195] Benveniste D, Sonntag H-J, Sanguinetti G, Sproul D. Transcription factor binding predicts histone modifications in human cell lines. Proc Natl Acad Sci USA 2014;111(37):13367–72. [196] Li J, Ching T, Huang S, Garmire LX. Using epigenomics data to predict gene expression in lung cancer. BMC Bioinform 2015;16(5):S10. [197] Baur B, Bozdag S. A feature selection algorithm to compute gene centric methylation from probe level methylation data. PLoS One 2016;11(2):e0148977. [198] Wang S, Sun H, Ma J, Zang C, Wang C, Wang J, et al. Target analysis by integration of transcriptome and ChIP-seq data with BETA. Nat Protoc 2013;8(12):2502–15. [199] Horvath S. DNA methylation age of human tissues and cell types. Genome Biol 2013;14(10):3156. [200] Wang J, Lunyak VV, Jordan IK. Chromatin signature discovery via histone modification profile alignments. Nucleic Acids Res 2012;40(21):10642–56. [201] Hon G, Ren B, Wang W. ChromaSig: a probabilistic approach to finding common chromatin signatures in the human genome. PLoS Comput Biol 2008;4(10):e1000201. [202] Younesy H, Nielsen CB, Lorincz MC, Jones SJM, Karimi MM, Möller T. ChAsE: chromatin analysis and exploration tool. Bioinformatics 2016;32(21):3324–6.

[203] Ernst J, Kellis M. ChromHMM: automating chromatin-state discovery and characterization. Nat Methods 2012;9(3):215–6. [204] Choi H, Fermin D, Nesvizhskii AI, Ghosh D, Qin ZS. Sparsely correlated hidden Markov models with application to genomewide location studies. Bioinformatics 2013;29(5):533–41. [205] Yu P, Xiao S, Xin X, Song C-X, Huang W, McDee D, et al. Spatiotemporal clustering of the epigenome reveals rules of dynamic gene regulation. Genome Res 2013;23(2):352–64. [206] Biesinger J, Wang Y, Xie X. Discovering and mapping chromatin states using a tree hidden Markov model. BMC Bioinform 2013;14(5):S4. [207] Song J, Chen KC. Spectacle: fast chromatin state annotation using spectral learning. Genome Biol 2015;16(1):33. [208] Yu H, Zhu S, Zhou B, Xue H, Han J-DJ. Inferring causal relationships among different histone modifications and gene expression. Genome Res 2008;18(8):1314–24. [209] Hoffman MM, Buske OJ, Wang J, Weng Z, Bilmes JA, Noble WS. Unsupervised pattern discovery in human chromatin structure through genomic segmentation. Nat Methods 2012;9(5):473–6. [210] Liu Y, Qiao N, Zhu S, Su M, Sun N, Boyd-Kirkup J, et al. A novel Bayesian network inference algorithm for integrative analysis of heterogeneous deep sequencing data. Cell Res 2013;23(3):440–3. [211] Lasserre J, Chung H-R, Vingron M. Finding associations among histone modifications using sparse partial correlation networks. PLoS Comput Biol 2013;9(9):e1003168. [212] Diehl AG, Boyle AP. Deciphering ENCODE. Trends Genet 2016;32(4):238–49. [213] Chari R, Coe BP, Wedseltoft C, Benetti M, Wilson IM, Vucic EA, et al. SIGMA 2: a system for the integrative genomic multi-dimensional analysis of cancer genomes, epigenomes, and transcriptomes. BMC Bioinform 2008;9(1):422. [214] Liu T, Ortiz JA, Taing L, Meyer CA, Lee B, Zhang Y, et al. Cistrome: an integrative platform for transcriptional regulation studies. Genome Biol 2011;12(8):R83. [215] Wrzodek C, Eichner J, Büchel F, Zell A. InCroMAP: integrated analysis of cross-platform microarray and pathway data. Bioinformatics 2013;29(4):506–8. [216] Font-Burgada J, Reina O, Rossell D, Azorín F. chroGPS, a global chromatin positioning system for the functional analysis and visualization of the epigenome. Nucleic Acids Res 2014;42(4):2126–37. [217] Pineda S, Real FX, Kogevinas M, Carrato A, Chanock SJ, Malats N, et al. Integration analysis of three omics data using penalized regression methods: an application to bladder cancer. PLoS Genet 2015;11(12):e1005689.

III.  Methods in Epigenetics

C H A P T E R

13 Epigenetic Modifications in Eukaryotes and Prokaryotes are Equally Intriguing Martin-Joseph Okpala, Zimuzoh Orakwue, Ifeanyi Okpala, Shiraz Mujtaba City University of New York, Brooklyn, NY, United States

O U T L I N E Introduction

193

Biological Significance of Acetylation in Eukaryotic Organisms 194 Acetylation of NonHistone Proteins in Regulating Cellular Functions

194

Role of Acetylation in Lower Organisms

195

Role of Lysine Methylation in Higher Eukaryotes

195

Role of Methylation in Lower Organisms

196

INTRODUCTION The sequential and combinatorial pattern of posttranslational modifications (PTMs) serve as a signature for regulating pivotal cellular processes, including gene expression, protein stability, and cell signaling [1]. These molecular events play a critical role not only during normal physiological course, including growth, development, and differentiation, but also during pathological situations, such as inflammation and cancer progression [2,3]. Notably, site-specific covalent modifications of amino acid residues expand the capacity of protein functions by altering the chemical environment, which facilitates new molecular interactions leading to selective downstream events [4]. Until now, atleast 200 PTMs have been reported, which are catalyzed mainly by the biochemical activities of enzymes [5]. A vast number of epigenetic modifications have been characterized that Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00013-4 Copyright © 2017 Elsevier Inc. All rights reserved.

Role of Deacetylation and Demethylation in Gene Regulation

198

Role of Phosphorylation in Higher Eukaryotes

199

Role of Phosphorylation in Lower Organisms

200

Cross-Talks Between Epigenetic Modifications

200

Conclusions and Future Perspectives

201

References

202

can direct a functional outcome alone or together with a distant modified site(s) [4]. Of all the PTMs discovered so far only a few have been linked to gene expression, which are mainly catalyzed by the enzymes that regulate epigenetic mechanisms [4]. Intriguingly, out of the 20 amino acids, only aspartate, glutamate, serine, threonine, tyrosine, histidine, lysine, cysteine, methionine, arginine, glutamine, proline, and glycine are susceptible for undergoing chemical modifications [6]. However, it is still unclear as to why, as compared to other amino acids, only a few amino acids facilitate the imprinting of PTM marks. It is likely that the presence of suitable nucleophilic side chains on these amino acids serve as an ideal substrate for imprinting the PTM marks [7]. For instance, the amino acids, lysine and cysteine have an amine group and a thiol group, respectively. Furthermore, a few of these amino acids participate in a mutually exclusive modifications depending upon the

193

194

13.  Epigenetic Modifications in Eukaryotes and Prokaryotes are Equally Intriguing

upstream cues and alteration in extracellular environment. For instance, the amino acid threonine undergoes glycosylation, as well as phosphorylation, which are involved in regulating protein structure and mediating signaling pathways during embryonic development, as well as regulation of gene expression [8]. Similarly, cysteine serves as a site for other PTMs including prenylation, nitrosylation, glutathionylation, acylation, and hydroxylation [9,10]. Taken together, PTMs are critical for many cellular processes, particularly, in eukaryotic cells. Although PTMs have been discovered in lower organisms, complete knowledge about their roles in epigenetic regulation remains unclear.

BIOLOGICAL SIGNIFICANCE OF ACETYLATION IN EUKARYOTIC ORGANISMS Acetylation of amino acid lysine was first reported in 1964 [11]. However, it was not until 1990s that many of these epigenetic modifications, including acetylation, became a major interest of biomedical research. Since then, due to its direct relevance in human diseases, it has become one of the most extensively investigated PTMs [12]. Biochemically, acetylation results in the addition of an acetyl group donated by the cofactor acetyl-CoA to the epsilon (ε) amino group of the lysine residue [11,13,14]. Until now, acetylation is known to play a pivotal role during chromatin remodeling, which facilitates transcriptional machinery to access gene promoters leading to gene activation [6,11]. A battery of enzymes that catalyzes acetylation of lysine residues was initially known as histone acetyltransferases [15]. However, recent nomenclature also describes them as lysine acetyltransferases as the acetylation is not restricted to chromatin, but also occurs on at least 2000 cellular proteins [16]. One of the major biochemical consequences of acetylation leads to weakening of the DNA/histone interactions [17]. This causes further relaxation of the compacted nucleosomal structure, which facilitates the binding of transcription factors and general transcriptional machineries to the promoter leading ultimately to gene activation. Acetylated-lysine moieties serve as a docking site for bromodomain-containing proteins and/or aid in the recruitment of histone deacetylases, which uniquely regulate the readout of gene expression in the context of upstream physiological stimuli (Fig. 13.1) [18]. For instance, acetylation of lysine 14 on histone (H3K14ac) facilitates the process of DNA repair upon DNA damage by ultraviolet light [19]. Further, H3K14ac acts in concert with the chromatin remodeler, remodel the structure of chromatin (RSC), which recruits repair proteins during a double strand break repair [19]. Additionally, H3K14ac strengthens the binding of RSC to nucleosomes, thereby facilitating repair by the enzyme DNA photolyase [19,20].

FIGURE 13.1  Acetylation of lysine leads to either recruitment of bromodomain-containing proteins (BrD) or HDACs, which could lead either to gene activation or repression.

Acetylation of lysine 9 on histone H3 (H3K9ac) has been shown to be enriched at the transcription start sites in NKG2D (natural-killer group 3, member D) genes in human CD8+ and natural killer cells [20]. Data from chromatin immunoprecipitation assay and subsequent PCR analysis showed that enrichment of acetylation in regions containing 382 base pairs of initiation site of the NKG2D gene was observed [20]. Using curcumin, which is a potential inhibitor of HAT, it was noted that expression of NKG2D was immensely reduced by the inducing of hypoacetylation, thereby, indicating the direct role of H3K9ac in the activation of the gene [20]. Another example involves the acetylation of lysine 5 on histone H4 (H4K5ac) by the histone acetyltransferase HBO1. H4K5ac triggers the activation of gene expression during postgastrulation embryonic development [21,22]. Also, a recent research has demonstrated that the acetylation of lysine 56 on histone H3 (H3K56ac) serves as a marker for genomic stability during DNA damage and subsequent repair, which is mainly catalyzed by the acetyltransferases p300/CBP [23]. During DNA damage by gamma radiation in mammalian cells, levels of H3K56ac increases, which triggers the assembly of DNA-damage responsive proteins, such as phospho-ATM, CHK2, and the tumor suppressor protein p53 [23]. H3K56ac site interaction with the chaperone, Asf1, enables chromatin reassembly, and checkpoint recovery during the S-phase of the cell cycle [23]. The level of H3K56ac is, however, reduced when cells transit into the G2 phase after the termination of DNA replication fork damage repair [23]. Clearly, lysine acetylation in eukaryotes is much better understood than in the lower organisms, possibly due to availability of proteomic tools as compared to prokaryotes.

ACETYLATION OF NONHISTONE PROTEINS IN REGULATING CELLULAR FUNCTIONS In addition to chromatin, many nonhistone proteins within and outside nucleus undergo acetylation [12,24–26]. Interestingly, the mechanism of acetylation

IV.  Model Organisms of Epigenetics



Role of Lysine Methylation in Higher Eukaryotes

has also been exploited by human viruses, for instance, acetylation of the transactivator protein activates rapid replication of human immunodeficiency virus in CD4+ cells [27]. Clearly, many viruses, including HIV exploit the transcriptional machinery of its host to establish their long-term infection [27,28]. Similarly, acetylation of Smc3 by acetyltransferase Eco1 is essential for the establishment of sister chromatid cohesion, which is a vital chromatin event [29]. There are two orthologs in human, namely, Esco1 and Esco2, however, Roberts syndrome is caused by a mutation in ESCO1 [29]. One of widely studied nonhistone proteins has been human tumor suppressor protein p53 in which acetylation was established [30,31]. Acetylation of p53 enhances its transcriptional functions and modulates the functions of downstream target genes to facilitate effective cellular response to maintain genomic integrity [26]. In addition to regulating gene expression, acetylation is also implicated in affecting protein stability and nuclear export of p53 [26]. It is speculated that since lower organisms lack nucleosomal architecture, it is possible that PTM of nonhistone protein could provide insights into the biological role of acetylation in bacterial proteins.

ROLE OF ACETYLATION IN LOWER ORGANISMS In the last two decades, extensive investigations have shed light on the role of acetylation in eukaryotic organisms. However, with the advancement in proteomic technologies our understanding of epigenetic modifications in lower organisms has now begun to increase. It was initially believed that only a few proteins undergo acetylation in bacteria, however, recent investigations revealed that a wide-range of bacterial proteins undergo acetylation [32]. Biochemically, though acetylation at the epsilon amino group of lysine is conserved in all domains of life, acetylation of alpha amino group appears to be a rare phenomenon in bacteria as compared to archaea and eukaryotes [33]. Understanding the significance of acetylation as a protein regulatory mechanism in bacteria was further enhanced by the observation that GCN5related N-acetyltransferase (GNAT) was also conserved in bacteria [33,34]. Most acetyltransferases involved in bacteria belong to the GNAT superfamily [33]. These enzymes are generally known to utilize an acyl coenzyme A as their acyl group donor during the process of acetylation [33]. Taken together, at this stage, several bacterial proteins, which undergo acetylation have been recognized, but the biological consequences of acetylation remains to be defined. In Salmonella enterica, acetylation has been shown to regulate the activity of acetyl-CoA synthetase (Acs) [33–35]. Acs is an enzyme that converts acetate to the

195

high energy acetyl-CoA, which is further utilized in other cellular processes [34]. The protein acetyltransferase Pat inactivates Acs by acetylation of the active site residue K609, which blocks formation of acetyl-AMP [35]. This serves as a mean of maintaining acetyl-CoA homeostasis, thereby, regulating the levels of acetate. The S. enterica Pat protein is a multidomain protein that is made up of C- and N-terminal domain, which is similar to nucleoside diphosphate forming acyl-CoA synthetase family [33,34]. A histidine residue is, however, lacking in this protein, which ultimately prevents it from producing acetyl-CoA from the common means, which is through acetate, adenosine triphosphate, and coenzyme A [33,35]. Recent investigations have found that S. enterica Pat also regulates activities in other metabolic enzymes, such as glyceraldehyde phosphate dehydrogenase, isocitrate lyase, and isocitrate dehydrogenase kinase/phosphatase [35]. Currently, only a few species have been shown to undergo acetylation, and even their consequences have not been fully understood. For instance, Alba proteins have been identified as being acetylated in euryarchaeotal and crenarchaeotal species [24]. Alba proteins are double stranded DNA binding proteins, which are conserved in most archaea genomes, including extremophiles [24,36]. Using mass spectrometry, it was demonstrated that acetylation of lysine 16 on the Alba protein of Salfolobus solfataricus reduces the affinity of the protein binding for DNA, thereby, preventing the inhibition of the helicase enzyme, MCM [24,37,38], as well as epsilon terminal acetylation, Alba proteins are also N-α-acetylated on its N-terminal serine [24,38]. Although only a few proteins undergoing acetylation have been discovered so far, in vitro acetylation increased the understanding of the mechanism of acetylation in most archaea proteins.

ROLE OF LYSINE METHYLATION IN HIGHER EUKARYOTES Methylation of lysine by Su(var)3-9, Enhancer-ofzeste, Trithorax (SET) domain plays a major role in several cellular processes, including growth, differentiation, immune regulation, cellular proliferation, and cancer progression [39]. Methylation, unlike acetylation, is biochemically more complex and has the potential to trigger a wide-range of cellular programs. Additionally, while acetylation generally serves as a mark for gene activation, lysine methylation, depending upon the position and degree, such as mono-, di-, or trimethylation (me1, me2 and me3), can lead to either gene activation or repression. Further, the biochemical effects of site- and position-specific methylation could lead to the formation of either euchromatin or heterochromatin [40,41]. SET domains catalyze the transfer of methyl group from the

IV.  Model Organisms of Epigenetics

196

13.  Epigenetic Modifications in Eukaryotes and Prokaryotes are Equally Intriguing

cofactor S-adenosyl methionine to the amino group of a lysine residue [40,41]. These methylated-lysine moieties further serve as recruitment sites for modular domains, including chromodomains and PHD fingers [42]. Methylation of histone H3 on lysine 4 (H3K4me) is a hallmark for transcriptional initiation and elongation [43]. For instance, H3K4me3 has been found in the enhancer regions of human brain genes [43]. Using the chromatin immunoprecipitation and next generation sequencing-based quantification of the nucleosomes, large peaks of H3K4me3 were observed at transcription start sites [19,43]. This also correlates with the occupation of RNA Polymerase II on the promoter regions [44]. Alterations in normal H3K4 methylation levels have been noticed in cerebral cortex of patients with autism and schizophrenia [43]. Methylation of lysine 27 on histone H3 (H3K27me) by Enhancer of Zeste Homolog 2 (EZH2) leads to the formation of hetrochromatin, particularly, during the inactivation of X chromosome, as well as by the recruitment of the polycomb repressor complex (PRC1) [45,46]. Polycomb group of proteins also modulates the activity of HOX genes during the course of development [45,46]. PRC1 then prevents access of nucleosome remodeling enzymes, leading to an overall transcriptional repressive state [45,46]. It is clear that methylation of eukaryotic proteins has been extensively investigated, however, despite the presence of SET proteins in prokaryotes and other lower organisms; it is still relatively less studied.

ROLE OF METHYLATION IN LOWER ORGANISMS Recent studies demonstrate that SET proteins also exist in lower organisms, including bacteria and archaea, as well as in viruses despite the lack of chromatin [47]. Unlike eukaryotes where SET domain is part of a larger protein, in prokaryotes, SET exists as an independent protein. Although SET proteins in lower organisms are relatively less investigated, emerging studies suggest that they play a crucial role in regulating the host’s transcription and immune system to support its own replication and pathogenicity [48,49]. Understanding the role of SET proteins in the interaction of pathogenhost interaction could be instrumental in identifying new pharmacological targets. Notably, the conservation of SET protein function in prokaryotes and lower eukaryotes [39], suggests an evolutionary significance of methyltransferases in regulating the cellular processes. Furthermore, understanding the role of SET proteins in prokaryotes will also reveal new nonhistone targets of methylation. Finally, the relative simplicity of prokaryotes provides an avenue to study the complexity of gene activation and repression in higher eukaryotes through SET domain-mediated methylation.

Interestingly, histone-like proteins have been identified in bacteria [50], which is also associated with DNA and undergo modifications that result in the alteration of nucleoid structure. These epigenetic changes lead to the regulation of gene expression, DNA recombination, and replication, as well as life cycle [50]. Further, some of these proteins also share structural similarities with eukaryotic histones. For instance, the amino acid composition of histone-like protein H is similar to histone H2A in eukaryotes [50]. Studies have shown that these histone-like proteins could be methylated in prokaryotes, such as Chlamydia and Synechocystis [51,52]. Lysine methylation of these histone-like proteins has been found to affect the adherence of some bacteria to their specific environment [51]. Bioinformatic investigations unraveled that there could be atleast 500 bacterial genome containing SET proteins (Fig. 13.2A–B) [53]. BaSET, a SET protein expressed by Bacillus anthracis have been shown to induce immunosuppression in its host to enhance its pathogenicity and growth [47,49]. This is accomplished by trimethylation of human histone H1 by BaSET, which is different from eukaryotic SET proteins that mostly target histones H3 and H4 [49,54]. BaSET induced H1-methylation in the nuclei of macrophages is able to suppress the host’s immunity through the repression of inflammatory genes. B. anthracis with a mutant SET protein not only shows a delayed growth, but, all the mice infected with it survived [49], which is in contrast to mice infected with B. anthracis that died within 4 days of postinfection [49]. Taken together, although we know that prokaryotic SET has ability to methylate human histone, we still do not quite understand the methylation targets within the bacillus. SET has also been found to be expressed in archaea [39]. For instance, Mechanosarcina mazei contains a SET domain-containing gene that was confirmed by amino acid alignment of histone-modifying proteins [39,48,49]. Go1-SET, which is expressed in M. mazei methylates DNAinteracting protein (MC1α) and regulates the chromatin structure, thereby, indicating its role in gene expression [50]. Further, reports have also identified another methyltransferase in archaea called archaeal methyltransferases (aKMT) [55,56]. These aKMT are similar to KMT4/ DOT1 (disruptor of telomeric silencing) family in eukaryotes, such as Saccharomyces cerevisiae [57]. They not only possess a DOT1 family methyltransferase catalytic core, but also their mechanism of action suggests their similarity to DOT1 methyltransferases in eukaryotes [58,59]. They methylate chromatin-associated proteins, such as Sul7d and Cren7 found in Sulfolobus, where they regulate gene expression during thermal adaptation [58]. Furthermore, recently Xia et al. reported that aKMT4 monomethylates minichromosome maintenance (MCM) in Sulfolobus, which enables maintenance of higher helicase activity at

IV.  Model Organisms of Epigenetics



Role of Methylation in Lower Organisms

197

FIGURE 13.2  (A) Multiple alignment of selected prokaryotic SET domain–containing proteins. (B) A phylogenetic tree of prokaryotic SET proteins mentioned in (A).

IV.  Model Organisms of Epigenetics

198

13.  Epigenetic Modifications in Eukaryotes and Prokaryotes are Equally Intriguing

higher temperatures [60]. Subsequently, methylated MCM was compared to an unmethylated MCM for their catalytic activities in vitro [60]. Data showed that there was a 50% increase in DNA unwinding in methylated MCM as compared to the unmethylated MCM [60]. This showed the evidence of direct impact on DNA unwinding by methylation in archaea. The presence of aKMT in archaea and their similarities to methyltransferases in lower eukaryotes suggests an evolutionary significance. Taken together, these findings suggests that SET domain proteins are capable of regulating gene expression and cellular processes in the absence of histone proteins in lower organisms, thereby underscoring the diversity in their mechanism of action. Additionally, lower eukaryotes, such as Saccharomyces cerevisiae also express SET domain proteins. The first histone H3 lysine 4 (H3K4) methylase was isolated from Saccharomyces cerevisiae [61]. Unlike prokaryotes, they express histone proteins in chromatin, which can be methylated by SET domain proteins to regulate processes, such as development/aging. A recent study found that Set2 histone methyltransferases regulate cellular lifespan and telomeric silencing in S. cerevisiae [62]. Though higher eukaryotes likely possess more SET domain proteins, lower eukaryotes may possess SET domain proteins that are similar in structure to those in higher eukaryotes. S. cerevisiae express six SET domain proteins compared to 60 human SET domain proteins, but its Set1 domain protein share about 44% homology with Drosophila Trithorax [63]. Histone methylation was halted in S. cerevisiae as a result of comparable mutation to Trithorax, which has been shown to lead to developmental disruption in Drosophila [64]. This suggests that while lower eukaryotes may have SET domain proteins that are not identical to those found in higher eukaryotes, they share some structural and functional attributes. Certain viruses have also been shown to express a SET protein, which plays a role that is distinct from those observed in eukaryotes. For example, a SET-protein, vSET in Paramecium bursaria chollera virus (PBCV-1), was found to function in the silencing of the host (Chlorella) transcriptional machinery [48]. This was triggered by dimethylation of its histone H3 on lysine 27, which is a sequence homolog of human histone H3 protein [61]. By silencing the host’s transcriptional machinery, it is able to facilitate its own replication. vSET methylation target was found to be identical to EZH2, a human H3 lysine 27 methyltransferase [48]. Other studies have shown supportive evidence that viral methyltransferases may function similarly to EZH2. One recent study showed that GF1 transcriptional activity could be repressed by human cytomegalovirus by targeting EZH2 [65]. This finding suggests that viral SET proteins could function similarly to human lysine methyltransferases, particularly EZH2, and they could also be instrumental in enhancing viral replication within the host.

The expression of SET protein in bacteria, archaea, and lower eukaryotes indicates its role in key cellular processes. While SET proteins might share similar motif, their mechanism of action could be distinctly selective to that organism. Investigations into activity of SET protein in lower organism could help us to identify new target(s) of methylation, as well as their functions. Together, this knowledge may be harnessed to understand the role of SET proteins in disease outcome and pathogenicity, and possibly, give us an insight into possible treatments.

ROLE OF DEACETYLATION AND DEMETHYLATION IN GENE REGULATION Deacetylation of histone proteins reverses the biochemical impact of acetylation on downstream gene activity, which results in the assembly of a repressor complex on the promoter region and condensation of chromatin leading to transcriptional repression [66,67]. Collectively, a delicate balance exists between acetylation and deacetylation, which serves as a pivotal regulatory mechanism during the control of gene expression [68]. The enzymes that catalyze histone deacetylation are commonly known as histone deacetylases (HDACs), which are essentially divided into 4 classes according to their homology with their yeast counterparts. Class I (HDAC 1, 2, 3, and 8) and class II (HDAC 4, 5, 6, 7, 9, and 10) HDACs are related to yeast RPD3 and yeast HDAC1, respectively [69]. Class III consists of seven sirtuins that depend upon NAD+ as their cofactor, while class IV consists of only HDAC11 [6,68,69]. Notably, it has been demonstrated that inhibition of HDAC activities by phenylbutyrate in patients with myelodysplastic syndrome showed clinical improvement [70]. Further, a patient with cutaneous T cell lymphoma had a positive response when administered the HDACi depsipeptide [70]. A number of investigations demonstrate that HDACs interact with transcription factors, which are linked with the regulation of cell cycle progression, as well as programmed cell death [11]. HDAC1 and HDAC2 interact with the retinoblastoma protein (Rb), which is involved in the repression of cell cycle progression [11]. Rb recruits HDACs to the promoter regions containing the E2F protein, the transcription factors involved in cell cycle regulation [11]. The E2F target genes are further repressed, which subsequently leads to cell cycle arrest [71]. In tumors, including colorectal, as well as ovarian, overexpression of HDAC1 and HDAC2 have been observed, where they are involved in the early stages of cancer progression [66,71]. Available data from cancer research have shown that there is a decrease in HAT activities due to mutations or simply an increase in HDAC activity [66]. Transcriptional repression of tumor-suppressor

IV.  Model Organisms of Epigenetics



Role of Phosphorylation in Higher Eukaryotes

genes could also promote the progression of cancer cells due to an increase in HDAC activity in their promoters [66]. For instance, cyclin-dependent kinase inhibitor 1 (CDKN1A), which is involved in the repression of the cell cycle in cancer cells, is repressed by an increase in HDAC recruitment [66]. Inhibition of HDACs leads to an increase in expression of this CDKN1A, and thereby inhibiting the development of cancer cells. Methylation of lysine was discovered in 1964, and was considered as an irreversible process until 2000 that demethylation was fully understood [40,72,73]. Demethylation, just like methylation, is a biochemically complex phenomenon that triggers both gene respression and activation, but through the removal of the methyl group [74]. The enzymes that catalyze this process are called lysine demethylases (KDM). Demethylase activity occurs because KDM1 has a C-terminal amine oxidase domain and is dependent on a flavin adenine dinucleotide mechanism [73]. A second family of demethylases was later discovered, which consist of the Jumonji C domain-containing proteins. These proteins demethylate mono-, di-, and trimethylated lysine residues via an oxygenase mechanism, using α-ketoglutarate and Fe(II) as cofactors [75]. There are 24 JmjC-domain proteins in humans, which are differentiated by domains, such as PHD, Tudor, FBOX, ARID, LRR, and JmjN [75]. Lysine demethylase activities have been reported to be overexpressed in prostate cancer cells [76]. Demethylation of H3K9me by LSD1 relieves nucleosomal repression, leading to the expression of androgen receptors in prostate cancer cells [76]. LSD1 acts as a cofactor that stimulates androgen-dependent gene transcription in prostate cancer cells [76]. LSD1, however, suppresses gene activation in embryonic cancer cells by reducing H3K4me2 to mono- and unmethylated residues [77]. The use of inhibitors was able to show reduced levels of LSD1 activities and an increase in methylation patterns, resulting in cancer progression [77]. JmjC-domaincontaining histone demethylase 2A is understood to be associated with gene repression [78]. It is known to specifically demethylate H3K9me1 and H3K9me2 in renal cancer cells, where upregulation of JHDM2A leads to cancer progression [78].

ROLE OF PHOSPHORYLATION IN HIGHER EUKARYOTES In addition to cellular proteins where phosphorylation plays a key role in signaling, phosphorylation also occurs on all of the four core histones, H2A, H2B, H3, and H4 [79,80], as well as on the linker histone H1 [79,81]. A phosphorylation mark occurs mainly on amino acids with a hydroxyl group on its side chain, such as serine, threonine, and tyrosine residues [82]. The process of

199

phosphorylation involves transfer of a phosphate group from the cofactor adenosine triphosphate to the target amino acid [83,84]. Subsequently, this alters the overall negative charge on the protein or chromatin, respectively [80]. Biochemically, phosphorylation is a reversible process mediated by a kinase and subsequently, a phosphatase, which are another class of enzymes that remove the phosphate group. Recruitment of kinases to the target protein is induced by extracellular signals, such as growth factors, genotoxic stress, as well as cytokines [85]. Subsequently, with a cascade of phosphorylation/dephosphorylation events extracellular message converges within the nucleus to activate the genes that are required to respond to the external changes. It was noted that phosphorylation triggers chromatin condensation during the course of mitosis and meiosis depending on what phase chromatin condensation is required [86]. For instance, phosphorylation of serine 10 on histone H3 is associated with chromatin condensation in the metaphase II during meiosis, which was detected in all mammals [80,86]. Although H3 phosphorylation was initially believed to be only involved in chromatin condensation, later investigations revealed it aided in the activation of gene expression as well [87]. For instance, phosphorylation of tyrosine 41 on histone 3 (H3Y41p) is associated with transcription, which is mediated by Janus Kinase 2 [87]. H3Y41p disrupts the formation of heterochromatin in the oncogene imo2 by inhibiting the binding of heterochromatin protein 1 [87]. Inhibition of HP1 leads to the activation of transcription in imo2 and other JAK2-related genes [87]. A number of proteins, such as BRCT, WW, FHA, WD40, 14-3-3 SH2 and LRR domains act as readers that recognize phosphorylated amino acid residues [88]. Recent investigations have been centered on finding new phosphorylation sites, as well as their biological functions. A recent study revealed that phosphorylation of threonine 3 on histone 3 (H3Tp) by Haspin plays a key role during chromosomal alignment in the metaphase of mitotic division [89]. Further, H3T3p in kinetochores serves as a mark for the recruitment of chromosomal passenger complex, which controls the alignment of chromosomes by reattaching improper kinetochorespindle fiber connections [89]. Inhibition or mutation within haspin kinase has potential to cause error during chromosomal segregation, which could result in miscarriage in pregnant women [89]. Phosphorylation of tyrosine 57 on histone H2A (H2AY57p) is linked to transcriptional elongation in lung, prostrate, and kidney cancers in mammals [90]. H2AY57p is catalyzed by a tyrosine kinase, Casein Kinase 2 (CK2) which can be located in the nucleus, as well as cytoplasm [91]. Inhibition of CK2 results in the decreased H2AY57p level, as well as transcription [90]. Mutation of tyrosine 57 negatively regulates H3K4me3 and H3K79me3,

IV.  Model Organisms of Epigenetics

200

13.  Epigenetic Modifications in Eukaryotes and Prokaryotes are Equally Intriguing

thereby, affecting the downstream transcription [90]. In addition, downregulation of H3K4me3 and H79me3, mutation of Y57 causes loss of monoubiquitylation on histone H2B [90]. DNA damage induces phosphorylation of H2A.X on Serine 139 (H2A.XS139p) that instantly acts as a recruiter of repair machineries [23]. H2A.XS139p site enables the retention of mediator of DNA damage checkpoint 1 in humans and 53BP1 in yeast in response to DNA damage [23]. γH2AX also recruits chromatin remodelers and promotes acetylation which allows chromatin around the breaks more attainable to the repair enzymes [23].

ROLE OF PHOSPHORYLATION IN LOWER ORGANISMS Ever since the discovery of covalent modifications, phosphorylation has become one of the most investigated PTMs in lower organisms, including bacteria. Since its discovery in the 1970s, serine/threonine phosphorylation has been linked to multiple cellular functions in bacteria. For instance, investigations have shown that phospho-regulatory mechanisms occur during bacterial cell cycle [92]. Phosphorylation of bacteria proteins are carried out by bacterial tyrosine kinases, known as BY-kinases. These proteins have also been linked with events, such as cell adaptation, virulence, as well as cell survival [92,93]. Extensive studies on the phosphoproteome performed on bacteria, such as Escherichia coli and Bacillus subtilis have shown that only 5% of their proteins can be phosphorylated on serine, threonine, and tyrosine residues [93]. Compared to eukaryotes, only a few functions have been matched to some phosphorylated sites. It is thought that phosphorylation may have a direct impact on DNA damage signaling and chromosomal replicating mechanisms [94]. For instance, the DNA recombinase RecA enzyme, which plays a critical role in the repair of double stranded breaks in B. subtilis, have been reported to be regulated by phosphorylation activities [94]. It is also linked with the maintenance of chromosome strength during sporulation [93,94]. Although this process is not fully understood, it is known that this phosphorylation event, which is catalyzed by the Ser/ Thr kinase YabT, further hinders sporulation when DNA damages are experienced [94]. Recent proteome investigations have shown that many adhesions, as well as metabolic proteins in Mycoplasma pneumoniae is phosphorylated [93]. Although about 18 phosphorylated proteins were discovered [93], their mechanism of phosphorylation, as well as their overall function remains to be fully understood. In M. pneumoniae, phosphorylation occurs in nearly 10% of its proteins, which is more compared to most bacteria proteons [93]. Phosphorylation in M. pneumoniae occurs in cell adhesion (HMW3

FIGURE 13.3  In Mycoplasma pneumoniae the cyto-adherence function of cell adhesion protein (HMW3) is regulated by phosphorylation by PrkC kinase and PrpC phosphatase. The yellow ball represents phospohorylation.

and P41) and cell surface proteins (MPN474), as with B. subtilis [93]. Repression or nonexpression of the phosphatase PrpC increases the phosphorylation of these proteins by the Kinase PrkC [93], which regulates its growth on a substrate, as well as induce cytotoxicity in host’s cells [95]. PrkC is one of the only two identified kinases in M. pneumoniae, and they are both responsible for the phosphorylation of 5 out of about 63 identified phosphorylation sites (Fig. 13.3) [93]. It is now understood that one kinase can phosphorylate different sites. The challenge has now been to identify more enzymes responsible for this modification, as well as their functions in M. pneumoniae and other lesser organisms.

CROSS-TALKS BETWEEN EPIGENETIC MODIFICATIONS So far it is understood that site-specific and combinatorial epigenetic modifications modulates the transcriptional outcome, such as gene activation and repression. However, it is not clearly known how a PTM on one site can interact with a PTM at a distant sites and how these biochemical processes impact the overall cellular processes. For instance, H3S10 phosphorylation of serine 10 (H3S10p) on histone H3 is linked to chromatin condensation during mitosis and meiosis, but H3S10p is also associated with transcription activation, possibly in conjunction with H3K14ac by GCN5 acetyltransferase in budding yeast because these two modifications occur in close proximity to each other [87,96,97]. The ability of histone methylation mediating gene activation or repression is enhanced by it cross-talk choices. For instance, an antagonistic relationship exists between H3K4me and H3K9me. Research has shown that H3K4me by SET7 represses H3K9me by SUV39H1 and further triggers the acetylation of H3 by the recruitment of the acetyltransferase p300/CBP, leading to gene expression [98,99]. In contrast, H3K9me negates the occurrence of acetylation on H3, leading to a condensed chromatin state and gene repression [98,99]. Surprisingly, presence of H3K9me mark also negates the functional outcome of H3K4me that results in gene repression [99].

IV.  Model Organisms of Epigenetics



Conclusions and Future Perspectives

201

FIGURE 13.4  (A) Phosphorylation of protein CheY by CheA kinase promotes recruitment of FliM, which regulates an increase in clockwise rotation in response to chemical stimuli. The yellow ball represents phospohorylation. (B) Acetylation of CheY blocks CheA to mediate phosphorylation and subsequent recruitment of FliM, which results in downregulation of chemotaxis. The green ball represents acetylation.

In lower organisms, phosphorylation also affects gene regulation by cross-talk events with other PTMs. For instance, H3S10p regulates gene activation by serving as a repressive mark for H3K9 methylation. This then triggers the acetylation on H3K9, which leads to chromatin remodeling and subsequent gene activation in mammalian cells [35,98,100]. In Saccharomyces cerevisiae, H3S10p by Snf1 regulates gene expression by acting in concert with the H3K14 acetylation, which is catalyzed by GCN5 HATs [98]. In E. coli, K91, K92, K109, K119, K122, and K126 residues on the chemotaxis response regulator protein, CheY, has been shown to undergo N-acetylation [35,37,101]. Biochemical data revealed that CheY can also be acetylated by using acetyl-CoA as its acetyl donor or Acs-catalyzed using acetate as its acetyl group donor [35,37,101]. Phosphorylation of CheY was initially believed to regulate rotation in response to chemical stimuli, such as amino acids, sugars, and electron acceptors in E. coli. Recent investigation has, however, shown that acetylation regulates phosphorylation of CheY. Phosphorylation induces binding of CheY to the FliM, which increases clockwise rotation in response to chemical stimuli [101]. Acetylation, however, reduces the binding affinity of CheY to its kinase, as well as the FliM protein, and subsequently its phosphatase [101]. This provides a slow regulation of chemotaxis induced by the metabolic state of the cell [101]. Interestingly, chemotaxis is not a metabolic process. Acetylation of CheY does, however, regulate other metabolic processes (Fig. 13.4A–B). It is now clear that phosphorylation and acetylation work together in response to environmental and physiological stimuli to regulate chemotactic response.

CONCLUSIONS AND FUTURE PERSPECTIVES Increasing our knowledge of epigenetic mechanisms that regulate the life cycle of lower organisms is essential but needs extensive investigations, which should be accompanied by the development of innovative proteomic tools. This notion, particularly, acquires more significance in light of emergence of new pathogens and progression of many infectious diseases. Since lower organisms do not possess a nucleosomal organization, it is likely that the functional outcome of acetylation in lower organisms could be similar to the acetylation of nonhistone proteins in eukaryotes. Phosphorylationmediated pathways that direct bacterial growth and division could serve as new targets for developing the new generation antibacterial medicines as it is expected in future that due to excessive consumption of antibiotics there will be emergence resistant strains globally. Further, given that lower organisms are sensitive to external nutritional environment, it is vital to probe the relation between metabolism and factors that participate in epigenetic regulation. Finally, DNA methylation and PTMs are central to the epigenetic mechanisms; however, molecular interactions that occur in proximity to the gene promoter also have ability to affect gene expression. Thus, epigenetic mechanisms serve as an umbrella for a number of gene modulator, which directly or indirectly impact protein synthesis and function. A highly conserved SET domain motif across species underscores their evolutionary significance and their fundamental biochemical role as a protein methylase

IV.  Model Organisms of Epigenetics

202

13.  Epigenetic Modifications in Eukaryotes and Prokaryotes are Equally Intriguing

involved in the regulation of gene, growth, and development. Moreover, a subtle structural difference around catalytic domain guarantees a functional diversity within the family of SET proteins/domains, particularly, the differential presence of cysteine rich motif on C-terminal. Biologically, this is exemplified by SET proteins ability to methylate a diverse set of histone and histone-like proteins. Though the biological role of methylation has established to an extent the epigenetic phenomenon, further studies are required to fully comprehend the biological mechanism, especially those that mutually direct exclusive modifications on a given lysine residue. As prokaryotes lack histone proteins, they could serve as a mechanistic model organism to further unravel the nonhistone protein targets of methylation. This will help to explain the role of SET proteins in prokaryotes. Lastly, in-depth understanding SET proteins in prokaryotes will help to elucidate the dynamics of gene activation and repression that can lead to the development of chronic diseases in higher eukaryotic organisms.

Acknowledgments MJO was supported by Minority Science and Engineering Improvement Program at MEC; ZO was supported by MEC-Research Initiative for Scientific Enhancement (NIGMS/NIH); SM was supported by NIH and Minority Science and Engineering Improvement Program.

References [1] Karve TM, Cheema AK. Small changes huge impact: the role of protein posttranslational modifications in cellular homeostasis and disease. J Amino Acids 2011;2011:207691. [2] Baylin SB, Jones PA. Epigenetic Determinants of Cancer. Cold Spring Harb Perspect Biol 2016;8:a019505. [3] Yang J, Liu Z, Xiao TS. Post-translational regulation of inflammasomes. Cell Mol Immunol 2016;14:65–79. [4] Sims RJ 3rd, Reinberg D. Is there a code embedded in proteins that is based on post-translational modifications? Nat Rev Mol Cell Biol 2008;9:815–20. [5] Prabakaran S, Lippens G, Steen H, Gunawardena J. Post-translational modification: nature’s escape from genetic imprisonment and the basis for dynamic information encoding. Wiley Interdiscip Rev Syst Biol Med 2012;4:565–83. [6] Duan G, Walther D. Computational phosphorylation network reconstruction: methods and resources. Methods Mol Biol 2015;1306:177–94. [7] Bischoff R, Schluter H. Amino acids: chemistry, functionality and selected non-enzymatic post-translational modifications. J Proteomics 2012;75:2275–96. [8] Walsh CT, Garneau-Tsodikova S, Gatto GJ Jr. Protein posttranslational modifications: the chemistry of proteome diversifications. Angew Chem Int Ed Engl 2005;44:7342–72. [9] Wani R, Nagata A, Murray BW. Protein redox chemistry: posttranslational cysteine modifications that regulate signal transduction and drug pharmacology. Front Pharmacol 2014;5:224. [10] Rothbart SB, Strahl BD. Interpreting the language of histone and DNA modifications. Biochim Biophys Acta 2014;1839:627–43.

[11] Fukuda H, Sano N, Muto S, Horikoshi M. Simple histone acetylation plays a complex role in the regulation of gene expression. Brief Funct Genomic Proteomic 2006;5:190–208. [12] Dekker FJ, Haisma HJ. Histone acetyl transferases as emerging drug targets. Drug Discov Today 2009;14:942–8. [13] Pougovkina O, te Brinke H, Ofman R, van Cruchten AG, Kulik W, Wanders RJ, et al. Mitochondrial protein acetylation is driven by acetyl-CoA from fatty acid oxidation. Hum Mol Genet 2014;23:3513–22. [14] Varland S, Osberg C, Arnesen T. N-terminal modifications of cellular proteins: the enzymes involved, their substrate specificities and biological effects. Proteomics 2015;15:2385–401. [15] Verdone L, Caserta M, Di Mauro E. Role of histone acetylation in the control of gene expression. Biochem Cell Biol 2005;83:344–53. [16] Choudhary C, Kumar C, Gnad F, Nielsen ML, Rehman M, Walther TC, et al. Lysine acetylation targets protein complexes and co-regulates major cellular functions. Science 2009;325:834–40. [17] Ettig R, Kepper N, Stehr R, Wedemann G, Rippe K. Dissecting DNA-histone interactions in the nucleosome by molecular dynamics simulations of DNA unwrapping. Biophys J 2011;101:1999–2008. [18] Patel J, Pathak RR, Mujtaba S. The biology of lysine acetylation integrates transcriptional programming and metabolism. Nutr Metab (Lond) 2011;8:12. [19] Duan MR, Smerdon MJ. Histone H3 lysine 14 (H3K14) acetylation facilitates DNA repair in a positioned nucleosome by stabilizing the binding of the chromatin remodeler RSC (remodels structure of chromatin). J Biol Chem 2014;289:8353–63. [20] Fernandez-Sanchez A, Baragano Raneros A, Carvajal Palao R, Sanz AB, Ortiz A, Ortega F, et al. DNA demethylation and histone H3K9 acetylation determine the active transcription of the NKG2D gene in human CD8+ T and NK cells. Epigenetics 2013;8:66–78. [21] Kueh AJ, Dixon MP, Voss AK, Thomas T. HBO1 is required for H3K14 acetylation and normal transcriptional activity during embryonic development. Mol Cell Biol 2011;31:845–60. [22] Mishima Y, Miyagi S, Saraya A, Negishi M, Endoh M, Endo TA, et al. The Hbo1-Brd1/Brpf2 complex is responsible for global acetylation of H3K14 and required for fetal liver erythropoiesis. Blood 2011;118:2443–53. [23] Cao LL, Shen C, Zhu WG. Histone modifications in DNA damage response. Sci China Life Sci 2016;59:257–70. [24] Soppa J. Protein acetylation in archaea, bacteria, and eukaryotes. Archaea 2010;2010:9. [25] Zhao S, Xu W, Jiang W, Yu W, Lin Y, Zhang T, et al. Regulation of cellular metabolism by protein lysine acetylation. Science 2010;327:1000–4. [26] Prives C, Manley JL. Why is p53 acetylated? Cell 2001;107:815–8. [27] Mujtaba S, He Y, Zeng L, Farooq A, Carlson JE, Ott M, et al. Structural basis of lysine-acetylated HIV-1 Tat recognition by PCAF bromodomain. Mol Cell 2002;9:575–86. [28] Mujtaba S, Zhou MM. Anti-viral opportunities during transcriptional activation of latent HIV in the host chromatin. Methods 2010;53:97–101. [29] Whelan G, Kreidl E, Wutz G, Egner A, Peters JM, Eichele G. Cohesin acetyltransferase Esco2 is a cell viability factor and is required for cohesion in pericentric heterochromatin. EMBO J 2012;31:71–82. [30] Li SD, Banck M, Mujtaba S, Zhou MM, Sugrue MM, Walsh MJ. p53-Induced growth arrest is regulated by the mitochondrial SirT3 deacetylase. PLoS One 2010;5:e10486. [31] Mujtaba S, He Y, Zeng L, Yan S, Plotnikova O, Sachchidanand, et al. Structural mechanism of the bromodomain of the coactivator CBP in p53 transcriptional activation. Mol Cell 2004;13: 251–63.

IV.  Model Organisms of Epigenetics



REFERENCES

[32] Liang W, Malhotra A, Deutscher MP. Acetylation regulates the stability of a bacterial protein: growth stage-dependent modification of RNase R. Mol Cell 2011;44:160–6. [33] Bernal V, Castano-Cerezo S, Gallego-Jara J, Ecija-Conesa A, de Diego T, Iborra JL, et al. Regulation of bacterial physiology by lysine acetylation of proteins. N Biotechnol 2014;31:586–95. [34] Thao S, Escalante-Semerena JC. Biochemical and thermodynamic analyses of Salmonella enterica pat, a multidomain, multimeric N(epsilon)-lysine acetyltransferase involved in carbon and energy metabolism. MBio 2011;2:e00216–11. [35] Thao S, Escalante-Semerena JC. Control of protein function by reversible nvarepsilon-lysine acetylation in bacteria. Curr Opin Microbiol 2011;14:200–4. [36] Eichler J, Adams MW. Posttranslational protein modification in Archaea. Microbiol Mol Biol Rev 2005;69:393–425. [37] Wardleworth BN, Russell RJ, Bell SD, Taylor GL, White MF. Structure of Alba: an archaeal chromatin protein modulated by acetylation. EMBO J 2002;21:4654–62. [38] Eichler J, Maupin-Furlow J. Post-translation modification in Archaea: lessons from Haloferax volcanii and other Haloarchaea. FEMS Microbiol Rev 2013;37:583–606. [39] Nwasike C, Ewert S, Jovanovic S, Haider S, Mujtaba S. SET domain-mediated lysine methylation in lower organisms regulates growth and transcription in hosts. Ann NY Acad Sci 2016;1376:18–28. [40] Allfrey VG, Faulkner R, Mirsky AE. Acetylation and methylation of histones and their possible role in the regulation of RNA synthesis. Proc Natl Acad Sci USA 1964;51:786–94. [41] Noh KM, Allis CD, Li H. Reading between the lines: “aDD”-ing histone and DNA methylation marks toward a new epigenetic “sum”. ACS Chem Biol 2016;11:554–63. [42] Yap KL, Zhou MM. Structure and mechanisms of lysine methylation recognition by the chromodomain in gene transcription. Biochemistry 2011;50:1966–80. [43] Shen E, Shulha H, Weng Z, Akbarian S. Regulation of histone H3K4 methylation in brain development and disease. Philos Trans R Soc Lond B Biol Sci 2014;369:20130514. [44] Pekowska A, Benoukraf T, Zacarias-Cabeza J, Belhocine M, Koch F, Holota H, et al. H3K4 tri-methylation provides an epigenetic signature of active enhancers. EMBO J 2011;30:4198–210. [45] Plath K, Fang J, Mlynarczyk-Evans SK, Cao R, Worringer KA, Wang H, et al. Role of histone H3 lysine 27 methylation in X inactivation. Science 2003;300:131–5. [46] Bracken AP, Pasini D, Capra M, Prosperini E, Colli E, Helin K. EZH2 is downstream of the pRB-E2F pathway, essential for proliferation and amplified in cancer. EMBO J 2003;22:5323–35. [47] Aravind L, Iyer LM. Provenance of SET-domain histone methyltransferases through duplication of a simple structural unit. Cell Cycle 2003;2:369–76. [48] Mujtaba S, Manzur KL, Gurnon JR, Kang M, Van Etten JL, Zhou MM. Epigenetic transcriptional repression of cellular genes by a viral SET protein. Nat Cell Biol 2008;10:1114–22. [49] Mujtaba S, Winer BY, Jaganathan A, Patel J, Sgobba M, Schuch R, et al. Anthrax SET protein: a potential virulence determinant that epigenetically represses NF-kappaB activation in infected macrophages. J Biol Chem 2013;288:23458–72. [50] Drlica K, Rouviere-Yaniv J. Histonelike proteins of bacteria. Microbiol Rev 1987;51:301–19. [51] Pennini ME, Perrinet S, Dautry-Varsat A, Subtil A. Histone methylation by NUE, a novel nuclear effector of the intracellular pathogen Chlamydia trachomatis. PLoS Pathog 2010;6:e1000995. [52] Kim YH, Park KH, Kim SY, Ji ES, Kim JY, Lee SK, et al. Identification of trimethylation at C-terminal lysine of pilin in the cyanobacterium Synechocystis PCC 6803. Biochem Biophys Res Commun 2011;404:587–92.

203

[53] Alvarez-Venegas R. Bacterial SET domain proteins and their role in eukaryotic chromatin modification. Front Genet 2014;5:65. [54] Black JC, Van Rechem C, Whetstine JR. Histone lysine methylation dynamics: establishment, regulation, and biological impact. Mol Cell 2012;48:491–507. [55] Sandman K, Reeve JN. Archaeal histones and the origin of the histone fold. Curr Opin Microbiol 2006;9:520–5. [56] Forbes AJ, Patrie SM, Taylor GK, Kim YB, Jiang L, Kelleher NL. Targeted analysis and discovery of posttranslational modifications in proteins from methanogenic Archaea by top-down MS. Proc Natl Acad Sci USA 2004;101:2678–83. [57] Baumbusch LO, Thorstensen T, Krauss V, Fischer A, Naumann K, Assalkhou R, et al. The Arabidopsis thaliana genome contains at least 29 active genes encoding SET domain proteins that can be assigned to four evolutionarily conserved classes. Nucleic Acids Res 2001;29:4319–33. [58] Niu Y, Xia Y, Wang S, Li J, Niu C, Li X, et al. A prototypic lysine methyltransferase 4 from Archaea with degenerate sequence specificity methylates chromatin proteins Sul7d and Cren7 in different patterns. J Biol Chem 2013;288:13728–40. [59] Botting CH, Talbot P, Paytubi S, White MF. Extensive lysine methylation in hyperthermophilic crenarchaea: potential implications for protein stability and recombinant enzymes. Archaea 2010;2010. [60] Xia Y, Niu Y, Cui J, Fu Y, Chen XS, Lou H, et al. The helicase activity of hyperthermophilic archaeal MCM is enhanced at high temperatures by lysine methylation. Front Microbiol 2015;6:1247. [61] Roguev A, Schaft D, Shevchenko A, Pijnappel WW, Wilm M, Aasland R, et al. The Saccharomyces cerevisiae Set1 complex includes an Ash2 homologue and methylates histone 3 lysine 4. EMBO J 2001;20:7137–48. [62] Ryu HY, Rhie BH, Ahn SH. Loss of the Set2 histone methyltransferase increases cellular lifespan in yeast cells. Biochem Biophys Res Commun 2014;446:113–8. [63] Stassen MJ, Bailey D, Nelson S, Chinwalla V, Harte PJ. The drosophila trithorax proteins contain a novel variant of the nuclear receptor type DNA binding domain and an ancient conserved motif found in other chromosomal proteins. Mech Dev 1995;52:209–23. [64] Nagy PL, Griesenbeck J, Kornberg RD, Cleary ML. A trithorax-group complex purified from Saccharomyces cerevisiae is required for methylation of histone H3. Proc Natl Acad Sci USA 2002;99:90–4. [65] Sourvinos G, Morou A, Sanidas I, Codruta I, Ezell SA, Doxaki C, et al. The downregulation of GFI1 by the EZH2-NDY1/KDM2BJARID2 axis and by human cytomegalovirus (HCMV) associated factors allows the activation of the HCMV major IE promoter and the transition to productive infection. PLoS Pathog 2014;10:e1004136. [66] Ropero S, Esteller M. The role of histone deacetylases (HDACs) in human cancer. Mol Oncol 2007;1:19–25. [67] Walkinshaw DR, Tahmasebi S, Bertos NR, Yang XJ. Histone deacetylases as transducers and targets of nuclear signaling. J Cell Biochem 2008;104:1541–52. [68] Delcuve GP, Khan DH, Davie JR. Roles of histone deacetylases in epigenetic regulation: emerging paradigms from studies with inhibitors. Clin Epigenetics 2012;4:5. [69] Wang YC, Peterson SE, Loring JF. Protein post-translational modifications and regulation of pluripotency in human stem cells. Cell Res 2014;24:143–60. [70] Johnstone RW, Licht JD. Histone deacetylase inhibitors in cancer therapy: is transcription the primary target? Cancer Cell 2003;4:13–8. [71] Palmer JM, Perrin RM, Dagenais TR, Keller NP. H3K9 methylation regulates growth and development in Aspergillus fumigatus. Eukaryot Cell 2008;7:2052–60.

IV.  Model Organisms of Epigenetics

204

13.  Epigenetic Modifications in Eukaryotes and Prokaryotes are Equally Intriguing

[72] Eissenberg JC, Shilatifard A. Histone H3 lysine 4 (H3K4) methylation in development and differentiation. Dev Biol 2010;339: 240–9. [73] Zhou X, Ma H. Evolutionary history of histone demethylase families: distinct evolutionary patterns suggest functional divergence. BMC Evol Biol 2008;8:294. [74] Wang J, Lu F, Ren Q, Sun H, Xu Z, Lan R, et al. Novel histone demethylase LSD1 inhibitors selectively target cancer cells with pluripotent stem cell properties. Cancer Res 2011;71:7238–49. [75] Cloos PA, Christensen J, Agger K, Helin K. Erasing the methyl mark: histone demethylases at the center of cellular differentiation and disease. Genes Dev 2008;22:1115–40. [76] Kahl P, Gullotti L, Heukamp LC, Wolf S, Friedrichs N, Vorreuther R, et al. Androgen receptor coactivators lysine-specific histone demethylase 1 and four and a half LIM domain protein 2 predict risk of prostate cancer recurrence. Cancer Res 2006;66:11341–7. [77] Yin F, Lan R, Zhang X, Zhu L, Chen F, Xu Z, et al. LSD1 regulates pluripotency of embryonic stem/carcinoma cells through histone deacetylase 1-mediated deacetylation of histone H4 at lysine 16. Mol Cell Biol 2014;34:158–79. [78] Guo X, Shi M, Sun L, Wang Y, Gui Y, Cai Z, et al. The expression of histone demethylase JMJD1A in renal cell carcinoma. Neoplasma 2011;58:153–7. [79] Cheema MS, Ausio J. The structural determinants behind the epigenetic role of histone variants. Genes (Basel) 2015;6:685–713. [80] Rossetto D, Avvakumov N, Cote J. Histone phosphorylation: a chromatin modification involved in diverse nuclear events. Epigenetics 2012;7:1098–108. [81] Zheng Y, John S, Pesavento JJ, Schultz-Norton JR, Schiltz RL, Baek S, et al. Histone H1 phosphorylation is associated with transcription by RNA polymerases I and II. J Cell Biol 2010;189:407–15. [82] Sawicka A, Seiser C. Sensing core histone phosphorylation: a matter of perfect timing. Biochim Biophys Acta 2014;1839:711–8. [83] Jenuwein T, Allis CD. Translating the histone code. Science 2001;293:1074–80. [84] Kar S, Sakaguchi K, Shimohigashi Y, Samaddar S, Banerjee R, Basu G, et al. Effect of phosphorylation on the structure and fold of transactivation domain of p53. J Biol Chem 2002;277:15579–85. [85] Thomson S, Mahadevan LC, Clayton AL. MAP kinase-mediated signalling to nucleosomes and immediate-early gene induction. Semin Cell Dev Biol 1999;10:205–14. [86] Wei Y, Yu L, Bowen J, Gorovsky MA, Allis CD. Phosphorylation of histone H3 is required for proper chromosome condensation and segregation. Cell 1999;97:99–109. [87] Hans F, Dimitrov S. Histone H3 phosphorylation and cell division. Oncogene 2001;20:3021–7.

[88] Yun M, Wu J, Workman JL, Li B. Readers of histone modifications. Cell Res 2011;21:564–78. [89] Nguyen AL, Gentilello AS, Balboula AZ, Shrivastava V, Ohring J, Schindler K. Phosphorylation of threonine 3 on histone H3 by haspin kinase is required for meiosis I in mouse oocytes. J Cell Sci 2014;127:5066–78. [90] Basnet H, Su XB, Tan Y, Meisenhelder J, Merkurjev D, Ohgi KA, et al. Tyrosine phosphorylation of histone H2A by CK2 regulates transcriptional elongation. Nature 2014;516:267–71. [91] Litchfield DW. Protein kinase CK2: structure, regulation and role in cellular decisions of life and death. Biochem J 2003;369:1–15. [92] Garcia-Garcia T, Poncet S, Derouiche A, Shi L, Mijakovic I, Noirot-Gros MF. Role of protein phosphorylation in the regulation of cell cycle and DNA-related processes in bacteria. Front Microbiol 2016;7:184. [93] Schmidl SR, Gronau K, Pietack N, Hecker M, Becher D, Stulke J. The phosphoproteome of the minimal bacterium Mycoplasma pneumoniae: analysis of the complete known Ser/Thr kinome suggests the existence of novel kinases. Mol Cell Proteomics 2010;9:1228–42. [94] Ramirez-Guadiana FH, Barajas-Ornelas Rdel C, Corona-Bautista SU, Setlow P, Pedraza-Reyes M. The recA-dependent SOS response is active and required for processing of DNA damage during Bacillus subtilis sporulation. PLoS One 2016;11:e0150348. [95] Schmidl SR, Gronau K, Hames C, Busse J, Becher D, Hecker M, et al. The stability of cytadherence proteins in Mycoplasma pneumoniae requires activity of the protein kinase PrkC. Infect Immun 2010;78:184–92. [96] Strahl BD, Allis CD. The language of covalent histone modifications. Nature 2000;403:41–5. [97] Hartzog GA, Tamkun JW. A new role for histone tail modifications in transcription elongation. Genes Dev 2007;21:3209–13. [98] Fischle W, Wang Y, Allis CD. Histone and chromatin cross-talk. Curr Opin Cell Biol 2003;15:172–83. [99] Wang H, Cao R, Xia L, Erdjument-Bromage H, Borchers C, Tempst P, et al. Purification and functional characterization of a histone H3-lysine 4-specific methyltransferase. Mol Cell 2001;8:1207–17. [100] Cheung P, Tanner KG, Cheung WL, Sassone-Corsi P, Denu JM, Allis CD. Synergistic coupling of histone H3 phosphorylation and acetylation in response to epidermal growth factor stimulation. Mol Cell 2000;5:905–15. [101] Liarzi O, Barak R, Bronner V, Dines M, Sagi Y, Shainskaya A, et al. Acetylation represses the binding of CheY to its target proteins. Mol Microbiol 2010;76:932–43.

IV.  Model Organisms of Epigenetics

C H A P T E R

14 Drosophila Epigenetics Ashley M. Karnay*, Felice Elefant** *Drexel University College of Medicine, Philadelphia, PA, United States; **Drexel University, Philadelphia, PA, United States

O U T L I N E Introduction: Drosophila as a Model Organism in Epigenetic Research

Epigenetic Modification of Histone Proteins Regulate Chromatin Packaging and Gene Control in Drosophila 207 Histone Acetylation 207 Histone Methylation 208 Position Effect Variegation Heterochromatin Within the Drosophila Genome Protein Regulators of Position Effect Variegation Epigenetic Histone Modifications Regulate PEV in Drosophila

209 209 210 210

The Role of Epigenetics During Drosophila Development: Epigenetic Memory 211 Epigenetic Modifications Maintain Patterns of HOX Gene Expression Throughout Drosophila Development 211 Mechanisms of Silencing and Activation 212 PcG Proteins and Gene Repression 213 TrxG Proteins and Gene Activation 214 PREs and PcG-TrxG–Mediated Gene Expression 214 Transgenerational Epigenetic Inheritance: “Epigenetic Memory” 215

INTRODUCTION: DROSOPHILA AS A MODEL ORGANISM IN EPIGENETIC RESEARCH Epigenetics is the study of phenotypic variations arising from the heritable changes in gene expression that result from alterations in the genome that do not Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00014-6 Copyright © 2017 Elsevier Inc. All rights reserved.

Epigenetics During S Phase Epigenetics During M Phase

205

Dosage Compensation The MSL Complex/DCC Regulation of DCC Targeting to the X Chromosome DCC Spreading and Target Recognition Transcriptional Activation of Active X-Linked Genes by the DCC The Epigenetic Language in Postmitotic Neurons Underlying Cognitive Function Activity-Dependent Histone Acetylation and Cognitive Function HAT: HDAC Interplay in Cognitive Function Deregulation of HAT Activity in the Etiology of Alzheimer’s Disease Environmental Enrichment Improves Learning and Memory in Drosophila

215 217 217 217 218 219 219 220 220 220 221 222

Conclusions

223

References

224

involve changes in the nucleotide sequence. Drosophila melanogaster has been utilized for over a century as a powerful model organism to help elucidate many of the fundamental concepts of human epigenetics. Conrad H. Waddington conducted studies in 1942 showing that altering the developmental environment of Drosophila embryos resulted in mature flies with varying thorax

205

206

14.  DROSOPHILA EPIGENETICS

and wing structures [1]. Herman J. Muller followed with his research using X-rays to induce genetic mutations, after which he observed a mosaic-like expression of the mutant and wildtype phenotype in the offspring, a phenomenon he termed “position effect variegation” (PEV) [2,3]. Findings from subsequent studies in Drosophila by Jack Shultz and later by Ed Lewis, led to the characterization of the family of genes whose protein products are responsible not only for the regulation of this epigenetic phenomenon, but also for the regulation of normal development and the transmission of epigenetic programs across generations [4]. There are many advantages that make Drosophila an ideal model system for the study of epigenetic gene control in a variety of biological processes. First, the fly genome is composed of four pairs of chromosomes containing approximately 17,000 genes [5], compared to the human genome, which is composed of roughly 20,000 protein-coding genes [6,7]. Despite this difference, more than 60% of the protein-coding genes within the fly genome have human homologs [8], and similarly, 61% of human disease–linked genes have homologs in Drosophila [9], making it an excellent model for translational or comparative genetic and epigenetic studies. Further, the polytene chromosome located within the larval salivary glands of Drosophila, provide scientists the unique ability to directly visualize structural changes that result from the epigenetic modifications of chromatin [10,11]. These chromosomes develop from normal chromosomes and are a result of specialized cells undergoing multiple rounds of replication without dividing. They have a characteristic banding pattern that arises as a result of regional chromatin condensation, with areas of densely packaged chromatin forming dark bands that contrast with areas of less compacted light bands. Epigenetic modifications to the chromatin result in the localized unwinding of the chromatin, which can be directly visualized as chromosome “puffs,” enabling the identification of regions of active transcription [12]. Recent findings indicate a correlation between the positioning of cis-regulatory elements, patterns of epigenetic modifications, and association with regulatory proteins between the polytene chromosome and diploid interphase chromosomes, making them an ideal model for the study of chromatin dynamics and regulation across cell types [13,14]. Another advantage of using Drosophila is its short life cycle, which translates to short generation times, allowing researchers the ability to rapidly assess the transgenerational effects of epigenetic modifications. The fly’s distinct morphological developmental stages provide mechanistic insight into the genetic and epigenetic changes that influence biochemical and physiological processes and how these changes impact developmental progress, survival, and longevity [10,15].

Drosophila has also proven to be a valuable tool for the study of epigenetic mechanisms underlying neurobiology and human neurodegenerative diseases. The fly nervous system is complex and highly evolved, and shares important regulatory genes and molecular pathways with the human nervous system [16]. The fly brain contains more than 100,000 neurons forming a complex synaptic network that mediates its ability to perform measurable higher-order functions similar to those in humans, such as learning and memory, feeding, courtship, aggression, and circadian rhythm, to name a few [15]. Many genes found in the Drosophila genome have structural and/or functional homologs in humans and as a result, specific anatomical structures within the fly have become attractive model systems for studying epigenetic gene control mechanisms underlying the underpinnings of biological processes mirrored in humans. The Drosophila neuromuscular junction is an excellent example of this. Genetic manipulation of synaptic components, coupled with sensitive electrophysiological techniques, have enabled its detailed characterization, making it a valuable model system for studying synapse function, regulation, and plasticity [17]. Further, fly synapses have a predictable structure, are relatively large in size (enabling easy manipulation), and are morphologically and functionally comparable to those of humans, making them pertinent for the study of numerous aspects of the mammalian central nervous system (CNS), including synaptogenesis, the events underlying synaptic transmission, and pathogenesis of neurodegenerative diseases [18]. The Drosophila visual system is also a powerful neural circuit used to provide insights into the overt toxic effects of individual human diseasecausing genes or neurodegenerative agents [8]. This is because eye degeneration defects are not lethal and can be easily quantitatively measured based on the direct visualization of cell loss within the organized pattern of cells. This system has widely been used to measure the neurodegenerative effects of expanded polyglutamine repeats characteristic of Huntington’s disease (HD), and spinocerebellar ataxia and amyloid beta (Aβ) aggregates in Alzheimer’s disease (AD) [19,20]. Importantly, this system has proved instrumental in enabling the identification of epigenetic modulators that can rescue such neurodegenerative defects. Epigenetic control mechanisms involved in the integration of higher-order sensory information and learning and memory have also been successfully studied in Drosophila using the mushroom body (MB) as a well-characterized cognitive model. The MB is a bilateral brain structure composed of a complex network of unmyelinated axons, glial processes, dendrites, and specialized neurons called Kenyon cells [21,22]. The Drosophila MB is known to regulate a range of behavioral

IV.  Model Organisms of Epigenetics



EPIGENETIC MODIFICATION OF HISTONE PROTEINS REGULATE CHROMATIN PACKAGING AND GENE CONTROL IN DROSOPHILA

and physiological functions that range from olfactory learning and memory and courtship conditioning to decision-making under uncertain conditions [23–27]. Coupled with the multitude of experimental paradigms aimed at analyzing the behavioral consequences of genetic interventions, the MB provides scientists with an easy-to-use cognition-linked circuit model of the brain [28]. Importantly, Drosophila can perform a multitude of tasks utilized in a variety of behavioral paradigms that can be studied and quantified. These behaviors range from the simple, such as aggression [29], to more complex functions, such as locomotor or circadian sleep/ wake cycles [30]. In particular, courtship conditioning in Drosophila is a complex behavioral learning paradigm that requires multimodal sensory input, involving chemosensory, mechanosensory, visual, and olfactory pathways and is thus well suited to study experiencedependent synaptic plasticity [31–34]. Complex behaviors, such as associative learning and memory, can be quantitatively studied using a variety of different assays that quantitatively measure courtship conditioning and the ability of Drosophila to learn to avoid a specific object or odor after experiencing it in conjunction with a painful stimulus [35]. In summary, as a result of the short life cycle, ease of environmental manipulation, the high degree of genetic similarity between fly and human, the facility of genetic manipulation, and the growing number of experimental paradigms available to study the consequences of such manipulations, D. melanogaster continues to be an ideal model organism in the study of epigenetics in humans.

EPIGENETIC MODIFICATION OF HISTONE PROTEINS REGULATE CHROMATIN PACKAGING AND GENE CONTROL IN DROSOPHILA Epigenetic alterations in gene expression profiles, which can be stable or transient, are the result of chromatin-remodeling factors and posttranslational modifications (PTM) to chromatin components that result in changes to the degree of chromatin compaction, which in turn impact the accessibility of the transcription machinery to the DNA template [36,37]. Such epigenetic gene regulatory mechanisms have long been studied effectively using Drosophila as a model system. In eukaryotic cells, the DNA is wrapped around octamers of four histone proteins (H2A, H2B, H3, and H4) to form the basic units of chromatin called nucleosomes [16,38]. This organization presents a major physical hurdle for the transcriptional machinery and must be overcome to enable gene transcription in vivo. One of the main mechanisms utilized by the cell to overcome the repressive nucleosomal environment is through the covalent PTM

207

of the core histone proteins. These modifications include acetylation, methylation, phosphorylation, ubiquitination, sumoylation, and ADP-ribosylation [39]. Histones are often concurrently modified on several residues and there is also a dynamic interplay between histone modifications and DNA modifications (such as DNA methylation), thus creating staggering combinatorial possibilities for gene regulation [40]. These modifications not only work to modulate the structure of the chromatin itself, but can also serve as distinct docking sites for other protein complexes that participate in gene transcription or repression [36]. Moreover, the recruitment of additional chromatin-interacting proteins often depends not solely on the presence of a particular PTM, but also relies on a combination of local marks and the overall context within which those marks are presented [41]. In this way, epigenetic patterns serve as a “code” that is generated by enzymes that function as “writers” to add specific PTMs to histones, recognized by “readers” that are comprised of regulatory proteins that control chromatin packaging, and removed by enzymes that serve as “erasers.” These chromatin-regulatory proteins work in concert to determine the transcriptional programs employed under transient cellular conditions, timing-specific developmental stages, in addition to maintaining patterns of transcription stably across generations [42]. Two of the most well-characterized epigenetic histone PTMs that are fully modeled in Drosophila will be discussed in the following sections.

Histone Acetylation One of the most extensively studied PTMs involved in gene regulation is histone acetylation. This epigenetic modification is generated by members of the family of histone acetyltransferase (HAT) enzymes. These enzymes catalyze the transfer of an acyl moiety from acyl coenzyme A to lysine residues on the N-terminal tails of the histones [37]. Acetylation alters the chromatin state in favor of a more relaxed, open conformation that is conducive to transcriptional activation which is achieved through a number of mechanisms (Fig. 14.1) [42,43]. Acetylation is believed to neutralize the positive charge on the lysine residue, thus reducing the affinity between histones and DNA, allowing for a more open or relaxed chromatin state [16,44]. More recent evidence points to acetylated lysine residues serving as molecular docking sites for proteins that act to further relax the chromatin structure, such as bromodomain-containing ATP-dependent chromatin-remodeling complexes and various members of the transcriptional machinery [43,45]. Indeed, Tat-interactive protein 60 kDa (Tip60), a member of the MYST HAT super family that is recruited to chromatin, is a member of the Tip60 protein complex shown to possess not only HAT activity, but also

IV.  Model Organisms of Epigenetics

208

14.  DROSOPHILA EPIGENETICS

FIGURE 14.1  Chromatin structure regulates transcriptional activity. Histone acetylation is regulated by the antagonistic activities of histone acetyltransferases (HATs) that acetylate histones and histone deacetylases (HDACs) that deacetylate histones. (A) Histone deacetylation induces a closed chromatin configuration with concomitant transcriptional repression. (B) Histone acetylation opens chromatin packaging to allow for transcriptional activation. Source: Adapted from Johnstone RW. Histone-deacetylase inhibitors: novel drugs for the treatment of cancer. Nat Rev Drug Discov 2002;1(4):287–99 [50].

ATPase, DNA helicase, as well as DNA-binding activities [45]. The relaxed hyperacetylated chromatin conformation [44] is marked by an increase in the sensitivity of the chromatin to DNase digestion, as well as to an increase in the flexibility of the nucleosomal DNA due to the reduction in conformation hindrance provided by nucleosomal compaction [46]. This chromatin unwinding allows access of the transcriptional machinery to the regulatory elements of the genes, facilitating their expression. It has been shown that TBP-associated factor (TAFII), which is part of the transcription initiation factor TFIID complex, has intrinsic HAT activity, thus enabling the unwinding of regions around specific genes to allow TFIID access to previously repressed genes [47]. Others have also shown that increased access to gene promoter regions via histone acetylation facilitates the ability of RNA POLII to initiate transcription at specific genetic loci [48]. Histone acetylation is a reversible process that is conversely regulated by histone deacetylase (HDACs) enzymes. HDACs catalyze the removal of the acyl moiety from the histone tail lysine residues, leading to the formation of a more compact and transcriptionally repressed chromatin state [49].

Histone Methylation Posttranslational methylation of histones is a wellstudied modification known to influence an array of biological processes, especially in the context of development. Lysine and arginine resides on histones 3 and 4 are are primary targets of methylation with

mono-, di-, or trimethylation being observed [51–53]. Methyl groups tend to exhibit a more stable association with histones relative to acetyl groups; therefore, it is no surprise that they are most often implicated in transgenerational epigenetic inheritance, a process studied very effectively in Drosophila. SET-domain containing lysine methyltransferases (KMTs) are capable of modifying both histone and nonhistone proteins, catalyzing the transfer of methyl moieties from S-adenosyl methionine to the ε-amino groups of lysine residues [51]. Proteins with PHD finger and chromodomains are most commonly associated with the binding of methyl residues; however, a variety of other methyl-binding domains exist [54]. The stepwise and variable pattern of histone methylation provides a heterogeneous platform for dynamically regulated transcription. As methylation can occur on multiple sites within the same histone and different lysine residues are capable of varying degrees of methylation, a combination of methyl moieties is possible, which translates to differential gene expression states. For example, H3K4me3 has been associated with regions of active transcription, while H3K27me3 marks are indicative of transcriptional silencing [55]. Leading to further complexity, methylation marks can be associated with opposing activities; H4K20me is highly enriched within the promoter and coding regions of actively transcribed genes [56], while H4K20me has also been shown to be a negative regulator of transcription, demonstrated in knockdown assays [57]. Methylation can also indirectly lead to the reorganization of the 3D

IV.  Model Organisms of Epigenetics



Position Effect Variegation

structure of chromatin, by designating regions of DNA for recognition by chromatin-remodeling complexes, resulting in the induction of higher-order chromatin conformations that suppress transcription [58–60]. Similar to acetylation, histone methylation is a reversible process, regulated by a variety of lysine demethylases, which are specific for each degree of methylation [41]. In summary, acetylation and methylation, as well as phosphorylation, ubiquitination, sumoylation, and ADP-ribosylation of histones, comprise a complex array of possible epigenetic modifications that confers flexibility and the ability to rapidly modulate according to the temporal and spatial requirements of gene transcription during development and environmental changes.

POSITION EFFECT VARIEGATION Covalent histone modifications have the ability to regulate complex gene expression profiles by modulating the degree of compaction of the chromatin fiber. Heterochromatin is characterized by a more compact chromatin state generally correlating to the repression of transcription. It is established and maintained by combinations of different marks dictated by the interaction between histone-modifying enzymes and chromosomal structural proteins, leading to demarcation of structurally and functionally distinct chromosomal domains. In contrast, euchromatin is less compacted and reflects areas of active transcription. Once heterochromatin formation is initiated, it is capable of spreading to and silencing nearby euchromatic genes, resulting in the formation of facultative heterochromatin, an epigenetic-mediated mechanism for permanent gene inactivation [61,62]. This process is well modeled in Drosophila, when genes that are normally expressed are aberrantly moved, either by rearrangement or transposition, to a site near the euchromatin–­ heterochromatin (Eu–Het) transition zone and as a result, are abnormally packaged into a heterochromatic form and silenced [62,63]. This phenomenon termed position effect variegation (PEV), was first observed in 1925 by Alfred Sturtevant while studying mutations in the Bar eye gene responsible for establishing the number of facets within the Drosophila eye during development. He observed that gene duplication mutations resulted in the unexpected increase in the number of copies of the Bar genes, but intriguingly the resultant phenotype was that of a reduction in the eye size. This suggested that the phenotype was not due to the number of copies of the gene, but rather to the positioning or arrangement of the genes within the chromosome, an influence he called “position effect” [64]. Following up on this work, Herman J. Muller used X-rays to induce mutations within the white gene, a mutation that results in the expression of a white

209

eye instead of the wildtype red eye. After X-ray exposure, he observed a unique phenotype in which each eye expressed a combination of the red and white pigmentation, that is, a “mottled” or “variegated” phenotype. He concluded that the gene itself was not mutated, evidenced by the successful expression in a subset of photoreceptor cells, but that aberrant repositioning of the gene likely led to abnormal silencing in some cells, resulting in the variegated phenotype [62,64]. Genetic rearrangements resulting in the repositioning of euchromatic genes near a Eu–Het transition zone are now known to lead to the repression of chromatin packaging and silencing of normally expressed genes, a phenomenon called PEV. The regulation and maintenance of these dynamic regions of the genome clearly have an important influence on gene expression, and Drosophila has proven to be an ideal model system for the analysis of the epigenetic processes controlling heterochromatin dynamics and PEV.

Heterochromatin Within the Drosophila Genome Heterochromatin, originally identified via cytological studies using the polytene chromosome, is defined as chromatin regions that maintain a dense staining and condensation pattern throughout the cell cycle [61,64]. These domains are composed mostly of tandem repeat motifs called “satellite” DNA sequences and transposable elements (TEs), such as DNA transposons and retroviruses. In contrast, euchromatin has a variable condensation and staining pattern, reflecting regions of active gene transcription [62,63]. In Drosophila, these chromatin domains can be characterized by their expression of unique combinations of histone marks. In an extensive analysis performed by Yin et al. [65] utilizing chromatin immunoprecipitation in combination with high-throughput sequencing (ChIP-Seq) in the adult fly, a high resolution whole-genome map of the key histone modifications corresponding to regions of euchromatin and heterochromatin regions was generated. In Drosophila, euchromatic regions are generally enriched in histone lysine methylation at H3K4, H3K79, and H3K36 [66]; histone acetylation of H3K9, H4K16, and H3K14 [63,66]; and phosphorylation of H3S10 [62,67]. In contrast, within the Drosophila genome there are distinct regions of heterochromatin, each displaying a unique combination of H3K9, H3K27, and H4K20 histone methylation marks [55]. For example, heterochromatin found at the chromosomal telomeres contain H3K9 mono-, di-, and trimethylated histone states [61]. Regions of heterochromatin within the chromosome’s centromeres are termed centric or chromocenter heterochromatin, and are specifically enriched in the marks H3K9me3 and H3K27me [55,61]. Pericentric heterochromatin, regions flanking centric heterochromatin, contains H4K20 trimethylation [61,66], H3K9 dimethylation, and H3K27 trimethylation

IV.  Model Organisms of Epigenetics

210

14.  DROSOPHILA EPIGENETICS

[16,55]. In contrast to euchromatin, histones within heterochromatic regions are generally hypoacetylated, with only H4K12ac found within chromocenter heterochromatin [68].

Protein Regulators of Position Effect Variegation Loss-of-function (LOF) mutational analysis has allowed identification of a number of genes whose protein products are associated with the establishment and maintenance of heterochromatin and PEV. These genes are classified as either enhancers of PEV [enhancer of variegation or E(var)], whose LOF results in an increase in gene silencing or suppressors of PEV [suppressor of variegation or Su(var)] whose LOF results in a loss of gene silencing [62]. The action of these proteins, which are components of heterochromatin, as well as enzymes involved in the modification of chromatin and histone proteins, correlates with changes in the chromatin state seen in PEV [61,62]. E(var) genes encode proteins responsible for chromatin decondensation and increased gene expression; therefore mutations in these genes lead to the enhancement of heterochromatic gene silencing. One such gene unique to Drosophila is E(var)3-9, encoding a zinc finger protein, which was originally identified following its LOF by the appearance of the variegated eye pigmentation phenotype, indicative of the induction of heterochromatin spreading into and silencing of the adjacent white (wt) gene [69]. E(var) proteins have been shown to counteract the suppressing effects of heterochromatin in a variety of ways. The main mechanism of action involves epigenetic-mediated chromatin remodeling, leading to increased accessibility of gene promoter regions to the transcriptional machinery [70]. Many of the identified E(var) genes, such as jumeaux and E(var)393E, encode general transcription factors, enhancing gene expression in a way independent of chromatin modification [71,72]. E(var) proteins have also been proposed to work by inhibiting the formation of heterochromatin-mediated silencing complexes, thereby maintaining a euchromatic state conducive to gene expression [73,74]. Conversely, LOF of Su(var) genes leads to the suppression of heterochromatic gene silencing, while additional copies of the genes enhance heterochromatin formation and spreading in Drosophila [61]. These proteins are either structural components of heterochromatin itself or histone-modifying enzymes responsible for the epigenetic-regulation of heterochromatin [75]. Three important Su(var) genes play key roles in PEV-mediated gene silencing in Drosophila: Su(var)3-9, Su(var)2-5, and Su(var)4-20. Su(var)3-9 encodes an H3K9 histone methyltransferase containing two evolutionarily conserved domains: [1] a C-terminal SET

domain, which shares homology with Su(var)3-9, Enhancer of zeste (EZ), and Trithorax proteins, and [2] a methyl-binding chromodomain. Both domains are essential for the association of Su(var)3-9 with heterochromatin [76]. Su(var)2-5 encodes the structural protein, heterochromatin protein 1 (HP1), which possesses multiple protein and methyl-binding domains. For example, its chromodomain allows it to bind to heterochromatin-specific H3K9me2/3 marks [61,62], while its C-terminal chromo shadow domain provides the ability to dimerize with itself and to interact with Su(var)3-9 [75]. The Su(var)4-20 protein is a H4K20specific histone ­ methyltransferase that controls the H4K20me3 characteristic of pericentric heterochromatin [62]. Additional genes implicated in the spread of pericentric heterochromatin observed in PEV include Su(var) 3-1 encoding the protein kinase JIL-1, Su(var)3-3 encoding a H3K4-specific demethylase homologous to the human LSD1 protein [77], and Su(var)3-26 whose protein product is the HDAC1 [78,79].

Epigenetic Histone Modifications Regulate PEV in Drosophila As different sets of specific histone modifications are found within euchromatin and heterochromatin, it is not surprising that the formation of pericentric heterochromatin and its encroachment into adjacent regions of euchromatin require a sequence of events involving the removal of one set and the replacement of another set of histone modifications (Fig. 14.2). To initiate the transition from euchromatin to heterochromatin, the active chromatin marks H3K9ac and H3K4me2 are removed by the histone deacetylase Su(var)3-26 (HDAC1) and the demethylase Su(var)3-3 (LSD1), respectively. Removal of these marks are prerequisites for the di- and trimethylation of H3K9, the histone modification hallmark of repressed chromatin regions, by Su(var)3-9 [80]. H3K9me2/3 then serves as a binding platform for HP1, which, through its multiple diverse domains, is capable of recruiting a variety of proteins to participate in the cisspreading of the newly initiated heterochromatin region into neighboring domains. Through a similar mechanism, TEs embedded within Drosophila euchromatin are epigenetically silenced by a class of intervening RNAs called piwi-interacting RNAs (piRNAs), which mediate the deposition of H3K9me marks and spreading of heterochromatin to adjacent gene regions, selectively suppressing TE’s deleterious effects on the Drosophila genome [62,81,82]. HP1 has been shown to recruit H4K20-specific HMTase Su(var)4-20, which catalyzes the trimethylation of H4K20, an evolutionarily conserved mark of pericentric heterochromatin [66]. Su(var)3-7, a zinc finger protein, binds directly with HP1 within multiple heterochromatic regions, as evidenced by

IV.  Model Organisms of Epigenetics



THE ROLE OF EPIGENETICS DURING DROSOPHILA DEVELOPMENT: EPIGENETIC MEMORY

211

FIGURE 14.2  Transition from euchromatin to heterochromatin. To initiate the transition from euchromatin to heterochromatin, the active chromatin marks H3K9ac and H3K4me2 are removed by the histone deacetylase Su(var)3-26 (HDAC1) and the demethylase Su(var)3-3 (LSD1), respectively. Removal of these marks allow for the deposition of the repressive marks di- and trimethylation of H3K9 by Su(var)3-9 [80]. ­H3K9me2/3 serves as a binding platform for HP1, which recruits a variety of proteins to participate in the cis-spreading of the newly initiated heterochromatin region into neighboring euchromatin domains. Source: Adapted from Grewal SI, Elgin SC. Transcription and RNA interference in the formation of heterochromatin. Nature 2007;447(7143):399–406 [85].

coimmunoprecipitation assays and direct visualization on polytene and embryonic chromosomes [83]. Based on the observation that HP1 and Su(var)3-7 remain closely associated with pericentric heterochromatin regions long after its induction, it has been proposed that these proteins serve as stable structural elements, maintaining the compacted chromatin state characteristic of these transcriptionally repressed regions; however, the exact role in PEV remains unknown [83,84]. It has been proposed that the genetic rearrangements resulting from aberrant gene translocations results in the removal of a specific barrier zone or boundary element demarcating adjacent euchromatin and heterochromatin regions, thus allowing the migration of chromatin compaction not normally seen in nuclei with structurally normal chromosomes [62,74]. Although the molecular nature of these buffering elements remains uncharacterized, it is speculated that their disruption or removal leads to the initiation of heterochromatin, which can subsequently propagate as far as 175 kb into nearby genes, resulting in the shifting of Eu–Het junctions relative to proximal genes and PEV gene silencing [86]. In Drosophila, this process is mediated both by HP1 itself and by the coordinated crosstalk between multiple histone modifications. The interdependence of various PTMs of histone proteins provides a means by which chromatin dynamics can be regulated throughout the Drosophila genome; PEV-mediated gene silencing is an important example. In summary, the large body of work in Drosophila on PEV has substantially elevated our understanding of

heterochromatin and euchromatin dynamics, and the regulatory influence of epigenetics.

THE ROLE OF EPIGENETICS DURING DROSOPHILA DEVELOPMENT: EPIGENETIC MEMORY Epigenetic Modifications Maintain Patterns of HOX Gene Expression Throughout Drosophila Development During the early stages of development, the identity and function of each cell within an organism is determined by a unique pattern of gene transcription determined via the integration of endogenous and exogenous signals. Initiated by transcription factors, these complex changes in the cellular transcriptome are epigenetically maintained and inherited as part of cellular differentiation and development in a process referred to as “epigenetic memory” [87,88]. The repressed and activated expression states of developmental genes are maintained by the antagonistic and highly conserved function of the Polycomb group (PcG) and Trithorax group (TrxG) proteins [58,89,90]. First characterized in Drosophila, PcG and TrxG proteins are required for the persistent regulation of appropriate expression patterns of the developmental homeotic (Hox) genes. Hox genes encode transcription factors whose regulated expression is responsible for the specification of segmental identity in the developing fly embryo [91]. These genes are clustered into two

IV.  Model Organisms of Epigenetics

212

14.  DROSOPHILA EPIGENETICS

separate gene complexes: the Antennapedia (ANT-C) and Bithorax (BX-C) complexes, with the former regulating development of the head and regions of the thorax and the latter regulating the abdominal sections and remaining regions of the thorax (Fig. 14.3) [91–93]. During the initial stages of development, after a period of transcriptional repression, maternally inherited transcription factors establish a spatial and temporal pattern of HOX gene expression that will eventually drive specific tissue differentiation and body segment development [94]. Multimeric PcG and TrxG protein complexes epigenetically mediate the long-term maintenance of these gene expression programs throughout the life of the fly [58,89–91].

Mechanisms of Silencing and Activation The temporal and spatial regulation of Hox gene expression correlates with specific histone modifications established and maintained by TrxG and PcG proteins. Gene silencing by PcG complexes relies on the deposition of repressive H3K27me3 and H2AK119ub marks while, in contrast, TrxG-mediated gene transcription is associated with the H3K4me3, H3K36me3, and H3K27ac chromatin marks. As described throughout this chapter, these modifications enhance or repress transcription by directly altering the 3D conformation of the chromatin fiber in addition to serving as binding platforms for chromatin-remodeling factors.

FIGURE 14.3  Organization of Hox clusters in Drosophila and mouse. The Drosophila Hox genes (top) are grouped into two genomic clusters: the Antennapedia (ANT-C) and Bithorax clusters (BX-C). Expression domains of the individual Hox genes within the ANT-C and BX-C along the anteroposterior (AP) axis of the fruitfly embryo match the array of the genes along the chromosome, displaying a property termed collinearity. Experiments in the late 1980s revealed that mice also possess a set of Hox genes (bottom) similar to those found in Drosophila and that their organization along the chromosome, as well as their order of expression along the AP axis also displayed collinearity. An important difference between these two systems is that the mouse has four clusters (Hoxa, Hoxb, Hoxc, and Hoxd) instead of one, as a result of two rounds of gene duplication. The Hox complement of the mouse also reveals that some individual Hox genes have been duplicated whereas others have been lost in each cluster. Based on sequence and genomic comparisons across a wide range of phyla, the structure of the cluster ancestral to insects and mammals can be inferred (middle). Source: Adapted from Mallo M, Alonso CR. The regulation of Hox gene expression during animal development. Development 2013;140(19):3951–63 [95].

IV.  Model Organisms of Epigenetics



THE ROLE OF EPIGENETICS DURING DROSOPHILA DEVELOPMENT: EPIGENETIC MEMORY

PcG Proteins and Gene Repression The Polycomb (Pc) system is made up of three protein complexes: Pho repressive complex (PhoRC), Polycomb repressive complex 2 (PRC2), and Polycomb repressive complex 1 (PRC1) [90]. Each complex is composed of a heterogenous group of factors that work cooperatively in a stepwise manner to modify and remodel the chromatin of repressed target Hox genes (Fig. 14.4A). PhoRC is the only PcG complex capable of directly binding to target Hox DNA, a unique function mediated by one of its members, Pho. PRC2 is recruited to the DNA bound PhoRC complex, and its subunit EZ, a SET domain KMT, catalyzes the trimethylation of H3K27 on proximal histones [58,89,90]. Additional PRC2 complexes are capable of binding to this modification thorough

213

the extra sex combs (Esc) subunit, proposed to result in the propagation of H3K27 trimethylation and the coating of the remainder of the gene body, as is commonly observed in repressed Hox chromatin [58]. The PRC1 complex recognizes and binds H3K27me3 through the chromodomain of its component Pc [88,89]. The main catalytic subunit of PRC1 is the E3 ubiquitin ligase sex combs extra (Sce), and is responsible for catalyzing the ubiquitylation of H2AK118, a histone modification associated with transcriptional repression [88,90,96]. It should be noted that additional proteins other than Pho may also be sufficient for PRC1/2 recruitment and tethering to target Hox genes. Several mechanisms underlying PcG-mediated negative regulation of transcription have been proposed based on the purification of each protein complex and

FIGURE 14.4  Mechanisms of silencing and activation by PcG and TrxG protein complexes. (A) Subunits of Polycomb repressive complex 1 (PRC1) and PRC2, a selected Trithorax group (TrxG) proteins are shown. The Drosophila protein names are shown with selected vertebrate homologs listed in brackets. The Polycomb group (PcG) and TrxG proteins are recruited to Pc and Trx response elements (PREs and TREs, respectively) through a platform of sequence-specific DNA-binding proteins. (B) Molecular mechanisms. PREs are thought to be depleted of nucleosomes, and chromatin modifications occur on the flanking nucleosomes. Self-reinforcing (top and bottom) and antagonistic (middle) interactions of PcG and TrxG with chromatin are depicted. Top: Pc-mediated silencing. Bottom: TrxG-mediated activation. Source: Adapted from Steffen PA, Ringrose L. What are memories made of? How Polycomb and Trithorax proteins mediate epigenetic memory. Nat Rev Mol Cell Biol 2014;15(5):340–56 [97].

IV.  Model Organisms of Epigenetics

214

14.  DROSOPHILA EPIGENETICS

the characterization of the interactions between each component protein and their target proteins or genes (Fig. 14.4B) [98]. First, the histone modifications catalyzed by PRC1 and 2 promote chromatin compaction and this transcriptionally repressive structural alteration is reinforced by the interaction of other members of the PcG system with these histone modifications. Second, PcG-mediated transcriptional repression may also be a consequence of its interference with various steps required for transcription. For example, PRC1-mediated H2A ubiquitylation has been shown to have a direct role in preventing activated RNA Pol II release from promoter regions, limiting transcription elongation [99]. Similarly, H3K27me3 deposition throughout HOX gene transcriptional units may also impede elongation and Pol II processivity [58]. Third, PcG complexes may also serve to silence gene expression by removing activating histone modifications and/or blocking the recruitment of chromatin-remodeling factors responsible for decondensing chromatin. For example, a variant of the PRC1 complex, CHRASCH, contains the HDAC1, responsible for the removal of activating histone acetyl marks and subsequent chromatin compaction [89]. Last, PRC1, specifically PSC, directly disrupts ATP-dependent nucleosome remodeling by the SWI/SNF complex, a homolog of the Drosophila BRM complex, preventing chromatin unwinding and thus maintaining a transcriptionally silent chromatin structure [100,101]. Within the last few years, a new model for the simultaneous and coordinated PcG-mediated silencing of multiple distant Hox genes has emerged. Electron microscopy (EM) has revealed that that the PRC1 complex, as well as multiple of its core components individually, is capable of converting the open “beads-on-a-string” conformation into a highly compacted, complex 3D chromatin structure [102]. It has been shown that H3K27me3 laid down by EZ bound to target sequences results in the formation of looping structures within the chromatin, restricting access to transcriptional machinery and silencing any promoters within [58–60]. These chromatin loops may facilitate the colocalization of multiple Hox genes normally located hundreds of base pairs apart, allowing their interaction and coordinated repression within distinct regions of the nucleus [91]. In agreement with this model, repressed genes from both the ANT-C and BX-C clusters, separated within the genome by approximately 10 Mb, have been found within the same PcG repressive structure in the nucleus of the Drosophila embryo [103,104].

TrxG Proteins and Gene Activation The TrxG system is composed of a family of proteins involved in maintaining the activation of HOX gene expression and opposing the gene silencing activity of

the Pc group (Fig. 14.4A–B). TrxG proteins maintain HOX gene expression by mediating the histone modification– based open chromatin formation and active chromatin remodeling [105,106]. Absent, small, or homeotic discs 1 (Ash1) is a SET domain KMT responsible for the trimethylation of H3K36 [89] and it, along with the TrxG protein Trx, is capable of methylating H3K4 [58,88]. These marks are not only associated with decondensed chromosomal states, but also directly inhibit the PRC2-mediated deposition of repressive H3K27 trimethylation marks [107,108]. Trx is also capable of directly binding the H3K27-specific HAT CREB-binding protein (CBP), whose deposition results in both chromatin unwinding and inhibition of PRC2-mediated histone methylation [88,109]. Similar to regulation by PcG proteins, the maintenance of the transcriptional state of TrxG target genes is a consequence of gene accessibility and Pol II processivity. The PTMs catalyzed by TrxG proteins or TrxG-associated proteins facilitate the reduction in electrostatic interactions between the histone proteins and DNA, thus increasing the accessibility of the DNA template to members of the transcriptional machinery. It has been proposed that TrxG proteins work to modify nucleosomes downstream of the transcription start site, thereby facilitating the unencumbered travel of RNA Pol II. Additionally, the vertebrate bromodomain-containing protein BRD4 has been shown to bind acetylated lysine residues and promote transcriptional elongation by phosphorylating the C-terminal domain of Pol II [110]. It is speculated that BRD4’s Drosophila homolog female sterile homeotic (Fsh) may function in a similar manner [88]. Additional TrxG genes have been identified by the ability of their encoded proteins to counteract PcG-mediated silencing and upon further analysis, their specific regulatory functions were shown to involve the modulation of nucleosome positioning in the context of chromatin. One such gene brahma (brm) encodes an ATPdependent chromatin-remodeling protein responsible for altering the chromatin structure of positively regulated Hox genes, facilitating gene expression [111].

PREs and PcG-TrxG–Mediated Gene Expression The PcG and TrxG proteins maintain their repressive or activating influence on their target genes by binding to Polycomb group response elements (PREs) or Trithorax group response elements (TREs), respectively [89,112]. Only characterized within Drosophila, PREs/TREs are required for the maintenance of Hox gene expression patterns through embryogenesis [88]. These cis-regulatory sequences have been shown to range in size from between a few hundred and a thousand base pairs [89] and are located dozens of kilobases away from HOX gene transcription start sites [98]. Shown to be devoid of nucleosomes, their regulatory function is based upon

IV.  Model Organisms of Epigenetics



THE ROLE OF EPIGENETICS DURING DROSOPHILA DEVELOPMENT: EPIGENETIC MEMORY

the recruitment and persistent binding of TrxG and PcG proteins, resulting in the stable modification and modulation of nearby histones and the establishment and reinforcement of expression patterns. These domains contain binding sites for DNA-binding proteins involved in the recruitment of PcG and TrxG protein complexes [89]. A few of these targeting proteins characterized in Drosophila include PHO, PHOL, PSQ, GAF, Zeste, GRH, and DPS1 [89,98]. It has been proposed that in the repressed state, DNA-binding proteins recruit PcG complexes to PREs, while in the active state DNA-binding proteins recruit TrxG proteins to TREs [106]. Taken together, recruitment of PcG or TrxG proteins to the cis-regulatory elements appears to be dependent upon the interaction between the DNA-binding proteins, as well as other local features of the chromatin itself, such as the existing chromatin landscape or the presence of nearby enhancer sequences [88,89]. Importantly, recent evidence suggests that the association of these recruiting proteins with PREs or TREs is not mutually exclusive and that both groups can be differentially and constitutively bound to both PREs and TREs. Additionally, chromatin-associated long noncoding RNA (lncRNA) has been implicated as a factor influencing the bidirectional modulation between TrxG-mediated gene expression and PcG-mediated repression through its interaction with various TrxG/PcG factors and regulatory elements. While not discussed here, the role of lncRNA in the maintenance of gene expression patterns can be found in a number of recent reviews [113–116]. Thus, it has been proposed that PcG and TrxG factors provide a regulatory platform that can be rapidly switched between expression states based on the developmental needs of the organism at a particular developmental time point [58,88]. While PREs and TREs have not yet been characterized in mammals, the idea that CpG islands or 1–2 kb DNA regions enriched in cytosine/guanine content may serve as the mammalian analogs is becoming increasingly popular. Evidence supporting this notion is reviewed in Refs. [117,118]. In summary, the TrxG and PcG are well-characterized families of enzymes responsible for the long-term epigenetic control of Hox gene transcription. Studies in Drosophila elucidating their individual and interdependent functions in modulating gene expression have served as an important foundation for further investigation into their roles in vertebrate and mammalian developmental gene expression.

Transgenerational Epigenetic Inheritance: “Epigenetic Memory” Once the epigenetic programs regulated by TrxG and PcG are established, they must be faithfully maintained

215

and replicated during division in every cell cycle in a process referred to as transgenerational epigenetic inheritance. The existence of this process is exemplified by the observation that an epigenetically silenced transgene can maintain its repressed state in cells propagated in vitro for more than a year [119]. The primary impediments against epigenetic inheritance during DNA replication and mitosis are the potential disruption of DNA–protein interactions during the passage of the replication machinery along the genome and the possible replacement of parental histones with newly synthesized, unmarked histone proteins. Drosophila provides an excellent model to study strategies that have evolved to ensure that epigenetic landscapes are successfully inherited.

Epigenetics During S Phase During the S phase, the parental histone octamers dissociate into H3–H4 tetramers and H2A–H2B dimers and are then, along with newly synthesized histones, randomly reassembled on the daughter DNA duplexes [120]. To maintain previously established gene expression patterns, the specific combinations of histone modifications must be reinstalled after replication. It was originally thought that all regulatory protein complexes, including histones and their modifications, were completely removed from parental DNA to allow the passage of the replication machinery and thus required de novo reestablishment after replication. These findings begged the question that if all of the marks are removed, how is the parental pattern of histone modifications reinstated in the absence of any previously established guiding landmarks? An insight into this question was provided by chromatin immunoprecipitation assays in S2 Drosophila cell cultures that revealed a dramatic increase in PRC2 binding to PREs in early S phase with a subsequent drop in binding in late S phase (Fig. 14.5A). These experiments suggested that faithful transmission of parental marks relies upon the excessive deposition of chromatin marks prior to S phase completion rather than de novo deposition mediated by the efforts of histone-modifying enzymes after replication [121]. However, a more recent alternative model suggests that histone-modifying enzymes may actually remain closely associated with the DNA during replication and thus are able to immediately direct the reestablishment of epigenetic marks onto the newly synthesized and assembled histone proteins [122]. In support of this model, Trx, Pc, and EZ are found within close proximity to the replication fork and all are bound to their respective PREs/TREs on the nascent DNA near the replication fork (Fig. 14.5B) [123]. A similar mechanism has been proposed for chromatin-remodeling proteins that are capable of directly binding to members of the replication machinery.

IV.  Model Organisms of Epigenetics

216

14.  DROSOPHILA EPIGENETICS

FIGURE 14.5  Replication and maintenance of silent and active states in chromatin. “Generic” active and silent chromatin is depicted, although specific modifications are differently distributed at Pc/Trx response elements (PREs/TREs), promoters, and gene bodies. PREs are thought to be depleted of nucleosomes, and chromatin modifications occur on the nucleosomes flanking them. Parental chromatin, newly replicated chromatin, and mature chromatin are shown. Source: Adapted from Steffen PA, Ringrose L. What are memories made of? How Polycomb and Trithorax proteins mediate epigenetic memory. Nat Rev Mol Cell Biol 2014;15(5):340–56 [97].

While Drosophila must ensure the proper transmission of proteins and histone modifications that activate euchromatic gene expression, it is equally important to direct the proper formation of heterochromatic locations within the daughter chromatin for sustained gene repression. HP1 directs the stable transmission of both itself and repressive H3K9 methylation marks throughout the cell cycle, thus facilitating the inheritance of the parental pattern of epigenetic marks and of structural proteins associated with heterochromatin. During nucleosome reorganization after replication fork passage, the Drosophila chaperone protein, CAF-1, guides newly synthesized histone proteins to

the replicated DNA, where they join inherited parental histones to form the daughter nucleosomes. HP1 forms a complex with CAF-1, functioning to guide HP1 to the site of replication where it binds inherited parental histones displaying the repressive H3K9me3 marks via its chromodomain [105]. In a process similar to heterochromatin spreading during PEV, HP1 binding then mediates the recruitment of the histone methyltransferase Su(var)3-9 catalyzing the deposition of additional H3K9me3 marks on to neighboring histones and facilitating the spreading of H3K9me3-rich heterochromatic domains onto the appropriate regions of daughter DNA [105,124].

IV.  Model Organisms of Epigenetics



Dosage Compensation

Epigenetics During M Phase After replication, mitosis requires the coordinated condensation of chromosomes that, once again, disrupts protein–chromatin interactions and higher-order chromatin structures [105]. Therefore, any proteins directly mediating the formation of compact or open daughter chromatin states, such as HP1 within heterochromatic regions or PRC1 dimers, must reassociate following cytokinesis. For HP1, its dynamic interaction with chromatin during M phase is dictated by an epigenetic switch referred to as the “methyl/phos switch” [125]. In Drosophila, a conserved family of protein kinases, called the Aurora-like kinases, directs proper chromosome segregation [126]. At the onset of mitosis, Aurora B catalyzes the phosphorylation of H3S10, the residue that neighbors H3K9. Intriguingly, H3S10ph deposition leads to the displacement of HP1 from chromatin, while conversely, depletion of Aurora B results in the aberrant association of HP1 with chromatin during mitosis [127]. Completion of cytokinesis results in the removal of H3S10ph by the protein phosphatase PP1, allowing HP1 to rebind to the H3K9 methyl marks [16,105]. Despite the removal of HP1, the H3K9 methylation mark previously bound by HP1 remains intact throughout mitosis, not only maintaining the epigenetic landscape, but also providing a scaffold for HP1 reassociation after cytokinesis; this allows it to resume its role as a heterochromatic marker in the newly divided cell. Similarly, H3S10ph mediates the removal of HP1 to allow proper M phase progression, while simultaneously serving as an epigenetic placeholder for HP1. Conversely, unlike the repressor HP1, some transcriptional activator proteins, such as ASH1 and Trx, have been shown to remain bound to the chromatin during mitosis to faithfully transmit the pattern of gene expression in newly divided cells by both maintaining the landscape of activating H3K4me3 marks and by antagonizing the deposition of PcG-mediated silencing marks within the next generation (Fig. 14.5B). In summary, Drosophila has served as an important model system to elucidate how epigenetic gene control mechanisms are faithfully inherited so that cell fate and more dynamic regulatory processes, such as signal transduction can be effectively maintained after cell division.

DOSAGE COMPENSATION In many organisms, evolution has resulted in the unequal distribution of the X and Y sex chromosomes with females having two X chromosomes and males having one X and one Y chromosome. This genetic aneuploidy requires a means of compensating for the resulting difference in X-linked gene representation between the sexes. Different organisms have developed different

217

mechanisms to adjust for this imbalance in processes termed dosage compensation. For example, many mammals including Mus musculus and humans randomly inactivate one of the two female X chromosomes [128]. In Drosophila, dosage compensation is achieved through the coordinated, multistep efforts of protein complexes and noncoding RNAs culminating in the epigenetically mediated enhanced transcription of active genes within the entire X chromosome in males (reviewed in Refs. [129,130]).

The MSL Complex/DCC The phenomenon of dosage compensation in Drosophila was first uncovered in the seminal finding by Mukherjee and Berman. They discovered that the level of [3H] uridine incorporation, indicative of RNA synthesis, by the male X chromosome was equivalent to that of the two female X chromosomes [129]. This was followed by the finding that LOF of four specific genes resulted in the equalization of [3H]uridine incorporation between the sexes and in a dramatic reduction in male viability without altering female phenotypes [131,132]. These genes, male-specific lethal (msl)1, msl2, msl3, and mle (maleless) have since been shown to be a part of the protein/RNA multisubunit complex responsible for dosage compensation in Drosophila called the MSL complex, also known as the dosage compensation complex (DCC) [4,130,133]. The DCC is made up of a multiprotein core, including MSL1, MSL2, MSL3, MLE, and males absent of first (MOF), along with one of two male-specific noncoding RNAs: roX1 and roX2 (RNA of the X) (Fig. 14.6) [130]. MSL2 has been implicated as the director of DCC assembly, initiating formation by binding to and stabilizing MSL1, possibly via the interaction between a RING finger domain of MSL2 and a coiled-coil domain within MSL1 [134,135]. This MSL2/MSL1 core regulates the chromatin-binding activity of the DCC and provides a platform for complex nucleation [135,136]. Mof encodes a member of the MYST family of HATs shown to direct the acetylation of H4K16, and plays an important role in DCC-mediated transcriptional enhancement [129]. The MLE protein has inherent ATPase helicase and single-stranded RNA- and DNA-binding abilities [137], and has been proposed to have a dual role within the DCC and in dosage compensation. Its helicase activity is important in DCC-mediated enhancement of transcription and in incorporation of the X chromosome– encoded roX RNAs into the protein complex [138]. The two noncoding RNAs are functionally redundant within the DCC, with single mutant males showing no phenotype [139]. However, their incorporation is critical for proper DCC targeting, as male flies containing mutant roX RNAs display a decrease in the localization of DCC proteins along the X chromosome [139].

IV.  Model Organisms of Epigenetics

218

14.  DROSOPHILA EPIGENETICS

FIGURE 14.6  A hierarchy of binding sites for the dosage compensation complex (DCC). (A) The DCC is made up of a multiprotein core, including MSL1, MSL2, MSL3, MLE, and MOF, along with one of two male-specific noncoding RNAs: roX1 and roX2 (RNA of the X) [130]. (B) Image depicts polytene chromosomes stained for MSL1 in red and DNA in blue. In the wildtype male, the DCC binds to hundreds of sites on the X chromosome. Source: Adapted from Gilfillan GD, Dahlsveen IK, Becker PB. Lifting a chromosome: dosage compensation in Drosophila melanogaster. FEBS Lett 2004;567(1):8–14 [144].

Although the two nocoding RNAs are different in size and in the majority of their sequence, they do share a sequence called the roX box, comprised of highly structured tandem stem loops [140], which enhance MOF-mediated histone acetylation in vivo [141]. While conflicting evidence exists regarding the specific DCC proteins responsible for roX RNA incorporation into and tethering to the DCC, it has been proposed that the inherent helicase activity of MLE works to fold the RNA into a 3D conformation, conducive to incorporation into the complex, while MOF and MLE may be responsible for tethering the folded RNA peripherally to the DCC [138,140,142,143].

Regulation of DCC Targeting to the X Chromosome Proper upregulation of active male X-linked genes is dependent upon the appropriate assembly of the DCC along the male X chromosome, while avoiding the female X chromosomes or the autosomes. The sexspecific chromosomal targeting of the DCC is crucial for viability [129] and thus, Drosophila has evolved a mechanism to ensure dosage compensation is present in males but absent in females. Sex lethal (SXL) is an RNA-binding protein whose fully functional form is only expressed in females. SXL recognizes and binds to polypyrimidine tracts within the 5′ and 3′ UTRs of the msl2 mRNA [145], where it recruits upstream N-Ras (UNR). With both bound to the msl2 transcript, they then act cooperatively to inhibit MSL2 translation in females [129,146,147]. Conversely, in Drosophila males, alternative splicing of the sxl transcript results in the expression of a nonfunctional SXL protein [148], thereby enabling successful MSL2 protein synthesis and DCC formation in males. Additionally, the assembly of the DCC complex exclusively on the X chromosome is a result of the recognition of specific sites within the X-encoded roX genes by members of the DCC along with the cotranscriptional incorporation of roX RNA into the DCC complex. Studies show that MSL2 is responsible for roX1 and roX2 expression [149,150], suggesting that this interaction may serve as

a nucleation site for X-linked DCC complex formation. Similarly, it has been demonstrated that the incorporation of roX RNAs into partial DCC complexes occurs at the site of roX RNA synthesis [138]. These findings support a model by which the role of X-encoded roX genes is to serve as the nucleation site for DCC assembly, mediating complex formation specifically on the X chromosome. Intriguingly, approximately 35 chromatin sites along the X chromosome have been shown to bind the MSL1/ MSL2 core with only two of those sites corresponding to the roX1 and roX2 genes [151], indicating that additional targeting sequences exist. Chromatin immunoprecipitation performed in conjunction with ChIP-seq identified 150 DCC targeting sequences dubbed chromatin entry sites (CES) or high-affinity sites (HAS) (Fig. 14.7), with further analysis characterizing a GA-repeat motif present in 91% of identified HASs and enriched on the X chromosome compared to autosomes [152,153]. This

FIGURE 14.7  Model of dosage compensation. The two-step model of MSL targeting to the X chromosome. (A) The MSL complex (red circles) targets a specific high-affinity site (HAS) on the X chromosome in a sequence-dependent manner. HASs represent a subset of the MSL-bound regions in wildtype Drosophila that also recruit the MSL complex under more stringent conditions (such as when inserted into an autosome or when integral subunits of the MSL complex are missing). (B) After initial targeting, the MSL complex spreads along in a cis-manner from the entry sites (shown by blue arrows), and predominantly binds to the 3′ end of actively transcribed genes. MSL binding causes acetylation at histone H4 and results in a global change of the chromatin structure, facilitating a twofold transcriptional upregulation of X-linked genes in males (green lines). Source: Adapted from Bachtrog D, Toda NR, Lockton S. Dosage compensation and demasculinization of X chromosomes in Drosophila. Curr Biol 2010;20(16):1476–81 [156].

IV.  Model Organisms of Epigenetics



Dosage Compensation

latter motif, named the MSL recognition element (MRE), was found to be, in addition to the roX DHS sequences, sufficient for DCC recognition and reporter gene expression [152]. Newly emerging and compelling evidence points to the zinc finger protein chromatin-linked adaptor for MSL proteins (CLAMP) as the main regulator of the chromosome-specific targeting of the DCC. The CLAMP protein is capable of both binding MREs and recruiting the DCC to these recognition elements located on the X chromosome, providing a direct link between DCC and the X chromosome [154,155]. Therefore, taken together, these findings suggest that sequences within roX genes, in concert with a defined distribution of HASs recognized by the CLAMP protein, likely play key roles as regulators of the recognition and localization of DCCs along the length of the X chromosome. While dosage compensation via assembly of the DCC on the X chromosome in males is critical, it is equally important that the autosomes are similarly untouched by this mechanism. Insights into how this chromosomespecific regulation occurs came with the discovery that the mof gene resides on the X chromosome and that the recruitment of the DCC complex proteins to the X chromosome is dependent upon its transcriptionally active state [157]. In addition, an alternative mechanism by which X-linked MOF expression aids in the targeting of DCC away from autosomes has been proposed. In this model, termed the inverse dosage model, expression and subsequent incorporation of MOF into the DCC, along with its association with the X chromosome, results in the sequestration of the MOF protein away from autosomes [157,158]. This alternative model suggests that the purpose of the DCC is not only to increase the expression of active X-linked genes, but also to prevent any H4K16acmediated transcriptional enhancement of autosomal genes by DCC-independent MOF activity [143].

DCC Spreading and Target Recognition While the roX gene and HAS sequences are critical for DCC complex formation and localization on the X chromosome, they are not universally found among the target active genes, implicating the existence of another set of targeting sequences that mediate DCC recognition of gene targets. It has been proposed that the DCC, after recognizing and binding to specific HASs on the X chromosome, initiates the epigenetic-based remodeling of the local chromatin, which would allow the diffusion or spreading of DCCs along the X chromosome to lower-affinity binding sites that correspond to the bodies of active genes (Fig. 14.7). Whether DCC spreading is a linear process involving the continuous DCC–chromatin association as it scans along the chromatin or one involving a bind-and-release mechanism of chromatin sampling remains unclear. However, this

219

process requires the action of MLE’s ATP-dependent helicase activity and/or MOF’s HAT activity, indicating that local chromatin remodeling is likely the underlying mechanism for DCC spreading [138,159–162]. It has been proposed that DNA duplex unwinding by MLE and chromatin conformation relaxation mediated by MOF-directed histone acetylation potentially results in both the increased ability of the DCC to travel unfettered along the chromatin and probably in exposing nearby low-affinity sites for DCC recognition and binding. Alternatively, it has been suggested that DCCs spreading away from HASs is a direct result of elevated DCC concentrations around the entry sites [143]. In this model, the HASs would serve as DCC “hubs,” constantly recruiting complexes to their location after the previous complex leaves, thereby driving the increase in the proximal concentrations of DCC.

Transcriptional Activation of Active X-Linked Genes by the DCC Although enhancement of transcription is the basis for dosage compensation, the exact mechanism by which this is achieved remains unclear. Two main theories, both involving epigenetic modifications, have been proposed and it is not unreasonable to believe that both can play a role in the targeted upregulation of transcription during dosage compensation. In the first model, epigeneticbased chromatin remodeling results in the unwinding of chromatin, providing the transcription machinery with a more open conformation and facilitating elongation and RNA Pol II progression through the target gene region. This hypothesis is based on the findings that high levels of H4K16ac marks are found mostly within the body and especially within the 3′ end of the active transcriptional units [163,164]. In fact, global run-on sequencing (GROseq) demonstrated more efficient RNA Pol II movement through X-linked gene transcribed units than through autosomal genes, and a dependence of efficiency on the presence of H2K16ac [165,166]. These findings support a model by which H4K16ac-dependent chromatin relaxation leads to transcription upregulation via facilitation of elongation by possibly presenting Pol II with fewer obstacles during elongation, providing a more favorable platform for gene transcription progression. In the second model, histone acetylation may actively counteract factors responsible for condensation of chromatin, which would normally disrupt DCC function. H4K16ac has been shown in vivo to antagonize the ISWI family of ATP-dependent chromatin-remodeling agents involved in the maintenance of higher-order chromatin structure [167] by establishing a chromatin structure inaccessible to ISWI proteins at gene promoters, thereby regulating its influence on the degree of chromatin compaction and gene expression during dosage compensation.

IV.  Model Organisms of Epigenetics

220

14.  DROSOPHILA EPIGENETICS

In summary, insights into the epigenetic mechanisms that underlie multiple facets of dosage compensation, including DCC formation, targeting, binding, and spreading, culminating in X chromosome–specific gene transcriptional activation and survival in male flies has been elucidated using Drosophila as an effective model system.

THE EPIGENETIC LANGUAGE IN POSTMITOTIC NEURONS UNDERLYING COGNITIVE FUNCTION Postmitotic neurons within the brain have the remarkable ability to shape and refine their synaptic connections in response to external stimuli. This adaptability relies on the precise regulation of specific gene programs and is the foundation of transient and sustained synaptic plasticity. Epigenetic histone modifications have emerged as the primary mechanism by which neuroadaptation is transcriptionally achieved in response to external cues. As the addition and removal of histone acetyl marks can be an extremely rapid process, it is well suited as the dynamic epigenetic link between the constantly changing environment and the molecular responses necessary for experience-driven behavioral changes. Furthermore, the balance between HAT and HDAC functions is important for establishing and maintaining the epigenetic landscape required for directing important neuronal functions, such as learning and memory. The disruption of this delicate balance results in the pathogenesis of neurodegenerative disorders. Drosophila has proven to be a powerful model system for elucidating epigenetic gene control mechanisms underlying both learning and memory, as well as cognitive deficit-associated neurological disorders [168–170].

Activity-Dependent Histone Acetylation and Cognitive Function Long-term synaptic plasticity is defined as the change in the synaptic strength in response to repeated patterns of synaptic transmission and is the underlying mechanism behind long-term memory (LTM) storage [171]. The initiation of these changes in synaptic strength occurs at the level of the individual synapse, where influences from the extracellular environment triggers downstream signaling pathways that eventually terminate in epigenetic-mediated expression alterations. Initiating this process, neurotransmitters, growth factors, or cytokines are released from surrounding neurons and facilitate the influx of calcium into the postsynaptic neuron through binding to their cognate receptors. Intracellular calcium then serves as a powerful secondary messenger, activating a multitude of downstream signaling cascades, some

of which have been shown to modulate the activity of specific HATs and HDACs, which can influence learning and memory via their specific enzymatic activities. The downstream events initiated by the transcription factor cAMP response element–binding protein (CREB) are a paradigmatic example of such a signaling cascade. CREB, activated by calcium-mediated phosphorylation, recruits the HAT CBP that in turn, epigenetically modifies the chromatin structure and induces the expression of specific genes [66,172,173]. In support of this, expression of an inhibitory isoform of CREB represses LTM in Drosophila [174,175], while expression of a constitutively active isoform enhances LTM [174,176]. Following disruption of CREB’s CBP-binding domain, not only is the expression of CREB target genes reduced, but LTM storage is also similarly impaired [177]. Furthermore, targeted knockdown of CBP HAT activity within the Drosophila MB results in the attenuation of LTM formation [175]. Activity-dependent transcriptional activation has also been suggested to be a consequence of HDAC export out of the nucleus, permitting HAT and transcription factor–mediated gene activation [178]. Taken together, these findings indicate that the activation of intracellular signaling pathways leads to enzymatic alterations in the epigenetic landscape, which mediate the transcriptional changes necessary for neuronal functional flexibility and cognitive function.

HAT: HDAC Interplay in Cognitive Function Drosophila is an attractive model for studies focused on epigenetic dissection of components of memory formation due to the availability of reproducible memory assays and genetic tools that enable restricting gene expression manipulation to specific subregions of the brain [179–181]. In particular, the Drosophila MB is known to regulate a variety of behavioral and physiological functions that range from learning and memory and courtship conditioning to decision-making under uncertain conditions [23–27]. While the cellular and molecular mechanisms involved in memory formation and maintenance have long been a topic of research, new evidence indicates that epigenetic regulation plays a key role. Specifically, the regulation of global histone acetylation by HATs and HDACs has been shown to be of critical importance in the epigenetic-based regulation of cognitive function. In a series of experiments within the MB of the fly, appropriate levels of Tip60 HAT activity was shown to be crucial for immediate recall memory [182] (Fig. 14.8). Conversely, the reduction in acetylation levels induced by the targeted overexpression of the HDAC Rpd3 and HDAC4 within the adult Drosophila MB resulted in severely impaired long-term courtship memory [112,183]. These studies support the premise that under normal conditions, maintaining the balance between HAT and HDAC levels and activities

IV.  Model Organisms of Epigenetics



The Epigenetic Language in Postmitotic Neurons Underlying Cognitive Function

FIGURE 14.8  Drosophila mushroom body (MB) neurons. Shown, are MB neurons coimmunostained with antibodies that label Tip60 shown in green and Fasciclin II (Fas II) shown in red. Fas II is a cell adhesion molecule that is expressed strongly in the MB α/β lobes and weakly in the γ lobe in the MB of the adult fly brain. Appropriate levels of Tip60 are required for axon outgrowth in the adult Drosophila MB.

is crucial for establishing appropriate histone modification patterns that serve to control both stable and rapidly changing gene expression profiles critical for both neuronal homeostasis and appropriate neurophysiological response outputs, such as long-term potentiation, and learning and memory, respectively [184]. Appropriate HAT levels within the CNS have been shown to be required for the proper development and function of neural morphology and functional processes implicated in cognition. Experiments in Drosophila have shown that reducing Tip60 HAT levels within the MB causes significantly shorter MB lobes in the adult fly brain, indicative of defects in axonal outgrowth. Further, loss of Tip60 HAT activity in motor neurons causes axonal transport stalling [182] and loss of Tip60 in the sLNv neurons that regulate sleep causes repressed axonal outgrowth. Functional consequences of these Tip60-mediated defects include impairment in short-term memory, sleep/wake cycles, and locomotor activity [168,185,186]. Additionally, Tip60 HAT activity has been shown to regulate proper synaptic bouton expansion at the Drosophila neuromuscular junction through modulation of the synaptic microtubule cytoskeleton [72]. Intriguingly, the activity-dependent remodeling events that shape the formation of neural circuits during development mirror the events underlying changes in synaptic connectivity required for learning and memory formation. Experience-dependent neuroadaptation requires increases in synaptic bouton densities and dendrite extensions that are similar to those essential during development [187]. Therefore, it has been hypothesized that the signaling cascades that result in experience-driven changes in cognitive function lead to the modulation of epigeneticmediated transcriptional control of genes that promote

221

synaptic growth and neuronal homeostasis. Although the identity of the genes responsible for the morphological and molecular changes underlying learning and memory have not been completely characterized, advances in high-throughput sequencing technology have begun to tease apart the epigenetically regulated genes involved in synaptic plasticity. For example, microarray analysis of Tip60 HAT-mutant flies revealed that Tip60 regulates the expression of genes whose functions include neuronal development and synaptic function [188]. Moreover, ChIP-Seq analysis in Drosophila reveals that Tip60 target genes are enriched for functions in cognitive processes and accordingly, key genes representing these pathways are misregulated in the Tip60 HAT mutant fly brain [168]. Together, these studies using the Drosophila model highlight the importance of specific HATs, such as Tip60, in the epigenetic regulation of cognitive function.

Deregulation of HAT Activity in the Etiology of Alzheimer’s Disease Precise regulation of epigenetic gene expression is critical for the maintenance of long-term neuronal survival and cognitive function. Consistent with this concept, skewing the HAT/HDAC balance leads to detrimental changes in the expression profiles of critical genes and subsequent cognitive decline and memory impairments, characteristic of neurodegenerative diseases, such as AD, a chronic disease affecting approximately 48 million people worldwide [71]. The selective atrophy of the hippocampus and cerebral cortex and the progressive cognitive deterioration characteristic of AD are causatively linked to aberrant intraneuronal aggregations of both Aβ plaques and Tau (neurofibrillary tangles) proteins [189]. Several powerful models of human AD have been developed using Drosophila as a model system, providing scientists with the ability to study the epigenetic mechanisms behind this neurodegenerative disease from early presymptomatic to late terminal stages [8]. These models utilize the overexpression of β-amyloid, amyloid precursor protein (APP), presinilin, and human tau proteins resulting in phenotypes ranging from physical abnormalities, such as eye degeneration and locomotor deficits, to molecular defects, such as accumulations of amyloid plaques, neurofibrillary tangles, and NMJ and axonal degeneration [8]. Accumulating evidence garnered from such AD Drosophila models suggests that the clinical symptoms associated with AD are the result of synaptic degeneration mediated by the impairment of epigenetic mechanisms of gene control [107,190]. For example, AD model flies that display deficits in both learning and short-term memory harbor a significant reduction in histone acetylation levels, loss of cognition-associated gene expression, and an increase in proapoptotic gene expression that induces neurodegeneration [169,185,191].

IV.  Model Organisms of Epigenetics

222

14.  DROSOPHILA EPIGENETICS

FIGURE 14.9  Modulation of specific HAT function like Tip60 displays neuroprotective effects under neurodegenerative conditions. Neurodegenerative diseases are characterized by impaired acetylation homeostasis that consequently leads to altered neuronal transcription profiles, resulting in attenuated expression of survival-associated genes, while simultaneously accentuating more proapoptotic genes. Conversely, modulation of cellular levels and/or enzymatic activity of specific HATs may enhance the expression of cassettes of specific genes that have neuroprotective effects as evidenced in the case of Tip60. Under amyloid precursor protein (APP)–induced neurodegenerative conditions, HAT competent Tip60 (Tip60WT), but not its HAT defective counterpart (Tip60mut), exerts neuroprotective defects by epigenetically regulating gene expression profiles that tip cell fate control balance toward cell survival. Thus, targeting specific HATs for therapeutic intervention may offer more promising alternatives for neurodegenerative diseases than currently available HDAC inhibitors. Source: Adapted from Pirooznia SK, Elefant F. Targeting specific HATs for neurodegenerative disease treatment: translating basic biology to therapeutic possibilities. Front Cell Neurosci 2013;7:30 [170].

Importantly, AD research specifically studying the HAT Tip60 has provided a wealth of intriguing and exciting information regarding the molecular mechanisms underlying AD pathogenesis, as well as potential therapeutic strategies for the future. These studies show that increasing Tip60 HAT levels in the Drosophila nervous system under APP-induced neurodegenerative conditions rescued AD-associated neuronal impairments, such as apoptotic neurodegeneration in the CNS [191], axonal outgrowth [186,192], and synaptic vesicle transport in motor neurons [185]. Excess Tip60 also restored associated disrupted complex functional abilities impaired in AD that include sleep cycles [186,192], locomotor function [185], and learning and memory [193] defects with concomitant induction of some genes critical for the function of these neural processes [185,191]. In direct contrast, loss of Tip60 HAT function in the fly nervous system causes gene misregulation and exacerbates such AD-associated impaired phenotypes [185,186,191–193]. Similarly, studies have shown that increasing CBP HAT levels is capable of rescuing an Aβ-mediated neurodegenerative phenotype modeled in the Drosophila eye [194]. Together, these findings demonstrate that specific HATs, such as Tip60, play a neuroprotective role in an array of cognition-associated neuronal processes that are impaired in AD, and exemplify Drosophila as an effective tool to dissect epigenetic mechanisms underlying neurodegenerative disorders, such as AD (Fig. 14.9).

While these studies suggest that histone acetylation upregulation holds a great potential in restoring AD-mediated disruptions in higher-order cognitive functions, the importance of a balance between HATs and HDACs has also been shown. Highlighting this concept, experimental manipulations of the HDAC Rpd3 within the MB of Drosophila demonstrated that both overexpression, as well as targeted knockdown of Rpd3 via RNAi, resulted in LTM impairment. In a separate study in Drosophila, overexpression and reduction of HDAC4 resulted in impairment of LTM [112]. Further, neuronal overexpression of Tip60 leads to increased apoptosis and lethality [191], while both targeted loss and gain of Tip60 levels within the MB resulted in the impairment of immediate recall memory [182]. Therefore, while it is believed that manipulation of HAT activators and HDAC inhibitors present promising therapeutic potential, the importance of their precise spatial and temporal regulatory balance remain to be delineated before their full potential can be determined.

Environmental Enrichment Improves Learning and Memory in Drosophila The ability of the neuronal circuitry to morphologically adapt to changing environmental conditions is the basis for experience-driven higher-order behavioral

IV.  Model Organisms of Epigenetics



223

CONCLUSIONS

functions, such as learning and memory. Enhanced sensory and motor stimulation through interactions with an “enriched” environment that is comprised of positive social reinforcements has been shown to lead to alterations in gene transcription profiles, modulation of the morphology of synaptic connection, and even facilitation of learning and memory, in a variety of organisms (reviewed in Ref. [195]). Environmental enrichment (EE) conditions have also been shown to have neuroprotective benefits under neuropathological conditions, such as AD [196–199]. These observations have allowed scientists to gain insights into the interactions between genes and the environment and the impact of EE on neuroadaptation. While experimental EE conditions may vary between studies exploring EE neuroadaptative benefits, one critical and nonvariable EE component widely conserved among species is social EE [200,201]. Well-established studies using Drosophila show that similar to mammals, social EE promotes significant beneficial structural changes in regions throughout the fly brain that include the MB that regulates a variety of behavioral and physiological functions ranging from olfactory learning and memory to decision-making under uncertain conditions [23,202–205]. Social EE promotes enhanced MB axon and dendrite formation, synaptic plasticity, and neuronal MB Kenyon cell growth [200,206]. Recent studies demonstrate that EE benefits require epigenetic gene regulation involving induction of specific histone acetylation profiles [197,207–209]. Nevertheless, how specific HATs mediate cognitive gene expression programs in response to changing environmental cues and the select HATs involved in this process remain largely unknown. Recent studies have exploited the power of Drosophila genetics and the behavioral and physiological conservation between flies and mammals in terms of their positive neuroadaptive response to investigate whether Tip60 HAT action is required for an EEinduced beneficial neuroadaptative response. These studies revealed that flies raised under EE conditions displayed enhanced MB axonal outgrowth, synaptic marker protein production, histone acetylation induction, and transcriptional activation of cognition-linked genes when compared to their genotypically identical siblings raised under isolated conditions. Further, these beneficial changes were impaired in both Tip60 HAT mutant flies and APP neurodegenerative flies. While EE conditions provided some beneficial neuroadaptive changes in the APP neurodegenerative fly MB, increasing MB Tip60 HAT levels significantly enhanced such positive changes. Together, these results implicated Tip60 as a critical mediator of EE-induced benefits, and provide broad insights into synergistic behavioral– and epigenetic-based therapeutic approaches for the treatment of cognitive impairment.

CONCLUSIONS In conclusion, D. melanogaster has proven to be a powerful model system for elucidating the fundamental roles epigenetic control mechanisms play in various conserved biological processes. In this chapter, we have summarized studies utilizing Drosophila that have proven invaluable for shedding light into fundamental concepts regarding the function of the generation and maintenance of specific epigenetic patterns in chromatin that govern processes unique to the fly, such as X-linked dosage compensation, as well as in processes conserved in humans, such as epigenetic mechanisms underlying gene control, PEV, developmental programming, epigenetic inheritance, and neuronal plasticity guiding learning and memory. Examples are also given pertaining to how Drosophila represents a uniquely powerful tool for analyzing the detrimental consequences of aberrations in epigenetic patterns in neurological disorders and thus represents a valuable resource for the development of potential epigenetic-based therapeutic interventions for cognitive deficits that that are a hallmark of neurodegenerative disorders. With regards to future directions in the field of Drosophila epigenetics, while it is clear that histone acetylation is a key mechanism through which the constantly changing environment can influence cognitive functions, exciting new research in Drosophila is aimed at teasing apart precisely how histone acetylation mediates EE-induced neuronal survival, and the mechanisms through which these changes can enhance learning and memory [210]. Conversely, understanding how disruption of the normally tightly regulated mechanisms of acetylation can result in abnormal changes in neuronal morphology and circuitry and lead to learning and memory impairment is key to understanding neurodegenerative diseases [168,182,210]. Recognition of the pivotal importance of histone acetylation has led to the hypothesis that HAT-based epigenetic therapeutic strategies can provide neuroprotective benefits in AD, a concept currently being developed in Drosophila and being ascertained in ongoing research for their applicability in mammalian systems.

Glossary Dosage compensation A term describing the mechanism of X-linked gene expression equalization between the sexes. Epigenetics The study of phenotypic variations arising from the heritable changes in gene expression that result from alterations in the genome that do not involve changes in the nucleotide sequence. Histone acetylation The transfer of an acyl moiety from acyl coenzyme A to lysine residues on the N-terminal tails of the histones by HAT enzymes. Epigenetic memory The maintenance of epigenetically mediated gene transcriptional states across cell generations.

IV.  Model Organisms of Epigenetics

224

14.  DROSOPHILA EPIGENETICS

Histone methylation The transfer of methyl moieties from S-adenosyl methionine to the ε-amino groups of lysine residues of histone proteins by KMT enzymes. Position effect variegation The process by which genes that are normally expressed are aberrantly moved, either by rearrangement or transposition, to a site near the Eu–Het transition zone, and as a result are abnormally packaged into a heterochromatic form and silenced. Transgenerational epigenetic inheritance The faithful maintenance and replication of epigenetic patterns through the cell cycle.

Abbreviations Aβ Amyloid beta AD Alzheimer’s disease APP Amyloid precursor protein Ash1 Absent, small, or homeotic discs 1 Brm Brahma CBP CREB-binding protein CES Chromatin entry site CLAMP Chromatin-linked adaptor for MSL protein CNS Central nervous system CREB cAMP response element–binding protein DCC Dosage compensation complex E(var) Enhancer of variegation EE Environmental enrichment EM Electron microscopy Esc Extra sex combs Eu–Het Euchromatin–heterochromatin EZ Enhancer of zeste Fsh Female sterile homeotic HAS High-affinity site HAT Histone acetyltransferase HD Huntington’s disease HDAC Histone deacetylase HOX genes Homeotic genes HP1 Heterochromatin protein 1 KMT Lysine methyltransferase lncRNA Long noncoding RNA LOF Loss of function LTM Long-term memory MB Mushroom body MLE Maleless MOF Males absent of first MRE MSL recognition element MSL Male-specific lethal Pc Polycomb PcG Polycomb group PEV Position effect variegation PhoRC Pho repressive complex piRNA Piwi-interacting RNA PRC1 Polycomb repressive complex 1 PRC2 Polycomb repressive complex 2 PRE Polycomb group response element PTM Posttranslational modification Sce Sex combs extra Su(var) Suppressor of variegation SXL Sex lethal TAFII TBP-associated factor TE Transposable element Tip60 Tat-interactive protein 60 kDa TRE Trithorax group response element Trx Trithorax TrxG Trithorax group UNR Upstream N-Ras

References [1] Noble D. Conrad Waddington and the origin of epigenetics. J Exp Biol 2015;218:816–8. [2] Rusche LN, Lynch, Partick J. Assembling heterochromatin in the appropriate places: a boost is needed. J Cell Physiol 2015;219(3):525–8. [3] Elgin SC, Reuter G. Position-effect variegation, heterochromatin formation, and gene silencing in Drosophila. Cold Spring Harb Perspect Biol 2013;5(8):a017780. [4] Rubin GM, Yandell MD, Wortman JR, Gabor Miklos GL, Nelson CR, Hariharan IK, et al. Comparative genomics of the eukaryotes. Science 2000;287(5461):2204–15. [5] Stapleton M, Liao G, Brokstein P, Hong L, Carninci P, Shiraki T, et al. The Drosophila gene collection: identification of putative full-length cDNAs for 70% of D. melanogaster genes. Genome Res 2002;12(8):1294–300. [6] Kim MS, Pinto SM, Getnet D, Nirujogi RS, Manda SS, Chaerkady R, et al. A draft map of the human proteome. Nature 2014;509(7502):575–81. [7] Ezkurdia I, Juan D, Rodriguez JM, Frankish A, Diekhans M, Harrow J, et al. Multiple evidence strands suggest that there may be as few as 19,000 human protein-coding genes. Human Mol Genet 2014;23(22):5866–78. [8] Pandey UB, Nichols CD. Human disease models in Drosophila melanogaster and the role of the fly in therapeutic drug discovery. Pharmacol Rev 2011;63(2):411–36. [9] Wangler MF, Yamamoto S, Bellen HJ. Fruit flies in biomedical research. Genetics 2015;199(3):639–53. [10] Min K. Drosophila as a model to study human brain degenerative diseases. Parkinsonism Relat Disord 2001;7(3):165–9. [11] Zhimulev IF, Boldyreva LV, Demakova OV, Poholkova GV, Khoroshko VA, Zykova TY, et al. Drosophila polytene chromosome bands formed by gene introns. Dokl Biochem Biophys 2016;466(1):57–60. [12] Zhimulev IF, Belyaeva ES, Vatolina TY, Demakov SA. Banding patterns in Drosophila melanogaster polytene chromosomes correlate with DNA-binding protein occupancy. Bioessays 2012;34(6):498–508. [13] Zhimulev IF, Zykova TY, Goncharov FP, Khoroshko VA, Demakova OV, Semeshin VF, et al. Genetic organization of interphase chromosome bands and interbands in Drosophila melanogaster. PLoS One 2014;9(7):e101631. [14] Eagen KP, Hartl TA, Kornberg RD. Stable chromosome condensation revealed by chromosome conformation capture. Cell 2015;163(4):934–46. [15] Mukherjee K, Twyman RM, Vilcinskas A. Insects as models to study the epigenetic basis of disease. Prog Biophys Mol Biol 2015;118(1–2):69–78. [16] Boros IM. Histone modification in Drosophila. Brief Funct Genomics 2012;11(4):319–31. [17] Budnik V. Synapse maturation and structural plasticity at Drosophila neuromuscular junctions. Curr Opin Neurobiol 1996;6(6):858–67. [18] Menon KP, Andrews S, Murthy M, Gavis ER, Zinn K. The translational repressors Nanos and Pumilio have divergent effects on presynaptic terminal growth and postsynaptic glutamate receptor subunit composition. J Neurosci 2009;29(17):5558–72. [19] Sarkar A, Irwin M, Singh A, Riccetti M, Singh A. Alzheimer’s disease: the silver tsunami of the 21(st) century. Neural Regen Res 2016;11(5):693–7. [20] Burr AA, Tsou WL, Ristic G, Todi SV. Using membrane-targeted green fluorescent protein to monitor neurotoxic proteindependent degeneration of Drosophila eyes. J Neurosci Res 2014;92(9):1100–9.

IV.  Model Organisms of Epigenetics



REFERENCES

[21] Ito K, Awano W, Suzuki K, Hiromi Y, Yamamoto D. The Drosophila mushroom body is a quadruple structure of clonal units each of which contains a virtually identical set of neurones and glial cells. Development 1997;124(4):761–71. [22] Barnstedt O, Owald D, Felsenberg J, Brain R, Moszynski JP, Talbot CB, et al. Memory-relevant mushroom body output synapses are cholinergic. Neuron 2016;89(6):1237–47. [23] Heisenberg M. Mushroom body memoir: from maps to models. Nat Rev Neurosci 2003;4(4):266–725. [24] Guven-Azkan T, D.R. L. Functional neuroanatomy of Drosophila olfactory memory formation. Learn Mem 2014;21(519–526):19. [25] Farris SM. Evolution of complex higher brain centers and behaviors: behavioral correlates of mushroom body elaboration in insects. Brain Behav Evol 2013;82(1):9–18. [26] Akalai DB, Wilson CF, Zong L, Tanaka NK, Ito K, Davis RL. Roles for Drosophila mushroom body neurons in olfactory learning and memory. Learn Mem 2006;13(5):659–68. [27] Margulies C TT, Dubnau J. Deconstructing memory in Drosophila. Curr Biol 2005;15(17):R700–R7013. [28] Heisenberg M. What do the mushroom bodies do for the insect brain? an introduction. Learn Mem 1998;5(1):1–10. [29] Lee G, Hall JC. A newly uncovered phenotype associated with the fruitless gene of Drosophila melanogaster: aggression-like head interactions between mutant males. Behav Genet 2000;30(4): 263–75. [30] Zordan MA, Benna C, Mazzotta G. Monitoring and analyzing Drosophila circadian locomotor activity. Methods Mol Biol 2007;362:67–81. [31] Mehren JE, Ejima A, Griffith LC. Unconventional sex: fresh approaches to courtship learning. Curr Opin Neurobiol 2004;14(6):745–50. [32] Keene AC, Waddell S. Drosophila olfactory memory: single genes to complex neural circuits. Nat Rev Neurosci 2007;8:341–54. [33] Busto GU, Cervantes-Sandoval I, Davis RL. Olfactory learning in Drosophila. Physiology 2010;26(6):338–46. [34] Dubnau J TT. Functional anatomy: from molecule to memory. Curr Biol 2001;11(6):R240–3. [35] Quinn WG, Harris WA, Benzer S. Conditioned behavior in Drosophila melanogaster. Proc Natl Acad Sci USA 1974;71(3):708–12. [36] Kuo MH, Allis CD. Roles of histone acetyltransferases and deacetylases in gene regulation. Bioessays 1998;20(8):615–26. [37] Grunstein M. Histone acetylation in chromatin structure and transcription. Nature 1997;389(6649):349–52. [38] Peterson CL, Laniel MA. Histones and histone modifications. Curr Biol 2004;14(14):R546–51. [39] Schneider A, Chatterjee S, Bousiges O, Selvi BR, Swaminathan A, Cassel R, et al. Acetyltransferases (HATs) as targets for neurological therapeutics. Neurotherapeutics 2013;10(4):568–88. [40] Wood MA, Hawk JD, Abel T. Combinatorial chromatin modifications and memory storage: a code for memory? Learn Mem 2006;13(3):241–4. [41] Scharf AN, Imhof A. Every methyl counts–epigenetic calculus. FEBS Lett 2011;585(13):2001–7. [42] Day JJ, Sweatt JD. Epigenetic treatments for cognitive impairments. Neuropsychopharmacology 2012;37(1):247–60. [43] Peixoto L, Abel T. The role of histone acetylation in memory formation and cognitive impairments. Neuropsychopharmacology 2013;38(1):62–76. [44] Krajewski WA, Becker PB. Reconstitution of hyperacetylated, DNase I-sensitive chromatin characterized by high conformational flexibility of nucleosomal DNA. Proc Natl Acad Sci USA 1998;95(4):1540–5. [45] Ikura T, Ogryzko VV, Grigoriev M, Groisman R, Wang J, Horikoshi M, et al. Involvement of the TIP60 histone acetylase complex in DNA repair and apoptosis. Cell 2000;102(4):463–73.

225

[46] Mizzen CA, Yang XJ, Kokubo T, Brownell JE, Bannister AJ, Owen-Hughes T, et al. The TAF(II)250 subunit of TFIID has histone acetyltransferase activity. Cell 1996;87(7):1261–70. [47] Nightingale KP, Wellinger RE, Sogo JM, Becker PB. Histone acetylation facilitates RNA polymerase II transcription of the Drosophila hsp26 gene in chromatin. EMBO J 1998;17(10):2865–76. [48] Liu D. From histone modification to gene repression: epigenetic regulation by holocarboxylase synthetase-containing repression complex and FAD-dependent lysine specific demethylase. Lincoln, NE: University of Nebraska-Lincoln; 2014. [49] Das C, Lucia MS, Hansen KC, Tyler JK. CBP/p300-mediated acetylation of histone H3 on lysine 56. Nature 2009;459(7243):113–7. [50] Johnstone RW. Hisonte-deacetylase inhibitors: novel drugs for the treatment of cancer. Nat Rev Drug Discov 2002;1:287–99. [51] Murray K. The occurrence of epsilon-N-methyl lysine in histones. Biochemistry 1964;3:10–5. [52] Hempel K, Lange HW, Birkofer L. [Epsilon-N-trimethyllysine, a new amino acid in histones]. Naturwissenschaften 1968;55(1):37. [53] Paik WK, Kim S. ε-N-dimethyllysine in histones. Biochem Biophys Res Commun 1967;27(4):479–83. [54] Greer EL, Shi Y. Histone methylation: a dynamic mark in health, disease and inheritance. Nat Rev Genet 2012;13(5):343–57. [55] Ebert A, Schotta G, Lein S, Kubicek S, Krauss V, Jenuwein T, et al. Su(var) genes regulate the balance between euchromatin and heterochromatin in Drosophila. Genes Dev 2004;18(23):2973–83. [56] Talasz H, Lindner HH, Sarg B, Helliger W. Histone H4-lysine 20 monomethylation is increased in promoter and coding regions of active genes and correlates with hyperacetylation. J Biol Chem 2005;280(46):38814–22. [57] Congdon LM, Houston SI, Veerappan CS, Spektor TM, Rice JC. PR-Set7-mediated monomethylation of histone H4 lysine 20 at specific genomic regions induces transcriptional repression. J Cell Biochem 2010;110(3):609–19. [58] Schuettengruber B, Chourrout D, Vervoort M, Leblanc B, Cavalli G. Genome regulation by polycomb and trithorax proteins. Cell 2007;128(4):735–45. [59] Cleard F, Moshkin Y, Karch F, Maeda RK. Probing long-distance regulatory interactions in the Drosophila melanogaster bithorax complex using Dam identification. Nat Genet 2006;38(8):931–5. [60] Comet I, Savitskaya E, Schuettengruber B, Negre N, Lavrov S, Parshikov A, et al. PRE-mediated bypass of two Su(Hw) insulators targets PcG proteins to a downstream promoter. Dev Cell 2006;11(1):117–24. [61] Ebert A, Lein S, Schotta G, Reuter G. Histone modification and the control of heterochromatic gene silencing in Drosophila. Chromosome Res 2006;14(4):377–92. [62] Elgin SC, Gunter R. Position-effect variegation, heterochromatin formation, and gene silencing in Drosophila. Cold Spring Harb Perspect Biol 2013;5(8):a017780. [63] Yasuhara JC, Wakimoto BT. Molecular landscape of modified histones in Drosophila heterochromatic genes and euchromatinheterochromatin transition zones. PLoS Genet 2008;4(1):e16. [64] Girton JR, Johansen KM. Chromatin structure and the regulation of gene expression: the lessons of PEV in Drosophila. Adv Genet 2008;61:1–43. [65] Yin H, Sweeney S, Raha D, Snyder M, Lin H. A high-resolution whole-genome map of key chromatin modifications in the adult Drosophila melanogaster. PLoS Genet 2011;7(12):e1002380. [66] Schotta G, Lachner M, Sarma K, Ebert A, Sengupta R, Reuter G, et al. A silencing pathway to induce H3-K9 and H4-K20 trimethylation at constitutive heterochromatin. Genes Dev 2004;18(11):1251– 62. [67] Zhang W, Deng H, Bao X, Lerach S, Girton J, Johansen J, et al. The JIL-1 histone H3S10 kinase regulates dimethyl H3K9 modifications and heterochromatic spreading in Drosophila. Development 2006;133(2):229–35.

IV.  Model Organisms of Epigenetics

226

14.  DROSOPHILA EPIGENETICS

[68] Turner BM, Birley AJ, Lavender J. Histone H4 isoforms acetylated at specific lysine residues define individual chromosomes and chromatin domains in Drosophila polytene nuclei. Cell 1992;69(2):375–84. [69] Weiler KS. E(var)3-9 of Drosophila melanogaster encodes a zinc finger protein. Genetics 2007;177(1):167–78. [70] Farkas G, Gausz J, Galloni M, Reuter G, Gyurkovics H, Karch F. The Trithorax-like gene encodes the Drosophila GAGA factor. Nature 1994;371(6500):806–8. [71] Seum C, Spierer A, Pauli D, Szidonya J, Reuter G, Spierer P. Position-effect variegation in Drosophila depends on dose of the gene encoding the E2F transcriptional activator and cell cycle regulator. Development 1996;122(6):1949–56. [72] Hofmann A, Brunner M, Korge G. The winged-helix transcription factor JUMU is a haplo-suppressor/triplo-enhancer of PEV in various tissues but exhibits reverse PEV effects in the brain of Drosophila melanogaster. Chromosome Res 2009;17(3):347–58. [73] Sass GL, Henikoff S. Comparative analysis of position-effect variegation mutations in Drosophila melanogaster delineates the targets of modifiers. Genetics 1998;148(2):733–41. [74] Locke J, Kotarski MA, Tartof KD. Dosage-dependent modifiers of position effect variegation in Drosophila and a mass action model that explains their effect. Genetics 1988;120(1):181–98. [75] Grewal SI, Elgin SC. Transcription and RNA interference in the formation of heterochromatin. Nature 2007;447(7143):399–406. [76] Schotta G, Ebert A, Krauss V, Fischer A, Hoffmann J, Rea S, et al. Central role of Drosophila SU(VAR)3-9 in histone H3-K9 methylation and heterochromatic gene silencing. EMBO J 2002;21(5):1121–31. [77] Rudolph T, Yonezawa M, Lein S, Heidrich K, Kubicek S, Schafer C, et al. Heterochromatin formation in Drosophila is initiated through active removal of H3K4 methylation by the LSD1 homolog SU(VAR)3-3. Mol Cell 2007;26(1):103–15. [78] Czermin B, Schotta G, Hulsmann BB, Brehm A, Becker PB, Reuter G, et al. Physical and functional association of SU(VAR)3-9 and HDAC1 in Drosophila. EMBO Rep 2001;2(10):915–9. [79] Mottus R, Sobel RE, Grigliatti TA. Mutational analysis of a histone deacetylase in Drosophila melanogaster: missense mutations suppress gene silencing associated with position effect variegation. Genetics 2000;154(2):657–68. [80] Vermaak D, Malik HS. Multiple roles for heterochromatin protein 1 genes in Drosophila. Annu Rev Genet 2009;43:467–92. [81] Lee YC. The role of piRNA-mediated epigenetic silencing in the population dynamics of transposable elements in Drosophila melanogaster. PLoS Genet 2015;11(6):e1005269. [82] de Vanssay A, Bouge AL, Boivin A, Hermant C, Teysset L, Delmarre V, et al. piRNAs and epigenetic conversion in Drosophila. Fly 2013;7(4):237–41. [83] Cleard F, Delattre M, Spierer P. SU(VAR)3-7, a Drosophila heterochromatin-associated protein and companion of HP1 in the genomic silencing of position-effect variegation. EMBO J 1997;16(17):5280–8. [84] James TC, Eissenberg JC, Craig C, Dietrich V, Hobson A, Elgin SC. Distribution patterns of HP1, a heterochromatin-associated nonhistone chromosomal protein of Drosophila. Eur J Cell Biol 1989;50(1):170–80. [85] Grewal SI, Elgin SC. Transcription and RNAi in the formation of heterochromatin. Nat Rev 2007;447:399–406. [86] Vogel MJ, Pagie L, Talhout W, Nieuwland M, Kerkhoven RM, van Steensel B. High-resolution mapping of heterochromatin redistribution in a Drosophila position-effect variegation model. Epigenetics Chromatin 2009;2(1):1. [87] Cao R, Tsukada Y, Zhang Y. Role of Bmi-1 and Ring1A in H2A ubiquitylation and Hox gene silencing. Mol Cell 2005;20(6):845–54. [88] Steffen PA, Ringrose L. What are memories made of? How Polycomb and Trithorax proteins mediate epigenetic memory. Nat Rev Mol Cell Biol 2014;15(5):340–56.

[89] Grimaud C, Negre N, Cavalli G. From genetics to epigenetics: the tale of Polycomb group and Trithorax group genes. Chromosome Res 2006;14(4):363–75. [90] Schwartz YB, Pirrotta V. Polycomb silencing mechanisms and the management of genomic programmes. Nat Rev Genet 2007;8(1):9–22. [91] Mallo M, Alonso CR. The regulation of Hox gene expression during animal development. Development 2013;140(19):3951–63. [92] Lewis RA, Wakimoto BT, Denell RE, Kaufman TC. Genetic analysis of the Antennapedia gene complex (Ant-C) and adjacent chromosomal regions of Drosophila melanogaster. II. Polytene chromosome segments 84A-84B1,2. Genetics 1980;95(2):383–97. [93] Lewis EB. A gene complex controlling segmentation in Drosophila. Nature 1978;276(5688):565–70. [94] Kmita M, Duboule D. Organizing axes in time and space; 25 years of colinear tinkering. Science 2003;301(5631):331–3. [95] Mallo M, Alonso CR. The regulation of Hox gene expression during animal development. Development 2013;140:3951–63. [96] Lavigne M, Francis NJ, King IF, Kingston RE. Propagation of silencing; recruitment and repression of naive chromatin in trans by polycomb repressed chromatin. Mol Cell 2004;13(3):415–25. [97] Steffen PA, Ringrose L. What are memories made of? How Polycomb and Trithorax proteins mediate epigenetic memory. Nat Rev Mol Cell Biol 2014;15:340–56. [98] Muller J, Kassis JA. Polycomb response elements and targeting of Polycomb group proteins in Drosophila. Curr Opin Genet Dev 2006;16(5):476–84. [99] Stock JK, Giadrossi S, Casanova M, Brookes E, Vidal M, Koseki H, et al. Ring1-mediated ubiquitination of H2A restrains poised RNA polymerase II at bivalent genes in mouse ES cells. Nat Cell Biol 2007;9(12):1428–35. [100] Shao Z, Raible F, Mollaaghababa R, Guyon JR, Wu CT, Bender W, et al. Stabilization of chromatin structure by PRC1, a Polycomb complex. Cell 1999;98(1):37–46. [101] Francis NJ, Saurin AJ, Shao Z, Kingston RE. Reconstitution of a functional core polycomb repressive complex. Mol Cell 2001;8(3):545–56. [102] Francis NJ, Kingston RE, Woodcock CL. Chromatin compaction by a polycomb group protein complex. Science 2004;306(5701):1574–7. [103] Bantignies F, Roure V, Comet I, Leblanc B, Schuettengruber B, Bonnet J, et al. Polycomb-dependent regulatory contacts between distant Hox loci in Drosophila. Cell 2011;144(2):214–26. [104] Tolhuis B, Blom M, Kerkhoven RM, Pagie L, Teunissen H, Nieuwland M, et al. Interactions among Polycomb domains are guided by chromosome architecture. PLoS Genet 2011;7(3):e1001343. [105] Blomen VA, Boonstra J. Stable transmission of reversible modifications: maintenance of epigenetic information through the cell cycle. Cell Mol Life Sci 2011;68(1):27–44. [106] Kiefer JC. Epigenetics in development. Dev Dyn 2007;236(4):1144–56. [107] Yuan W, Xu M, Huang C, Liu N, Chen S, Zhu B. H3K36 methylation antagonizes PRC2-mediated H3K27 methylation. J Biol Chem 2011;286(10):7983–9. [108] Schmitges FW, Prusty AB, Faty M, Stutzer A, Lingaraju GM, Aiwazian J, et al. Histone methylation by PRC2 is inhibited by active chromatin marks. Mol Cell 2011;42(3):330–41. [109] Tie F, Banerjee R, Stratton CA, Prasad-Sinha J, Stepanik V, Zlobin A, et al. CBP-mediated acetylation of histone H3 lysine 27 antagonizes Drosophila Polycomb silencing. Development 2009;136(18):3131–41. [110] Devaiah BN, Lewis BA, Cherman N, Hewitt MC, Albrecht BK, Robey PG, et al. BRD4 is an atypical kinase that phosphorylates serine2 of the RNA polymerase II carboxy-terminal domain. Proc Natl Acad Sci USA 2012;109(18):6927–32. [111] Armstrong JA, Papoulas O, Daubresse G, Sperling AS, Lis JT, Scott MP, et al. The Drosophila BRM complex facilitates global

IV.  Model Organisms of Epigenetics



[112]

[113]

[114] [115]

[116] [117]

[118] [119]

[120] [121]

[122]

[123]

[124]

[125]

[126]

[127]

[128]

[129] [130]

[131] [132] [133] [134]

REFERENCES

transcription by RNA polymerase II. EMBO J 2002;21(19): 5245–54. Fitzsimons HL, Schwartz S, Given FM, Scott MJ. The histone deacetylase HDAC4 regulates long-term memory in Drosophila. PLoS One 2013;8(12):e83903. Geisler SJ, Paro R. Trithorax and Polycomb group-dependent regulation: a tale of opposing activities. Development 2015;142(17):2876–87. Brockdorff N. Noncoding RNA and Polycomb recruitment. RNA 2013;19(4):429–42. Schuettengruber B, Martinez AM, Iovino N, Cavalli G. Trithorax group proteins: switching genes on and keeping them active. Nat Rev Mol Cell Biol 2011;12(12):799–814. Meller VH, Joshi SS, Deshpande N. Modulation of Chromatin by Noncoding RNA. Annu Rev Genet 2015;49:673–95. Bauer M, Trupke J, Ringrose L. The quest for mammalian Polycomb response elements: are we there yet? Chromosoma 2016;125(3):471–96. Aranda S, Mas G, Di Croce L. Regulation of gene transcription by Polycomb proteins. Sci Adv 2015;1(11):e1500737. Feng YQ, Desprat R, Fu H, Olivier E, Lin CM, Lobell A, et al. DNA methylation supports intrinsic epigenetic memory in mammalian cells. PLoS Genet 2006;2(4):e65. Ehrenhofer-Murray AE. Chromatin dynamics at DNA replication, transcription and repair. Eur J Biochem 2004;271(12):2335–49. Lanzuolo C, Lo Sardo F, Diamantini A, Orlando V. PcG complexes set the stage for epigenetic inheritance of gene silencing in early S phase before replication. PLoS Genet 2011;7(11):e1002370. Petruk S, Black KL, Kovermann SK, Brock HW, Mazo A. Stepwise histone modifications are mediated by multiple enzymes that rapidly associate with nascent DNA during replication. Nat Commun 2013;4:2841. Petruk S, Sedkov Y, Johnston DM, Hodgson JW, Black KL, Kovermann SK, et al. TrxG and PcG proteins but not methylated histones remain associated with DNA through replication. Cell 2012;150(5):922–33. Huang H, Yu Z, Zhang S, Liang X, Chen J, Li C, et al. Drosophila CAF-1 regulates HP1-mediated epigenetic silencing and pericentric heterochromatin stability. J Cell Sci 2010;123(16):2853–61. Fischle W, Wang Y, Allis CD. Binary switches and modification cassettes in histone biology and beyond. Nature 2003;425(6957):475–9. Giet R, Glover DM. Drosophila aurora B kinase is required for histone H3 phosphorylation and condensin recruitment during chromosome condensation and to organize the central spindle during cytokinesis. J Cell Biol 2001;152(4):669–82. Hirota T, Lipp JJ, Toh BH, Peters JM. Histone H3 serine 10 phosphorylation by Aurora B causes HP1 dissociation from heterochromatin. Nature 2005;438(7071):1176–80. Heard E, Disteche CM. Dosage compensation in mammals: fine-tuning the expression of the X chromosome. Genes Dev 2006;20(14):1848–67. Lucchesi JC, Kuroda MI. Dosage compensation in Drosophila. Cold Spring Harb Perspect Biol 2015;7(5). Conrad T, Akhtar A. Dosage compensation in Drosophila melanogaster: epigenetic fine-tuning of chromosome-wide transcription. Nat Rev Genet 2011;13(2):123–34. Belote JM, Lucchesi JC. Male-specific lethal mutations of Drosophila melanogaster. Genetics 1980;96(1):165–86. Belote JM, Lucchesi JC. Control of X chromosome transcription by the maleless gene in Drosophila. Nature 1980;285(5766):573–5. Lucchesi JC. Dosage compensation in Drosophila and the “complex” world of transcriptional regulation. Bioessays 1996;18(7):541–7. Scott MJ, Pan LL, Cleland SB, Knox AL, Heinrich J. MSL1 plays a central role in assembly of the MSL complex, essential for dosage compensation in Drosophila. EMBO J 2000;19(1):144–55.

227

[135] Copps K, Richman R, Lyman LM, Chang KA, RampersadAmmons J, Kuroda MI. Complex formation by the Drosophila MSL proteins: role of the MSL2 RING finger in protein complex assembly. EMBO J 1998;17(18):5409–17. [136] Hallacli E, Lipp M, Georgiev P, Spielman C, Cusack S, Akhtar A, et al. Msl1-mediated dimerization of the dosage compensation complex is essential for male X-chromosome regulation in Drosophila. Mol Cell 2012;48(4):587–600. [137] Lee CG, Chang KA, Kuroda MI, Hurwitz J. The NTPase/helicase activities of Drosophila maleless, an essential factor in dosage compensation. EMBO J 1997;16(10):2671–81. [138] Meller VH, Gordadze PR, Park Y, Chu X, Stuckenholz C, Kelley RL, et al. Ordered assembly of roX RNAs into MSL complexes on the dosage-compensated X chromosome in Drosophila. Curr Biol 2000;10(3):136–43. [139] Meller VH, Rattner BP. The roX genes encode redundant malespecific lethal transcripts required for targeting of the MSL complex. EMBO J 2002;21(5):1084–91. [140] Ilik IA, Quinn JJ, Georgiev P, Tavares-Cadete F, Maticzka D, Toscano S, et al. Tandem stem-loops in roX RNAs act together to mediate X chromosome dosage compensation in Drosophila. Mol Cell 2013;51(2):156–73. [141] Park SW, Kang Y, Sypula JG, Choi J, Oh H, Park Y. An evolutionarily conserved domain of roX2 RNA is sufficient for induction of H4-Lys16 acetylation on the Drosophila X chromosome. Genetics 2007;177(3):1429–37. [142] Akhtar A, Zink D, Becker PB. Chromodomains are protein-RNA interaction modules. Nature 2000;407(6802):405–9. [143] Tie F, Banerjee R, Conrad PA, Scacheri PC, Harte PJ. Histone demethylase UTX and chromatin remodeler BRM bind directly to CBP and modulate acetylation of histone H3 lysine 27. Mol Cell Biol 2012;32(12):2323–34. [144] Gilfillan GD, Dahlsveen IK, Becker PB. Lifting a chromosome: dosage compensation in Drosophila melanogaster. FEBS Lett 2004;(567):8–14. [145] Bashaw GJ, Baker BS. The msl-2 dosage compensation gene of Drosophila encodes a putative DNA-binding protein whose expression is sex specifically regulated by Sex-lethal. Development 1995;121(10):3245–58. [146] Abaza I, Coll O, Patalano S, Gebauer F. Drosophila UNR is required for translational repression of male-specific lethal 2 mRNA during regulation of X-chromosome dosage compensation. Genes Dev 2006;20(3):380–9. [147] Beckmann K, Grskovic M, Gebauer F, Hentze MW. A dual inhibitory mechanism restricts msl-2 mRNA translation for dosage compensation in Drosophila. Cell 2005;122(4):529–40. [148] Salz HK. Sex determination in insects: a binary decision based on alternative splicing. Curr Opin Genet Dev 2011;21(4):395–400. [149] Bai X, Alekseyenko AA, Kuroda MI. Sequence-specific targeting of MSL complex regulates transcription of the roX RNA genes. EMBO J 2004;23(14):2853–61. [150] Rattner BP, Meller VH. Drosophila male-specific lethal 2 protein controls sex-specific expression of the roX genes. Genetics 2004;166(4):1825–32. [151] Kelley RL, Meller VH, Gordadze PR, Roman G, Davis RL, Kuroda MI. Epigenetic spreading of the Drosophila dosage compensation complex from roX RNA genes into flanking chromatin. Cell 1999;98(4):513–22. [152] Alekseyenko AA, Peng S, Larschan E, Gorchakov AA, Lee OK, Kharchenko P, et al. A sequence motif within chromatin entry sites directs MSL establishment on the Drosophila X chromosome. Cell 2008;134(4):599–609. [153] Straub T, Grimaud C, Gilfillan GD, Mitterweger A, Becker PB. The chromosomal high-affinity binding sites for the Drosophila dosage compensation complex. PLoS Genet 2008;4(12): e1000302.

IV.  Model Organisms of Epigenetics

228

14.  DROSOPHILA EPIGENETICS

[154] Kuzu G, Kaye EG, Chery J, Siggers T, Yang L, Dobson JR, et al. Expansion of GA dinucleotide repeats increases the density of CLAMP binding sites on the X-chromosome to promote Drosophila dosage compensation. PLoS Genet 2016;12(7):e1006120. [155] Soruco MM, Chery J, Bishop EP, Siggers T, Tolstorukov MY, Leydon AR, et al. The CLAMP protein links the MSL complex to the X chromosome during Drosophila dosage compensation. Genes Dev 2013;27(14):1551–6. [156] Bachtrog D, Toda NRT, Lockton S. Dosage compensation and demasculinization of X chromosomes in Drosophila. Curr Biol 2010;16:1–16. [157] Kind J, Akhtar A. Cotranscriptional recruitment of the dosage compensation complex to X-linked target genes. Genes Dev 2007;21(16):2030–40. [158] Bhadra U, Pal-Bhadra M, Birchler JA. Role of the male specific lethal (msl) genes in modifying the effects of sex chromosomal dosage in Drosophila. Genetics 1999;152(1):249–68. [159] Morra R, Smith ER, Yokoyama R, Lucchesi JC, The MLE. subunit of the Drosophila MSL complex uses its ATPase activity for dosage compensation and its helicase activity for targeting. Mol Cell Biol 2008;28(3):958–66. [160] Gu W, Wei X, Pannuti A, Lucchesi JC. Targeting the chromatinremodeling MSL complex of Drosophila to its sites of action on the X chromosome requires both acetyl transferase and ATPase activities. EMBO J 2000;19(19):5202–11. [161] Gu W, Szauter P, Lucchesi JC. Targeting of MOF, a putative histone acetyl transferase, to the X chromosome of Drosophila melanogaster. Dev Genet 1998;22(1):56–64. [162] Gelbart ME, Kuroda MI. Drosophila dosage compensation: a complex voyage to the X chromosome. Development 2009;136(9):1399–410. [163] Smith ER, Allis CD, Lucchesi JC. Linking global histone acetylation to the transcription enhancement of X-chromosomal genes in Drosophila males. J Biol Chem 2001;276(34):31483–6. [164] Bell O, Conrad T, Kind J, Wirbelauer C, Akhtar A, Schubeler D. Transcription-coupled methylation of histone H3 at lysine 36 regulates dosage compensation by enhancing recruitment of the MSL complex in Drosophila melanogaster. Mol Cell Biol 2008;28(10):3401–9. [165] Larschan E, Bishop EP, Kharchenko PV, Core LJ, Lis JT, Park PJ, et al. X chromosome dosage compensation via enhanced transcriptional elongation in Drosophila. Nature 2011;471(7336): 115–8. [166] Gilfillan GD, Straub T, de Wit E, Greil F, Lamm R, van Steensel B, et al. Chromosome-wide gene-specific targeting of the Drosophila dosage compensation complex. Genes Dev 2006;20(7):858–70. [167] Corona DF, Clapier CR, Becker PB, Tamkun JW. Modulation of ISWI function by site-specific histone acetylation. EMBO Rep 2002;3(3):242–7. [168] Xu S, Wilf R, Memon T, Sarthi J, Elefant F. Epigenetic regulation of learning and memory by Tip60 HAT action. Genetics 2014;4:1571–86. [169] Pirooznia K, Elefant F. Modulating epigenetic HAT activity: a promising therapuetic option for neurological disease? J Mol Cloning Genet Recomb 2012;1(1):1–3. [170] Pirooznia K, Elefant F. Targeting specific HATs for neurodegenerative disease treatment: translating basic biology to therapeutic possibilities. Front Cell Neurosci 2013;7(30):1–18. [171] Kandel ER, Pittenger C. The past, the future and the biology of memory storage. Philos Trans R Soc Lond B 1999;354(1392):2027–52. [172] Lonze BE, Ginty DD. Function and regulation of CREB family transcription factors in the nervous system. Neuron 2002;35(4):605–23. [173] Watts RJ, Hoopfer ED, Luo L. Axon pruning during Drosophila metamorphosis: evidence for local degeneration and requirement of the ubiquitin-proteasome system. Neuron 2003;38(6):871–85.

[174] Yin JC, Wallach JS, Del Vecchio M, Wilder EL, Zhou H, Quinn WG, et al. Induction of a dominant negative CREB transgene specifically blocks long-term memory in Drosophila. Cell 1994;79(1):49–58. [175] Hirano Y, Masuda T, Naganos S, Matsuno M, Ueno K, Miyashita T, et al. Fasting launches CRTC to facilitate long-term memory formation in Drosophila. Science 2013;339(6118):443–6. [176] Yin JC, Del Vecchio M, Zhou H, Tully T. CREB as a memory modulator: induced expression of a dCREB2 activator isoform enhances long-term memory in Drosophila. Cell 1995;81(1):107–15. [177] Wood MA, Attner MA, Oliveira AM, Brindle PK, Abel T. A transcription factor-binding domain of the coactivator CBP is essential for long-term memory and the expression of specific target genes. Learn Mem 2006;13(5):609–17. [178] Chawla S, Vanhoutte P, Arnold FJ, Huang CL, Bading H. Neuronal activity-dependent nucleocytoplasmic shuttling of HDAC4 and HDAC5. J Neurochem 2003;85(1):151–9. [179] Kahsai L, Zars T. Learning and memory in Drosophila: behaviour, genetics and neural systems. Int Rev Neurobiol 2011;99:139–67. [180] Fiala A. Olfaction and olfactory learning in Drosophila: recent progress. Curr Opin Neurobiol 2007;17(6):720–6. [181] Siwicki KK, Ladewski L. Associative learning and memory in Drosophila: beyond olfactory conditioning. Behav Processes 2003;64(2):225–38. [182] Xu S, Elefant F. Tip off the HAT—epigenetic control of learning and memory by Drosophila Tip60. Fly 2015;9(1):22–8. [183] Fitzsimons HL, Scott MJ. Genetic modulation of Rpd3 expression impairs long-term courtship memory in Drosophila. PLoS One 2011;6(12):e29171. [184] Saha RN, Pahan K. HATs and HDACs in neurodegeneration: a tale of disconcerted acetylation homeostasis. Cell Death Differ 2006;13(4):539–50. [185] Johnson AA, Sarthi J, Pirooznia SK, Reube W, Elefant F. Increasing Tip60 HAT levels rescues axonal transport defects and associated behavioral phenotypes in a Drosophila Alzheimer’s disease model. J Neurosci 2013;33(17):7535–47. [186] Pirooznia SK, Chiu K, Chan MT, Zimmerman JE, Elefant F. Epigenetic regulation of axonal growth of Drosophila pacemaker cells by histone acetyltransferase tip60 controls sleep. Genetics 2012;192(4):1327–45. [187] Mayford M, Siegelbaum SA, Kandel ER. Synapses and memory storage. Cold Spring Harb Perspect Biol 2012;4(6):a005751. [188] Lorbeck M, Pirooznia K, Sarthi J, Zhu X, Elefant F. Microarray analysis uncovers a role for Tip60 in nervous system function and general metabolism. PLoS One 2011;6(4):e18412. [189] Pirooznia SK, Elefant F. Targeting specific HATs for neurodegenerative disease treatment: translating basic biology to therapeutic possibilities. Front Cell Neurosci 2013;7:30. [190] Francis NJ, Follmer NE, Simon MD, Aghia G, Butler JD. Polycomb proteins remain bound to chromatin and DNA during DNA replication in vitro. Cell 2009;137(1):110–22. [191] Pirooznia SK, Sarthi J, Johnson AA, Toth MS, Chiu K, Koduri S, et al. Tip60 HAT activity mediates APP induced lethality and apoptotic cell death in the CNS of a Drosophila Alzheimer’s disease model. PLoS One 2012;7(7):e41776. [192] Pirooznia SK, Elefant F. A HAT for sleep? Epigenetic regulation of sleep by Tip60 in Drosophila. Fly 2013;7(2):99–104. [193] Xu S, Wilf R, Menon T, Panikker P, Sarthi J, Elefant F. Epigenetic control of learning and memory in Drosophila by Tip60 HAT action. Genetics 2014;198(4):1571–86. [194] Cutler T, Sarkar A, Moran M, Steffensmeier A, Puli OR, Mancini G, et al. Drosophila eye model to study neuroprotective role of CREB binding protein (CBP) in Alzheimer’s disease. PLoS One 2015;10(9):e0137691. [195] Nithianantharajah J, Hannan AJ. Enriched environments, experience-dependent plasticity and disorders of the nervous system. Nat Rev Neurosci 2006;7(9):697–709.

IV.  Model Organisms of Epigenetics



REFERENCES

[196] Carulli D, Foscarin S, Rossi F. Activity-dependent plasticity and gene expression modifications in the adult CNS. Front Mol Neurosci 2011;4(50):1–11. [197] Sweatt JD. Experience-dependent epigenetic modifications in the central nervous system. Biol Psychiatry 2009;65(3):191–7. [198] Fischer A, Sananbenesi F, Wang X, Dobbin M, Tsai LH. Recovery of learning and memory is associated with chromatin remodelling. Nature 2007;447(7141):178–82. [199] Praag Hv, Kempermann G, Gage FH. Neural consequences of environmental enrichment. Nat Neurosci 2000;1:191–8. [200] Donlea JM, Shaw PJ. Sleeping together: using social interactions to understand the role of sleep in plasticity. Adv Genet 2009;68:57–81. [201] Heisenberg M, Heusipp M, Wanke C. Structural plasticity in the Drosophila brain. J Neurosci 1995;15(3 Pt. 1):1951–60. [202] Aso Y, Grubel K, Busch S, Friedrich AB, Siwanowicz I, Tanimoto H. The mushroom body of adult Drosophila characterized by GAL4 drivers. J Neurogenet 2009;23(1–2):156–72. [203] Margulies C, Tully T, Dubnau J. Deconstructing memory in Drosophila. Curr Biol 2005;15(17):R700–13.

229

[204] Guven-Ozkan T, Davis R. Functional neuroanatomy of Drosophila olfactory memory formation. Learn Mem 2014;21(10):519–26. [205] Farris S. Evolution of complex higher brain centers and behaviors: behavioral correlates of mushroom body elaboration in insects. Brain Behav Evol 2013;82(1):9–18. [206] Technau GM. Fiber number in the mushroom bodies of adult Drosophila melanogaster depends on age, sex and experience. J Neurogenet 2007;21:183–96. [207] Lopez-Atalaya JP, Ciccarelli A, Viosca J, Valor LM, JimenezMinchan M, Canals S, et al. CBP is required for environmental enrichment-induced neurogenesis and cognitive enhancement. EMBO J 2011;30(20):4287–98. [208] Sweatt JD. Creating stable memories. Science 2011;331:869–71. [209] Penner MR, Roth TL, Barnes CA, Sweatt JD. An epigenetic hypothesis of aging-related cognitive dysfunction. Front Aging Neurosci 2010;2:9. [210] Xu S, Pankikker P, Iqbal S, Elefant F. Tip60 HAT action mediates environmental enrichment induced cognitive restoration. PLoS One 2016;11:1–26.

IV.  Model Organisms of Epigenetics

C H A P T E R

15 Mouse Models of Epigenetic Inheritance: Classification, Mechanisms, and Experimental Strategies Shaoshuai Mao*, Yongqin Li*, Bo Liu*, Tian Chi*,** *ShanghaiTech University, Shanghai, China; **Yale University Medical School, New Haven, CT, United States

O U T L I N E Two Types of Epigenetic Inheritance (EI): Mitotic Versus Generational

231

EI Signals: Cis Versus Trans

232

Mice as Model Mammals

233

Mouse EI Models: A Brief History

233

Conventional Mouse EI Models Spontaneous Random Epigenetic Variations Experimentally Induced EI

234 234 237

TWO TYPES OF EPIGENETIC INHERITANCE (EI): MITOTIC VERSUS GENERATIONAL Transient signals, including external cues (delivered by extracellular agents, such as hormones or environmental toxicans) and cell-intrinsic “noises” (emerging stochastically from ongoing biological processes inside the cells), can trigger sustained changes in gene functions without mutating DNA. Such stable changes in the functional states of genes, called epigenetic changes, is reversible but can be durable enough to be passed onto daughter cells following mitosis, and even to the offspring of an organism if the changes occur in the germline. The propagation of epigenetic states from a somatic cell to its progeny cells, or from a gamete to the resultant Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00015-8 Copyright © 2017 Elsevier Inc. All rights reserved.

Targeting Epigenetic Modifications to Reporter Genes: A Unique EI Model Tractable for Mechanistic Studies

238

Conclusions

240

References

241

organism, is each an example of epigenetic inheritance (EI), with former called “Mitotic EI” in this chapter and the latter, “Generational EI” [1,2] (Fig. 15.1). Generational EI includes at least two subtypes: “Intergenerational EI,” where the epigenetic state is transmitted from the individual directly experiencing environmental challenges to the immediate offspring (F1 to F2 in Fig. 15.1), and “Transgenerational EI,” defined as “the germline transmission of epigenetic information between generations in the absence of any environmental exposure” [3] (F2 to F3 in Fig. 15.1; F1 to F2 is not transgenerational because the F1 germline has been exposed). Of note, the definition of EI in this chapter echoes an early definition of epigenetics as “the study of mitotically and/or meiotically heritable changes in gene function that cannot be explained by changes in DNA sequence” [4]. However,

231

232

15.  Mouse Models of Epigenetic Inheritance: Classification, Mechanisms, and Experimental Strategies

FIGURE 15.1  A hypothetical model of EI. (A) Normal development of an embryo (F1) inside the mother (F0). The embryo consists of soma and germline, the latter conventionally considered F2. The somatic cells undergo mitosis as the embryo develops and grows into the adult. Germline development is more complex. In males, the primordial germ cells (PGCs) emerge at embryonic day 7 (E7) and become mitotically arrested “prospermatogonia” at E14. Shortly after birth, prospermatogonia become spermatogonial stem cells (SSCs), which proliferate and mature into primary spermatocytes before the latter undergo meiosis to become sperm at about 1 month after birth. In females, the occytes have already developed in the embryo, but are arrested in meiosis till fertilization. The gametes (sperm and egg) develop into the offspring via numerous steps (dashed arrow). (B) An environmental factor acting on the embryo (lightening bolt) inflicts a similar epigenetic lesion (explosion) on both the soma (e.g., brain) and the germline (top right image). The lesion in the brain survives mitosis and is transmitted to the brain of the adult F1 mouse (Mitotic EI, left panel). On the other hand, the lesion in the fetal germline survives both mitosis and meiosis to emerge in the mature gametes in the adult F1 mouse, and is subsequently passed onto the offspring via the gametes (Generational EI, right panel).

we add the term “inheritance” to “epigenetic” to emphasize the heritable nature of the EI process. This emphasis is necessary because nowadays, “epigenetic” is also loosely used to refer to any gene regulatory process that occurs at the level of chromatin, be it heritable or not [5]. We also wish to emphasize that by definition, to qualify as “epigenetically heritable,” the functional state of the gene must persist through cell division, rather than get erased during the division and then reinstated thereafter. This distinction is not only conceptually important, but also medically relevant, as seen at the human tumor suppressor MHL. A certain allele of this gene is methylated and silenced in the somatic cells in the family members, but the functional state is cleared in the germline of the parents before being reestablished in the next generation, which could dictate a different therapeutic strategy than if the silencing is epigenetically inherited [6]. Generational EI is obviously much harder to achieve than Mitotic EI, but the basic mechanisms underlying the two forms of EI are similar, as detailed here.

EI SIGNALS: CIS VERSUS TRANS In general, there are two mechanisms whereby a transient stimulus can induce heritable changes in gene function [7]. First, the stimulus causes modifications

to chromatin (histones and/or DNA), which are then recognized and perpetuated by cellular machinery capable of maintaining and replicating the modifications (Fig. 15.2A). This EI system works in cis in that the chromatin marking is a molecular signature of genes, affecting only the alleles it physically associates with. In contrast, the second EI system works in trans. Here, the transient stimulus triggers the expression of a transacting factor (such as transcription activator or RNA) capable of maintaining its own expression via, e.g., an autoregulatory loop; as the factor is freely diffusible, both the alleles will be affected (Fig. 15.2B). Mitotic and generational EI can each be mediated by cis or trans signals. For Mitotic EI, it is difficult to distinguish between the two EI signals, except in the case of monoallelic expression (e.g., genomic imprinting and X-inactivation), which by necessity is under the control of cis-signals. In contrast, the signals underlying Generational EI of a phenotype are readily distinguishable by out-crossing inbred epigenetically modified mice and their progeny to genetically identical control mice. If the phenotype in the offspring is inherited via cis-signals, its penetrance will decline progressively in successive generations as the epigenetically marked allele is increasingly diluted out by the “naïve” control allele. In contrast, the penetrance can remain unchanged if the inheritance is directed by strong, durable trans-signals.

IV.  Model Organisms of Epigenetics



Mouse EI Models: A Brief History

233

FIGURE 15.2  Maintenance and propagation of transcription states via cis- or trans-acting epigenetic signals, as represented by CpG methylation and transcription activator, respectively. (A) Cis-acting mechanism. A transient stimulus (lightening bulb) induces the methylase DNMT3-catalyzed de novo symmetric CpG methylation (CH3) on both strands of the DNA at a target gene; only one allele is shown for simplicity. Immediately following DNA replication, the DNA is hemimethylated, but the methylation tag preexisiting at the parental strand quickly helps recruit the maintenance methylase DNMT1 to copy the methylation to the daughter strand (blue), thus restoring the intact epigenetic signal. A similar mechanism might be in place if a methyl group on one of the two strands is accidentally lost in resting cells (bracket). Other chromatin features, including histone modifications, nonhistone chromatin proteins, higher-order chromatin structure, and nuclear localization are also heritable but the mechanisms underlying their heritance are poorly understood. (B) Trans-acting mechanism. A transient stimulus triggers the expression of an activator from its target gene, which then binds its own promoter to sustain its expression in the absence of the initial trigger; only one allele is shown for simplicity. During cell division, the activator, probably detached from the DNA, is partitioned to daughter cells, where it reestablishes its own expression.

MICE AS MODEL MAMMALS The laboratory mouse is the premier mammalian model organism for biomedical research. Mouse and human shares 99% of the genes and most physiological and pathological features, with similarities found in all systems including the nervous, cardiovascular, endocrine, immune, musculoskeletal systems [8]. Mice are small, docile, prolific, and easy to rear in controlled environments. Hundreds of inbred lines with divergent traits that suit various experimental purposes have arisen through a combination of natural evolution and human-directed breeding. Sophisticated mouse genetic tools have been developed to manipulate the mouse genome and define its function, which include insertion of exogenous genes via transgenesis and deletion of endogenous genes via homologous recombination [9]. Genome manipulation and phenotypic characterization are greatly facilitated with the advent of Cas9-based gene editing and high-throughput sequencing methods [10,11]. A concerted global effort, under the auspices of the International Mouse Phenotype Consortium, is in progress to “discover functional insight for every gene

by generating and systematically phenotyping 20,000 knockout mouse strains” (http://www.mousephenotype.org). Furthermore, large scale high-throughput random mutagenesis in mice is feasible using transposons (Sleeping Beauty and piggyback) or the chemical mutagen N-ethyl-N-nitrosourea (ENU), which enables both forward (gene-driven) and reverse (phenotype-driven) genetic screen [12–15]. Within the near future, the functions of the majority of protein-coding genes in mice are expected to be deciphered, and the noncoding RNAs are following suit. In many aspects, mouse is the most studied and best understood mammal, and the insights obtained from mouse research has proven invaluable for understanding human health and disease.

MOUSE EI MODELS: A BRIEF HISTORY The use of mice to study EI dates back at least to the late 19th century. At that time, the Lamarckian principle that “acquired characteristics are heritable” is accepted by some people, who would cite, as the key supporting evidence, anecdotal reports claiming that

IV.  Model Organisms of Epigenetics

234

15.  Mouse Models of Epigenetic Inheritance: Classification, Mechanisms, and Experimental Strategies

the phenotypes caused by body injury or mutilation are sometimes transmitted to the offspring. To refute the claim, August Weismann did a famous experiment in which he cut off mouse tails over successive generations and determined its effects on the offspring. Specifically, “901 young were produced by five generations of artificially mutilated parents, and yet there was not a single example of a rudimentary tail or of any other abnormality in this organ” [16]. Furthermore, as Weismann pointed out, the ceremonial mutilation of certain parts of human body from times immemorial (such as circumcision) has “not in a single instance led to the malformation or reduction in the part” [16]. Collectively these data demonstrate that the effect of mutilation is not heritable. Unfortunately, this conclusion was overly generalized to cover all acquired traits, thus putting a final end to Lamarckism. However, over the past few decades, there has been a renewed interest in Lamarckism, as evidence accumulates that ancestral environmental exposures can affect the progeny via the germline, and more strikingly, in some cases, the phenotypes developed in the offspring more or less recapitulate that directly induced by the environmental factors in the ancestors [1,2,17–24]. Although Generational EI has been observed in numerous species [23], murine models have contributed most extensively to our knowledge in this area. Mice have also contributed greatly to our knowledge about Mitotic EI, especially heritable gene silencing. Indeed, several classic examples of heritable gene silencing in mammals, namely X-inactivation, parental imprinting, and metastable epialleles (see section "Conventional Mouse EI Models" ), were all discovered in mice (between 1960s and 1990s) [25] Finally, since 1990s, after gene targeting in mice became feasible, many genes involved in epigenetic processes, including histone modifications, chromatin remodeling, and noncoding RNA metabolism, have been knocked out and their functions subsequently revealed, which is indispensable for defining the genetic basis of EI. In summary, mouse has historically served as a key model for exploring all three aspects of EI in mammals: Generational EI, Mitotic EI and the genetic basis of epigenetic control. We will outline the mouse models of Mitotic and Generational EI in the upcoming section.

CONVENTIONAL MOUSE EI MODELS For the most part, EI phenomena in mice might be classified roughly into three broad catagories. The first involves standard developmental processes, such as lineage differentiation, parental imprinting, and X-inactivation, which are regulated by elaborate developmental signals and executed by sophisticated epigenetic machinery, with well-defined biological functions. The

second type of EI phenomena are in conflict with the first. Here, the cells exposed to identical developmental signals or isogenic mice raised under identical conditions paradoxically show significant phenotypic variations. Such variations emerge randomly during normal development as a result of the noise inherent in cellular biological processes, and their physiological significance remain largely unclear. Finally, in the third type of EI phenomena, mice are experimentally challenged with environmental factors capable of fine-tuning or disrupting normal development, leading to long-lasting consequences, with obvious biological and clinical implications. The first type of EI phenomena has been extensively characterized and reviewed [26–29] (see also Chapters 18, 24, and 25), and so we will discuss only the second and third types (i.e., spontaneous random variation and experimentally induced variation), which involve noncanonical developmental processes and are poorly understood.

Spontaneous Random Epigenetic Variations We will classify the mouse models of such variations based on whether the variations are mitotically or generationally heritable. Mitotic EI Models In these models, epigenetic variations can occur at the cellular level within a mouse (variegated expression), or at the organismal level among individual mice (intangible variance). VARIEGATED EXPRESSION: DIFFERENTIAL GENE EXPRESSION IN THE CELLS OF THE SAME TYPE WITHIN A MOUSE

This is often observed in transgenic lines, where a transgene is silenced in a subset of cells of the same type [30,31]. Endogenous genes in genetically modified mice can also display variegated expression. For example, deleting a Cd8 enhancer or mutating the Cd4 silencer turns uniform CD8 and CD4 expression into variegated one [32,33]. Variegated expression at endogenous genes is clinically relevant, as exemplified in facioscapulohumeral dystrophy (FSHD), a genetic disease caused by contraction of the D4Z4 macrosatellite repeat array on chromosome 4 [34]. This contraction leads to ectopic, stochastic derepression of the transcription factor DUX4 in a small fraction of skeletal muscle cells, which causes apoptosis as well as other pathological changes in the muscle. Variegated expression has been most extensively studied at “metastable epialleles,” the mammalian alleles displaying labile epigenetic states [35,36]. Metastable epialleles are associated with retrotransposons that are young in evolutionary terms and so might

IV.  Model Organisms of Epigenetics



Conventional Mouse EI Models

235

FIGURE 15.3  Epigenetic phenomena at Avy. (A) Locus structure. Agouti signals hair follicular melanocytes to make yellow pigment instead of black. In WT mice, Agouti transcription is controlled by a promoter in exon 2 of the agouti gene (black arrow). This promoter is active only at a specific stage of hair growth, and the resultant transient AGOUTI expression causes only a subapical yellow band on each black hair. At Avy, a contraoriented intracisternal A particle (IAP) retrotransposon (broad red arrow) is inserted into the Agouti locus. A cryptic promoter in IAP drives ectopic agouti transcription(red arrow) and hence ectopic yellow coloration, but the promoter is subject to variable degrees of silencing, resulting in coat color variations. (B) Coat color spectrum in isogenic Avy mice of the same age and sex. The colors are linked to the CpG methylation status of the cryptic IAP promoter, with yellow being hypomethylated and brown fully methylated. The promoter activity can also differ among individual cells within the same mouse, producing epigenetically mosaic mouse with a mixture of yellow and brown hair (mottled). Of note, the yellow color is associated with obesity due to pleotropic effects of Agouti. (C) Generational EI of coat color. Avy females (circles) are bred with mice lacking functional agouti gene (not shown) to produce offspring. The result indicates that the coat colors are partially heritable. Furthermore, the effect lasts at least another generation (not shown). Interestingly, the father’s colors are not heritable (not shown). (D) Lasting consequences of maternal diet. Avy mice are exposed to excessive methyl donors (delivered through maternal diet) during development and lactation. The transient exposure leads to permanent shift of coat colors, which is associated with IAP CpG hypermethylation. Source: Part B, Reproduced from Jirtle, RL. The Agouti mouse: a biosensor for environmental epigenomics studies investigating the developmental origins of health and disease. Epigenomics 2014; 6: 447–450 [38]. Part C, Adapted from Rosenfeld C, Tollesfbol T, editors. Transgenerational epigenetic inheritance: evidence and debates. Elsevier; 2014: 123–145; Morgan H.D., Sutherland H.G., Martin D.I., Whitelaw E. Epigenetic inheritance at the agouti locus in the mouse. Nat Genet 1999; 23: 314–318 [60,62]. Part D, Adapted from Barouki R., Gluckman P.D., Grandjean P., Hanson M., Heindel J.J. Developmental origins of non-communicable disease: implications for research and public health. Environ Health 2012; 11:42; Waterland R.A., Jirtle R.L. Transposable elements: targets for early nutritional effects on epigenetic gene regulation. Mol Cell Biol 2003; 23:5293–5300 [70,72].

still possess some ability to escape silencing [25,37], as revealed at the prototypical metastable epiallele Agouti viable yellow (Avy) [38]. Avy carries such a retrotransposon inserted upstream of the Agouti transcription start site (Fig. 15.3A). A cryptic promoter at the retrotransposon causes aberrant Agouti transcription in the hair follicles, resulting in the yellow coat color. However, this promoter is silenced stochastically (associated with CpG methylation) in variable numbers of cells in the hair follicle, which is established before gastrulation and persists into adulthood (and even transmitted to the offspring via the eggs; see further). If the promoter is silenced in all the cells, the coat color becomes brown, but if only patches of the cells undergo silencing, the mice are epigenetically mosaic, displaying mottled colors resulting from a mixture of brown and yellow cells, which has become the poster boy of variegated expression because of its macroscopic visibility (Fig. 15.3B).

Interestingly, the degrees of silencing vary dramatically even among genetically identical littermates (Fig. 15.3B). Such differential gene expression among isogenic individuals is termed “variable expressivity” [36], a form of “intangible variance” (see following). INTANGIBLE VARIANCE: PHENOTYPIC VARIATION IN ISOGENIC LITTERMATES RAISED UNDER THE SAME CONDITION

For inbred mice raised under the standard conditions, many quantitative traits (such as body weight, size, behavior, and stress response) conform to a normal distribution (bell-shaped curve) despite tightly controlled environment and genetic background [25]. Thus, beside genotype and environment, a “third component” [39], named “intangible variance” [40], is also a key determinant of traits; in fact, intangible variance contributes to ∼80% of the variance in body

IV.  Model Organisms of Epigenetics

236

15.  Mouse Models of Epigenetic Inheritance: Classification, Mechanisms, and Experimental Strategies

weight [25,39]. Intangible variance is established by the 8-cell stage during embryogenesis, and then persists for life [39]. Multiple other cases of intangible variance have been reported, including regression of a random subset of embryos in utero [41], incomplete penetrance of phenotypes caused by mutations [42], and the aforementioned variable expressivity at Avy [38]. Variable expressivity has also been reported in transgenic lines. Indeed, shortly after the publication of the first transgenic line in 1980, it was already realized that transgene expression can vary not only among different founders due to random integration of the transgenes (i.e., position-effects), but also among different offsprings derived from the same founders, which cannot be explained by position effects [31]. The molecular basis of variable expressivity, as in the case of variegated expression, is stochastic, heritable gene silencing [25,35,36], and such silencing might also contribute to other forms of intangible variance. RELATIONSHIP BETWEEN VARIEGATED EXPRESSION AND VARIABLE EXPRESSIVITY

To our knowledge, metastable epialleles and transgenes that display variable expressivity always show variegated expressions. The reverse is true for metastable epialleles, but not for transgenes, as transgenic lines often display variegated expression without concomitant variable expressivity [30]. Furthermore, for endogenous genes, mutations at cis-regulatory elements can cause variegated expression without concomitant variable expressivity [32,33]. Thus, variegated expression seems a basic form of variation, whereas variable expressivity requires additional conditions. It is unclear why the two forms of variations are linked at metastable epialleles but not the other genes. This dichotomy is perhaps based on differences in the mechanisms of silencing. Metastable epiallele silencing occurs at the retrotransposons, and so reflect genome defense against invading parasites [43]. In contrast, transgene silencing results from adjacent heterochromatin (position effects) [44,45], repeats within transgene concatemers (repeat-induced gene silencing) [46–48], or prokaryotic vector sequences associated with the transgenes [49]. Note that of the three scenarios causing transgene silencing, the last one is reminiscent of that at metastable epialleles, as both involve foreign DNA. It would be interesting to determine whether the transgenes that display both variegated expression and variable expressivity are indeed silenced due to vector sequences. BIOLOGICAL SIGNIFICANCE OF PHENOTYPIC NOISE

Variegated expression and intangible variation are phenotypic noise rooted in stochasticity in gene expression [25]. Gene expression is intrinsically stochastic,

which arises from random fluctuations in transcription and translation due ultimately to the fact that biological processes are fundamentally driven by random collisions between macromolecules [50,51]. Noise is generally considered harmful, as it can cause death of cell or embryo and increase disease susceptibility [50]. Indeed, in normal mice, noise is actively and constantly buffered, as mutations have been identified that lead to its exacerbation. For example, mutations at the cis-acting elements turn the uniform CD4/CD8 expression into variegated one [32,33], and depletion of chromatin modifiers increases quantitative trait variations [52]. However, regulated noise and the resultant phenotypic diversification can also be beneficial by e.g., establishing heterogeneity in an initially homogenous cell population, which seems prerequisite for lineage differentiation [51,53]. Thus, noise may be biologically important [51,53,54]. Studies of variegated expression and intangible variation have illuminated, and should continue to illuminate, the mechanisms of phenotypic noise, with a wide range of implications. Generational EI Models In Generational EI, the epigenetic states established in the gametes inside the parents are (partially) transmitted to the offspring. Gametes are highly differentiated cells, and their epigenetic states must be reset shortly following fertilization (at the preimplantation stage) to confer totipotency on the early fetal cells, a prerequisite for their subsequent differentiation [55]. However, the resetting is incomplete, as demonstrated by the fact that CpG methylation at all imprinted and some nonimprinted genes are (partially) resistant to the reprogramming [56–59]. Thus, physiological mechanisms exist that enable genes to escape reprogramming, which makes Generational EI possible. Generational EI has been clearly observed at two metastable epialleles (Avy and AxinFu) [25,35,60] and four transgenes [61]. At Avy, the epigenotypes of the mothers shift the spectrum of epigenotypes of the pups: yellow mothers (with active Avy) produce 57% yellow, 43% mottled but no brown pups while brown mothers (with inactive Avy) produce 20% brown but only 40% yellow pups [62] (Fig. 15.3C). The four transgenic lines known to display naturally occurring Generational EI, E36 [63–65], 239B [66], TKZ751 [67,68], and Mta#7 [69], all carry transgenes that are expressed in the founders but heritably silenced in their offspring. The heritable silencing is established either in the germline of the founders, and/or in the germline of their children expressing genetic modifiers introduced by crossing the founders with breeders on different genetic backgrounds. The mechanisms of induction and inheritance of silencing at these transgenes are ill-defined, although CpG methylation is found associated with silencing [61].

IV.  Model Organisms of Epigenetics



Conventional Mouse EI Models

Experimentally Induced EI

237

Traditional toxicology and teratology teach us that environmental factorssuch as toxin and virus are harmful, and embryos are particularly vulnerable, as adverse gestational exposures can cause grotesque birth defects and even death of the fetus. However, over the past few decades, it has become clear that environmental factors can also act covertly, causing subtle yet lasting epigenetic defects that predispose the individual to later-onset diseases, which in humans can occur years and even decades after the exposures. Furthermore, the environmental factors can even impact the offspring for multiple generations via the germline [1,2,17–24]. This mode of action of environmental exposures may underlie the modern pandemic of many noncommunicable diseases, such as obesity, type 2 diabetes, and cardiovascular diseases [70]. Numerous mouse models have been established over the past few decades to study the long-term effects of environmental exposures. These models are highly diverse in terms of the environmental agents used, the timing of exposures, the phenotypes produced and the patterns of inheritance of phenotypes [17,19,24,60]. It is noteworthy that one of the earliest models, which is still popular today, reveals the ability of maternal dietary supplements (methyl donor) to shift the coat colors of the Avy pups [71,72] (Fig. 15.3D). Here we will focus on some crucial issues regarding the mechanisms underlying the induction and inheritance of epigenetic states, which are not usually discussed in the literature. For the sake of discussion, consider the “ideal” EI model depicted in Fig. 15.1. Here, the environmental insult induces similar epigenetic lesions in (some) somatic lineages and germline, the latter transmitted to the offspring for two generations. So far, no single animal model is known to completely conform to this hypothetical scenario, in part because testing this model would require the analysis of the epigenetic states of both the germline and the somatic cells for three generations (F1– F3), which has not been performed. However, incomprehensive analyses in diverse models suggests that some EI models do seem to fit at least portions of the “ideal” picture described in Fig. 15.1, and various models collectively would embody the entire picture, even though the evidence is largely indirect, as outlined below.

animals (F1 in Fig. 15.1), which reflect the direct effects of environmental factors on somatic cells, can be (partly) recapitulated in the F2 mice. Thus, in these models, the F1 and F2 somatic cells, and by inference, the F1 germline, presumably share some epigenetic lesions underlying the phenotypes. Indeed, prenatal viral immune activation (by injection of a viral mimic on Embryonic day 9 or E9) induces, in the ensuing adult males, behavioral abnormalities in three areas (sociability, cue fear expression, and sensorimotor gating), the first two abnormalities recapitulated in the F2 (and F3) mice [73]. Consistent with this, the expression of 2217 and 4015 genes in the amygdalar complex is altered in the F1 and F2 males, with a remarkable overlap of 1132 genes (although oddly the expression of some of these 1132 genes are changed in the opposite directions). Behavioral defects induced by postnatal stress are also (partially) heritable [74–76]. Similar to behavioral defects, some metabolic defects can be passed onto the offspring. For example, feeding young male mice with high fat diet (from 5 weeks of age for 10 weeks) induces obesity, which is recapitulated in the children (and to less extents grandchildren) [77], and a recent study using in vitro fertilization (IVF) demonstrates that such heritable effects are mediated exclusively by the gametes from obese parents [78]. The metabolic effects of poor nutrition are also partially heritable: maternal caloric restriction during gestation causes three defects in the pups (low birth weight, obesity and glucose intolerance), the first two transmitted to their offspring via the paternal line [78,79]. These data beg the question: how could environmental insults paradoxically cause similar epigenetic changes in somatic cells and the germline? To address this, it is necessary to identify primary epigenetic lesions in the somatic cells and germline. This is challenging because environmental factors are pleotropic, affecting thousands of genes most of which are presumably secondary or even tertiary targets. Further complicating the analysis, endogenous genes are often large in size and regulated by complex and poorly defined mechanisms. A mouse model where epigenetic lesions can be selectively targeted to a simple reporter gene in both somatic cells and germline would help confront these challenges (see section "Targeting Epigenetic Modifications to Reporter Genes: A Unique EI Model Tractable for Mechanistic Studies").

Induction of Epigenetic Lesion: Soma Versus Germ Perhaps the most striking aspect of the “ideal” scenario is that an environmental insult could inflict comparable epigenetic lesion on the somatic cells and germline. This is striking partly because of the dramatic differences in the ways the epigenetic states are established and maintained in the somatic cells versus germline (especially the male germline). However, many studies demonstrate that the phenotypes in directly exposed

Propagation of Epigenetic Lesion Across Generations: Mode of Inheritance and Determinant of Durability The epigenetic lesions in our hypothetic EI model can be transmitted by cis- or trans- mechanisms. Both mechanisms have been implicated in Generational EI models, but significant knowledge gaps and caveats exist. CpG methylation is perhaps the first assay one would do to analyze the epigenetic states, and indeed,

IV.  Model Organisms of Epigenetics

238

15.  Mouse Models of Epigenetic Inheritance: Classification, Mechanisms, and Experimental Strategies

where examined, aberrant methylation is often linked to the phenotypes. However, to our knowledge, in no experimentally induced Generational EI models, has CpG methylation been clearly shown to be the primary, generationally heritable epigenetic lesion (i.e., the cis-signal), and in some cases, the methylation indeed turns out to be a secondary lesion. For example, early postnatal life trauma (via maternal separation) alters DNA methylation in several candidate genes in the sperm of the separated males, and comparable changes are present in the brain of the offspring [74]. However, the aberrant methylation is not the cis-signal because the penetrance of the phenotype does not decline in the outcross experiments [74], indicating that a transmechanism is at work. Indeed, the same group subsequently found that the trauma also alters the sperm microRNA (miRNAs) in the exposed males, and injecting total RNA from these sperm into normal zygotes recapitulates the effects of trauma in the ensuing adult [80], suggesting the aberrant methylation is secondary to abnormal RNAs. In contrast to the uncertainty surrounding the cissignals, RNAs can clearly serve as the trans-signal for intergenerational EI from directly exposed individuals to the immediate offspring (F1 to F2 in Fig. 15.1). Specifically, environmental exposures such as stress and diet can alter RNA contents in the sperm, and injecting relevant RNAs from such sperm into normal fertilized eggs can reproduce the effects of environmental exposures; the RNAs injected include total RNA as aforementioned [80], a set of nine miRNAs [81] and tRNA-derived small RNAs [82,83]. These miRNAs and tRNA fragments are able to degrade stored maternal mRNAs and to regulate endogenous retroelements in the early zygotes, respectively, which presumably alter the developmental trajectory of the embryos, leading ultimately to adult phenotypes [81,82] Sperm RNAs delivered to the eggs are minute in quantities, and mammals do not have RNA-dependent RNA polymerase to replicate RNA as do lower organisms. Thus, as the eggs divide, the paternal RNAs rapidly disappear. Indeed, while early life stress alters miRNAs in the sperm of the exposed males (F1 in Fig. 15.1), the defect is absent from F2 sperm [80]. Paradoxically, in this model, the F2 sperm can pass on the behavior phenotype to F3, suggesting that the aberrant RNAs in F1 sperm have relayed the information to unknown carriers in the F2 sperm [80]. Sperm RNAs might use another strategy to affect F3, as illustrated by the following study [84]. Specifically, injecting a single miRNA (miR-1) into fertilized eggs suffices to cause cardiac hypertrophy in 90% of the ensuing adults (F1 in Fig. 15.1), which is heritable for at least 3 generations (from F2 to F4) with no decline in penetrance. Amazingly, miR-1 is elevated in both F2 and F3 sperm (F1 sperm somehow not tested), suggesting that the miR-1 injected into the eggs, which should

rapidly vanish, is regenerated in the F1, F2, and F3 sperm presumably via an autoregulatory loop (Fig. 15.2B) [84]. Of note, although it is clear that RNAs can be sufficient to replace the environmental factors in triggering the phenotypes, their necessity has yet to be shown; this would require depleting the RNA in sperm in the exposed mice. In addition to the nature of the signals directing the inheritance of epigenetic lesions, another mystery surrounding these lesions concerns their durability. In our hypothetical scenario, the epigenetic lesion in the F1 germline can be faithfully transmitted to F2 and F3. This is certainly not generally true. For example, prenatal exposure to environmental toxicans [on E8-20 during primordial germ cell (PGC) reprograming] induces multiple alterations in transcription and methylation profiles in the ensuing prospermatogonia, but none of these changes are inherited by the prospermatogonia of the immediate offspring [85]. The failure to inherit some epigenetic lesions is also consistent with the observation that often, only a subset of the environmentally induced phenotypes is generationally heritable. So what determine the durability of epigenetic lesions? Our lack of knowledge here is glaring. This issue is hard to address in part due to the pleotropic nature of environmental factors and complexity of endogenous genes. Further complicating the problem, the pleotropic epigenetic alterations in conventional EI models can trigger secondary DNA mutations [86]. These limitations again calls for mouse models allowing for selective induction of epigenetic lesions at simple target genes.

TARGETING EPIGENETIC MODIFICATIONS TO REPORTER GENES: A UNIQUE EI MODEL TRACTABLE FOR MECHANISTIC STUDIES To overcome the aforementioned limitations in conventional EI models, we have established mouse models where heritable chromatin states can be pharmacologically induced at simple reporter genes in both somatic cells and the germline [87]. Our models, quite consistent with the “ideal” model depicted in Fig. 15.1, greatly facilitates dissection of various parameters potentially influencing the establishment and propagation of epigenetic lesions, including timing of exposure, nature of epigenetic modifications, and genomic location of the target genes. These models are based on the well-defined tetracycline (tet)-sensitive gene regulatory system. Initially, we used mice ubiquitously expressing the reverse tetregulated transcription activator (rtTA) and carrying a transgene bearing its cognate DNA binding sites (tet-O) upstream of the human CMV minimal promoter and GFP (Fig. 15.4A); the transgene is inserted as a singlecopy gene into the Col1a1 locus via flippase-catalyzed

IV.  Model Organisms of Epigenetics



Targeting Epigenetic Modifications to Reporter Genes: A Unique EI Model Tractable for Mechanistic Studies

239

FIGURE 15.4  Epigenetic phenomenon at the tetO-CMV-GFP reporter gene. (A) The experimental system consists of Dox-activated transcription factor rtTA, which is ubiquitously expressed from a transgene inserted into the Rosa26 locus, and a simple reporter gene bearing rtTA binding site (TetO) upstream of the CMV minimal promoter and GFP, which is inserted into the Col1a1 locus. Depicted is the scenario in adult mice, where Dox reversibly induces GFP. (B) Fetal Dox exposure leads to widespread GFP silencing in the ensuing adults, which is transgenerationally heritable via the female germline. Prenatally exposed females (F1) are mated with naive, nontransgenic males to generate F2. F2 females are in turn mated with control males to generate F3. Adult mice are challenged with Dox before imaging. Of note, brain is not included in the photographs because Dox can not induce GFP in adult brain even in the control mice, which is due to blood-brain barrier in adults. However, the transgene in the brain from prenatally exposed mice is highly methylated, suggesting it is also silenced [87]. (C) EI pattern at the tetO-CMV-GFP reporter. Fetal Dox exposure induces widespread epigenetic silencing, but for simplicity, only the brain and the germline are shown affected. The silencing can be transmitted to F5 (unpublished). (D) Effects of rtTA on the tetO-CMV-GFP transgene in ES cells expressing rtTA. Cells are cultured in the presence of Dox for various days before analysis. The top panel displays micrographs of ES cells following 2 (left) and 15 (right) days of Dox exposure. The bottom panel is flow cytometrical quantification of GFP expression, with Dox stimulation times and percentage of GFP-expressing cells indicated at left and right, respectively. (E) Fetal Dox exposure facilitates CD4 induction in the ensuing adult mice. (Top) CD4 transcription is controlled by enhancer (E), promoter (P) and silencer (S). tetO (red box) is inserted into the Cd4 locus so that Cd4 expression is subject to regulation by Dox. The numbers denote the distance from the transcription start site (arrow). (Bottom) Flow cytometrical quantification of CD4 expression in CD8 lymphocytes in mice expressing rtTA. CD4 is normally repressed in CD8 cells. Dox administration (via drinking water) induces CD4, but slowly and inefficiently, with only 46% CD8 cells expressing CD4 following even prolonged (17 days) induction (left panel). Prenatal Dox exposure (E0-20) dramatically increases both the kinetics and the extent of CD4 induction: the proportion of CD4-expressing CD8 cells already reaches 53% within 2 days of induction, and by Day 17, the majority of CD8 cells (85%) are expressing CD4 (right panel). Source: Parts B, D, and E, Reproduced from Wan M, et al. Inducible mouse models illuminate parameters influencing epigenetic inheritance. Development 2013; 140: 843–852 [87].

site-specific recombination [88]. rtTA binds and activates target genes only in the presence of tet or its derivative doxycycline (Dox). The mice were created in the Jaenisch lab, who demonstrated that Dox administration (via drinking water) activates GFP expression in adult mice [88]. Our goal was to determine whether transient Dox preexposure in adults can facilitate the second round of Dox induction later. We observed no such epigenetic memory. Unexpectedly, when the mice were preexposed to Dox during fetal development, the gene became completely refractory to Dox challenge in the ensuing adults, and furthermore, the refractory state was maternally

transmitted to ∼30% of offspring for at least two generations (Figs. 15.4B–C). The silencing is associated with DNA hypermethylation and loss of an activating histone mark (H3K4me2) at the CMV promoter in the somatic cells in the directly exposed mothers (F1) and their offspring (F2 and F3), raising the possibility that similar epigenetic lesions have been induced in the F1 somatic cells and eggs and transmitted to F2 eggs. Furthermore, the silenced transgene can not affect the “naïve” transgene, demonstrating that the inheritance is not mediated by a trans signal. Thus, fetal Dox exposure causes transgenerationally heritable chromatin marking at this simple,

IV.  Model Organisms of Epigenetics

240

15.  Mouse Models of Epigenetic Inheritance: Classification, Mechanisms, and Experimental Strategies

single-copy reporter gene. To our knowledge, this is the first time a cis-acting mechanism is clearly shown to underlie the inheritance in experimentally induced Generational EI models. How rtTA paradoxically triggers epigenetic silencing is unclear, but the phenomenon was seen only in fetus and ES cells, where the silencing was preceded by GFP induction, suggesting it is an indirect effect of rtTA activation that requires tissue-specific repressors [87] (Fig. 15.4D). We then used this model to define the timing of establishment of the silenced states. We exposed the mice to Dox at various stages of fetal development. Dox exposure for the first 4 days was critical but insufficient for producing transgenerationally heritable silencing, but exposure for 10 days was partially sufficient. Since epigenetic reprogramming in PGCs begins around E8 in mice, the epigenetic silencing must have been largely established before the onset of this reprogramming, and once established, it must be partially resistant to the reprogramming. We have thus identified E0-10 as an essential and partially sufficient time window to induce transgenerationally heritable epigenetic lesion, which has never been reported for other EI models. To dissect other parameters influencing Generational EI, we inserted into the Col1a1 locus a very different Dox-responsive transgene, which contains the CD4 regulatory elements (promoter/enhancer/silencer) instead of the CMV minimal promoter. Despite the genetic differences, fetal Dox exposure similarly produced a transgenerationally heritable phenotype, but of the opposite nature: facilitation of GFP induction associated with activating histone modifications (H3K9a and H3K4me2). This is the first demonstration that activating chromatin marks could be transgenerationally heritable in mammals. Since both activating and repressive perturbations are heritable at the Col1a1 locus, and since the two transgenes are genetically distinct, the Col1a1 locus (rather than the nature of epigenetic modifications or the DNA sequences of the transgenes) seems the key determinant of Generational EI. In support of this, fetal Dox exposure also caused mitotically heritable silencing of a randomly integrated CMV promoter (not shown) and triggered mitotically heritable activation of a Dox-responsive endogenous Cd4 allele (Fig. 15.4E), but neither effect was intergenerationally heritable. These data also demonstrate that fetal epigenome is extremely malleable, where epigenetic perturbations, once induced at proper times during embryogenesis, can generally be transmitted to adults regardless of the nature or location of the perturbations. Consistent with this, a Dox-regulated transcription repressor (tTS) can cause mitotically stable silencing at two distinct, randomly integrated promoters if and only if the repressor acts during the first few days of fetal development [89]. Thus, mitotically heritable epigenetic perturbations can be readily established as long as the

inducers act at the critical times during embryogenesis, whereas Generational EI additionally requires special locus environment. What is special about the Col1a1 locus? Our current hypothesis is that the locus harbors insulator-like elements that protect epigenetic modifications (whether silencing or activating) from reprogramming enzymes and efforts are underway to identify the putative elements. In summary, we have established a unique approach to study EI, which has already provided insights into the mechanisms of induction and durability of cis-epigenetic signals that are hard to obtain with conventional EI models. Clearly, this approach is also applicable to transsignals.

CONCLUSIONS Ever increasing numbers of mouse models have been established to probe long-term effects of environmental factors on human health and to illuminate basic EI mechanisms, but many crucial issues remain, a few of them discussed below. First, environmental factors (mainly psychological stress and nutritional imbalance), even when acting postnatally, can apparently produce similar epigenetic changes in the soma and the germline, despite the dramatic differences between the two. Psychological stress and nutritional imbalance are not evolutionary novel, and so the organisms may have evolved mechanisms not only to confront these challenges themselves, but also to prepare the offspring for the same challenges predicted to befall them. Thus, the generational EI in such cases may be part of a physiological adaptive response (predictive adaptive response), that is hardwired in the genome [90]. It will be of great interest to decipher the mechanisms of such a response, which, to Weismann, must be “immensely complex, nay! altogether inconceivable” [16]. However, given the rapid progress in the field, it is not inconceivable that the mechanisms will be revealed in the near future, which would surely make both Weismann and Lamarck roll in their graves. Second, the parameters governing the induction and transmission of cis- or trans-epigenetic signals are poorly understood, partly due to the complexity of the endogenous genes and confounding secondary and parallel epigenetic lesions that presumably coexist with the primary epigenetic signals in conventional EI models. Mouse models with targeted epigenetic lesions would be invaluable here. Finally, little is known about the genetic basis of EI, including how genetic backgrounds affect EI, an issue of medical importance given the genetic diversity of humans. Large-scale screens using ENU have successfully uncovered multiple genetic modifiers of epigenetic

IV.  Model Organisms of Epigenetics



REFERENCES

reprograming [25,41,91–97], but this approach is laborintensive and costly. An alternative strategy would be to use mosaic analysis, which requires much fewer mice. Such screens may be feasible given the availability of lentiviral sgRNAs libraries and Cas9-expressing mice.

Glossary Epigenetic inheritance (EI) signals  Signals responsible for EI, based on chemical modifications of DNA and/or histone (cis-signal), or on diffusible molecules, such as RNAs (trans-signal). Epigenetic inheritance  Transmission of epigenetic information through cell divisions. Epigenetics  The study of mitotically and/or meiotically heritable changes in gene function that cannot be explained by changes in DNA sequence. Generational EI, intergenerational EI, transgenerational EI There seems no consensus on their definitions. We propose, based in part on literature survey, to define the three terms as “germline transmission of epigenetic information between generations,” “germline transmission of epigenetic information from an environmentally exposed organism to the immediate offspring,” and “the germline transmission of epigenetic information between generations in the absence of any environmental exposure,” respectively (see main text). Intangible variance Phenotypic variation among genetically identical individuals raised under the same condition (e.g., littermates of inbred mice). Variable expressivity is a form of intangible variance. Metastable epialleles  Epigenetically labile alleles displaying variegated expression and variable expressivity due to variable epigenetic modifications, which are stochastically established during early development. Mitotic EI  Transmission of epigenetic information from parental cells to daughter cells through mitosis. Variable expressivity  Variable expression of a gene between genetically identical individuals. Variegated expression  Variable expression of a gene in a single cell type within the same individual.

References [1] Skinner MK. Environmental epigenetic transgenerational inheritance and somatic epigenetic mitotic stability. Epigenetics 2011;6:838–42. [2] Tollefsbol T. . In: Tollesfbol T, editor. Transgenerational epigenetic inheritance: evidence and debates. Boston, MA: Elsevier; 2014. p. 1–8. [3] Skinner MK. What is an epigenetic transgenerational phenotype? F3 or F2. Reprod Toxicol 2008;25:2–6. [4] Riggs A, Porter T. Overview of epigenetic mechanisms. In: Russo VEA, Martienssen R, Riggs AD, editors. Epigenetic mechanisms of gene regulation. Cold Spring Harbor, NY: Cold Spring Harbor Laboratory Press; 1996. p. 29–45. [5] Felsenfeld G. A brief history of epigenetics. Cold Spring Harb Perspect Biol 2014;6:a018200. [6] Hitchins MP, et al. Dominantly inherited constitutional epigenetic silencing of MLH1 in a cancer-affected family is linked to a single nucleotide variant within the 5′UTR. Cancer Cell 2011;20:200–13. [7] Bonasio R, Tu S, Reinberg D. Molecular signals of epigenetic states. Science 2010;330:612–6. [8] Beckers J, Wurst W, de Angelis MH. Towards better mouse models: enhanced genotypes, systemic phenotyping and envirotype modelling. Nat Rev Genet 2009;10:371–80.

241

[9] van der Weyden L, White JK, Adams DJ, Logan DW. The mouse genetics toolkit: revealing function and mechanism. Genome Biol 2011;12:224. [10] Kaczmarczyk L, Jackson WS. Astonishing advances in mouse genetic tools for biomedical research. Swiss Med Wkly 2015;145:w14186. [11] Singh P, Schimenti JC, Bolcun-Filas E. A Mouse geneticist’s practical guide to CRISPR applications. Genetics 2015;199:1–15. [12] Carlson CM, et al. Transposon mutagenesis of the mouse germline. Genetics 2003;165:243–56. [13] Ding S, et al. Efficient transposition of the piggyBac (PB) transposon in mammalian cells and mice. Cell 2005;122:473–83. [14] Dupuy AJ, Akagi K, Largaespada DA, Copeland NG, Jenkins NA. Mammalian mutagenesis using a highly mobile somatic sleeping beauty transposon system. Nature 2005;436:221–6. [15] Kile BT, Hilton DJ. The art and design of genetic screens: mouse. Nat Rev Genet 2005;6:557–67. [16] Weismann, A. The Supposed Transmission of Mutilations. A lecture delivered at the Meeting of the Association of German Naturalists at Cologne, September 1888; 1888. [17] Bohacek J, Mansuy IM. Molecular insights into transgenerational non-genetic inheritance of acquired behaviours. Nat Rev Genet 2015;16:641–52. [18] Burggren W. Epigenetic inheritance and its role in evolutionary biology: re-evaluation and new perspectives. Biology 2016;5:24. [19] Hanson MA, Skinner MK. Developmental origins of epigenetic transgenerational inheritance. Environ Epigenet 2016;2:dvw002. [20] Heard E, Martienssen RA. Transgenerational epigenetic inheritance: myths and mechanisms. Cell 2014;157:95–109. [21] Jablonka E. Epigenetic inheritance and evolution The Lamarckian dimension. New York: Oxford University Press; 1999. [22] Jablonka E. Epigenetic inheritance and plasticity: the responsive germline. Prog Biophys Mol Biol 2013;111(2–3):99–107. [23] Jablonka E, Raz G. Transgenerational epigenetic inheritance: prevalence, mechanisms, and implications for the study of heredity and evolution. Q Rev Biol 2009;84:131–76. [24] Rando OJ, Simmons RA. I’m eating for two: parental dietary effects on offspring metabolism. Cell 2015;161:93–105. [25] Blewitt M, Whitelaw E. The use of mouse models to study epigenetics. Cold Spring Harb Perspect Biol 2013;5:a017939. [26] Lee JT, Bartolomei MS. X-Inactivation, imprinting, and long noncoding RNAs in health and disease. Cell 2013;152:1308–23. [27] Barlow DP, Bartolomei MS. Genomic imprinting in mammals. Cold Spring Harb Perspect Biol 2014;6:a018382. [28] Boland MJ, Nazor KL, Loring JF. Epigenetic regulation of pluripotency and differentiation. Circ Res 2014;115:311–24. [29] Brockdorff N, Turner BM. Dosage compensation in mammals. Cold Spring Harb Perspect Biol 2015;7:a019406. [30] Martin DI, Whitelaw E. The vagaries of variegating transgenes. Bioessays 1996;18:919–23. [31] Palmiter RD, Brinster RL. Germ-line transformation of mice. Annu Rev Genet 1986;20:465–99. [32] Garefalaki A, et al. Variegated expression of CD8 alpha resulting from in situ deletion of regulatory sequences. Immunity 2002;16:635–47. [33] Taniuchi I, Sunshine MJ, Festenstein R, Littman DR. Evidence for distinct CD4 silencer functions at different stages of thymocyte differentiation. Mol Cell 2002;10:1083–96. [34] Tawil R, van der Maarel SM, Tapscott SJ. Facioscapulohumeral dystrophy: the path to consensus on pathophysiology. Skelet Muscle 2014;4:12. [35] Kellermayer R. . In: Tollesfbol T, editor. Transgenerational epigenetic inheritance: evidence and debates. Elsevier; 2014. p. 69–73. [36] Rakyan VK, Blewitt ME, Druker R, Preis JI, Whitelaw E. Metastable epialleles in mammals. Trends Genet 2002;18:348–51.

IV.  Model Organisms of Epigenetics

242

15.  Mouse Models of Epigenetic Inheritance: Classification, Mechanisms, and Experimental Strategies

[37] Whitelaw E, Martin DI. Retrotransposons as epigenetic mediators of phenotypic variation in mammals. Nat Genet 2001;27:361–5. [38] Jirtle RL. The Agouti mouse: a biosensor for environmental epigenomics studies investigating the developmental origins of health and disease. Epigenomics 2014;6:447–50. [39] Gärtner K. A third component causing random variability beside environment and genotype. A reason for the limited success of a 30 year long effort to standardize laboratory animals? Int J Epidemiol 2012;41:335–41. [40] Falconer D. Introduction to quantitative genetics. 3rd ed. New York, NY: Wiley; 1989. [41] Blewitt ME, et al. An N-ethyl-N-nitrosourea screen for genes involved in variegation in the mouse. Proc Natl Acad Sci USA 2005;102:7629–34. [42] Biben C, et al. Cardiac septal and valvular dysmorphogenesis in mice heterozygous for mutations in the homeobox gene Nkx2-5. Circ Res 2000;87:888–95. [43] Crichton JH, Dunican DS, MacLennan M, Meehan RR, Adams IR. Defending the genome from the enemy within: mechanisms of retrotransposon suppression in the mouse germline. Cell Mol Life Sci 2014;71:1581–605. [44] Liu C. Strategies for designing transgenic DNA constructs. Methods Mol Biol 2013;1027:183–201. [45] Wilson C, Bellen HJ, Gehring WJ. Position effects on eukaryotic gene expression. Annu Rev Cell Biol 1990;6:679–714. [46] Garrick D, Fiering S, Martin DI, Whitelaw E. Repeat-induced gene silencing in mammals. Nat Genet 1998;18:56–9. [47] Henikoff S. Conspiracy of silence among repeated transgenes. Bioessays 1998;20:532–5. [48] McBurney M. Evidence for repeat-induced gene silencing in cultured mammalian cells: inactivation of tandem repeats of transfected genes. Exp Cell Res 2002;274:1–8. [49] Kjer-Nielsen L, Holmberg K, Perera JD, McCluskey J. Impaired expression of chimaeric major histocompatibility complex transgenes associated with plasmid sequences. Transgenic Res 1992;1:182–7. [50] Kaern M, Elston TC, Blake WJ, Collins JJ. Stochasticity in gene expression: from theories to phenotypes. Nat Rev Genet 2005;6:451–64. [51] Pujadas E, Feinberg AP. Regulated noise in the epigenetic landscape of development and disease. Cell 2012;148:1123–31. [52] Whitelaw NC, et al. Reduced levels of two modifiers of epigenetic gene silencing, Dnmt3a and Trim28, cause increased phenotypic noise. Genome Biol 2010;11:R111. [53] Whitelaw NC, Chong S, Whitelaw E. Tuning in to noise: epigenetics and intangible variation. Dev Cell 2010;19:649–50. [54] Eldar A, Elowitz MB. Functional roles for noise in genetic circuits. Nature 2010;467:167–73. [55] Feng S, Jacobsen SE, Reik W. Epigenetic reprogramming in plant and animal development. Science 2010;330:622–7. [56] Borgel J, et al. Targets and dynamics of promoter DNA methylation during early mouse development. Nature 2010;42:1093–100. [57] Guibert S, Forné T, Weber M. Global profiling of DNA methylation erasure in mouse primordial germ cells. Genome Res 2012;22:633–41. [58] Hackett JA, et al. Germline DNA demethylation dynamics and imprint erasure through 5-hydroxymethylcytosine. Science 2013;339:448–52. [59] Smallwood SA, et al. Dynamic CpG island methylation landscape in oocytes and preimplantation embryos. Nature 2011;43:811–4. [60] Rosenfeld C, Tollesfbol T, editors. Transgenerational epigenetic inheritance: evidence and debates. Boston, MA: Elsevier; 2014. p. 123–45. [61] Kaundal R, Yang Y, Nottoli T, Chi T. . In: Tollesfbol T, editor. Transgenerational epigenetic inheritance: evidence and debates. Boston, MA: Elsevier; 2014. p. 75–83.

[62] Morgan HD, Sutherland HG, Martin DI, Whitelaw E. Epigenetic inheritance at the agouti locus in the mouse. Nat Genet 1999;23:314–8. [63] Hadchouel M, Farza H, Simon D, Tiollais P, Pourcel C. Maternal inhibition of hepatitis B surface antigen gene expression in transgenic mice correlates with de novo methylation. Nature 1987;329:454–6. [64] Pourcel C, Tiollais P, Farza H. Transcription of the S gene in transgenic mice is associated with hypomethylation at specific sites and with DNase I sensitivity. J Virol 1990;64:931–5. [65] Schweizer J, Valenza-Schaerly P, Goret F, Pourcel C. Control of expression and methylation of a hepatitis B virus transgene by strain-specific modifiers. DNA Cell Biol 1998;17:427–35. [66] Kearns M, Preis J, McDonald M, Morris C, Whitelaw E. Complex patterns of inheritance of an imprinted murine transgene suggest incomplete germline erasure. Nucleic Acids Res 2000;28:3301–9. [67] Allen ND, Norris ML, Surani MA. Epigenetic control of transgene expression and imprinting by genotype-specific modifiers. Cell 1990;61:853–61. [68] Pickard B, et al. Epigenetic targeting in the mouse zygote marks DNA for later methylation: a mechanism for maternal effects in development. Mech Dev 2001;103:35–47. [69] Sutherland HG, et al. Reactivation of heritably silenced gene expression in mice. Mamm Genome 2000;11:347–55. [70] Barouki R, Gluckman PD, Grandjean P, Hanson M, Heindel JJ. Developmental origins of non-communicable disease: implications for research and public health. Environ Health 2012;11:42. [71] Wolff GL, Kodell RL, Moore SR, Cooney CA. Maternal epigenetics and methyl supplements affect agouti gene expression in Avy/a mice. FASEB J 1998;12:949–57. [72] Waterland RA, Jirtle RL. Transposable elements: targets for early nutritional effects on epigenetic gene regulation. Mol Cell Biol 2003;23:5293–300. [73] Weber-Stadlbauer U, et al. Transgenerational transmission and modification of pathological traits induced by prenatal immune activation. Mol Psychiatry 2016;22(1):102–12. [74] Franklin TB, et al. Epigenetic transmission of the impact of early stress across generations. Biol Psychiatry 2010;68:408–15. [75] Rodgers AB, Morgan CP, Bronson SL, Revello S, Bale TL. Paternal stress exposure alters sperm MicroRNA content and reprograms offspring HPA stress axis regulation. J Neurosci 2013;33:9003–12. [76] Saavedra-Rodríguez L, Feig LA. Chronic Social Instability Induces Anxiety and Defective Social Interactions Across Generations. Biol Psychiatry 2013;73:44–53. [77] Fullston T, et al. Paternal obesity initiates metabolic disturbances in two generations of mice with incomplete penetrance to the F2 generation and alters the transcriptional profile of testis and sperm microRNA content. FASEB J 2013;27:4226–43. [78] Huypens P, et al. Epigenetic germline inheritance of diet-induced obesity and insulin resistance. Nat Genet 2016;48:497–9. [79] Jimenez-Chillaron JC, et al. Intergenerational Transmission of Glucose Intolerance and Obesity by In Utero Undernutrition in Mice. Diabetes 2009;58:460–8. [80] Gapp K, et al. Implication of sperm RNAs in transgenerational inheritance of the effects of early trauma in mice. Nat Neurosci 2014;17:667–9. [81] Rodgers AB, Morgan CP, Leu NA, Bale TL. Transgenerational epigenetic programming via sperm microRNA recapitulates effects of paternal stress. Proc Natl Acad Sci USA 2015;112:13699– 704. [82] Sharma U, et al. Biogenesis and function of tRNA fragments during sperm maturation and fertilization in mammals. Science 2016;351:391–6. [83] Chen Q, et al. Sperm tsRNAs contribute to intergenerational inheritance of an acquired metabolic disorder. Science 2015;351(6271):397–400.

IV.  Model Organisms of Epigenetics



243

REFERENCES

[84] Wagner KD, et al. RNA induction and inheritance of epigenetic cardiac hypertrophy in the mouse. Dev Cell 2008;14:962–9. [85] Iqbal K, et al. Deleterious effects of endocrine disruptors are corrected in the mammalian germline by epigenome reprogramming. Genome Biol 2015;16:59. [86] Guerrero-Bosagna C, Skinner MK. Environmentally induced epigenetic transgenerational inheritance of phenotype and disease. Mol Cell Endocrinol 2011;354(1–2):3–8. [87] Wan M, et al. Inducible mouse models illuminate parameters influencing epigenetic inheritance. Development 2013;140:843–52. [88] Beard C, Hochedlinger K, Plath K, Wutz A, Jaenisch R. Efficient method to generate single-copy transgenic mice by site-specific integration in embryonic stem cells. Genesis 2006;44:23–8. [89] Wiznerowicz M, et al. The Kruppel-associated box repressor domain can trigger de novo promoter methylation during mouse early embryogenesis. J Biol Chem 2007;282:34535–41. [90] Hanson MA, Gluckman PD. Early developmental conditioning of later health and disease: physiology or pathophysiology? Physiol Rev 2014;94:1027–76. [91] Bultman SJ, Gebuhr TC, Magnuson T. A Brg1 mutation that uncouples ATPase activity from chromatin remodeling reveals an

[92]

[93] [94]

[95]

[96]

[97]

essential role for SWI/SNF-related complexes in beta-globin expression and erythroid development. Genes Dev 2005;19:2849–61. Percec I, Plenge RM, Nadeau JH, Bartolomei MS, Willard HF. Autosomal dominant mutations affecting X inactivation choice in the mouse. Science 2002;296:1136–9. Percec I, et al. An N-ethyl-N-nitrosourea mutagenesis screen for epigenetic mutations in the mouse. Genetics 2003;164:1481–94. Schumacher A, Faust C, Magnuson T. Positional cloning of a global regulator of anterior-posterior patterning in mice. Nature 1996;384:648. Wilson L, et al. Random mutagenesis of proximal mouse chromosome 5 uncovers predominantly embryonic lethal mutations. Genome Res 2005;15:1095–105. Jun JE, et al. Identifying the MAGUK protein Carma-1 as a central regulator of humoral immune responses and atopy by genome-wide mouse mutagenesis. Immunity 2003;18:751–62. Carpinelli MR, et al. Suppressor screen in Mpl-/- mice: c-Myb mutation causes supraphysiological production of platelets in the absence of thrombopoietin signaling. Proc Natl Acad Sci USA 2004;101:6553–8.

IV.  Model Organisms of Epigenetics

C H A P T E R

16 Plant Epigenomics Venu (Kal) Kalavacharla, Mayavan Subramani, Vasudevan Ayyappan, Mollee C. Dworkin, Rita K. Hayford Center for Integrated Biological & Environmental Research (CIBER), Delaware State University, Dover, DE, United States

O U T L I N E Introduction

245

DNA Methylation Functions and Mechanisms of DNA Methylation Interactions With Other Epigenetic Factors Methods To Interrogate DNA Methylation Stability, Heritability, and Agricultural Significance

246 246 247 247 247

Histone Modifications in Plants 248 Linker Histones in Plants 248 Types of Histone Modifications 249 Mechanism of Action of Histone Modifying Enzymes 249 Role of Histone Modifying Enzymes in Developmental Processes 249 Regulation of Gene Expression by Histone Modifying Enzymes During Biotic and Abiotic Stress 249

INTRODUCTION Epigenomics focuses on the analysis of global epigenetic changes that provides important clues regarding mechanisms and function of gene regulation across many genes in a cell or organism. Epigenomic approaches have drawn major attention in the present “postgenomics era” to unravel mechanisms involved in gene regulation of an organism. These mechanisms have fascinated scientists for many years and which do not necessarily follow Mendelian rules [1]. Plants are an ideal system for epigenomics research. Discoveries made in plant genomics have opened new avenues to explore epigenomics. Features of plant epigenomics are shared with other organisms but with some unique features that may help explore this in further detail. These advantages include Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00016-X Copyright © 2017 Elsevier Inc. All rights reserved.

Crosstalk of Histone Modification and DNA Methylation

250

Role of Epigenetic Marks in Plant Development

250

Histone Modifications and Gene Expression in Response to Abiotic and Biotic Stress

251

Epigenome Editing—A New Approach for Crop Improvement Tools for Epigenome Editing Epigenome Editing Using Genome Editing Tools

252 252 252

Future Perspectives

253

References

254

for most plants a relatively quick generation cycle, ease of imparting abiotic, and biotic stresses, production of seed/clones that can serve to implement various experimental conditions, and finally ease of growth in varied environments. Like in animals, plant miRNAs and siRNAs are involved in regulating gene expression via histone modification and DNA methylation. However, unlike animals, DNA methylation responsible for epigenetic inheritance in plants is regained after fertilization [2]. Available whole genome sequences from model plants have facilitated the interpretation of genomewide histone modifications, DNA methylation, and their relation to coding and noncoding RNAs. Development of genome editing tools, such as Zinc-finger nucleases (ZFNs), transcription activator-like effectors nucleases (TALENs) and the CRISPR-Cas9 system have opened

245

246

16.  Plant Epigenomics

new possibilities for the targeted modification of histone tails like histone methylation, demethylation, histone acetylation, and hydroxymethylation [3]. Thus, epigenomics research in plants will help in understanding control mechanisms of plant responses to environmental stressors and developmental cues, while enabling the future development of valid epigenome based-molecular markers that add to the available genome and transcriptome-based molecular markers. This research would help better understand crop plant genomes and epigenomes for various purposes including biofuel production, producing crop varieties that can withstand biotic and abiotic stresses, and breeding climate resilient crops to cope with the changing climatic conditions and enhance food security. We have attempted to provide an overview of plant epigenomics in relation to crop improvement by focusing on the various mechanisms involved and their interplay.

DNA METHYLATION DNA organization and compaction play a major role in plant gene regulation. A combination of DNA modifications (methylation), DNA-protein interactions (transcription factors, histones/nucleosomes), and protein–protein interactions (promoter–enhancer contact via transcription factor binding), directly control DNA accessibility, and thus gene expression [4]. One of the primary marks dictating the accessibility of DNA by transcription factors is DNA methylation. DNA methylation typically refers to the addition of a methyl (CH3) group to the 5′ carbon of a cytosine, which is by far the most studied and well-characterized nucleic acid modification [5]. There are other nucleic acid methylation modifications, such as adenine methylation (6 mA), which is present in mRNA and low amounts in genomic DNA. The variation of 6 mA in mRNA derived from different tissues was found to correlate with expression levels of the plant methyltransferase mRNA adenosine methylase (MTA) [6,7]. MTA inactivation was correlated with abnormal organ development in Arabidopsis [6]. Another form of DNA methylation is 5-hydroxymethylcytosine, which is considered as an intermediate of active demethylation, and is found in very low amounts in Arabidopsis [8]. Most prokaryotes and eukaryotes contain some level of DNA methylation. There are some species that lack DNA methylation and are usually considered to be loss-of-function, including those found in Caenorhabditis elegans and Saccharomyces cerevisiae [9,10]. Though most eukaryotic cells contain methylated DNA, there are considerable differences in the context and presence of DNA methylation among species [11–13]. Specific roles and functions of DNA methylation can vary greatly between different plant species and even within different tissues

in the same organism. Plants possess specific RNA polymerases (IV and V) that are responsible for introducing de novo DNA methylation, which contributes to the increased presence of non-CG methylation in plants [14].

Functions and Mechanisms of DNA Methylation Though non-CG methylation is found in mammalian cells, it is usually observed in a cell-type specific manner, and is particularly high levels in stem cells and neural cells, while very low levels are seen in somatic cells [15,16]. It is reported that though non-CG methylation is observed in certain tissues of mammals, although the specific function in each cell type remains to be elucidated [15,16]. Plants have the most diverse DNA methylation contexts throughout most tissues and have been shown to possess considerable levels of DNA methylation in the CHG and CHH trinucleotides. DNA methylation varies in tissues within the same plant as seen in soybean roots, stems leaves, and cotyledons [17], cotton plantlet and mature plant tissues [18], vegetative and floral stages of azalea buds [19], shoot, embryo, and endosperm in maize [20,21], between six different cell types making up the Arabidopsis root meristem [22], and among different Arabidopsis accessions [23,24]. General trends have been observed regarding the presence of methylated cytosines in plants. DNA methylation in plants is frequently found in the gene body of constitutively expressed genes, duplicate genes, at intron junctions, in suppressed promoters, transposons, and heterochromatin [11,25,26]. The boundaries of active transcription start sites (TSSs) and transcription termination sites (TTSs) have a dramatic decrease in DNA methylation. Methylation of promoters generally decreases expression of adjacent genes, especially in the CG context [17]. DNA methylation, particularly in the CG context is widely conserved among orthologs across plant species [11]. The overall levels and prominence of non-CG contexts was shown to vary radically among plants. MET1 is responsible for maintaining CG methylation. CMT3 is unique to plants and is responsible for non-CG methylation and de novo methylation. DRM1 and DRM2 are redundantly involved in non-CG methylation and de novo methylation. RNA-directed DNA methylation (RdDM) is a process that introduces new DNA methylation by plant-specific RNA polymerases (IV [NRPD] and V [NRPE]). RNA Pol IV and V are involved in the biogenesis and targeting of small interfering RNAs (siRNAs) to their genomic targets [14]. siRNAs are generated by Pol IV and are targeted to genomic DNA via Pol V, which recruits the DNA methylation enzyme DRM2 and other chromatin remodeling factors [14]. Comprehensive bisulfite-sequencing studies profiling Arabidopsis mutants have determined genes responsible

IV.  Model Organisms of Epigenetics



DNA Methylation

for CG, CHG, and CHH DNA methylation regulation. CG methylation is maintained by MET1 [27–29]. Additionally, vim1 vim2 vim3 triple mutants showed highly reduced CG methylation. Interestingly TEs of these mutants were found to have normal methylation levels, while genes were severely altered from the wildtype [29]. ddm1 mutants lose heterochromatic CG DNA methylation, implying that these enzymes work nonredundantly in CG methylation [29]. Arabidopsis CHG methylation is greatly reduced in cmt3 mutants, as well as histone methyltransferase triple mutant suvh4 suvh5 suvh6 [29]. The remaining CHG site methylation was lost in drm1/2 mutants [29]. CHH methylation was lost in Arabidopsis suvh4, suvh5, suvh6, and cmt3 mutants [29]. Loss of CHH methylation was found to be partially correlated with the loss of methylation in the symmetrical contexts, CG and CHG [29]. Mutants in RdDM- and RNAi-related genes ranged from having little to no effect on DNA methylation, to very high levels of DNA methylation loss. Gain of methylation in all three contexts has been observed in some Arabidopsis DNA demethylase mutants, ros1 dml dml3 [29].

Interactions With Other Epigenetic Factors DNA methylation and other epigenetic modifications do not occur independently of each other. There is growing evidence of crosstalk and reinforcing loops; for example one of the most well studied interactions is the recognition of methylated histones (H3K9) by CMT3, which leads to DNA methylation [30]. Histone modification mutants, suvh4, suvh5, and suvh6 were found to lack CMT2/3-dependent DNA methylation [30]. SUVH4 (also known as KYP) recruits CMT3 to methylate H3K9 histone tails, causing methylation of the adjacent DNA [31]. siRNAs are also associated with H3K9, leading to non-CG methylation [31]. H3K4 methylation and DNA methylation have an indirect relationship in plants and mammals [32]. A study in an Arabidopsis met1 mutant showed an interesting relationship between DNA methylation and histone modification-H3K27me3 [31]. Induced hypomethylation of CGs and H3K9me2 sites in transposons showed a hypermethylation at H3k9me3, suggesting that H3K9me3 and DNA methylation/H3K9me2 have an indirect relationship as repressive marks in Arabidopsis heterochromatic regions [31]. Many other methylation and acetylation modifications can occur on plant histones. DNA methylation-histone modification crosstalk is certainly a major factor in the formation of heterochromatin and open chromatin conformations. The interactions between DNA methylation, histone modification, siRNAs, and other factors are complex and will require many more studies to dissect precise cause and effect mechanisms.

247

Methods To Interrogate DNA Methylation Many techniques exist for analyzing DNA methylation, which range in sensitivity of detection, from targeted analysis of individual PCR products to whole genome analysis as reviewed by Laird in 2010 [33]. Briefly, these are sodium bisulfite-based, affinity enrichment-based, and enzyme digestion-based techniques for both sequencing and microarray strategies [33,34]. A frequently used enzyme-based technique is Methylseq, which uses a methylation-sensitive restriction enzyme, such as MspI or HpaII, followed by sequencing, or microarray [33]. Affinity-based techniques include methylated DNA immunoprecipitation-sequencing (MeDIP-Seq) [35,36] and Methyl-CpG-binding domain sequencing (MBD-seq) [37]. Bisulfite-based techniques are also commonly used; whole-genome bisulfitesequencing (WGBS) [38] yields information about individual cytosines, across the whole genome, while reduced representation bisulfite sequencing (RRBS) [39] uses a methylation-sensitive enzyme in order to enrich for methylated DNA [33]. The cheapest and lowest throughput bisulfite method would be targeted bisulfite sequencing. This involves bisulfite conversion, PCR, and Sanger sequencing.

Stability, Heritability, and Agricultural Significance DNA methylation plays an important role in plants, as its disruption is known to cause developmental deformities [10]. Flowering time is regulated by the Flowering Wageningen (FWA) gene, and its expression is known to be controlled by DNA methylation [40]. A lack of CG methylation at the promoter of this locus caused late flowering, which is a heritable phenotype [40]. Transposable elements (TEs) are generally heavily methylated in plants for genomic stability [41]. Loss of enzymes responsible for DNA methylation and chromatin remodeling factors (such as DDM1) increases the activation of TEs. The movement of transposons has been implicated largely with negative effects, but there are beneficial effects in multiple plants. For example, the insertion of an LTR retrotransposon within an intron of a soybean phytochrome A photoreceptor resulted in a protein mutation, which made it photoperiod-insensitive [42]. Consequently, this extends the latitude range for crop cultivation regardless of day length difference. Due to this phenomenon and other cases of favorable phenotypes, targeting TEs has been proposed as a tool for plant breeding [41]. Though there is some promise in exploiting controlled retrotransposon insertions, there may be unforeseen consequences during breeding. Since TEs are methylated as a means to control genome stability, methylation of a

IV.  Model Organisms of Epigenetics

248

16.  Plant Epigenomics

newly inserted TE can have an effect on the expression of adjacent genes as well [42]. LTR retrotransposons can be suppressed differently between plant tissues; for example, Arabidopsis columella cells were found to have the highest level of CHH methylation and the genes that were not expressed or lowly expressed were correlated with hypermethylation of the CHH context [22]. Additionally, genes containing TEs were found among the group of lowest expression, which also had relatively higher methylation in the CHH context [22]. Alternatively, some maize genes, such as those involved in abiotic stress response, may be activated by upstream insertion of certain families of TEs [43]. The mechanism could be that the TE causes enhancer activity that stimulates expression of stress-responsive genes [43]. In rice, genome-wide 5-hydroxymethylcytosine profiling was found to highly correlate with heterochromatic regions containing TEs that were not expressed [21]. Another study found that new mPing TE insertions throughout the rice genome either increased nearby gene expression or had no effect [44]. The authors noted that the rapid increase in copy number of TEs (TE bursts) could contribute to the fast response to stress and enhanced survival [44,45]. With TEs being unpredictable by nature, and displaying diverse effects among plant species, the consequences of breeding in rice and other crops could yield higher copies of TEs after each generation. This could potentially cause random and unexpected mutations and effects on global gene expression. Heritable changes in DNA methylation have been seen to occur due to mutation of DNA methyltransferases and other chromatin remodeling factors. One example is the loss of methylation at the FWA promoter, where the change in promoter methylation was not corrected after a functional copy of the DNA methylation-related genes was reintroduced back into the genome [40]. The authors suggested that CG methylation could not be rescued due to a lack of template for the maintenance of CG methylation by DDM1 and MET1 [40]. ddm1 mutations are also associated with a decrease in the histone modificationmethylation of H3K9 [32]. RNA-directed DNA methylation (RdDM) plays a role in defense against viruses and bacteria [14,46]. Hypomethylation of promoters and genes that are responsible for resistance, or hypersensitivity responses, is usually observed after inoculation with pathogens [47]. Transgenerational resistance attributed to conserve changes in DNA methylation was observed during the systemic acquired resistance (SAR) immunity response to Psuedomonas syringae inoculation [48]. In addition to biotic resistance, abiotic tolerance is also affected by or results in changes in DNA methylation [47]. Localized and genome-wide DNA methylation changes have been observed in a variety of abiotic stress conditions,

including cold, metal, salt/osmotic, hormone, and other stresses [47]. Studying these changes can provide insight as to which genes are directly involved in resistance and tolerance responses to biotic and abiotic stresses. Ultimately, controlled changes of DNA methylation or TE insertion could potentially cause heritable enhanced characteristics for plant breeding. However, both of these mechanisms are subject to constant change and require further research.

HISTONE MODIFICATIONS IN PLANTS The nucleosome is a subunit of chromatin and is composed of approximately 146 bp of DNA wrapped around a histone octamer (consisting of two copies each of the four core histones; H2A, H2B, H3, and H4) [49]. Histones are small proteins between 11 and 21.5 kDa and are structurally highly conserved through evolution from yeast to mammals [49,50]. A globular carboxyl domain and amino (N-terminal) tail make up histones. The N-terminal tails extend from the surface of nucleosomes and attract proteins, complexes, and remodeling enzymes with specialized enzyme activities [51].

Linker Histones in Plants Nucleosomes are linked by linker histones (H1) which further compact chromatin by binding DNA. Plant linker histones are more diverged from animals compared to their core histones. Linker histones have three domains: a globular core, which is highly conserved, N-terminal domain (NTD), and a C-terminal domain (CTD). Similar to core histones, histone linkers are highly basic and are rich in proline and lysine. However, they are more variable in eukaryotes compared to core histones, which are highly conserved. Some organisms have multiple histone variants while others have single histone variants, as in Drosophila melanogaster [52]. Histone linker H1.3 from Arabidopsis is a drought-inducible histone variant and binds closely to chromatin. Drought-inducible histone linkers have been identified in tomato and they provide resistance during drought conditions by enhancing a reduction of stomatal conductance, transpiration, and photosynthesis [53]. However, HIC and HID homologs of drought-inducible histone variant H1 from tobacco [54] have been identified but were not induced by drought suggesting the variety of roles that these histone linkers play in plants. Furthermore, the involvement of histone linkers in development, have also been reported [52]. Like in animals, linker histone variants have been identified in plants and their expression in different tissues and developmental stages of plants have been reported. They have been found to be induced by environmental factors, however, in most cases, the function of these

IV.  Model Organisms of Epigenetics



Histone Modifications in Plants

proteins remains unknown and more work is needed to fully understand their roles in plants.

Types of Histone Modifications Histone modification occurs at the amino termini of the histone core molecules, which is a significant part of the group of histones. The N-terminal tails of core histones undergo posttranslational modification, such as acetylation, methylation, phosphorylation, ubiquitination, ADP-ribosylation and sumoylation, and contribute to the regulation of transcription [55]. Histone modifying enzymes alter histones and cause changes in other nearby nucleosomes, as well as other regulatory proteins. These changes in effect regulate transcription by changing chromatin conformation [56].

Mechanism of Action of Histone Modifying Enzymes The activity of histone modifying enzymes is mediated by direct binding of transcription factors to recruit these enzymes into specific sites. Transcription factors also bind indirectly through coregulators to recruit regulatory proteins and other complexes for their activity. In addition to transcription factors, scaffolds of noncoding RNAs recruit histone modifying enzymes for their activity [57]. Additionally, some histone modifications serve as a target for the activation or inhibition of other histone modifying complexes. Phosphorylation of H3 serine activates the acetylation of lysine 14 of the same histone by GCN5 in yeast [58]. The association of multiple histone modifying enzymes provides specific targets for activity of the histone modifying complexes. For example, the existence of three specific complexes in Arabidopsis PRC2 is reported to regulate distinct target genes via H3K27. Another mode of action of histone modifiers is through posttranslational modification. For example, Hda1 from maize is activated through proteolytic cleavage during the time of seed germination [59].

Role of Histone Modifying Enzymes in Developmental Processes Acetylation of the lysine residues of core histones is associated with gene activation and expression. Histone acetyltransferases (HATs) are enzymes that mediate histone acetylation in eukaryotes. The reversal of histone acetylation is catalyzed by histone deacetylases (HDACs) through deacetylation and is associated with gene suppression. HATs and HDACs are the most studied enzymes involved in histone modifications. Four HAT families have been identified in plants including GCN5-related N-terminal acetyltransferase family (GNAT), MYST (MOZ, Ybf2 (Sas3), Sas2, and

249

Tip60), CREB-binding protein (CBP), and TATA-binding protein-associated factors (TAFII250). Twelve HATs and 18 HDACs have been identified in Arabidopsis [60]. The rice genome contains at least 8 HAT and 19 HDAC genes [61]. GCN5 is a transcriptional activator and various GCN5 coactivator complexes have been identified in plants and reported to play a role in plant development [62]. For example, GCN5 and its associated complexes are involved in floral organogenesis, cell differentiation, and meristem cell function. Similarly, a CAMP-responsive protein binding element, HAC1 from Arabidopsis plays a part in root elongation, de novo regeneration of shoots and flowering [63]. There are three major families of HDACs namely, the reduced potassium dependency protein 3 (RPD3)/ HDA1, silent information regulator protein 2 (SIR2), and the plant specific family of histone deacetylase 2 (HD2). The role of HDACs in plant developmental processes has been elaborated. Various developmental abnormalities were observed in hda19 mutant of Arabidopsis. These included abnormal development in flowers and siliques, reduction in male and female fertility, and premature death of seedlings and defects in the embryo. HDA19 and HDA6 belong to the RPD3 family of HDACs and are generally characterized as gene repressors. HDA6 is important for flower, seed, and leaf development in Arabidopsis. Mutants of hda6 in Arabidopsis showed late onset of flowering [64]. ZmRpd31 (RPD3-like enzyme) from maize and maize related protein (ZmRBR1) are involved in regulating cell cycle progression, senescence, flowering, and suppression of embryonic features [65,66]. HDA15, also a member of the RPD3 family of HDACs interacts with phytochrome interacting factor 3 (PIF3) to repress biosynthesis of chlorophyll and photosynthesis genes; HD2 is involved in leaf polarity in Arabidopsis [63].

Regulation of Gene Expression by Histone Modifying Enzymes During Biotic and Abiotic Stress Histone modifications in plants control gene expression when exposed to biotic and abiotic stress factors. Histone acetyltransferase (HAT) gene HAC1 is essential for transcriptional upregulation of heat-shock gene HSP17. In addition, the expression of genes involved in pathogenesis is regulated by HDA19 and useful for providing resistance to fungal pathogens [59]. There are reports on the opposing roles of GCN5 (HAT) gene and HD1 (plant specific histone deacetylases) from Arabidopsis in the control of various light-responsive genes with distinct effects on histone acetylation [67]. A group of stress-inducible genes was upregulated in gcn5 mutants. In Arabidopsis, SGF29a encodes GCN5 subunit complex and its mutants exhibited salt tolerance compared to the wild type. Furthermore, in ada2b mutants, subunits of

IV.  Model Organisms of Epigenetics

250

16.  Plant Epigenomics

GCN5 regulate COR6.6, RAB18, and RD29b by causing reduction of expression of these genes in Arabidopsis. There are reports of positive regulation of cold inducible genes by GCN5 and its complexes (GCN5 and ADA2b). Rice HATs; OsHAC701, OsHAC703, OsHAC704, and OsHAG703 were upregulated and downregulated with salt and cold treatments, respectively. Another HAT complex is the Elongator complex, which consists of six subunits (ELP1–ELP6). These Elongators are highly conserved in eukaryotes and have been shown to be involved in abiotic stress responses for example during drought conditions, and also with abscisic acid (ABA) and oxidative stress response [63]. HDACs play key roles in plant responses to biotic and abiotic stresses. In an earlier study, HDA6 and HDA19 showed significant involvement in jasmonic acid and plant-pathogen interaction pathways. In Arabidopsis, over expression of HDA19 resulted in an upregulation of ERF1 and increased resistance to the pathogen Alternaria brassicicola. Additionally, there are reports on the interactions of HDA19, ERF, and WRKY transcription factors during abiotic and biotic stresses. The amount of salicylic acid increased during loss-of-function of HDA19, which further raised the resistance of Arabidopsis to the pathogen Pseudomonas syringae [63]. There was an increase in transcripts of HDA6 at 12 h after inoculation of Phaseolus vulgaris inoculated with the bean rust pathogen Uromyces appendiculatus [68]. Another family of HDACs, SRT2 from Arabidopsis is reported to be involved in basal defense and induced during inoculation with Pseudomonas syringae pv. Tomato DC3000 (Pst DC3000) [63]. HDA6 is involved in cold and salt tolerance in Arabidopsis. Rice HDACs are reported to be involved in abiotic and biotic stresses; for example, HDACs transcripts in rice including OsHDA714, OsSRT701, and OsSRT702 are regulated by cold, salt, and mannitol stresses. HD2 and HDA6 interact to provide functional response to ABA and salt stresses. HD2 genes in barley are also reported to respond to biotic and abiotic stress. Stress-related plant hormones were induced during treatments with abscisic acid, jasmonic acid, and salicylic acid in barley [69]. The trithorax family of histone methyltransferases, such as Arabidopsis thaliana ATX1 and Hordeum vulgare HvTX1 play significant roles in plants in response to stresses. For example, ATX1 was induced during infection with bacterial pathogen [70] and HvTX1 during drought [71].

Crosstalk of Histone Modification and DNA Methylation Histone modification and DNA methylation link together to regulate epigenetic control of gene expression in eukaryotes. Histone modification and DNA methylation in plants associate together in a codependent feed forward loop, with both mechanisms

enhancing RNA-directed DNA methylation [30]. During salinity stress in soybean, transcription factors including AP2/EREB, bZIP, NAC, and MYB were activated with a reduction in DNA methylation and at the same time increase of histone modification marks H3K4me3 and H3K9ac. In addition, there are reports on the inclusion of HDA6 as an important component in gene silencing, involved with RNA-directed DNA methylation (RdDM) and in the maintenance of transposable elements and ribosomal RNA silencing [63].

ROLE OF EPIGENETIC MARKS IN PLANT DEVELOPMENT In eukaryotes including plants, chromatin structure plays an essential role in regulating gene expression. Therefore, it is important to study the mechanisms involved in altering chromatin structure thereby ultimately regulating the transcriptional activity of genes in various environmental conditions. The correlation between epigenetic landscapes and transcriptome alteration in plant growth and development has been reported in previous studies [72]. Chromatin also plays a dynamic role in cell cycle regulation. Therefore, some specific chromatin modifications are crucial to progress through the cell cycle [73]. The cell cycle phases associated with histone modifications has been extensively studied [74–76]. In the G1-phase, there is a strict balance between acetylases and deacetylases that plays a major role in maintaining acetylation patterns at target promoters thereby contributing to an open or closed chromatin configuration has been reported [77–79]. Full genome duplication in S-phase is maintained by the structural organization of the genome, largely due to epigenetic information in the modified histones. In a previous study, H3K18ac and H4K16ac have been shown to increase during S-phase progression in Arabidopsis [80]. Similarly, H4K5ac, H3K8ac, and H4K12ac in barley cells [81] and H5K5ac in onion cells [82] have also been associated with increases in acetylation in S-phase. Like in the G1-phase, the chromatin modifications associated with transcriptional activation or repression also take place in G2/M genes. Chromosomal packaging associated with a decrease in transcriptional activity and overall increase in histone acetylation has been observed in mitotic cells at H3K18 and K23, and H4K5, K8, K12, and K16 residues [83]. In another report, global changes in the heterochromatin mark H3K9me2 were not altered during the differentiation process in protoplasts [84]. In a recent study, it was found that four histones modifications namely, H3K4me2, H3K4me3, H3K9ac, and H3K27ac are associated with active transcription regions in rice [85]. Histone acetylation is known to induce transcription, as well as other cellular processes, such as DNA replication, repair,

IV.  Model Organisms of Epigenetics



Histone Modifications and Gene Expression in Response to Abiotic and Biotic Stress

and recombination [86]. In Arabidopsis, H3K9me3 levels were correlated with transcriptional activation [87]. Similarly, H3K36me3 was associated with induced activation of MADS box genes expression that is involved in flowering-time and floral organ development in Arabidopsis [88]. The enrichment of H3K4me3 around transcription regions indicates its crucial role in transcriptional initiation and early elongation, and was seen to be involved in early stages of xylogenesis (wood formation) in Eucalyptus grandis [89]. The genome-wide identification of activating (H3K4me3) and silencing (H3K27me3) marks on the differential expression of senescence-related genes has been documented in Arabidopsis [90]. Genes encoding lysine-specific histone demethylase 1 (LSD1), putative nucleic acid methyl transferases and binding proteins, RNA methyltransferase, and ribosomal RNA FtsJ-like methyltransferase were highly expressed in the progeny of the primed plants than nonprimed plants under postanthesis high-temperature stress [91].

HISTONE MODIFICATIONS AND GENE EXPRESSION IN RESPONSE TO ABIOTIC AND BIOTIC STRESS Understanding the molecular mechanisms of abiotic and biotic stress responses, tolerance, and adaptation in plants is important [92]. The mechanism by which plants adapt to unfavorable conditions involves the transcriptional regulation of genes and transcription factors. Plants adapt to abiotic and biotic stresses by modulating the expression of genes, which are often associated with epigenetic modifications. The distribution of histone marks that turn on or turn off the gene depends on environmental conditions. Chromatin structure plays a major role in regulating physiological mechanisms in response to the environment. There are several reports indicating that histone methylation changes are associated with various abiotic and biotic stress conditions. There are 28 histone modification sites identified in Arabidopsis [93]. Modification in chromatin structure is related to changes in posttranslational modifications (PTMs) of amino acid residues in N-terminal tails of histones [94]. Through the activity of particular histone-modifying enzymes, these histones become either acetylated or methylated depending on the particular chromatin structure needed for DNA accessibility. In plants, changes in chromatin structure, was related to significant effects on biological processes including plant growth and development, flowering, organogenesis, and embryogenesis [95–97]. It is reasonably clear that transcriptional regulation along with DNA methylation and DNA-histone modifications play a major role in stress response and adaptation [98]. In an earlier study, we reported the genome-wide mapping of histone modifications and gene expression in

251

response to bean rust infection in common bean [99]. Genome-wide histone methylation related to leaf senescence has been reported in Arabidopsis [90]. Genome-wide analyses have been performed in Arabidopsis, rice, maize, common bean, and woody plants using ChIP-chip or to high-throughput sequencing (ChIP-Seq). Acetylation of histones has been reported to play a role in flowering [100] and cold stress tolerance [101]. The enrichment of drought-responsive genes, RD29A, RD29B, RD20, and RAP2.4 has been associated with H3K4me3 and H3K9 acetylation and activation in 15-day-old Arabidopsis seedlings in response to dehydration stress [93]. In another study, histone H3K9ac was enriched by drought and this mark was rapidly removed following rehydration [102]. The synergetic relationships of three activating marks H3K4me3, H3K27ac, and H3K9ac and the induction of drought and developmental transition genes has been documented in Physcomitrella patens [103]. The activation by histone mark H3K9ac showed less enrichment in the progeny of salt-stressed plants [104]. H3K4me3 methylation affected only a small portion of stress-responsive genes under drought stress [105]. This modification positively and significantly enhanced transcript levels only for a subset of genes. Mutation in the Arabidopsis PICKLE (PKL) gene encoding a chromatin modifier showed increased and sustained expression of ABI3 and ABI5 to exogenous abscisic acid (ABA) treatment with reduced levels of H3K9me2 and H3K27me2 [106]. Dijk and coworkers in 2010 observed a strong correlation between H3K4me3 abundance and transcripts from drought-stress responsive genes [107]. They further suggested the involvement of histone demethylase in modulating the expression of a subset of stressresponsive genes. In response to submergence, the dimethylated state was changed to trimethylated state in the 5′ and 3′-coding regions of alcohol dehydrogenase 1 (ADH1) and pyruvate decarboxylase 1 (PDC 1) genes [108]. In Arabidopsis, the induction of the active histone modification mark H3K4me3 and decrease in the repressive mark H3K9me2 has been observed upon treatment with salt or ABA. The induction was correlated with the transcription of abiotic stress-responsive genes (ABI1, ABI2, KAT1, KAT2, DREB2A, RD29A, and RD29B) [109]. In previous reports, it was seen that JmjCdomain-containing histone demethylases JMJ14, JMJ15, and JMJ18 regulate H3K4me2/3 demethylase activity to maintain chromatin function and development during stress [110–112]. Fertilization-Independent Endosperm 1 (OsFIE1) lines showed reduced DNA methylation and H3K9 methylation upon heat stress in rice. OsFIE1 is a potential key component involved in thermal sensitivity and seed enlargement during endosperm development [113]. Pollen is highly sensitive to environmental disturbance, especially with high temperature as the process of male sporogenesis and male gametogenesis involves

IV.  Model Organisms of Epigenetics

252

16.  Plant Epigenomics

pollen development. The epigenetic mechanisms of thermotolerance during pollen development in plants have been extensively studied [114]. There are several reports indicating that epigenetic processes are involved in heat stress responses in pollen development [115– 119]. Cold stress is reported to change histone methylation in cold-responsive genes. During cold stress, the repression mark H3K27me decreased for cold-responsive genes, cold-regulated 15A (COR15A) and galactinol synthase 3 (ATGOL3) [120]. Significant changes in histone acetylation mark H4K12ac and histone methylation mark H3K9me2 and their correlation with gene expression have been reported in fungal rust-infected common bean [99].

EPIGENOME EDITING—A NEW APPROACH FOR CROP IMPROVEMENT Epigenome editing involves the characteristics feature of writing or erasing epigenetic marks in order to modulate gene expression at specific genic sites. This process has the potential advantage of passing these changes to the next generation (heritable) [121]. An important concept of epigenome editing is the fusion of a gene-specific DNA binding domain (DBD) with epigenetic effectors domain in order to modulate specific gene expression by the presence of enforced epigenetic effectors domains at the transcription start site [121]. A complete picture of epigenome editing can only be derived by acquiring a better understanding about genome editing [122]. Epigenome editing shares many similarities with genome editing as they fall under similar core biotechnology concepts and tools (Table 16.1).

Tools for Epigenome Editing In previous studies, chemicals, such as 5-Azacytidine (5-azaC; trade name Vidaza) and its deoxy derivative 5-aza-2′-deoxycytidine (Decitabine; trade name Dacogen) were used to modify the epigenetic state of plant genomes.

These chemicals were used to inhibit cytosine methylation in plants [123]. There is a growing list of cytosine methylation inhibitor compounds. However, these compounds have raised concern over their toxicity effects. On the other hand, epigenetic states of nuclear genomes have also been modified through enzymatic machinery, but there have not been reports of site specific epigenetic changes achieved using either chemical or enzymatic methods [124]. Successful accomplishment of site-specific changes in gene expression has been established by coupling a DNA binding domain with epigenetic modifiers [125,126]. These site-specific epigenetic modifications can be confirmed by methods, such as ChIP-seq, Bisulfite-PCR amplification, and quantification of gene expression [127]. Genome editing tools, such as zinc finger nucleases (ZFNs), transcription activator like effector’s nucleases (TALENs), and clustered regularly interspaced short palindromic repeats (CRISPR) have been used as DNA binding domains, which can be readily reengineered for epigenome editing [128].

Epigenome Editing Using Genome Editing Tools Targeted expression of enhanced DNA methylation and histone 3 lysine 9 (H3K9) methylation in vascular endothelial growth factor A (VEGF-A) promoter was achieved through fusing catalytic domains of DNA or H3K9 methyltransferases with zinc finger protein [129]. In another study, zinc fingers were efficiently employed for the suppression of oncogenes by fusing with DNA methyltransferase 3a to enable methylation and to downregulate gene expression at the promoters of Mapsin and SOX2 [130]. However, the designing of multizinc finger modules is hindered by the complex nature of interaction with targeted DNA sequence [131]. Furthermore, the practical application of ZFNs is limited by higher cost and altered sequence recognition due to presence of additional genomic and chromatin content surrounding the target site [122]. On the other hand, TALENs, are easier to design and can be constructed using simplified DNA assembly techniques,

TABLE 16.1 Overview of Epigenome and Genome Editing Epigenome editing

Genome editing

Targeting enzymatic activity at defined genomic target sites to introduce or erase chromatin mark

DNA recognition domains are fused with nuclease to induce changes in DNA sequence at target sites

Required components: DNA binding element, effector domain, and deactivated nuclease

Required components: DNA binding element and a nuclease

No alteration in genomic sequence

Alteration in genomic sequence

Chromatin modifying enzyme domains (DNA methyltransferase, demethylase, histone deacetylase) as fusion partner of DNA binding domain

Nuclease domain as fusion partner of DNA binding domain

IV.  Model Organisms of Epigenetics



Future Perspectives

such as golden gate cloning [132]. Fusion of ten-eleven translocation 1 (TET1) hydroxylase catalytic domain with TALE leads to substantial increase of endogenous genes by targeted demethylation at CpGs in human cells [126]. Generalized inactivation of enhancer-associated chromatin was achieved at targeted human cells by combining TALE repeat arrays and lysine-specific demethylase 1 (LSD1). These studies showed efficient and selective inactivation of its target enhancer [133]. However, TALE-based targeting, are highly sensitive to CpG methylation, which makes it unsuitable for methylating promoters to control gene inheritance [3]. Finally, selection of an ideal DNA-binding domain is directed toward CRISPR/CAS9-based epigenome editing because of the possibility of multiple target sites by guide RNAs and insensitivity to CpG methylation embedded in many promoters [3]. Additionally, different target specificity can be achieved by simple modification of 20-nucleotide in the gRNA sequence [132]. The employment of CRISPR/CAS9 for the specific expression of gene has been reported successfully [134]. Which involved fusion of nuclease deficient (dCAS9) with VP64 transactivation domain to activate specifically human vascular endothelial growth factor A (VEGFA) gene. In another study, site-specific gene regulation with minimal off target effects was achieved by fusion of effectors domain (KRAB) with dCAS9 [135]. A recent study by Hilton and coworkers in 2015 revealed target specific epigenome editing by combining nuclease nulldCAS9 with catalytic core of human acetyltransferase p300 [136]. This result supported highly specific targeted acetylation to regulate genes across the genome. Targeted methylation at larger portions of the promoter was achieved by combining deactivated Cas9 (dcas9) with the catalytic domain of the DNA methyltransferase DNMT3A. This methylation activity was site specific and heritable across mitotic divisions in human embryonic cells [3]. Although epigenome editing has been successfully applied in mammals, it is yet to be explored in plants [124], which could be a significant opportunity to understand the regulation of plant genomes by epigenome editing.

FUTURE PERSPECTIVES DNA methylation: plants develop germ cells late in development, which allows for DNA methylation “mutations” to develop during the plant’s life cycle and to be passed on to the next generation [23]. Since it has been shown that heritable changes in the epigenome are possible, this opens up the opportunity for introducing desirable phenotypes in crops. DNA methylation also occurs in plant organelles. Plants, unlike humans, have been found to lack 5mC in mitochondria; however,

253

plants do possess m6A, in small amounts [110]. Other organelles, such as chloroplasts and amyloplasts also contain m6A and 5mC [110]. Low levels of chloroplast DNA (5mC) methylation were discovered in Arabidopsis vegetative cells, excluding gametes [111]. Atypical mitochondrial DNA methylation in humans has been linked to diseases; therefore mutations in plant organelle DNA methylation can likely cause unfavorable conditions as well [112]. Differences in DNA methylation and ultimately gene expression can sometimes be attributed to proximally located TEs. Epialleles are known to contribute to differences in phenotype [113]. Taken together, it is possible that plant breeding may incorporate aspects of DNA methylation, as well as other epigenetic factors. With decreasing costs in sequencing and the availability of DNA methylation-sensitive enzymes, it may eventually become another tool in the toolbox for plant biologists. Histone modifications: Enzymes involved in histone modifications regulate various biological processes including development and stress responses in plants. However, information on the regulatory pathways of these enzymes is limited. Future research can be directed at determining the biological functions of the histone modifying enzymes which may help in understanding why plants are resistant or tolerant to abiotic and biotic stresses. Epigenome editing: Genome editing tools have potential application in basic and applied research in plant biology. Among the various genome-editing tools, CRISPR/CAS9 is preferable due to simplicity, efficiency, high specificity, minimal off-target effects and ease of multiplexing. Hence, CRISPR/CAS9 may be a very important tool in plant biology, which may aid in tagging endogenous genes with fluorescence proteins to monitor gene expression in vivo. Using fluorescent tagged dCAS9, changes in genomic structure and chromosome dynamics and gene responses to the environmental stimuli can be explored [131]. The dCAS9 can also act as a platform to recruit transcription activation/ repression domains to specific sites of genomes to regulate gene expression. The study conducted by Lowder and coworkers in 2015 in tobacco showed enhanced transcriptional activation of protein coding genes by dCAS9 fused with VP64 transcriptional domain [137]. It may also be interesting to use the transcriptional activation domain in epigenome editing via recruitment of histone modifying or DNA methylation proteins for the precise modulation of chromatin state. Targeted epigenome editing in human cells was successfully accomplished using dCAS9 fused with the catalytic core domain of acetyltransferase p300 [136]. Thus, introduction of epigenetic variation in plants using the CRISPR/ CAS9 system provides a platform to understand how plant genomes are regulated by epigenomes.

IV.  Model Organisms of Epigenetics

254

16.  Plant Epigenomics

Glossary 6mA  N6-methyladenine, the addition of a methyl group to the nitrogen-6 position ATX-1  Arabidopsis thaliana Trithorax 1: Is an Arabidopsis homolog of Trithorax and it controls the development of floral organs Cas9  CRISPR associated protein 9 CMT2/3  Chromomethylase 2/3 is responsible for non-CG DNA methylation and is involved in de novo DNA methylation via RdDM CPB  CREB-binding protein, is one of the families of histone acetyltransferases CRISPR  Clustered regularly interspaced short palindromic repeats DDM1  Decreased DNA methylation 1, a helicase that plays a role in CG and CHG methylation via RdDM, as well as histone modifications DRM1/2  Domains rearranged 1/2 are redundantly involved in de novo DNA methylation of non-CG sites via RdDM Epialleles  A gene variant that has identical gene sequence, but different epigenetic modification(s) FWA  Flowering wageningen, also known as homeobox-leucine zipper protein HDG6, a protein responsible for normal flowering time GCN5 General control nonrepressible 5, a histone acetyltransferase gene and belong to the GNAT superfamily GNAT Gcn5-related N-acetyltransferase (GNAT) superfamily of N-acetyltransferases H3K27  Histone 3, lysine at the 27th position, a target of HMTs to induce gene repression (di- and trimethylation), while mono-Methylation or acetylation induce gene transcription H3K4  Histone 3, lysine at the 4th position, a target of HMTs to induce gene expression (mono-, di-, and trimethylation) H3K9  Histone 3, lysine at the 9th position, a target of HMTs to induce gene repression (di- and trimethylation), while mono-Methylation or acetylation induce gene transcription HAT  Histone acetyltransferase HDAC  Histone deacetylase HIC and HID  Minor histone variants in tobacco HvTX1  Hordeum vulgare Trithorax 1: Is a homolog of Trithorax from Barley LTR retrotransposon  Long terminal repeat retrotransposon, a class I transposable element MBD-seq  Methyl-CpG-binding domain sequencing, a DNA methylation profiling method that involves the use of methyl-CpG-binding proteins MeDIP-seq  Methylated DNA immunoprecipitation sequencing, a DNA methylation profiling method that involves the use of an anti5-methylcytosine antibody, followed by high-throughput sequencing MET1  Methyltransferase 1 (MET1) maintains 5-methylcystosine at CG dinucleotides Methyl-seq  A DNA methylation profiling method that involves the use of methylated cytosine-sensitive restriction enzymes, followed by high-throughput sequencing mPing  A high copy-number DNA transposable element in rice MTA  mRNA adenosine methylase, the gene of a plant methyltransferase, which methylates an adenosine in an mRNA MYST  MOZ, Ybf2 (Sas3), Sas2, and Tip60, is a family of histone acetyltransferase MYST  MYST is a family of histone acetyltransferase formed from four members (MOZ, Ybf2, Sas2 and Tip60) NRPD  Nuclear RNA polymerase D (also known as RNA Pol IV), a plant-specific RNA polymerase involved in RdDM NRPE  Nuclear RNA polymerase E D (also known as RNA Pol V), a plant-specific RNA polymerase involved in RdDM RdDM  RNA-directed DNA methylation, an epigenetic process that introduces de novo DNA methylation RNAi  RNA interference, a posttranscriptional gene silencing mechanism RPD3  Reduced Potassium Dependency Protein 3-Is a major family of histone deacetylases

RRBS  Reduced representation bisulfite sequencing, a methylated DNA profiling method that involves the use of DNA methylation-sensitive restriction enzymes, followed by bisulfite conversion, and finally high-throughput sequencing SIR2  Silent Information Regulator Protein 2- is one of the major families of histone deacetylases siRNA  Small interfering RNA, double stranded RNA that is a length of approximately 20–25 base pairs, involved in RNAi and RdDM SOX2  Sex determining region Y box 2 SUVH4  Su(var)3–9 homolog protein 4, a histone methyltransferase that causes the epigenetic mark of methylation at the H3K9 position SUVH5  Su(var)3–9 homolog protein 5, a histone methyltransferase that causes the epigenetic mark of methylation at the H3K9 position SUVH6  Su(var)3–9 homolog protein 6, a histone methyltransferase that causes the epigenetic mark of methylation at the H3K9 position TAFII250  TATA-binding protein associated factors II 250, is a family of histone acetyltransferase TALENs  Transcription activator-like effectors nucleases TE  Transposable element, also known as a transposon, a mobile DNA element that can change its location throughout a genome Transposon  Also known as a transposable element (see TE) TSS  Transcription start sites, where DNA begins transcription into RNA TTS  Transcription termination sites, where DNA ends transcription into RNA VIM 1/2/3  Variant in methylation1/2/3, involved in CG DNA methylation WGBS  Whole-genome bisulfite sequencing, a methylated DNA profiling method that involves bisulfite treatment of DNA before sequencing ZFNs  Zinc finger nucleases

References [1] Grant-Downton RT, Dickinson HG. Epigenetics and its implications for plant biology. 1. The epigenetic network in plants. Ann Bot 2005;96:1143–64. [2] Calarco JP, Borges F, Donoghue MTA, Van Ex F, Jullien PE, Lopes T, et al. Reprogramming of DNA methylation in pollen guides epigenetic inheritance via small RNA. Cell 2012;151:194–205. [3] Vojta A, Dobrinic´ P, Tadic´ V, Bocˇkor L, Korac´ P, Julg B, et al. Repurposing the CRISPR-Cas9 system for targeted DNA methylation. Nucleic Acids Res 2016;44:5615–28. [4] Risca VI, Greenleaf WJ. Unraveling the 3D genome: genomics tools for multiscale exploration. Trends Genet 2015;31:357–72. [5] Saze H, Tsugane K, Kanno T, Nishimura T. DNA methylation in plants: relationship to small RNAs and histone modifications, and functions in transposon inactivation. Plant Cell Physiol 2012;53:766–84. [6] Luo G-Z, MacQueen A, Zheng G, Duan H, Dore LC, Lu Z, et al. Unique features of the m6A methylome in Arabidopsis thaliana. Nat Commun 2014;5:5630. [7] Luo G-Z, Blanco MA, Greer EL, He C, Shi Y. DNA N(6)-methyladenine: a new epigenetic mark in eukaryotes? Nat Rev Mol Cell Biol 2015;16:705–10. [8] Erdmann RM, Souza AL, Clish CB. Gehring M. 5-hydroxymethylcytosine is not present in appreciable quantities in Arabidopsis DNA. G3 (Bethesda) 2014;5:1–8. [9] Yi S. Birds do it, bees do it, worms and ciliates do it too: DNA methylation from unexpected corners of the tree of life. Genome Biol 2012;13:174. [10] He X-J, Chen T, Zhu J-K. Regulation and function of DNA methylation in plants and animals. Cell Res 2011;54:442–65. [11] Niederhuth CE, Bewick AJ, Ji L, Alabady MS, Kim K Do, Li Q, et al. Widespread natural variation of DNA methylation within angiosperms. Genome Biol 2016;17:194.

IV.  Model Organisms of Epigenetics



255

REFERENCES

[12] Feng S, Cokus SJ, Zhang X, Chen P-Y, Bostick M, Goll MG, et al. Conservation and divergence of methylation patterning in plants and animals. Proc Natl Acad Sci USA 2010;107:8689–94. [13] Gao F, Liu X, Wu X-P, Wang X-L, Gong D, Lu H, et al. Differential DNA methylation in discrete developmental stages of the parasitic nematode Trichinella spiralis. Genome Biol 2012;13:R100. [14] Matzke Ma, Kanno T, Matzke AJM. RNA-directed DNA methylation: the evolution of a complex epigenetic pathway in flowering plants. Annu Rev Plant Biol 2014;66:1–25. [15] Butcher LM, Ito M, Brimpari M, Morris TJ, Soares FAC, ÄhrlundRichter L, et al. Non-CG DNA methylation is a biomarker for assessing endodermal differentiation capacity in pluripotent stem cells. Nat Commun 2016;7:10458. [16] Guo W, Zhang MQ, Wu H. Mammalian non-CG methylations are conserved and cell-type specific and may have been involved in the evolution of transposon elements. Sci Rep 2016;6:32207. [17] Song Q-X, Lu X, Li Q-T, Chen H, Hu X-Y, Ma B, et al. Genomewide analysis of DNA methylation in soybean. Mol Plant 2013;6:1961–74. [18] Osabe K, Clement JD, Bedon F, Pettolino FA, Ziolkowski L, Llewellyn DJ, et al. Genetic and DNA methylation changes in cotton (Gossypium) genotypes and tissues. PLoS One 2014;9:1–11. [19] Meijón M, Valledor L, Santamaría E, Testillano PS, Risueño MC, Rodríguez R, et al. Epigenetic characterization of the vegetative and floral stages of azalea buds: dynamics of DNA methylation and histone H4 acetylation. J Plant Physiol 2009;166:1624–36. [20] Zhang M, Xie S, Dong X, Zhao X, Zeng B, Chen J, et al. Genomewide high resolution parental-specific DNA and histone methylation maps uncover patterns of imprinting regulation in maize. Genome Res 2014;24:167–76. [21] Wang P, Xia H, Zhang Y, Zhao S, Zhao C, Hou L, et al. Genomewide high-resolution mapping of DNA methylation identifies epigenetic variation across embryo and endosperm in Maize (Zea may). BMC Genom 2015;16:21. [22] Kawakatsu T, Stuart T, Valdes M, Breakfield N, Schmitz RJ, Nery JR, et al. Unique cell-type-specific patterns of DNA methylation in the root meristem. Nat Plants 2016;2:16058. [23] Kawakatsu T, Huang S, shan C, Jupe F, Sasaki E, Schmitz RJ, Urich MA, et al. Epigenomic diversity in a global collection of Arabidopsis thaliana accessions. Cell 2016;166:492–506. [24] Dubin MJ, Zhang P, Meng D, Remigereau M-S, Osborne EJ, Paolo Casale F, et al. DNA methylation in Arabidopsis has a genetic basis and shows evidence of local adaptation. Elife 2015;4:e05255. [25] Li X, Zhu J, Hu F, Ge S, Ye M, Xiang H, et al. Single-base resolution maps of cultivated and wild rice methylomes and regulatory roles of DNA methylation in plant gene expression. BMC Genom 2012;13:300. [26] Wang J, Marowsky NC, Fan C. Divergence of gene body DNA methylation and evolution of plant duplicate genes. PLoS One 2014;9:e110357. [27] Cokus SJ, Feng S, Zhang X, Chen Z, Merriman B, Haudenschild CD, et al. Shotgun bisulphite sequencing of the Arabidopsis genome reveals DNA methylation patterning. Nature 2008;452:215–9. [28] Lister R, O’Malley RC, Tonti-Filippini J, Gregory BD, Berry CC, Millar AH, et al. Highly integrated single-base resolution maps of the epigenome in Arabidopsis. Cell 2008;133:523–36. [29] Stroud H, Greenberg MVC, Feng S, Bernatavichute YV, Jacobsen SE. Comprehensive analysis of silencing mutants reveals complex regulation of the Arabidopsis methylome. Cell 2013;152: 352–64. [30] Du J, Johnson LM, Jacobsen SE, Patel DJ. DNA methylation pathways and their crosstalk with histone methylation. Nat Rev Mol Cell Biol 2015;16:519–32. [31] Deleris A, Stroud H, Bernatavichute Y, Johnson E, Klein G, Schubert D, et al. Loss of the DNA methyltransferase MET1

[32]

[33] [34]

[35] [36]

[37]

[38]

[39]

[40] [41] [42]

[43]

[44]

[45] [46]

[47]

[48] [49] [50]

[51] [52] [53]

[54]

induces H3K9 hypermethylation at PcG target genes and redistribution of H3K27 trimethylation to transposons in Arabidopsis thaliana. PLoS Genet 2012;8:e1003062. Law JA, Jacobsen SE. Establishing, maintaining and modifying DNA methylation patterns in plants and animals. Nat Rev Genet 2010;11:204–20. Laird PW. Principles and challenges of genome-wide DNA methylation analysis. Nat Rev Genet 2010;11:191–203. Olkhov-Mitsel E, Bapat B. Strategies for discovery and validation of methylated and hydroxymethylated DNA biomarkers. Cancer Med 2012;1:237–60. Mohn F, Weber M, Schübeler D, Roloff T-C. Methylated DNA Immunoprecipitation (MeDIP). Methods Mol Biol 2009;507:55–64. Taiwo O, Wilson G, Morris T, Seisenberger S, Reik W, Pearce D, et al. Methylome analysis using MeDIP-seq with low DNA concentrations. Nat Protoc 2012;7:617–36. Grafi G, Zemach A, Pitto L. Methyl-CpG-binding domain (MBD) proteins in plants. Biochim Biophys Acta—Gene Struct Exp 2007;1769:287–94. Urich MA, Nery JR, Lister R, Schmitz RJ, Ecker JR. MethylC-seq library preparation for base-resolution whole-genome bisulfite sequencing. Nat Protoc 2015;10:475–83. Meissner A, Gnirke A, Bell GW, Ramsahoye B, Lander ES, Jaenisch R. Reduced representation bisulfite sequencing for comparative high-resolution DNA methylation analysis. Nucleic Acids Res 2005;33:5868–77. Iwasaki M, Paszkowski J. Epigenetic memory in plants. EMBO J 2014;33:1–12. Paszkowski J. Controlled activation of retrotransposition for plant breeding. Curr Opin Biotechnol 2015;32:200–6. Grandbastien MA. LTR retrotransposons, handy hitchhikers of plant regulation and stress response. Biochim Biophys Acta— Gene Regul Mech 2015;1849:403–16. Makarevitch I, Waters AJ, West PT, Stitzer M, Hirsch CN, Ross-­Ibarra J, et al. Transposable elements contribute to activation of maize genes in response to abiotic stress. PLoS Genet 2015;11:e1005566. Naito K, Zhang F, Tsukiyama T, Saito H, Hancock CN, Richardson AO, et al. Unexpected consequences of a sudden and massive transposon amplification on rice gene expression. Nature 2009;461:1130–4. Naito K, Monden Y, Yasuda K, Saito H, Okumoto Y. mPing: the bursting transposon. Breed Sci 2014;64:109–14. López A, Ramírez V, García-Andrade J, Flors V, Vera P. The RNA silencing enzyme RNA polymerase V is required for plant immunity. PLoS Genet 2011;7:e1002434. Peng H, Zhang J. Plant genomic DNA methylation in response to stresses: potential applications and challenges in plant breeding. Prog Nat Sci 2009;19:1037–45. Luna E, Bruce TJA, Roberts MR, Flors V, Ton J. Next-generation systemic acquired resistance. Plant Physiol 2012;158:844–53. Rosa S, Shaw P. Insights into chromatin structure and dynamics in plants. Biology (Basel) 2013;2:1378–410. Mahrez W. Histone modifications and transcript processing during Arabidopsis development. Diss., Eidgenössische Technische Hochschule ETH Zürich No. 21256, 2013, 140p. Bannister AJ, Kouzarides T. Regulation of chromatin by histone modifications. Cell Res 2011;21:381–95. Over RS, Michaels SD. Open and closed: the roles of linker histones in plants and animals. Mol Plant 2014;7:481–91. Scippa GS, Di Michele M, Onelli E, Patrignani G, Chiatante D, Bray EA. The histone-like protein H1-S and the response of tomato leaves to water deficit. J Exp Bot 2004;55:99–109. Przewloka M, Wierzbicki A, Ślusarczyk J, Kuraś M, Grasser K, Stemmer C, et al. The “drought-inducible” histone H1s of tobacco play no role in male sterility linked to alterations in H1 variants. Planta 2002;215:371–9.

IV.  Model Organisms of Epigenetics

256

16.  Plant Epigenomics

[55] Natsume-Kitatani Y, Shiga M, Mamitsuka H. Genome-wide integration on transcription factors, histone acetylation and gene expression reveals genes co-regulated by histone modification patterns. PLoS One 2011;6:e22281. [56] Roth SY, Denu JM, Allis CD. Histone Acetyltransferases. Annu Rev Biochem 2001;70:81–120. [57] Holoch D, Moazed D. RNA-mediated epigenetic regulation of gene expression. Nat Rev Genet 2015;16:71–84. [58] Kurdistani SK, Grunstein M. Histone acetylation and deacetylation in yeast. Nat Rev Mol Cell Biol 2003;4:276–84. [59] Pfluger J, Wagner D. Histone modifications and dynamic regulation of genome accessibility in plants. Curr Opin Plant Biol 2007;10:645–52. [60] Pandey R, Müller A, Napoli CA, Selinger DA, Pikaard CS, Richards EJ, et al. Analysis of histone acetyltransferase and histone deacetylase families of Arabidopsis thaliana suggests functional diversification of chromatin modification among multicellular eukaryotes. Nucleic Acids Res 2002;30:5036–55. [61] Liu X, Yang S, Zhao M, Luo M, Yu C-W, Chen C-Y, et al. Transcriptional repression by histone deacetylases in plants. Mol Plant 2014;7:764–72. [62] Kaldis A, Tsementzi D, Tanriverdi O, Vlachonasios KE. Arabidopsis thaliana transcriptional co-activators ADA2b and SGF29a are implicated in salt stress responses. Planta 2011;233:749–62. [63] Yuan L, Liu X, Luo M, Yang S, Wu K. Involvement of histone modifications in plant abiotic stress responses. J Integr Plant Biol 2013;55:892–901. [64] Wang Z, Cao H, Chen F, Liu Y. The roles of histone acetylation in seed performance and plant development. Plant Physiol Biochem PPB 2014;84:125–33. [65] Rossi V, Locatelli S, Lanzanova C, Boniotti MB, Varotto S, Pipal A, et al. A maize histone deacetylase and retinoblastoma-related protein physically interact and cooperate in repressing gene transcription. Plant Mol Biol 2003;51:401–13. [66] Tanaka M, Kikuchi A, Kamada H. The Arabidopsis histone deacetylases HDA6 and HDA19 contribute to the repression of embryonic properties after germination. Plant Physiol 2008;146:149–61. [67] Eckardt NA. Facets of histone acetylation required for lightresponsive gene expression. Plant Cell (Online) 2006;18:2868. [68] Melmaiee K, Kalavacharla VK, Brown A, Todd A, Thurston Y, Elavarthi S. Quantification and gene expression analysis of histone deacetylases in common bean during rust fungal inoculation. Int J Genom 2015;2015:153243. [69] Demetriou K, Kapazoglou A, Tondelli A, Francia E, Stanca MA, Bladenopoulos K, et al. Epigenetic chromatin modifiers in barley: I. Cloning, mapping and expression analysis of the plant specific HD2 family of histone deacetylases from barley, during seed development and after hormonal treatment. Physiol Plant 2009;136:358–68. [70] Alvarez-Venegas R, Abdallat AAl, Guo M, Alfano JR, Avramova Z. Epigenetic control of a transcription factor at the cross section of two antagonistic pathways. Epigenetics 2007;2:106–13. [71] Shvarts IB, Kahn TG, Pirrotta V. Polycomb and trithorax control genome expression by determining the alternative epigenetic states of chromatin for key developmental regulators. Genetika 2010;46:1413–6. [72] Lauria M, Rossi V. Epigenetic control of gene regulation in plants. Biochim Biophys Acta 2011;1809:369–78. [73] Costas C, Desvoyes B, Gutierrez C. A chromatin perspective of plant cell cycle progression. Biochim Biophys Acta 2011;1809:379–87. [74] Taubert S, Gorrini C, Frank SR, Parisi T, Fuchs M, Chan H-M, et al. E2F-dependent histone acetylation and recruitment of the Tip60 acetyltransferase complex to chromatin in late G1. Mol Cell Biol 2004;24:4546–56.

[75] Guo L, Yu Y, Law JA, Zhang X. SET DOMAIN GROUP2 is the major histone H3 lysine 4 trimethyltransferase in Arabidopsis. Proc Natl Acad Sci USA 2010;107:18557–62. [76] Fleury D, Himanen K, Cnops G, Nelissen H, Boccardi TM, Maere S, et al. The Arabidopsis thaliana homolog of yeast BRE1 has a function in cell cycle regulation during early leaf and root growth. Plant Cell 2007;19:417–32. [77] Zhang HS, Dean DC. Rb-mediated chromatin structure regulation and transcriptional repression. Oncogene 2001;20:3134–8. [78] Macaluso M, Montanari M, Giordano A. Rb family proteins as modulators of gene expression and new aspects regarding the interaction with chromatin remodeling enzymes. Oncogene 2006;25:5263–7. [79] Desvoyes B, Sanchez MP, Ramirez-Parra E, Gutierrez C. Impact of nucleosome dynamics and histone modifications on cell proliferation during Arabidopsis development. Heredity (Edinb) 2010;105:80–91. [80] Jasencakova Z, Soppe WJJ, Meister A, Gernand D, Turner BM, Schubert I. Histone modifications in Arabidopsis—high methylation of H3 lysine 9 is dispensable for constitutive heterochromatin. Plant J 2003;33:471–80. [81] Jasencakova Z, Meister A, Schubert I. Chromatin organization and its relation to replication and histone acetylation during the cell cycle in barley. Chromosoma 2001;110:83–92. [82] Mayr C, Jasencakova Z, Meister A, Schubert I, Zink D. Comparative analysis of the functional genome architecture of animal and plant cell nuclei. Chromosome Res 2003;11:471–84. [83] Bonenfant D, Towbin H, Coulot M, Schindler P, Mueller DR, van Oostrum J. Analysis of dynamic changes in post-translational modifications of human histones during cell cycle by mass spectrometry. Mol Cell Proteomics 2007;6:1917–32. [84] Tessadori F, Chupeau M-C, Chupeau Y, Knip M, Germann S, van Driel R, et al. Large-scale dissociation and sequential reassembly of pericentric heterochromatin in dedifferentiated Arabidopsis cells. J Cell Sci 2007;120:1200–8. [85] Du Z, Li H, Wei Q, Zhao X, Wang C, Zhu Q, et al. Genome-wide analysis of histone modifications: H3K4me2, H3K4me3, H3K9ac, and H3K27ac in Oryza sativa L. Japonica. Mol Plant 2013;6:1463–72. [86] Lee KK, Workman JL. Histone acetyltransferase complexes: one size doesn’t fit all. Nat Rev Mol Cell Biol 2007;8:284–95. [87] Roudier F, Ahmed I, Bérard C, Sarazin A, Mary-Huard T, Cortijo S, et al. Integrative epigenomic mapping defines four main chromatin states in Arabidopsis. EMBO J 2011;30:1928–38. [88] Xu L, Zhao Z, Dong A, Soubigou-Taconnat L, Renou J-P, Steinmetz A, et al. Di- and tri- but not monomethylation on histone H3 lysine 36 marks active transcription of genes involved in flowering time regulation and other processes in Arabidopsis thaliana. Mol Cell Biol 2008;28:1348–60. [89] Hussey SG, Mizrachi E, Groover A, Berger DK, Myburg AA. Genome-wide mapping of histone H3 lysine 4 trimethylation in Eucalyptus grandis developing xylem. BMC Plant Biol 2015;15:117. [90] Brusslan JA, Rus Alvarez-Canterbury AM, Nair NU, Rice JC, Hitchler MJ, Pellegrini M. Genome-wide evaluation of histone methylation changes associated with leaf senescence in Arabidopsis. PLoS One 2012;7:e33151. [91] Wang X, Xin C, Cai J, Zhou Q, Dai T, Cao W, et al. Heat priming induces trans-generational tolerance to high temperature stress in wheat. Front Plant Sci 2016;7:501. [92] Bita CE, Gerats T. Plant tolerance to high temperature in a changing environment: scientific fundamentals and production of heat stress-tolerant crops. Front Plant Sci 2013;4:273. [93] Kim J-M, To TK, Ishida J, Morosawa T, Kawashima M, Matsui A, et al. Alterations of lysine modifications on the histone H3 N-tail under drought stress conditions in Arabidopsis thaliana. Plant Cell Physiol 2008;49:1580–8.

IV.  Model Organisms of Epigenetics



REFERENCES

[94] Fischle W, Wang Y, Allis CD. Histone and chromatin cross-talk. Curr Opin Cell Biol 2003;15:172–83. [95] Shen W-H, Xu L. Chromatin remodeling in stem cell maintenance in Arabidopsis thaliana. Mol Plant 2009;2:600–9. [96] Berr A, Shen WH. Molecular Mechanisms in Epigenetic Regulation of Plant Growth and Development. Plant Dev Biol Biotechnol Perspect, Berlin, Heidelberg: Springer Berlin Heidelberg; 2010, pp. 325–344. [97] Nelissen H, Boccardi TM, Himanen K, Van Lijsebettens M. Impact of core histone modifications on transcriptional regulation and plant growth. CRC Crit Rev Plant Sci 2007;26:243–63. [98] Pikaard CS, Mittelsten Scheid O. Epigenetic regulation in plants. Cold Spring Harb Perspect Biol 2014;6:a019315. [99] Ayyappan V, Kalavacharla V, Thimmapuram J, Bhide KP, Sripathi VR, Smolinski TG, et al. Genome-wide profiling of histone modifications (H3K9me2 and H4K12ac) and gene expression in rust (Uromyces appendiculatus) inoculated common bean (Phaseolus vulgaris L.). PLoS One 2015;10:e0132176. [100] Bond DM, Dennis ES, Pogson BJ, Finnegan EJ. Histone acetylation, VERNALIZATION INSENSITIVE 3, FLOWERING LOCUS C, and the vernalization response. Mol Plant 2009;2:724–37. [101] Zhu J, Jeong JC, Zhu Y, Sokolchik I, Miyazaki S, Zhu J-K, et al. Involvement of Arabidopsis HOS15 in histone deacetylation and cold tolerance. Proc Natl Acad Sci USA 2008;105:4945–50. [102] Kim J-M, To TK, Ishida J, Matsui A, Kimura H, Seki M. Transition of chromatin status during the process of recovery from drought stress in Arabidopsis thaliana. Plant Cell Physiol 2012;53:847–56. [103] Widiez T, Symeonidi A, Luo C, Lam E, Lawton M, Rensing SA. The chromatin landscape of the moss Physcomitrella patens and its dynamics during development and drought stress. Plant J 2014;79:67–81. [104] Farahani BNM. Epigenetics of stress adaptation in Arabidopsis: the case of histone modifications. Electron Thesis Diss Repos 2013;1174:1–93. [105] Zong W, Zhong X, You J, Xiong L. Genome-wide profiling of histone H3K4-tri-methylation and gene expression in rice under drought stress. Plant Mol Biol 2013;81:175–88. [106] Perruc E, Kinoshita N, Lopez-Molina L. The role of chromatinremodeling factor PKL in balancing osmotic stress responses during Arabidopsis seed germination. Plant J 2007;52:927–36. [107] van Dijk K, Ding Y, Malkaram S, Riethoven J-JM, Liu R, Yang J, et al. Dynamic changes in genome-wide histone H3 lysine 4 methylation patterns in response to dehydration stress in Arabidopsis thaliana. BMC Plant Biol 2010;10:238. [108] Tsuji H, Saika H, Tsutsumi N, Hirai A, Nakazono M. Dynamic and reversible changes in histone H3-Lys4 methylation and H3 acetylation occurring at submergence-inducible genes in rice. Plant Cell Physiol 2006;47:995–1003. [109] Chen L-T, Luo M, Wang Y-Y, Wu K. Involvement of Arabidopsis histone deacetylase HDA6 in ABA and salt stress response. J Exp Bot 2010;61:3345–53. [110] Deleris A, Greenberg MVC, Ausin I, Law RWY, Moissiard G, Schubert D, et al. Involvement of a Jumonji-C domain-containing histone demethylase in DRM2-mediated maintenance of DNA methylation. EMBO Rep 2010;11:950–5. [111] Lu F, Cui X, Zhang S, Liu C, Cao X. JMJ14 is an H3K4 demethylase regulating flowering time in Arabidopsis. Cell Res 2010;20: 387–90. [112] Yang H, Mo H, Fan D, Cao Y, Cui S, Ma L. Overexpression of a histone H3K4 demethylase, JMJ15, accelerates flowering time in Arabidopsis. Plant Cell Rep 2012;31:1297–308. [113] Folsom JJ, Begcy K, Hao X, Wang D, Walia H. Rice fertilizationIndependent Endosperm1 regulates seed size under heat stress by controlling early endosperm development. Plant Physiol 2014;165:238–48.

257

[114] Chen Y, Müller F, Rieu I, Winter P. Epigenetic events in plant male germ cell heat stress responses. Plant Reprod 2016;29: 21–9. [115] Pecinka A, Mittelsten Scheid O. Stress-induced chromatin changes: a critical view on their heritability. Plant Cell Physiol 2012;53:801–8. [116] Solís M-T, Rodríguez-Serrano M, Meijón M, Cañal M-J, Cifuentes A, Risueño MC, et al. DNA methylation dynamics and MET1a-like gene expression changes during stress-induced pollen reprogramming to embryogenesis. J Exp Bot 2012;63: 6431–44. [117] Migicovsky Z, Yao Y, Kovalchuk I. Transgenerational phenotypic and epigenetic changes in response to heat stress in Arabidopsis thaliana. Plant Signal Behav 2014;9:e27971. [118] McCue AD, Panda K, Nuthikattu S, Choudury SG, Thomas EN, Slotkin RK. ARGONAUTE 6 bridges transposable element mRNA-derived siRNAs to the establishment of DNA methylation. EMBO J 2015;34:20–35. [119] Bokszczanin KL. Solanaceae pollen thermotolerance initial training network (SPOT-ITN) consortium, Fragkostefanakis S. perspectives on deciphering mechanisms underlying plant heat stress response and thermotolerance. Front Plant Sci 2013;4:315. [120] Kwon CS, Lee D, Choi G, Chung W-I. Histone occupancydependent and -independent removal of H3K27 trimethylation at cold-responsive genes in Arabidopsis. Plant J 2009;60: 112–21. [121] de Groote ML, Verschure PJ, Rots MG. Epigenetic editing: targeted rewriting of epigenetic marks to modulate expression of selected target genes. Nucleic Acids Res 2012;40:10596–613. [122] Laufer BI, Singh SM. Strategies for precision modulation of gene expression by epigenome editing: an overview. Epigen Chromatin 2015;8:34. [123] Arase S, Kasai M, Kanazawa A. In planta assays involving epigenetically silenced genes reveal inhibition of cytosine methylation by genistein. Plant Methods 2012;8:10. [124] Puchta H. Using CRISPR/Cas in three dimensions: towards synthetic plant genomes, transcriptomes and epigenomes. Plant J 2016;87:5–15. [125] Jurkowski TP, Ravichandran M, Stepper P, Gurdon J, Waddington C, Consortium E, et al. Synthetic epigenetics—towards intelligent control of epigenetic states and cell identity. Clin Epigen 2015;7:18. [126] Maeder ML, Angstman JF, Richardson ME, Linder SJ, Cascio VM, Tsai SQ, et al. Targeted DNA demethylation and activation of endogenous genes using programmable TALE-TET1 fusion proteins. Nat Biotechnol 2013;31:1137–42. [127] Cano-Rodriguez D, Rots MG. Epigenetic editing: on the verge of reprogramming gene expression at will. Curr Genet Med Rep 2016;4:170–9. [128] Day JJ. New approaches to manipulating the epigenome. Dialogues Clin Neurosci 2014;16:345–57. [129] Kungulovski G, Nunna S, Thomas M, Zanger UM, Reinhardt R, Jeltsch A. Targeted epigenome editing of an endogenous locus with chromatin modifiers is not stably maintained. Epigen Chromatin 2015;8:12. [130] Rivenbark AG, Stolzenburg S, Beltran AS, Yuan X, Rots MG, Strahl BD, et al. Epigenetic reprogramming of cancer cells via targeted DNA methylation. Epigenetics 2012;7:350–60. [131] Kumar V, Jain M. The CRISPR-Cas system for plant genome editing: advances and opportunities. J Exp Bot 2015;66:47–57. [132] Bortesi L, Fischer R. The CRISPR/Cas9 system for plant genome editing and beyond. Biotechnol Adv 2014;33:41–52. [133] Mendenhall EM, Williamson KE, Reyon D, Zou JY, Ram O, Joung JK, et al. Locus-specific editing of histone modifications at endogenous enhancers. Nat Biotechnol 2013;31:1133–6.

IV.  Model Organisms of Epigenetics

258

16.  Plant Epigenomics

[134] Maeder ML, Linder SJ, Cascio VM, Fu Y, Ho QH, Joung JK. CRISPR RNA-guided activation of endogenous human genes. Nat Methods 2013;10:977–9. [135] Gilbert LA, Larson MH, Morsut L, Liu Z, Brar GA, Torres SE, et al. CRISPR-mediated modular RNA-guided regulation of transcription in eukaryotes. Cell 2013;154:442–51.

[136] Hilton IB, D’Ippolito AM, Vockley CM, Thakore PI, Crawford GE, Reddy TE, et al. Epigenome editing by a CRISPR-Cas9based acetyltransferase activates genes from promoters and enhancers. Nat Biotechnol 2015;33:510–7. [137] Lowder LG, Zhang D, Baltes NJ, Paul JW, Tang X, Zheng X, et al. A CRISPR/Cas9 toolbox for multiplexed plant genome editing and transcriptional regulation. Plant Physiol 2015;169:971–85.

IV.  Model Organisms of Epigenetics

C H A P T E R

17 Dynamic Changes in DNA Modifications During Key Embryonic Transitions Chris O’Neill Kolling Institute for Medical Research, Royal North Shore Hospital, NSW, Australia

O U T L I N E Introduction

261

Roles of TET-Mediated Modifications of 5meC

267

Embryo Development

262

Epigenetic Transitions in the Early Embryo

262

Stress During Key Embryonic Transitions Have Lifelong Consequences

270

Setting the Embryonic Epigenetic Ground State

263

Conclusions

271

Changes Accompanying Differentiation

265

References

272

The Roles for the Global Demethylation of Pluripotent Cells

266

INTRODUCTION Developmental epigenetics is the study of the processes that govern the stable patterns of gene expression which define each of the many cell types that go to make up the body. Each cell of the early embryo has the same developmental potential—the potential to form any of the cells of the body. This capacity is termed as pluripotency. As development progresses, differentiation into a number of discrete cell populations occurs and each of these populations have a progressively restricted range of developmental outcomes. Differentiation is a result of changed patterns of gene expression and these patterns define not only the structure and function of the cell but also the range of developmental outcomes of each cell lineage. Underlying these changes in gene expression are alterations to the structure and function of chromatin, which confer these stable, but lineage-specific, patterns of gene expression. In one of the cell populations within the embryo, the primordial germ cells (PGCs), this progressive pattern of differentiation is interrupted Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00017-1 Copyright © 2017 Elsevier Inc. All rights reserved.

by the regeneration of the pluripotent state, prior to their differentiation into the gametes. There are many epigenetic processes including covalent modifications of cytosine nucleotides and histone proteins, the actions of a range of nonhistone proteins within euchromatin, and noncoding RNAs. DNA methylation was one of the first epigenetic modifications studied and has a broad range of actions detailed elsewhere in this book. This chapter will focus on the reprogramming of DNA methylation (and related modifications) during key developmental transitions during embryo growth. Errors in epigenetic reprogramming are likely the cause of embryo loss and some forms of congenital abnormalities [1]. The embryo and fetus’ environment during critical epigenetic transitions can create subtle alterations to epigenetic settings that can cause longterm changes in normal postnatal development [2–5]. As a result, the environment experienced during development can have important lifelong impacts on the health and disease predisposition of offspring, a phenomenon referred to as the developmental origins of health and

261

262

17.  Dynamic Changes in DNA Modifications During Key Embryonic Transitions

disease. Currently knowledge of epigenetic mechanisms it is not sufficient to allow a detailed understanding of how this long-term molecular memory is achieved, and this limits our capacity for therapeutic interventions, and is an important research objective of the field. Improved methods for analysis of epigenetic changes in small numbers of cells, together with the molecular characterization of embryonic stem cells (ESCs) and induced pluripotent stem cells (iPS cells) are now providing an “embryonic” understanding of some of the epigenetic mechanisms that underpin key developmental transitions. This chapter will identify the broad trends in our current understanding of changes in DNA methylation required for the formation of pluripotent cells. Much of the discussion will be on evidence from the mouse, unless otherwise indicated.

EMBRYO DEVELOPMENT Embryo development is initiated by a series of rapid cell divisions of the single cell embryo (zygote) that is formed at fertilization. In mammals, the first few cell cycles are directed by mRNA and protein inherited primarily from the oocyte. In the mouse, the vast majority of this inherited information is degraded by the end of the 2-cell stage of development and the initiation of new transcription from the embryonic genome at this time is required for further normal development [6,7]. This new transcriptome is different in many respects from that inherited from the gametes [7,8] and creates a state of totipotency, whereby the resulting cells have the unique capacity to form all of the specialized cell types of the embryo, the extraembryonic membranes, and the placenta. As the number of cells within the embryo increases, subsets of cells progressively begin to specialize in function and fate by the process of differentiation. In the mouse, the first differentiation event is completed several days after fertilization when embryos contain around 60 cells. This results in the formation of the blastocyst stage embryo, which possesses two distinct populations, the pluripotent inner cell mass (ICM) and the multipotent trophectoderm (TE) [9]. The TE cells have a marked restriction in their developmental potential, being capable of only forming the several cell types within the placenta. As such, these cells are classed as multipotent. As the blastocyst expands, the cells of the ICM resolve into two distinct layers: the pluripotent epiblast and the multipotent hypoblast (also known as primitive endoderm). At gastrulation, the cells of the epiblast differentiate to form the three primary germ layers: the mesoderm, the ectoderm, and the endoderm. It is from these three multipotent germ layers that have the

capacity to differentiate into each of the cells and tissues of the body by progressive rounds of differentiation [10]. The second round of reprogramming to pluripotency commences on day 7.5 of development in the mouse, when a small cluster [30–50] of cells in the extraembryonic mesoderm commence migration from the dorsal aspect of the hindgut toward the genital ridge [11]. Upon arrival within the genital ridge, they continue to proliferate and contribute to the formation of the sex-specific gonads. These migrating cells are the primordial germ cells (PGCs) and undergo a reversion to the pluripotent state during migration [12]. Once established within the developing gonads they commence differentiation into the gametes: the oocytes and spermatozoa [11,12].

EPIGENETIC TRANSITIONS IN THE EARLY EMBRYO These processes of embryo development require three major rounds of epigenetic reprogramming. The first round occurs immediately after fertilization and prepares the new zygotic genome [13] to produce a totipotent transcriptome. The second is the formation of the pluripotent ICM/epiblast of the blastocyst [14]. The third of these reprogramming events occurs during the formation of the pluripotent PGCs [12]. Currently, there is a poor understanding of the nature and epigenetic regulation of totipotency, but the formation of the ICM and PGCs are both considered to reset the epigenetic landscape to a pluripotent ground state, from which differentiation to progressively more advanced lineages can occur. In both cases profound changes have been observed in a range of epigenetic properties, and there is some similarity in both reprogramming events. Reprogramming to these two epigenetic ground states, however, results in profoundly different developmental outcomes. PGCs differentiate into the gametes (sperm and eggs) while the pluripotent cells of the epiblast differentiate into all of the cells of the organism, including the PGCs. These different developmental fates may suggest that the epigenetic programs created during these two reprogramming events are profoundly different. Yet, experimental results suggest that this is not the case. The most powerful finding is that when ESC (derived from the ICM) or PGC cells are isolated and then reintroduced into the early embryo, both cell types are capable of becoming incorporated into all the tissues of the body of the resulting offspring [15,16]. The chimeric individuals formed (including germ cell chimerism) are considered the gold-standard proof of the pluripotency of the tested cells. Furthermore, ESCs derived from the ICM can be induced to form PGC-like cells which, when transferred to the gonads, differentiate into gametes [17].

V.  Factors Influencing Epigenetic Changes



Setting the Embryonic Epigenetic Ground State

Such results argue that the epigenetic ground state of both the PGCs and ICM of the early embryo are sufficiently similar to allow the same developmental outcomes when the cells are placed within the same context. Thus, the different developmental outcomes that are derived from PGCs and ICM result primarily from their different developmental context rather than any fundamental difference between these two pluripotent lineages. A pluripotent cell within the gonad receives the positional information that directs differentiation to the gametes, yet when placed within the early embryo these same cells receive different positional cues that results in differentiation along other embryonic lineages. These results demonstrate that the interface between the epigenetic settings created within a cell and their developmental fate depends on the environmental cues normally provided by their position within the space-time of embryo development. In one important respect though the pluripotent state of the ICM and PGCs are markedly different. The epigenetic reprogramming to pluripotency that occurs in the ICM occurs without the loss of the parent-of-origin epigenetic marking of imprinted loci, whereas these imprints are removed in PGCs [18]. Imprinting of target loci is reestablished in a sex (parent)-specific manner during differentiation of the germ cells into gametes within the gonads. Hence, the pattern of gene imprinting in the oocyte and spermatozoa differs and this provides an epigenetic molecular memory of the parent-of-origin for each allele at imprinted loci. A detailed description

263

of the mechanisms of imprinting is found elsewhere in this volume, but it is evident that by allelic differences in the methylation of 5′-cytosine (5meC) within cytosineguanosine dinucleotides (CpGs) it is of fundamental importance to the monoallelic expression of these loci [19]. As imprinting was one of the first [20] and most fully investigated epigenetic phenomena proposed and investigated in mammals, this has meant that DNA methylation has been prominent in the investigation of developmental epigenetics.

SETTING THE EMBRYONIC EPIGENETIC GROUND STATE It has been long considered that a characteristic feature of the epigenetic reprogramming in the early embryo is a marked global reduction of DNA methylation immediately after fertilization (for review, see ref. [21]), however, this model has come under reassessment in recent years. The emerging new view on the dynamics of global changes in methylation during development is illustrated in Fig. 17.1. Spermatozoa have relatively high overall levels of genomic methylation, with similar levels to most somatic cells, while the ovulated oocyte has somewhat lower levels. Upon fertilization, methylation of the paternally acquired genome is remodeled with the result that by the time of syngamy the methylation levels of the paternally and maternally acquired DNA are similar [13,14,22–24]. Upon syngamy, there is some

FIGURE 17.1  A schematic representation of our new understanding of the changes in the global levels of 5meC during key epigenetic and developmental transitions across embryo development in the mouse. Neither axis is drawn to scale. ICM, Inner cell mass; PGC, primordial germ cells.

V.  Factors Influencing Epigenetic Changes

264

17.  Dynamic Changes in DNA Modifications During Key Embryonic Transitions

further modest remodeling of the genome over the next several cell-cycles [13,22,23], but no evidence for a global failure of maintenance methylation. Following the compaction of the mouse embryo at the 8-cell stage, there is a significant loss of methylation from the inner cells of the embryo [14] (Fig. 17.2). These inner cells are determined to the ICM fate while the outer cells are determined to the TE lineage [25]. After implantation of the embryo into the lying of the uterus, a significant increase in global methylation levels occurs in the embryo [13,22]. To date it has not been possible to assess changes in each of the emerging cell lineages of the embryo after implantation, however, an in vitro model of implantation shows that there is considerable difference between each of the lineages [14] (Fig. 17.2). Thus, the pluripotent epiblast that forms from the ICM retains a globally hypomethylated state, the newly formed hypoblast becomes remethylated, while the trophoblast and trophoblast giant cells show high levels of global methylation [14]. Base-level analysis of methylation status has been possible on ∼90% of all CpGs within the mouse preimplantation embryo and gametes [22]. At most CpGs

there was a general trend for average methylation levels being higher in sperm than oocytes, and the level in 2-cell was a little lower than in oocytes. There was a further small decline in most classes of genomic elements in 4-cell embryo and a large decline in the ICM [22]. This trend was evident for all classes of repeat elements, but some specific elements, such as IAPs, showed a much more modest level of change during development. Methylation levels in enhancer elements also showed the general pattern of extensive demethylation in the ICM cells. Intriguingly, this general trend was not as strongly evident for CpG islands (CGIs), which generally showed low levels of methylation at each stage of early development, which is only modestly further reduced in pluripotent cells [22]. Imprinted loci escape this round of demethylation [18]. The selective demethylation of the emergent ICM may be associated with the reduced expression of DNMT3B in ICM compared to the TE [14,26], suggesting that differential levels of de novo methylation are important in creating these differences between the ICM and TE. Base-level analysis has also been undertaken in human embryos [27] and broadly

FIGURE 17.2  Patterns of 5meC and 5hmC staining in the oocyte and different stages of preimplantation embryo development. (A) z-Stack projection images of staining for DNA (DAPI, blue), 5meC (green), 5hmC (red), and their merged (merge) images in the oocyte, zygote (PN1 and PN3), 2 cell, 4 cell, 8 cell, morula (ML), and blastocyst (BL) stages of development, and a blastocyst stained with nonimmune IgG. Asterisk indicates the decondensing sperm head (in newly fertilized oocyte) or the male pronucleus (in PN3 zygotes). Star indicates a female pronucleus. Triangle indicates the extruded polar body. (B) Sequential single confocal sections through the blastocyst, showing high levels of 5meC and 5hmC in the TE and low levels in the ICM. White asterisk is the blastocoel cavity, dotted line outlines the ICM. Scale bars 10 µm. Reprinted in modified form with permission [14]. 5hmC, 5′-Hydroxymethyl cytosine; 5meC, methylation of 5′-cytosine.

V.  Factors Influencing Epigenetic Changes



Changes Accompanying Differentiation

reflects the patterns observed in mice. However, it is important to note that by necessity analysis is on human embryos created by IVF. As will be discussed in a later section, the stresses associated with these procedures cause a range of aberrations in the pattern of normal methylation and it is well to take this into consideration when considering the normal processes of remodeling in human. The relevance of this profound hypomethylation in the ICM/epiblast for pluripotency has been challenged by observations that seemingly pluripotent ESCs have high global methylation levels [28]. Indeed, the methylation levels of ESCs were not markedly different from that of most differentiated somatic cells. Recently, however, it has been shown that the pluripotency of ESCs cultured under conventional conditions was more apparent than real. ESCs in conventional culture are in an epigenetically metastable state and are considered to be “primed” for differentiation rather than in the true pluripotent ground state [28,29]. New culture methods enhance the proportion of cells in the pluripotent ground state [30] and this is associated with the onset a marked global hypomethylation [28,29]. This hypomethylation is more consistent with characteristic hypomethylation of the pluripotent ICM [28] and PGCs [12]. Thus, ESC in the primed state has around 4% of all cytosines (∼80% of all CpGs) methylated and this is reduced to only 1% of cytosines within ESCs in ground state pluripotency [22]. Conversely, the transition of ESCs from “ground state pluripotency” to “primed for differentiation” is accompanied by a fourfold increase in global 5meC levels and an upregulation in the expression of the DNA methylases Dnmt3a, 3b, and their regulatory component Dnmt3l [28]. Thus, global hypomethylation, affecting most classes of genetic elements, is a defining feature of the pluripotent state. PGCs also undergo global erasure of DNA methylation. Loss of DNA methylation is reported [12] to commence by at least day 10.5 of development during the migration of the PGCs from the hindgut to the genital ridge. Demethylation seems to be largely complete by day 13.5 and occurs in PGCs of both male and female embryos. In contrast to the pluripotent ICM, methylation erasure in PGCs includes the imprinted loci insert space [12]. Since the discovery of the differential hypomethylation of the ICM is only recent [14], a molecular description of the control of this process is awaited. The marked differential expression of DNMT3B between the ICM and TE suggests that the relative absence of this methylase from the ICM may play an important role. The demethylation in PGCs is reported to be accounted for by a failure of maintenance methylation, yet this is not mediated by an absence of DNMT1. It has been proposed that the absence of UHRF1 (a recruitment factor for DNMT1) which limits the function of DNMT1 [31]. It has also

265

been shown that there is a relative absence of DNMT3A and 3B from the PGCs [32], so perhaps the interplay of reduced activity of all three DNMTs account for this loss of methylation. In contrast to the reprogramming that occurs in the formation of the ICM, the demethylation that occurs in PGCs also involves loss of methylation at imprinted alleles. This seems to be a late event in reprogramming. It occurs after the global demethylation during migration and becomes evident after the entry of PGCs into the genital ridge [11]. Base-level methylation analysis of CpG’s [12] shows that while the demethylation in PGCs affects most regions of the genome, including many of the repeat elements, some regions, such as some introns and intergenic regions, are more resistant to demethylation. There are also PGC sex-specific differences in the global and base-level patterns of demethylation. As is the case for the preimplantation embryo, the CGIs in most promoters are relatively hypomethylated throughout this reprogramming of PGCs and most show only modest differential methylation during germ cell specification and maturation [12].

CHANGES ACCOMPANYING DIFFERENTIATION Extensive remodeling of CpG methylation occurs with differentiation of the pluripotent cells into more specialized cells in the embryo. Whole genome methylation analysis shows that the methylation status of only 6.8% of CpGs within the sperm genome remains stable by day 7.5 of development. Of the stable CpGs in sperm, most were unmethylated (mainly in CGIs and promoters) and a smaller number of stably methylated CpGs were overrepresented in repeats or introns. Similar trends were observed for oocyte CpGs [22]. The first differentiation event in the pluripotent ICM is the formation of the hypoblast (also known as the primitive endoderm). This lineage is multipotent and contributes to the formation of the extraembryonic tissue, especially the visceral endoderm. Formation of the hypoblast is accompanied by an increase in global methylation [14] and an increased level of DNMT3B [26] relative to low levels in the pluripotent ICM/epiblast. Differentiation of PGCs around day 16.5 is accompanied by a marked remethylation of the germ cell genome as gametogenesis commences. This remethylation is greater in male than female germ cells, probably accounting for the higher levels of global methylation in sperm compared with oocytes, and occurs early in development in the male gametes [12]. This results in new patterns of methylation at imprinted loci being laid down [12]. The centrality of remethylation to the differentiation from pluripotency is illustrated by observations that mouse

V.  Factors Influencing Epigenetic Changes

266

17.  Dynamic Changes in DNA Modifications During Key Embryonic Transitions

ESCs lacking DNMTs are viable and proliferative yet lacking the capacity to escape pluripotency or undergo normal differentiation [33,34]. The maintenance of human ESCs, by contrast, is independent of Dnmt3 and Dnmt3b, but requires the activity of Dnmt1 [35]. Differentiation from the pluripotent state is characterized by defined changes in methylation patterns. Upon differentiation from the epiblast each of the endodermal, ectodermal, and mesodermal primary germ lines show sets of 2000–4000 differentially methylated regions, and these tend to be enriched for transcription factor binding sites [36]. Studies in ESC indicate that methylation at transcription factor binding sites plays an important role in constraining the range of genes that can be expressed in a tissue-specific manner [37]. Analysis of a range of differentiated somatic cells showed that the developmental origin of the cell type is a major determinant of the pattern of differential methylation detected. This lineage-dependent effect shows cells that have a common ontological heritage share differentially methylated regions. These regions are enriched for genetic regulatory elements and lineage-defining genes [36]. Thus, the onset of differentiation is characterized by increased DNA methylase activity that confers lineage-specific patterns of methylation that serve to define each of the primary germ layers. Further differentiation within each of these lineages may build on these patterns of methylation in an iterative manner. The action of sets of core transcription factors appears to be primarily responsible for defining the lineage-specific patterns of transcription. Epigenetic processes define the targets and levels of binding of these transcription factors and transcription binding sites overlap with regions of dynamic reprogramming of CpG methylation during differentiation. Genome-wide mapping of trimethylated histone 3 lysine 4 (H3K4me3) and trimethylated (H3K27me3) in ESCs and differentiated cell-lineages demonstrate that genes, which carry H3K4me3, but not H3K27me3, are generally actively expressed in ESCs [38–41]. These include the core pluripotency transcription factor genes, Oct4, Sox2, and Nanog. Many genes in ESCs, however, do not have either a definitive active or repressed chromatin state but instead have both the active and repressive epigenetic signatures. This bivalent state is thought to generate a condition whereby the gene is “poised” for activation and held in a transcriptionally ready state in pluripotent cells. Many of the genes possessing this bivalent domain are those that govern lineage specification and initiation of lineage specification is associated with resolution of this bivalency to either a stably repressed or active state [38–41]. A key feature in stabilizing these new epigenetic states with differentiation may be the methylation of these genes regulatory regions. In human ESCs the induction of differentiation of ESC into each of the three

primary germlines is associated with increased methylation of many of the binding sites for the pluripotency core transcription factors. This is strongly associated with the reduced expression of their target genes [42]. Conversely, differentiation into each of these primary germ lines in vitro is associated with targeted loss of methylation at binding sites for the core transcription factors for these lineages [42]. At this time the molecular mechanisms for these targeted changes in methylation is yet to be defined, and the functional interrelationship between changes in histone modifications, DNA methylation, and lineage-specific gene expression is yet to be fully defined. In considering these roles of methylation in cellular differentiation it's well to consider some limitations in our current understanding. Much of the analyses use ESCs because of the technical and logistical challenges associated with the analysis of these changes in the embryo in situ. Almost all detailed reports to date utilize ESC that are primed for differentiation rather than those that are in the naïve ground state of pluripotency. Thus, the relationships between pluripotency and differentiated lineages may be more complex then currently indicated by the data. A second major current limitation is that many reports of methylation changes actually use methods that do not distinguish between 5meC and 5′-hydroxymethyl cytosine. Since these two CpG modifications almost certainly have profoundly differing epigenetic functions, this is a significant caveat to interpretation of current data.

THE ROLES FOR THE GLOBAL DEMETHYLATION OF PLURIPOTENT CELLS Global hypomethylation is a hallmark of pluripotency and contrasts with the relatively hypermethylated state of most differentiated somatic cell types. The purpose and role of this global demethylation is yet to be understood. Until recently much of the attention to CpG methylation has focused on the differential levels of methylation of CGIs at gene promoters. This focus originally stemmed from knowledge of the role of differential methylation of monoallelically imprinted loci [18] and the clear negative correlation between promoter methylation and levels of gene expression [22]. With respect to this, the discovery that the expression of critical pluripotency specifying genes, such as Oct-4, are negatively associated with their levels of promoter methylation [43,44] further focuses attention to this level of control. This gene level focus, however, may not be fully warranted. In the early embryo much of the loss of methylation occurs at sites other than the CGIs [22], and even where extensive loss of methylation at promoters occurs

V.  Factors Influencing Epigenetic Changes



ROLES OF TET-MEDIATED MODIFICATIONS OF 5meC

in the embryo there still remains a strong negative association between (these lower) levels of methylation and the levels of gene expression [22]. Furthermore, the vast majority of CGIs are hypomethylated in most cell types and show relatively limited differential methylation during epigenetic reprogramming to pluripotency, compared to very large changes in methylation of most other elements within the genome [22]. It is also noteworthy that the methylation of Oct4 that occurs with differentiation is a lagging event and occurs after the repression of Oct4 gene expression [45]. It has been suggested that this reflects a mechanism for stabilizing or locking the gene in a repressed state, rather than as a primary repressive event [46]. The observation that the enhancer regions (and particularly the super-enhancers—which are important in lineage specification [47]) show the general pattern of a marked hypomethylation during reprogramming to pluripotency [22] may suggest that these are potentially important targets of epigenetic reprogramming. Yet, detailed mapping of pluripotent PGCs [46] shows that the expression of relatively few of their genes are dependent upon differential methylation, and of those that are dependent, most function to protect cells from the genomic instability that normally results from demethylation rather than obvious lineage-specifying functions [46]. So this raises the question, if the role of global demethylation in pluripotency is not primarily to reprogram CGIs, what is the purpose? Induced demethylation of somatic cells by knockout of Dnmts results in a marked change in gene expression, but more importantly leads to a profound loss of genomic stability and typically leads to cell death [48]. Yet the genetic removal of Dnmts from ESCs does not influence their viability or proliferative capacity but does prevent their normal capacity for differentiation [33]. This may argue that while differential methylation is required for lineage differentiation, the hypomethylated state of pluripotent cells has other as yet undefined functions. A large component (∼40%) of the mammalian genome consists of transposable elements [49]. The expression of these elements can lead to transposition and hence potentially harmful mutations [50]. In somatic cells, the typically high levels of methylation of these elements are considered to be fundamentally important to the integrity of the genome by ensuring the transcriptional silencing of these elements [51]. If this is true than the developmentally associated rounds of global demethylation in the pluripotent ICM and PGCs, which involves many of the transposable elements, presents a particular challenge to genomic integrity. In PGCs a small number of loci have been identified that do rely primarily on DNA methylation for their expression [46]. These are enriched for genes that are associated with genome defense against transposable elements. It seems likely that the

267

demethylation of the promoters of these genes results in their increased expression and so serves to protect the genome from the dangers of inadvertent transposable element expression that might otherwise occur due to the hypomethylated state of these pluripotent cells. These methylation-dependent genes therefore act as part of a fail-safe mechanism to repress transposable element activity upon global demethylation [46]. The significant nature of the genetic risks created by hypomethylation and the mobilization of protected measures in pluripotent cells provides a compelling argument that to warrant such a risk of harm global demethylation during pluripotency must provide essential epigenetic information to the developing embryo. To date, however, there is no unequivocal evidence of what this essential information is. One possibility may be that embryo cells do not have a mechanism that can allow them to precisely target loss of methylation at the large number of specific loci and elements that define the sperm and oocyte cells. If so, a global loss of methylation may be the only mechanism capable of ensuring loss of methylation at all the required sites across a few cell cycles. Perhaps the only mechanism available to the cell for such large scale reprogramming is to rather nonselectively wipe the epigenetic slate relatively clean during these key embryonic transitions, which then allows the setting of a new patterns of methylation with each differentiation event during development. The question then becomes how and why are some regions of the genome protected from this demethylation during reprogramming to pluripotency? For instance, the imprinted loci are protected from demethylation during reprogramming to pluripotency in the ICM, but not the PGCs. A mechanistic basis for this selectivity is not completely described. In the early embryo, PGC7 binds to dimethylated histone 3 lysine 7 and protects some regions of the genome from demethylation [52,53]. It has also been proposed that low levels of the related epigenetic modification, 5′-hydroxymethyl cytosine (5hmC) is present at imprinted loci in most tissue and may contribute to their relative resistance to demethylation. By contrast, these loci have increased levels of 5hmC in PGCs prior to the loss of imprinting [54]. This may indicate that modulating 5hmC levels at particular elements can determine whether they are subjected to demethylation at major genetic transitions.

ROLES OF TET-MEDIATED MODIFICATIONS OF 5meC The TET class of enzymes acts to convert 5meC to 5hmC [and sequentially to 5′-formylcytosine (5fC) and 5′-carboxylcytosine (5caC)] by a Fe(II) and α-ketogluteratedependent mechanism [55,56]. Three gene members of this

V.  Factors Influencing Epigenetic Changes

268

17.  Dynamic Changes in DNA Modifications During Key Embryonic Transitions

class have been identified (Tet 1, 2, or 3). Overexpression of the Tet genes can cause increased levels of 5hmC and is associated with a corresponding decrease in 5meC [55]. Such findings prompted the suggestion that TET might provide a basis for active demethylation [57]. TET catalysed modifications are substrates for thymine-DNA glycosylase [58–60]. This enzyme conventionally acts as a G/T mismatch-specific thymine DNA glycosylase, but can also efficiently excise 5fC and 5caC [59,61,62]. Repair of the resulting basic sites by the insertion of cytosine completes the demethylation process (Fig. 17.3). This pathway indicates a potential role for these TET-mediated metabolites to serve as intermediates in a demethylation pathway, yet it is also increasingly recognized that 5hmC can act as a stable cytosine modification that serves a host of epigenetic roles in its own right [24,63–65]. TET1 is present in nuclei of 1-cell through to blastocysts and is enriched in ICM relative to TE [55] and TET3 is present in the oocyte and zygote (TET3) [66–68]. There is a significant amount of 5hmC in the embryo and throughout preimplantation development [14,22,24,66]. Mass spectrometry estimates that 0.09% of possible CpGs in zygotes are hydroxymethylated [69]. Genomewide base level analysis [22] gave a lower estimate of ∼0.049% (∼13% of all methylated CpGs) in paternal and 0.02% (∼4%) in the maternal genomes of the early mouse embryo. Most of classes of genomic elements contained hydroxymethylation (but it was lower in 5′-UTRs and CGIs) [22]. This same analysis found 0.95 × 106 CpGs

were enriched for 5fC, with a similar pattern of distribution across elements as 5hmC. Quantitative analysis of 5caC in embryos is yet to be successfully achieved, but this modification is detected in the early embryo by immunolabeling [22]. A negative correlation between the levels of 5meC and 5hmC might be expected if a primary role for the TET pathway is global demethylation. Yet, changes in global 5hmC levels in the early embryo broadly parallels the changes observed for 5meC [14] (Fig. 17.2). Both are relatively high in the early embryo and both show a marked loss in the cells determined to the ICM [14]. The increased global DNA methylation of the cells of embryo during the first differentiation events was accompanied by increased levels of hydroxymethylation [14]. The relatively high levels of 5hmC, 5fC, and 5caC that persist over the first several cell-cycles of embryo development suggests that TET-mediated modifications are of themselves not a sufficient precondition for demethylation. Furthermore, immunolocalization indicates that while much of the methylated cytosine tended to accumulate within high-intensity staining foci characteristic of heterochromatin, 5hmC showed a more uniform pattern of distribution throughout the nucleoplasm characteristic of euchromatin [14] (Fig. 17.2). This may be consistent with observations in ESCs where 5hmC is particularly associated with genes bodies and gene regulatory elements [70,71], and the well-known association of 5meC with heterochromatin [72,73].

FIGURE 17.3  5meC and its TET-mediated products provide different epigenetic information during development. A schematic representation of the structural and functional interplay between 5meC and its TET-mediated products. TET products are potential targets for base excision allowing conversion back to cytosine. This pathway provides the potential for active demethylation. 5meC and 5hmC are also stable epigenetic modifications that can act as docking sites for selective binding proteins, such as MBD1 and MBD3, respectively. These binding proteins can in turn recruit a range of epigenetic modifiers, which can cause to large-scale conformational changes to chromatin structure, as indicated. 5caC, 5′-Carboxylcytosine; 5fc, 5′-formylcytosine; 5hmC, 5′-hydroxymethyl cytosine; MBD1, methyl binding protein one.

V.  Factors Influencing Epigenetic Changes



ROLES OF TET-MEDIATED MODIFICATIONS OF 5meC

The methyl group added to 5′-cytosine alters the 3D structure of DNA, potentially changing the affinity for transcription factors and other DNA binding proteins, and methylated cytosines also modify the nucleosome stability and so likely influences accessibility in this way [74]. An important feature of methylation is that this modification lies within the major groove of DNA [74] and so provides a significant topological feature. As such it is a docking site for a range of binding proteins that contain the methyl-CpG binding domain (MBD) [75]. An example of the functional significance of these interactions is the binding of methyl binding protein one (MBD1) to 5meC. MBD1 in turn recruits a range of chromatin modifies, including the histone 3 lysine 9 methylase, SETDB1, which induces the formation of stable heterochromatin in regions of DNA methylation [76,77]. In some contexts the loss of MBD1 can cause failure of normal differentiation [78] indicating that such interactions are one mechanism that decodes the methylomes epigenetic information. The range of binding proteins for 5hmC is less well established. MBD3 is a likely target and appears to be enriched within euchromatin [79–81]. SALL4 is downstream pluripotency factor and is shown to preferentially associate with 5hmC in vitro [82]. It binds to enhancers in ES cells where it prevents precocious activation of differentiation pathways [83]. SALL4 binding to enhancers is dependent upon Tet1 in ESCs, and it seems that this binding facilitates further oxidation of 5hmC in a TET2-dependent manner [82]. SALL4 is known to interact with NANOG in ESCs [82] and in this way perhaps refines the targeting of enhancers. So it seems that SALL4 provides a potential mechanism for site-specific stepwise oxidation of 5meC at enhancers that may allow the demethylation of these sites. Further research is required to determine whether other such chromatin factors target 5hmC oxidation across the range of genetic elements. While it has been long known that DNA methylation plays essential roles in maintaining genomic stability, 5hmC localizes to sites of DNA damage and repair and TET deficiency causes segregation defects in response to replication stress [81]. Thus, the Tet mediated modifications are likely to play multiple roles in the creation of a cell’s epigenetic landscape, with important roles in defining large-scale chromatin structure and defining genomic stability, as well as facilitating site-specific demethylation. A role in global demethylation processes is yet to be unequivocally established. Evidence for a role for TET in the generation of the pluripotent state is equivocal. The expression of Tet1 and Tet2 occurs in ES cells [55]. The knockdown of TET1 (but not 2 or 3) resulted in a modest decrease in pluripotency markers and favored the formation of the trophectoderm over ICM in blastocysts [55]. The induc-

269

tion of differentiation in ESCs was accompanied by a reduction in Tet1 and TET1. TET1 binds to proximal and distal tissue-specific differentially methylated regions of the pluripotency gene Nanog [55]. TET1 and NANOG cooccupy loci associated with both pluripotency and lineage commitment in embryonic stem cells, and TET1 binding to these sites is reduced upon NANOG depletion [84]. In overexpression models, the important epigenetic modifier, PRDM14, is found to coimmunoprecipitate with TET1 and TET2 through its N-terminal domain, and this is enhanced by the binding of PRDM14 to DNA [85]. This may be of particular relevance since PRDM14 expression shows a strong negative associated with the expression of Dnmt3a and 3b in embryos [86], raising the possibility that a role for the TET enzymes is to modulate 5meC levels via the repression of de novo methylases. TET1 can also cooperate with the polycomb protein, PRC2, and the corepressor SIN3A to regulate gene expression [85]. Thus, the potential exists for the TET enzymes to regulate methylation more indirectly by regulating the de novo DNA methylases and by guiding histone modifications and recruitment of corepressors. ESCs with the triple Tet deletion show complete loss of 5hmC from cells, and some increase in global levels of 5meC [87]. These cells maintained normal ESC morphology, expression of pluripotency markers Oct4 and Nanog, and capacity to form embryoid bodies [87]. As was the case for the loss of all Dnmts from ESCs [34], triple knockout of the Tets resulted in cells with a poor capacity to differentiate in vitro and limited capacity for chimera formation. Base-level analysis showed that in embryoid bodies the absence of TETs resulted in the hypermethylation of the promoter regions of a subset of developmentally important genes [87]. This effect of the TET enzyme family may account for the aberrations in the differentiation capacity of the cells. Studies of induced pluripotent cells, iPS, implicate a role for TET activity in there reprogramming to pluripotency. During iPS reprogramming there is increased expression of Tet1 and Tet2 but a decrease in Tet3 [88]. The knockdown of Tet1 reduces the efficiency of reprogramming, while the methylated regions of Oct4 in somatic cells were converted to 5hmC during iPS cell reprogramming [88]. Interestingly, ectopic expression of Tet1 can replace Oct4 in iPS reprogramming and induces 5mC to 5hmC conversion at Oct4 regulatory regions [45], suggesting important interactions between TET and the regulation of key pluripotency genes [45]. Tet1/2-null embryos develop to seemingly normal blastocysts and efficient derivation of pluripotent ESCs from Tet-null blastocysts can be achieved despite the complete depletion of 5hmC from the cells [89]. This finding seemed consistent with the marked loss of 5hmC from the ICM [14] (Fig. 17.2). Tet-double knockout

V.  Factors Influencing Epigenetic Changes

270

17.  Dynamic Changes in DNA Modifications During Key Embryonic Transitions

mice are also fertile indicating normal development of PGCs, despite the absence of 5hmC from these mutant cells. Some embryos, however, showed developmental defects in mid-gestation [89] perhaps suggesting some specific roles for TET enzymes during differentiation of some lineages. One important phenotype of these Tet1/2 double knockout mice was the observation of an increased incidence of imprinting errors [89], implicating an important role for these enzymes in the normal establishment of parent-of-origin imprints. This finding is consistent with observations that imprinted loci are relatively depleted of 5hmC in somatic cell types but become highly enriched for 5hmC in PGCs around the time of imprint erasure [54]. In Tet1/2 double knockout ESCs there is a compensatory increase in the expression of Tet3 yet the cells remain depleted of 5hmC [89], indicating that TET 1 and 2 play the dominant role in pluripotent cells. Tet1/2 double knockout mice have detectable 5hmC in some somatic cells, including sperm, indicating a role for TET3 in some differentiated tissue [89]. Tet3-null embryos are neonatal lethal, but they show seemingly normal preimplantation embryo development [67]. TET3 is present in the oocyte and zygote and the absence of Tet3 is reported to prevent the demethylation of pluripotency genes Nanog and Lemd1, and retards the demethylation of Oct4 [67]. In contrast to these claims for a role of Tet3 in global demethylation in the embryo are the observations that change in paternal 5mC in the zygote do not require 5hmC formation [69]. Inhibition of Tet3 reduces 5hmC accumulation but does affect the change in paternal 5mC. 5hmC in the zygotes is dependent on the activity of zygotic Dnmt3a and Dnmt1 suggesting perhaps that Tet3 has a role in targeting de novo methylation in the early embryo [69]. Yet, the biological significance of TET action in the embryo remains unclear given the apparent viability of the knockout of individual Tets within embryos. Triple Tet knockout embryos do seem to have limited viability to the blastocyst stage [87], although the nature of the apparent lethality has not been defined [87]. Despite some tantalizing experimental implication of TET-mediated modifications of cytosine in epigenetic reprogramming to pluripotency, currently an unequivocal role for TETs not established. The evidence points to a role in the differentiation of some cell lineages, sitespecific demethylation at some developmentally important genes, and in the clearance of imprinting marks in PGCs. It is an open question as to whether TET acts as a primary epigenetic mediator and/or has other important functions in chromatin organization in early embryo cells. The balance of evidence seems to point to TET-mediated modifications of 5meC creating information important for differentiation, with considerable redundancy in the actions between each member of this class of enzymes.

STRESS DURING KEY EMBRYONIC TRANSITIONS HAVE LIFELONG CONSEQUENCES The sheer magnitude of changes to CpG modifications across the genome during epigenetic reprogramming to and from pluripotency point to quality assurance challenges second only to that involved in DNA replication. While we now have a deep understanding of the quality assurance of DNA replication there is little corresponding knowledge about the processes that ensure the fidelity of epigenetic programming. Twin studies show that a considerable amount of the epigenetic variability between individuals is heritable. There is a much stronger correlation in methylation patterns between monozygotic than dizygotic twins at birth [90–92]. Analysis of monozygotic twins that were discordant for birth weight, however, yielded greater variability in methylation patterns [93]. Heritability of DNA methylation in complex family pedigrees indicates that around 80% of the variability in CpG methylation is primarily determined by the genotype and environmental influences are assumed to account for much of the remaining 20% [91]. It is also clear, however, that some aspects of development have a significant stochastic element. There is, for example, random variability in transcription, translation, and epigenetic modifications [92,94,95], and these can potentially amass in a drift-like fashion [92]. This stochastic developmental variation is considered as the third source of phenotypic variation [96]. Some of the epigenetic discordance observed between monozygotic twins may be accounted for by this stochasticity. The relative contribution of environmental effects and stochastic drift are at this time difficult to quantify. The extent to which epigenetic noise is accounted for by imprecision in the targeting of methylation, errors in proofreading and editing of the process during DNA replication, or errors in site-specific demethylation during differentiation remain largely open questions. It is important to gain an understanding of whether the environmentally induced epigenetic variability that occurs during development is due to exacerbation of these processes, by independent mechanisms, or some combination of both. Analysis of individuals that were conceived during the Dutch Hunger Winter in 1944–45 have been particularly informative on the environmental effects on epigenetic programming. These studies show that early-life environmental conditions, such as nutrition level, can cause epigenetic changes in humans that persist throughout their life and cause marked changes in homeostatic stability [97]. Importantly, famine exposure during gestation weeks 1–10, but not weeks 11–20, 21–30, or 31-delivery, was associated with an increase in DNA methylation at a range of CpG dinucleotides in adults

V.  Factors Influencing Epigenetic Changes



271

Conclusions

[98]. This evidence points to the period during development, when greatest epigenetic reprogramming occurs, as the period that is most sensitive to environmental perturbation. Another natural experiment involves examination of the offspring of rural Gambian women who are exposed to seasonal differences in methyl-donor intake in their diet, when this occurs during the periconceptual period, this seasonal effect resulted in persistent and systemic changes in the patterns of methylation to a range of metastable epialleles in their children [99]. Thus, changes in nutrition during the times of epigenetic reprogramming can change the patterns of this reprogramming and result in lifelong changes in homeostasis. A clinically relevant model that involves profound alterations to embryo nutrition is their creation by assisted reproductive technologies (ART), such as in vitro fertilization. These technologies, particularly in ovine and bovine models, cause a proportion of progeny to have “large offspring syndrome” (LOS) [100]. This is manifested as fetal and pediatric overgrowth and is generally associated with a range of congenital malfunctions and poor health outcomes [101]. Although most evident in assisted reproduction in ruminants, it has been identified to occur to some extent in all mammalian species studied. In general, it seems that the more extreme the intervention the greater the risk and degree of fetal overgrowth. For example, in somatic cell nuclear transfer, where an unnatural process of epigenetic reprogramming occurs, there is commonly a high incidence of large offspring syndrome [102]. Epigenetic reprogramming errors are evident in some human IVF offspring, with reports of an elevated occurrence of fetal overgrowth syndromes [103,104]. In humans, disorders such as Beckwith–Wiedemann syndrome (BWS) are considered to be manifestation of LOS. BWS has a low frequency in the general population but ART is associated with 3–16fold increased incidence [105]. Most BWS cases are related with loss of methylation at the KvDMR1 locus and a smaller proportion have a gain of methylation of the H19/IGF2 locus (20) [105,106]. LOS may result primarily from failure of the normal processes of parental imprinting of genes. The detected increase rate of imprinting errors in ART offspring may reflect the ready detection of such errors due to the normal pattern of monoallelic expression of these genes. This raises the question, does a wider range of epigenetic perturbation to biallelic loci occur but are less phenotypically evident. IVF or culture of embryos in vitro causes a change in the pattern of expression of a range of genes that are not imprinted [107,108] and it will be of interest to determine how much of this transcriptional effect has an epigenetic basis. Mouse models show that ART leads to altered postnatal growth trajectories and changed organ allometery [5]. Most importantly, this effect can persist across into the next generation, even after natural conception.

This transgenerational effect is strongly suggestive of an epigenetic origin of these changes [5]. The analysis of epigenetically labile genes is useful in addressing this point. These loci are known as metastable epialleles [109]. In mice, a canonical example of this class of genes is the agouti viable yellow (Avy) locus [110]. Genetically identical mice carrying this locus show varying phenotypes; with changed coat color, body weight, and predisposition to diabetes [110]. These changes are determined by the methylation status of the promoter of the agouti gene. In this model, embryo culture from the zygote to blastocyst stage caused a 3–4 fold change in the extent of methylation reprogramming of this gene compared to embryos subjected to embryo transfer without culture [4]. Infinium HumanMethylation450k BeadChip assays [111] of human postnatal blood reveals that ART causes differential methylation at CpG clusters across the genome. This was more common within gene bodies than promoter regions and was notable across a range of imprinted loci. A range of metastable epialleles was also found to be enriched for differential methylation after ART [111]. The strong and building body of evidence for a degree reduced fidelity of epigenetic reprogramming squarely places this as a predisposing factor for some serious diseases, including a range of chronic conditions such as diabetes and cardiovascular disease. Important outstanding questions are whether the epigenetic changes detected in response to environmental stress represent: (1) programmed adaptive alterations to the epigenetic program in response to the environmental stress that become maladaptive postnataly; (2) exacerbation of the normal level of epigenetic noise within biological systems; (3) increased errors in the implementation of the epigenetic program, and/or (4) failures in the normal processes of quality control of epigenetic programming.

CONCLUSIONS DNA methylation is a key epigenetic modification essential for the maintenance of genomic integrity, and has an important role in the formation of several pervasive epigenetic features, including heterochromatin formation, parent-of-origin genomic imprinting, and dosagecompensation X-chromosome inactivation (in females). The formation of the two-pluripotent cell populations within the early embryo is accompanied by the profound loss of methylation across much of the genome. The formation of differentiated cell lineages from these pluripotent cells is accompanied by extensive lineagespecific patterns of remethylation and this defines the patterns of gene expression required for lineage specification. The extensive level of remodeling of methylation during transition to the pluripotent state creates a

V.  Factors Influencing Epigenetic Changes

272

17.  Dynamic Changes in DNA Modifications During Key Embryonic Transitions

susceptibility to maladaptive reprogramming. A host of common cellular stressors can alter the fidelity of reprogramming and can predispose individuals to lifelong patterns of homeostatic instability and chronic disease. Recent reanalyses of demethylation in the preimplantation embryo shows that it is an event more closely associated with the formation of the pluripotent state rather than the totipotent state, and this is consistent with the demethylation that is associated with the formation of the pluripotent PGCs. The global demethylation of pluripotent cells and their remethylation with differentiation provides a strong circumstantial case for an essential role for this epigenetic modification in embryo development. For many years the role of methylation in the differential activity of gene promoters, such as Oct4, has been a favorite hypothetical target of regulation. Yet, recent genome-wide analysis indicates that changes in methylation of promoters are rather modest compared with most other elements within the genome. These new results indicate that new models for thinking about the processes of epigenetic reprogramming during embryo development are required. Currently, a major limiting factor to the field is a detailed understanding of the processes for regulating site-specific demethylation during major epigenetic transitions. One possible explanation for the pervasive global loss of methylation associated with the formation of pluripotent ICM and PGC cells may be that it is the most efficient method to remodel a large number of elements over a few cell-cycles, and in so doing removes the epigenetic memory of the sperm and egg, except at the small number of imprinted loci. Differentiation of pluripotent cells can then be achieved by transcription factor-directed reestablishment of site-specific methylation as is required for each lineage’s pattern of gene expression. The manner by which context-specific cues that direct differentiation are decoded with high fidelity to produce lineage-specific patterns of remethylation is a major question for the field. The advent of powerful new analytical tools will ensure this is a field of exponential growth of insights in the coming decade.

References [1] Xie P, Zang L-Q, Li X-K, Shu Q. An epigenetic view of developmental diseases: new targets, new therapies. World J Pediatr 2016;12(3):291–7. [2] Dolinoy DC, Weidman JR, Jirtle RL. Epigenetic gene regulation: linking early developmental environment to adult disease. Reprod Toxicol 2007;23(3):297–307. [3] Dolinoy DC, Huang D, Jirtle RL. Maternal nutrient supplementation counteracts bisphenol A-induced DNA hypomethylation in early development. Proc Nat Acad Sci 2007;104(32):13056–61. [4] Morgan HD, Jin XL, Li A, Whitelaw E, O’Neill C. The culture of zygotes to the blastocyst stage changes the postnatal expression of an epigentically labile allele, agouti viable yellow, in mice. Biol Reprod 2008;79(4):618–23.

[5] Mahsoudi A, Li A, O’Neill C. Assessment of the long-term and transgenerational consequences of perturbing preimplantation embryo development in mice. Biol Reprod 2007;77:889–96. [6] Schultz RM. The molecular foundations of the maternal to zygotic transition in the preimplantation embryo. Hum Reprod Update 2002;8(4):323–31. [7] Latham KE, Schultz RM. Embryonic genome activation. Front Biosci 2001;6:D748–59. [8] Evsikov AV, de Vries WN, Peaston AE, Radford EE, Fancher KS, Chen FH, et al. Systems biology of the 2-cell mouse embryo. Cytogenet Genome Res 2004;105:240–50. [9] Rossant J. Developmental biology: a mouse is not a cow. Nature 2011;471(7339):457–8. [10] Gardner RL. Asymmetry and prepattern in the mammalian development. In: Rossant J, Tam P, editors. Mouse Development: Patterning, Morphogenesis, and Organogenesis. San Diego: Academic Press; 2002. [11] Hajkova P, Erhardt S, Lane N, Haaf T, El-Maarri O, Reik W, et al. Epigenetic reprogramming in mouse primordial germ cells. MechDev 2002;117(1–2):15–23. [12] Kobayashi H, Sakurai T, Miura F, Imai M, Mochiduki K, Yanagisawa E, et al. High-resolution DNA methylome analysis of primordial germ cells identifies gender-specific reprogramming in mice. Genome Res 2013;23(4):616–27. [13] Smith ZD, Chan MM, Mikkelsen TS, Gu H, Gnirke A, Regev A, et al. A unique regulatory phase of DNA methylation in the early mammalian embryo. Nature 2012;484(7394):339–44. [14] Li Y, Seah MKY, O’Neill C. Mapping global changes in nuclear cytosine base modifications in the early mouse embryo. Reproduction 2016;151(2):83–95. [15] Stewart CL, Gadi I, Bhatt H. Stem-cells from primordial germcells can reenter the germ-line. Dev Biol 1994;161(2):626–8. [16] Wood SA, Allen ND, Rossant J, Auerbach A, Nagy A. Non-injection methods for the production of embryonic stem cell-embryo chimaeras. Nature 1993;365:87–9. [17] Hayashi K, Ogushi S, Kurimoto K, Shimamoto S, Ohta H, Saitou M. Offspring from oocytes derived from in vitro primordial germ cell-like cells in mice. Science 2012;338(6109):971–5. [18] Barlow DP, Bartolomei MS. Genomic imprinting in mammals. Cold Spring Harb Perspect Biol 2014;6(2):1–20. [19] Bartolomei MS, Webber AL, Brunkow ME, Tilghman SM. Epigenetic mechanisms underlying the imprinting of the mouse H19 gene. Genes Dev 1993;7(9):1663–73. [20] Holliday R, Pugh JE. DNA modification mechanisms and gene activity during development. Science 1975;187(4173):226–32. [21] Morgan HD, Santos F, Green K, Dean W, Reik W. Epigenetic reprogramming in mammals. Hum Mol Genet 2005;14(Suppl. 1):R47–58. [22] Wang L, Zhang J, Duan J, Gao X, Zhu W, Lu X, et al. Programming and inheritance of parental DNA methylomes in mammals. Cell 2014;157(4):979–91. [23] Li Y, O’Neill C. Persistence of cytosine methylation of DNA following fertilisation in the mouse. PLoS One 2012;7(1):e30687. [24] Li Y, O’Neill C. 5-methylcytosine and 5-hydroxymethylcytosine each provide epigenetic information to the mouse zygote. PLoS One 2013;8(5):e63689. [25] Cockburn K, Rossant J. Making the blastocyst: lessons from the mouse. J Clin Invest 2010;120(4):995–1003. [26] Hirasawa R, Sasaki H. Dynamic transition of Dnmt3b expression in mouse pre- and early post-implantation embryos. Gene Expr Patterns 2009;9(1):27–30. [27] Smith ZD, Chan MM, Humm KC, Karnik R, Mekhoubad S, Regev A, et al. DNA methylation dynamics of the human preimplantation embryo. Nature 2014;511:611–5. [28] Habibi E, Brinkman AB, Arand J, Kroeze LI, Kerstens HHD, Matarese F, et al. Whole-genome bisulfite sequencing of two

V.  Factors Influencing Epigenetic Changes



REFERENCES

[29]

[30]

[31]

[32]

[33]

[34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

[45]

[46]

distinct interconvertible DNA methylomes of mouse embryonic stem cells. Cell Stem Cell 2013;13(3):360–9. Leitch HG, McEwen KR, Turp A, Encheva V, Carroll T, Grabole N, et al. Naive pluripotency is associated with global DNA hypomethylation. Nat Struct Mol Biol 2013;20:311–6. Ying Q-L, Wray J, Nichols J, Batlle-Morera L, Doble B, Woodgett J, et al. The ground state of embryonic stem cell self-renewal. Nature 2008;453(7194):519–23. Kagiwada S, Kurimoto K, Hirota T, Yamaji M, Saitou M. Replication-coupled passive DNA demethylation for the erasure of genome imprints in mice. EMBO J 2013;32(3):340–53. Seki Y, Hayashi K, Itoh K, Mizugaki M, Saitou M, Matsui Y. Extensive and orderly reprogramming of genome-wide chromatin modifications associated with specification and early development of germ cells in mice. Dev Biol 2005;278(2):440–58. Jackson M, Krassowska A, Gilbert N, Chevassut T, Forrester L, Ansell J, et al. Severe global DNA hypomethylation blocks differentiation and induces histone hyperacetylation in embryonic stem cells. Mol Cell Biol 2004;24(20):8862–71. Tsumura A, Hayakawa T, Kumaki Y, Takebayashi S-i, Sakaue M, Matsuoka C, et al. Maintenance of self-renewal ability of mouse embryonic stem cells in the absence of DNA methyltransferases Dnmt1, Dnmt3a and Dnmt3b. Genes Cells 2006;11(7):805–14. Liao J, Karnik R, Gu H, Ziller MJ, Clement K, Tsankov AM, et al. Targeted disruption of DNMT1, DNMT3A and DNMT3B in human embryonic stem cells. Nat Genet 2015;47(5):469–78. Roadmap Epigenomics C, Kundaje A, Meuleman W, Ernst J, Bilenky M, Yen A, et al. Integrative analysis of 111 reference human epigenomes. Nature 2015;518(7539):317–30. Oda M, Kumaki Y, Shigeta M, Jakt LM, Matsuoka C, Yamagiwa A, et al. DNA methylation restricts lineage-specific functions of transcription factor Gata4 during embryonic stem cell differentiation. PLoS Genet 2013;9(6):e1003574. Mikkelsen TS, Ku M, Jaffe DB, Issac B, Lieberman E, Giannoukos G, et al. Genome-wide maps of chromatin state in pluripotent and lineage-committed cells. Nature 2007;448(7153):553–60. Zhao XD, Han X, Chew JL, Liu J, Chiu KP, Choo A, et al. Wholegenome mapping of histone H3 lys4 and 27 trimethylations reveals distinct genomic compartments in human ebryonic stem cells. Cell Stem Cell 2007;1(3):286–98. Pan G, Tian S, Nie J, Yang C, Ruotti V, Wei H, et al. Whole-genome analysis of histone H3 lysine 4 and Lysine 27 methylation in human embryonic stem cells. Cell Stem Cell 2007;1(3):299– 312. Wong JCY, Jack MM, Li Y, O’Neill C. The epigenetic bivalency of core pancreatic β-cell transcription factor genes within mouse pluripotent embryonic stem cells is not affected by knockdown of the polycomb repressive complex 2, SUZ12. PLoS One 2014;9(5):e97820. Tsankov AM, Gu H, Akopian V, Ziller MJ, Donaghey J, Amit I, et al. Transcription factor binding dynamics during human ES cell differentiation. Nature 2015;518(7539):344–9. Gidekel S, Bergman Y. A unique developmental pattern of Oct3/4 DNA methylation is controlled by a cis-demodification element. J Biol Chem 2002;277(37):34521–30. Hattori N, Nishino K, Ko YG, Ohgane J, Tanaka S, Shiota K. Epigenetic Ccntrol of mouse Oct-4 gene expression in embryonic stem cells and trophoblast stem cells. J Biol Chem 2004;279(17):17063–9. Radzisheuskaya A, Silva JCR. Do all roads lead to Oct4? The emerging concepts of induced pluripotency. Trends Cell Biol 2014;24(5):275–84. Hackett JA, Reddington JP, Nestor CE, Dunican DS, Branco MR, Reichmann J, et al. Promoter DNA methylation couples genomedefence mechanisms to epigenetic reprogramming in the mouse germline. Develop 2012;139(19):3623–32.

273

[47] Whyte WA, Orlando DA, Hnisz D, Abraham BJ, Lin CY, Kagey MH, et al. Master transcription factors and mediator establish super-enhancers at key cell identity genes. Cell 2013;153(2):307– 19. [48] Jackson-Grusby L, Beard C, Possemato R, Tudor M, Fambrough D, Csankovszki G, et al. Loss of genomic methylation causes p53-dependent apoptosis and epigenetic deregulation. Nat Genet 2001;27(1):31–9. [49] Waterston R, Lindblad-Toh K, Birney E, Rogers J, Abril JF, Agarwal P, et al. Initial sequencing and comparative analysis of the mouse genome. Nature 2002;420(6915):520–62. [50] Huang CRL, Burns KH, Boeke JD, NB A, SA S, DK G, et al. Active transposition in genomes rodent L1 evolution has been driven by a single dominant lineage that has repeatedly acquired new transcriptional regulatory sequences. Ann Rev Genet 2012;46(1):651–6575. [51] Yoder JA, Walsh CP, Bestor TH. Cytosine methylation and the ecology of intragenomic parasites. Trends Genet 1997;13(8):335– 40. [52] Nakamura T, Liu Y-J, Nakashima H, Umehara H, Inoue K, Matoba S, et al. PGC7 binds histone H3K9me2 to protect against conversion of 5mC to 5hmC in early embryos. Nature 2012;486(7403):415–9. [53] Nakamura T, Arai Y, Umehara H, Masuhara M, Kimura T, Taniguchi H, et al. PGC7/Stella protects against DNA demethylation in early embryogenesis. Nat Cell Biol 2006;9:64–71. [54] Hackett JA, Dietmann S, Murakami K, Down TA, Leitch HG, Surani MA. Synergistic mechanisms of DNA demethylation during transition to ground-state pluripotency. Stem Cell Rep 2013;1(6):518–31. [55] Ito S, D’Alessio AC, Taranova OV, Hong K, Sowers LC, Zhang Y. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature 2010;466(7310):1129–33. [56] Tahiliani M, Koh KP, Shen Y, Pastor WA, Bandukwala H, Brudno Y, et al. Conversion of 5-methylcytosine to 5-hydroxymethylcytosine in mammalian DNA by MLL partner TET1. Science 2009;324(5929):930–5. [57] Auclair G, Weber M. Mechanisms of DNA methylation and demethylation in mammals. Biochimie 2012;94(11):2202. [58] Kohli RM, Zhang Y. TET enzymes, TDG and the dynamics of DNA demethylation. Nature 2013;502:472–9. [59] He Y-F, Li B-Z, Li Z, Liu P, Wang Y, Tang Q, et al. Tet-mediated formation of 5-carboxylcytosine and its excision by TDG in mammalian DNA. Science 2011;333(6047):1303–7. [60] Shen L, Wu H, Diep D, Yamaguchi S, D’Alessio Ana C, Fung H-L, et al. Genome-wide analysis reveals TET- and TDG-dependent 5-methylcytosine oxidation dynamics. Cell 2013;153(3):692–706. [61] Zhang L, Lu X, Lu J, Liang H, Dai Q, Xu G-L, et al. Thymine DNA glycosylase specifically recognizes 5-carboxylcytosinemodified DNA. Nat Chem Biol 2012;8(4):328–30. [62] Cortellino S, Xu J, Sannai M, Moore R, Caretti E, Cigliano A, et al. Thymine DNA glycosylase is essential for active DNA demethylation by linked deamination-base excision repair. Cell 2011;146(1):67–79. [63] Bachman M, Uribe-Lewis S, Yang X, Williams M, Murrell A, Balasubramanian S. 5-hydroxymethylcytosine is a predominantly stable DNA modification. Nat Chem 2014;6(12):1049–55. [64] Kubiura M, Okano M, Kimura H, Kawamura F, Tada M. Chromosome-wide regulation of euchromatin-specific 5mC to 5hmC conversion in mouse ES cells and female human somatic cells. Chromosome Res 2012;20:1–12. [65] Vella P, Scelfo A, Jammula S, Chiacchiera F, Williams K, Cuomo A, et al. Tet proteins connect the O-linked N-acetylglucosamine transferase Ogt to chromatin in embryonic stem cells. Mol Cell 2013;49(4):645–56.

V.  Factors Influencing Epigenetic Changes

274

17.  Dynamic Changes in DNA Modifications During Key Embryonic Transitions

[66] Iqbal K, Jin S-G, Pfeifer GP, Szabó PE. Reprogramming of the paternal genome upon fertilization involves genomewide oxidation of 5-methylcytosine. Proc Natl Acad Sci USA 2011;108(9):3642–7. [67] Gu T-P, Guo F, Yang H, Wu H-P, Xu G-F, Liu W, et al. The role of Tet3 DNA dioxygenase in epigenetic reprogramming by oocytes. Nature 2011;477:606–10. [68] Inoue A, Matoba S, Zhang Y. Transcriptional activation of transposable elements in mouse zygotes is independent of Tet3-mediated 5-methylcytosine oxidation. Cell Res 2012;22(12):1640–9. [69] Amouroux R, Nashun B, Shirane K, Nakagawa S, Hill PWS, D’Souza Z, et al. De novo DNA methylation drives 5hmC accumulation in mouse zygotes. Nat Cell Biol 2016;18(2):225–33. [70] Song C-X, Szulwach Keith E, Dai Q, Fu Y, Mao S-Q, Lin L, et al. Genome-wide profiling of 5-formylcytosine reveals its roles in epigenetic priming. Cell 2013;153(3):678–91. [71] Yu M, Hon GC, Szulwach KE, Song C-X, Zhang L, Kim A, et al. Base-resolution analysis of 5-hydroxymethylcytosine in the mammalian genome. Cell 2012;149(6):1368–80. [72] Rose NR, Klose RJ. Understanding the relationship between DNA methylation and histone lysine methylation. Biochim Biophys Acta 2014;1839(12):1362–72. [73] Celik S, Li Y, O’Neill C. The exit of mouse embryonic fibroblasts from the cell-cycle changes the nature of solvent exposure of the 5′-methylcytosine epitope within chromatin. PLoS One 2014;9(4):e92523. [74] Dantas Machado AC, Zhou T, Rao S, Goel P, Rastogi C, Lazarovici A, et al. Evolving insights on how cytosine methylation affects protein–DNA binding. Brief Funct Genomics 2015;14(1):61–73. [75] Zou X, Ma W, Solov’yov IA, Chipot C, Schulten K. Recognition of methylated DNA through methyl-CpG binding domain proteins. Nucleic Acids Res 2012;40(6):2747–58. [76] Sarraf SA, Stancheva I. Methyl-CpG binding protein MBD1 couples histone H3 methylation at lysine 9 by SETDB1 to DNA replication and chromatin assembly. Mol Cell 2004;15(4):595–605. [77] Fujita N, Watanabe S, Ichimura T, Tsuruzoe S, Shinkai Y, Tachibana M, et al. Methyl-CpG binding domain 1 (MBD1) interacts with the Suv39h1-HP1 heterochromatic complex for DNA methylation-based transcriptional repression. J Biol Chem 2003;278(26):24132–8. [78] Liu C, Teng Z-Q, Santistevan NJ, Szulwach KE, Guo W, Jin P, et al. Epigenetic regulation of miR-184 by MBD1 governs neural stem cell proliferation and differentiation. Cell Stem Cell 2010;6(5):433–44. [79] Shimbo T, Du Y, Grimm SA, Dhasarathy A, Mav D, Shah RR, et al. MBD3 localizes at promoters, gene bodies and enhancers of active genes. PLoS Genet 2013;9(12):e1004028. [80] Yildirim O, Li R, Hung J-H, Chen PB, Dong X, Ee L-s, et al. Mbd3/NURD complex regulates expression of 5-hydroxymethylcytosine marked genes in embryonic stem cells. Cell 2011;147(7):1498–510. [81] Kafer GR, Li X, Horii T, Suetake I, Tajima S, Hatada I, et al. 5-Hydroxymethylcytosine marks sites of DNA damage and promotes genome stability. Cell Rep 2016;14(6):1283–92. [82] Xiong J, Zhang Z, Chen J, Huang H, Xu Y, Ding X, et al. Cooperative action between SALL4A and TET Proteins in stepwise oxidation of 5-methylcytosine. Mol Cell 2016;64(5):913–25. [83] Miller A, Ralser M, Kloet SL, Loos R, Nishinakamura R, Bertone P, et al. Sall4 controls differentiation of pluripotent cells independently of the nucleosome remodelling and deacetylation (NuRD) complex. Develop 2016;143(17):3074–84. [84] Costa Y, Ding J, Theunissen TW, Faiola F, Hore TA, Shliaha PV, et al. NANOG-dependent function of TET1 and TET2 in establishment of pluripotency. Nature 2013;495(7441):370–4. [85] Okashita N, Kumaki Y, Ebi K, Nishi M, Okamoto Y, Nakayama M, et al. PRDM14 promotes active DNA demethylation through

[86]

[87]

[88]

[89]

[90]

[91]

[92]

[93]

[94] [95]

[96]

[97]

[98]

[99]

[100]

[101]

[102] [103] [104]

the ten-eleven translocation (TET)-mediated base excision repair pathway in embryonic stem cells. Develop 2014;141(2):269–80. Burton A, Muller J, Tu S, Padilla-Longoria P, Guccione E, TorresPadilla M-E. Single-cell profiling of epigenetic modifiers identifies PRDM14 as an inducer of cell fate in the mammalian embryo. Cell Rep 2013;5(3):687–701. Dawlaty MM, Breiling A, Le T, Barrasa MI, Raddatz G, Gao Q, et al. Loss of Tet enzymes compromises proper differentiation of embryonic stem cells. Dev Cell 2014;29(1):102–11. Gao Y, Chen J, Li K, Wu T, Huang B, Liu W, et al. Replacement of Oct4 by Tet1 during iPSC induction reveals an important role of DNA methylation and hydroxymethylation in reprogramming. Cell Stem Cell 2013;12(4):453–69. Dawlaty MM, Breiling A, Le T, Raddatz G, Barrasa MI, Cheng AW, et al. Combined deficiency of Tet1 and Tet2 causes epigenetic abnormalities but is compatible with postnatal development. Dev Cell 2013;24(3):310–23. Ollikainen M, Smith KR, Joo EJ-H, Ng HK, Andronikos R, Novakovic B, et al. DNA methylation analysis of multiple tissues from newborn twins reveals both genetic and intrauterine components to variation in the human neonatal epigenome. Hum Mol Genet 2010;19(21):4176–88. Day K, Waite LL, Alonso A, Irvin MR, Zhi D, Thibeault KS, et al. Heritable DNA methylation in CD4+ cells among complex families displays genetic and non-genetic effects. PLoS One 2016;11(10):e0165488. Czyz W, Morahan JM, Ebers GC, Ramagopalan SV. Genetic, environmental and stochastic factors in monozygotic twin discordance with a focus on epigenetic differences. BMC Medicine 2012;10(1):93. Chen M, Baumbach J, Vandin F, Röttger R, Barbosa E, Dong M, et al. Differentially methylated genomic regions in birth-weight discordant twin pairs. Ann Hum Genet 2016;80(2):81–7. Losick R, Desplan C. Stochasticity and cell fate. Science 2008;320(5872):65–8. Raj A, Rifkin SA, Andersen E, van Oudenaarden A. Variability in gene expression underlies incomplete penetrance. Nature 2010;463(7283):913–8. Vogt G. Stochastic developmental variation, an epigenetic source of phenotypic diversity with far-reaching biological consequences. J Biosci 2015;40(1):159–204. Heijmans BT, Tobi EW, Stein AD, Putter H, Blauw GJ, Susser ES, et al. Persistent epigenetic differences associated with prenatal exposure to famine in humans. Proc Natl Acad Sci USA 2008;105(44):17046–9. Tobi EW, Slieker RC, Stein AD, Suchiman HED, Slagboom PE, van Zwet EW, et al. Early gestation as the critical time-window for changes in the prenatal environment to affect the adult human blood methylome. Int J Epidemiol 2015;44(4):1211–23. Dominguez-Salas P, Moore SE, Baker MS, Bergen AW, Cox SE, Dyer RA, et al. Maternal nutrition at conception modulates DNA methylation of human metastable epialleles. Nat Commun 2014;5:3746. Behboodi E, Anderson GB, BonDurant RH, Cargill SL, Kreuscher BR, Medrano JF, et al. Birth of large calves that developed from in vitro-derived bovine embryos. Theriogenology 1995;44(2):227–32. Chavatte-Palmer P, Camous S, Jammes H, Le Cleac’h N, Guillomot M, Lee RSF. Review: Placental perturbations induce the developmental abnormalities often observed in bovine somatic cell nuclear transfer. Placenta 2012;33(Suppl. 0):S99–S104. Young LE, Sinclair KD, Wilmut I. Large offspring syndrome in cattle and sheep. Rev Reprod 1998;3:155–63. Chen Z, Robbins KM, Wells KD, Rivera RM. Large offspring syndrome. Epigenetics 2013;8(6):591–601. Halliday J, Oke K, Breheny S, Algar E, Amor DJ. Beckwith-Wiedemann syndrome and IVF: a case-control study. Am J Hum Genet 2004;75(3):526.

V.  Factors Influencing Epigenetic Changes



REFERENCES

[105] White CR, Denomme MM, Tekpetey FR, Feyles V, Power SGA, Mann MRW. High frequency of imprinted methylation errors in human preimplantation embryos. Sci Rep 2015;5:17311. [106] Chen Z, Hagen DE, Elsik CG, Ji T, Morris CJ, Moon LE, et al. Characterization of global loss of imprinting in fetal overgrowth syndrome induced by assisted reproduction. Proc Nat Acad Sci 2015;112(15):4618–23. [107] Katari S, Turan N, Bibikova M, Erinle O, Chalian R, Foster M, et al. DNA methylation and gene expression differences in children conceived in vitro or in vivo. Hum Mol Genet 2009;18(20):3769–78.

275

[108] Rinaudo P, Schultz RM. Effects of embryo culture on global pattern of gene expression in preimplantation mouse embryos. Reproduction 2004;128(3):301–11. [109] Rakyan VK, Blewitt ME, Druker R, Preis JI, Whitelaw E. Metastable epialleles in mammals. Trends Genet 2002;18(7):348–51. [110] Morgan HD, Sutherland HGE, Martin DIK, Whitelaw E. Epigenetic inheritance at the agouti locus in the mouse. Nat Genet 1999;23(3):314–8. [111] Estill MS, Bolnick JM, Waterland RA, Bolnick AD, Diamond MP, Krawetz SA. Assisted reproductive technology alters deoxyribonucleic acid methylation profiles in bloodspots of newborn infants. Fertil Steril 2016;106(3). 629.e10-639.e10.

V.  Factors Influencing Epigenetic Changes

C H A P T E R

18 Epigenetic Biomarkers Xiaotong Hu, Shuiping Liu Sir Run Run Shaw Hospital, Zhejiang University, Hangzhou, Zhejiang, China

O U T L I N E Introduction

277

Epigenetic Biomarkers Offer Distint Advantages Over Genetic Biomarkers

278

Minimally Invasive Tissues are Suitable for Detecting Epigenetic Biomarkers

278

Field Cancerization and Epigenetic Biomarkers

280

Potential Methylation Biomarkers in Cancer and Other Disease

281

Potential Histone Modification Biomarkers in Cancer and Other Diseases

282

INTRODUCTION An increasing amount of evidence has supported a key role of epigenetic processes in cell biology, physiology, and disease development both in individual genes and at genome-wide scale. Epigenetic mechanisms determine the phenotype without changes in the genotype. DNA methylation is the best studied epigenetic mechanism and acts together with other epigenetic entities: histone modification and noncoding RNAs (ncRNAs) to shape the chromatin structure of DNA according to its functional state. The epigenetic interface sits between the genetic blueprints stored in DNA sequences and phenotypes dictated by the pattern of gene expression. Patients’ individual epigenetic profiling, which is more susceptible to the changing environment, determines disease development and response to treatment. For example, recent studies in gastric cancer (GC) have strongly suggested for the association of the risk factors, such as Helicobacter Pylori infection and a unique aberrant Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00018-3 Copyright © 2017 Elsevier Inc. All rights reserved.

Potential Noncoding RNA Biomarkers in Diseases Potential miRNA Biomarkers in Cancer and Other Diseases Potential Long Noncoding RNA Biomarkers in Cancer and Other Diseases

282 283 285

Epigenetic Biomarker Detection Methods in Clinical Application

287

Challenges and Future Perspectives

288

References

289

DNA methylation pattern [1]. Another example is that Mastroeni et al. determined the status of DNA methylation in monozygotic twins, where genetic similarities exist but discordant for Alzheimer’s disease (AD) proved that epigenetic mechanisms may mediate the effects of life events on AD risk [2]. Epigenetic biomarkers refer to the measurement of epigenetic alterations in tissues or peripheral fluids, such as urine, blood, plasma, serum, and stool samples, as markers of disease detection, progression, and therapy response. Although most research about epigenetic biomarkers has focused on cancers, epigenetic biomarkers have also been used as markers of other diseases, such as neurological, metabolic and cardiovascular disorders, autoimmune diseases, and genomic imprinting disorders. Many diseases may remain asymptomatic and are always diagnosed at later stage lesions and lose the chances for successful interventions. As genetic defects, the epigenetic abnormalities can occur at all the stages of disease progression, such as the dysplasia, local benign

277

278

18.  Epigenetic Biomarkers

and malignant tumor, and finally metastatic lesions of cancer. Moreover, epigenetic alterations are environmentally regulated and provide a plausible link between the environment epigenetics and disease susceptibility. In addition, epigenetic changes are reversible makes them an attractive target for disease therapy. So the epigenetic biomarkers have a great value in disease screening, early detection, diagnosis, staging, risk stratification, and treatment monitoring. Nowadays, knowledge of epigenetic alterations in disease is rapidly increasing owing to the development of genome-wide techniques for their identification. Epigenetic biomarkers offer distinct advantages over genetic biomarkers and they can be detected and quantified by numerous technologies including genome-wide screening methods, as well as locus- or gene-specific high-resolution analysis in different tissue samples and body fluids obtained through minimally invasive procedures making them very suitable biomarkers. However, many studies were carried out in small samples and only a few results have been appropriately validated in independent cohorts or investigated in large perspective clinical trials. It is important to notice that there are a number of limitations to overcome before these potential biomarkers reach the clinic. This chapter will present systemic information of resent research progress on epigenetic biomarkers, summarize the advantages of epigenetic biomarkers over other biomarkers such as genetic biomarkers, the performance in liquid biopsy, and their potential clinical application and challenges in cancer and other diseases.

EPIGENETIC BIOMARKERS OFFER DISTINCT ADVANTAGES OVER GENETIC BIOMARKERS First, genetic biomarkers (such as loss of heterozygosity and mutation) are always masked by normal cells while epigenetic biomarkers are not [3]. The readout of epigenetic biomarkers doesn’t suffer from sample contamination and illegitimate expression. Second, the genetic defects within a gene may involve numerous regions, not to mention its noncoding region, which has not or rarely been surveyed. For example, over 23,544 somatic mutations and 376 germline mutations have been reported for the p53-coding region in cancer. At the present time, it is not possible to cover all the known mutation in genes analyzed even with the most powerful platform technology. On the contrary, epigenetic defects like hypermethylation are always within a defined promoter-region CpG island and easier to be detected. In addition, epigenetic defects are specific for diseases and occur in higher percentages than genetic variations. For example, aberrant methylation is found at early stag-

es of carcinogenesis and distinct types of cancer exhibit specific patterns of methylation changes [4]. Also, cancer deregulated miRNAs are demonstrated to significantly correlate with clinically relevant information, such as disease aggressiveness and metastatic potential, patient response to therapy, time-to-relapse, and overall survival [5]. Meanwhile, diseases other than cancer reveal profound alterations in epigenetic defects, whereas genetic alterations such as mutations are rarely detected, which makes epigenetic analysis applicable to a broader range of diseases [6]. Furthermore, comparing DNA-, RNA-, and proteinbased biomarkers, RNA- and protein-based biomarkers are biologically and biochemically unstable. This feature prevents their use in clinical practice although they have been most intensively evaluated for more than a decade. DNA based biomarkers are stable, so utility of the DNA methylation biomarkers have been seriously considered [7]. Nowadays, genome-wide analysis has revealed that about 70% of the genome is actively transcribed into ncRNAs while less than 2% of the sequences encode proteins [8]. The ncRNAs comprise a very complex and diverse class, which has been arbitrarily subdivided into two major groups: small ncRNAs including microRNAs (miRNAs) and long noncoding RNAs (lncRNAs). One of the most remarkable aspects of miRNAs is their high stability with resistance to RNase digestion and other harsh conditions, such as extreme pH, boiling, extended storage, and multiple freezethaw cycles outside the intracellular environment and this aspect makes them potential biomarkers for clinical use. The stability is due to several factors, which prevent their degradation from both cellular and environmental degradation. First, miRNAs can be packaged in microvesicles, such as exosomes and apoptotic bodies. Second, most miRNAs are bound to RNA binding proteins such as AGO2 or lipoprotein complexes such as high density lipoproteins (HDLs) [9]. On the other hand, lncRNAs expression are more specific to organs, tissues, cell types, developmental stages, and human diseases than protein-coding RNAs [10], making them promising candidates as diagnostic and prognostic biomarkers.

MINIMALLY INVASIVE TISSUES ARE SUITABLE FOR DETECTING EPIGENETIC BIOMARKERS Many tissue types are suitable for the discovery of epigenetic biomarkers including cell-based samples such as tumor tissue biopsy and cell-free samples like plasma. Biomarkers detected in tissue biopsy or postoperation tissues are helpful to identify high-risk patients who need advanced treatment or intensive follow-up. For example, stage I nonsmall cell lung cancer (NSCLC)

V.  Factors Influencing Epigenetic Changes



Minimally Invasive Tissues are Suitable for Detecting Epigenetic Biomarkers

patients with recurrence after surgery showed a higher frequency of DNA methylation in CDKN2A and Cadherin 13 (CDH13) than patients without recurrences [11]. However one of the standards of good disease biomarkers is that collection of the detected tissue should be simple and minimally invasive. For early diagnosis or curative effect monitoring of disease, a less invasive liquid biopsy is considered more effective. These minimally invasive tissues mainly include blood, urine, stools, ductal lavages, bronchial aspirate, sputumis, and pleural effusions. The presence of cell-free DNA or circulatingfree DNA (cfDNA) in human plasma has been known for some time [12]. cfDNA in blood is always released from normal and cancer cells undergoing apoptosis or necrosis (Fig. 18.1). cfDNA possesses tumor-related genetic alterations, such as mutations and loss of heterozygosity. Similarly, epigenetic alterations including DNA methylation are also detected in cfDNA from patients, thus detecting tumor-specific epigenetic biomarkers in the liquid biopsy is a potential tool for cancer diagnosis [13]. Analysis of DNA methylation in tumor-derived DNA from stools is a potent noninvasive approach for the diagnosis of colorectal cancer (CRC). Analysis of DNA methylation in pleural effusions is useful for the differential diagnosis of lung cancers and other thoracic malignancies. Knowledge regarding the change of histone modifications in disease is accumulating. Additionally, histone modification alterations can be detected on nucleosomes

279

released by apoptotic and necrotic cells in the blood of patients having cancer. As nucleosomes are stable structures in the circulation [14], utility of detecting histone alterations, such as methylation and acetylation on circulating nucleosomes shows considerable potential as biomarkers. In addition, apart from their release by apoptotic and necrotic cells, miRNAs can also be actively secreted into the blood circulation by exosomes. Exosomes are endosome-derived membraneous nanosized microvesicles actively secreted by a wide range of normal cells, as well as malignant cells and are also found in most biological fluids including blood, urine, ascites, and cerebrospinal fluid. Exosomes contain cargoes, such as miRNAs, proteins, mRNA, and so on. Exosomal cargoes including miRNAs are thought to play an important role in cellto-cell communication and have potential roles during disease development. Recently, a great interest in identification of exosomal biomarkers from blood samples for both diagnosis and stratification of tumor patients has been observed. Some pioneer study results have been reported, such as exosomal miR-21 as biomarker in hepatocellular carcinoma (HCC) [15] and exosomal miR-373 as biomarker in breast cancer [16]. Furthermore, exosomes are robust and can be stored for extended periods without significantly affecting the integrity of encapsulated miRNAs. This feature of exosomes increases their potential applicability in the clinical interface. So the detection of exosomal miRNAs in body fluids is a

FIGURE 18.1  Tumor cells release cell-free DNA into the blood stream from normal and cancer cells undergoing apoptosis or necrosis. Circulating tumor cells, noncoding RNAs, exosomes, and nucleosomes can also be released. Epigentic alteration can be detected using these noninvasive liquid biopsy. Ac, Acetylation; Me, methylation; P, phosphorylase.

V.  Factors Influencing Epigenetic Changes

280

18.  Epigenetic Biomarkers

FIGURE 18.2  Field cancerization leads to disease recurrence after surgical resection and development of local recurrence or independent secondary tumors.

booming research field and could be a potential gold mine to advance diagnostics and monitor therapeutic efficiency with a higher sensitivity and specificity. Similarly, an advantage of lncRNAs is that they may be used in clinic as minimally invasive biomarkers due to their presence in body fluids such as serum/plasma and urine in a stable form protected from endogenous RNase, although the precise mechanism of lncRNA release into the extracellular environment is not completely understood. Some study results have shown that exosomes also contain lncRNAs. For example, Li et al. [17] analyzed the expression level of LINC00152 in plasma and plasma-derived exosomes from GC patients and found that between the two samples there is no statistically expression difference, suggesting that the majority of plasma LINC00152 is derived from exosomes. So these evidences showing epigenetic alterations in disease could be detected in minimally invasive tissues emphasize the crucial role of these epigenetic alterations for clinical diagnosis, prognosis, and prediction of response to therapies.

FIELD CANCERIZATION AND EPIGENETIC BIOMARKERS Epigenetic alterations are believed to be the early events in disease especially cancer development. It has been generally accepted that the pathologically defined neighboring noncancerous tissues represent the cells residing at the early stages of carcinogenesis. Previous studies have shown that normal adjacent cells to tumors frequently harbor genetic alterations, such as loss of heterozygosity, microsatellite and chromosome instability, and gene mutations [18]. And now epigenetic alterations are also presented in premalignant tissues [19]. This phenomenon was first described in 1953 by Slaughter group using the term “field cancerization” (also known as field effect or field defect), which refers to the presence of

cancer causing molecular abnormalities in histologically normal-appearing tissue surrounding a neoplasm, some of which are the same as those in the tumors [20,21]. Field cancerization reflects the susceptibility of normal tissue to undergo early development of cancer as a result of exposure to carcinogens. Finally, development of second primary tumors and recurrences can be explained by the presence of residual abnormal tissue following surgery, which is pathologically normal (Fig. 18.2). Now epigenetic markers were found to be useful as a quantitative measure for field cancerization. Epigenetic alterations in field cancerization has recently been shown in various types of human cancers, such as gastric, colorectal, esophageal, liver, and renal cancers [22,23]. For example, Issa et al. reported that four specific genes (ESR, MYOD1, CDKN2A, and VCAN) were highly methylated in the normal appearing epithelium from ulcerative colitis patients with high-grade dysplasia or cancer than in nonulcerative colitis individuals, indicating that methylation precedes dysplasia development [24]. Hypomethylation also occur as marker of epigenetic field defects. Milicic et al. observed field cancerization in CRC indicating that demethylation of P-cadherin might be an early event in the colorectal mucosa, prior to any morphological changes [25]. Another example is that Sato et al. examined miRNA expression profiling in paired tumor and nontumor liver tissue samples from 73 HCC patients and found that the expression patterns of tumor-derived miRNAs tended to predict early recurrence better than late recurrence, whereas those of nontumor-derived miRNAs tend to predict late recurrence after hepatic resection for HCC. These researchers suggested that miRNA expression profiling in nontumor liver tissue would reflect the accumulation of genome abnormalities (the field cancerization) in the noncancerous liver cells [26]. Since epigenetic changes always occur before histopathological changes, a better understanding of epigenetic field cancerization may represent a useful translational opportunity for cancer risk assessment.

V.  Factors Influencing Epigenetic Changes



281

Potential Methylation Biomarkers in Cancer and Other Disease

TABLE 18.1 Commercially Available Tests for DNA Methylation Detection in Clinic DNA methylation biomarkers

Cancer type

Applications

Sources

SEPT9

Colorectal

Detection

Blood

NDRG4, BMP3, (KRAS mutation)

Colorectal

Detection

Stool

VIM

Colorectal

Detection

Stool

SHOX2

Lung

Detection

Bronchial fluid

MGMT

Brain

Prediction

Tumor

MGMT, O6-methylguanine DNA methyltransferase; SEPT9, septin 9; SHOX2, short stature homeobox 2.

POTENTIAL METHYLATION BIOMARKERS IN CANCER AND OTHER DISEASE Cancer is a leading cause of death worldwide, and largely due to the lack of early detection methods, many cancers are diagnosed at advanced stages with poor prognosis. Nowadays, most of the existing cancer biomarkers are based on proteins, such as CEA, PSA, TPA, CA125, and CA19-9, which have low specificity. As for the existing screening tests, such as fecal occult-blood testing, colonoscopy, and stool DNA testing have low sensitivity and reproducibility. Therefore, the discovery of new biomarkers in cancer is particularly important for patient survival. Cancer is a heterogeneous disease and it is widely accepted that tumorigenesis is triggered by the accumulation of both genetic and epigenetic alterations. Epigenetics is an integral part of cancer initiation, development, and recurrence. The tumor microenvironment is mainly affected by all of the epigenomic components in the cells. At the global level, DNA is often hypomethylated in cancer. Hypomethylation can cause activation of the normally silent regions of the genome or genes that would normally be silent during development, including protooncogenes. Previous studies strongly support the hypothesis of early global hypomethylation in carcinogenesis. Hypomethylation and activation of long interspersed nucleotide element-1 (LINE-1), which comprises approximately 18% of the human genome, lead to chromosomal instability and is a biomarker unfavorable prognosis in many cancers, such as CRC [27] or lung neoplasia [28]. However, in clinical applications, DNA hypomethylation used as a biomarker is limited since a reduction of DNA methylation is technically more complicated to detect than a gain of a signal. So more important methylation changes in cancer are characterized by local hypermethylation of individual genes, which is always associated with tumor suppressor genes (TSG) downregulation [29]. During the past decade, many hypermethylated genes have been identified

and a number of putative DNA methylation markers have been evaluated in a variety of clinical specimens for cancer diagnosis, prognosis, and predictive information. Currently, DNA methylation appears as one of the most promising epigenetic biomarkers [30,31]. Actually, several commercial DNA methylation tests have been available for clinical applications (Table 18.1). DNA methylation of Septin 9 (SEPT9) is a valuable marker for the diagnosis of CRC and is now used as a test in the clinic (Epi proColon) [32]. In a prospective study of SEPT9, the sensitivity and specificity of this marker to diagnose CRC were 48.2 and 91.5%, respectively [33]. The sensitivity of the SEPT9 test (73.3%) was not inferior to faecal immunochemical test (FIT, 68.0%) for the detection of colorectal, however the specificity of SEPT9 was lower than FIT (81.5% vs. 97.4%) [34]. Other commercially available DNA methylation tests for diagnosing CRC patients are DNA methylation of VIM and multitarget DNA test in the stool. The sensitivity and specificity of VIM methyaltion detection were 46%–81% and 82%–90% [35] while the sensitivity and specificity of the minimally invasive multitarget stool DNA test, which includes KRAS mutation, aberrant DNA methylation of N-Myc, NDRG4, BMP3, and β-actin (as a reference gene for DNA quantity) with a hemoglobin immunoassay, reached 92.3 and 86.6%, respectively [36]. DNA methylation of short stature homeobox 2 (SHOX2) is a commercially available biomarker for the diagnosis of lung cancer. Using bronchial aspirates, SHOX2 methylation was found in 62% of lung cancers, which were negative by cytological analysis [37]. SHOX2 methylation was also detected in plasma of lung cancer patients with a sensitivity of 60% and specificity of 90% [38]. MGMT methylation was one of the first DNA methylation biomarkers to be identified. O6-methylguanine DNA methyltransferase is a DNA repair protein that is encoded by the MGMT gene and is capable of removing alkyl residues directly from the O6-position of guanines. DNA methylation status of MGMT gene might offer a

V.  Factors Influencing Epigenetic Changes

282

18.  Epigenetic Biomarkers

tailored therapy indicates therapeutic choice in glioma and MGMT methylation predicts survival benefit in glioma patients treated with alkylating agents [39]. We can see the use of DNA methylation biomarkers in cancer looks very promising. Further validation studies of the potential cancer methylation biomarkers in prospective cohorts will find the most robust methylation biomarkers in clinical application. Though using methylation biomarkers in cancer is best characterized, these concepts are also applicable to other diseases, such as neurological, metabolic and cardiovascular disorders, autoimmune diseases, and genomic imprinting disorders. For example, assays of H19/IGF2 aberrant DNA methylation as a risk factor for colon cancer are commercially available for the neurodevelopmental disorders of genomic imprinting Prader–Willi syndrome (PWS) and Angelman syndrome (AS) [40]. AD is the most common age-dependent neurodegenerative disorder. In light of recent studies, epigenetic modification has emerged as one of the pathogenic mechanisms of AD. Understanding of the epigenomes and transcriptomes of AD may warrant future identification of novel biological markers. The methylation frequency of HTERT, a mRNA component of telomerase, is higher in AD patients compared to elderly controls [41]. The association of HTERT methylation status with AD indicates that this gene may be involved in higher telomerase activity and immune dysfunctions in AD pathogenesis. In addition, other inflammatory genes, such as iNOS, IL-1, and TNF-α are hypomethylated in the AD cortex [42]. This study suggests that many inflammatory genes are activated by epigenetic alterations and are involved in the pathogenesis of AD. Genomic imprinting disorders include PWS, AS, Beckwith–Wiedemann syndrome (BWS), and Silver– Russell syndrome (SRS). For diagnosis of PWS or AS, methylation status of the small nuclear ribonucleoprotein polypeptide N (SNRPN) gene locus or the methylation status and gene copy number changes at several sites across the region are commonly used [43]. BWS and SRS are imprinting disorders associated with imprinted genes on two differentially methylated regions (DMRs) of chromosome region 11p15.5. DRM1 consists of the imprinted insulin-like growth factor gene 2 (IGF2) and the noncoding RNA H19. DRM2 contains several imprinted genes including cyclin-dependent kinase inhibitor 1C (CDKN1C), potassium voltage-gated channel, KQT like subfamily member 1 (KCNQ1), and KCNQ1 opposite strand/antisense transcript 1 (KCNQ1OT1). Loss of methylation at DMR2 (KCNQ1OT1 hypomethylation) is the most frequent alteration in around 50% of BWS patients [44], whereas loss of methylation at DMR1 (H19 hypomethylation) is found in around 40% of SRS patients [45,46].

POTENTIAL HISTONE MODIFICATION BIOMARKERS IN CANCER AND OTHER DISEASES Histone modifications regulate the structure of chromatin, and play fundamental roles in gene regulation and expression. So it is not surprising that aberrant patterns of histone marks are found in cancer and other diseases. For example in CRC, two different studies reported that global levels of acetylation of H4K12 (H4K12ac) and H3K18 (H3K18ac) increased in adenocarcinomas with respect to normal tissue or adenoma [47,48]. Lower levels of dimethylation of H3K4 (H3K4me2) correlated with a poor survival rate. The multivariate survival analysis showed that H3K4me2 status is an independent prognostic factor for patients with CRC [49]. For clinical application, the detection of histone methylation in blood circulation was carried out [50]. Two histone methylation marks, H3K9me3 and H4K20me3, were investigated in blood circulating nucleosomes. H3K9me3 and H4K20me3 were found to be lower in patients with CRC in comparison with healthy controls [51]. Similar results were obtained in a further study including patients with CRC, breast, lung, and benign gastrointestinal disease [52]. All these data suggest the biomarker potential of H3K9me3 and H4K20me3-related nuclesomes in CRC. The changes in histone modifications are also detected in other diseases, such as neurodegenerative, autoimmune, and cardiovascular disorders. Many studies show that aberrant expression of enzymes performing the covalent histone modifications, including histone deacetylases (HDACs), histone methylases, and demethylases is often checked in these diseases [53]. However, the biomarker potential of histone marks faces more technical challenges compared with the diagnostic potential of DNA methylation, as histone modifications are less-stable modifications and have been shown to be more dynamic [54,55]. The assessment of histone marks is more difficult to apply and to standardize because it relies on using antibodies, which vary in performance.

POTENTIAL NONCODING RNA BIOMARKERS IN DISEASES NcRNAs comprise a very complex and diverse class, which has been arbitrarily subdivided into two major groups, small ncRNAs and lncRNAs group, according to their size based on a cutoff of 200 nucleotides. Small ncRNAs includes well-defined miRNAs, siRNAs, promoter- or antisense-associated short RNAs, and “housekeeping” RNAs. While the lncRNAs are classified into

V.  Factors Influencing Epigenetic Changes



Potential Noncoding RNA Biomarkers in Diseases

five general categories according to their location on the genome: sense or antisense, intronic, bidirectional, and intergenic. Although functions of most ncRNAs are still largely uncharacterized, a number of studies provide evidences that ncRNAs are involved in many diseases and illustrated the potential of these ncRNAs as novel biomarkers for diseases [56,57]. However, nowadays only few ncRNAs have been appropriately validated in independent cohorts or investigated for their function in disease development.

Potential miRNA Biomarkers in Cancer and Other Diseases miRNAs (∼22 nucleotides in length) are short endogenous nonprotein coding single stranded RNA molecules that play roles in posttranscriptional gene regulation through translational repression or inducing degradation of target messenger RNAs [58,59]. miRNAs contribute to various biological processes, such as development, differentiation, and carcinogenesis [60]. Bioinformatic data indicates that each miRNA can control hundreds of gene targets, underscoring the potential influence of miRNAs on almost every genetic pathway. Extensive studies have shown that the circulating miRNAs in a range of body fluids are useful biomarkers for the diagnosis and prognosis of disease [56–63]. In 2008, Chim et al. first identified the expression of miRNAs in the circulation by qRT-PCR [63]. In the same year, serum high miR-21 expression was found to be linked to the relapse-free individual though comparing B cell lymphoma patients with healthy controls [62]. MiR-21 is one of the first miRNAs, which is associated with a variety of cancers, such as prostate cancer, pancreas cancer, liver cancer, GC, CRC, and NSCLC [64]. These studies opened the door to the possibility of circulating miRNAs as potent, minimally invasive analytical, or diagnostics markers for different diseases, especially cancers. These potential biomarker miRNAs are also correlated with disease status, stage, aggressiveness, and response to therapy. Later, serum miR-141 was identified and could distinguish prostate cancer patients from healthy controls, providing evidence that tumor-restricted miRNAs may serve as blood-based biomarkers [65]. Recently, Chen et al. performed miRNA profiling in plasma from patients with prostate cancer or benign prostatic hyperplasia (BPH) and identified that a 5 miRNA-model (let-7c, let-7e, miR-30c, miR-622, miR1285) could differentiate prostate cancer from BPH with AUC of 0.924 and prostate cancer from healthy individuals with AUC of 0.860. These miRNAs were shown to improve the diagnostic performance of the PSA test [66].

283

In pancreatic ductal adenocarcinoma (PDAC), miRNAs were reported as useful circulating biomarkers for the first time in 2009. Wang et al reported a panel of four overexpressed plasma miRNAs (miR-21, -210, -155, and -196a) could clearly distinguish patients from healthy controls with a sensitivity of 64% and a specificity of 89% [67]. Recently, Kojima et al. detected 2555 human miRNAs using 100 serum samples from PDAC patients and 150 from healthy controls and found that a combination of 8 miRNAs (miR-6075, -4294, -6880-5p, -6799-5p, -125a3p, -4530, -6836-3p, and -4476) had very high sensitivity (80.3%) and specificity (97.6%) [68]. Intriguingly, Ng et al. [69] reported the combination of plasma miR-145 and -451 levels provided the best markers for breast cancer prediction with sensitivity of 90% and specificity of 92% in discriminating breast cancer from control subjects including all other types of cancers recruited in their study. It is showed that miRNAs in plasma could also serve as an early detection marker for GC and CRC. miR-175p, -21, -106a, and -106b were upregulated while let-7a was downregulated in plasma of GC patients compared with healthy controls. The ratio of miR-106b/let-7a yielded the highest sensitivity and specificity with an AUC value of 0.879 [70]. More interestingly the diagnostic potential of another panel of five circulating miRNAs (miR-1, -20a, -27a, -34, and -423-5p) for GC was greater than the standard testing with CEA and CA19-9 [71]. For CRC, elevated levels of plasma miR-29a and -92a showed as potential early diagnostic markers even for advanced adenoma patients [72]. Although most miRNAs studies are also focused on cancer, current studies have revealed that a number of miRNAs may act as potential biomarkers for other diseases, such as metabolic disorders, autoimmune, inflammatory diseases. For instance, miR-19b, -106a, and -629 or -191could distinguish Crohn’s disease from ulcerative colitis [73,74]. Moreover, miRNA let-7d and let-7e were identified as possible therapeutic biomarkers in patients with Crohn’s disease [75]. Circulating miRNAs may also represent good markers for heart failure (HF) and for the response to therapeutic treatment. In humans, urinary and plasma miR-21 are associated with severe acute kidney injury and poor outcomes after adult cardiac surgery [76]. Circulating levels of miR-423-5p and miR-18b* were able to distinguish HF patients from controls. While miR-423-5p even could differentiate HF patients from non-HF patients with other causes of dyspnea [77]. In addition, miR-1 is not only a potential indicator of acute myocardial infarction (AMI) but also a determinant of AMI size and recovery [78–80]. For reference, more potential miRNA biomarkers are listed in Table 18.2.

V.  Factors Influencing Epigenetic Changes

284

18.  Epigenetic Biomarkers

TABLE 18.2 Representative Potential miRNA Biomarkers in Cancer and Other Diseases miRNAs

Related diseases

Applications

miR-21

NSCLC, CRC, laryngeal squamous cell carcinoma, hepatocellular cancer, esophageal cancer, squamous cell carcinoma, GC, ovarian cancer, breast cancer, pancreatic denocarcinoma, cervical cancer, and glioblastoma, B cell lymphoma

Diagnostic and prognostic biomarker

miR-9

Osteosarcoma, nasopharyngeal carcinoma, NSCLC, biliary tract cancer

Prognostic biomarker

miR-210

CRC, cell renal cell carcinoma

Diagnostic and prognostic biomarker

miR-200c

GC, CRC

Diagnostic and prognostic biomarker

miR-205

Bladder cancer, prostate cancer

Diagnostic and prognostic biomarker

miR-150

CRC

Diagnostic and prognostic biomarker

miR-1825/miR-484

Prostate cancer

Diagnostic biomarker

miR-103

Breast cancer

Potential diagnostic biomarker

miR-16

Melanoma

Diagnostic and prognostic biomarker

miR-10a

Cervical cancer, breast cancer, bladder cancer, GC

Potential biomarkers of aggressive progression and poor prognosis

miR-18a

Esophageal cancer, pancreatic cancer, hepatocellular cancer, CRC

Potential diagnostic biomarker

miR-155

B-cell malignancies, chronic lymphocytic leukemia, breast cancer, esophageal cancer

Diagnostic and prognostic biomarker

miR-221

Larynx cancer, breast cancer

Diagnostic and prognostic biomarker

miR-20a

Cervical cancer, multiple myeloma, rectal cancer, prostate cancer, GC,

Potential biomarkers of aggressive progression and poor prognosis

miR-152

Bladder cancer, NSCLC

Diagnostic biomarker

let-7c

NSCLC

Diagnostic biomarker

miR-125b

Ewing’s sarcoma, NSCLC, rectal adenocarcinoma

Diagnostic and prognostic biomarker

miR-126

Colorectal, esophageal cancer, hepatocellular carcinoma, acute myeloid leukemia, GC, thyroid cancer, breast cancer, oral cancer, and prostate cancer

Diagnostic and prognostic biomarker

miR-1260b

CRC

Potential prognostic biomarker

miR-22

Osteosarcoma

Diagnostic and prognostic biomarker

miR-744

Pancreatic cancer

Diagnostic and prognostic biomarker

miR-25

Oesophageal squamous cell carcinoma

Diagnostic and prognostic biomarker

miR-196a

Breast cancer

Prognostic biomarker

miR-193b

Ovarian cancer

Diagnostic and prognostic biomarker

miR-34a

Prostate cancer

Diagnostic and prognostic biomarker

miR-221

Larynx cancer, breast cancer

Diagnostic and prognostic biomarker

miR-19b/miR-106a/miR-629

CD

Diagnostic biomarker

miR-191

Ulcerative colitis, CD

Diagnostic biomarker

let-7c

Cervical intraepithelial lesions

Diagnostic biomarker

let-7d

CD

Prognostic biomarker

let-7e

CD

Prognostic biomarker

miR-10a

Coronary artery disease

Potential diagnostic biomarker

V.  Factors Influencing Epigenetic Changes



Potential Noncoding RNA Biomarkers in Diseases

285

TABLE 18.2 Representative Potential miRNA Biomarkers in Cancer and Other Diseases (cont.) miRNAs

Related diseases

Applications

miR-210

Preeclampsia

Diagnostic and prognostic biomarker

miR-155

Ankylosing spondylitis, male fertility

Diagnostic and prognostic biomarker

miR-150

Inflammatory active disease

Early diagnostic biomarker

miR-150-5p

Myasthenia gravis

Diagnostic biomarker

miR-23a

Prediabetes and type 2 diabetes

Diagnostic biomarker

miR-125b

AD

Diagnostic biomarker

AD, Alzheimer’s disease; CD, crohn’s disease; CRC, colorectal cancer; GC, gastric cancer; miRNA, microRNAs; NSCLC, nonsmall cell lung cancer.

Potential Long Noncoding RNA Biomarkers in Cancer and Other Diseases lncRNAs play important regulatory roles in transcription, translation, chromatin modification, and cellular organization. Misregulation of lncRNAs is found associated with various human diseases. Though lncRNAs are only recently discovered, at least 321 experimentally verified lncRNAs associated with 221 various types of diseases, which are most related to cancer [81]. The lncRNAs represent another group of potential biomarkers for cancer diagnosis and prognosis. Prostate cancer antigen 3 (PCA3 lncRNA or DD3) is a well-studied lncRNA and the most specific to prostate cancer (PCa) as it is not expressed in other normal human tissues. PROGENSA PCA3 test is the first urinebased molecular diagnostic test approved by the Food and Drug Administration [82]. The sensitivity and specificity of urine PCA3 expression for PCa diagnosis reach 62 and 75%, respectively, supporting PCA3 as a reasonable marker for prostate cancer diagnosis [83]. Urothelial carcinoma-associated 1 (UCA1) is a lncRNA dysregulated in a broad range of human cancers [84]. UCA1 was shown to be a promising biomarker for bladder cancer diagnosis and therapeutic monitoring. UCA1 level in blood or urine of bladder cancer patients could distinguish bladder cancer from other urinary tract diseases, such as neurogenic bladder, renal cell carcinoma, and upper urinary tract restriction or reflux, with an overall AUC of 0.882 and an overall specificity of 91.8% [85,86]. UCA1 has also been detected to be overexpressed in serum of HCC patients than in that from patients with chronic hepatitis C virus infection or healthy volunteers [87] and in gastric juice from GC patients than in gastric juice from normal subjects [88]. Recently, a metaanalysis was conducted to assess the association between UCA1 and cancer prognosis. UCA1 levels were significantly and negatively associated with overall survival times in CRC, NSCLC, ovarian cancer, and GC [89].

HOX transcript antisense RNA (HOTAIR) is another well-studied lncRNA. In cervical cancers, high serum levels of HOTAIR were significantly correlated with tumor recurrence and shorter overall survival [90]. Studies have shown that the combination of several lncRNAs could considerably increase the biomarker efficiency. For example, Tang et al. showed that plasma RP11-160H22.5, XLOC_014172, and LOC149086 transcripts also overexpressed in HCC. The sensitivity and specificity for HCC diagnosis reached 82 and 73% with a merged AUC of 0.896. More interestingly, lncRNAs XLOC_014172 and LOC149086 also have prognostic value for metastasis prediction, showing a sensitivity and specificity of 91 and 90% [91]. Dong et al. found that the combination of CUDR, LSINCT-5, and PTENP1 provided the best diagnostic value in GC with an AUC of 0.92, a sensitivity of 74.1%, and a specificity of 100%. They were also sufficiently sensitive and specific for early GC detection and distinguishing benign peptic ulcers from GC [92]. In addition, Shi et al. found a panel of three lncRNAs (XLOC_006844, LOC152578, and XLOC_000303) significantly upregulated in CRC plasma samples compared to those from healthy controls [93]. They have a strong diagnostic value with an AUC of 0.919 in the training set and an AUC of 0.975 in the validation set. Dysregulation of lncRNAs was involved not only in cancer processes but also in a range of other disease types, in which two top categories are cardiovascular diseases (including HF, atherosclerosis, and coronary artery disease) and neurodegenerative (including schizophrenia, AD, and Huntington’s disease). It is reported that serum lncRNA LIPCAR is considered a novel biomarker of cardiac remodeling and is predictive of mortality in HF patients [94]. In patients affected by ischemic HF, three lncRNAs (CDKN2B-AS1/ ANRIL, HOTAIR, and LOC285194/TUSC7) showed similar modulation in peripheral blood mononuclear cells and heart tissue, suggesting a potential role as disease biomarkers [95]. As well, the circulating UCA1

V.  Factors Influencing Epigenetic Changes

286

18.  Epigenetic Biomarkers

could be used as a promising novel biomarker for the diagnosis and/or prognosis of AMI [96]. In neurodegenerative diseases, LncRNAs FMR1-AS1 and FMR6 expressed in peripheral blood leukocytes and may represent early diagnostic biomarkers for fragile X syndrome (FXS) and fragile X tremor/ataxia syndrome (FXTAS). Kraus et al. reported that five lncRNAs are significantly differentially expressed in Parkinson’s disease (PD), of which, H19 upstream conserved 1 and 2 is significantly downregulated while lincRNA-p21, MALAT1, SNHG1, and TncRNA are significantly upregulated. They may serve as potential new biomarkers even in early PD [97]. As for other diseases, the expression profile of lncRNAs in labial salivary glands of primary Sjögren’s syndrome (pSS) patients revealed that many novel lncRNA transcripts that play important roles in the pathogenesis of pSS were dysregulated in pSS will aid in the

development of new diagnostic biomarkers [98]. Xu et al. [99] found that the expression of two lncRNAs (Vax2os1 and Vax2os2) were significantly upregulated in the aqueous humor of choroidal neovascularization patients, making them predictive biomarkers for the diagnosis of ocular neovascular diseases [100]. In addition, MALAT1 is upregulated in peripheral blood samples from proliferative vitreo-retinopathy (PVR) patients, implying that it may represent an easily detectable biomarker for minimally invasive diagnosis to identify high-risk PVR patients [101]. Moreover, recent a genome-wide long noncoding RNA analysis identified that a panel of five circulating lncRNAs (NR_038395, NR_038452, ENST00000482343, ENST00000544649, and ENST00000393610) was a potential minimally invasive biomarker for endometriosis [102]. For reference, more potential lncRNA biomarkers are listed in Table 18.3.

TABLE 18.3 Representative Potential lncRNA Biomarkers in Cancer and Other Diseases LncRNAs

Related diseases

Applications

HOTAIR

Lung cancer, breast cancer, colorectal cancer, esophageal cancer, laryngeal cancer, nasopharyngeal cancer, hepatocellular cancer, GC, pancreatic carcinoma, NSCLC, mesenchymalglioma

Prognostic biomarker of metastasis

SRA

Breast cancer, uterus cancer, ovary carcinomas

Coactivator of estrogen receptor alpha

PANDAR

Bladder cancer, GC, HCC, NSCLC

Diagnostic and prognostic biomarker

MALAT1

Lung cancer, breast cancer, CRC, bladder cancer, kidney cancer, endometrium cancer, pancreatic carcinoma

Predict poor overall survival

LincRNA-p21

Lung cancer, CRC, leukemia, head and neck cancer

Predict the treatment response

H19

Prostate cancer, glioma cancer, GC, breast cancer, bladder cancer, kidney cancer

Putative susceptibility and diagnostic biomarker

GAS5

Prostate cancer, breast cancer, kidney cancer

Predict metastasis and poor prognosis

ANRIL

Breast cancer, melanoma, leukemia

Prognostic biomarker

MEG3

Breast cancer, lung cancer, bladder cancer, kidney cancer, leukemia, osteosarcoma

Predictor biomarker in progression and poor prognosis

XIST

Prostate cancer, breast cancer

Diagnostic and prognostic biomarker

GCAT1

GC

Potential diagnostic and prognostic biomarker

SUMO1P3

GC, bladder cancer

Potential diagnostic and prognostic biomarker

CCAT1

GC, CRC, breast cancer, HCC, gallbladder cancer

Potential diagnostic and prognostic biomarker

ENST00000435885.1/XLOC 013014/ ENST00000547963.1

Esophagus cancer

Prognostic biomarker

uc.73

CRC

Prognostic biomarker

HULC

Liver cancer, GC, osteosarcoma

Diagnostic and prognostic biomarker

MALAT-1

Liver cancer, lung adenocarcinomas, bladder cancer

Predict poor prognosis

V.  Factors Influencing Epigenetic Changes



Epigenetic Biomarker Detection Methods in Clinical Application

287

TABLE 18.3 Representative Potential lncRNA Biomarkers in Cancer and Other Diseases (cont.) LncRNAs

Related diseases

Applications

HOTTIP

Liver cancer, GC, prostate cancer, CRC, osteosarcoma, HCC

Predicts poor prognosis

HOXA13

Liver cancer, NSCLC, GC, glioma

Diagnostic biomarker and promotes invasion

TUG1

NSCLC, CRC, GC

Poor prognosis biomarker

BANCR

NSCLC, melanoma,

Associated with clinical progression and poor prognosis

LincRNA-RoR

Breast cancer, CRC, glioma, pancreatic cancer

Promotes invasion, metastasis and tumor growth

UCA1

Osteosarcoma, bladder cancer, renal cell carcinoma, epithelial ovarian cancer, GC

Poor prognostic biomarker

linc-UBC1

Bladder cancer, GC

Negative prognostic biomarker

PCA3/TMPRSS2-ERG

Prostate cancer

Diagnostic biomarkers

HOTAIRM1

Leukemia, CRC

Potential diagnostic biomarker

UCA1

Acute myocardial infarction

Diagnostic biomarker

LIPCAR

Heart failure patients

Diagnostic and prognosticbiomarker

ANRIL/HOTAIR/TUSC7

Ischemic heart failure

Potential role as disease biomarkers

H19 upstream conserved 1 and 2/lincRNA-p21/Malat1/ SNHG1/TncRNA

PD

Potential diagnostic biomarkers

MALAT1

Proliferative vitreoretinopathy

Diagnostic biomarker

Vax2os1/Vax2os2

Ocular neovascular diseases

Diagnostic biomarker

NR_038395/NR_038452/ ENST00000482343/ ENST00000544649/ ENST00000393610

Endometriosis

Diagnostic biomarker

HOTAIR, HOX transcript antisense RNA; lncRNA, long noncoding RNAs; PCA3, prostate cancer antigen 3; PD, Parkinson’s disease; UCA1, urothelial carcinomaassociated 1.

The future development of lncRNAs applications relies on technologies to identify and validate their functions, structures, and mechanisms. Comprehensive understanding of genome-wide interaction networks of lncRNAs with proteins, chromatins, and other RNAs in regulating cellular processes will allow personalized medicine to use lncRNAs as highly specific biomarkers in diagnosis, prognosis, and therapeutic targets.

EPIGENETIC BIOMARKER DETECTION METHODS IN CLINICAL APPLICATION Methods in epigenetics have been discussed systematically in section ”Methods in Epigenetics” so in this chapter we just focused on epigenetic biomarker detection methods and the challenges faced by the researchers in clinical application.

In general, these techniques can be divided into various types of pretreatment (enzyme digestion, affinity enrichment, sodium bisulfite) followed by different analytical steps (locus-specific, gel-based, array-based, and next-generation sequencing-based analysis) [103]. Bisulfite treatment of DNA is widely used for analysis of DNA methylation status. During bisulfite treatment, unmethylated cytosines are converted into uracils, whereas methylated cytosines remain intact. However, after bisulfite conversion, 5-hydroxymethylcytosines also remain intact, indicating overestimation of the DNA methylation level in some cases. In addition, Bisulfite treatment always leads to significant DNA degradation. So methylation-sensitive restriction enzyme–mediated methods are considered to be more sensitive than bisulfite treatment methods for detecting low-abundance methylated TSGs in the liquid biopsy cfDNA. Conventional methylation-specfic PCR (MSP) is frequently used for detection of DNA methylation. But it is

V.  Factors Influencing Epigenetic Changes

288

18.  Epigenetic Biomarkers

qualitative analysis, not suitable for clinical setting. By contrast, quantitative MSP analysis, such as Methylight, pyrosequencing or SMART-MSP might be more suitable for clinical application. However, the limit of this detection is as low as 5% for the each CpG, it might not have enough power to assess the low amount of methylated DNA fragment in cfDNA. Now some new quantitative DNA methylation assays are showing, such as MethylBEAMing based on emulsion PCR assay [104] and MethyLight coupled with digital PCR [105]. They are all reported as more sensitive digital approaches. These methods might have less bias during sample preparation and give more accurate data, making a next revolution in DNA methylation analysis. Anyway, there is another important issue that deserves attention when using PCR amplification analysis. Multiple splice variants of the detected genes exist and specific primers should be designed accordingly. In addition, heterogeneity at the subject and tissue levels can also affect reproducibility in target-gene focused investigations. Since research of DNA methylation has evolved from locus specific approach to genome-wide determination of methylome with base pair resolution. Array-based technology displaying a high reproducibility and accuracy is widely used for analysis of clinical samples [106]. The more comprehensive DNA methylation analysis is whole genome shotgun bisulfite sequence (WGBS-seq). However, it remains still time consuming and too expensive for analyzing multiple clinical samples [107]. To reduce the amount of sequencing, targeted bisulfite sequencing methods have been developed using array capture or affinity enrichment of methylated DNA using antibody for methylated cytosine or methyl-binding proteins before sequencing [103]. Reduced representation bisulfite sequencing (RRBS) is a semitargeted bisulfite sequencing method, which analyze DNA fragments after methylation-insensitive restriction enzyme (MspI) digestion [108]. Nowadays, next generation sequencing (NGS) has greatly enhanced precision and resolution in preclinical and clinical epigenetic studies. It is combined with other techniques to identify epigenetic marks on a genomic scale. Targeted DNA methylation sequencing by NGS is reported as a novel DNA methylation analysis for cfDNA [109]. However, further validation studies may be still required when applying this method for clinical application. In terms of histone modification biomarkers, the antibodies to acetylated histone H3 and H4 have been used for ChIP to determine histone acetylation in specific regions of gene promoter and other regulatory regions. Genome-wide analysis was initially carried out using a technique known as ChIP-chip. ChIP-chip is now being superseded by ChIP-seq, which combines ChIP with NGS technologies involving high-throughput sequencing of all enriched DNA sequences. It provides the

opportunity for increased sensitivity, higher resolution, and more comprehensive screening of genomic profiles while requiring less input material than ChIP-chip. However, since NGS generally produces a notoriously large amount of data, more powerful bioinformatics support is essential for data processing and analysis [110]. miRNA or lncRNA biomarkers can be easily measured by qRT-PCR using the TaqMan PCR method with stem-looped RT primers or oligo-dT primers. Several novel detection methods have been developed to allow for rapid quantification of RNA, improve the sensitivity and specificity including the use of simple molecular beacons, enzymatic luminescence, nanoparticle-based probes, and different forms of electrophoresis [111]. High throughput miRNA and lncRNA quantification technologies including comprehensive RNA sequence analysis by next generation sequencing have developed to be a convenient clinical measuring tool. But as for exosomal miRNA detection in clinical setting, there remains a need for a rapid and inexpensive standard procedure that generates an effective, pure isolation of exosomes, and accurate genome-wide quantification [112]. So currently the technical analyses of epigenetic alterations are still under development and researchers face a number of challenges when studying epigenetic changes. Technique selection depends on the quality and quantity of input DNA needed, purity and type of tissue or fluid DNA that is being extracted, extent of genome coverage, and overall assay reproducibility, sensitivity, specificity, accuracy, and quantification. Improvements in extraction and amplification procedures and growing availability of next-generation sequencing technologies may allow for the discovery of more specific and sensitive RNA markers in the near future.

CHALLENGES AND FUTURE PERSPECTIVES Though studies have revealed a number of epigenetic changes could be used as disease biomarkers providing some clinical value in diagnosis and prognosis. Only a few results have been appropriately validated in independent cohorts or investigated in large, perspective clinical trials. Many results are only from primary tissue biopsy and may totally different from liquid biopsy of the same patient. Moreover, for those minimally invasive epigenetic biomarker detection, insufficient material, and technical shortcomings are the main challenges faced by researchers. For example, now most studies still use the bisulfite-treated DNA methylation detection. Then a considerable loss of initial input DNA can occur and especially in those samples where only a limited amount of genomic DNA is available, such as those from very small biopsies. In addition, nowadays an agreed

V.  Factors Influencing Epigenetic Changes



REFERENCES

methodology as the gold standard for epigenetic alterations is lacking. So it is important to notice that there are a number of limitations to overcome before these potential biomarkers reach the clinic. So in future, advancements in study and clinical trial design will certainly help to speed up validation and clinical implementation of potential epigentic biomarkers. Standardization of specimen type and of sample preparation is also necessary if results from different groups or different samples are to be meaningfully obtained and compared. Ideally, epigenetic biomarker detection methods for clinical settings should be low cost, easy to use, automatable, and capable of processing many samples in parallel to minimize costs. There is an urgent need to develop cost-effective genome-wide technologies with high sensitivity and specificity and used with minimally invasive samples since a panel of multiple disease-specific biomarkers will diagnose and predict disease prognosis more accurately than a single marker. Moreover, it is known that genetics and epigenetics both act together and take advantage of each other to play an important role in disease development, the combination of different epigenetic, genetic, or protein biomarkers will significantly increase sensitivity and specificity. In addition, powerful bioinformatic platforms will play a crucial role in this type of research since now bioinformatic tools have made it possible to address all the aspects of miRNA research. It is fully anticipated that there will be a rapid improvement in technological capability and in the gaining of new insights into the role and mechanisms of epigenetic phenomena in both development and disease. More and more epigenetic biomarkers will be successfully translated into clinical practice in the near future.

Abbreviations Ac Acetylation AD Alzheimer’s disease AMI Acute myocardial infarction AS Angelman syndrome BPH Benign prostatic hyperplasia BWS Beckwith–Wiedemann syndrome CDKN1C Cyclin-dependent kinase inhibitor 1C cfDNA Cell-free DNA or circulating free DNA CRC Colorectal cancer DMRs Differentially methylated regions FXS Fragile X syndrome FXTAS Fragile X tremor/ataxia syndrome HCC Hepatocellular carcinoma HDACs Histone deacetylases HF Heart failure IGF2 Insulin-like growth factor gene 2 KCNQ1 KQTlike subfamily, member 1 KCNQ1OT1 KCNQ1 opposite strand/antisense transcript 1 lncRNAs Long noncoding RNAs Me Methylation MGMT O6-methylguanine DNA methyltransferase

289

miRNAs microRNAs MSP Methylation-specfic PCR NGS Next generation sequencing NSCLC Nonsmall cell lung cancer P Phosphorylase PCa Prostate cancer PCA3 Prostate cancer antigen 3 PD Parkinson’s disease PDAC Pancreatic ductal adenocarcinoma pSS Sjögren’s syndrome PVR Proliferative vitreo-retinopathy PWS Prader–Willi syndrome RRBS Reduced representation bisulfite sequencing SHOX2 Short stature homeobox 2 SNRPN Small nuclear ribonucleoprotein polypeptide N SRS Silver–Russell syndrome UCA1 Urothelial carcinoma-associated 1 WGBS-seq Whole genome shotgun bisulfite sequence

References [1] Ushijima T, Nakajima T, Maekita T. DNA methylation as a marker for the past and future. J Gastroenterol 2006;41(5):401–7. [2] Mastroeni D, McKee A, Rogers J, Coleman PD. Epigenetic differences in cortical neurons from a pair of monozygotic twins discordant for Alzheimer’s disease. PLoS One 2009;4(8):e6617. [3] Kolble K, Ullrich OM, Pidde H, Barthel B, Diermann J, Rudolph B, et al. Microsatellite alterations in serum DNA of patients with colorectal cancer. Lab Invest 1999;79(9):1145–50. [4] Paluszczak J, Baer-Dubowska W. Epigenetic diagnostics of cancer–the application of DNA methylation markers. J Appl Genet 2006;47(4):365–75. [5] Ferracin M, Negrini M. Micromarker 2.0: an update on the role of microRNAs in cancer diagnosis and prognosis. Expert Rev Mol Diagn 2015;15(10):1369–81. [6] Volkmar M, Dedeurwaerder S, Cunha DA, Ndlovu MN, Defrance M, Deplus R, et al. DNA methylation profiling identifies epigenetic dysregulation in pancreatic islets from type 2 diabetic patients. EMBO J 2012;31(6):1405–26. [7] Jubb AM, Quirke P, Oates AJ. DNA methylation, a biomarker for colorectal cancer: implications for screening and pathological utility. Ann NY Acad Sci 2003;983(1):251–67. [8] Sanchez Y, Huarte M. Long non-coding RNAs: challenges for diagnosis and therapies. Nucleic Acid Ther 2013;23(1):15–20. [9] Khoury S, Tran N. Circulating microRNAs: potential biomarkers for common malignancies. Biomark Med 2015;9(2):131–51. [10] Djebali S, Davis CA, Merkel A, Dobin A, Lassmann T, Mortazavi A, et al. Landscape of transcription in human cells. Nature 2012;489(7414):101–8. [11] Brock MV, Hooker CM, Ota-Machida E, Han Y, Guo M, Ames S, et al. DNA methylation markers and early recurrence in stage I lung cancer. N Engl J Med 2008;358(11):1118–28. [12] Schwarzenbach H, Hoon DS, Pantel K. Cell-free nucleic acids as biomarkers in cancer patients. Nature Rev Cancer 2011;11(6): 426–37. [13] Laird PW. The power and the promise of DNA methylation markers. Nature Rev Cancer 2003;3(4):253–66. [14] Holdenrieder S, Von Pawel J, Nagel D, Stieber P. Long-term stability of circulating nucleosomes in serum. Anticancer Res 2010;30(5):1613–5. [15] Wang H, Hou L, Li A, Duan Y, Gao H, Song X. Expression of serum exosomal microRNA-21 in human hepatocellular carcinoma. BioMed Res Int 2014;2014:864894. [16] Eichelser C, Stückrath I, Müller V, Milde-Langosch K, Wikman H, Pantel K, et al. Increased serum levels of circulating exosomal

V.  Factors Influencing Epigenetic Changes

290

[17]

[18]

[19] [20]

[21]

[22] [23] [24]

[25]

[26]

[27]

[28]

[29]

[30] [31] [32]

[33]

[34]

[35]

[36]

18.  Epigenetic Biomarkers

microRNA-373 in receptor-negative breast cancer patients. Oncotarget 2014;5(20):9650–63. Li Q, Shao Y, Zhang X, Zheng T, Miao M, Qin L, et al. Plasma long noncoding RNA protected by exosomes as a potential stable biomarker for gastric cancer. Tumor Biol 2015;36(3): 2007–12. Braakhuis BJ, Tabor MP, Kummer JA, Leemans CR, Brakenhoff RH. A genetic explanation of Slaughter’s concept of field cancerization: evidence and clinical implications. Cancer Res 2003;63(8):1727–30. Ushijima T. Epigenetic field for cancerization. J Biochem Mol Biol 2007;40(2):142–50. Dakubo GD, Jakupciak JP, Birch-Machin MA, Parr RL. Clinical implications and utility of field cancerization. Cancer Cell Int 2007;7:2. Slaughter DP, Southwick HW, Smejkal W. Field cancerization in oral stratified squamous epithelium; clinical implications of multicentric origin. Cancer 1953;6(5):963–8. Ramachandran K, Singal R. DNA methylation and field cancerization. Epigenomics 2012;4(3):243–5. Ushijima T. Epigenetic field for cancerization: its cause and clinical implications. BMC Proc 2013;7(Suppl. 2):K22. Issa JPJ, Ahuja N, Toyota M, Bronner MP, Brentnall TA. Accelerated age-related CpG island methylation in ulcerative colitis. Cancer Res 2001;61(9):3573–7. Milicic A, Harrison LA, Goodlad RA, Hardy RG, Nicholson AM, Presz M, et al. Ectopic expression of P-cadherin correlates with promoter hypomethylation early in colorectal carcinogenesis and enhanced intestinal crypt fission in vivo. Cancer Res 2008;68(19):7760–8. Sato F, Hatano E, Kitamura K, Myomoto A, Fujiwara T, Takizawa S, et al. MicroRNA profile predicts recurrence after resection in patients with hepatocellular carcinoma within the Milan Criteria. PloS one 2011;6(1):e16435. Hur K, Cejas P, Feliu J, Moreno-Rubio J, Burgos E, Boland CR, et al. Hypomethylation of long interspersed nuclear element-1 (LINE-1) leads to activation of proto-oncogenes in human colorectal cancer metastasis. Gut 2014;63(4):635–46. Daskalos A, Nikolaidis G, Xinarianos G, Savvari P, Cassidy A, Zakopoulou R, et al. Hypomethylation of retrotransposable elements correlates with genomic instability in non-small cell lung cancer. Int J Cancer 2009;124(1):81–7. Park SY, Kim BH, Kim JH, Cho NY, Choi M, Yu EJ, et al. Methylation profiles of CpG islands loci in major types of human cancer. J Korean Med Sci 2007;22(2):311–7. Heyn H, Esteller M. DNA methylation profiling in the clinic: applications and challenges. Nat Rev Genet 2012;13(10):679–92. Paska AV, Hudler P. Aberrant methylation patterns in cancer: a clinical view. Biochem Med (Zagreb) 2015;25(2):161–76. Tetzner R, Model F, Weiss G, Schuster M, Distler J, Steiger KV, et al. Circulating methylated SEPT9 DNA in plasma is a biomarker for colorectal cancer. Clin Chem 2009;55(7):1337–46. Church TR, Wandell M, Lofton-Day C, Mongin SJ, Burger M, Payne SR, et al. Prospective evaluation of methylated SEPT9 in plasma for detection of asymptomatic colorectal cancer. Gut 2014;63(2):317–25. Johnson DA, Barclay RL, Mergener K, Weiss G, König T, Beck J, et al. Plasma Septin9 versus fecal immunochemical testing for colorectal cancer screening: a prospective multicenter study. PloS one 2014;9(6):e98238. Ned RM, Melillo S, Marrone M. Fecal DNA testing for Colorectal Cancer Screening: the ColoSureTM test. PLOS Curr 2011;3: RRN1220. Imperiale TF, Ransohoff DF, Itzkowitz SH, Levin TR, Lavin P, Lidgard GP, et al. Multitarget stool DNA testing for colorectal-cancer screening. N Engl J Med 2014;2014(370):1287–97.

[37] Schmidt B, Liebenberg V, Dietrich D, Schlegel T, Kneip C, Seegebarth A, et al. SHOX2 DNA methylation is a biomarker for the diagnosis of lung cancer based on bronchial aspirates. BMC Cancer 2010;10:600. [38] Kneip C, Schmidt B, Seegebarth A, Weickmann S, Fleischhacker M, Liebenberg V, et al. SHOX2 DNA methylation is a biomarker for the diagnosis of lung cancer in plasma. J Thorac Oncol 2011;6(10):1632–8. [39] Hegi ME, Diserens AC, Gorlia T, Hamou MF, de Tribolet N, Weller M, et al. MGMT gene silencing and benefit from temozolomide in glioblastoma. N Engl J Med 2005;352(10):997–1003. [40] Wojdacz TK, Dobrovic A, Algar EM. Rapid detection of methylation change at H19 in human imprinting disorders using methylation-sensitive high-resolution melting. Hum Mutat 2008;29(10):1255–60. [41] Silva PNO, Gigek CO, Leal MF, Bertolucci PHF, de Labio RW, Payao SLM, et al. Promoter methylation analysis of SIRT3, SMARCA5, HTERT and CDH1 genes in aging and Alzheimer’s disease. J Alzheimer’s Dis 2008;13(2):173–6. [42] Akiyama H, Barger S, Barnum S, Bradt B, Bauer J, Cole GM, et al. Inflammation and Alzheimer’s disease. Neurobiol Aging 2000;21(3):383–421. [43] Ramsden SC, Clayton-Smith J, Birch R, Buiting K. Practice guidelines for the molecular analysis of Prader-Willi and Angelman syndromes. BMC Med Genet 2010;11:70. [44] Smilinich NJ, Day CD, Fitzpatrick GV, Caldwell GM, Lossie AC, Cooper PR, et al. A maternally methylated CpG island in KvLQT1 is associated with an antisense paternal transcript and loss of imprinting in Beckwith-Wiedemann syndrome. Proc Natl Acad Sci USA 1999;96(14):8064–9. [45] Eggermann T, Begemann M, Binder G, Spengler S. Silver-Russell syndrome: Genetic basis and molecular genetic testing. Orphanet J Rare Dis 2010;5:19. [46] Horike S, Ferreira JC, Meguro-Horike M, Choufani S, Smith AC, Shuman C, et al. Screening of DNA methylation at the H19 promoter or the distal region of its ICR1 ensures efficient detection of chromosome 11p15 epimutations in Russell-Silver syndrome. Am J Med Genet A 2009;149A(11):2415–23. [47] Ashktorab H, Belgrave K, Hosseinkhah F, Brim H, Nouraie M, Takkikto M, et al. Global histone H4 acetylation and HDAC2 expression in colon adenoma and carcinoma. Dig Dis Sci 2009;54(10):2109–17. [48] Nakazawa T, Kondo T, Ma D, Niu D, Mochizuki K, Kawasaki T, et al. Global histone modification of histone H3 in colorectal cancer and its precursor lesion. Hum Pathol 2012;43(6):834–42. [49] Tamagawa H, Oshima T, Shiozawa M, Morinaga S, Nakamura Y, Yoshihara M, et al. The global histone modification pattern correlates with overall survival in metachronous liver metastasis of colorectal cancer. Oncol Rep 2012;27(3):637–42. [50] Deligezer U, Akisik EE, Erten N, Dalay N. Sequence specific histone methylation is detectable on circulating nucleosomes in plasma. Clin Chem 2008;54(7):1125–31. [51] Deligezer U, Akisik EZ, Akisik EE, Kovancilar M, Bugra D, Erten N, et al. H3K9me3/H4K20me3 ratio in circulating nucleosomes as potential biomarker for colorectal cancer. In: Gahan PB, editor. Circulating nucleic acids in plasma and serum. Dordrecht: Springer; 2011. p. 97–103. [52] Leszinski G, Gezer U, Siegele B, Stoetzer OJ, Holdenrieder S. Histone modifications H3K9me3 and H4K20me3 on circulating nucleosomes in cancer disease. Anticancer Res 2012;32(5):2199–206. [53] Halusková J. Epigenetic studies in human diseases. Folia Biol (Praha) 2010;56(3):83–96. [54] Feinberg AP. Phenotypic plasticity and the epigenetics of human disease. Nature 2007;447(7143):433–40. [55] Berdasco M, Esteller M. Aberrant epigenetic landscape in cancer: how cellular identity goes awry. Dev Cell 2010;19(5):698–711.

V.  Factors Influencing Epigenetic Changes



REFERENCES

[56] Chen X, Ba Y, Ma L, Cai X, Yin Y, Wang K, et al. Characterization of microRNAs in serum: a novel class of biomarkers for diagnosis of cancer and other diseases. Cell Res 2008;18(10): 997–1006. [57] Gilad S, Meiri E, Yogev Y, Benjamin S, Lebanony D, Yerushalmi N, et al. Serum microRNAs are promising novel biomarkers. PloS one 2008;3(9):e3148. [58] Esquela-Kerscher A, Slack FJ. Oncomirs-microRNAs with a role in cancer. Nat Rev Cancer 2006;6(4):259–69. [59] Chen K, Rajewsky N. The evolution of gene regulation by transcription factors and microRNAs. Nat Rev Genet 2007;8(2): 93–103. [60] Yan W, Qian L, Chen J, Chen W, Shen B. Comparison of prognostic microRNA biomarkers in blood and tissues for gastric cancer. J Cancer 2016;7(1):95–106. [61] Gilad S, Meiri E, Yogev Y, Benjamin S, Lebanony D, Yerushalmi N, et al. Serum microRNAs are promising novel biomarkers. PLoS One 2008;3(9):e3148. [62] Lawrie CH, Gal S, Dunlop HM, Pushkaran B, Liggins AP, Pulford K, et al. Detection of elevated levels of tumour-associated microRNAs in serum of patients with diffuse large B-cell lymphoma. Br J Haematol 2008;141(5):672–5. [63] Chim SS, Shing TK, Hung EC, Leung TY, Lau TK, Chiu RW, et al. Detection and characterization of placental microRNAs in maternal plasma. Clin Chem 2008;54(3):482–90. [64] Wang W, Li J, Zhu W, Gao C, Jiang R, Li W, et al. MicroRNA-21 and the clinical outcomes of various carcinomas: a systematic review and meta-analysis. BMC Cancer 2014;14:819. [65] Mitchell PS, Parkin RK, Kroh EM, Fritz BR, Wyman SK, Pogosova-Agadjanyan EL, et al. Circulating microRNAs as stable bloodbased markers for cancer detection. Proc Natl Acad Sci USA 2008;105(30):10513–8. [66] Chen ZH, Zhang GL, Li HR, Luo JD, Li ZX, Chen GM, et al. A panel of five circulating microRNAs as potential biomarkers for prostate cancer. Prostate 2012;72(13):1443–52. [67] Wang J, Chen J, Chang P, LeBlanc A, Li D, Abbruzzesse JL, et al. MicroRNAs in plasma of pancreatic ductal adenocarcinoma patients as novel blood-based biomarkers of disease. Cancer prev res 2009;2(9):807–13. [68] Kojima M, Sudo H, Kawauchi J, Takizawa S, Kondou S, Nobumasa H, et al. MicroRNA markers for the diagnosis of pancreatic and biliary-tract cancers. PLoS One 2015;10(2): e0118220. [69] Ng EK, Li R, Shin VY, Jin HC, Leung CP, Ma ES, et al. Circulating microRNAs as specific biomarkers for breast cancer detection. PloS one 2013;8(1):e53141. [70] Tsujiura M, Ichikawa D, Komatsu S, Shiozaki A, Takeshita H, Kosuga T, et al. Circulating microRNAs in plasma of patients with gastric cancers. Br J Cancer 2010;102(7):1174–9. [71] Liu R, Zhang C, Hu Z, Li G, Wang C, Yang C, et al. A fivemicroRNA signature identified from genomewide serum microRNA expression profiling serves as a fingerprint for gastric cancer diagnosis. Eur J Cancer 2011;47(5):784–91. [72] Huang Z, Huang D, Ni S, Peng Z, Sheng W, Du X. Plasma microRNAs are promising novel biomarkers for early detection of colorectal cancer. Int J Cancer 2010;127(1):118–26. [73] Chen WX, Ren LH, Shi RH. Implication of miRNAs for inflammatory bowel disease treatment: systematic review. World J Gastrointest Pathophysiol 2014;5(2):63–70. [74] Glinsky GV. Disease phenocode analysis identifies SNP-guided microRNA maps (MirMaps) associated with human “master” disease genes. Cell Cycle 2008;7(23):3680–94. [75] Fujioka S, Nakamichi I, Esaki M, Asano K, Matsumoto T, Kitazono T. Serum microRNA levels in patients with Crohn’s disease during induction therapy by infliximab. J Gastroenterol Hepatol 2014;29(6):1207–14.

291

[76] Du J, Cao X, Zou L, Chen Y, Guo J, Chen Z, et al. MicroRNA-21 and risk of severe acute kidney injury and poor outcomes after adult cardiac surgery. PloS one 2013;8(5):e63390. [77] Tijsen AJ, Creemers EE, Moerland PD, de Windt LJ, van der Wal AC, Kok WE, et al. MiR423-5p as a circulating biomarker for heart failure. Circ Res 2010;106(6):1035–9. [78] Wang GK, Zhu JQ, Zhang JT, Li Q, Li Y, He J, et al. Circulating microRNA: a novel potential biomarker for early diagnosis of acute myocardial infarction in humans. Eur Heart J 2010;31(6):659–66. [79] Cheng Y, Tan N, Yang J, Liu X, Cao X, He P, et al. A translational study of circulating cell-free microRNA-1 in acute myocardial infarction. Clin Sci (Lond) 2010;119(2):87–95. [80] Ai J, Zhang R, Li Y, Pu J, Lu Y, Jiao J, et al. Circulating microRNA-1 as a potential novel biomarker for acute myocardial infarction. Bioch Biophys Res Commun 2010;391(1):73–7. [81] Chen G, Wang Z, Wang D, Qiu C, Liu M, Chen X, et al. LncRNADisease: a database for long-non-coding RNA-associated diseases. Nucleic Acids Res 2013;41(Database issue):D983–6. [82] Lee GL, Dobi A, Srivastava S. Prostate cancer: diagnostic performance of the PCA3 urine test. Nat Rev Urol 2011;8(3):123–4. [83] Xue WJ, Ying XL, Jiang JH, Xu YH. Prostate cancer antigen 3 as a biomarker in the urine for prostate cancer diagnosis: a metaanalysis. J Cancer Res Ther 2014;10(7):218–21. [84] Xue P, Chen W, Li X. Urothelial cancer associated 1: a long noncoding RNA with a crucial role in cancer. J Cancer Res Clin Oncol 2016;142(7):1407–19. [85] Wang XS, Zhang Z, Wang HC, Cai JL, Xu QW, Li MQ, et al. Rapid identification of UCA1 as a very sensitive and specific unique marker for human bladder carcinoma. Clin Cancer Res 2006;12(16):4851–8. [86] Zhang Q, Su M, Lu G, Wang J. The complexity of bladder cancer: long noncoding RNAs are on the stage. Mol Cancer 2013;12(1):101. [87] Kamel MM, Matboli M, Sallam M, Montasser IF, Saad AS, El-Tawdi AH. Investigation of long noncoding RNAs expression profile as potential serum biomarkers in patients with hepatocellular carcinoma. Transl Res 2016;168:134–45. [88] Zheng Q, Wu F, Dai WY, Zheng DC, Zheng C, Ye H, et al. Aberrant expression of UCA1 in gastric cancer and its clinical significance. Clin Transl Oncol 2015;17(8):640–6. [89] Hong HH, Hou LK, Pan X, Wu CY, Huang H, Li B, et al. Long non-coding RNA UCA1 is a predictive biomarker of cancer. Oncotarget 2016;7(28):44442–7. [90] Li J, Wang Y, Yu J, Dong R, Qiu H. A high level of circulating HOTAIR is associated with progression and poor prognosis of cervical cancer. Tumor Biol 2015;36(3):1661–5. [91] Tang J, Jiang R, Deng L, Zhang X, Wang K, Sun B. Circulation long non-coding RNAs act as biomarkers for predicting tumorigenesis and metastasis in hepatocellular carcinoma. Oncotarget 2015;6(6):4505–15. [92] Dong L, Qi P, Xu MD, Ni SJ, Huang D, Xu QH, et al. Circulating CUDR, LSINCT-5 and PTENP1 long noncoding RNAs in sera distinguish patients with gastric cancer from healthy controls. Int J Cancer 2015;137(5):1128–35. [93] Shi J, Li X, Zhang F, Zhang C, Guan Q, Cao X, et al. Circulating lncRNAs associated with occurrence of colorectal cancer progression. Am J Cancer Res 2015;5(7):2258–65. [94] Kumarswamy R, Bauters C, Volkmann I, Maury F, Fetisch J, Holzmann A, et al. Circulating long noncoding RNA, LIPCAR, predicts survival in patients with heart failure. Circ Res 2014;114(10): 1569–75. [95] Greco S, Zaccagnini G, Perfetti A, Fuschi P, Valaperta R, Voellenkle C, et al. Long noncoding RNA dysregulation in ischemic heart failure. J Transl Med 2016;14(1):183. [96] Yan Y, Zhang B, Liu N, Qi C, Xiao Y, Tian X, et al. Circulating Long Noncoding RNA UCA1 as a Novel Biomarker of Acute Myocardial Infarction. Biomed Res Int 2016;2016:8079372.

V.  Factors Influencing Epigenetic Changes

292

18.  Epigenetic Biomarkers

[97] Kraus TF, Haider M, Spanner J, Steinmaurer M, Dietinger V, Kretzschmar HA. Altered Long Noncoding RNA Expression Precedes the Course of Parkinson’s Disease-a Preliminary Report. Mol Neurobiol 2016;2016:1–9. [98] Shi H, Cao N, Pu Y, Xie L, Zheng L, Yu C. Long non-coding RNA expression profile in minor salivary gland of primary Sjogren’s syndrome. Arthritis Res Ther 2016;18(1):109. [99] Xu XD, Li KR, Li XM, Yao J, Qin J, Yan B. Long non-coding RNAs: new players in ocular neovascularization. Mol Biol Rep 2014;41(7):4493–505. [100] Meola N, Pizzo M, Alfano G, Surace EM, Banfi S. The long noncoding RNA Vax2os1 controls the cell cycle progression of photoreceptor progenitors in the mouse retina. RNA 2012;18(1): 111–23. [101] Zhou RM, Wang XQ, Yao J, Shen Y, Chen SN, Yang H, et al. Identification and characterization of proliferative retinopathyrelated long noncoding RNAs. Biochem Biophys Res Commun 2015;465(3):324–30. [102] Wang WT, Sun YM, Huang W, He B, Zhao YN, Chen YQ. Genome-wide long non-coding RNA analysis identified circulating LncRNAs as novel non-invasive diagnostic biomarkers for gynecological disease. Sci Rep 2016;6:23343. [103] Laird PW. Principles and challenges of genomewide DNA methylation analysis. Nat Rev Genet 2010;11(3):191–203. [104] Li M, Chen WD, Papadopoulos N, Goodman SN, Bjerregaard NC, Laurberg S, et al. Sensitive digital quantification of DNA methylation in clinical samples. Nat Biotechnol 2009;27(9): 858–63.

[105] Campan M, Moffitt M, Houshdaran S, Shen H, Widschwendter M, Daxenbichler G, et al. Genome-scale screen for DNA methylation-based detection markers for ovarian cancer. PLoS One 2011;6(12):e28141. [106] Dedeurwaerder S, Defrance M, Calonne E, Denis H, Sotiriou C, Fuks F. Evaluation of the infinium methylation 450K technology. Epigenomics 2011;3(6):771–84. [107] Lee EJ, Luo J, Wilson JM, Shi H. Analyzing the cancer methylome through targeted bisulfite sequencing. Cancer Lett 2013;340(2):171–8. [108] Meissner A, Mikkelsen TS, Gu H, Wernig M, Hanna J, Sivachenko A, et al. Genomescale DNA methylation maps of pluripotent and differentiated cells. Nature 2008;454(7205):766–70. [109] Vaca-Paniagua F, Oliver J, Nogueira da Costa A, Merle P, McKay J, Herceg Z, et al. Targeted deep DNA methylation analysis of circulating cell-free DNA in plasma using massively parallel semiconductor sequencing. Epigenomics 2015;7(3):353–62. [110] Jang H, Shin H. Current trends in the development and application of molecular technologies for cancer epigenetics. World J Gastroenterol 2013;19(7):1030–9. [111] Castoldi M, Schmidt S, Benes V, Hentze MW, Muckenthaler MU. miChip: an array-based method for microRNA expression profiling using locked nucleic acid capture probes. Nat Protoc 2008;3(2):321–9. [112] Thind A, Wilson C. Exosomal miRNAs as cancer biomarkers and therapeutic targets. J Extracell Vesicles 2016;5:31292.

V.  Factors Influencing Epigenetic Changes

C H A P T E R

19 Metabolic Regulation of DNA Methylation in Mammalian Cells Regan Vryer, Richard Saffery Murdoch Children’s Research Institute, Parkville, VIC, Australia; University of Melbourne, Parkville, VIC, Australia

O U T L I N E DNA Methylation as a Dynamic Epigenetic Process Requiring Constant Methyl Donors 293 Mechanisms of DNA Demethylation

294

One-Carbon Metabolism as the Source of Methyl Donors for DNA Methylation

295

S-Adenosylmethionine: Master of the Methylome

296

Regulators of SAM Biochemical Inhibitors

298 298

DNA METHYLATION AS A DYNAMIC EPIGENETIC PROCESS REQUIRING CONSTANT METHYL DONORS Over 100 methyltransferase enzymes (including 5 DNA methyltransferases) have been described in mammals, which are responsible for the transfer of methyl groups to a large array of proteins, phospholipids, and nucleotides. These reactions are fundamental to many cellular functions, and it is therefore not surprising that insufficient methyl donor availability has the potential to disrupt a wide variety of biological processes, including DNA and RNA nucleotide synthesis; protein, RNA, and DNA methylation (Fig. 19.1); and gene expression (among others). The production of sufficient methyldonors (primarily S-adenosyl methionine; also known as SAM, SAMe, adomethionine, and adoMet) is therefore of critical importance for faithful cell division and development. Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00019-5 Copyright © 2017 Elsevier Inc. All rights reserved.

Genetic Variation and SAM

299

Methyl Donor/Cofactor Availability and DNA Methylation

299

Disruption of One-Carbon Metabolism, DNA Methylation, and Disease: What is the Link?

300

Concluding Remarks

302

References

302

Transfer of a single methyl group from SAM to a cytosine residue in DNA occurs primarily at CpG dinucleotides in mammals, though non-CpG cytosine and N(6)-adenine methylation have also been reported [1]. The maternal and paternal genomes are relatively hypomethylated at fertilization, though maintaining a gamete-specific pattern of DNA methylation [2]. Celltype specific patterns of methylation are then reestablished progressively starting in the preimplantation period with the specification of the trophectoderm and inner cell mass [3]. The de novo establishment of DNA methylation is carried out by DNMT3A and -3B methyltransferases and is modulated by DNMT3L, lacking direct catalytic activity [4] (Fig. 19.2). Very recently, another de novo methyltransferase, DNMT3C, was described with a specific role in silencing of transposable elements in the mammalian male germline [5]. In addition to the de novo establishment and removal of DNA methylation markings during early

293

294

19.  Metabolic Regulation of DNA Methylation in Mammalian Cells

MECHANISMS OF DNA DEMETHYLATION

FIGURE 19.1  DNA methylation is a covalent modification utilizing a methyl group from S-adenosyl-l-methionine (AdoMet or SAM), initiated by DNA methyltransferase enzymes (DNMTs) resulting in 5-methylcytosine (5mC).

development, the genomewide methylation profile of dividing cells is faithfully copied in newly synthesized DNA strands in daughter cells following cell division. This is carried out by the maintenance DNA methyltransferase, DNMT1 [6] (Fig. 19.2). In addition to the well-established highly dynamic nature of the embryonic DNA profile, several studies have now confirmed an ongoing “drift” in DNA methylation throughout the lifetime, resulting in the concept of an epigenetic “clock” of biological aging [7]. Mounting data has also shown dynamic remodeling of epigenetic markings during the cell cycle [8–10], highlighting the potentially highly dynamic nature of DNA methylation in certain contexts. Most recently, this has been extended to the delineation of a mitotic epigenetic “clock” that acts as a marker of cumulative cell divisions over time and is universally accelerated in human cancers [11].

In certain circumstances DNA methylation is thought to be highly dynamic, with both addition and removal of methylation at specific sequences. Although the mechanism by which methyl groups are added to CpG dinucleotides has been known since the early 1990's, the primary mechanism of DNA demethylation in mammalian cells is only now beginning to be understood. Two basic pathways have been described: passive loss of methylation and an active removal process. Passive loss is predicted to arise in the absence of the maintenance methyltransferase, DNMT1, during cell division [12], most likely via a decrease in enzyme level or activity. 5-Aza-2′-deoxycytidine (AzdC or decitabine) is a wellknown DNMT inhibitor, associated with passive DNA demethylation over time [13]. The potential for active demethylation has been the subject of intense debate for many years, but has largely remained controversial. The discovery of 5-hydroxymethylcytosine (5hmC) [14], 5-formylcytosine (5fC), and 5-carboxycytosine (5caC) [15], in addition to the TET family of enzymes [16], has rapidly changed this situation. It is now apparent that highly specific DNA demethylation cycle exists in several cell types and developmental settings. The TET group of enzymes converts 5-methylcytosine (5mC) to 5hmC and then to 5fC and 5caC [15,17]. The latter is then deaminated to uracil and replaced with cytosine by the cellular DNA repair machinery [15,17,18] (Fig. 19.3).

FIGURE 19.2  Maintenance versus de novo DNA methylation. (A) DNMT1 is a maintenance methyltransferase that recognizes hemimethyated DNA following DNA replication to ensure faithful maintenance of DNA methylation following cell division. (B) In contrast, DNMT3A, -3B, and -3C are de novo methyltransferases with the capacity to add methylation groups to umethylated CpG sites to DNA in a replication-independent manner. DNMT3A–C act in distinct developmental and genomic contexts and utilize DNMT3L (lacking enzyme activity) as a coregulator.

V.  Factors Influencing Epigenetic Changes



One-Carbon Metabolism as the Source of Methyl Donors for DNA Methylation

295

FIGURE 19.3  The DNA methylation cycle. A proposed DNA methylation cycle involving DNMT addition of methyl groups from SAM to cytosine to generated 5mC. The ten-eleven translocation (TET) family of enzymes oxidize 5mC to 5-hydroxymethylcytosine (5hmC), 5-formylcytosine (5fC), and 5-carboxycytosine (5caC). Both 5fC and 5caC are templates for DNA base excision repair (BER) and are replaced by unmethylated cytosine. Blue arrows indicate a direct conversion of one modified nucleotide to another by conversion enzymes (DMNT or TET). Dashed blue lines indicate molecules acting as substrates for downstream reactions (BER).

ONE-CARBON METABOLISM AS THE SOURCE OF METHYL DONORS FOR DNA METHYLATION As around 4% of all cytosine nucleotides within genomic DNA of mammalian somatic cells are methylated [19], the constant availability of a pool of methyl donors in dividing cells is necessary if the methylation pattern (methylome) is to be faithfully replicated. The methyl groups required for establishment and maintenance of DNA methylation are derived solely from dietary methyl donors in association with specific enzymes and associated cofactors [20]. DNMTs catalyze the transfer of methyl groups from a common methyl donor produced by one-carbon metabolism, SAM. This molecule represents the universal methyl donor in all cells [21] (Fig. 19.1). Folates are the primary methyl donors and key mediators of one-carbon metabolic pathways along with choline and other cofactors, such as B-group vitamins B2, B6, and B12. Folate in blood plasma exists predominantly as polyglutamated methyl-tetrahydrofolate (methyl THF; ① in Fig. 19.4). Following transport into cytoplasm, primarily by the reduced folate carrier (RFC), methyl THF acts as a methyl group donor for the production of tetrahydrofolate (THF; ② in Fig. 19.4) and a precursor for homocysteine (2-amino-4-mercaptobutyric acid) conversion to methionine (③ in Fig. 19.4). The process is catalyzed by methionine synthase (MS; ④ in Fig. 19.4),

and requires cofactor vitamin B12. As MS is the only reaction that utilizes methyl THF, folates remain “trapped” in this form by any impairment of the MS-catalyzed reaction [22,23]. Methionine is further activated to SAM by methionine adenosyltransferase (MAT; ⑤ in Fig. 19.4). SAM synthesis represents the preferred catabolic pathway for methionine in the liver, where up to half of the daily intake of methionine is converted to SAM and the majority of all subsequent methylation reactions take place [24]. Such reactions invariably result in conversion of SAM to S-adenosylhomocysteine (SAH; ⑥ in Fig. 19.4), which is further hydrolyzed back to homocysteine by SAH hydrolase (SAHH; ⑦ in Fig. 19.4). SAH is a potent competitive inhibitor of transmethylation reactions; disruption of the SAH:SAM ratio, through an increase in SAH or a decrease in SAM, leads to inhibition of transmethylation reactions [25,26]. For this reason, SAH removal is essential and is carried out by SAHH, but only in the absence of downstream products adenosine and homocysteine [27]. Folic acid is a synthetic form of folate used widely in food fortification and dietary supplements. Unlike folate, it is reduced prior to entering the one-carbon cycle, first to dihydrofolate (DHF) and then to THF (⑧ in Fig. 19.4). Both processes are mediated by dihydrofolate reductase (DHFR). THF formed in this manner is further converted and recycled to 5,10-methylene THF with the aid of vitamin B6 (⑨ in Fig. 19.4), and to 5-methyl THF by methylene

V.  Factors Influencing Epigenetic Changes

296

19.  Metabolic Regulation of DNA Methylation in Mammalian Cells

FIGURE 19.4  Schematic illustration of the one-carbon metabolic pathway. Precursors and cofactors obtained exclusively from the diet are highlighted in orange. Red dashed lines indicate inhibition of enzymatic reactions: acetylaldehyde directly inhibits folate and absorption of folate. It also downregulates methionine synthase (MS) [80,88]. High levels of S-adenosylhomocysteine (SAH) downregulate the activity of methyltransferase (MT). Increased concentration of SAM inhibits the activity of methylene tetrahydrofolate (THF) reductase (MTHFR), limiting the bioavailability of 5-methyl THF [20]. Cadmium is an inhibitor of DNMT activity [122]. Green dashed line indicates upregulation of specific pathways. All mammalian tissues express methionine adenosyltransferase (MAT) and MS, whereas betaine homocysteine methyltransferase (BHMT) is found only in the liver and kidney. S-adenosylmethionine (SAM) inhibits MTHFR and MS and activates cystathionine beta-synthase (CBS), leading to homocysteine channeling down the transsulfuration pathway. DHF, Dihydrofolate; DMG, dimethylglycine; FR, folate receptor; RFC, reduced folate carrier; SHMT, serine hydroxymethyltransferase; TS, thymidylate synthase; X, substrates for methylation.

THF reductase (MTHFR) with the aid of vitamin B2 (⑩ in Fig. 19.4). The former reaction is associated with conversion of serine to glycine, catalyzed by serine hydroxymethyltransferase (SHMT). Folic acid may have a negative effect on the absorption of naturally occurring folates, as it binds endogenous receptors with a higher affinity than THF [28]. Chemical structural formulas for the key compounds involved in this pathway are shown in Table 19.1.

S-ADENOSYLMETHIONINE: MASTER OF THE METHYLOME SAM is an essential precursor molecule and the principal methyl donor in all organisms (transmethylation). It is also the precursor of aminopropyl groups

(aminopropylation) and of glutathione (transsulfuration) in the liver [27]. Thought to be second only to ATP in terms of the number of reactions it facilitates, SAM is critical for a wide range of metabolic reactions, such as nucleic acid synthesis and histone methylation, in addition to DNA methylation. Further, SAM has an important role in the regulation of activity of several enzymes (discussed earlier), and evidence is mounting for independent actions on cell growth, apoptosis, and differentiation, independent of its role as a methyl donor [29]. The structure of SAM was first reported by Catoni in 1951 [30]. This revealed the presence of a high-energy sulfonium ion (Fig. 19.5) that activates each of the attached carbon atoms toward nucleophilic attack, making it an excellent substrate for distinct biochemical reactions, such as DNA methylation. SAM is produced following

V.  Factors Influencing Epigenetic Changes



S-Adenosylmethionine: Master of the Methylome

TABLE 19.1 Chemical Structures of the Key Compounds Involved in the One-Carbon Metabolic Pathway Compound name

Chemical structures

Folate

THF

DHF

5-Methyl THF

5,10-Methylene THF

Homocysteine

Methionine

DHF, Dihydrofolate; THF, tetrahydrofolate.

V.  Factors Influencing Epigenetic Changes

297

298

19.  Metabolic Regulation of DNA Methylation in Mammalian Cells

FIGURE 19.5  Conversion of SAM to SAH. The high-energy sulfonium ion (orange shading) of SAM activates each of the attached carbons toward nucleophilic attack, facilitating methyl donor transfer.

transfer of the adenosyl moiety of ATP to methionine by the MAT enzyme. In the majority of instances SAM reacts by transfer of the S-associated methyl group to acceptor molecules as part of transmethylation reactions [27]. The common product for all reactions is SAH. Most transmethylation reactions are inhibited by SAH, whereas decreasing SAM stimulates enzyme activity [26]. In this way, SAM inhibits formation of THF, involved in transmethylation, while at the same time stimulates the transsulfuration pathway that converts homocysteine (a byproduct of transmethylation reactions) to cystathione and cysteine [31] (⑪ in Fig. 19.4). Interestingly, oxidative stress may play an important role in determining the balance between transsulfuration and transmethylation reactions. The MS enzyme is inactivated by oxidation and requires reductive methylation for reactivation [32].

REGULATORS OF SAM The production of SAM as a methyl donor for DNA methylation involves a complex interplay between metabolic, cofactor bioavailability and enzymatic activity, the latter being also influenced by genetic variability.

Biochemical Inhibitors As SAH, produced during DNA methylation, functions as an inhibitor of SAM-dependent methyltransferases (Fig. 19.4), removal of this byproduct is a prerequisite if the methylation demand within the cell is to be met [20]. Homocysteine represents a point of intersection between the methionine cycle and transsulfuration pathway. Homocysteine is a nonessential amino acid, derived from metabolism of dietary methionine [20], which causes toxicity upon accumulation [33]. Homocysteine is produced following hydrolysis of SAH by SAHH [34]. Homocysteine produced via hydrolysis of

SAH can either undergo remethylation, yielding methionine, or transsulfuration, yielding cysteine [20,34]. It can also be exported into extracellular fluids [34]. The liver is thought to make the greatest contribution to the plasma homocysteine level via homocysteine export [35]. Regulation of homocysteine distribution between methylation and transsulfuration by SAM is speculated to occur primarily in the liver, which is unique in its ability to synthesize excess SAM in response to methionine availability. In nonhepatic tissues the level of SAM is tightly regulated and SAH may be the dominant effector [25]. The basic methionine cycle occurs in all mammalian cells. In contrast, transsulfuration has a limited distribution in mammalian cells, and those lacking this pathway require an exogenous source of cysteine [25]. Remethylation of homocysteine to methionine can be catalyzed by MS, which requires vitamin B12 as a cofactor [20]. This is recycled via the action of MTHFR, an important reaction in one-carbon metabolism with the potential to influence DNA methylation [20]. In the liver, homocysteine can also be remethylated to methionine by betaine homocysteine methyltransferase (BHMT) [34]. As the pathways for the metabolism of folate and choline intersect at the conversion of homocysteine to methionine, folate and choline act together to decrease homocysteine concentration. Impairment of these pathways, as a result of dietary deficiency or genetic polymorphisms that reduce enzyme activity (see further), leads to elevated plasma homocysteine concentrations. Hyperhomocysteinemia is an independent risk factor for several diseases, including cancer, mental health disorders, and vascular disease [36–40]; however, it remains unclear whether this is a causative agent or a marker of specific pathologies. Glycine N-methyltransferase (GNMT) is a SAMdependent enzyme present in the liver, kidney, and pancreas. It plays an important role in regulating the ratio of SAM to SAH within the cell [34], which can be considered a “methylation potential” or indicator of capacity

V.  Factors Influencing Epigenetic Changes



Methyl Donor/Cofactor Availability and DNA Methylation

for functional SAM activity [41]. GNMT activity is inhibited by allosteric binding of 5-methyl THF, the synthesis of which is inhibited by SAM [20]. The activity of GNMT can therefore be regulated by the intracellular concentrations of SAM and 5-methyl THF. The SAM/SAH ratio is regulated via inhibition of SAM by MTHFR, and of GNMT by folate compounds [34]. Folate deficiency in rats leads to increased GNMT activity and a decrease in the SAM/SAH ratio [42].

Genetic Variation and SAM The rate of passage through the one-carbon cycle can be influenced by genetic polymorphisms in genes encoding the enzymes involved in this pathway [43]. The C to T substitution at nucleotide 677 of the MTHFR gene (677C>T; rs1801133) has been most widely studied in this regard. This common SNP results in a more thermolabile enzyme with lower catalytic activity, lower levels of 5-methyl THF, and increasing levels of 5,10-methylene THF and plasma homocysteine. TT homozygous individuals generally have lower levels of DNA methylation relative to CC homozygotes [43,44]. This variant of MTHFR has been linked to an increased risk of adverse pregnancy outcomes, various cancers, coronary disease, atherosclerosis, alcoholic liver disease, Down Syndrome, and neuropsychiatric conditions in humans, in both case–control studies and metaanalyses [45–48]. A second common variant of MTHFR (1298C>A; rs1801131) has also been widely studied and shown to reduce enzymatic activity, without altering thermolability [49,50]. Interestingly, opposite effects on red blood cell folate levels have been reported for each of the common MTHFR variants [51], and protective effects against disease have also been reported for these variants in some studies. A recent genomewide association study confirmed genetic variation in the MTHFR gene as a major determinant of serum homocysteine levels [52]. Other previously described genetic variants with the potential to disrupt one-carbon metabolic include MTHFD1 (rs2236225, rs1950902), MTHFD2 (rs1667627), MTRR (rs1532268), MTR (rs1805087), BHMT (rs3733890), RFC-1 (rs1051266), and SHMT (rs1979277) [53–58]. In addition to the proposed link between genetic variants of enzymes regulating one-carbon metabolism, variants in DNMTs have also been identified as risk factors for disease, including DNMT1 in systemic lupus erythematosus [59]. Genetic deficiency of DNMT3B causes a recessive human disorder characterized by immunodeficiency, centromere instability, and facial anomalies [60]. Variants in other DNMTs have been associated with increased cancer risk [61–64], and are also widely found in adult cancers, particularly DNMT3A mutations in adult acute myeloid leukemias [65,66].

299

METHYL DONOR/COFACTOR AVAILABILITY AND DNA METHYLATION Available evidence from both animal and human studies suggests that the effects of folate deficiency on DNA methylation are very complex, being cell type and target organ dependent and gene and site specific [67,68]. Both circumstantial and direct evidence exists for a link between disruptions in the folate pathway (Fig. 19.4), changes in DNA methylation, altered gene expression, and disease predisposition. Animal studies have shown that prenatal feeding of a methyl-supplemented diet can increase DNA methylation and decrease expression of genes in offspring [69–71], while limiting folate supply in humans results in increasing levels of homocysteine and reduced DNA methylation [40,72–74]. For example, colonic DNA methylation in humans has been positively correlated with serum and RBC folate concentrations and negatively correlated with plasma homocysteine concentrations [75]. Conversely, folate supplementation at 2.5–25 times the daily requirement for 3–12 months significantly increases genomic DNA methylation in subjects with resected colorectal adenoma or cancer [75–78]. In the Caenorhabditis elegans model silencing of the genes involved with folate metabolism pathways occurred, when oversupplementation of folate decreased MTHFR and MS but favored thymidylate synthase (TS), thus disrupting folate homeostasis and inducing DNA damage. mRNA expression is adapted and impaired during both high and low levels of folate [79]. Dietary folates (THFs) are required as cofactors for reactions involved in one-carbon metabolism. 5,10-Methylene THF is reduced to 5-methyl THF by MTHFR. 5-Methyl THF is required for conversion of homocysteine to methionine, and therefore influences the availability of methyl donor SAM [33]. Dietary choline is oxidized to betaine, which can be utilized in an alternative pathway for the conversion of homocysteine to methionine, also influencing availability of SAM [80]. Low dietary intake of folate and choline decreases concentrations of SAM, resulting in hypomethylation of DNA. Numerous studies have demonstrated that additional dietary or behavioral factors, such as alcohol consumption, can affect the bioavailability of SAM, and therefore DNA methylation levels [80–84]. Alcohol antagonizes one-carbon metabolism by preventing the uptake of folate in cells [85]. In addition, alcohol has been reported to adversely affect both the availability and metabolism of folate through diminished or inadequate dietary intake, impaired intestinal absorption, and increased loss via renal excretion [86]. Direct effects of alcohol on folate metabolism occur primarily via inhibition of MS, resulting in decreased concentrations of downstream products, methionine and SAM (Fig. 19.4), and increased concentrations of

V.  Factors Influencing Epigenetic Changes

300

19.  Metabolic Regulation of DNA Methylation in Mammalian Cells

precursors, homocysteine, and SAH [86]. A study utilizing a rodent model of alcoholism found that prolonged heavy alcohol consumption resulted in decreased tissue levels of SAM, increased tissue levels of SAH, and substantial global hypomethylation of DNA in the colonic mucosa [87]. In another study, inhibition of MS by alcohol in a rodent model also resulted in an increase in BHMT activity and a decrease in betaine level [88]. Alcoholism is associated with alterations in DNA methylation at both global and gene-specific levels [89]. It is also linked to increased risk of cancer, especially in the colon and liver. Paternal chronic alcohol consumption has been linked to impaired DNMT function in sperm. This further leads to the disruption of genomic imprinting and altered fetal growth [90].

DISRUPTION OF ONE-CARBON METABOLISM, DNA METHYLATION, AND DISEASE: WHAT IS THE LINK? Mounting evidence suggests a direct link between disruption of one-carbon metabolism and disease pathogenesis. This can arise either through genetic variability (described earlier), environmental factors (dietary, lifestyle, or exposures), or a combination of both. There is little doubt that the downstream consequences of this disruption are manifold and include changes to genomic DNA methylation and an imbalance in homocysteine levels, each of which has been associated with an increasing disease risk in numerous studies. Perhaps the most compelling evidence for a link between altered one-carbon metabolism and disease comes from the success attributed to folic acid supplementation in reducing the rate of neural tube defects (NTDs) [86]. However, given the interplay between one-carbon production, DNA methylation, and other fundamental cellular activities, the direct contribution of DNA methylation to human disease risk remains controversial. Irrespective of this, animal studies have proven to be invaluable in confirming much of the circumstantial data obtained in human studies linking insufficient methyl donors, DNA hypomethylation, and increasing disease risk. For example, insufficient zinc intake is associated with decreasing levels of SAM in the liver, DNA hypomethylation, low birth weight, and reduced growth in rats [81,91]. Furthermore, in colonic and gastric cancer cell lines treated with SAM, hypermethylation of the oncogenes c-myc and H-ras promoters was observed, indicating a potential therapeutic utilization of SAM and providing evidence for its role in the regulation of oncogenic factors [92]. Similarly, another cosubstrate of onecarbon metabolism, selenium, is also shown to induce hypomethylation in rat colon and liver with increased homocysteine levels in the plasma [93]. A similar finding

has been reported in a study using a human colon cancer cell line (Caco-2) [94]. An interesting study has also found that the group of rats fed with selenium were less susceptible to colon cancer than the control group, suggesting adequate selenium intake may have a protective effect against cancer [95]. Despite these findings, the mechanism(s) by which selenium interacts with other enzymes and substrates in the one-carbon metabolism still remains to be clarified. There is compelling evidence for folate deficiency increasing disease risk. In general, most studies have found that folate deficiency is directly associated with low SAM levels. This limits the availability of the universal methyl donor and induces global hypomethylation, in addition to specific promoter hypomethylation of certain genes associated with tumorigenesis, which is associated with some cancers [23,96–98]. Interestingly, some suggest that the link between folate intake and risk of cancer may be dosage and time dependent [23,97,98]. Inadequate folate may contribute to cancer progression, but high folate levels may also promote cancer growth once established [97,98]. This may be due to the role of folate in synthesizing nucleotides that are required for rapid DNA replication and cell division in cancer; thus high folate levels may enhance tumor progression [23,99]. Therefore, it has been suggested that there exists an optimal folate intake for cancer prevention [97]. Additionally, the current evidence suggests that inadequate intake of vitamins B6 or B12, along with low folate, decreases the SAM levels, which, in turn, causes increased levels of homocysteine in blood plasma due to a block in the one-carbon cycle. Both choline and methionine deficiencies have also been linked to increased risk of cancer with a further elevated risk when the deficiencies occur in combination with each other or with folate deficiency [81]. The relationship between deficiencies in one-carbon metabolites/cofactors and cancer development/progression seems to be complex and requires further investigation [97]. Given the accumulated data linking alcohol and folate antagonism through reduced folate absorption, increased folate excretion, and direct alteration of MS activity [81,89], it is not surprising that supplementing one-carbon metabolism has been investigated as a treatment for diseases associated with chronic alcohol consumption. Whereas inhibition of MS activity by chronic overconsumption of alcohol creates a condition that cannot be overcome by dietary folate supplementation alone, betaine supplementation can alleviate alcohol-induced alterations to one-carbon metabolism in the liver by restoring levels of SAM and preventing increased export of homocysteine [34]. Direct administration of SAM to subjects with alcoholic liver disease has been associated with improvements in survival [100], and can prevent

V.  Factors Influencing Epigenetic Changes



Disruption of One-Carbon Metabolism, DNA Methylation, and Disease: What is the Link?

the development of liver cancer in at least some rodent models [101], but does not appear to reverse cancer progression once established [102]. Additionally, rodent models have shown a wide range of adverse phenotypes associated with cardiovascular disease due to low maternal folate status, including hypertension and increased insulin resistance [103]. Indeed, in humans hyperhomocysteinemia is an independent risk factor for several diseases, including cancer, mental health disorders, and vascular disease [36–40,104]; however, it remains unclear whether this is a causative agent or a marker of specific pathologies. Prenatal influences on fetal development have been a widely growing field, made prominent by the Development Origins of Health and Disease (DOHAD) hypothesis. The interplay between genetics, dietary exposures, epigenetics, and adverse developmental outcomes typifies DOHAD research. The primary outcome of folate deficiency during pregnancy is NTDs, leading to fetal and infant morbidity. Epidemiologically, studies have shown preconceptional fortification of folate to have protective effects against NTDs [105], and a significant reduction in the prevalence of spina bifida and encephaly [106]. NTDs arise from a failure of the neural tube to close, during the first few weeks of embryonic development, but mechanistically, is not fully understood. Epigenetic mechanisms can explain the progression of this developmental aberration. Aberrant DNA methylation has been characterized in brain tissue from human fetuses suffering from a NTD, as well as a folate-deficient intrauterine environment [107]. A suggested, but not exhaustive, role of folate in prevention of NTDs is in the upregulation of KDM6B expression, a gene that is involved with decreasing H3K27 methylation on the Hes1 and Neurog2 genes, both tightly involved in the formation of the neural tube [108]. The maternal/fetal interface has been suggested as a contributor of disease and cancer progression via an epigenetic pathway, with some studies linking the methylation of genes involved with endocrine function and cofactor availability [109]. One-carbon metabolite interactions during pregnancy are increasingly of interest, given the common shared metabolic pathways. Cobalamin deficiency is strongly associated with increased total homocysteine [110] and reduced folate associated with a higher DMG/betaine ratio [111]. The implications of these interactions on developmental health is yet to be elucidated, though some studies do show a dichotomous effect between folate and cobalamin with regards to insulin resistance [112] and high adiposity [113] in children and gestational diabetes [114]. Conversely, high levels of folate and cobalamin during pregnancy are associated with greater instances of atopic dermatitis in children [115]. While many of the maternal pre- and perinatal influences have been investigated and characterized, paternal

301

influences are not well understood, but are an increasingly researched area in the DOHAD field. In the initial phases of zygote formation, the paternal and maternal genomes are demethylated, with some exceptions (imprinted genes), while new patterns of methylation are also established (de novo) [116]. Maintenance of these parental epigenomes is crucial for optimal fetal development. Rat models show that chronic exposure to alcohol leads to reduced mRNA of cytosine methyltransferases in sperm, potentially mediating a loss of methylation in key imprinted genes, leading to impaired imprinting in the offspring [117]. Paternal diet in rodent models indicate that deficiency in folate can lead to a reduction in fertility, and an increase in birth defects in the offspring [118]. The general mechanism of such changes occurs through changes in the sperm. During cell replication and folate deficiency, there is a lack of nucleotide subunits available for DNA synthesis, leading to misincorporation of uracil to form DNA, and subsequent DNA damage [119]. In addition, the epigenome of the sperm from a folate-deficient mouse is differentially methylated relative to sperm from a nutrient-sufficient mouse, including in regions involved in the nervous system, kidney, spleen, gastrointestinal tract, and muscle function, as well as in genes associated with chronic conditions, such as cancer, diabetes, and autism [118]. Future studies into nongenetic paternal inheritance should examine the effect of one-carbon pathway metabolites and their effects on the gametic epigenome closely. The past decade in epigenetic studies has revealed the importance and potentiality of epigenomics. Instead of identifying individual epigenetic variation implicated in disease, an epigenomic perspective provides a snapshot of an epigenetic profile. The DNA methylome, or the interrogation of methylation across the entire genome, has been examined in the context of dietary exposures and their downstream effects. Folate-sensitive regions have been observed that show differential methylation at gene regions, such as ZFP57, implicated in regulation of DNA methylation and imprinting [120]. More recently, large-scale epigenomewide analyses of cord blood using data pooled samples from two large cohorts (n = 1275 and 713) identified 443 CpG sites of differential methylation. Counterintuitively, elevated levels of maternal plasma folate were associated with a decrease in methylation at 416 CpG sites, with only 27 CpG sites showing a loss of methylation [121]. These were enriched in genes implicated in cell, embryonic, and neural development. In identifying a candidate pathway gene, the most significantly differentially methylated sites were found in APC2, implicated in brain development and wnt signaling [121]. Additional differentially methylated genes included SLC16A12, KLK4, LHX1, IHH, ROBO3, PCSK9, FAM83H, and GJA3, all implicated in non-NTD developmental abnormalities.

V.  Factors Influencing Epigenetic Changes

302

19.  Metabolic Regulation of DNA Methylation in Mammalian Cells

In the examination of other one-carbon metabolites, B12 and choline were not found to confound the association between differential methylation and maternal folate. While it would be expected that increased folate increased methylation levels generally, this was not apparent. One explanation posits that at high folic acid levels, the inhibitory effect of SAM on MTHFR is disrupted, leading to a reduction in the SAM/SAH ratio and decreased methylation [121]. Further epigenomic and metabolic studies investigating these dynamics are required to fully elucidate the effect of folate and supplemented folic acid on methylation.

CONCLUDING REMARKS The metabolic pathways that generate the primary methyl donors needed for the de novo establishment and maintenance of the DNA methylation profile are complicated and do not exist in isolation from other essential biochemical reactions necessary for cell survival, division, and differentiation. However, despite the “mountains” of conflicting experimental data, the significance of methyl donor production and insufficiency is unequivocal. To date, unraveling the mechanisms underlying disease risk associated with deficiencies in these pathways has proven problematic, and in many cases, controversial. The recent rise in reports providing novel insights into the interaction of environmental, genetic, and epigenetic determinants suggest that an understanding of these processes is likely to be achieved in the not too distant future. Application of emergent technologies that can be applied to high-resolution studies of DNA methylation across the genome (e.g., highthroughput sequencing), will further speed progress toward this important goal.

References [1] Wu TP, Wang T, Seetin MG, Lai Y, Zhu S, Lin K, et al. DNA methylation on N(6)-adenine in mammalian embryonic stem cells. Nature 2016;532(7599):329–33. [2] Wang L, Zhang J, Duan J, Gao X, Zhu W, Lu X, et al. Programming and inheritance of parental DNA methylomes in mammals. Cell 2014;157(4):979–91. [3] Kelsey G. Genomic imprinting—roles and regulation in development. Endocr Dev 2007;12:99–112. [4] Aapola U, Kawasaki K, Scott HS, Ollila J, Vihinen M, Heino M, et al. Isolation and initial characterization of a novel zinc finger gene, DNMT3L, on 21q22.3, related to the cytosine-5-methyltransferase 3 gene family. Genomics 2000;65(3):293–8. [5] Barau J, Teissandier A, Zamudio N, Roy S, Nalesso V, Herault Y, et al. The DNA methyltransferase DNMT3C protects male germ cells from transposon activity. Science 2016;354(6314):909–12. [6] Bestor TH. The DNA methyltransferases of mammals. Human Mol Genet 2000;9(16):2395–402. [7] Jones MJ, Goodman SJ, Kobor MS. DNA methylation and healthy human aging. Aging Cell 2015;14(6):924–32.

[8] Brown SE, Fraga MF, Weaver ICG, Berdasco M, Szyf M. Variations in DNA methylation patterns during the cell cycle of HeLa cells. Epigenetics 2007;2(1):54–65. [9] Kangaspeska S, Stride B, Métivier R, Polycarpou-Schwarz M, Ibberson D, Carmouche RP, et al. Transient cyclical methylation of promoter DNA. Nature 2008;452(7183):112–5. [10] Métivier R, Gallais R, Tiffoche C, Le Péron C, Jurkowska RZ, Carmouche RP, et al. Cyclical DNA methylation of a transcriptionally active promoter. Nature 2008;452(7183):45–50. [11] Yang Z, Wong A, Kuh D, Paul DS, Rakyan VK, Leslie RD, et al. Correlation of an epigenetic mitotic clock with cancer risk. Genome Biol 2016;17(1):205. [12] Patra SK, Patra A, Rizzi F, Ghosh TC, Bettuzzi S. Demethylation of (Cytosine-5-C-methyl) DNA and regulation of transcription in the epigenetic pathways of cancer development. Cancer Metastasis Rev 2008;27(2):315–34. [13] Yoo CB, Jones PA. Epigenetic therapy of cancer: past, present and future. Nat Rev Drug Discov 2006;5(1):37–50. [14] Kriaucionis S, Heintz N. The nuclear DNA base 5-hydroxymethylcytosine is present in Purkinje neurons and the brain. Science 2009;324(5929):929–30. [15] Ito S, Shen L, Dai Q, Wu SC, Collins LB, Swenberg JA, et al. Tet proteins can convert 5-methylcytosine to 5-formylcytosine and 5-carboxylcytosine. Science 2011;333(6047):1300–3. [16] Ono R, Taki T, Taketani T, Taniwaki M, Kobayashi H, Hayashi Y. LCX, leukemia-associated protein with a CXXC domain, is fused to MLL in acute myeloid leukemia with trilineage dysplasia having t(10;11)(q22;q23). Cancer Res 2002;62(14):4075–80. [17] Ito S, D’Alessio AC, Taranova OV, Hong K, Sowers LC, Zhang Y. Role of Tet proteins in 5mC to 5hmC conversion, ES-cell self-renewal and inner cell mass specification. Nature 2010;466(7310):1129–33. [18] Pastor WA, Aravind L, Rao A. TETonic shift: biological roles of TET proteins in DNA demethylation and transcription. Nat Rev Mol Cell Biol 2013;14(6):341–56. [19] Gama-Sosa MA, Midgett RM, Slagel VA, Githens S, Kuo KC, Gehrke CW, et al. Tissue-specific differences in DNA methylation in various mammals. Biochim Biophys Acta 1983;740(2):212–9. [20] Ulrey CL, Liu L, Andrews LG, Tollefsbol TO. The impact of metabolism on DNA methylation. Human Mol Genet 2005;14:R139–47. [21] Stanger O. Physiology of folic acid in health and disease. Curr Drug Metab 2002;3(2):211–23. [22] Ulrich CM, Reed MC, Nijhout HF. Modeling folate, one-carbon metabolism, and DNA methylation. Nutr Rev 2008;66(Suppl. 1):S27–30. [23] Smith AD, Kim Y-I, Refsum H. Is folic acid good for everyone? Am J Clin Nutr 2008;87(3):517–33. [24] Mudd SH, Poole JR. Labile methyl balances for normal humans on various dietary regimens. Metabolism 1975;24(6):721–35. [25] Finkelstein JD. Methionine metabolism in mammals. J Nutr Biochem 1990;1(5):228–37. [26] Mato JM, Alvarez L, Ortiz P, Pajares MA. S-adenosylmethionine synthesis: molecular mechanisms and clinical implications. Pharmacol Therapeut 1997;73(3):265–80. [27] Lu SC. S-Adenosylmethionine. Int J Biochem Cell Biol 2000;32(4):391–5. [28] Qiu A, Jansen M, Sakaris A, Min SH, Chattopadhyay S, Tsai E, et al. Identification of an intestinal folate transporter and the molecular basis for hereditary folate malabsorption. Cell 2006;127(5):917–28. [29] Lu SC, Mato JM. S-Adenosylmethionine in cell growth, apoptosis and liver cancer. J Gastroenterol Hepatol 2008;23(Suppl. 1):S73–7. [30] Catoni GL. Methylation of nicotinamide with a soluble enzyme system from rat liver. J Biol Chem 1951;189:203–16.

V.  Factors Influencing Epigenetic Changes



REFERENCES

[31] Selhub J. Homocysteine metabolism. Annu Rev Nutr 1999;19:217–46. [32] Chen Z, Banerjee R. Purification of soluble cytochrome b5 as a component of the reductive activation of porcine methionine synthase. J Biol Chem 1998;273(40):26248–55. [33] Beaudin AE, Stover PJ. Folate-mediated one-carbon metabolism and neural tube defects: balancing genome synthesis and gene expression. Birth Defects Research C 2007;81(3):183–203. [34] Grillo MA, Colombatto S. S-adenosylmethionine and its products. Amino Acids 2008;34(2):187–93. [35] Stead LM, Brosnan ME, Brosnan JT. Characterization of homocysteine metabolism in the rat liver. Biochem J 2000;350(Pt. 3):685–92. [36] Boushey CJ, Beresford SA, Omenn GS, Motulsky AG. A quantitative assessment of plasma homocysteine as a risk factor for vascular disease. Probable benefits of increasing folic acid intakes. JAMA 1995;274(13):1049–57. [37] Fanapour PC, Yug B, Kochar MS. Hyperhomocysteinemia: an additional cardiovascular risk factor. WMJ 1999;98(8):51–4. [38] Herrmann W. The importance of hyperhomocysteinemia as a risk factor for diseases: an overview. Clin Chem Lab Med 2001;39(8):666–74. [39] Wu LL, Wu JT. Hyperhomocysteinemia is a risk factor for cancer and a new potential tumor marker. Clin Chim Acta 2002;322(1–2):21–8. [40] Muskiet FAJ. The importance of (early) folate status to primary and secondary coronary artery disease prevention. Reprod Toxicol 2005;20(3):403–10. [41] Mato JM, Alvarez L, Ortiz P, Mingorance J, Durán C, Pajares MA. S-adenosyl-l-methionine synthetase and methionine metabolism deficiencies in cirrhosis. Adv Exp Med Biol 1994;368:113–7. [42] Ogawa H, Gomi T, Takusagawa F, Fujioka M. Structure, function and physiological role of glycine N-methyltransferase. Int J Biochem Cell Biol 1998;30(1):13–26. [43] Friso S, Choi S-W, Girelli D, Mason JB, Dolnikowski GG, Bagley PJ, et al. A common mutation in the 5,10-methylenetetrahydrofolate reductase gene affects genomic DNA methylation through an interaction with folate status. Proc Natl Acad Sci USA 2002;99(8):5606–11. [44] Castro R, Rivera I, Ravasco P, Camilo ME, Jakobs C, Blom HJ, et al. 5,10-Methylenetetrahydrofolate reductase (MTHFR) 677C>T and 1298A>C mutations are associated with DNA hypomethylation. J Med Genet 2004;41(6):454–8. [45] Loenen WAM. S-adenosylmethionine: jack of all trades and master of everything? Biochem Soc Transact 2006;34(Pt. 2):330–3. [46] Ueland PM, Hustad S, Schneede J, Refsum H, Vollset SE. Biological and clinical implications of the MTHFR C677T polymorphism. Trends Pharmacol Sci 2001;22(4):195–201. [47] Kim Y-I. 5,10-Methylenetetrahydrofolate reductase polymorphisms and pharmacogenetics: a new role of single nucleotide polymorphisms in the folate metabolic pathway in human health and disease. Nutr Rev 2005;63(11):398–407. [48] Shi J, Gershon ES, Liu C. Genetic associations with schizophrenia: meta-analyses of 12 candidate genes. Schizophrenia Res 2008;104(1–3):96–107. [49] van Gabreels F, Stevens EM, Smeitink JA, Trijbels FJ, der Put NM, Eskes TK, et al. A second common mutation in the methylene tetrahydrofolate reductase gene: an additional risk factor for neural-tube defects? Am Genet 1998;62:1044–51. [50] Weisberg I, Tran P, Christensen B, Sibani S, Rozen R. A second genetic polymorphism in methylenetetrahydrofolate reductase (MTHFR) associated with decreased enzyme activity. Mol Genet Metab 1998;64(3):169–72. [51] Parle-McDermott A, Mills JL, Molloy AM, Carroll N, Kirke PN, Cox C, et al. The MTHFR 1298CC and 677TT genotypes have opposite associations with red cell folate levels. Mol Genet Metab 2006;88(3):290–4.

303

[52] Tanaka T, Scheet P, Giusti B, Bandinelli S, Piras MG, Usala G, et al. Genome-wide association study of vitamin B6, vitamin B12, folate, and homocysteine blood concentrations. Am J Human Genet 2009;84(4):477–82. [53] Fredriksen A, Meyer K, Ueland PM, Vollset SE, Grotmol T, Schneede J. Large-scale population-based metabolic phenotyping of thirteen genetic polymorphisms related to one-carbon metabolism. Human Mutat 2007;28(9):856–65. [54] Gast A, Bermejo JL, Flohr T, Stanulla M, Burwinkel B, Schrappe M, et al. Folate metabolic gene polymorphisms and childhood acute lymphoblastic leukemia: a case-control study. Leukemia 2007;21(2):320–5. [55] Koushik A, Kraft P, Fuchs CS, Hankinson SE, Willett WC, Giovannucci EL, et al. Nonsynonymous polymorphisms in genes in the one-carbon metabolism pathway and associations with colorectal cancer. Cancer Epidemiol Biomarkers Prevent 2006;15(12):2408–17. [56] Ulvik A, Ueland PM, Fredriksen A, Meyer K, Vollset SE, Hoff G, et al. Functional inference of the methylenetetrahydrofolate reductase 677C>T and 1298A>C polymorphisms from a largescale epidemiological study. Human Genet 2007;121(1):57–64. [57] Molloy AM. Folate and homocysteine interrelationships including genetics of the relevant enzymes. Curr Opin Lipidol 2004;15(1):49–57. [58] Lucock M, Yates Z. Folic acid—vitamin and panacea or genetic time bomb? Nat Rev Genet 2005;6(3):235–40. [59] Park BL, Kim LH, Shin HD, Park YW, Uhm WS, Bae S-C. Association analyses of DNA methyltransferase-1 (DNMT1) polymorphisms with systemic lupus erythematosus. J Human Genet 2004;49(11):642–6. [60] Hansen RS, Wijmenga C, Luo P, Stanek AM, Canfield TK, Weemaes CM, et al. The DNMT3B DNA methyltransferase gene is mutated in the ICF immunodeficiency syndrome. Proc Natl Acad Sci USA 1999;96(25):14412–7. [61] Montgomery KG, Liu MCP, Eccles DM, Campbell IG. The DNMT3B C>T promoter polymorphism and risk of breast cancer in a British population: a case-control study. Breast Can Res 2004;6(4):R390–4. [62] Lee SJ, Jeon H-S, Jang J-S, Park SH, Lee GY, Lee B-H, et al. DNMT3B polymorphisms and risk of primary lung cancer. Carcinogenesis 2005;26(2):403–9. [63] Kelemen LE, Sellers TA, Schildkraut JM, Cunningham JM, Vierkant RA, Pankratz VS, et al. Genetic variation in the onecarbon transfer pathway and ovarian cancer risk. Cancer Res 2008;68(7):2498–506. [64] Cebrian A, Pharoah PD, Ahmed S, Ropero S, Fraga MF, Smith PL, et al. Genetic variants in epigenetic genes and breast cancer risk. Carcinogenesis 2006;27(8):1661–9. [65] Walter MJ, Ding L, Shen D, Shao J, Grillot M, McLellan M, et al. Recurrent DNMT3A mutations in patients with myelodysplastic syndromes. Leukemia 2011;25(7):1153–8. [66] Ley TJ, Ding L, Walter MJ, McLellan MD, Lamprecht T, Larson DE, et al. DNMT3A mutations in acute myeloid leukemia. N Engl J Med 2010;363(25):2424–33. [67] Kim Y-I. Folate and DNA methylation: a mechanistic link between folate deficiency and colorectal cancer? Cancer Epidemiol Biomarkers Prevent 2004;13(4):511–9. [68] Jhaveri MS, Wagner C, Trepel JB. Impact of extracellular folate levels on global gene expression. Mol Pharmacol 2001;60(6):1288–95. [69] Cooney CA, Dave AA, Wolff GL. Maternal methyl supplements in mice affect epigenetic variation and DNA methylation of offspring. J Nutr 2002;132(8 Suppl.):2393S–400S. [70] Lillycrop KA, Phillips ES, Jackson AA, Hanson MA, Burdge GC. Dietary protein restriction of pregnant rats induces and folic acid supplementation prevents epigenetic modification of hepatic gene expression in the offspring. J Nutr 2005;135(6):1382–6.

V.  Factors Influencing Epigenetic Changes

304

19.  Metabolic Regulation of DNA Methylation in Mammalian Cells

[71] Cropley JE, Suter CM, Beckman KB, Martin DIK. Germ-line epigenetic modification of the murine Avy allele by nutritional supplementation. Proc Natl Acad Sci USA 2006;103(46):17308–12. [72] Singh SM, Murphy B, O’Reilly R. Epigenetic contributors to the discordance of monozygotic twins. Clin Genet 2002;62(2): 97–103. [73] Jacob RA, Gretz DM, Taylor PC, James SJ, Pogribny IP, Miller BJ, et al. Moderate folate depletion increases plasma homocysteine and decreases lymphocyte DNA methylation in postmenopausal women. J Nutr 1998;128(7):1204–12. [74] Rampersaud GC, Kauwell GP, Hutson AD, Cerda JJ, Bailey LB. Genomic DNA methylation decreases in response to moderate folate depletion in elderly women. Am J Clin Nutr 2000;72(4):998–1003. [75] Pufulete M, Al-Ghnaniem R, Khushal A, Appleby P, Harris N, Gout S, et al. Effect of folic acid supplementation on genomic DNA methylation in patients with colorectal adenoma. Gut 2005;54(5):648–53. [76] Cravo M, Fidalgo P, Pereira AD, Gouveia-Oliveira A, Chaves P, Selhub J, et al. DNA methylation as an intermediate biomarker in colorectal cancer: modulation by folic acid supplementation. Eur J Cancer Prevent 1994;3(6):473–9. [77] Cravo ML, Pinto AG, Chaves P, Cruz JA, Lage P, Nobre Leitão C, et al. Effect of folate supplementation on DNA methylation of rectal mucosa in patients with colonic adenomas: correlation with nutrient intake. Clin Nutr 1998;17(2):45–9. [78] Kim YI, Baik HW, Fawaz K, Knox T, Lee YM, Norton R, et al. Effects of folate supplementation on two provisional molecular markers of colon cancer: a prospective, randomized trial. Am J Gastroenterol 2001;96(1):184–95. [79] Ortbauer M, Ripper D, Fuhrmann T, Lassi M, Auernigg-Haselmaier S, Stiegler C, et al. Folate deficiency and over-supplementation causes impaired folate metabolism: regulation and adaptation mechanisms in Caenorhabditis elegans. Mol Nutr Food Res 2016;60(4):949–56. [80] Niculescu MD, Zeisel SH. Diet, methyl donors and DNA methylation: interactions between dietary folate, methionine and choline. J Nutr 2002;132(8 Suppl.):2333S–5S. [81] Davis CD, Uthus EO. DNA methylation, cancer susceptibility, and nutrient interactions. Exp Biol Med 2004;229(10):988–95. [82] McGowan PO, Meaney MJ, Szyf M. Diet and the epigenetic (re) programming of phenotypic differences in behavior. Brain Res 2008;1237:12–24. [83] Ross SA. Diet and DNA methylation interactions in cancer prevention. Ann NY Acad Sci 2003;983:197–207. [84] Ross SA, Milner JA. Epigenetic modulation and cancer: effect of metabolic syndrome? Am J Clin Nutr 2007;86(3):s872–7. [85] Halsted CH, Villanueva JA, Devlin AM, Chandler CJ. Metabolic interactions of alcohol and folate. J Nutr 2002;132(8 Suppl.):2367S–72S. [86] Mason JB, Choi S-W. Effects of alcohol on folate metabolism: implications for carcinogenesis. Alcohol 2005;35(3):235–41. [87] Choi SW, Stickel F, Baik HW, Kim YI, Seitz HK, Mason JB. Chronic alcohol consumption induces genomic but not p53-specific DNA hypomethylation in rat colon. J Nutr 1999;129(11): 1945–50. [88] Barak AJ, Beckenhauer HC, Tuma DJ. Hepatic transmethylation and blood alcohol levels. Alcohol Alcohol 1991;26(2):125–8. [89] Bleich S, Hillemacher T. Homocysteine, alcoholism and its molecular networks. Pharmacopsychiatry 2009;42(Suppl. 1):S102–9. [90] Bielawski DM, Zaher FM, Svinarich DM, Abel EL. Paternal alcohol exposure affects sperm cytosine methyltransferase messenger RNA levels. Alcohol Clin Exp Res 2002;26(3):347–51. [91] Wallwork JC, Duerre JA. Effect of zinc deficiency on methionine metabolism, methylation reactions and protein synthesis in isolated perfused rat liver. J Nutr 1985;115(2):252–62.

[92] Luo J, Li YN, Wang F, Zhang WM, Geng X. S-adenosylmethionine inhibits the growth of cancer cells by reversing the hypomethylation status of c-myc and H-ras in human gastric cancer and colon cancer. Int J Biol Sci 2010;6(7):784–95. [93] Uthus EO, Ross SA, Davis CD. Differential effects of dietary selenium (se) and folate on methyl metabolism in liver and colon of rats. Biol Trace Element Res 2006;109(3):201–14. [94] Davis CD, Uthus EO, Finley JW. Dietary selenium and arsenic affect DNA methylation in vitro in Caco-2 cells and in vivo in rat liver and colon. J Nutr 2000;130(12):2903–9. [95] Finley JW, Davis CD, Feng Y. Selenium from high selenium broccoli protects rats from colon cancer. J Nutr 2000;130(9):2384–9. [96] Mason JB, Cole BF, Baron JA, Kim YI, Smith AD. Folic acid fortification and cancer risk. Lancet 2008;371(9621):1335. [97] Ulrich CM. Folate and cancer prevention: a closer look at a complex picture. Am J Cin Nutr 2007;86(2):271–3. [98] Ulrich CM, Potter JD. Folate and cancer—timing is everything. JAMA 2007;297(21):2408–9. [99] Kim YI. Folate: a magic bullet or a double edged sword for colorectal cancer prevention? Gut 2006;55(10):1387–9. [100] Mato JM, Cámara J, Fernández de Paz J, Caballería L, Coll S, Caballero A, et al. S-adenosylmethionine in alcoholic liver cirrhosis: a randomized, placebo-controlled, double-blind, multicenter clinical trial. J Hepatol 1999;30(6):1081–9. [101] Pascale RM, Marras V, Simile MM, Daino L, Pinna G, Bennati S, et al. Chemoprevention of rat liver carcinogenesis by S-adenosyl-l-methionine: a long-term study. Cancer Res 1992;52(18):4979–86. [102] Lu SC, Ramani K, Ou X, Lin M, Yu V, Ko K, et al. S-adenosylmethionine in the chemoprevention and treatment of hepatocellular carcinoma in a rat model. Hepatology 2009;50(2):462–71. [103] Pravenec M, Kozich V, Krijt J, Sokolova J, Zidek V, Landa V, et al. Folate deficiency is associated with oxidative stress, increased blood pressure, and insulin resistance in spontaneously hypertensive rats. Am J Hypertens 2013;26(1):135–40. [104] Shenoy V, Mehendale V, Prabhu K, Shetty R, Rao P. Correlation of serum homocysteine levels with the severity of coronary artery disease. Indian J Clin Biochem 2014;29(3):339–44. [105] De Wals P, Tairou F, Van Allen MI, Uh SH, Lowry RB, Sibbald B, et al. Reduction in neural-tube defects after folic acid fortification in Canada. N Engl J Med 2007;357(2):135–42. [106] Boulet SL, Yang Q, Mai C, Kirby RS, Collins JS, Robbins JM, et al. Trends in the postfortification prevalence of spina bifida and anencephaly in the United States. Birth Defects Res A 2008;82(7):527–32. [107] Chang H, Zhang T, Zhang Z, Bao R, Fu C, Wang Z, et al. Tissuespecific distribution of aberrant DNA methylation associated with maternal low-folate status in human neural tube defects. J Nutr Biochem 2011;22(12):1172–7. [108] Ichi S, Costa FF, Bischof JM, Nakazaki H, Shen YW, Boshnjaku V, et al. Folic acid remodels chromatin on Hes1 and Neurog2 promoters during caudal neural tube development. J Biol Chem 2010;285(47):36922–32. [109] Ba Y, Yu H, Liu F, Geng X, Zhu C, Zhu Q, et al. Relationship of folate, vitamin B12 and methylation of insulin-like growth factor-II in maternal and cord blood. Eur J Clin Nutr 2011;65(4):480–5. [110] Refsum H, Yajnik CS, Gadkari M, Schneede J, Vollset SE, Orning L, et al. Hyperhomocysteinemia and elevated methylmalonic acid indicate a high prevalence of cobalamin deficiency in Asian Indians. Am J Clin Nutr 2001;74(2):233–41. [111] Fernandez-Roig S, Cavalle-Busquets P, Fernandez-Ballart JD, Ballesteros M, Berrocal-Zaragoza MI, Salat-Batlle J, et al. Low folate status enhances pregnancy changes in plasma betaine and dimethylglycine concentrations and the association between betaine and homocysteine. Am J Clin Nutr 2013;97(6):1252–9.

V.  Factors Influencing Epigenetic Changes



REFERENCES

[112] Yajnik CS, Deshpande SS, Jackson AA, Refsum H, Rao S, Fisher DJ, et al. Vitamin B(12) and folate concentrations during pregnancy and insulin resistance in the offspring: the Pune Maternal Nutrition Study. Diabetologia 2008;51(1):29–38. [113] Dabelea D, Hanson RL, Lindsay RS, Pettitt DJ, Imperatore G, Gabir MM, et al. Intrauterine exposure to diabetes conveys risks for type 2 diabetes and obesity: a study of discordant sibships. Diabetes 2000;49(12):2208–11. [114] Krishnaveni GV, Hill JC, Veena SR, Bhat DS, Wills AK, Karat CL, et al. Low plasma vitamin B12 in pregnancy is associated with gestational ‘diabesity’ and later diabetes. Diabetologia 2009;52(11):2350–8. [115] Kiefte-de Jong JC, Timmermans S, Jaddoe VW, Hofman A, Tiemeier H, Steegers EA, et al. High circulating folate and vitamin B-12 concentrations in women during pregnancy are associated with increased prevalence of atopic dermatitis in their offspring. J Nutr 2012;142(4):731–8. [116] Bernal AJ, Jirtle RL. Epigenomic disruption: the effects of early developmental exposures. Birth Defects Res A 2010;88(10):938–44. [117] Ouko LA, Shantikumar K, Knezovich J, Haycock P, Schnugh DJ, Ramsay M. Effect of alcohol consumption on CpG methylation in the differentially methylated regions of H19 and IG-DMR in male gametes: implications for fetal alcohol spectrum disorders. Alcohol Clin Exp Res 2009;33(9):1615–27.

305

[118] Lambrot R, Xu C, Saint-Phar S, Chountalos G, Cohen T, Paquet M, et al. Low paternal dietary folate alters the mouse sperm epigenome and is associated with negative pregnancy outcomes. Nat Commun 2013;4:2889. [119] Blount BC, Mack MM, Wehr CM, MacGregor JT, Hiatt RA, Wang G, et al. Folate deficiency causes uracil misincorporation into human DNA and chromosome breakage: implications for cancer and neuronal damage. Proc Natl Acad Sci USA 1997;94(7): 3290–5. [120] Amarasekera M, Martino D, Ashley S, Harb H, Kesper D, Strickland D, et al. Genome-wide DNA methylation profiling identifies a folate-sensitive region of differential methylation upstream of ZFP57-imprinting regulator in humans. FASEB J 2014;28(9):4068–76. [121] Joubert BR, den Dekker HT, Felix JF, Bohlin J, Ligthart S, Beckett E, et al. Maternal plasma folate impacts differential DNA methylation in an epigenome-wide meta-analysis of newborns. Nat Commun 2016;7:10577. [122] Takiguchi M, Achanzar WE, Qu W, Li G, Waalkes MP. Effects of cadmium on DNA-(cytosine-5) methyltransferase activity and DNA methylation status during cadmium-induced cellular transformation. Exp Cell Res 2003;286(2):355–65.

V.  Factors Influencing Epigenetic Changes

C H A P T E R

20 Dietary and Metabolic Compounds Affecting Covalent Histone Modifications Fermin I. Milagro, J. Alfredo Martinez Centre for Nutrition Research, University of Navarra, Pamplona, Spain; Institute of Health Carlos III, Madrid, Spain

O U T L I N E Introduction

307

Other Histone Modifications

316

Histone Acetylation Histone Deacetylases Histone Acetyltransferases

309 311 313

Conclusions

316

References

319

Histone Methylation

313

INTRODUCTION Eukaryotic DNA is intimately associated with histones, highly basic proteins, forming chromatin. Two different forms of chromatin have been described: heterochromatin is a tightly compacted form of DNA usually associated with transcriptionally silent genomic regions, whereas euchromatin is a lightly packed form of chromatin usually under active transcription. The term chromatin remodeling encompasses a wide variety of changes in chromatin structure, but can be defined as a discernible change in histone–DNA contacts. This molecular mechanism is regulated by epigenetic processes, such as DNA methylation, noncoding RNAs (micro RNAs, long noncoding RNAs and others), Polycombgroup proteins, and covalent histone modifications [1]. Five major families of histones (encoded in humans by 76 different genes) exist: histones H2A, H2B, H3, and H4, which are known as the core histones, and histone H1, which is the linker histone [2]. Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00020-1 Copyright © 2017 Elsevier Inc. All rights reserved.

The fundamental unit of chromatin is the nucleosome, which consists of approximately 147 base pairs of DNA wrapped around a histone octamer containing two H2A– H2B dimers and a H3–H4 tetramer [2]. Chromatin is further compacted by the incorporation of the linker histone H1, which has been reported to have at least eight isoforms in higher eukaryotes [3]. There are also variant histone subspecies that are recognized by differences in their amino acid sequence relative to the major histone species [4]. Each protein has both a histone fold domain, which mediates the histone–histone and histone–DNA interactions that are crucial for the assembly of the nucleosome core particle, and a flexible amino-terminal tail domain, which protrudes from the nucleosome core particle [5]. There are various histone posttranscriptional modifications that decorate the canonical histones (Fig. 20.1). The combination of modifications (marks) produced by specific enzymes has been proposed to constitute a code (histone code) that regulates downstream processes, such as gene transcription, DNA repair, and apoptosis [6,7].

307

308

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

FIGURE 20.1  The histone code, including the main histone modifications in histones H2A, H2B, H3, and H4, and the amino acids affected.

Altering chromatin structure at the level of histone modifications involves the participation of multiple enzymes that induce biochemical changes on precise amino acids in a posttranscriptional manner. In this context, enzymes have been identified for lysine acetylation, lysine, and arginine methylation, serine, tyrosine and threonine phosphorylation, and lysine biotinylation (Fig. 20.2), but also for citrullination, ubiquitination, SUMOylation, ADP-ribosylation, deimination, carbonylation, and proline isomerization [8]. Functions of these enzymes depend on cofactors, such as acetyl CoA, ATP, biotin, NAD, and S-adenosyl methionine (SAM), with the intranuclear levels of them depending on the metabolic, inflammatory, or redox state of the cell (Fig. 20.2). More subtle effects can be exerted by naturally occurring inhibitors, such as short chain fatty acids or nicotinamide. In any case, aging, environmental exposure and lifestyle changes (including the diet) are closely associated with an epigenetic drift [9,10]. There are even transient, reversible circadian epigenetic patterns controlled by chromatin remodeling that are sensitive to environmental cues [11]. Especially important in epigenetic programming is the nutritional status of the mother during pregnancy and lactation, as well as the maternal behavior during interactions with pups [12], being most of these studies performed in animal models. Most modifications are localized to the amino- and carboxy-terminal histone tails, particularly in H3 and H4 histones, which have long tails protruding from the

nucleosome. Only a few of the modifications are localized to the histone globular domains. There are more than 60 different residues on histones where modifications have been detected, either by specific antibodies or by mass spectrometry. Extra complexity partly comes from the fact that methylation at lysines or arginines may be one of three different forms: mono-, di-, or trimethyl for lysines, and mono- or di- (asymmetric or symmetric) for arginines. This vast array of modifications gives enormous potential for functional adaptive responses, but not all these modifications will be on the same histone at the same time. The majority of these histone changes regulate gene transcription, which has been explained by two different mechanisms. The first one proposes that chromatin packaging is directly altered (either by a change in electrostatic charge or through internucleosomal contacts) to open or close the DNA polymer, thus controlling access of DNA-binding proteins, such as transcription factors. The other one postulates that the attached chemical moieties alter the nucleosome surface to promote the association of chromatin binding proteins. The effects of chromatin modifications on transcription regulation of the protein-encoding genome have been broadly classified into repressing and activating (Fig. 20.3). The repressing marks have been associated with gene silencing in heterochromatin, whereas the activating marks have been related to euchromatin. In other words, they correlate with, and perhaps directly regulate, gene repression and induction. In general, histone

V.  Factors Influencing Epigenetic Changes



309

Histone Acetylation

FIGURE 20.2  Main histone modifications, enzymes involved, and donors. Modifications: aK, acetyl lysine; N-me, methyl arginine, or methyl lysine (mono-, di-, or trimethylation); uK, ubiquitinated lysine; pS, phosphorylated serine; bK, biotinylated lysine; AR, ADP-ribosylation. Enzymes and substrates: histone methyltransferases (KMTs) and histone lysine demethylases (KDMs); histone acetyltransferases (HATs) and histone deacetylases (HDACs); SAM, S-adenosyl methionine.

acetylation and phosphorylation act as activators of gene expression, whereas histone deacetylation, biotinylation, and SUMOylation inhibit gene expression. Methylation and ubiquitination can act as repressor or activator of gene expression depending on the histone residue being targeted [2,13,14]. For example, trimethylation of both H3 lysine 4 (H3K4Me3) and H3 lysine 36 (H3K36Me3) are particularly associated with active transcription, whereas trimethylation of H3 lysine 27 (H3K27Me3) and H4 lysine 20 (H4K20Me3), or di and trimethylation of H3 lysine 9 (H3K9Me2/3) are particularly associated with repressed genes. The involvement of histone modifications in transcription factor binding processes provides a potentially enormous repertoire of signals. In fact, more than 100 possible chemical modifications of histones are known (acetylation, monomethylation, dimethylation, trimethylation, etc.), performed by different families of modifying and demodifying enzymes. Many of these enzymes require common metabolites as substrates or cofactors for regulating the “histone code” and the way the genes are regulated. Furthermore, changes in cellular

metabolic compounds can affect enzyme functions, acting as substrates or inhibitors of the modification of the N-terminal tail domains of the core histones. Cellular metabolic compounds can be affected in general by diet and nutrients, metabolic status of the body (hypoxia, hyperglycemia, redox status, inflammation, oxidative stress), and also by endocrine unbalances and diseases that, in turn, could alter mRNA and protein levels of histone-modifying enzymes (Fig. 20.4). This chapter is focused on the effect of different metabolic and dietary compounds, as well as the metabolic state, on relevant posttranslational histone modifications, with emphasis on acetylation and methylation of histones H3 and H4 as these modifications are thought to correlate best with gene activation and repression.

HISTONE ACETYLATION Histones are targets for acetylation and deacetylation, especially on lysine residues in the N-terminal tail. This regulatory mechanism is catalyzed by two types of

V.  Factors Influencing Epigenetic Changes

310

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

FIGURE 20.3  Histone modifications involved in the repression of gene transcription (heterochromatin, left) or gene transcription activation (euchromatin, right).

FIGURE 20.4  Influence of dietary factors, metabolic compounds and diseases on the activity of histone-modifying enzymes, the histone code and gene transcription regulation.

V.  Factors Influencing Epigenetic Changes



Histone Acetylation

enzyme, histone acetyltransferases (HATs) and histone deacetylases (HDACs), which not only act on histone substrates but also on nonhistone proteins, such as scaffold proteins, DNA polymerase, heterochromatin protein 1 (HP1) and Polycomb-group proteins. In the acetylation process, acetyl-coenzyme A is the donor of the acetyl group, whereas this group is transferred to coenzyme A [15]. In fact, nuclear acetyl-CoA synthesis is a rate limiting step for histone acetylation. Since acetyl-CoA metabolism is directly linked to chromatin regulation, it may affect diverse cellular processes where acetylation and metabolism intersect, such as disease states and aging [16]. As a result, fluctuations in the concentration of acetyl-CoA affect acetylation of proteins. However, it remains to be elucidated if all HATs and all acetylation sites are affected by acetyl-CoA fluctuations to the same extent, or whether the acetylation status of different lysine residues differs in responsiveness to acetylCoA fluctuations. Acetyl-CoA is produced by glycolysis and other catabolic pathways, such as the beta-oxidation of fatty acids, and it has an important regulatory role as is used as a substrate for the citric acid cycle and as a precursor in the synthesis of fatty acids and steroids (and in other anabolic pathways). Many stimuli can activate histone acetylation (Fig. 20.5), including inflammation (i.e., IL-1β, TNFα, and LPS), oxidative stress, UV light, bacteria, viruses, cigarette smoke [17], aging [18], hypoxia [19], cell adhesion-mediated signaling [20], biotin deficiency [21], quercetin [22], caffeine [23],

311

or glucose availability [24]. In relation to inflammation, dietary polyphenols, such as resveratrol, curcumin, and catechins, are shown to modulate NF-kB activation and chromatin remodeling through modulation of SIRT1 and HDAC2 activity [25]. Indeed, histone acetylation in mammalian cells is dependent on ATP-citrate lyase, the enzyme that converts glucose-derived citrate into acetyl CoA [26]. Regarding hormones, estrogens (i.e., 17betaE2) attenuate H3 and H4 histone acetylation [27]. Finally, different metals are involved in histone acetylation, such as zinc, nickel, arsenite, chromium, and copper [28].

Histone Deacetylases Histone deacetylases (HDACs) are specific deacetylases that remove the acetyl groups of lysine residues of histone tails leading to chromatin compaction and transcriptional repression [17,29,30]. HDACs are categorized into five main classes based on their sequence homology and expression patterns. Class I, IIA, IIB, and IV HDACs are Zn-dependent deacetylases, whereas the Class III HDACs, also called sirtuins in mammals, are NADdependent deacetylases [31]. Different HDAC inhibitors have been clinically used for cancer therapy and psychiatry, and proposed for the treatment of metabolic diseases, such as stroke and cardiovascular diseases, asthma, arthritis, type 2 diabetes, and neurodegenerative processes [30,32,33]. These inhibitors include hydroxamates, such as trichostatin A

FIGURE 20.5  Dietary and metabolic factors involved in histone acetylation and deacetylation.

V.  Factors Influencing Epigenetic Changes

312

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

(TSA) and vorinostat (SAHA); short chain fatty acids, such as butyrate and valproic acid; cyclic tetrapeptides, such as apicidin or depsipeptide, benzamides, epoxides, amides, ketones, and lactones [29,30,32]. Different dietary agents, such as biotin, lipoic acid, garlic organosulfur compounds (i.e., diallyl disulfide), and vitamin E metabolites show structural features compatible with HDAC inhibition [34]. For example, diallyl disulfide is an organosulfur compound found in garlic that increases histone H3 and H4 acetylation and p21waf1/cip1 expression in human colon tumor cell lines [35] and, like allyl mercaptan and butanethiol, in liver and Morris hepatoma cells [36]. Sulforaphane (SFN) is an isothiocyanate found in cruciferous vegetables that weakly inhibits HDAC activity in human cell lines while suppressing tumorigenesis in Apc-minus mice by increasing acetylated histone H3 associated with the promoter region of bax and P21 genes [37]. Isothiocyanates with a similar spacer length to that of sulforaphane, such as sulforaphene, erucin, and phenylbutyl isothiocyanate (all found in plants form the Brassicaceae family), exhibit comparable HDAC inhibitory activities. Even compounds with a longer or shorter spacer length, such as sulforaphane [6,9], 6-erucin and phenethyl isothiocyanate, display a similar effect [38]. Likewise, other natural organoselenium compounds found in yeasts (methylselenocysteine and selenomethionine), in addition to targeting redox-sensitive signaling proteins and transcription factors, also act as HDAC inhibitors with potential anticancer properties in human prostate cancer cells [39]. The isoflavone genistein, with antiestrogenic activities and abundant in soybeans, is able to inhibit HDAC6, which is the main cytoplasmic deacetylase in mammalian cells, downregulating thus androgen receptor [40]. Other plant compounds with effects on the histone acetylation process are flavones from Acca sellowiana [41], isoliquiritigenin (from liquorice), dihydrocoumarin (from sweet clover), quercetin (abundant in onions, citrus rind, garlic, and the Brassica genus) and psammaplin A (from a marine sponge) [5]. Concerning green tea polyphenols, an activatory role of green tea’s theophylline on HDAC activity has been reported in epithelial cells and macrophages [42]; however, a mix of epicatechin monomers from green tea has been reported to inhibit the growth of melanoma cell lines through inhibition of class I HDACs resulting in a decrease in the levels of cyclins and cyclin-dependent kinases of G1 phase of cell cycle and an upregulation of tumor suppressor proteins (Cip1/WAF1/p21, p16, and p53) [43]. Some of these compounds and their chemical structural formulae have been perfectly revised by Meeran et al. [44]. However, despite the ability of these substances to induce epigenetic changes, it has to be clarified whether they should be considered as real epigenetic modifiers [45]. And, in some cases, one bioactive compounds can

act through several epigenetic mechanisms; this is the case of genistein, which affects DNA methylation, histone modifications controlling chromatin accessibility (i.e., upregulating HATs, including HAT1, p300, PCAF, CBP, and downregulating HDAC6 and SIRT1), and noncoding RNAs (reviewed by Ref. [46]). Regarding metabolism, oxidative stress inhibits HDAC activity [47,48]. On the other hand, it has been suggested that, in cancer, HDAC inhibitors and sirtuin inhibitors induce apoptosis through oxidative stress and DNA damage mechanisms in addition to direct activation of apoptosis-inducing genes [49]. Function of Class IIa HDAC, at least in mouse models, is regulated by neuronal stimuli, immune system, physical exercise, and perhaps fasting. Smoking alters the expression of Class I/II HDACs and a high-salt diet induces the expression of a kinase that can block the functioning of Class IIa HDACs in rats [50]. Finally, as all the above HDACs are zinc-dependent hydrolases [51], zinc levels in the diet could affect their activity. Concerning Class III HDACs or sirtuins, resveratrol is a polyphenolic activator of SIRT1 found in red wine and different vegetable products that shows neuroprotective, antioxidant and antiaging effects [52]. However, as occurs with other bioactive compounds, resveratrol binds to many diverse proteins (more than 20) and is therefore often difficult to separate epigenetic effects from effects on other cellular structures [53]. As reviewed by Britton et al [53], resveratrol can act through many pathways including antiinflmmatory mechanisms, cell cycle, and programmed cell death pathways, calorie restriction mimetics via antioxidant and prooxidant properties, effects on mTOR signaling, miRNA modulation, AMP-activated protein kinase (AMPK) activation, NF-kB activation, nitric oxide production, cytochrome P450 inhibition, and HDAC inhibition. Other SIRT1-activating compounds are the plant polyphenols butein (found in Toxicodendron vernicifluum, Dahlia, Butea monosperma, and Coreopsis) and fisetin (a flavonoid abundant in strawberries, apples, persimmons, onions, and cucumbers) [54], the resveratrol metabolite piceatannol (found in passion fruit, grapes, and Japanese knotweed), the liquorice flavonoid isoliquiritigenin, and the flavonoid luteolin, found in many plants including celery, broccoli, carrots, green peppers, oranges, perilla, and olive oil [55]. Similarly, calorie restriction upregulates SIRT1 in human mononuclear cells [56], fat, muscle, and liver [57], contributing thus to extend lifespan and ameliorate insulin resistance. However, as sirtuins need to bind to NAD+, low levels of NAD+ inhibit their activity [58]; this is the reason why high nicotinamide concentrations (the amide of nicotinic acid, vitamin B3), which act by interfering unspecifically with NAD+dependent reactions, can act as a sirtuin inhibitor [52]. Finally, high glucose levels have been also reported to

V.  Factors Influencing Epigenetic Changes



313

Histone Methylation

act as a sirtuin inhibitor, accelerating endothelial progenitor cells senescence via repression of FoxO1 transcriptional activity [59].

Histone Acetyltransferases Histone acetyltransferases (HATs) are acetyltransferases that acetylate lysines within lysine-rich aminoterminal tails of histone proteins, resulting in charge neutralization and a more relaxed, open, and transcriptionally active chromatin structure. They transfer an acetyl group from acetyl CoA to form ε-N-acetyllysine. HATs are broadly classified into two different classes, based on their functional localization: Type A HATs are located in the nucleus and contain a bromodomain that helps them recognize and bind to acetylated lysine residues, whereas type B HATs are cytoplasmic and modify the newly synthesized histones before their assembly into nucleosomes [60]. HAT inhibitors seem promising for the treatment of human diseases, such as Alzheimer’s disease, hyperlipidemias, and diabetes [61]. Despite that, few natural inhibitors of HATs have been reported. Anacardic acid, a phenolic lipid isolated from the shell of the cashew nut (Anacardium occidentale), has been identified as a potent noncompetitive inhibitor of both p300 and PCAF HAT activity in vitro [62]. A polyprenylated benzophenone known as garcinol, isolated from Garcinia indica fruit rind, has been also identified as an inhibitor of p300 and PCAF HAT activity [63]. The polyphenol curcumin is a p300/CREB-binding protein-specific inhibitor of acetyltransferase that exhibits a variety of pharmacological effects including antitumor, antiinflammatory, and antiinfectious activities [64,65]. The catechin epigallocatechin-3-gallate, abundant in tea and with potent antioxidant properties, acts also as a histone acetyltransferase inhibitor [66]. Although not natural products, several isothiazolone-based HAT inhibitors have shown inhibitory effects on PCAF and p300 [64], in a similar way to some alpha-methylene butyrolactones, benzylidene acetones, and alkylidene malonates [67]. Finally, there are different synthetic HAT inhibitors that are specific for p300 (i.e., lysyl CoA) and for PCAF (i.e., H3-CoA-20) [68]. Concerning HAT activators, glucose induces overall acetylation of H3K9, K18, and K27 and H4K5, K8, and K12 in yeasts, probably by regulating HAT activity [69]. Retinoic acid, an oxidized form of vitamin A, is also able to acetylate H3K9 [70]. Ethanol has been reported to acetylate H3K9 in rat hepatocytes, as well as other liver nuclear and nonnuclear proteins, which probably contributes to alcohol-induced hepatotoxicity [71]. Surrogate alcohols, such as 1-propanol, 1-butanol, and isopentanol, modulate H3K9 via increasing HAT activity [72]. A high carbohydrate/fat ratio diet induces H3K9

acetylation on the SGLT1 gene and its expression in the jejunum [73]. Other physiological factors that could participate in the acetylation process are exposure to cold [74] and aging [75,76], at least for H4K16, whose acetylation plays a critical role in lifespan regulation, transcription activation, and protein interactions [77,78]. Chronic hypoxia induces H3K9 acetylation in prostate cancer cells, with several histone acetyltransferase enzymes (i.e., CBP and p300) being components of the HIF-1α complex (that is responsible for many hypoxia-induced effects, such as glucose use and uptake, angiogenesis, inflammation, cell survival, tumor growth and apoptosis resistance), therefore potentiating the HIF-1α gene expression response to hypoxia [79]. Finally, apart from HATs and HDACs, other posttranslational modifications regulate cellular acetylation. For example, phosphorylation has been found to activate HAT function and repress HDACs and methylation has also been reported to control HAT function [60].

HISTONE METHYLATION Methylation of lysine or arginine residues can occur in several modification states due to specific histone-lysine N-methyltransferases and histone-arginine N-methyltransferases. Lysine residues can house either mono-, di-, or trimethyl moieties on their amine group, whereas arginine residues can carry mono- or dimethyl groups on their guanidinyl group [80,81]. The dimethyl arginine state is further defined by whether the modification exists in the symmetric (me2s) or the asymmetric (me2a) configuration [80]. These modifications in histone methylation states can have different and profound implications for the function of chromatin, conforming per se an internal cryptogram within the histone code. Histone arginine methylation is more dynamic, positively correlating with gene activation, whereas its loss from target arginines in H3 and H4 correlates with gene inactivation [82]. In contrast, lysine methylation seems to be a more stable mark, but with a more complicated readout [83]. For example, although methylation of H3K4 and H3K79 correlates with gene activation, methylation of H3K27 correlates with repression [82]. Changes at any particular amino acid are determined by the dynamic equilibrium between the activities of modifying and demodifying enzymes. Although histone methylation was discovered in the 1960s, histone methyltransferases or HMTs (including lysine methyltransferases or KMTs, and arginine-specific methyltransferases or PRMTs) and histone lysine demethylases (KDMs) have only recently been identified [81]. The most thoroughly studied lysine methylation marks in histones are found in H3K4, K9, K27, K36, K79, and

V.  Factors Influencing Epigenetic Changes

314

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

H4K20. In general, H3K4, K36, and K79 methylation are localized near active or poised transcriptional units, whereas H3K9 and H4K20 modifications are hallmarks of silenced or heterochromatic regions. There are six families of KDMs (KDM1, 2, 3, 4, 5, and 6, which act on different substrates and play different roles in cellular function), which require either flavin adenine dinucleotide (FAD) or α-ketoglutarate, oxygen and Fe2+, depending on whether they act on mono- and dimethyl lysine (e.g., KDME1/LSD1) or trimethyl lysine (e.g., jumonji domain enzymes) [84]. On the other hand, SAM serves as a cofactor and methyl donor group for KMTs. Histone arginine methylation occurs on H3R2, R8, R17, and R26 and H4R3. As lysine methylation, it can be either activatory or repressive for transcription, and the enzymes (PRMTs) are recruited to promoters by transcription factors [80]. The arginine demethylation process has recently been described as performed by deiminases and jumonji-group demethylases [85]. For example, a member of the JmjC family (JMJD6) has been shown to apparently reverse histone methylation. As different authors have described [86–89], epigenetic mechanisms could be directly controlled by metabolic and dietary compounds, metabolic state, endocrine unbalances, and diseases (Fig. 20.4). Deepening into histone methylation, these studies postulated early on possible effects of one-carbon and redox metabolism on methyltransferases action in addition to effects of oxygen on demethylases. Both, DNA and histone methyltransferases (KMT and PRMTs), all use SAM as methyl donor [88], which can directly induce epigenetic marks. The bioavailability of SAM is directly influenced by the diet since SAM is formed from methyl groups derived from choline, methionine, or methyl-tetrahydrofolate. Betaine, choline, methionine, zinc, and folate are necessary for homocysteine conversion into methionine [1]. These nutrients interact among them to alter epigenetic DNA and histone methylation processes. In fact, the perturbation of the metabolism of one of the methyl donors results in compensatory changes in the other methyl donors because of the intermingling of these metabolic pathways [90]. Several examples have been published in the literature showing a direct effect of metabolic compounds on methyltransferase activity. For example, Clarke et al. [91] described methyltransferases as being inhibited by the cellular ratio of SAM and S-adenosyl-homocysteine (SAH), especially by the inhibitory capacity of the latter. More recently, it has been observed that a methyldeficient diet reduces protein and mRNA levels of some lysine and arginine methyltransferases in mice [91–94]. Also, a diet low in methionine and cysteine increases histone H1 relative to the total content of histones in rat liver [95]. On the other hand, gestational choline supplementation [96] and the in vitro experiments carried

out by Ara et al. [97] increase lysine methyltransferase mRNA levels. Finally, it is interesting to remark the putative clinical relevance of asymmetric dimethylarginine, a product of PRMTs, in cardiovascular diseases [98,99]. Sometimes, changes in protein/mRNA levels of histone modifying enzymes are not experimentally related to changes in posttranslational modifications of histones. In this context, changes in the levels of methylated histones (lysines and arginines), in a global manner or on a specific locus, are more relevant (Table 20.1) in order to discriminate the direct effect of a metabolic compound. Changes in the methylation levels in lysine and arginine residues of histones 3 and 4 have been described as a result of changes in methyl donors in the diet. For example, the levels of H3K9me2 and H3K27me3, tags of transcriptionally repressed chromatin, were upregulated by choline supplementation, whereas the levels of H3K4me2, associated with active promoters, were highest in choline-deficient rats [96]. A methyl-deficient diet increases considerably the level of histone macroH2A and H3K9me3 in mice, and reduces the levels of H3K27me3, H3K4me2, and H4K20me3, being sometimes related to changes in enzymes or not [94,100]. Similarly, methyldeficient diet leads to progressive loss of H4K20me3, H3K9me2, and H3K9me3 [100,101]. In this sense, actions of choline on human cognition (both during gestation or adulthood) have been suggested to be mediated by epigenomic mechanisms, including changes in the methylation patterns of H3K4, H3K9, and H3K29, and the expression of DNA (Dnmt1, Dnmt3a) and histone (G9a/Ehmt2/Kmt1c, Suv39h1/Kmt1a) methyltransferases [102]. Food restriction and/or protein restriction increases H3K27me3, H3K9me3, and H3K9me1, and reduces H3K4me3, being related to changes in glucocorticoid receptor and IGf2 mRNA gene expression [103,104]. On the other hand, high-fat diets have been found to alter the methylation status of histones in animal models. For example, Suter et al. [105] reported that postweaning exposure to this dietary pattern decreased the levels of H3K4me3 in monkeys, whereas in utero exposure to a maternal high-fat diet increased those of H3K9me3 in mice [106]. Interestingly, most of the JmjC domain-containing proteins are upregulated by hypoxia [107], and hypoxia induces an increase in methylated histones: Johnson et al. [108] reported that hypoxia, in hepatic cells, increased H3K14ac, decreased H3K9ac and increased H4R3me2, H3K4me2, H3K4me3, and H3K79me2, most of which generally correlate with transcriptional activation. On the other hand, the same study reported that hypoxia increased other histone modifications correlated with repression of transcription, such as H3K27me3 and H3K4me1. Xia et al. [109] confirmed that that histone demethylation decreased as oxygen levels dropped, increasing particularly overall methylation of histone

V.  Factors Influencing Epigenetic Changes



315

Histone Methylation

TABLE 20.1 Examples of Experimental Changes in the Level of Histone Methylations (Lysines, K and Arginines, R) by Different Metabolic Models Effect

Residue/methylation

Experimental model

References

Increase

H3K4me1

Insulin via ROS, high glucose, choline deficiency

[96,114,117]

H3K4me2

Methyl deficiency, hypoxia, hexavalent chromium, aging, glucose

[100,108,109,121,127]

H3K4me3

Hexavalent chromium, arsenite, nickel, LPS, aging, hypoxia, glucose, ischemia, TNFα

[97,108,109,121,122,116,125,126,127]

H3K9me1

Protein restriction, nickel, insulin via ROS with hyperglycemia, glucose deprivation, gestational choline supply, hypoxia, chromium, high glucose

[96,103,109,113,114, 121,123,119]

H3K9me3

Food restriction, methyl deficient model, hexavalent chromium, maternal high-fat diet

[94,101,104,106,110]

H3K27me3

Food restriction, hypoxia, gestational choline supply

[96,103,108]

H3K36me3

Hypoxia

[109]

H3K79me2

Hypoxia

[108]

H4R3me2

Hypoxia

[108]

H3K4me1

Insulin via ROS with hyperglycemia

[114]

H3K4me2

High glucose, glucose induced cAMP, diabetic state

[116,119,120]

H3K4me3

Food restriction, chromium, SAM, LPS, high-fat diet

[97,104,105,109]

H3K9me1

Insulin via ROS

[114]

H3K9me2

Protein restriction, choline/methionine restriction, high glucose

[100,103,117,119]

H3K9me3

db/db mice, high glucose

[117,118]

H3K27me3

Hypoxia, hexavalent chromium, methyl deficient diet

[94,108,121]

H3R17

Insulin

[115]

H3R2me2

Hexavalent chromium

[121]

H4K20me3

Food restriction, methyl deficient diet, glucoseinduced cAMP

[94,101,104,116]

Reduction

H3 lysines 4, 9, and 36. The complex interaction between hypoxia and histone methylation has been comprehensively reviewed by Hancock et al. [110]; basically, hypoxia induces histone methylation in all circumstances except for H3K79me2, H3K27me3, and H3K9me2, that depend on cell line and oxygen partial pressure. On the other hand, demethylase levels are regulated by oxygen partial pressure [109,111,112]. These findings suggest that an overexpression of jumonji histone demethylases helps to compensate for decreased levels of molecular oxygen in maintaining H3K4 methylation as a target of HIF-1 transactivation [109]. Also, glucose deprivation induces a higher NAD+/ NAD ratio that is related to a redox-induced increase in the histone methyltransferase Suv39H1 and in H3K9me2 [113]. Finally, results from Kabra et al. [114] have shown that insulin, via ROS, increases H3K4me1 and H3K9me1, demonstrating that theoretical previous studies were

right concerning the control of epigenetic mechanisms by redox status. On the other hand, insulin reduces the methylation levels at H3R17 in relation to a downregulation of PEPCK and G6Pase [115]. Glucose and cAMP produce opposite effects on the methylation status of histone H3 associated with the L-PK promoter and coding regions [116]. The sustained upregulation of the NF-kB-p65 gene as a result of ambient or prior hyperglycemia has been associated with increased H3K4me1 and the suppression of H3K9me2 and H3K9me3 methylation on the p65 promoter [117]. Remarkably, diabetic state induces a decrease of H3K9me3 in vascular smooth muscle cells from db/db mice [118] and a decrease of H3K9me2 and an increase of H3K4me2 in human peripheral blood cells [119]. In this context, the same group [120] described an epigenetic role for the increase of histone H3K9me2 in the lymphocytes of type 1 diabetic patients.

V.  Factors Influencing Epigenetic Changes

316

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

These results open the door to the possibility of the control of histone modifications by other metabolic compounds. Different minerals, such as arsenite, nickel, selenite, or chromium, also induce changes in histone methylations (H3K4, H3K9, and H3K27), although most of the studies are related to toxicological and teratogenic effects and more research is needed to establish whether the toxic effects are mainly mediated by histone methylation changes [121–123]. In a similar way, high doses of some polyphenols, such as genistein [124], have been related to histone methylation. Inflammation, as a complex biological response to harmful stimuli like pathogens, damaged cells or irritants, has recently been related to the onset and complications of different diseases including obesity and cardiovascular diseases. There is evidence regarding the epigenetic control of inflammation. For example, LPS increases the binding of H3K4me3 and histone methyltransferases Set1 and myeloid/lymphoid leukemia to iNOS and TNFα gene promoters [97]. TNFα induces the recruitment of p65, p300, and SET7/9 (H3K4 methyltransferase) at the MCP-1 and TNFα promoters, along with increased H3K4me levels (especially trimethylation) [125]. In this sense, a recent study reports that ischemic renal injury activates proinflammatory genes and progressive elevations of H3K4me3, H2A.Z, and SET1 methyltransferase in mice, suggesting that inflammation could be a trigger of methyl-­histone modifications [126]. Epigenetic and transcriptional mechanisms also contribute to dysregulated inflammatory and autoimmune responses associated with aging, especially in changes associated to H3K4me2 and H3K4me3 [127].

OTHER HISTONE MODIFICATIONS Other important posttranslational histone modifications are led by enzymatic reactions. For instance, phosphorylation, usually an activator of gene expression, can be regulated by different dietary compounds, including genistein [128], sulforaphane [129], or the fungal estrogenic contaminant of moldy feeds, zearalenone [130]. Other histone modifications influenced by dietary compounds and metabolic processes are carbonylation, reduced by aging and calorie restriction [131], biotinylation, influenced by dietary biotin levels [132], and ubiquitination, regulated by nickel [133]. Finally, other less known histone modifications (as they have not been comprehensively studied) are ADP-ribosylation [134], SUMOylation [135], deimination or citrullination [136], proline isomerization [137], and lysine propionylation and butyrylation [138]. Although they have been less exhaustively analyzed, some metabolic and dietary activators and inhibitors have been

described (Table 20.2), which is new promising targets for studies concerning histone modifications by the metabolic state.

CONCLUSIONS It is well established that epigenetic processes allow plasticity of phenotype in a fixed genotype [139]. For example, only environmental (including nutritional) factors are able to explain the phenotypical and epigenetic differences reported in monozygotic twins, which increase over the years [9]. These altered epigenetic marks are implicated in the etiology of several diseases including cancer and chronic metabolic diseases [140]. The metabolic compounds and physiological situations that regulate these enzymatic modifications are being thoroughly studied in order to understand the intrinsic mechanisms that are involved in histone action in human diseases (Table 20.3). Although research in nutritional epigenomics is just beginning, the number of histone modifications whose transcriptional effects are known is increasing every year. In summary, the coming years will see more research in the field of epigenetic mechanisms that influence human health and disease susceptibility. Molecules generated by intermediary metabolism, diet and other lifestyle factors are important regulators of some of these processes. In this sense, a diet rich in phytogenic compounds, such as resveratrol, sulforaphane, diallyl disulfide, genistein, curcumin, and epicatechins, can positively modulate the epigenome and lead to health benefits [141]. Also, although there is not complete consensus, it has been suggested [142] that the reduction of calories, proteins and glucose, and the modulation of methyl donors in the diet, could reduce the incidence of cancer, neurodegenerative disorders and metabolic diseases, acting at least in part by regulating gene expression through histone tail posttranscriptional modifications. The knowledge of the doses, combinations and periods of life in which these dietary modifications occur is of special importance and must be elucidated over the next years, especially in relation to gestation and perinatal age. Some of this information is revised in Chapter 27 of this book. In summary, further advances in the therapeutic applications of these epigenetic mechanisms are foreseen in the coming years, not only in the form of dietary and metabolic advices, but also with the development of new drugs. For example, many epigenetic drugs are being explored for treating oncologic, neurodegenerative, immunologic, metabolic, inflammatory and cardiovascular disorders, including HDAC inhibitors (particularly selective HDAC6 and sirtuin inhibitors), but also DNA methylation inhibiting drugs, protein methyltransferase inhibitors, histone methylation inhibitors, bromodomain inhibitors, and

V.  Factors Influencing Epigenetic Changes



TABLE 20.2 Inductors of the Different Histone Modifications, Enzymes Implicated, and Amino Acidic Residues Affected ADP-Ribosylation

SUMOylation

Ubiquitination

Deimination

Biotinylation

Proline ­Isomerization

Serine kinases

Poly-ADPribose polymerase

Sumo-protein ligases

Ubiquitin ligase

Peptidylarginine deiminases (PAD 1–6)

Biotinidase

Cyclophilins

Serine phosphatases

Mono-ADP-ribosyltransferases

Holocarboxylase synthetase

FK506-binding proteins

ADP-ribosyl cyclases

Pin1

Sirtuin SirT4 Amino acids

Serine, threonine, tyrosine

Arginine, tryptophan, glutamic acid, phosphoserine

Inductors

DNA breaks

DNA breaks

Ginkgolic acid

Oxidative stress

Oxidative stress

Anacardic acid

Bisphenol A

Nicotinamide (niacin)

Kerriamycin B

Lysine

Arginine

Lysine

Proline

PYR-41

Inflammation

Biotin

Juglone

Oxidative stress

Oxidative stress

Conclusions

V.  Factors Influencing Epigenetic Changes

Enzymes

Phosphorylation

Cell proliferation

+

Benzene

NAD

UVA

Genotoxics

Hypoxia Butyrate (inhibitor)

317

318

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

TABLE 20.3 Examples of Different Metabolic and Dietary Conditions, as Well as Several Plant Compounds, in the Level of Histone Methylation and Acetylation and the Activities of Related Enzymes

Metabolism

Diet

Phytogenic compounds

Effector

Effect

Insulin

Increases H3K4me1 and H3K9me1 reduces H3R17

High glucose

Inhibits sirtuins and induce acetylation increases H3K4me2 decreases H3K9me2 and H3K9me3

Glucose deprivation

Increases Suv39H1 and H3K9me2 increases demethylase activities

Diabetes

Decreases H3K9me3 and H3K4me2 increases H3K9me2

Hypoxia

Activates histone acetylation (i.e., H3K9) activates JmjC-containing proteins Increases H3K9me2 and G9a

Oxygen

Activates KDMs

Aging

Acetylates H4K16 increases H3K4me2 and H3K4me3

Inflammation

Activates histone acetylation increases H3K4me3 and H3K4 Recruits 65, p300, and SET7/9 Increases H2A.Z and SET1 MT

Oxidative stress

Inhibits HDAC activity

Estrogens

Decreases H3 and H4 acetylation

Cold exposure

Acetylates H4K16

Methyl deficient diet

Downregulates methyltransferases loss of H4K20me3 and H3K9me3 increases macroH2A and H3K9me3 Reduces H3K27me3 and H4K20me3 Increases histone H1 proportion

Choline deficiency

Increases H3K4Me2

Choline supplementation

Upregulates lysine methyltransferase upregulates H3K9Me2, H3K27Me3

Calorie restriction

Upregulates SIRT1

Food/protein restriction

Increases H3K27me3, H3K9me1/me3 reduces H3K4me3

High-fat diet

Increases hepatic H3K14ac and H3K9me3 in liver

High carbohydrate/fat ratio

Acetylates H3K9

Biotin deficiency

Activates histone acetylation

High-salt diet

Inhibits class IIa HDACs

Zinc deficiency

Inhibits class I and II HDACs

Genistein

Activates histone methylation inhibits HDAC6

Diallyl disulfide

Increases H3 and H4 acetylation

Sulforaphane

Inhibits HDAC activity

Theophylline

Increases HDAC activity

Quercetin

Activates histone acetylation

Resveratrol

Activates SIRT1

Luteolin

Activates SIRT1

Butein and fisetin

Activates SIRT1

Anacardic acid

Inhibits HAT activity

Garcinol

Inhibits HAT activity

Curcumin

Inhibits HAT activity

Epigallocatechin-3-gallate

Inhibits HAT activity

Retinoic acid

Acetylates H3K9

Ethanol

Acetylates H3K9

Surrogate alcohols

Acetylates H3K9

V.  Factors Influencing Epigenetic Changes



REFERENCES

HAT inhibitors [143]. However, the many physiological involvements of epigenetic mechanisms and enzymes, as well as the lack of specificity of many of these drugs and dietary compounds, point to possibly many sideeffects that must be carefully evaluated before initiating any therapy.

Glossary Histone acetyltransferase Histone deacetylase Hypoxia-inducible factor Histone methyltransferase Interleukin 1-beta bifunctional arginine demethylase and lysyl-hydroxylase KDM Histone lysine demethylase KDM1A/LSD1 Lysine-specific histone demethylase 1A or lysine (K)specific demethylase 1A KMT Histone lysine methyltransferase LPS Lipopolysaccharide PCAF P300/CBP-associated factor PRMT Protein arginine methyltransferase SAH S-adenosyl-homocysteine SAHA Vorinostat SAM S-adenosyl methionine SFN Sulforaphane SGLT1 Sodium/glucose cotransporter 1 SIRT1 Sirtuin 1 TNFα Tumor necrosis factor alpha TSA Trichostatin A HAT HDAC HIF HMT IL-1β JMJD6

Acknowledgments The authors acknowledge financial support from CIBERobn, Linea Especial about Nutrition, Obesity and Health (University of Navarra LE/97), MINECO (AGL2013-45554-R) and IdiSNA.

References [1] Campion J, Milagro FI, Martinez JA. Individuality and epigenetics in obesity. Obesity Revi 2009;10:383–92. [2] Hake SB, Xiao A, Allis CD. Linking the epigenetic language of covalent histone modifications to cancer. Br J Cancer 2007;96(Sup pl):R31–9. [3] Kimmins S, Sassone-Corsi P. Chromatin remodelling and epigenetic features of germ cells. Nature 2005;434:583–9. [4] Sarma K, Reinberg D. Histone variants meet their match. Nat Rev Mol Cell Biol 2005;6:139–49. [5] Delage B, Dashwood RH. Dietary manipulation of histone structure and function. Ann Rev Nutr 2008;28:347–66. [6] Ballestar E, Esteller M. Epigenetic gene regulation in cancer. Adv Genet 2008;61:247–67. [7] Strahl BD, Allis CD. The language of covalent histone modifications. Nature 2000;403:41–5. [8] Cosgrove MS, Wolberger C. How does the histone code work? Biochem Cell Biol 2005;83:468–76. [9] Fraga MF, Ballestar E, Paz MF, Ropero S, Setien F, Ballestar ML, et al. Epigenetic differences arise during the lifetime of monozygotic twins. Proc Natl Acad Sci USA 2005;102:10604–9. [10] Wang SC, Oelze B, Schumacher A. Age-specific epigenetic drift in late-onset Alzheimer’s disease. PLoS ONE 2008;3:e2698.

319

[11] Gallou-Kabani C, Vige A, Junien C. Lifelong circadian and epigenetic drifts in metabolic syndrome. Epigenetics 2007;2:137–46. [12] Gallou-Kabani C, Junien C. Nutritional epigenomics of metabolic syndrome: new perspective against the epidemic. Diabetes 2005;54:1899–906. [13] Ito T. Role of histone modification in chromatin dynamics. J Biochem 2007;141:609–14. [14] Weake VM, Workman JL. Histone ubiquitination: triggering gene activity. Mol Cell 2008;29:653–63. [15] Sadoul K, Boyault C, Pabion M, Khochbin S. Regulation of protein turnover by acetyltransferases and deacetylases. Biochimie 2008;90:306–12. [16] Takahashi H, McCaffery JM, Irizarry RA, Boeke JD. Nucleocytosolic acetyl-coenzyme A synthetase is required for histone acetylation and global transcription. Mol Cell 2006;23:207–17. [17] Adcock IM, Lee KY. Abnormal histone acetylase and deacetylase expression and function in lung inflammation. Inflamm Res 2006;55:311–21. [18] Suo L, Meng QG, Pei Y, Yan CL, Fu XW, Bunch TD, et al. Changes in acetylation on lysine 12 of histone H4 (acH4K12) of murine oocytes during maternal aging may affect fertilization and subsequent embryo development. Fertility Sterility 2010;93:945–51. [19] Taylor PB, McDonald GW. Hypoxia stimulates myocardial RNA synthesis and histone acetylation. Res Commun Chem Pathol Pharmacol 1983;40:349–52. [20] Kim YB, Yu J, Lee SY, Lee MS, Ko SG, Ye SK, et al. Cell adhesion status-dependent histone acetylation is regulated through intracellular contractility-related signaling activities. J Biol Chem 2005;280:28357–64. [21] Petrelli F, Coderoni S, Moretti P, Paparelli M. Effect of biotin on phosphorylation, acetylation, methylation of rat liver histones. Mol Biol Rep 1978;4:87–92. [22] Jia J, Chen J. Histone hyperacetylation is involved in the quercetininduced human leukemia cell death. Pharmazie 2008;63:379–83. [23] Mukwevho E, Kohn TA, Lang D, Nyatia E, Smith J, Ojuka EO. Caffeine induces hyperacetylation of histones at the MEF2 site on the Glut4 promoter and increases MEF2A binding to the site via a CaMK-dependent mechanism. Am J Physiol Endocrinol Metab 2008;294:E582–8. [24] Mosley AL, Ozcan S. Glucose regulates insulin gene transcription by hyperacetylation of histone h4. J Biol Chem 2003;278:19660–6. [25] Rahman I, Chung S. Dietary polyphenols, deacetylases and chromatin remodeling in inflammation. World Revi Nutr Diet 2010;101:84–94. [26] Wellen KE, Hatzivassiliou G, Sachdeva UM, Bui TV, Cross JR, Thompson CB. ATP-citrate lyase links cellular metabolism to histone acetylation. Science 2009;324:1076–80. [27] Adamski J, Ma Z, Nozell S, Benveniste EN. 17beta-Estradiol inhibits class II major histocompatibility complex (MHC) expression: influence on histone modifications and cbp recruitment to the class II MHC promoter. Mol Endocrinol 2004;18:1963–74. [28] Baccarelli A, Bollati V. Epigenetics and environmental chemicals. Curr Opin Pediat 2009;21:243–51. [29] Fraga MF, Agrelo R, Esteller M. Cross-talk between aging and cancer: the epigenetic language. Annals NYAcad Sci 2007;1100:60–74. [30] Langley B, Brochier C, Rivieccio MA. Targeting histone deacetylases as a multifaceted approach to treat the diverse outcomes of stroke. Stroke 2009;40:2899–905. [31] Smith KT, Workman JL. Histone deacetylase inhibitors: anticancer compounds. Int J Biochem Cell Biol 2009;41:21–5. [32] Gluckman PD, Hanson MA, Buklijas T, Low FM, Beedle AS. Epigenetic mechanisms that underpin metabolic and cardiovascular diseases. Nat Rev Endocrinol 2009;5:401–8. [33] Halili MA, Andrews MR, Sweet MJ, Fairlie DP. Histone deacetylase inhibitors in inflammatory disease. Curr Topics Med Chem 2009;9:309–19.

V.  Factors Influencing Epigenetic Changes

320

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

[34] Myzak MC, Dashwood RH. Histone deacetylases as targets for ­dietary cancer preventive agents: lessons learned with butyrate, diallyl disulfide, and sulforaphane. Curr Drug Targets 2006;7: 443–52. [35] Druesne-Pecollo N, Chaumontet C, Pagniez A, Vaugelade P, Bruneau A, Thomas M, et al. In vivo treatment by diallyl disulfide increases histone acetylation in rat colonocytes. Biochem Biophys Res Commun 2007;354:140–7. [36] Lea MA, Randolph VM. Induction of histone acetylation in rat liver and hepatoma by organosulfur compounds including diallyl disulfide. Anticancer Res 2001;21:2841–5. [37] Myzak MC, Dashwood WM, Orner GA, Ho E, Dashwood RH. Sulforaphane inhibits histone deacetylase in vivo and suppresses tumorigenesis in Apc-minus mice. Faseb J 2006;20:506–8. [38] Myzak MC, Ho E, Dashwood RH. Dietary agents as histone deacetylase inhibitors. Mol Carcinogen 2006;45:443–6. [39] Lee JI, Nian H, Cooper AJ, Sinha R, Dai J, Bisson WH, et al. Alpha-keto acid metabolites of naturally occurring organoselenium compounds as inhibitors of histone deacetylase in human prostate cancer cells. Cancer Prevent Res 2009;2:683–93. [40] Basak S, Pookot D, Noonan EJ, Dahiya R. Genistein down-­ regulates androgen receptor by modulating HDAC6-Hsp90 chaperone function. Mol Cancer Therap 2008;7:3195–202. [41] Bontempo P, Mita L, Miceli M, Doto A, Nebbioso A, De Bellis F, et al. Feijoa sellowiana derived natural Flavone exerts anti-cancer action displaying HDAC inhibitory activities. Int J Biochem Cell Biol 2007;39:1902–14. [42] Ito K, Lim S, Caramori G, Cosio B, Chung KF, Adcock IM, et al. A molecular mechanism of action of theophylline: induction of histone deacetylase activity to decrease inflammatory gene expression. Proc Natl Acad Sci USA 2002;99:8921–6. [43] Prasad R, Katiyar SK. Polyphenols from green tea inhibit the growth of melanoma cells through inhibition of class I histone deacetylases and induction of DNA damage. Genes Cancer 2015;6:49–61. [44] Meeran SM, Ahmed A, Tollefsbol TO. Epigenetic targets of bioactive dietary components for cancer prevention and therapy. Clin Epigen 2010;1:101–16. [45] Calvanese V, Lara E, Kahn A, Fraga MF. The role of epigenetics in aging and age-related diseases. Age Res Rev 2009;8:268–76. [46] Pudenz M, Roth K, Gerhauser C. Impact of soy isoflavones on the epigenome in cancer prevention. Nutrients 2014;6:4218–72. [47] Bartling TR, Drumm ML. Oxidative stress causes IL8 promoter hyperacetylation in cystic fibrosis airway cell models. Am J Respir Cell Mol Biol 2009;40:58–65. [48] Rahman I, Marwick J, Kirkham P. Redox modulation of chromatin remodeling: impact on histone acetylation and deacetylation, NFkappaB and pro-inflammatory gene expression. Biochem Pharmacol 2004;68:1255–67. [49] Robert C, Rassool FV. HDAC inhibitors: roles of DNA damage and repair. Adv Cancer Res 2012;116:87–129. [50] Yang XJ, Seto E. HATs and HDACs: from structure, function and regulation to novel strategies for therapy and prevention. Oncogene 2007;26:5310–8. [51] Attenni B, Ontoria JM, Cruz JC, Rowley M, Schultz-Fademrecht C, Steinkuhler C, et al. Histone deacetylase inhibitors with a primary amide zinc binding group display antitumor activity in ­xenograft model. Bioorg Med Chem Lett 2009;19:3081–4. [52] Pallas M, Verdaguer E, Tajes M, Gutierrez-Cuesta J, Camins A. Modulation of sirtuins: new targets for antiageing. Recent Patents CNS Drug Discov 2008;3:61–9. [53] Britton RG, Kovoor C, Brown K. Direct molecular targets of resveratrol: identifying key interactions to unlock complex mechanisms. Annals NY Acad Sci 2015;1348:124–33. [54] Yang J, Kong X, Martins-Santos ME, Aleman G, Chaco E, Liu GE, et al. Activation of SIRT1 by resveratrol represses transcription

[55] [56]

[57]

[58]

[59]

[60]

[61]

[62]

[63]

[64]

[65]

[66]

[67]

[68]

[69]

[70]

[71]

of the gene for the cytosolic form of phosphoenolpyruvate carboxykinase (GTP) by deacetylating hepatic nuclear factor 4alpha. J Biol Chem 2009;284:27042–53. Adams J, Klaidman L. Sirtuins, nicotinamide and aging: a critical review. Lett Drug Design Discov 2007;4:44–8. Crujeiras AB, Parra D, Goyenechea E, Martinez JA. Sirtuin gene expression in human mononuclear cells is modulated by caloric restriction. Europ J Clin Invest 2008;38:672–8. Chen D, Bruno J, Easlon E, Lin SJ, Cheng HL, Alt FW, et al. ­Tissue-specific regulation of SIRT1 by calorie restriction. Genes Dev 2008;22:1753–7. Green KN, Steffan JS, Martinez-Coria H, Sun X, Schreiber SS, Thompson LM, et al. Nicotinamide restores cognition in Alzheimer’s disease transgenic mice via a mechanism involving sirtuin inhibition and selective reduction of Thr231-phosphotau. J ­Neurosci 2008;28:11500–10. Balestrieri ML, Rienzo M, Felice F, Rossiello R, Grimaldi V, Milone L, et al. High glucose downregulates endothelial progenitor cell number via SIRT1. Biochim Biophys Acta 2008;1784: 936–45. Selvi RB, Kundu TK. Reversible acetylation of chromatin: implication in regulation of gene expression, disease and therapeutics. Biotechnol J 2009;4:375–90. Manzo F, Tambaro FP, Mai A, Altucci L. Histone acetyltransferase inhibitors and preclinical studies. Expert Opin Therap Patents 2009;19:761–74. Balasubramanyam K, Swaminathan V, Ranganathan A, Kundu TK. Small molecule modulators of histone acetyltransferase p300. J Biol Chem 2003;278:19134–40. Balasubramanyam K, Altaf M, Varier RA, Swaminathan V, Ravindran A, Sadhale PP, et al. Polyisoprenylated benzophenone, garcinol, a natural histone acetyltransferase inhibitor, represses chromatin transcription and alters global gene expression. J Biol Chem 2004;279:33716–26. Balasubramanyam K, Varier RA, Altaf M, Swaminathan V, ­Siddappa NB, Ranga U, et al. Curcumin, a novel p300/CREBbinding protein-specific inhibitor of acetyltransferase, represses the acetylation of histone/nonhistone proteins and histone acetyltransferase-dependent chromatin transcription. J Biol Chem 2004;279:51163–71. Morimoto T, Sunagawa Y, Kawamura T, Takaya T, Wada H, Nagasawa A, et al. The dietary compound curcumin inhibits p300 histone acetyltransferase activity and prevents heart failure in rats. J Clin Invest 2008;118:868–78. Choi KC, Jung MG, Lee YH, Yoon JC, Kwon SH, Kang HB, et al. Epigallocatechin-3-gallate, a histone acetyltransferase inhibitor, inhibits EBV-induced B lymphocyte transformation via suppression of RelA acetylation. Cancer Res 2009;69:583–92. Stimson L, Rowlands MG, Newbatt YM, Smith NF, Raynaud FI, Rogers P, et al. Isothiazolones as inhibitors of PCAF and p300 histone acetyltransferase activity. Mol Cancer Therap 2005;4: 1521–32. Gorsuch S, Bavetsias V, Rowlands MG, Aherne GW, Workman P, Jarman M, et al. Synthesis of isothiazol-3-one derivatives as inhibitors of histone acetyltransferases (HATs). Bioorg Med Chem 2009;17:467–74. Lau OD, Kundu TK, Soccio RE, Ait-Si-Ali S, Khalil EM, Vassilev A, et al. HATs off: selective synthetic inhibitors of the histone acetyltransferases p300 and PCAF. Mol Cell 2000;5:589–95. Friis RM, Wu BP, Reinke SN, Hockman DJ, Sykes BD, Schultz MC. A glycolytic burst drives glucose induction of global histone acetylation by picNuA4 and SAGA. Nucleic Acids Res 2009;37: 3969–80. Shepard BD, Tuma PL. Alcohol-induced protein hyperacetylation: mechanisms and consequences. World J Gastroenterol 2009;15:1219–30.

V.  Factors Influencing Epigenetic Changes



REFERENCES

[72] Choudhury M, Shukla SD. Surrogate alcohols and their metabolites modify histone H3 acetylation: involvement of histone acetyl transferase and histone deacetylase. Alcohol Clin Exp Res 2008;32:829–39. [73] Honma K, Mochizuki K, Goda T. Inductions of histone H3 ­acetylation at lysine 9 on SGLT1 gene and its expression by feeding mice a high carbohydrate/fat ratio diet. Nutrition 2009;25: 40–4. [74] Kulkarni-Shukla S, Barge AP, Vartak RS, Kar A. Cold-induced alteration in the global structure of the male sex chromosome of In1BM2(reinverted) of Drosophila melanogaster is associated with increased acetylation of histone 4 at lysine 16. J Genet 2008;87:235–40. [75] Dang W, Steffen KK, Perry R, Dorsey JA, Johnson FB, Shilatifard A, et al. Histone H4 lysine 16 acetylation regulates cellular lifespan. Nature 2009;459:802–7. [76] Manosalva I, Gonzalez A. Aging alters histone H4 acetylation and CDC2A in mouse germinal vesicle stage oocytes. Biol Reprod 2009;81:1164–71. [77] Dou Y, Milne TA, Tackett AJ, Smith ER, Fukuda A, Wysocka J, et al. Physical association and coordinate function of the H3 K4 methyltransferase MLL1 and the H4 K16 acetyltransferase MOF. Cell 2005;121:873–85. [78] Shogren-Knaak M, Ishii H, Sun JM, Pazin MJ, Davie JR, Peterson CL. Histone H4-K16 acetylation controls chromatin structure and protein interactions. Science 2006;311:844–7. [79] Watson JA, Watson CJ, McCrohan AM, Woodfine K, Tosetto M, McDaid J, et al. Generation of an epigenetic signature by chronic hypoxia in prostate cells. Hum Mol Genet 2009;18:3594–604. [80] Bedford MT, Clarke SG. Protein arginine methylation in mammals: who, what, and why. Mol Cell 2009;33:1–13. [81] Nottke A, Colaiacovo MP, Shi Y. Developmental roles of the histone lysine demethylases. Development 2009;136:879–89. [82] Bannister AJ, Schneider R, Kouzarides T. Histone methylation: dynamic or static? Cell 2002;109:801–6. [83] Zhang Y, Reinberg D. Transcription regulation by histone methylation: interplay between different covalent modifications of the core histone tails. Genes Dev 2001;15:2343–60. [84] Shi Y. Histone lysine demethylases: emerging roles in development, physiology and disease. Nat Rev Genet 2007;8:829–33. [85] Ng SS, Yue WW, Oppermann U, Klose RJ. Dynamic protein methylation in chromatin biology. Cell Mol Life Sci 2009;66:407–22. [86] Dashwood RH, Ho E. Dietary agents as histone deacetylase inhibitors: sulforaphane and structurally related isothiocyanates. Nutr Rev 2008;66(Suppl 1):S36–8. [87] Hitchler MJ, Domann FE. Metabolic defects provide a spark for the epigenetic switch in cancer. Free Rad Biol Med 2009;47:115–27. [88] Huang S. Histone methyltransferases, diet nutrients and tumour suppressors. Nat Rev Cancer 2002;2:469–76. [89] Junien C, Gallou-Kabani C, Vige A, Gross MS. Nutritional epigenomics of metabolic syndrome. Med Sci (Paris) 2005;21:396–404. [90] Zeisel SH. Epigenetic mechanisms for nutrition determinants of later health outcomes. Am J Clin Nutr 2009;89:1488S–93S. [91] Clarke S, Banfield K. Homocysteine in Health and Disease. Cambridge: Cambridge University Press; 2001. [92] Horiguchi K, Yamada M, Satoh T, Hashimoto K, Hirato J, Tosaka M, et al. Transcriptional activation of the mixed lineage leukemiap27Kip1 pathway by a somatostatin analogue. Clin Cancer Res 2009;15:2620–9. [93] Osborne TC, Obianyo O, Zhang X, Cheng X, Thompson PR. Protein arginine methyltransferase 1: positively charged residues in substrate peptides distal to the site of methylation are important for substrate binding and catalysis. Biochemistry 2007;46:13370–81. [94] Pogribny IP, Tryndyak VP, Bagnyukova TV, Melnyk S, Montgomery B, Ross SA, et al. Hepatic epigenetic phenotype predetermines individual susceptibility to hepatic steatosis in mice fed a lipogenic methyl-deficient diet. J Hepatol 2009;51:176–86.

321

[95] Norell M, von der Decken A. Effect of a diet low in methioninecysteine on rat liver chromatin and nuclear proteins. Cell Mol Biol 1989;35:63–74. [96] Davison JM, Mellott TJ, Kovacheva VP, Blusztajn JK. Gestational choline supply regulates methylation of histone H3, expression of histone methyltransferases G9a (Kmt1c) and Suv39h1 (Kmt1a), and DNA methylation of their genes in rat fetal liver and brain. J Biol Chem 2009;284:1982–9. [97] Ara AI, Xia M, Ramani K, Mato JM, Lu SC. S-adenosylmethionine inhibits lipopolysaccharide-induced gene expression via modulation of histone methylation. Hepatology 2008;47: 1655–66. [98] Nicholson TB, Chen T, Richard S. The physiological and pathophysiological role of PRMT1-mediated protein arginine methylation. Pharmacol Res 2009;60:466–74. [99] Puchau B, Zulet MA, Urtiaga G, Navarro-Blasco I, Martinez JA. Asymmetric dimethylarginine association with antioxidants intake in healthy young adults: a role as an indicator of metabolic syndrome features. Metabolism 2009;58:1483–8. [100] Dobosy JR, Fu VX, Desotelle JA, Srinivasan R, Kenowski ML, Almassi N, et al. A methyl-deficient diet modifies histone methylation and alters Igf2 and H19 repression in the prostate. Prostate 2008;68:1187–95. [101] Pogribny IP, Tryndyak VP, Muskhelishvili L, Rusyn I, Ross SA. Methyl deficiency, alterations in global histone modifications, and carcinogenesis. J Nutr 2007;137:216S–22S. [102] Blusztajn JK, Mellott TJ. Choline nutrition programs brain development via DNA and histone methylation. Central Nerv Syst Agents Med Chem 2012;12:82–94. [103] Lillycrop KA, Slater-Jefferies JL, Hanson MA, Godfrey KM, Jackson AA, Burdge GC. Induction of altered epigenetic regulation of the hepatic glucocorticoid receptor in the offspring of rats fed a protein-restricted diet during pregnancy suggests that reduced DNA methyltransferase-1 expression is involved in impaired DNA methylation and changes in histone modifications. Brit J Nutr 2007;97:1064–73. [104] Sharif J, Nakamura M, Ito T, Kimura Y, Nagamune T, Mitsuya K, et al. Food restriction in pregnant mice can induce changes in histone modifications and suppress gene expression in fetus. Nucl Acids Symp Series (Oxf) 2007;51:125–6. [105] Suter MA, Takahashi D, Grove KL, Aagaard KM. Postweaning exposure to a high-fat diet is associated with alterations to the hepatic histone code in Japanese macaques. Ped Res 2013;74: 252–8. [106] Suter MA, Ma J, Vuguin PM, Hartil K, Fiallo A, Harris RA, et al. In utero exposure to a maternal high-fat diet alters the epigenetic histone code in a murine model. Am J Obstetrics Gynecol 2014;210:463. e1–e11. [107] Yang J, Ledaki I, Turley H, Gatter KC, Montero JC, Li JL, et al. Role of hypoxia-inducible factors in epigenetic regulation via histone demethylases. Annals NY Acad Sci 2009;1177: 185–97. [108] Johnson AB, Denko N, Barton MC. Hypoxia induces a novel signature of chromatin modifications and global repression of transcription. Mutat Res 2008;640:174–9. [109] Xia X, Lemieux ME, Li W, Carroll JS, Brown M, Liu XS, et al. Integrative analysis of HIF binding and transactivation reveals its role in maintaining histone methylation homeostasis. Proc Natl Acad Sci USA 2009;106:4260–5. [110] Hancock RL, Dunne K, Walport LJ, Flashman E, Kawamura A. Epigenetic regulation by histone demethylases in hypoxia. Epigenomics 2015;7:791–811. [111] Pollard PJ, Loenarz C, Mole DR, McDonough MA, Gleadle JM, Schofield CJ, et al. Regulation of Jumonji-domain-containing histone demethylases by hypoxia-inducible factor (HIF)-1alpha. Biochem J 2008;416:387–94.

V.  Factors Influencing Epigenetic Changes

322

20.  Dietary and Metabolic Compounds Affecting Covalent Histone Modifications

[112] Wellmann S, Bettkober M, Zelmer A, Seeger K, Faigle M, Eltzschig HK, et al. Hypoxia upregulates the histone demethylase JMJD1A via HIF-1. Biochem Biophys Res Commun 2008;372:892–7. [113] Murayama A, Ohmori K, Fujimura A, Minami H, YasuzawaTanaka K, Kuroda T, et al. Epigenetic control of rDNA loci in response to intracellular energy status. Cell 2008;133:627–39. [114] Kabra DG, Gupta J, Tikoo K. Insulin induced alteration in posttranslational modifications of histone H3 under a hyperglycemic condition in L6 skeletal muscle myoblasts. Biochim Biophys Acta 2009;1792:574–83. [115] Sun H, Zhou X, Chen H, Li Q, Costa M. Modulation of histone methylation and MLH1 gene silencing by hexavalent chromium. Toxicol Appl Pharmacol 2009;237:258–66. [116] Zhou X, Li Q, Arita A, Sun H, Costa M. Effects of nickel, chromate, and arsenite on histone 3 lysine methylation. Toxicol Appl Pharmacol 2009;236:78–84. [117] Chen H, Ke Q, Kluz T, Yan Y, Costa M. Nickel ions increase histone H3 lysine 9 dimethylation and induce transgene silencing. Mol Cell Biol 2006;26:3728–37. [118] Kikuno N, Shiina H, Urakami S, Kawamoto K, Hirata H, Tanaka Y, et al. Genistein mediated histone acetylation and demethylation activates tumor suppressor genes in prostate cancer cells. Int J Cancer 2008;123:552–60. [119] Hall RK, Wang XL, George L, Koch SR, Granner DK. Insulin represses phosphoenolpyruvate carboxykinase gene transcription by causing the rapid disruption of an active transcription complex: a potential epigenetic effect. Mol Endocrinol 2007;21: 550–63. [120] Burke SJ, Collier JJ, Scott DK. cAMP prevents glucose-mediated modifications of histone H3 and recruitment of the RNA polymerase II holoenzyme to the L-PK gene promoter. J Mol Biol 2009;392:578–88. [121] Brasacchio D, Okabe J, Tikellis C, Balcerczyk A, George P, Baker EK, et al. Hyperglycemia induces a dynamic cooperativity of histone methylase and demethylase enzymes associated with gene-activating epigenetic marks that coexist on the lysine tail. Diabetes 2009;58:1229–36. [122] Villeneuve LM, Reddy MA, Lanting LL, Wang M, Meng L, Natarajan R. Epigenetic histone H3 lysine 9 methylation in metabolic memory and inflammatory phenotype of vascular smooth muscle cells in diabetes. Proc Natl Acad Sci USA 2008;105:9047–52. [123] Miao F, Wu X, Zhang L, Yuan YC, Riggs AD, Natarajan R. Genome-wide analysis of histone lysine methylation variations caused by diabetic conditions in human monocytes. J Biol Chem 2007;282:13854–63. [124] Miao F, Smith DD, Zhang L, Min A, Feng W, Natarajan R. Lymphocytes from patients with type 1 diabetes display a distinct profile of chromatin histone H3 lysine 9 dimethylation: an epigenetic study in diabetes. Diabetes 2008;57:3189–98. [125] Li Y, Reddy MA, Miao F, Shanmugam N, Yee JK, Hawkins D, et al. Role of the histone H3 lysine 4 methyltransferase, SET7/9, in the regulation of NF-kappaB-dependent inflammatory genes. Relevance to diabetes and inflammation. J Biol Chem 2008;283:26771–81.

[126] Zager RA, Johnson AC. Renal ischemia-reperfusion injury upregulates histone-modifying enzyme systems and alters histone expression at proinflammatory/profibrotic genes. Am J Physiol Renal Physiol 2009;296:F1032–41. [127] El Mezayen R, El Gazzar M, Myer R, High KP. Aging-dependent upregulation of IL-23p19 gene expression in dendritic cells is associated with differential transcription factor binding and histone modifications. Aging Cell 2009;8:553–65. [128] Cappelletti V, Fioravanti L, Miodini P, Di Fronzo G. Genistein blocks breast cancer cells in the G(2)M phase of the cell cycle. J Cell Biochem 2000;79:594–600. [129] Davie JR, Chadee DN. Regulation and regulatory parameters of histone modifications. J Cell Biochem 1998;30–31:203–13. [130] Ahamed S, Foster JS, Bukovsky A, Wimalasena J. Signal transduction through the Ras/Erk pathway is essential for the mycoestrogen zearalenone-induced cell-cycle progression in MCF-7 cells. Mol Carcinogen 2001;30:88–98. [131] Sharma R, Nakamura A, Takahashi R, Nakamoto H, Goto S. Carbonyl modification in rat liver histones: decrease with age and increase by dietary restriction. Free Rad Biol Med 2006;40:1179–84. [132] Hassan YI, Zempleni J. Epigenetic regulation of chromatin structure and gene function by biotin. J Nutr 2006;136:1763–5. [133] Ke Q, Davidson T, Chen H, Kluz T, Costa M. Alterations of histone modifications and transgene silencing by nickel chloride. Carcinogenesis 2006;27:1481–8. [134] Hassa PO, Haenni SS, Elser M, Hottiger MO. Nuclear ADPribosylation reactions in mammalian cells: where are we today and where are we going? Microbiol Mol Biol Rev 2006;70: 789–829. [135] Fukuda I, Ito A, Hirai G, Nishimura S, Kawasaki H, Saitoh H, et al. Ginkgolic acid inhibits protein SUMOylation by blocking formation of the E1-SUMO intermediate. Chem Biol 2009;16: 133–40. [136] Neeli I, Khan SN, Radic M. Histone deimination as a response to inflammatory stimuli in neutrophils. J Immunol 2008;180: 1895–902. [137] Lu KP, Finn G, Lee TH, Nicholson LK. Prolyl cis–trans isomerization as a molecular timer. Nat Chem Biol 2007;3:619–29. [138] Chen Y, Sprung R, Tang Y, Ball H, Sangras B, Kim SC, et al. Lysine propionylation and butyrylation are novel post-translational modifications in histones. Mol Cell Proteomics 2007;6:812–9. [139] Mathers JC. Session 2: personalised nutrition. Epigenomics: a basis for understanding individual differences? Proc Nutr Soc 2008;67:390–4. [140] Feinberg AP. Phenotypic plasticity and the epigenetics of human disease. Nature 2007;447:433–40. [141] Daniel M, Tollefsbol TO. Epigenetic linkage of aging, cancer and nutrition. J Exp Biol 2015;218:59–70. [142] Longo VD, Antebi A, Bartke A, Barzilai N, Brown-Borg HM, Caruso C, et al. Interventions to slow aging in humans: are we ready? Aging Cell 2015;14:497–510. [143] Heerboth S, Lapinska K, Snyder N, Leary M, Rollinson S, Sarkar S. Use of epigenetic drugs in disease: an overview. Genet Epigen 2014;6:9–19.

V.  Factors Influencing Epigenetic Changes

C H A P T E R

21 Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders Bhairavi Srinageshwar*, Panchanan Maiti*,**, Gary L. Dunbar*,**, Julien Rossignol* *Central Michigan University, Mt. Pleasant, MI, United States; **Field Neurosciences Institute, Saginaw, MI, United States

O U T L I N E Introduction to Epigenetics

323

Epigenetics and the Human Brain Stem Cells Eukaryotic Chromosomal Organization

324 324 325

INTRODUCTION TO EPIGENETICS Epigenetics is defined as structural and functional changes occurring in histones and DNA, in the absence of alterations of the DNA sequence, which, in turn, has a significant impact on how gene expression is altered in a cell [1]. The term “epigenetics” was coined by the famous developmental biologist, Cornard Hal Waddington, as “the branch of biology that studies the causal interactions between genes and their products, which bring the phenotype into being”[2]. Epigenetics bridge the gap between the environment and gene expression, which was once believed to function independently [3]. Epigenetic changes can lead to increase or decrease in gene expression, thereby activating and/or deactivating genes, depending on the nature of the epigenetic control. Some of the most important histone modifications include: (1) methylation; (2) acetylation; (3) phosphorylation; (4) ubiquitination; and (5) SUMOylation and DNA modification, including DNA methylation. These Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00021-3 Copyright © 2017 Elsevier Inc. All rights reserved.

Histones and Their Structure Epigenetics and Neurological Disorders Conclusions References

325 326 335 336

changes are discussed in detail elsewhere [3,4], but are briefly described later as an overview for this chapter. DNA methylation. DNA methylation and some of the histone modifications are interdependent and play an important role in gene activation and repression during development [5]. DNA methylation reactions are catalyzed by a family of enzymes called DNA methyl transferases (DNMTs), which add methyl groups to a cytosine base of the DNA at the 5’-end, giving rise to the 5’-methyl cytosine. This reaction can either activate or repress gene expression, depending on the site of methylation and it can also determine how well the enzymes for gene transcription can access the DNA that is wrapped around the histone [6]. Histone methylation. Trimethylation of lysine at position 4 on histone 3 (H3K4me3) promotes gene transcription (i.e., gene activation), whereas trimethylation of lysine at position 27 on histone 3 (H3K27me3) inhibits gene transcription (i.e., gene silencing). Alternate gene activation and repression promote a balanced dose of gene

323

324

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

expression, which is required for those genes involved in overall development and maturation of organisms. This ensures that only appropriate genes are turned “ON” and “OFF” at any given point of time [7]. Huntington’s disease (HD), Parkinson’s disease (PD), and multiple sclerosis (MS) are some of the diseases that are caused by abnormal DNA methylation pattern. These are discussed in more detail in later sections of this chapter. Histone acetylation. DNA acetylation involves the acetylation of lysine residue, which is catalyzed by the enzymes, histone acetyltransferases (HATs), and histone deacetylases (HDACs), which have opposing effects on each other. HATs transfer acetyl groups to lysine, whereas HDACs remove acetyl groups from lysine. However, presence or absence of an acetyl group (CH3CO) on a lysine residue alters the charge on the amino acid and can decrease the interaction of the N-terminal region of histones with the negatively charged phosphate groups of DNA. These events are involved in transformation of the condensed chromatin into a more relaxed structure, which can induce gene expression [8]. Histone phosphorylation. Phosphorylation involves addition of phosphate group to threonine, serine, and tyrosine residues. Phosphorylation of serine at position 139 on histone 2 (H2) occurs as a response to DNA damage during cell cycling. This relaxes the chromatin; thereby the proteins or factors responsible for repairing the damaged regions of the DNA have greater access to the DNA, which aids in the recovery of the DNA damage. Moreover, phosphorylation of threonine and serine residues on histone 3 (H3) facilitates regulation of gene expression [9]. Histidine phosphorylation is one of the epigenetic modifications occurring in the prokaryotic cells and in lower eukaryotes that plays a major role in cell signaling. The histidine is phosphorylated at the imidazole ring, but occurs only on those nitrogen atoms that are unprotonated [10]. Histone ubiquitination and SUMOylation. Ubiquitination and SUMOylation are associated with posttranslational modifications that regulate transcription of gene and protein translation activities. It is well established that the addition of ubiquitin molecules to proteins facilitate targeted protein degradation. In addition to ubiquitin, various small ubiquitin-like molecules (SUMO) are observed in cells, known as small ubiquitin-related modifier. These molecules have activities, which are similar to ubiquitin and can attach covalently to those proteins, which are involved in changing chromatin structure and gene expression [3].

EPIGENETICS AND THE HUMAN BRAIN Generation of neurons and glial cells from progenitor cells involve epigenetic mechanisms that take place throughout the developmental stages of the brain.

The outer layer of the embryo, called ectoderm, forms the central nervous system, and during the process of development, DNA methylation controls the epigenetic mechanisms of the embryonic cells. For example, to prevent the differentiation of nonneuronal cells into mature neurons, the proneural genes, as well as gene which are associated with proteins involved with neurogenesis, such as the brain-derived neurotrophic factor (BDNF) gene, are silenced by DNA methylation at their promoter region. However, DNA remethylation can take place on a neuronal gene, such as Sox2 [11], which allows for the selective initiation of neuronal development. Similarly, during initial development of the cortex, the genes involved in the formation of glial cells are suppressed by DNA methylation, promoting the formation of more neurons during the early stages of neuronal development. Eventually, during the later stages of cortical development, the DNA methylation is reversed, leading to the generation of glial cells [12,13]. Some of the genes involved in postnatal neurogenesis are Sox2, Dlx2, Sp8, and Neurog2 and those involved in the formation of glial cells are Sparcl1 and Nkx2-2. It has been shown that during the differentiation of neural stem cells (NSCs) or progenitor cells, DNA methylation is facilitated by DNMT3a, which silences the genes Sparcl1 and Nkx2-2, leading to the inhibition of glial cell formation and promotion of mature neuron development from NSCs [14]. It is also believed that the normal aging process in humans is associated with modification of the epigenome in the brain, affecting certain genes related to neurogenesis, especially within the cortex [15]. This process leads to the disruption of synapses, abnormal neurotransmission, and is associated to age-related disorders [16]. However, a comprehensive description of the epigenetic mechanisms related to aging is beyond the scope of this chapter.

Stem Cells Dysregulation of epigenetic mechanisms has a direct impact on gene expression patterns that lead to abnormal gene functions, which form the basis of most of the genetic disorders (monogenic or polygenic) in humans. Environmental stress can also alter epigenetic mechanisms, which could become a cause for the predisposition of certain diseases, such as autism, schizophrenia, and congenital heart disease [17]. This chapter focuses on the role of epigenetics and epigenetic changes that take place during stem cell differentiation, which can be used as a potential therapy for neurological diseases. Stem cells have a unique property of proliferation, differentiation, and self-renewal. Stem cell plasticity is an important characteristic and is based on the degree of pluripotency, which is the ability of a cell to differentiate into another cell lineage. Stem cells can be classified

V.  Factors Influencing Epigenetic Changes



Epigenetics and the Human Brain

325

TABLE 21.1 An Overview of Various Epigenetic Mechanisms Associated With Neurodegenerative Diseases Neurodegenerative disease

Stem cell based therapy

Epigenetic mechanism involved with disease

HD

BM-MSCs

Histone 3 (H3) methylation leading to reduced trophic factors [20]

PD

NSCs

DNA methylation leading to metabolic defects [21]

RTT

iPSCs

Point mutations in MeCP2 gene leading to defective epigenetic regulatory molecules [22]

SCA

UC-MSCs

Methylation and acetylation of histone leading to reduced RNA expression [23,24]

MS

HSCs

DNA methylation, histone acetylation and posttranscriptional modification by miRNA leading to compromised immune response [25–27]

HD, Huntington’s disease; MS, multiple sclerosis; PD, Parkinson’s disease; RTT, Rett syndrome; SCA, spinocerebellar ataxias.

as: (1) totipotent, such as embryonic stem cells (ESCs) of the morula, which can be differentiated into any cell type, including placental cells; (2) pluripotent, such as induced pluripotent stem cells (iPSCs), which can be differentiated into any cell type, except for placental cells; and (3) multipotent, such as mesenchymal stem cells (MSCs) and neural stem cells (NSCs), which can be differentiated into many, but not all, cell types. ESCs are highly unspecialized and can form any type of specialized cells under appropriate conditions and environments. ESCs divide and renew themselves, which make them appropriate candidates for regenerative medicine or cell replacement therapy [18]. Previously, we have reviewed the role of epigenetics in MSCs, NSCs, and iPSCs and their association with neurodegenerative diseases [19]. In addition, the basics of stem cells, and the various epigenetic mechanisms associated with them, are explained comprehensively in the previous edition of this book. Therefore, the aim of the present chapter is to discuss the role that epigenetics of stem cells might play in a subset of neurological diseases. Some of the inherited genetic disorders that occur as a consequence of abnormal epigenetic mechanisms include: (1) HD; (2) PD; (3) Rett Syndrome (RTT); (4) spinocerebellar ataxia (SCA); and (5) MS. Currently, stem cell–based therapies are being tested as a potential treatment for these diseases. A sampling of these treatment strategies are outlined in this chapter and include examples of: (1) transplants of bone marrow-derived mesenchymal stem cells (BM-MSCs) for treating HD; (2) transplants of neural stem cells (NSCs) for treating PD; (3) transplants of induced pluripotent stem cells (iPSCs) for treating RTT; (4) transplants of umbilical cord-derived mesenchymal stem cells (UC-MSCs) for treating SCA; and (5) transplants of hematopoietic stem cells (HSCs) for treating MS. These treatment strategies and the epigenetic mechanisms involved with these disorders are summarized in Table 21.1. To better understand the role of epigenetics in neurological disorders, a working knowledge about histones,

chromosomes, and the general hierarchy in the chromosomal organization, is needed, which is briefly reviewed in the next section.

Eukaryotic Chromosomal Organization The strands of DNA are about 2 nm wide and are packed around the core of four pairs of histone proteins (H2A, H2B, H3, and H4), forming the nucleosome, which is the first level of chromosomal organization (Fig. 21.1). These nucleosomes are the building blocks of the chromatin structure, which is about 30 nm in diameter. The formation of chromatin structure involves the fifth histone, H1, which is near the adjacent nucleosome, thereby compacting the nucleosome or chromatin to form chromatin coils, which are about 300 nm in diameter. These fibers are further condensed to make loops of 700 nm in diameter, which, in turn, form the intact metaphase chromosomes, which are about 1400 nm wide [28].

Histones and Their Structure The genome consists of two molecules of each histone protein (H2A, H2B, H3, and H4) giving rise to an octamer, which is made of about 130 amino acids. The nucleosome, as discussed before, consists of DNA, that is, 146–147 nucleotides long, making about 1.65 turns around the octamer. The core histones are highly conserved in eukaryotes, having a “tail” at their N-terminal end where the epigenetic modifications, such as methylation, acetylation, and/or phosphorylation, take place. These regulate the chromatin structure, which has an impact on recruiting various proteins involving activation and repression of gene expression [28,29]. Defects in chromatin organization and deficiency of enzymes lead to various forms of human diseases, such as RTT, Rubinstein–Taybi syndrome, and Coffin–Lowry syndrome [30].

V.  Factors Influencing Epigenetic Changes

326

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

FIGURE 21.1  The eukaryotic chromosomal organization. The steps showing the chromosomal organization involved in condensation of a 2 nm wide DNA wrapping around the histone molecules to form the nucleosome, which in turn twists to form coils (30 nm) and loops (300 nm) which further condenses into a 1400 nm wide chromosome [28].

Epigenetics and Neurological Disorders Huntington’s Disease and Mesenchymal Stem Cells Huntington’s disease (OMIM #143100) is a devastating, fatal, autosomal dominant neurodegenerative disorder, which most profoundly affects the striatal region of the brain. The disease is characterized by cognitive, motor, and psychiatric disturbances [31]. There is atrophy and loss of the medium spiny GABAergic neurons in the caudate and putamen regions of the striatum, which leads to motor and cognitive impairment [32]. The disease is due to the expansion of CAG repeats in the Huntingtin gene (HTT), which produces mutant huntingtin protein (mHTT) that is highly toxic, leading to the signs and symptoms of the disease [33].

Currently our laboratory is investigating the role of transplanting BM-MSCs and UC-MSCs as a potential treatment for HD. We have shown that BM-MSCs create an optimal microenvironment in the striatum that slows the progression of neuronal loss and dysfunction by restoring the various neurotrophic factors, including BDNF, which is down regulated in HD. In our previous studies, we transplanted BM-MSCs, which were genetically altered to overexpress BDNF, into the striata of YAC128 mice (a slowly progressing, transgenic HD mouse model, which carries the entire human mHTT gene) and R6/2 mice, (fast-progressing, transgenic HD mouse model, carrying exon 1 of human mHTT gene), and observed profound neuroprotective effects, including a significant reduction in the motor symptoms of the

V.  Factors Influencing Epigenetic Changes



Epigenetics and the Human Brain

disease. These observations suggest that manipulation of BM-MSCs could be a potential therapeutic for alleviating HD pathology [34–36]. The Huntingtin gene. The normal HTT protein is localized in the cytoplasm, whereas mHTT is found in the nucleus, as well as in the cytoplasm. Though both wild type HTT and the mHTT can interact and inhibit acetyltransferase function, the actual presence of the mHTT in the nucleus was shown to be specifically responsible for interfering with the acetylation of histones [37]. The cognitive symptoms observed in HD are linked to the hypermethylated status of certain genes, such as Sox2 and Pax6. Because these genes play vital roles in the proliferation and maintenance of neural stem cells, which eventually become neurons, alterations of their structure may contribute to subsequent dysfunctions, such as an impairment of hippocampal neurogenesis [38]. Similarly, the hypoacetylation of certain genes, such as BDNF gene, has been found in HD patients and animal models of HD. The CREB binding protein (also known as CREBBP) is a histone acetyltransferase and a transcriptional cofactor that regulates histone acetylation and gene activation. When the mHTT interacts with the CREB binding protein, it loses its transcriptional activator and HAT functions, which lead to hypoacetylation of histones. This in turn, leads to transcriptional dysregulation of certain genes in the neurons in HD brain [39]. Later, it was found that defects, other than DNA methylation and acetylation, are involved in HD. Lee and coworkers [40] investigated the role of microRNA molecules (miRNA) and found lower levels of 9 types of miRNA in 12-month-old YAC128 and 10-week-old R6/2 mice. Understanding the defective epigenetic mechanisms in HD has led to the use of HDAC inhibitors and miRNA as potential treatments for this disease. HD and BDNF. BDNF is an important neurotropic factor, which is expressed abundantly in different brain regions. H3K4me3 is a histone involved in transcription of the BDNF gene, which is significantly reduced in the cortex of the HD brain. BDNF is expressed in the cortical neurons that project into the striatum and has been shown to be essential for the survival of striatal neurons. As mentioned earlier, histone methylation leads to gene silencing and histone acetylation leads to gene activation. However, H3K4 is an exception because either methylation or acetylation of this histone leads to gene activation. Modification of H3K4me3 is widely studied and is of interest to researchers because H3 is associated with the promoters of the genes that are actively transcribed (e.g., BDNF) [41]. Research using chromatin immunoprecipitation (ChIP) analysis has revealed a correlation between BDNF expression and H3K4me3 levels in HD [20]. In the brain, repressor element-1 silencing transcription factor/neuron-restrictive silencer factor (REST/NRSF) is the main factor that recruits other

327

cofactors to regulate neuronal gene expression. REST, a transcriptional repressor, plays a major role in the regulation of the BDNF gene. Under normal circumstances, REST binds with the HTT in the cytoplasm, but in the case of mHTT, there is no interaction between the REST and HTT (Fig. 21.2). This leads to nuclear translocation of the REST, which then inhibits BDNF expression, resulting in neurodegeneration. REST not only targets BDNF gene, but also influence other neuronal genes that are down regulated in HD [41]. These findings show that rescue of REST-regulated genes may prove to have a promising therapeutic effect on HD. H3K4me3 levels are significantly lower at the REST region of the BDNF gene in the cortex when comparisons are made between 8 and 12-week old R6/2 mice. H3K4me3 not only regulates BDNF gene expression, but also has an impact on other genes, a postulate that was confirmed by transcriptome analysis and genome-wide analysis, which revealed that there are about 98 major genes showing differential expressions in the cortex and the striatum in 12-week old R6/2 mice. In the cortex, the differentially expressed genes are associated with neurotransmission (e.g., Grla3, Grm4, and Bagra5), G-protein signaling (e.g., Rgs9 and Arpp21), synaptic transmission (e.g., Snap25 and Rph3a), inflammation (e.g., C4a and Dusp6), and calcium signaling (e.g., Scn4b, Hpca, and Itpr1). In the striatum, the differentially expressed genes are associated with neurotransmission (e.g., Drd2, Grm3, and Gabrd), G-protein signaling (e.g., Rgs9 and Arpp21), synaptic transmission (e.g., Snapr and Dlg4), and calcium signaling (e.g., Scn4b, Hpca, and Itpr1) [20]. Stem cell therapy for HD. MSCs are adult stem cells that are abundantly present in bone-marrow (BM). MSCs can also be derived from umbilical cord (UC) and adipose tissue (AT). The MSCs, derived from BM and AT, have a greater survival rate, when compared to the MSCs derived from other sources [43]. BM-MSCs can differentiate into osteogenic, adipogenic, and chondrogenic lineages. However, by triggering an epigenetic mechanism, these MSCs can differentiate into a neuronal-like lineage.

FIGURE 21.2  Role of REST in downregulation of BDNF in HD. (A) The wild-type HTT protein binds to the REST protein in the cytoplasm, thereby preventing the REST molecule to bind to the BDNF promoter. (B) The mHTT fails to bind to the REST which causes REST to bind to BDNF promoter and inhibits trophic factor transcription leading to reduced BDNF expression as seen in HD brain [42].

V.  Factors Influencing Epigenetic Changes

328

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

This can be achieved by passaging the MSCs and every time the MSCs get passaged, a series of methylation and acetylation reactions take place [44]. Recently, we have found that MSCs at higher passages (passaged about 40–50 times prior to transplantation) have more therapeutic efficacy for alleviating motor symptoms in the R6/2 mice, compared to lowerpassaged MSCs (passaged about 3–8 times prior to transplantation). This study showed that passaging the cells up to 40–50 times produces a subpopulation of MSCs that have the potential to create an optimal environment within the transplantation site in the striatum, as well as generating BDNF, which is usually deficient in HD brains [36]. Genes that are associated with steering the MSCs toward osteogenic, adipogenic, and chondrogenic lineages, such as osteopontin (OPN), peroxisome proliferator-activated receptors gamma 2 (PPAR-γ2), and fatty acid binding protein 4 (FABP4), undergo methylation and histone acetylation, as the cells are passaged, thereby reducing differentiation into osteogenic, adipogenic, or chondrogenic lineages [44,45]. The acetylation status of H3K9 at the promoter regions of these genes undergoes changes, thereby leading to less activation [46]. Therefore, when higher passaged MSCs are transplanted into animal models, the environment at the transplantation site is more favorable for increasing BDNF, which is required for neuronal survival and alleviating symptoms of HD [36]. Similarly, in another study, we genetically modified the BM-MSCs to overexpress BDNF and transplanted these cells into the striatum of the YAC128 mice. The mice, which received BM-MSCs that overexpressed BDNF showed improvement in motor coordination, compared to YAC128 mice, which did not receive BMMSCs that overexpressed BDNF [34]. Interestingly, one of our previous studies involved transplantation of UC-MSCs into the striata of the R6/2 mice and though these mice showed reduction in both spatial memory and motor deficits, the extent of behavioral sparing was slightly higher when BM-MSCs were used, suggesting that the source of the MSCs may affect their efficacy when transplanted [35]. Conclusions for epigenetics in HD. The epigenetic alterations of genes, such as OPN, PPAR-γ2, and FABP4, steer MSCs away from osteogenic, adipogenic, and chondrogenic lineages as a function of cell passaging and may play an important role in driving these stem cells into a neuronal-like lineage. The aforementioned studies show that MSCs have a therapeutic effect on HD and by utilizing higher-passaged MSCs, the transplants appear to be more efficacious than using lower-passaged MSCs. However, the higher-passaged MSCs may have less clinical utility, because they are more susceptible to other epigenetic effects and tumor formation following transplantation, as they have been shown to carry

chromosomal abnormalities that may adversely affect survivability and successful engraftment at the transplantation site [47]. Therefore, even though higherpassaged MSCs may have a stronger therapeutic effect for HD, it is important to study the epigenetic mechanisms of MSCs at different passages to select a subpopulation of cells that will show their maximum therapeutic effects, without adverse effects. Epigenetic mechanisms play an important role in the dysregulation of genes that are associated with the secretion of trophic factors, such as BDNF, as described previously. Therefore, targeting epigenetic markers and improving the expression of BDNF may prove to be beneficial in alleviating the signs and symptoms of HD [20]. Although various studies have described histone modifications and histone variants that are associated with HD, including phosphorylation of histone 2 variants observed in HD cell line and in R6/2 mouse model [48], a detailed description of this and many other histone variations is beyond the scope of this chapter. Parkinson’s Disease and Neural Stem Cells Parkinson’s disease (OMIM #168600) is a late-onset neurodegenerative disease, mainly affecting individuals of about 65–85 years of age. The disease is characterized by impairment of both motor and nonmotor symptoms, including rigidity, bradykinesia, tremor, postural instability, depression, abnormal sleep patterns, cognitive dysfunction, and autonomic insufficiency [21,49]. PD is due to the degeneration of dopaminergic neurons in substantia nigra, pars compacta (SNpc). The genetic factors, such as mutations in PARK genes and environmental factors, such as aging and exposure to neurotoxins, contribute to the disease [50,51]. As such, PD can be caused either by sporadic mutations or can be inherited. Major gene candidates that are associated with PD, include PARK genes, leucine-rich repeat kinase 2 gene (LRRK2), and the α-synuclein gene (SNCA). The mutations in PARK gene family (PARK 1–15) and SNCA show Mendelian inheritance patterns, suggesting familial PD. Sporadic PD is caused by variants found in SNCA and LRRK 2 genes. PD also shows polygenic and complex inheritance patterns, combined with environmental factors [21]. Impairment of one-carbon metabolism in PD. The group of metabolic reactions consisting of various enzymes and coenzymes that are involved in various biological functions that involve lipid metabolism, redox reaction, and methylation reaction is known as one-carbon metabolism. These metabolic activities take place by utilizing glucose, amino acids, such as serine and glycine, and vitamins, such as B12 and B6 [52]. Therefore, impairment in DNA methylation is a part of one-carbon metabolism that is found in PD.

V.  Factors Influencing Epigenetic Changes



Epigenetics and the Human Brain

As mentioned earlier, DNA methylation is one of the major epigenetic modifications that ensures condensation or relaxation of chromatin structure, depending on the methylation status of the DNA that is wrapped around the histones. The chemical reaction is catalyzed by DNMT1, where cytosine gets methylated to form 5′methyl-cytosine, a reaction that involves two major molecules, S-adenosylmethionine (SAM), and S-adenosylhomocysteine (SAH). SAM is a universal methylating agent produced from folate and homocysteine (HCY), which methylates histones and DNA [21]. SAM, SAH, and HCY are some of the biomarkers of the metabolic pathway associated with the one-carbon metabolism [53]. Hence, the DNA methylation potential depends on the levels of SAM and the potential of DNMT1 to catalyze the reaction and transfer the methyl group to cytosine. This reaction was found to be impaired in many neurodegenerative diseases, especially in genes associated with Alzheimer’s disease (AD) and PD. Due to defects in one-carbon metabolism, the rate of methylation decreases, leading to increased gene expression. An example of such defects in metabolism was found during the analysis of methylation status of the SNCA gene in PD patients [21]. The SNCA gene. The α-synuclein protein is important for dopaminergic neurogenesis during early embryonic development. The Lewy bodies found in PD consist of SNCA and protein inclusion bodies, which lead to disease pathogenesis. Lower levels of SNCA are associated with loss of dopaminergic neurons during embryonic stage, whereas increased expression of SNCA may be a risk factor or a threat during later stages [54]. The genetic abnormalities associated with SNCA that lead to PD are the point mutations and copy number variants. Previous studies have reported that the mutation of the SNCA genes found in dopaminergic neuron, results in decreased methylation, leading to mono-allelic expression of the gene. However, mRNA levels from this single allele exceed that of the control subjects having normal biallelic expression. Methylation status of the intronic regions (human intron 1 having 66 CpG sites) of the SNCA gene was found to be similar to those reported in previous studies, especially in reference to the cortex and substantia nigra regions of the brain [55,56]. The miRNA also plays a role in regulating the gene expression of SNCA. Doxakis [57] has analyzed two of the major miRNA of the brain, mi-RNA7 and mi-RNA153, and found that overexpression of these miRNAs in neuronal culture lowered the SNCA levels. The PARK gene. The majority of PARK gene mutations are also associated with juvenile form of PD. Cai and coworkers [58] investigated about 33 CpG regions on the PARK gene promoter in three groups of individuals, including PD patients carrying PARK gene mutations, PD patients without PARK gene mutations,

329

and age-matched controls. DNA methylation analysis revealed that there was no significant difference in methylation status between the three groups, which suggests that PARK gene methylation does not contribute in the PD pathogenesis. Stem cell therapy for PD. More than half of the dopaminergic neurons are lost in SNpc before the actual onset of PD [59]. There are various studies and literature reviews that have investigated the role of neural stem cells in PD [49,54,60] and the importance of neurotrophic factors, particularly glial cell line-derived neurotrophic factor (GDNF). The NSCs are multipotent stem cells that are specifically found in the subventricular zone (SVZ), subgranular zone (SGZ), and the dentate gyrus (DG) of the hippocampus. The environment at these regions are the most favorable for the differentiation of NSCs into neurons [61]. Sanberg [60] has discussed the role of transplantation of undifferentiated NSCs into the striatum of a primate model of PD and indicated that NSCs were able to survive and migrate to the site of neurodegeneration and replace the lost neurons. The animals recovered from their behavioral deficits. Redmond and coworkers [62] also transplanted undifferentiated NSCs into the primate model of PD and found that though, these undifferentiated NSCs had therapeutic effect, only a small population of these NSCs partially differentiated into dopaminergic neurons, due to a less-than optimal microenvironment at the site of transplantation. These studies indicate that the partially differentiated NSCs migrate to the substantia nigra, through the nigrostriatal pathway, following their unilateral transplantation into the striatum. However, these finding suggest that cell replacement therapy provides only minimal neuroprotective effects. Other approaches, such as dopamine replacement therapy and deep brain stimulation (which decrease tremor and rigidity), also failed to show neuroprotective effects. GDNF is well known to increase survivability of dopaminergic neurons, and delivering GDNF into the brain of PD animal models, such as MPTP (1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine)-treated rhesus monkey [63] and 6-OHDA (6-hydroxydopamine) injected rats [64,65], were shown to be neuroprotective. Overall, GDNF is critical for establishing the nigrostriatal dopamine system during development and plays an important role in protecting dopaminergic neurons from degeneration by maintaining their morphology and neurochemical and biochemical reactions that are taking place in them, as well as ensuring proper neuronal differentiation and long-term survivability of neurons [66]. Open-labeled clinical trials using GDNF showed tolerance and clinical benefits in patients within 3 months of treatment, but randomized clinical trials failed to reveal significant benefits [67]. Deng and coworkers [68] have shown that the cotransplantation of dopaminergic

V.  Factors Influencing Epigenetic Changes

330

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

neurons and NSCs can reduce motor symptoms in a rodent model of PD. However, to increase the number of NSCs that differentiate into dopaminergic neurons, the overexpression of nuclear receptor (Nurr1) and factors, derived from local type 1 astrocytes, are necessary. Wagner and coworkers [69] showed that NSCs having these factors were able to differentiate into dopaminergic neurons, compared to NSCs that did not express these factors. Similarly, Nurr1 and Pitx3 are needed to produce dopaminergic neurons from embryonic stem cells. Nurr1 is a nuclear hormone receptor that is involved in the dopaminergic neurogenesis, while Pitx3 is a transcription factor that is important for the differentiation and maintenance of dopaminergic neurons in the mid-brain. Nurr 1, along with Pitx3, influences the expression of some of the genes involved in production of dopamine, such as tyrosine hydroxylase (TH) and dopamine transporter (DAT), which are associated with dopamine signaling [70]. Nurr1 is usually present in a silenced state when not combined with Pitx3. This is due to the binding of silencing mediator of retinoic acid and thyroid hormone receptor (SMRT) that leads to HDACmediated silencing of Nurr1. However, in the presence of Pitx3, the binding of SMRT is reduced and Nurr1 is activated, thereby the Nurr1-Pitx3 complex can bind to the promoter of TH and DAT genes and activate them. Therefore, Nurr1, on its own, cannot activate the target genes associated with dopamine production [71]. Conclusions for epigenetics in PD. The familial form of PD involves defects in epigenetic mechanisms, such as one-carbon metabolism reaction, which is associated with defects in rate of DNA methylation that leads to PD. The rate of transfer of methyl group to the cytosine by DNMT1 has been highly reduced, thereby interfering with the one-carbon metabolism, which was confirmed by measuring SAM and SAH, the biomarkers of the one-carbon metabolism associated with DNA methylation. Similarly, abnormal increase in expression of certain miRNAs leads to decreased SNCA gene expression, which is also associated with this disease. Cell replacement therapy to increase production of GDNF through transplantation of partially differentiated NSCs has proven, thus far, to have only limited efficacy. To achieve conversion of NSCs into dopaminergic neurons, the expression of Nurr1 and Pitx3 are required. The complex formed between Nurr1 and Pitx3 is associated with increased gene expression of TH and DAT, which, in turn, increases the production of dopamine. Therefore, complexing Nurr1 and Pitx3 is necessary, since Nurr1, per se, is considered to be in a silenced state, and binding with Pitx3 leads to the activation of the complex. Rett Syndrome and Induced Pluripotent Stem Cells Rett syndrome (OMIM #312750) is a X-linked autosomal dominant progressive neurodevelopmental

disorder affecting, predominantly, the female population and classified as one of the autism-spectrum disorders (ASDs). The classical symptoms of this disease include speech disability, stereotypic hand use, autistic characteristics, and seizures that gradually develop after 18 months of age [22]. The cause of the disease is associated with point mutations in methyl-CpG binding protein 2 gene (MeCP2), although there are other sets of genes and environmental factors that contribute to the onset of this disease. Most of the mutations observed in MeCP2 gene are point mutations (missense or nonsense). Some of the hotspots include: (1) p. R133C and p. T158M, found in the methyl binding domain (MBD); and (2) p. R306C, p. R168X, p. R294X, and p. R255X located in transcriptional repression domain (TRD) [72–75]. The MeCP2 gene has a major role to play in the epigenetic regulation of various gene expressions related to ASD [72]. The domains of MeCP2 gene, such as the MBD and TRD, are involved in chromatin remodeling and protein interactions, respectively [76]. MeCP2 gene and its function. MeCP2 gene is involved in coding an epigenetic regulatory molecule, and mutations or large-scale deletions, duplications, and insertions of MeCP2 cause RTT syndrome. RTT can be classified into two categories, either atypical RTT or classical RTT. More than 95% of the patients having mutation in MeCP2 gene are considered having classical RTT. In general, MeCP2 binds to the DNA via the MBD and silences the gene. Similarly, the MeCP2 protein helps with chromatin remodeling by binding to the DNA via its TRD. The methylation of histone 3 at lysine position 9 (H3K9me) is achieved by MeCP2, thereby silencing the gene to which it is bound [22]. However, Yasui and coworkers [77] have extensively studied the MeCP2 binding sites on the genes and found that only about 6% of the CpG islands are bound by MeCP2. Their study indicated that: (1) the main function of MeCP2 is not associated with silencing the methylated regions of the gene and (2) the genes having the maximum methylation status are not bound by MeCP2. Previous publications have shown that the human genome has methylated cytosine as 5-hydroxymethylcytosine (5hmC) and 5-methylcytosine (5mC). It has been shown that 5hmC is found abundantly in neuronal genes that are active and that MeCP2 has a very high affinity toward 5hmC compared to 5mC, which plays a major role in how the gene expressions are regulated in neurons (Fig. 21.3). An interesting finding is that the MeCP2 competes with the histone, H1, to bind to the nucleosome, indicating that the levels of H1 and MeCP2 are not always corelated with each other, especially in neurons. The finding that MeCP2 gene is associated with activating genes when bound to 5hmC, as well as with silencing the genes when linked with H3K9, underlies the dual nature of the protein. Therefore, there are some genes that are

V.  Factors Influencing Epigenetic Changes



Epigenetics and the Human Brain

331

FIGURE 21.3  Function of MeCP2 gene. MeCP2 regulates the transcription of certain genes in the brain by binding to 5-methylcytosine (5mC). The main function of MeCP2 gene is not just silencing the gene to which it is bound, but also has the ability to increase the transcription of genes, such as BDNF, ORPM1, and CREB 1 [15].

down regulated when MeCP2 is lost and upregulated with increased MeCP2 gene expression [78]. iPSC models of RTT. Induced pluripotent stem cells have been used as a cell model for RTT [79]. Takahashi and Yamanaka [80] first reprogrammed the somatic cells into iPSCs by overexpressing four major genes, such as Oct3/4, Sox2, Klf4, and c-Myc, which are now collectively known as the Yamanaka factors. Reprogramming of iPSCs involves loss and gain of DNA methylation on H3 at lysine positions 27 (H3K27me3) and 4 (H3K4me3), as the cells get transformed from somatic stage, to pluripotent stage as has been discussed in detail in our previously work [19]. The iPSCs derived from the RTT patients have been reprogrammed to form neurons, which show significant pathological changes, such as reduced nuclear size, lower expression of neuronal markers, reduced dendritic spine density, loss of synapses, and lower levels of intracellular calcium. Electrophysiological analysis of neurons obtained from the RTT-derived iPSCs show decreased excitatory and inhibitory postsynaptic potential (EPSP and IPSP). Although cell replacement research for RTT using iPSC transplantation has not been translated to the clinic, the in vitro remodeling of fibroblasts derived from the RTT patients to form iPSCs have been successfully achieved by Marchetto and coworkers [81]. These findings are paving the way toward a better understanding of the pathophysiology of the disease and for identifying drugs or treatments that are patientor mutation-specific [81,82]. Conclusions for epigenetics in RTT. In general, it has been assumed that MeCP2 usually silences the gene to which it is associated, but subsequent studies have indicated that this might not always be the case. The MeCP2 gene can increase, as well as decrease, the gene expression depending on the methylation

status, thereby emphasizing the dual role of the gene. Although both RTT syndrome and iPSCs have wellknown and strong epigenetic components, the use of iPSCs for treating RTT has not been investigated. The stem-cell-based model of RTT is very useful to study a specific mutation leading to a phenotype-genotype correlation and correcting the mutated RTT-iPSCs, in vitro, and then transplanting them may prove to be a future treatment for RTT. Given that highly specific and targeted cell replacement therapy can be achieved using the corrected iPSCs, utilizing epigenetically corrected iPSCs for treating RTT syndrome is worthy of further investigation as such an approach has significant promise. The Spinocerebellar Ataxia and Mesenchymal Stem Cells SCAs are a group of neurodegenerative disorders that are caused by trinucleotide repeat expansions (mainly CAG expansions that lead to elongated polyglutamine tracts). There are about 30 different genes responsible for the disease that is inherited in an autosomal dominant pattern. SCA patients have neuronal degeneration in cerebellum, brain stem, and spinal cord. The main characteristics of this disease are related to retinopathy, neuropathy, cognitive dysfunction, and dementia [83]. Unfortunately, there is no cure for SCAs. Approximately 28 different types of SCAs have been discovered so far, with the most common forms being SCA1, SCA2, SCA3, and SCA7. A detailed description of the different types of SCAs is discussed by Paulson [84]. Interestingly, SCA1, SCA2, SCA3, SCA6, SCA7, and SCA17 are due to the repeat expansion in the coding regions of the gene, whereas SCA8, SCA10, SCA12, and SCA31 have repeat expansions in the noncoding regions of the gene [84].

V.  Factors Influencing Epigenetic Changes

332

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

SCA1 (OMIM #164400);

SCA1 mainly affects the brain stem and Purkinje cells of cerebellum and is characterized by ataxia of limbs and abnormal gait, leading to chorea. Assessments of expanded Ataxin1 gene having 82 CAG repeats in the cerebellar Purkinje cells in SCA1 animal models have shown reduced gene expression, which is involved in signal transduction and calcium homeostasis [85]. Ataxin 1 interacts with two major proteins, the retinoid acid receptor-related orphan receptor α (RORA, a type of HAT), and acetyltransferase tat-interactive protein 60 (Tip60, MW 60kDa), which is a nuclear receptor coactivator that helps in the interaction of RORA [using ATXN1 with HMG-box protein 1 (AXH) domain of ATXN1]. However, in the presence of the mutant Ataxin 1 gene or protein, the interactions with Tip60 and RORAα are disrupted, which lead to disease pathogenesis [86]. SCA 7 (OMIM # 164500);

The symptoms associated with SCA 7 include cerebellar neurodegeneration and retinal degeneration. Similar to SCA 1, SCA7 also interacts with transcription factor cone-rod homeobox protein (CRX). CRX is a transcription activator of genes that are involved in formation of photoreceptors in the eyes. The interactions of mutant SCA7 with CRX cause abnormal formation of the photoreceptors, leading to retinal degeneration. Although the actual function of Ataxin 7 gene is unknown, it has recently been shown that Ataxin 7 is a subunit of HATs [23]. Therefore, the mutant form of the protein is involved in the alteration of HATs, which eventually leads to the disruption of histone acetylation. SCA 8 (OMIM # 608768);

SCA 8 is due to the combined repeat expansion of CTG and CAG in the Ataxin 8 gene. Chen and coworkers [24] studied the SCA 8 cell line, or transcript known as ATXN8OS. Epigenetic analysis of this transcript revealed increased levels of H3K9me2 and hypoacetylation of H3K14 that led to repression of ATXN8OS RNA in cell lines that have 157 repeats. Similarly, methylation of arginine residues and phosphorylation of serine or threonine were found in the cell lines with about 88 repeats, which eventually led to decreased RNA expression. Stem cell therapy for SCA. Although preclinical and clinical trials have been conducted using drugs, antioxidants, and neurotrophic factors, none of these trials were successful in alleviating the symptoms in SCA patients [87]. However, UC-MSCs transplantations in a mouse model of SCA produced promising effects. The major advantages of using UC-MSCs are: (1) there are no ethical issues that arise from their use; (2) they are highly multipotent stem cells; and (3) they have immunosuppressive properties, resulting in reduced risk of tumor formation posttransplantation. Using SCA mice, Zhang

and coworkers [88] showed that UC-MSCs provided a therapeutic effect on these animals, including the restoration of motor functions at 8 weeks posttransplantation. The results of this study also showed that the cerebellar atrophy and the number of cells undergoing apoptosis were reduced. There was also an increased production of growth factors, such as insulin-like growth factor-1 (IGF-1) and vascular endothelial growth factor (VEGF). Jin and coworkers [89] performed intravenous and intrathecal transplantation of UC-MSCs into patients affected with SCA and found that UC-MSCs are safe and have the capacity to alleviate the symptoms in SCA patients. Another clinical study, conducted by Dongmei and coworkers [90] showed that transplantation of UC-MSCs could alleviate the symptoms of SCA, without any side effects, when evaluated using International Cooperative Ataxia Rating Scale (ICARS) and Activity of Daily Living Scale (ADL). Most importantly, these clinical findings provide converging evidence with the preclinical experimental results that UC-MSCs are safe and capable of alleviating the symptoms of SCA, indicating a potentially safe and efficacious therapy for SCA. Conclusions for epigenetics in SCA. There is a strong correlation between epigenetic defects in the genes Ataxin 1, Ataxin 7, and Ataxin 8 and the neuropathological phenotypes associated with SCA in patients. Each ataxin type has its own associated epigenetic mechanism. For example, the Ataxin 1 is a complex that is associated with RORA and Tip60, which are acetyltransferases and the interaction with them is disrupted in the presence of mutant Ataxin 1. Similarly, Ataxin 7 is a part of HAT and presence of mutant Ataxin 7 leads to improper histone acetylation. Detailed analysis of SCA 8 cell line revealed hyper- and hypomethylation of H3K9 and H3K14, respectively. [24] To the best of our knowledge, there are no epigenetic mechanisms related to the UC-MSCs that would drive the cells to take on a specific neuronal phenotype that would confer significant therapeutic effects. However, previous studies have shown a favorable safety profile and beneficial therapeutic effects of UC-MSCs for SCA. Multiple Sclerosis and Hematopoietic Stem Cells Multiple sclerosis (OMIM #126200) is an autoimmune neurological disease characterized by the loss of myelin sheath, leading to demyelination and neurodegeneration of brain and spinal cord. The majority of the affected individuals are females between 20 and 40 years of age. In MS, the T-lymphocytes become stimulated by various factors, which, in turn, activate the inflammatory pathways, leading to the symptoms of the disease [91]. MS involves the genetic, epigenetic, and environmental factors (nutritional status). Epigenetic causes include: (1) DNA methylation; (2) posttranscriptional modification by miRNA; and (3) histone acetylation,

V.  Factors Influencing Epigenetic Changes



Epigenetics and the Human Brain

such as what occurs with the HLA-DRB 1 (human leukocyte antigen having beta chain) gene on chromosome 6, which is responsible for the production of major histocompatibility complex class II (MHC class-II) antigen and which plays an important role in immune response mechanisms [25,92]. The environmental factors include vitamin D deficiency and frequent smoking which, in turn, leads to epigenetic changes observed in MS. DNA methylation in MS. Baranzini and coworkers [93] were the first to study the RNA transcriptome sequences and the epigenome sequences of CD4+ T-lymphocytes from three sets of MS-discordant, monozygotic twins, and found that there were no significant differences in the DNA methylation patterns. However, based on this study, alterations in DNA methylation patterns, as a cause of the disease cannot be ruled out, because the sample size was too small to make definitive conclusions. Other studies have shown that DNA methylation could be the cause of MS [94]. The methylation of CpG islands in some of the genes may be responsible for the disease, because the methylation pattern determines how the two different types of T-helper-cells (Th1 and Th2) are formed, which, in turn, gives rise to cytokines, such as interferon-gamma (IFNγ), interleukin-2 (IL-2), interleukin-4 (IL-4), and tumor necrosis factor-α (TNF-α). In MS, there is a dominance of IFN-γ expression associated with Th1, compared to IL-4 molecules that are associated with Th2. Therefore, abnormal DNA methylation pattern that is found in the promoter region of IFN-γ may explain why the immune response by Th1 is greater than Th2, leading to the pathophysiology of the disease (Fig. 21.4). Similarly, DNA methylation or histone deacetylation is associated with the IL-4 genes, leading to the gene silencing. HDAC inhibitors are given as a potential treatment for MS, because they lead to reduction in inflammation, demyelination, and neuronal degeneration [95]. Epigenetics not only control cytokine gene activation in MS, but also affects myelin structure, by regulating myelin basic protein (MBP). For example, hypomethylation of an enzyme, called peptidyl-arginine deiminase 2 (PAD 2), has been observed in MS [96]. This enzyme is responsible for conversion of arginine to citrullin and its level is increased in MS. Due to overactivation of PAD 2, the MBP is citrullinated and becomes vulnerable for degradation by the myelin-associated proteases, such as cathepsin D. This further leads to a reduced binding capacity of the MBP, causing lipid vesicle fragmentation, which, in turn, leads to the myelin breakdown observed in MS [96] (Fig. 21.5). Micro-RNA. In vitro analysis of MS induced lesions and pooled cells have shown that there were about ten miRNAs that were upregulated, especially mi-RNA155 and mi-RNA326. However, mi-RNA326 is more commonly associated with disease relapse, rather than being

333

FIGURE 21.4  Impact of DNA methylation on T-cells. Compromised immune response exerted by Th1 and Th2 cells (dominance of IFN-γ over IL-4) is due to DNA methylation that leads to MS.

directly associated with the initiation of the disease. It was found that the patients have more severe form of disease; following relapse, and showed a high expression of mi-RNA326 (Table 21.2). This, in turn, leads to higher activation of T-cells and increased expression of inflammatory cytokines, which then results in abnormal immune response, leading to MS [26]. Histone Acetylation. In 2007, the International Multiple Sclerosis Genetics Consortium (IMSGC) conducted a genome-wide association study (GWAS) that included about 12,000 subjects. It was found that two genes, other than HLA-DRB 1, are associated with MS. Single nucleotide polymorphisms (SNPs) in interleukin-2 receptor α gene (IL2RA) and interleukin-7 receptor α gene (IL7RA) were found to be risk factors for MS. Again, in 2011, IMSGC conducted another GWAS study and found that abnormalities in some of the genes involved in cytokines and signal transduction pathways cause MS, reinforcing the fact that the most commonly affected gene by histone acetylation was the HLA-DRB 1 gene [25]. Stem cell therapy for MS. Among different stem cell populations, HSCs have drawn a special attention for use as a potential therapy of MS, because of their multipotency. These stem cells differentiate into a very large population of cells, including all functional types of blood cells, B-cells, T-cells, and many other cell types, which make them a promising candidate for therapy of many diseases [97]. The first therapy trials using HSCs was started

V.  Factors Influencing Epigenetic Changes

334

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

FIGURE 21.5  Comparison of citrullinated myelin in normal and MS patients. Hypomethylation causes increase in the production of enzyme peptidyl-arginine deiminase 2 (PAD 2) leading to increased citrullinated myelin basic protein, thus gives abnormal structure and thereby myelin breakdown in MS patients [92].

TABLE 21.2 Roles of Different Types of miRNA in MS miRNA

Role in MS

mi-RNA155

Dysregulation of gene expression in CD4+ cells and peripheral blood mononuclear cells [27]

mi-RNA326

Dysregulation of gene expression in CD4+ [27]

mi-RNA18b, mi-RNA493, and mi-RNA599

Increased expression in relapse remitting MS [27]

in 1995, followed by a second successful therapy trial on humans in 1998. These studies revealed that scores on the Expanded Disability Status Scale (EDSS) improved for patients who received the HSCs, confirming their use as a potential treatment for this disease [98,99]. However, subsequent clinical trials [100] revealed that patients who have a severe form of this disease (based on their EDSS score) and who received the transplants did not show significant benefits from the cell treatment. It is possible that these stem cells may exert their beneficial effects only during the early stages of disease and not after the disease has progressed to its severe stage [100]. Most of the transplants associated with the HSCs are autogenic in nature, because allogeneic or HLA-matched transplants cause an increase in mortality [91]. Atkins and Freedman [101] reviewed the pros and cons of using HSCs as a potential treatment for MS. One of the major advantages of using HSCs is that once

they are extracted from bone-marrow and transplanted into MS patients, the B-cells, and, especially the T-cells, mature and become activated, which play a major role in boosting the immune levels. However, because it is important to ensure that other immune cells, such as macrophages, do not populate the regions near the graft, it is advisable to use immune ablative conditioning regimens to achieve the maximum benefit from the T-cells. A phase II clinical trial conducted by Mancardi and coworkers [102] was reported recently in 2015, which showed that autologous HSC transplantation resulted in suppression of lesion-induced inflammation. These results are based on a 4-year follow up of patients having a progressive or a relapsing form of the disease. Collectively, these studies show that HSCs, due to their multipotent and immunomodulatory properties, have a significant promise for producing an effective therapy for MS.

V.  Factors Influencing Epigenetic Changes



Conclusions

Conclusions for epigenetics in MS. The compromised immune response seen in MS is due to the epigenetic mechanisms that affect the T-cells and the cytokines, which form the major molecules or proteins of the immune system. Hypomethylation of PAD 2 enzyme is observed in patients affected with MS, which leads to the breakdown of the MBP. Further analysis showed that miRNA, such as mi-RNA155, mi-RNA326, miRNA18b, mi-RNA493, and mi-RNA599 are associated with MS. The GWAS study and other studies have shown that HLA-DRB 1 is the main candidate gene for MS. However, there are some SNPs associated other genes, such as IL2RA and IL7RA, that pose a risk factor for this disease. HSCs have been used as a stem-cellbased therapy for MS, whereby their activation of the T-cells improve immune responses, producing favorable outcomes.

CONCLUSIONS Epigenetics not only play a role in determining the stem cell fate, but also form the underlying bases of neurodegenerative diseases, such as HD, PD, RTT, SCA, and MS, as discussed in this chapter. Although the genetic bases of these diseases vary, the epigenetic causes are similar. For example, alterations in the levels of DNA methylation and histone acetylation gives rise to SCA and MS. Understanding the epigenetic mechanisms of stem cells is an important aspect that needs to be carefully considered when designing strategies to achieve optimal efficacy and efficiency when cell replacement therapy for neurodegenerative diseases is being considered. In order to translate and improve the outcomes of clinical trials, further research on how epigenetics drive stem cell fate is necessary. For example, because the use of MSCs for treating HD has been predominantly used in preclinical trials, researchers should be cognizant that the subpopulation of cells being used is the direct function of the number of passages these cells have undergone. Though higher passaged MSCs have proven to be highly beneficial in restoring the motor symptoms associated with HD, there is also a risk of chromosomal abnormalities that would lead to tumor formation or other adverse effects. Therefore, studying the epigenetic mechanism of MSCs to provide an effective cell replacement therapy is necessary. Some diseases, such as RTT have an epigenetic basis, which makes them excellent candidates for the yet-to-bedeveloped epigenetic-driven stem-cell replacement therapies. As such, a mutation corrected RTT-iPSC cell line for the transplantation may prove to be a promising therapeutic approach. Similarly, identifying the

335

dysregulated epigenetic markers in each disease (such as H3K4me3 in HD and H3K9 and H3K14 in SCA8) may lead to designing a more targeted therapy for such neurological disorders. Addressing these types of issues and furthering our knowledge about the epigenetic mechanisms of stem cells and the diseases associated with epigenetic alterations could pave way toward developing more effective and long-lasting therapeutic approaches.

Abbreviations ADL Activity of Daily Living Scale ASDs Autism-spectrum disorders AT Adipose tissue BDNF Brain derived neurotrophic factor BM Bone marrow BM-MSCs Bone marrow-derived MSCs cAMP Cyclic adenosine monophosphate CBP CREB-binding protein ChIP Chromatin immunoprecipitation CRE cAMP response element CREB CRE-binding protein CRX Cone-rod homeobox protein DAT Dopamine transporter DNMTs DNA methyl transferases EDSS Expanded Disability Status Scale EPSP Excitatory post-synaptic potential ESCS Embryonic stem cells GDNF Glial derived neurotrophic factor GWAS Genome-wide association study HATs Histone acetyltransferases HD Huntington’s disease HDACs Histone deacetylases HLA-DRB Human leukocyte antigen having beta chain HSCs Hematopoietic stem cells HSCs 6-OHDA 6-Hydroxydopamine 5hmC 5-Hydroxymethylcytosine ICARS International Cooperative Ataxia Rating Scale IFN α Interferon α IGF-1 Insulin-like growth factor-1 IL-2 Interleukin-2 IL-4 Interleukine-4 iPSCs Induced pluripotent stem cells LRRK2 Leucine-rich repeat kinase 2 gene MBD Methyl binding domain MBS Myelin basic protein MeCP2 Methyl-CpG binding protein 2 gene MHC class-II Histocompatibility complex class II 5mC 5-Methylcytosine mHtt Mutant huntingtin protein miRNA MicroRNA MPTP 1-Methyl-4-phenyl-1,2,3,6-tetrahydropyridine MS Multiple sclerosis MSC Mesenchymal stem cells NRSF Neuron-restrictive silencer factor NSCs Neural stem cells PAD 2 Peptidyl-arginine deiminase 2 PD Parkinson’s disease PSP Inhibitory post-synaptic potential REST Repressor element-1 silencing transcription factor RTT Rett syndrome

V.  Factors Influencing Epigenetic Changes

336

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

SAH SAM SCA SGZ SMRT

S-adenosylhomocysteine S-adenosylmethionine Spinocerebellar ataxia Subgranular zone Silencing mediator of retinoic acid and thyroid hormone receptor SNCA α-Synuclein SNpc Substantia nigra, pars compacta SUMO Small ubiquitin-like molecules SVZ Subventricular zone Th T-helper-cells TH Tyrosine hydroxylase TNF-α Tumor necrosis factor-α TRD Transcriptional repression domain UC Umbilical cord UC-MSCs Umbilical cord derived MSCs VEGF Vascular endothelial growth factor

References [1] Landgrave-Gómez J, Mercado-Gómez O, Guevara-Guzmán R. Epigenetic mechanisms in neurological and neurodegenerative diseases. Front Cell Neurosci 2015;9:58. Available from: http:// www.ncbi.nlm.nih.gov/pmc/articles/PMC4343006/ [2] Goldberg AD, Allis CD, Bernstein E. Epigenetics: a landscape takes shape. Cell 2007;128(4):635–8. [3] Gill G. SUMO and ubiquitin in the nucleus: different functions, similar mechanisms? Genes Dev 2004;18(17):2046–59. [4] Sadakierska-Chudy A, Filip M. A comprehensive view of the epigenetic landscape. Part II: Histone post-translational modification, nucleosome level, and chromatin regulation by ncRNAs. Neurotox Res 2015;27:172–97. [5] Cedar H, Bergman Y. Linking DNA methylation and histone modification: patterns and paradigms. Nat Rev Genet 2009;10(5):295–304. [6] Jin B, Li Y, Robertson KD. DNA methylation. Genes Cancer 2011;2(6):607–17. [7] Voigt P, Tee W-W, Reinberg D. A double take on bivalent promoters. Genes Dev 2013;27(12):1318–38. [8] Bannister AJ, Kouzarides T. Regulation of chromatin by histone modifications. Cell Res 2011;21(3):381–95. [9] Rossetto D, Avvakumov N, Côté J. Histone phosphorylation: a chromatin modification involved in diverse nuclear events. Epigenetics 2012;7(10):1098–108. [10] Puttick J, Baker EN, Delbaere LTJ. Histidine phosphorylation in biological systems. Biochim Biophys Acta 2008;1784(1):100–5. [11] Ellis P, Fagan BM, Magness ST, Hutton S, Taranova O, Hayashi S, et al. SOX2, a persistent marker for multipotential neural stem cells derived from embryonic stem cells, the embryo or the adult. Dev Neurosci 2004;26(2–4):148–65. [12] Gapp K, Woldemichael BT, Bohacek J, Mansuy IM. Epigenetic regulation in neurodevelopment and neurodegenerative diseases. Neuroscience 2014;264:99–111. [13] Hirabayashi Y, Gotoh Y. Epigenetic control of neural precursor cell fate during development. Nat Rev Neurosci 2010;11(6):377–88. [14] Wu H, Coskun V, Tao J, Xie W, Ge W, Yoshikawa K, et al. Dnmt3adependent nonpromoter DNA methylation facilitates transcription of neurogenic genes. Science 2010;329(5990):444–8. [15] Jakovcevski M, Akbarian S. Epigenetic mechanisms in neurological disease. Nat Med 2012;18(8):1194–204. [16] Azpurua J, Eaton BA. Neuronal epigenetics and the aging synapse. Front Cell Neurosci 2015;9:208. Available from: http:// www.ncbi.nlm.nih.gov/pmc/articles/PMC4444820/.

[17] Brookes E, Shi Y. Diverse epigenetic mechanisms of human disease. Annu Rev Genet 2014;48:237–68. [18] Ul Hassan A, Hassan G, Rasool Z. Role of stem cells in treatment of neurological disorder. Int J Health Sci (Qassim) 2009 Jul;3(2):227–33. [19] Srinageshwar B, Maiti P, Dunbar GL, Rossignol J. Role of epigenetics in stem cell proliferation and differentiation: implications for treating neurodegenerative diseases. Int J Mol Sci 2016;17(2):e199. [20] Vashishtha M, Ng CW, Yildirim F, Gipson TA, Kratter IH, Bodai L, et al. Targeting H3K4 trimethylation in Huntington disease. Proc Natl Acad Sci USA 2013;110(32):E3027–36. [21] Coppedè F. Genetics and epigenetics of Parkinson’s disease. ScientificWorldJournal 2012;2012:489830. [22] Kubota T, Miyake K, Hirasawa T. Role of epigenetics in Rett syndrome. Epigenomics 2013;5(5):583–92. [23] Helmlinger D, Hardy S, Sasorith S, Klein F, Robert F, Weber C, et al. Ataxin-7 is a subunit of GCN5 histone acetyltransferasecontaining complexes. Hum Mol Genet 2004;13(12):1257–65. [24] Chen I-C, Lin H-Y, Lee G-C, Kao S-H, Chen C-M, Wu Y-R, et al. Spinocerebellar ataxia type 8 larger triplet expansion alters histone modification and induces RNA foci. BMC Mol Biol 2009;10:9. [25] Küçükali CI˙, Kürtüncü M, Çoban A, Çebi M, Tüzün E. Epigenetics of multiple sclerosis: an updated review. Neuromolecular Med 2015;17(2):83–96. [26] Du C, Liu C, Kang J, Zhao G, Ye Z, Huang S, et al. MicroRNA miR-326 regulates TH-17 differentiation and is associated with the pathogenesis of multiple sclerosis. Nat Immunol 2009;10(12):1252–9. [27] Huynh JL, Casaccia P. Epigenetic mechanisms in multiple sclerosis: implications for pathogenesis and treatment. Lancet Neurol 2013;12(2):195–206. [28] Alberts B, Johnson A, Lewis J, Raff M, Roberts K, Walter P. Chromosomal DNA and its packaging in the chromatin fiber. 2002. Available from: http://www.ncbi.nlm.nih.gov/books/ NBK26834/ [29] Mariño-Ramírez L, Kann MG, Shoemaker BA, Landsman D. Histone structure and nucleosome stability. Expert Rev Proteomics 2005;2(5):719–29. [30] Hendrich B, Bickmore W. Human diseases with underlying defects in chromatin structure and modification. Hum Mol Genet 2001;10(20):2233–42. [31] Roos RA. Huntington’s disease: a clinical review. Orphanet J Rare Dis 2010;5:40. [32] Walker FO. Huntington’s disease. Lancet 2007;369(9557):218–28. [33] The Huntington’s disease collaborative research group. A novel gene containing a trinucleotide repeat that is expanded and unstable on Huntington’s disease chromosomes. Cell 1993;72(6):971–83. [34] Dey ND, Bombard MC, Roland BP, Davidson S, Lu M, Rossignol J, et al. Genetically engineered mesenchymal stem cells reduce behavioral deficits in the YAC 128 mouse model of Huntington’s disease. Behav Brain Res 2010;214(2):193–200. [35] Fink KD, Rossignol J, Crane AT, Davis KK, Bombard MC, Bavar AM, et al. Transplantation of umbilical cord-derived mesenchymal stem cells into the striata of R6/2 mice: behavioral and neuropathological analysis. Stem Cell Res Ther 2013;4(5):130. [36] Rossignol J, Fink KD, Crane AT, Davis KK, Bombard MC, Clerc S, et al. Reductions in behavioral deficits and neuropathology in the R6/2 mouse model of Huntington’s disease following transplantation of bone-marrow-derived mesenchymal stem cells is dependent on passage number. Stem Cell Res Ther 2015;6:9. [37] Urdinguio RG, Sanchez-Mut JV, Esteller M. Epigenetic mechanisms in neurological diseases: genes, syndromes, and therapies. Lancet Neurol 2009;8(11):1056–72.

V.  Factors Influencing Epigenetic Changes



REFERENCES

[38] Ng CW, Yildirim F, Yap YS, Dalin S, Matthews BJ, Velez PJ, et al. Extensive changes in DNA methylation are associated with expression of mutant huntingtin. Proc Natl Acad Sci USA 2013;110(6):2354–9. [39] Lee J, Hwang YJ, Kim KY, Kowall NW, Ryu H. Epigenetic mechanisms of neurodegeneration in Huntington’s disease. Neurotherapeutics 2013 Oct;10(4):664–76. [40] Lee S-T, Chu K, Im W-S, Yoon H-J, Im J-Y, Park J-E, et al. Altered microRNA regulation in Huntington’s disease models. Exp Neurol 2011;227(1):172–9. [41] Moumné L, Betuing S, Caboche J. Multiple aspects of gene dysregulation in Huntington’s disease. Front Neurol 2013;4:127. [42] Bithell A, Johnson R, Buckley NJ. Transcriptional dysregulation of coding and non-coding genes in cellular models of Huntington’s disease. Biochem Soc Trans 2009;37(Pt 6):1270–5. [43] Kern S, Eichler H, Stoeve J, Klüter H, Bieback K. Comparative analysis of mesenchymal stem cells from bone marrow, umbilical cord blood, or adipose tissue. Stem Cells 2006;24(5):1294–301. [44] Li Z, Liu C, Xie Z, Song P, Zhao RCH, Guo L, et al. Epigenetic dysregulation in mesenchymal stem cell aging and spontaneous differentiation. PLoS One 2011;6(6):e20526. [45] Phinney DG. Functional heterogeneity of mesenchymal stem cells: implications for cell therapy. J Cell Biochem 2012;113(9):2806–12. [46] Guo L, Zhou Y, Wang S, Wu Y. Epigenetic changes of mesenchymal stem cells in three-dimensional (3D) spheroids. J Cell Mol Med 2014;18(10):2009–19. [47] Moll G, Rasmusson-Duprez I, von Bahr L, Connolly-Andersen A-M, Elgue G, Funke L, et al. Are therapeutic human mesenchymal stromal cells compatible with human blood? Stem Cells 2012;30(7):1565–74. [48] Sadri-Vakili G, Cha J-HJ. Histone deacetylase inhibitors: a novel therapeutic approach to Huntington’s disease (complex mechanism of neuronal death). Curr Alzheimer Res 2006;3(4):403–8. [49] Han F, Baremberg D, Gao J, Duan J, Lu X, Zhang N, et al. Development of stem cell-based therapy for Parkinson’s disease. Translational Neurodegeneration 2015;4:16. [50] Klein C, Westenberger A. Genetics of Parkinson’s disease. Cold Spring Harb Perspect Med 2012;2(1). Available from: http:// www.ncbi.nlm.nih.gov/pmc/articles/PMC3253033/ [51] Lesage S, Brice A. Parkinson’s disease: from monogenic forms to genetic susceptibility factors. Hum Mol Genet 2009;18(R1) :R48–59. [52] Mentch SJ, Locasale JW. One-carbon metabolism and epigenetics: understanding the specificity. Ann NY Acad Sci 2016;1363:91–8. [53] King WD, Ho V, Dodds L, Perkins SL, Casson RI, Massey TE. Relationships among biomarkers of one-carbon metabolism. Mol Biol Rep 2012;39(7):7805–12. [54] Chou C-H, Fan H-C, Hueng D-Y, Chou C-H, Fan H-C, Hueng D-Y. Potential of neural stem cell-based therapy for Parkinson’s disease. Parkinson’s Dis 2015;2015:e571475. 2015,. [55] Matsumoto L, Takuma H, Tamaoka A, Kurisaki H, Date H, Tsuji S, et al. CpG demethylation enhances alpha-synuclein expression and affects the pathogenesis of Parkinson’s disease. PLoS One 2010;5(11):e15522. [56] Jowaed A, Schmitt I, Kaut O, Wüllner U. Methylation regulates alpha-synuclein expression and is decreased in Parkinson’s disease patients’ brains. J Neurosci 2010;30(18):6355–9. [57] Doxakis E. Post-transcriptional regulation of alpha-synuclein expression by mir-7 and mir-153. J Biol Chem 2010;285(17):12726– 34. [58] Cai M, Tian J, Zhao G, Luo W, Zhang B. Study of methylation levels of parkin gene promoter in Parkinson’s disease patients. Int J Neurosci 2011;121(9):497–502. [59] Fitzmaurice AG, Rhodes SL, Lulla A, Murphy NP, Lam HA, O’Donnell KC, et al. Aldehyde dehydrogenase inhibition as a

337

pathogenic mechanism in Parkinson disease. Proc Natl Acad Sci USA 2013;110(2):636–41. [60] Sanberg PR. Neural stem cells for Parkinson’s disease: To protect and repair. PNAS 2007;104(29):11869–70. [61] Ma DK, Marchetto MC, Guo JU, Ming G, Gage FH, Song H. Epigenetic choreographers of neurogenesis in the adult mammalian brain. Nat Neurosci 2010;13(11):1338–44. [62] Redmond DE, Bjugstad KB, Teng YD, Ourednik V, Ourednik J, Wakeman DR, et al. Behavioral improvement in a primate Parkinson’s model is associated with multiple homeostatic effects of human neural stem cells. Proc Natl Acad Sci USA 2007 Jul 17;104(29):12175–80. [63] Gash DM, Zhang Z, Ovadia A, Cass WA, Yi A, Simmerman L, et al. Functional recovery in parkinsonian monkeys treated with GDNF. Nature 1996;380(6571):252–5. [64] Sullivan AM, Opacka-Juffry J, Blunt SB. Long-term protection of the rat nigrostriatal dopaminergic system by glial cell linederived neurotrophic factor against 6-hydroxydopamine in vivo. Eur J Neurosci 1998;10(1):57–63. [65] Kirik D, Georgievska B, Björklund A. Localized striatal delivery of GDNF as a treatment for Parkinson disease. Nat Neurosci 2004;7(2):105–10. [66] d’Anglemont de Tassigny X, Pascual A, López-Barneo J. GDNFbased therapies, GDNF-producing interneurons, and trophic support of the dopaminergic nigrostriatal pathway. Implications for Parkinson’s disease. Front Neuroanat 2015;9:10. [67] Patel NK, Gill SS. GDNF delivery for Parkinson’s disease. Acta Neurochir Suppl 2007;97(Pt 2):135–54. [68] Deng X, Liang Y, Lu H, Yang Z, Liu R’en, Wang J, et al. Cotransplantation of GDNF-overexpressing neural stem cells and fetal dopaminergic neurons mitigates motor symptoms in a rat model of Parkinson’s disease. PLoS One 2013;8(12):e80880. [69] Wagner J, Akerud P, Castro DS, Holm PC, Canals JM, Snyder EY, et al. Induction of a midbrain dopaminergic phenotype in Nurr1-overexpressing neural stem cells by type 1 astrocytes. Nat Biotechnol 1999;17(7):653–9. [70] van Heesbeen HJ, Mesman S, Veenvliet JV, Smidt MP. Epigenetic mechanisms in the development and maintenance of dopaminergic neurons. Development 2013;140(6):1159–69. [71] Jacobs FMJ, van Erp S, van der Linden AJA, von Oerthel L, Burbach JPH, Smidt MP. Pitx3 potentiates Nurr1 in dopamine neuron terminal differentiation through release of SMRT-mediated repression. Development 2009;136(4):531–40. [72] Amir RE, Van den Veyver IB, Wan M, Tran CQ, Francke U, Zoghbi HY. Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein 2. Nat Genet 1999;23(2):185–8. [73] Trappe R, Laccone F, Cobilanschi J, Meins M, Huppke P, Hanefeld F, et al. MECP2 mutations in sporadic cases of Rett syndrome are almost exclusively of paternal origin. Am J Hum Genet 2001;68(5):1093–101. [74] Girard M, Couvert P, Carrié A, Tardieu M, Chelly J, Beldjord C, et al. Parental origin of de novo MECP2 mutations in Rett syndrome. Eur J Hum Genet 2001;9(3):231–6. [75] Zhang X, Zhao Y, Bao X, Zhang J, Cao G, Wu X. Genetic features and mechanism of Rett syndrome in Chinese population. Zhonghua Yi Xue Yi Chuan Xue Za Zhi 2014;31(1):1–5. [76] Bienvenu T, Chelly J. Molecular genetics of Rett syndrome: when DNA methylation goes unrecognized. Nat Rev Genet 2006;7(6):415–26. [77] Yasui DH, Peddada S, Bieda MC, Vallero RO, Hogart A, Nagarajan RP, et al. Integrated epigenomic analyses of neuronal MeCP2 reveal a role for long-range interaction with active genes. Proc Natl Acad Sci USA 2007;104(49):19416–21. [78] Mellen M, Ayata P, Dewell S, Kriaucionis S, Heintz N. MeCP2 binds to 5hmc enriched within active genes and accessible chromatin in the nervous system. Cell. 2012;151(7):1417–30.

V.  Factors Influencing Epigenetic Changes

338

21.  Epigenetics, Stem Cells, Cellular Differentiation, and Associated Hereditary Neurological Disorders

[79] Dajani R, Koo S-E, Sullivan GJ, Park I-H. Investigation of Rett syndrome using pluripotent stem cells. J Cell Biochem 2013;114(11):2446–53. [80] Takahashi K, Yamanaka S. Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 2006;126(4):663–76. [81] Marchetto MCN, Carromeu C, Acab A, Yu D, Yeo GW, Mu Y, et al. A model for neural development and treatment of Rett syndrome using human induced pluripotent stem cells. Cell 2010;143(4):527–39. [82] Kim K-Y, Hysolli E, Park I-H. Neuronal maturation defect in induced pluripotent stem cells from patients with Rett syndrome. Proc Natl Acad Sci USA 2011;108(34):14169–74. [83] Teive HAG. Spinocerebellar ataxias. Arq Neuropsiquiatr 2009;67(4):1133–42. [84] Paulson HL. The Spinocerebellar Ataxias. J Neuroophthalmol 2009;29(3):227–37. [85] Lin X, Antalffy B, Kang D, Orr HT, Zoghbi HY. Polyglutamine expansion down-regulates specific neuronal genes before pathologic changes in SCA1. Nat Neurosci 2000;3(2):157–63. [86] Gehrking KM, Andresen JM, Duvick L, Lough J, Zoghbi HY, Orr HT. Partial loss of Tip60 slows mid-stage neurodegeneration in a spinocerebellar ataxia type 1 (SCA1) mouse model. Hum Mol Genet 2011;20(11):2204–12. [87] Ogawa M. Pharmacological treatments of cerebellar ataxia. Cerebellum 2004;3(2):107–11. [88] Zhang M-J, Sun J-J, Qian L, Liu Z, Zhang Z, Cao W, et al. Human umbilical mesenchymal stem cells enhance the expression of neurotrophic factors and protect ataxic mice. Brain Res 2011;1402:122–31. [89] Jin J-L, Liu Z, Lu Z-J, Guan D-N, Wang C, Chen Z-B, et al. Safety and efficacy of umbilical cord mesenchymal stem cell therapy in hereditary spinocerebellar ataxia. Curr Neurovasc Res 2013;10(1):11–20. [90] Dongmei H, Jing L, Mei X, Ling Z, Hongmin Y, Zhidong W, et al. Clinical analysis of the treatment of spinocerebellar ataxia and multiple system atrophy-cerebellar type with umbilical cord mesenchymal stromal cells. Cytotherapy 2011;13(8):913–7. [91] Bakhuraysah MM, Siatskas C, Petratos S. Hematopoietic stem cell transplantation for multiple sclerosis: is it a clinical reality? Stem Cell Res Ther 2016;7:12. Available from: http://www.ncbi. nlm.nih.gov/pmc/articles/PMC4715306/

[92] Koch MW, Metz LM, Kovalchuk O. Epigenetic changes in patients with multiple sclerosis. Nat Rev Neurol 2013;9(1):35–43. [93] Baranzini SE, Mudge J, van Velkinburgh JC, Khankhanian P, Khrebtukova I, Miller NA, et al. Genome, epigenome and RNA sequences of monozygotic twins discordant for multiple sclerosis. Nature 2010;464(7293):1351–6. [94] Liggett T, Melnikov A, Tilwalli S, Yi Q, Chen H, Replogle C, et al. Methylation patterns of cell-free plasma DNA in relapsingremitting multiple sclerosis. J Neurol Sci 2010;290(1–2):16–21. [95] Kürtüncü M, Tüzün E. Multiple sclerosis: could it be an epigenetic disease? Med Hypotheses 2008;71(6):945–7. [96] Mastronardi FG, Noor A, Wood DD, Paton T, Moscarello MA. Peptidyl argininedeiminase 2 CpG island in multiple sclerosis white matter is hypomethylated. J Neurosci Res 2007;85(9):2006–16. [97] Seita J, Weissman IL. Hematopoietic stem cell: self-renewal versus differentiation. Wiley Interdiscip Rev Syst Biol Med 2010;2(6):640–53. [98] Burt RK, Traynor AE, Cohen B, Karlin KH, Davis FA, Stefoski D, et al. T cell-depleted autologous hematopoietic stem cell transplantation for multiple sclerosis: report on the first three patients. Bone Marrow Transplant 1998;21(6):537–41. [99] Burt RK, Burns W, Hess A. Bone marrow transplantation for multiple sclerosis. Bone Marrow Transplant 1995;16(1):1–6. [100] Mandalfino P, Rice G, Smith A, Klein JL, Rystedt L, Ebers GC. Bone marrow transplantation in multiple sclerosis. J Neurol 2000;247(9):691–5. [101] Atkins HL, Freedman MS. Hematopoietic stem cell therapy for multiple sclerosis: top 10 lessons learned. Neurotherapeutics 2013;10(1):68–76. [102] Mancardi GL, Sormani MP, Gualandi F, Saiz A, Carreras E, Merelli E, et al. Autologous hematopoietic stem cell transplantation in multiple sclerosis: a phase II trial. Neurology 2015;84(10):981–8.

Further Reading Coppede F. Genetics and epigenetics of Parkinson’s disease. ScientificWorldJournal 2012;2012:e489830.

V.  Factors Influencing Epigenetic Changes

C H A P T E R

22 Epigenetic Regulation of Neuron’s Regenerative Abilities After Spinal Cord Injury Michael I. Shifman Shriners Hospitals Pediatric Research Center (Center for Neural Repair and Rehabilitation), Philadelphia, PA, United States; Temple University School of Medicine, Philadelphia, PA, United States

O U T L I N E Introduction

339

DNA Methylation in Axon Growth and Regeneration

340

Histone Posttranslational Modifications in Axon Growth and Regeneration

341

Advantages of Studying Axonal Regeneration in Larval Lamprey CNS

342

Changes in HATs and HDACs mRNA Expression in Lamprey CNS After Spinal Cord Transection

342

INTRODUCTION Spinal cord injury (SCI) is a devastating clinical problem manifested by irreversible paralysis below the level of injury. The main cause of disability is interruption of axon tracts connecting different levels of the spinal cord with each other and with the brain. Although axons attempt to regrow, regeneration, and synaptic reconnection is not successful in mammals, including humans. Previous work has shown that successful axon regeneration is dependent upon the transcription of a large number of regeneration-associated genes (RAGs) and transcription factors (TF) [1]. Current thinking in the field of axonal regeneration is that one of the major Handbook of Epigenetics. http://dx.doi.org/10.1016/B978-0-12-805388-1.00022-5 Copyright © 2017 Elsevier Inc. All rights reserved.

Chromatin-Independent Effects of Protein Acetylation on Axon Growth and Regeneration

342

The Role of Long Noncoding RNAs (lncRNAs) in Nerve Regeneration

343

Conclusions

344

References

345

differences in regenerative potential between the regenerating peripheral nervous system (PNS) neuron and the quiescent central nervous system (CNS) neuron may reflect major differences in intrinsic transcriptional networks, rather than individual genes [2,3]. Indeed, recent large-scale screening approaches using a combination of microarrays with phosphoproteomics have identified 39 injury-inducible TF [4] while the comparative transcriptome microarray analysis of peripheral versus central nerve injured models of dorsal root ganglion (DRG) neurons found 30 candidate TFs [5]. These injury-inducible TFs are presumed to control hundreds of transcriptional targets of multiple regeneration-associated signaling pathways [3].

339

340

22.  Epigenetic Regulation of Neuron’s Regenerative Abilities After Spinal Cord Injury

While adult mammalian CNS neurons have limited regenerative ability, developing CNS neurons regenerate axons after injury [6,7]. Therefore, it was hypothesized that during development, neurons undergo a transcriptionally regulated switch that limits transcription of a large number of RAGs and TF and, therefore, their regenerative capacity [3,8]. The process of gene expression in eukaryotic cells is governed by the action of the transcriptional machinery of the cell (RNA polymerases, TF, and chromatin remodeling enzymes). There are two primary and interconnected epigenetic mechanisms—DNA methylation and covalent modification of histones—that result in changes in the chromatin structure, which in turn influence gene transcription. In addition, it has become apparent in the past few years that noncoding RNA is also intimately involved in this process. Chromatin-based epigenetic mechanisms underlie important aspects of CNS function [9,10], but little is known about the role of epigenetic alterations in the response to SCI. But this situation is beginning to change and several review papers were published recently to alleviate the shortage of reliable information concerning the role of epigenetic modifications in SCI mechanisms [11–15].

DNA METHYLATION IN AXON GROWTH AND REGENERATION DNA methylation is an epigenetic modification critical to normal genome functions and regulation and also a major modulator of gene expression. In the mammalian genome, about 70% of CpG dinucleotides are methylated. Many of the remaining nonmethylated CpGs are in CpG islands typically found in functional promoter regions [16]. DNA methylation has long been viewed as an epigenetic marker of gene repression, and genes with abundant 5-methylcytosine in their promoter region are usually transcriptionally silent. DNA methylation is carried out by a family of enzymes called DNA methyltransferases (DNMTs). DNMTs transfer a methyl group from S-adenosyl-methionine to cytosine in CpG dinucleotides. Three DNMTs (DNMT1, DNMT3a, and DNMT3b) are required to establish and maintain DNA methylation patterns. DNMT1 maintains the pattern of DNA methylation (i.e., methylation of hemimethylated sites) while DNMT3a and 3b establish new or de novo DNA methylation patterns. The vitamin folate pathway is a key player in mechanisms of DNA methylation [17]. The first report that folic acid is effective for enhancing growth, repair, and recovery in the injured adult CNS indicated that intraperitoneal treatment of adult rats with folic acid significantly improves regrowth of sensory spinal axons into a grafted segment of peripheral nerve in vivo [18]. Retinal

ganglion cell (RGC) axon growth into a graft is also enhanced, though less than spinal axons. Moreover, folic acid treatment enhances neurological recovery from SCI. Subsequent research from the same group focused on potential mechanisms of folic acid treatment. As DNMTs directly methylate CpG dinucleotides in the mammalian genome, the authors studied the contributions of the several DNMTs to CNS injury and repair [19]. DNMT1, DNMT3a, and DNMT3b protein levels were measured by Western blot prior to and following combined spinal cord and peripheral nerve injury. DNMT1 protein level was unchanged after this injury, but DNMT3a and DNMT3b protein levels were downregulated in the spinal cord; folic acid treatment restored both DNMT3a and DNMT3b to control levels. Combined spinal cord and peripheral nerve injury decreased global genomic methylation by 40%, an effect that folic acid treatment prevented. This data has yet to be reconciled with the knowledge that DNA methylation, particularly of CpG islands colocalized with promoters or other regulatory regions, can repress transcription. This repression could be either direct, by precluding transcriptional activators, or indirect, by promoting recruitment of repressor complexes that usually contain histone deacetylases (HDACs) [10]. It is plausible that if SCI leads to hypomethylation, then this will enhance gene expression of RAGs and regeneration. Moreover, significance of this data was decreased somehow by the fact that authors used whole spinal cord for their analysis, therefore important information concerning cell-type-specific (neurons, glia, nonneuronal cells) DNA methylation effect on SCI was missing. Therefore, conclusion by authors that folic acid inhibits a hypomethylation response caused by injury, and DNA methylation, together with DNMT3a and DNMT3b protein levels, correlate with the ability of spinal neurons to regenerate their injured axons [19] need to be verified by analyses of specific cell types. Indeed, a recent quantitative reverse transcription polymerase chain reaction (RT–PCR) analysis of differentially methylated genes and expression of DNA methyltransferases did not find a consistent correlation between DNA methylation levels and gene expression in DRG neurons after injury [20]. Moreover, DNA methylation arrays displayed only a modest number of genes differentially methylated in DRG neurons following peripheral sciatic nerve axotomy (SNA) as compared to central dorsal column axotomy (DCA). In addition, none of the RAG genes (Gap-43, Galanin, Sprr1a, and BDNF) exhibited significant levels of methylation nor were they differentially methylated by SNA and DCA. Therefore, the authors concluded that DNA methylation is not a key factor in the differential regenerative response between CNS and PNS injuries of the DRG [20]. More research will be necessary to clarify the role of DNA methylation in the regenerative response after injury.

V.  Factors Influencing Epigenetic Changes



Histone Posttranslational Modifications in Axon Growth and Regeneration

HISTONE POSTTRANSLATIONAL MODIFICATIONS IN AXON GROWTH AND REGENERATION Histones pack DNA into nucleosomes, the building blocks of chromatin. Histones are characterized by a large number of posttranslational modifications, which divide the genome into active regions (euchromatin), where DNA is accessible for transcription, and inactive regions (heterochromatin), where DNA is more compact and therefore less accessible for transcription. At least nine different types of histone modifications have been described, each catalyzed by a specific set of enzymes. The best understood modifications are lysine acetylation, lysine and arginine methylation, serine/threonine/ tyrosine phosphorylation, and serine/threonine ubiquitylation. Other modifications include GlcNAcylation, citrullination, krotonilation, and proline isomerization. Acetylation, one of the most widely studied histone modifications, is linked to transcriptional regulation. The enzymes responsible for regulating the acetylation of histone tails are histone acetyltransferases (HATs), which add acetyl groups to lysine residues, and HDACs, which remove the acetyl groups [21,22]. In mammals, Class I HDACs are ubiquitously expressed nuclear enzymes, and Class II HDACs, including Class IIb HDACs, shuttle between nucleus and cytoplasm. Class IIa HDACs are signal transducers characterized by the presence in their regulatory N-terminal domains of two or three conserved serine residues subject to reversible phosphorylation. Phosphorylation leads to binding of the 14-3–3 proteins, nuclear export of HDACs and derepression of their target genes. HDAC6 is an α-tubulin deacetylase and thus involved in cytoskeleton regulation [23,24]. As axons fail to regenerate in the CNS, the vast majority of experiments studying the role of epigenetic histone modifications in axonal regeneration have examined PNS neurons. Direct pharmacological treatment of dissociated murine DRG neurons with HDAC inhibitor MS-275 produced no significant differences in the mean axonal length, or the percentage of axon-bearing neurons compared with controls. However, in vivo injections of HDAC inhibitors (TSA or MS-275) to mice caused DRG neurons to grow longer axons than vehicle-treated mice [25]. This result might be due to nonspecific effects of the HDAC inhibitors [21]. In marked contrast to results obtained with DRG culture, inhibition of HDACs I and II by broad-spectrum inhibitor TSA in cultured cerebellar granule neurons significantly increased average neurite length and numbers, and reduced growth cone collapse on both “permissive” (poly-d-lysine) and “nonpermissive” substrates (myelin and chondroitin sulphate) [26]. Similarly, treatment with an HDAC inhibitor, valproic acid (VPA), enabled CNS neurons

341

cultured from embryonic spinal cord or hippocampus to overcome Nogo-A inhibition of neurite outgrowth. Interestingly, the Cavalli laboratory recently reported that peripheral axotomy of sensory neurons caused a calcium wave to travel retrogradely from the injury site to the soma, where it induced nuclear export of HDAC5 that enhanced histone acetylation and activated a proregenerative gene-expression program. This HDAC5 pathway functioned only in injured PNS and was not activated by a CNS injury. Increasing HDAC5 nuclear export promotes axon regeneration in vitro and in vivo, whereas overexpression of a nuclear-trapped HDAC5 mutant prevents axon regeneration [27]. Several line of evidence points out to the fact that different classes of neurons may use different epigenetic mechanisms to regulate axon growth potential. Recent work has shown that the levels of acetylated histone 4 (AcH4) were low in control, uninjured adult DRG neurons but markedly increased in DRG neurons conditioned by peripheral axotomy. Interestingly, AcH4 was selectively enriched at the promoters of a subgroup of early RAGs (Smad1, Atf3, Sprr1a, Npy, and Galanin), but not of housekeeping genes (Gapdh and Rpl13a) or of a RAG induced later (Gap 43) [25]. In vivo axotomy of facial motor neurons led to formation of a transcriptional complex between p53 and acetyltransferases CBP/p300 that acetylated p53 at K372-382, upregulated axonal GAP 43, and enhanced axon outgrowth [28]. Moreover, initiation of transcription of both CBP and p53 indicated that the p53/GAP-43 transcriptional module was specifically switched on during axon regeneration in vivo [28]. Very different results were obtained in the injured CNS: SCI led to marked reduction in histone acetylation. Levels of acetylated histone H3 (Ac-H3) and H4 (Ac-H4) were significantly decreased as early as first day after SCI in rat, and remained below those in uninjured controls for at least 2 weeks after SCI. Injection of HDAC inhibitor, VPA prevented these reductions of Ac-H3 and Ac-H4, reduced apoptosis and promoted locomotor recovery [29]. Similarly, treatment with an HDAC inhibitor, VPA, enabled CNS neurons cultured from embryonic spinal cord or hippocampus to overcome Nogo-A inhibition of neurite outgrowth. Moreover, VPA treatment augmented the level of histone acetylation and expression of BDNF in spinal neurons [30]. In addition, experiments conducted by Di Giovanni and coworkers showed that hyperacetylation induced by HDAC I/II inhibition stimulated neurite outgrowth of primary cerebellar granule neurons (CGN) by enhancing gene expression on the HATs (CBP/p300 and P/CAF) promoters by acetylation of histone H3 and by enhancing p53-dependent promoter reactivity that in turn increased expression of axonal progrowth genes [26]. If we summarize experimental finding discussed in this section, it is becoming clear that histone acetylation

V.  Factors Influencing Epigenetic Changes

342

22.  Epigenetic Regulation of Neuron’s Regenerative Abilities After Spinal Cord Injury

is appearing to play an important role in PNS regeneration. Injury to PNS neurons lead to the enhanced histone acetylation and activation a proregenerative gene-expression programs. Those acetylation pathways operated only in injured PNS and not activated by a CNS injury that may contributes to CNS regenerative failure. Therefore, modulating HDACs activity through pharmacological modulation with specific inhibitors might promote regeneration of CNS neurons after SCI.

ADVANTAGES OF STUDYING AXONAL REGENERATION IN LARVAL LAMPREY CNS Whereas mammals and other high vertebrates undergo a developmental switch in which growth-associated genes are transcriptionally silenced, in fish many of these genes are reactivated upon CNS injury. The sea lamprey is an early evolved vertebrate that recovers behaviorally following SCI [31,32] and in which axons regenerate selectively in their correct paths [33,34]. Yet the large, identified reticulospinal (RS) neurons in the lamprey brain vary greatly in their regenerative abilities: some are good regenerators (probability of regeneration >50%) and others are bad regenerators (