Circular RNAs [2 ed.] 1071636774, 9781071636770

Citation preview

Methods in Molecular Biology 2765

Christoph Dieterich Marie-Laure Baudet  Editors

Circular RNAs Second Edition

METHODS

IN

MOLECULAR BIOLOGY

Series Editor John M. Walker School of Life and Medical Sciences University of Hertfordshire Hatfield, Hertfordshire, UK

For further volumes: http://www.springer.com/series/7651

For over 35 years, biological scientists have come to rely on the research protocols and methodologies in the critically acclaimed Methods in Molecular Biology series. The series was the first to introduce the step-by-step protocols approach that has become the standard in all biomedical protocol publishing. Each protocol is provided in readily-reproducible step-bystep fashion, opening with an introductory overview, a list of the materials and reagents needed to complete the experiment, and followed by a detailed procedure that is supported with a helpful notes section offering tips and tricks of the trade as well as troubleshooting advice. These hallmark features were introduced by series editor Dr. John Walker and constitute the key ingredient in each and every volume of the Methods in Molecular Biology series. Tested and trusted, comprehensive and reliable, all protocols from the series are indexed in PubMed.

Circular RNAs Second Edition

Edited by

Christoph Dieterich Klaus Tschira Institute for Integrative Computational Cardiology, University Heidelberg, Heidelberg, Germany; Department of Internal Medicine III (Cardiology, Angiology, and Pneumology), University Hospital Heidelberg, Heidelberg, Germany; German Centre for Cardiovascular Research (DZHK)-Partner Site Heidelberg/Mannheim, Heidelberg, Germany

Marie-Laure Baudet Department of Cellular, Computational and Integrative Biology – CIBIO, University of Trento, Trento, Italy

Editors Christoph Dieterich Klaus Tschira Institute for Integrative Computational Cardiology University Heidelberg Heidelberg, Germany

Marie-Laure Baudet Department of Cellular, Computational and Integrative Biology – CIBIO University of Trento Trento, Italy

Department of Internal Medicine III (Cardiology, Angiology, and Pneumology) University Hospital Heidelberg Heidelberg, Germany German Centre for Cardiovascular Research (DZHK)-Partner Site Heidelberg/Mannheim Heidelberg, Germany

ISSN 1064-3745 ISSN 1940-6029 (electronic) Methods in Molecular Biology ISBN 978-1-0716-3677-0 ISBN 978-1-0716-3678-7 (eBook) https://doi.org/10.1007/978-1-0716-3678-7 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Cover designed by Barbara Poli from an original idea by Efrem Bertini This Humana imprint is published by the registered company Springer Science+Business Media, LLC, part of Springer Nature. The registered company address is: 1 New York Plaza, New York, NY 10004, U.S.A. Paper in this product is recyclable.

Preface In the rapidly evolving field of molecular biology, a captivating class of non-coding RNAs has emerged as a subject of intense research and fascination—Circular RNAs (circRNAs). These enigmatic molecules, once considered mere byproducts of splicing errors, have now emerged as pivotal players in diverse cellular processes, challenging conventional paradigms of genetic regulation. The current literature highlights the role of circular RNAs for diagnostic purposes, as therapeutic targets and tools, revolutionizing our understanding of both normal physiology and disease pathogenesis. Total transcriptome sequencing has rediscovered circular RNAs by identifying their characteristic back-splicing junctions across all eukaryotes (protists, fungi, plants, animals). The advent of tailored algorithmic approaches to mine the growing database of RNA-seq data along with better protocols to isolate circular RNAs led to comprehensive catalogs and circular RNA inventories. Although we have made significant progress in identifying characteristic back-splicing junctions through RNA-sequencing data [1], many challenges remain concerning circular RNAs’ form and function. The current volume addresses some of these challenges, providing up-to-date answers on circular RNA purification (Chap. 1) and in silico characterization (Chaps. 2 and 3). Moreover, recent research attention has shifted toward better characterizing circular RNAs in terms of localization, digital counting, primary sequence, and secondary structure. Chapters 4–8 present up-to-date protocols and approaches for circRNA detection, sequence validation, and quantification. Additionally, significant advancements have been achieved in creating tools to study the functional relevance of circular RNAs [2, 3]. A substantial part of this volume is devoted to techniques related to gain- and loss-of-function approaches (see Chaps. 9–14). This also encompasses circular RNA synthesis and application by different means, e.g., split ligation, engineering, and nanoparticle packaging. Chapter 15 delves into the effect of RNA modifications on circular RNA biogenesis and RNA translation potential, whereas the last part of the book (Chaps. 16 and 17) focuses on interaction partners of circular RNAs and neuronal phenotypes of circular RNA perturbations in dedicated cell culture systems. The book’s overarching goal is to provide a comprehensive toolkit for studying circRNAs, spanning their biogenesis, regulatory mechanisms, functions, and emerging roles in health and disease. Most intriguing, first vaccines based on circular RNAs and several other applications clearly demonstrate its potential as a biomedical tool [4, 5].

v

vi

Preface

The journey into the world of circular RNAs has been made possible through the contributions of more than 50 scientists who authored this new edition of Methods in Molecular Biology: Molecular and Functional Dissection of Circular RNAs. Their passion for unraveling the complexities of circular RNA life has allowed us to compile this second volume to advance circular RNA research. Join us as we explore the remarkable world of circular RNAs! Heidelberg, Germany Trento, Italy

Christoph Dieterich Marie-Laure Baudet

References 1. Vromman M, Anckaert J, Bortoluzzi S, Buratin A, Chen CY, Chu Q, Chuang TJ, Dehghannasiri R, Dieterich C, Dong X, Flicek P, Gaffo E, Gu W, He C, Hoffmann S, Izuogu O, Jackson MS, Jakobi T, Lai EC, Nuytens J, Salzman J, Santibanez-Koref M, Stadler P, Thas O, Vanden Eynde E, Verniers K, Wen G, Westholm J, Yang L, Ye CY, Yigit N, Yuan GH, Zhang J, Zhao F, Vandesompele J, Volders PJ (2023) Large-scale benchmarking of circRNA detection tools reveals large differences in sensitivity but not in precision. Nat Methods 20(8):1159–1169. https://doi.org/10.1038/s41592-023-01944-6. Epub 2023 Jul 13. PMID: 37443337 2. Liu CX, Chen LL (2022) Circular RNAs: characterization, cellular roles, and applications. Cell 185(12):2016–2034. https://doi.org/10.1016/j.cell.2022.04.021. Epub 2022 May 17. Erratum in: Cell 185(13):2390. PMID: 35584701 3. Nielsen AF, Bindereif A, Bozzoni I, Hanan M, Hansen TB, Irimia M, Kadener S, Kristensen LS, Legnini I, Morlando M, Jarlstad Olesen MT, Pasterkamp RJ, Preibisch S, Rajewsky N, Suenkel C, Kjems J (2022) Best practice standards for circular RNA research. Nat Methods 19(10):1208–1220. https://doi.org/10.1038/s41592-022-01487-2. Epub 2022 May 26. PMID: 35618955; PMCID: PMC9759028 4. Chen R, Wang SK, Belk JA, Amaya L, Li Z, Cardenas A, Abe BT, Chen CK, Wender PA, Chang HY (2023) Author Correction: Engineering circular RNA for enhanced protein production. Nat Biotechnol 41(2):293. https://doi.org/10.1038/s41587-022-01472-2. Erratum for: Nat Biotechnol. 2023 Feb;41(2):262–272. PMID: 35978134; PMCID: PMC9931571 5. Liu C, Shi Q, Huang X, Koo S, Kong N, Tao W (2023) mRNA-based cancer therapeutics. Nat Rev Cancer 23(8):526–543. https://doi.org/10.1038/s41568-023-00586-2. Epub 2023 Jun 13. PMID: 37311817

Contents Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Contributors. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

PART I

CIRCRNA

PURIFICATION

1 Purification of Circular RNAs Using Poly(A) Tailing Followed by RNase R Digestion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Mei-Sheng Xiao and Jeremy E. Wilusz

PART II

3

IN SILICO CIRCRNA PROFILING

2 State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Tobias Jakobi 3 Exploring Circular RNA Profile and Expression in Extracellular Vesicles . . . . . . . Jingjing Zhao, Qiaojuan Li, and Shenglin Huang

PART III

v ix

CIRCRNA AND

23 47

DETECTION, SEQUENCE VALIDATION

QUANTIFICATION

4 In Situ Hybridization of circRNAs in Cells and Tissues through BaseScope™ Strategy. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63 Eleonora D’Ambra, Erika Vitiello, Tiziana Santini, and Irene Bozzoni 5 Sanger Sequencing to Determine the Full-Length Sequence of Circular RNAs . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 93 Suman Singh, Aniruddha Das, and Amaresh C. Panda 6 Droplet Digital PCR for the Detection and Quantification of Bona Fide CircRNAs. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 107 Linda Masante, Giorgia Susin, and Marie-Laure Baudet 7 Targeted Sequencing of Circular RNAs for Illumina-Based Counting and Nanopore Structure Determination . . . . . . . . . . . . . . . . . . . . . . . . . . 127 Isabel S. Naarmann-de Vries and Christoph Dieterich 8 Nanopore-Mediated Sequencing of Circular RNA . . . . . . . . . . . . . . . . . . . . . . . . . . 143 Morten Trillingsgaard Venø, Junyi Su, Yan Yan, and Jørgen Kjems

PART IV

FUNCTIONAL STUDY: TOOLS FOR LOSS- AND GAIN-OF FUNCTION ANALYSIS

9 In Vivo Tissue-Specific Knockdown of circRNAs Using shRNAs in Drosophila melanogaster . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 161 Ines Lucia Patop, Michael Canori, and Sebastian Kadener

vii

viii

Contents

10

The Functional Circular RNA Screening via RfxCas13d/BSJ-gRNA System. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Hao Wu and Ling-Ling Chen 11 Base-Editor-Mediated circRNA Knockout by Targeting Predominantly Back-Splice Sites . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Xu-Kai Ma, Xiang Gao, Mei Cao, and Li Yang 12 Directed Circularization of a Short RNA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ¨ ller, Cathrin E. Hansen, Danilo Springstubbe, Sabine Mu and Sonja Petkovic 13 Engineering Synthetic circRNAs for Efficient CNS Expression . . . . . . . . . . . . . . . Katie N. Clements, Trevor J. Gonzalez, and Aravind Asokan 14 Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via Microfluidic Mixing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Malte Juchem, Sarah Cushman, Dongchao Lu, Shambhabi Chatterjee, a r, and Thomas Thum Christian B€

PART V 15

17

193 209

227

247

FUNCTIONAL STUDY: BIOGENESIS AND TRANSLATION

Characterizing Post-transcriptional Modifications of circRNAs to Investigate Biogenesis and Translation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 263 Gaia Di Timoteo, Dario Dattilo, and Irene Bozzoni

PART VI 16

173

CIRCRNA

MECHANISMS OF ACTION

Native Circular RNA Pulldown Method to Simultaneously Profile RNA and Protein Interactions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 299 Marta M. Gabryelska, Stuart T. Webb, He Lin, Laura Gantley, Kirsty Kirk, Ryan Liu, Brett W. Stringer, Vanessa M. Conn, and Simon J. Conn Screening and Characterization of Functional circRNAs in Neuronal Cultures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 311 Darren Kelly and Gerhard Schratt

Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

325

Contributors ARAVIND ASOKAN • Department of Surgery, Duke University School of Medicine, Durham, NC, USA; Department of Molecular Genetics and Microbiology, Duke University School of Medicine, Durham, NC, USA; Department of Biomedical Engineering, Duke University, Durham, NC, USA CHRISTIAN BA€ R • Institute of Molecular and Translational Therapeutic Strategies (IMTTS), Hannover Medical School, Hannover, Germany; Fraunhofer Institute for Toxicology and Experimental Medicine (ITEM), Hannover, Germany MARIE-LAURE BAUDET • Department of Cellular, Computational and Integrative Biology – CIBIO, University of Trento, Trento, Italy IRENE BOZZONI • Center for Life Nano- & Neuro-Science, Fondazione Istituto Italiano di Tecnologia (IIT), Rome, Italy; Center for Human Technology, Istituto Italiano di Tecnologia (IIT), Genoa, Italy; Department of Biology and Biotechnologies “Charles Darwin”, Sapienza University of Rome, Rome, Italy MICHAEL CANORI • Brandeis University, Waltham, MA, USA MEI CAO • Center for Molecular Medicine, Children’s Hospital, Fudan University and Shanghai Key Laboratory of Medical Epigenetics, International Laboratory of Medical Epigenetics and Metabolism, Institutes of Biomedical Sciences, Fudan University, Shanghai, China SHAMBHABI CHATTERJEE • Institute of Molecular and Translational Therapeutic Strategies (IMTTS), Hannover Medical School, Hannover, Germany; Fraunhofer Institute for Toxicology and Experimental Medicine (ITEM), Hannover, Germany LING-LING CHEN • Key Laboratory of RNA Science and Engineering, CAS Center for Excellence in Molecular Cell Science, Shanghai Institute of Biochemistry and Cell Biology, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China; School of Life Science and Technology, ShanghaiTech University, Shanghai, China; School of Life Science, Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences, Hangzhou, China KATIE N. CLEMENTS • Department of Surgery, Duke University School of Medicine, Durham, NC, USA SIMON J. CONN • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia VANESSA M. CONN • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia SARAH CUSHMAN • Institute of Molecular and Translational Therapeutic Strategies (IMTTS), Hannover Medical School, Hannover, Germany ELEONORA D’AMBRA • Center for Life Nano- & Neuro-Science, Fondazione Istituto Italiano di Tecnologia (IIT), Rome, Italy ANIRUDDHA DAS • Institute of Life Sciences, Nalco Square, Bhubaneswar, Odisha, India DARIO DATTILO • Department of Biology and Biotechnology Charles Darwin, Sapienza University of Rome, Rome, Italy GAIA DI TIMOTEO • Department of Biology and Biotechnology Charles Darwin, Sapienza University of Rome, Rome, Italy

ix

x

Contributors

CHRISTOPH DIETERICH • Klaus Tschira Institute for Integrative Computational Cardiology, University Heidelberg, Heidelberg, Germany; Department of Internal Medicine III (Cardiology, Angiology, and Pneumology), University Hospital Heidelberg, Heidelberg, Germany; German Centre for Cardiovascular Research (DZHK)-Partner Site Heidelberg/Mannheim, Heidelberg, Germany MARTA M. GABRYELSKA • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia LAURA GANTLEY • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia XIANG GAO • Key Laboratory of Systems Health Science of Zhejiang Province, School of Life Science, Hangzhou Institute for Advanced Study, University of Chinese Academy of Sciences, Hangzhou, China; State Key Laboratory of Molecular Biology, Shanghai Key Laboratory of Molecular Andrology, CAS Center for Excellence in Molecular Cell Science, Shanghai Institute of Biochemistry and Cell Biology, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China TREVOR J. GONZALEZ • Department of Molecular Genetics and Microbiology, Duke University School of Medicine, Durham, NC, USA CATHRIN E. HANSEN • University Hospital Schleswig-Holstein, Campus Lu¨beck, Germany SHENGLIN HUANG • Fudan University Shanghai Cancer Center and Institutes of Biomedical Sciences, Shanghai Medical College, Fudan University, Shanghai, China TOBIAS JAKOBI • Department of Internal Medicine and the Translational Cardiovascular Research Center, University of Arizona – College of Medicine – Phoenix, Phoenix, AZ, USA MALTE JUCHEM • Institute of Molecular and Translational Therapeutic Strategies (IMTTS), Hannover Medical School, Hannover, Germany; Fraunhofer Institute for Toxicology and Experimental Medicine (ITEM), Hannover, Germany SEBASTIAN KADENER • Brandeis University, Waltham, MA, USA DARREN KELLY • Lab of Systems Neuroscience, Institute for Neuroscience, Department of Health Science and Technology, Swiss Federal Institute of Technology ETH, Zurich, Switzerland KIRSTY KIRK • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia JØRGEN KJEMS • Interdisciplinary Nanoscience Center, Aarhus University, Aarhus, Denmark; Department of Molecular Biology and Genetics, Aarhus University, Aarhus, Denmark QIAOJUAN LI • Fudan University Shanghai Cancer Center and Institutes of Biomedical Sciences, Shanghai Medical College, Fudan University, Shanghai, China HE LIN • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia RYAN LIU • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia DONGCHAO LU • Institute of Molecular and Translational Therapeutic Strategies (IMTTS), Hannover Medical School, Hannover, Germany XU-KAI MA • Center for Molecular Medicine, Children’s Hospital, Fudan University and Shanghai Key Laboratory of Medical Epigenetics, International Laboratory of Medical Epigenetics and Metabolism, Institutes of Biomedical Sciences, Fudan University, Shanghai, China LINDA MASANTE • Department of Cellular, Computational and Integrative Biology – CIBIO, University of Trento, Trento, Italy

Contributors

xi

SABINE MU¨LLER • University Hospital Schleswig-Holstein, Campus Lu¨beck, Germany AMARESH C. PANDA • Institute of Life Sciences, Nalco Square, Bhubaneswar, Odisha, India INES LUCIA PATOP • Brandeis University, Waltham, MA, USA SONJA PETKOVIC • University Hospital Schleswig-Holstein, Campus Lu¨beck, Germany TIZIANA SANTINI • Department of Biology and Biotechnologies “Charles Darwin”, Sapienza University of Rome, Rome, Italy GERHARD SCHRATT • Lab of Systems Neuroscience, Institute for Neuroscience, Department of Health Science and Technology, Swiss Federal Institute of Technology ETH, Zurich, Switzerland SUMAN SINGH • Institute of Life Sciences, Nalco Square, Bhubaneswar, Odisha, India DANILO SPRINGSTUBBE • University Hospital Schleswig-Holstein, Campus Lu¨beck, Germany BRETT W. STRINGER • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia JUNYI SU • Interdisciplinary Nanoscience Center, Aarhus University, Aarhus, Denmark GIORGIA SUSIN • Department of Cellular, Computational and Integrative Biology – CIBIO, University of Trento, Trento, Italy THOMAS THUM • Institute of Molecular and Translational Therapeutic Strategies (IMTTS), Hannover Medical School, Hannover, Germany MORTEN TRILLINGSGAARD VENØ • Omiics ApS, Aarhus, Denmark ERIKA VITIELLO • Center for Human Technology, Istituto Italiano di Tecnologia (IIT), Genoa, Italy; Department of Biology and Biotechnologies “Charles Darwin”, Sapienza University of Rome, Rome, Italy ISABEL S. NAARMANN-DE VRIES • Klaus Tschira Institute for Integrative Computational Cardiology, University Heidelberg, Heidelberg, Germany; Department of Internal Medicine III (Cardiology, Angiology, and Pneumology), University Hospital Heidelberg, Heidelberg, Germany; German Centre for Cardiovascular Research (DZHK)-Partner Site Heidelberg/Mannheim, Heidelberg, Germany STUART T. WEBB • Flinders Health and Medical Research Institute (FHMRI), College of Medicine and Public Health, Flinders University, Adelaide, SA, Australia JEREMY E. WILUSZ • Verna and Marrs McLean Department of Biochemistry and Molecular Pharmacology, Therapeutic Innovation Center, Baylor College of Medicine, Houston, TX, USA HAO WU • Key Laboratory of RNA Science and Engineering, CAS Center for Excellence in Molecular Cell Science, Shanghai Institute of Biochemistry and Cell Biology, University of Chinese Academy of Sciences, Chinese Academy of Sciences, Shanghai, China MEI-SHENG XIAO • RNA Biology Laboratory, Center for Cancer Research, National Cancer Institute, National Institutes of Health, Frederick, MD, USA YAN YAN • Omiics ApS, Aarhus, Denmark LI YANG • Center for Molecular Medicine, Children’s Hospital, Fudan University and Shanghai Key Laboratory of Medical Epigenetics, International Laboratory of Medical Epigenetics and Metabolism, Institutes of Biomedical Sciences, Fudan University, Shanghai, China JINGJING ZHAO • Fudan University Shanghai Cancer Center and Institutes of Biomedical Sciences, Shanghai Medical College, Fudan University, Shanghai, China

Part I circRNA Purification

Chapter 1 Purification of Circular RNAs Using Poly(A) Tailing Followed by RNase R Digestion Mei-Sheng Xiao and Jeremy E. Wilusz Abstract Thousands of eukaryotic protein-coding genes can be alternatively spliced to yield linear mRNAs and circular RNAs (circRNAs). Some circRNAs accumulate to higher levels than their cognate linear mRNAs, but the vast majority are expressed at low levels. Hence, for most circRNAs, only a handful of sequencing reads, if any, that span the backsplicing junction are observed in standard RNA-seq libraries. It thus has become common to use the 3′–5′ exonuclease ribonuclease R (RNase R) to selectively degrade linear RNAs when aiming to prove transcript circularity or biochemically enrich circRNAs. However, RNase R fails to degrade linear RNAs with structured 3′ ends or internal G-quadruplex structures. To overcome these shortcomings, we describe an improved protocol for circRNA purification from total RNA that employs a poly(A) tailing step prior to RNase R digestion, which is performed in a Li+ containing buffer (rather than K+) to destabilize G-quadruplexes. This biochemical method enables higher enrichment (two- to threefold) of circRNAs to be obtained compared to standard RNase R protocols due to more efficient removal of linear RNAs. By then performing quantitative RT-PCR (RT-qPCR) or generating RNA-seq libraries, the expression of individual circRNAs can be examined or the entire set of expressed circRNAs defined using established annotation algorithms. We describe step-by-step methods for annotating circRNAs using the CIRI2 and CIRCexplorer2 algorithms. In total, this overall approach can be used to enrich for circRNAs from any total RNA sample, thereby enabling one to quickly identify and validate circRNAs of interest for functional studies. Key words Backsplicing, CircRNA, CiRNA, G-quadruplex, Poly(A) polymerase, Ribonuclease R, RNA-seq, RT-qPCR, CIRI2, CIRCexplorer2

1

Introduction Most eukaryotic genes are split by intronic sequences that need to be removed from nascent pre-messenger RNAs (pre-mRNAs) [1, 2]. Introns are typically removed in a sequential order as lariat structures when the spliceosome joins a splice donor (5′ splice site) to a downstream splice acceptor (3′ splice site), leading to production of a mature linear mRNA (Fig. 1, top). However, it is now recognized that the spliceosome can alternatively join a splice

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_1, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

3

4

Mei-Sheng Xiao and Jeremy E. Wilusz mRNA m7G

AAAAAA

circRNA

Fig. 1 Eukaryotic genes can be alternatively spliced to generate linear and circular RNAs. The pre-mRNA splicing machinery can generate a canonical linear RNA when splice sites (ss) are joined in a linear order (top). Alternatively, a 5′ ss can be joined to an upstream 3′ ss via backsplicing to produce a circular RNA with covalently linked ends (bottom)

donor to an upstream splice acceptor in a process known as “backsplicing” to generate an exonic circular RNA (circRNA) with covalently linked ends [reviewed in 3–6] (Fig. 1, bottom). Backsplicing reactions are often much less efficient than canonical splicing events [7] and the vast majority of circRNAs in cells are of low abundance. Nonetheless, thousands of circRNAs have been identified across eukaryotes [8–11] and a subset are expressed at much higher levels than their cognate linear mRNAs [9], in part because circRNAs are naturally resistant to degradation by exonucleases. The set of circRNAs expressed in a given cell type is most commonly defined using RNA-seq followed by identification of sequencing reads that span backsplicing junctions [reviewed in 12, 13]. Multiple algorithms to identify such reads have been developed [e.g., 14, 15], but the low expression level of most circRNAs means that only a handful of reads (if any) spanning their backsplicing junctions are observed in standard RNA-seq libraries. To comprehensively identify circular transcripts, it has thus become common to use ribonuclease R (RNase R), a 3′–5′ exonuclease, to selectively degrade most linear RNAs (e.g., SEC61G mRNA, Fig. 2a) prior to RNA-seq library preparation [8, 16, 17]. Doing so enables increased representation in the sequencing library of exonic circRNAs (e.g., from the CDYL gene, Fig. 2b) and circular intronic RNAs (ciRNAs), which are derived from intron lariats that fail to be debranched (e.g., from intron 2 of the ANKRD52 gene, Fig. 2c).

Biochemical Purification of Circular RNAs

A 1500

SEC61G

2 kb

CDYL

50 kb

5

Control 1500

RNase R (KCl) 1500

A-Tailing + RNase R (LiCl)

B 2000

Control 2000

RNase R (KCl) 2000

A-Tailing + RNase R (LiCl)

circRNA

C 1500

ANKRD52

2 kb

HIST2H4B

100 bp

MBNL1

20 kb

Control 1500

RNase R (KCl) 1500

A-Tailing + RNase R (LiCl)

ciRNA

D 9000

Control 9000

RNase R (KCl) 9000

A-Tailing + RNase R (LiCl)

E 500

Control 500

RNase R (KCl) 500

A-Tailing + RNase R (LiCl)

G-quad pA

Fig. 2 A-tailing followed by RNase R treatment enables better linear RNA degradation and circular RNA enrichment compared to a standard RNase R protocol. (a–e) RNA-seq data were generated from HeLa total RNA that had been subjected to a control treatment (red), digested at 37 °C for 15 min with RNase R in KCl

6

Mei-Sheng Xiao and Jeremy E. Wilusz

RNase R has helicase activity that enables it to degrade many highly structured linear RNAs, including ribosomal RNAs, as long as a single-stranded 3′ overhang of at least seven nucleotides is present [18–20]. Nonetheless, a number of abundant cellular RNAs, including small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNAs), tRNAs, and replication-dependent histone mRNAs (e.g., HIST2H4B, Fig. 2d), have structured 3′ ends and are poor substrates for RNase R. This causes these linear transcripts to become enriched by standard RNase R treatments (Fig. 2d). The inclusion of an in vitro polyadenylation step prior to RNase R treatment greatly improves the ability of the enzyme to degrade such transcripts [16, 17]. We previously noticed that RNase R also stalls within the bodies of hundreds of polyadenylated mRNAs at G-quadruplex structures (e.g., in the 3′ UTR of MBNL1, Fig. 2e) [16]. Monovalent cations, such as K+ (which are present in the most commonly used RNase R reaction buffer: 20 mM Tris–HCl, pH 8, 0.1 M KCl, and 0.1 mM MgCl2), intercalate into the central core and stabilize G-quadruplex structures, while cations with a smaller ionic radius, such as Li+, do not [reviewed in 21]. Hence, by changing the buffer used in the RNase R reaction from one that contains K+ to one that contains Li+ (20 mM Tris–HCl, pH 8, 0.1 M LiCl, and 0.1 mM MgCl2), we were able to show that RNase R can now efficiently digest G-quadruplex containing mRNAs (Fig. 2e) [16]. Note that RNase R has reduced activity in a Li+ containing buffer and longer incubation times are needed, e.g., 90 min rather than 15 min in a K+ containing buffer. Here, we describe in detail this improved method for circRNA identification/quantification that takes advantage of biochemical enrichment of circRNAs from isolated total RNA using a poly (A) tailing step followed by RNase R digestion in a Li+ containing buffer (Fig. 3) [16]. Compared to standard RNase R protocols that use a K+ containing buffer, this protocol enables higher enrichment (two to threefold) of exonic circRNAs (Fig. 2b) and ciRNAs (Fig. 2c) to be obtained. This is due to more efficient removal of linear RNAs with structured ends (Fig. 2d) and/or G-quadruplex structures (Fig. 2e). The obtained RNAs that are enriched in circRNAs can be used in quantitative RT-PCR (RT-qPCR) assays to examine individual transcripts or used as input for RNA-seq library preparation. The latter approach enables the global set of circRNAs present in cells to be identified, annotated using the CIRI2 [14] ä Fig. 2 (continued) containing buffer (blue), or subjected to A-tailing followed by RNase R digestion at 37 °C for 90 min in LiCl containing buffer (this protocol, green) [16]. Normalized sequencing read coverage was calculated and the SEC61G (a), CDYL (b), ANKRD52 (c), HIST2H4B (d), and MBNL1 (e) loci are shown. Regions in CDYL and ANKRD52 that generate a circRNA or a circular intronic RNA (ciRNA), respectively, are denoted. For MBNL1, the polyadenylation (pA) signal as well a region of the transcript that forms a G-quadruplex (G-quad) are noted. Gray arrow below each of the gene models indicates the direction of transcription

Biochemical Purification of Circular RNAs

7

Culture cells

Extract total RNA using TRIzol

Confirm RNA quality using 1% agarose bleach gel Poly(A) tailing of total RNA using E-PAP - 10 min at 37°C Purify RNA Degrade linear RNAs using RNase R in LiCl containing buffer - 90 min at 37 °C Purify RNA

Validate efficiency of circRNA enrichment using RT-qPCR

RNA-seq library preparation and sequencing

Annotation of circRNAs using CIRI2 and CIRCexplorer2 algorithms

Visualization of circular RNAs on genome browser

Fig. 3 Workflow for circular RNA enrichment, sequencing, and annotation. See text for details

and CIRCexplorer2 [15] algorithms, and visualized on the UCSC genome browser. For global circRNA annotation, we recommend the use of two circRNA detection algorithms as the sets of transcripts identified by each approach can be quite divergent [reviewed in 12, 13]. Only the intersecting set of circRNAs should be used for downstream analysis. In total, this method can be used to quickly identify circRNAs present in cells and help prioritize candidates for functional studies.

2 2.1

Materials Cell Culture

1. Cell culture room equipped with laminar flow hood and CO2 incubator. 2. HeLa cells (ATCC). 3. HeLa culture medium: Dulbecco’s modified Eagle’s medium (DMEM) (Thermo Fisher Scientific 25030081) with 10% (v/v) heat-inactivated fetal bovine serum (FBS) (HyClone SH30396.03) and 1% (v/v) penicillin-streptomycin (10,000 U/mL) solution (Thermo Fisher Scientific 15140122). 4. 10 cm tissue culture plates (Corning 430167 or similar).

8

Mei-Sheng Xiao and Jeremy E. Wilusz

2.2 Total RNA Extraction and Confirmation of RNA Quality

1. TRIzol Reagent (Thermo Fisher Scientific 15596018). 2. Orbital shaker. 3. Chloroform. 4. Isopropanol. 5. 75% (v/v) ethanol in water. 6. Nuclease-free water. 7. Sterile 1.7 mL microtubes. 8. Centrifuge equipped with rotor for 1.7 mL microtubes. 9. UV-vis spectrophotometer with microvolume capacity (e.g., NanoDrop instrument from Thermo Fisher Scientific). 10. Standard reagents electrophoresis.

and

equipment

for

agarose

gel

11. 1x TBE (Tris/borate/EDTA) buffer. 12. Agarose bleach gel: 1% (w/v) agarose, 1% (v/v) bleach, ethidium bromide, in 1× TBE buffer [22]. 13. 10x RNase R Buffer: 0.2 M Tris–HCl (pH 8.0), 1 mM MgCl2, 1 M LiCl. 2.3

Poly(A) Tailing

1. Poly(A) Tailing Kit (Thermo Fisher Scientific AM1350). 2. RNase Inhibitor (Thermo Fisher Scientific N8080119 or similar). 3. 37 °C incubator (e.g., heat block or PCR machine). 4. RNA Clean & Concentrator-5 kit for RNA clean up (Zymo Research R1013). 5. Centrifuge equipped with rotor for 1.7 mL microtubes. 6. Nuclease-free water.

2.4 RNase R Digestion

1. 20 U/μL RNase R (Lucigen RNR07250 or equivalent). 2. 10x RNase R Buffer: 0.2 M Tris–HCl (pH 8.0), 1 mM MgCl2, 1 M LiCl. 3. 37 °C incubator (e.g., heat block or PCR machine). 4. RNA Clean & Concentrator-5 kit for RNA clean up (Zymo Research R1013). 5. Centrifuge equipped with rotor for 1.7 mL microtubes. 6. Nuclease-free water.

2.5

RT-qPCR

1. SuperScript III first-strand synthesis kit for RT-PCR (Thermo Fisher Scientific 18080051) or equivalent. 2. Power SYBR Green PCR Master Mix (Thermo Fisher Scientific 4367659) or equivalent. 3. 96-well qPCR plates and optical adhesive film covers. 4. Centrifuge equipped with rotor for microplates.

Biochemical Purification of Circular RNAs

9

5. Real-time qPCR instrument, e.g., LightCycler 96 Real-Time PCR System (Roche) or equivalent. 6. Nuclease-free water. 7. Gene-specific oligonucleotide primers to amplify transcripts of interest (see Note 1). 2.6 RNA-seq Library Preparation and Illumina Sequencing

1. Automated electrophoresis system and associated kit to determine RNA quality and quantity (e.g., Agilent BioAnalyzer or TapeStation). 2. Total RNA-seq library preparation kit that uses rRNA depletion step (e.g., TruSeq Stranded Total RNA Library Prep Gold (Illumina 20020598)). 3. (Optional) ERCC Spike-In Mix (Thermo Fisher Scientific 4456740). 4. Illumina NextSeq 500/550 system (see Note 2). 5. Illumina flow cell and sequencing reagents kit: TG NextSeq 500/550 High Output Kit v2.5 (150 cycles) (see Note 3).

2.7 RNA-seq Data Analysis

1. High-performance computing server or computer. 2. fastp (https://github.com/OpenGene/fastp). 3. BWA-MEM2 (https://github.com/lh3/bwa). 4. STAR (https://github.com/alexdobin/STAR). 5. perl (https://www.perl.org/get.html). 6. CIRI2 (https://sourceforge.net/projects/ciri/files/CIRI2/). 7. CIRCexplorer2, version 2.2.7 (https://circexplorer2. readthedocs.io/en/2.2.7/). 8. genePredToGtf tool (http://hgdownload.soe.ucsc.edu/ admin/exe/). 9. samtools (https://www.htslib.org/download/). 10. bedtools (https://bedtools.readthedocs.io/en/latest/con tent/installation.html).

3

Methods

3.1 Culture and Extraction of Total RNA from HeLa Cells

1. Grow HeLa cells at 37 °C and 5% CO2 in culture media in 10 cm tissue culture plates until cells are nearly confluent. 2. Remove media and add TRIzol reagent (3 mL/plate) to lyse cells. Place plates on an orbital shaker for 5 min to facilitate lysis. 3. Transfer 1 mL cell lysate into a sterile 1.7 mL microtube (see Note 4). 4. Add 200 μL chloroform, close tube lid, and shake vigorously for 20 s. Incubate samples at room temperature for 5 min. Centrifuge samples for 15 min at 12,000× g at 4 °C.

10

Mei-Sheng Xiao and Jeremy E. Wilusz

5. Carefully take tubes out of the centrifuge and transfer the colorless upper aqueous phase that contains RNA to a new 1.7 mL microtube. It is important to avoid pipetting from the interphase or the lower red phenol-chloroform layer. 6. Add 0.5 mL isopropanol to the aqueous phase. Mix well by inverting multiple times and incubate at room temperature for 10 min. Centrifuge for 15 min at 12,000× g at 4 °C. 7. Remove and discard supernatant with a micropipettor or vacuum without disturbing the white RNA pellet at the bottom of the tube. 8. Add 1 mL 75% ethanol to the RNA pellet. Invert tubes several times and centrifuge for 5 min at 12,000× g at 4 °C. 9. Remove and discard all of the 75% ethanol with a micropipettor or vacuum without disturbing the white RNA pellet at the bottom of the tube. Dry the pellet for 5 min at room temperature. 10. Resuspend the RNA pellet in nuclease-free water by pipetting up and down many times. 11. Determine RNA concentration from absorbance at 260 nm (1 OD260 = 40 μg/mL). 12. To confirm RNA quality and the lack of contaminating nucleases, set up 20 μL reactions that include (i) 5 μg total RNA in 1x RNase R buffer or (ii) 5 μg total RNA in nucleasefree water. Incubate samples for at least 30 min at 37 °C. Run reactions on 1% bleach agarose gel in 1x TBE buffer [22]. For high quality RNA samples, the 28S and 18S ribosomal RNA bands should be distinct and there should be no quantitative differences between samples incubated in water or RNase R buffer. 3.2 Poly(A) Tailing of Total RNA

1. Set up a 50 μL reaction for polyadenylation of 10 μg total RNA using a poly(A) tailing kit that takes advantage of E. coli poly (A) polymerase (E-PAP). Volume

Component

X

Total RNA (10 μg)

10 μL

5x E-PAP Buffer

5 μL

25 mM MnCl2

5 μL

10 mM ATP

2 μL

RNase Inhibitor (20 U/μL)

26 μL – X

Nuclease-free Water

Biochemical Purification of Circular RNAs

11

2. Add 2 μL E-PAP enzyme and mix the reaction by pipetting up and down multiple times. Incubate for 10 min at 37 °C (see Note 5). 3. Place reaction on ice. If desired, reactions can be stored at 20 °C for later use. 4. Use the RNA Clean & Concentrator-5 kit according to the manufacturer’s instructions to clean up the RNA and remove the E-PAP enzyme. Elute RNA by adding 12 μL nuclease-free water (see Note 6). 5. Place purified RNA on ice or store at -20 °C for later use. 3.3 RNase R Digestion to Remove Linear RNAs

1. Use the purified RNA that had been subjected to poly (A) tailing in the previous section as a substrate in the following 20 μL reaction (see Note 7) Volume

Component

11 μL

Polyadenylated RNAs (from previous section)

2 μL

10x RNase R Buffer

0.5 μL

RNase R (20 U/μL)

6.5 μL

Nuclease-free water

2. Incubate for 90 min at 37 °C. 3. Place tube on ice and use the RNA Clean & Concentrator-5 kit according to the manufacturer’s instructions to clean up the RNA and remove RNase R enzyme. Elute the purified RNA by adding 25 μL nuclease-free water (see Note 6). 4. Determine RNA concentration from absorbance at 260 nm (1 OD260 = 40 μg/mL). Use RNA immediately or store at -80 °C. 3.4 RT-qPCR to Examine Individual Transcript Levels

1. Use equal amounts of total RNA (e.g., 300 ng) that had been subjected to the polyadenylation/RNase R treatments and control treatments as templates for first-strand cDNA synthesis. Use random hexamers and SuperScript III in 20 μL reactions following the manufacturer’s instructions. Once reactions are complete, dilute cDNA fivefold with nucleasefree water. Diluted cDNA can be stored at -20 °C for later use if desired. 2. Set up RT-qPCR reactions using Power SYBR Green PCR Master Mix. We typically set up two or three technical replicates for each sample.

12

Mei-Sheng Xiao and Jeremy E. Wilusz

Volume

Component

7.5 μL

2x Power SYBR Green PCR Master Mix

6 μL

Primer pairs (1.5 μM)

1 μL

Diluted cDNA

0.5 μL

Nuclease-free water

Seal plate with optical adhesive film and briefly centrifuge. 3. Place plate in real-time qPCR system and use the following cycling conditions: 95 °C for 10 min, 40 amplification cycles of 95 °C for 15 s followed by 60 °C for 1 min, and a final melting cycle of 95 °C for 10 s, 65 °C for 1 min, and 97 °C for 1 s. A melting curve analysis step should be included to verify that a single peak is observed for each primer pair. 4. Calculate relative transcript levels using the 2-ΔΔCT method. Linear RNAs (e.g., GAPDH mRNA) should have been efficiently degraded by the polyadenylation/RNase R treatments while circular RNAs should have become enriched (see Note 8). 3.5 RNA-seq Library Preparation and Illumina Sequencing

1. Check quality and quantity of polyadenylation/RNase R treated RNAs using Bioanalyzer or TapeStation systems according to the manufacturer’s instructions. As most of the ribosomal RNA should have been degraded by RNase R, a low RNA integrity number (RIN), e.g., 4, is normal. 2. Use 500 ng RNA as starting material for ribosomal RNA depletion and RNA-seq library preparation using a commercial kit following the manufacturer’s instructions (see Note 9). 3. RNA-seq libraries are sequenced from a single end for 150 nt on the Illumina NextSeq 500/550 system.

3.6 Annotation of Circular RNAs from RNA-seq Data Using CIRI2

1. Filter and trim the sequencing reads in fastq format with fastp [23] to remove reads with low quality scores (Q scores 30), step 1 will not

Biochemical Purification of Circular RNAs

13

significantly affect the set of circular RNA candidates that are identified. 2. Install BWA aligner tool on computer or high-performance computing server following the developer’s documentation. 3. Use the Wget network utility to download the whole human genome sequence (GRCh38 assembly) as a reference for mapping from UCSC: $ wget https://hgdownload.soe.ucsc.edu/goldenPath/hg38/ bigZips/hg38.fa.masked.gz This file should then be decompressed and renamed as hg38_ref.fa with Gunzip and the mv command in Linux, respectively: $ gunzip hg38.fa.masked.gz $ mv hg38.fa.masked hg38_ref.fa 4. Use the Wget network utility to download the human genome annotation file from GENCODE with the basic gene annotation for all chromosomes: $ wget https://ftp.ebi.ac.uk/pub/databases/gencode/ Gencode_human/release_42/gencode.v42.basic.annota tion.gtf.gz Decompress the gtf file using the Gunzip command in Linux: $ gunzip gencode.v42.basic.annotation.gtf.gz Rename the decompressed gtf file as hg38.gtf using the mv command in Linux: $ mv gencode.v42.basic.annotation.gtf hg38.gtf 5. Generate genome index files for bwa mapping using hg38_ref. fa file from step 3: $ bwa index -a bwtsw hg38_ref.fa This step will generate multiple files with hg38_ref.fa as a prefix for each generated index file. 6. Map filtered RNA-seq reads generated in step 1 (read1.flt. fastq) to the genome with the bwa mem algorithm that performs local alignments. We use a minimum acceptable alignment score (the -T parameter) of 19. $ bwa mem -T 19 hg38_ref.fa read1.flt.fastq > SE_alignments. sam An alignment output file called SE_alignments.sam will be generated. 7. Download and install perl on computer or high-performance computing server following the developer’s documentation.

14

Mei-Sheng Xiao and Jeremy E. Wilusz

8. Use the Wget network utility to download the latest version of CIRI2 (using 2.0.6 as an example) following the developer’s documentation: $ wget https://sourceforge.net/projects/ciri/files/CIRI2/ CIRI_v2.0.6.zip Decompress the file: $ unzip CIRI_v2.0.6.zip The executable file CIRI2.pl will be found in the folder CIRI_v2.0.6 and no installation is required. 9. Use perl to run CIRI2.pl to identify putative circular RNAs in the mapping output file from step 6 (SE_aligments.sam). The genome annotation file from step 4 (hg38.gtf) and genome sequence file from step 3 (hg38_ref.fa) are used: $ perl ./CIRI_v2.0.6/CIRI2.pl -I SE_aligments.sam -O SE_circRNA -F hg38_ref.fa -A hg38.gtf 10. The output file (SE_circRNA) generated is a tab separated txt file which contains all identified circular RNAs (see Note 11). 3.7 Annotation of Circular RNAs from RNA-seq Data Using CIRCexplorer2

1. Install CIRCexplorer2 (version 2.2.7) on computer or highperformance computing server following the developer’s documentation. 2. Install the STAR aligner for sequence read mapping according to the developer’s documentation (see Note 12). 3. Use the Wget network utility to download the version of genePredToGtf tool that matches your computing system (linux.x86_64 is used here as an example). The genePredToGtf tool will be used to convert gene annotations from UCSC to GTF format. $ wget http://hgdownload.soe.ucsc.edu/admin/exe/linux. x86_64/genePredToGtf Use the Linux chmod command to make the file executable: $ chmod +x genePredToGtf 4. Use the fetch_ucsc.py script that is built into CIRCexplorer2 to download the human gene annotation and reference genome files that will be used for mapping. Obtain RefSeq gene annotation: $ fetch_ucsc.py hg38 ref ucsc_hg38_ref.txt Obtain human reference genome file: $ fetch_ucsc.py hg38 fa ucsc_hg38.fa Convert gene annotation file (ucsc_hs38_ref.txt) to GTF format using the cut command in Linux as well as the genePredToGtf tool:

Biochemical Purification of Circular RNAs

15

$ cut -f 2-11 ucsc_hg38_ref.txt|./genePredToGtf file stdin ucsc_hg38_ref.gtf 5. Generate index files for mapping RNA-seq reads to the genome with STAR. First, create a directory named ucsc_hg38_STAR_Index for storing the index to be generated: $ mkdir ucsc_hg38_STAR_Index Run STAR to generate the genome index for mapping: $ STAR --runThreadN 8 --runMode genomeGenerate -genomeDir ucsc_hg38_STAR_Index --genomeFastaFiles ucsc_hg38.fa --sjdbGTFfile ucsc_hg38_ref.gtf -sjdbOverhang 149 In this case, “8” indicates that eight threads will be used for the run, but note that this number should be set to the desired number of available cores on the server node. The genomeGenerate option will generate the index for the genome and will deposit it in the ucsc_hg38_STAR_Index directory generated above. The reference genome file (ucsc_hg38.fa) and gene annotation file in GTF format (ucsc_hg38_ref.gtf) were generated in step 4. “149” is calculated by subtracting 1 from the sequencing read length (here it is 150 nt - 1 = 149). 6. Map RNA-seq reads with STAR: $ STAR --chimSegmentMin 10 --runThreadN 10 --genomeDir ucsc_hg38_STAR_Index --readFilesIn read1.flt.fastq Here, detection of chimeric (fusion) alignments is allowed with a minimum mapped length of 10 nt for each of the two segments. Ten threads are used for the run. The ucsc_hg38_STAR_Index genome index directory that was generated in step 5 is used for mapping. The input RNA-seq filtered read file (read1.flt.fastq) was generated in step 1 of Subheading 3.6. 7. Use the CIRCexplorer2 parse tool to identify reads that span putative backsplicing junctions using the Chimeric.out.junction file that was generated in step 6: $ CIRCexplorer2 parse -t STAR Chimeric.out.junction > CIRCexplorer2_parse.log This step will generate a folder called circ_out which contains a bed format file (fusion_junction.bed). 8. Use the CIRCexplorer2 annotate tool to annotate circular RNAs based on the fusion_junction.bed file generated in step 7: $ CIRCexplorer2 annotate -r ucsc_hg38_ref.txt -g ucsc_hg38. fa circ_out > CIRCexplorer2_annotate.log The gene annotation file (ucsc_hg38_ref.txt) and reference genome file (ucsc_hg38.fa) were generated in step 4. This step

16

Mei-Sheng Xiao and Jeremy E. Wilusz

will generate another default folder (annotation) within the circ_out folder from step 7. In the annotation folder, one of the output files is annotated_fusion.txt contains all the annotated fusion junction information, while circ_fusion.txt contains all circular RNAs identified by CIRCexplorer2 (see Note 13). 3.8 Generate Genome Browser Tracks for Visualization of Circular RNAs

1. Use samtools to sort sam files from the mapping outputs of either STAR or bwa and output it as bam format. Here, the STAR mapping output file (Aligned.out.sam) that was generated in step 6 of Subheading 3.7 is used as an example: $ samtools sort -T tmp Aligned.out.sam -o Aligned.out.sorted. bam 2. Install bedtools which will be used to convert the bam file to bedGraph format. 3. Use bedtools to convert the sorted bam format file (Aligned. out.sorted.bam) from step 1 to bedGraph format: $ genomeCoverageBed -bg -split -ibam Aligned.out.sorted. bam > Aligned.out.bedGraph 4. Add a track definition line that enables multiple different options for controlling how the track will be displayed. The following command adds a simple definition line at the beginning of the bedGraph file and then compresses it: $ echo -e ’track type=bedGraph name=“NAME” visibility=2 color=225,0,0’ |cat - Aligned.out.bedGraph|gzip -c > Aligned.out.bedGraph.gz Note that NAME should be changed to an appropriate descriptive title for the track. Visibility mode of 2 enables the initial display to be in full mode. Track color consists of three comma separated RGB values from 0 to 255. Default is 0,0,0 (black). 5. The compressed bedGraph file (Aligned.out.bedGraph.gz) can be uploaded to the UCSC genome browser (https://genome. ucsc.edu) for visualization of the sequencing data like in Fig. 2.

4

Notes 1. Primers that amplify across the backsplicing junction should be used to detect and quantify circRNAs. Make sure that each PCR amplicon is between 100 and 200 nt. RT-qPCR primers for a set of human linear RNAs, circRNAs, and ciRNAs were previously published [16]. 2. RNA-seq libraries can be sequenced on other Illumina platforms, e.g., NextSeq 1000/2000 or NovaSeq 6000, if higher

Biochemical Purification of Circular RNAs

17

sequencing depth is desired. A system-appropriate sequencing flow cell and reagent kit should be used. 3. To identify circular RNAs, longer sequencing reads are preferred over shorter reads (e.g., 75 cycles) to more efficiently identify reads that span backsplicing junctions. 4. Transfer each mL of cell lysate in TRIzol to a separate 1.7 mL nuclease-free microtube. Lysates can be processed immediately or stored at -80 °C for future use. 5. A control reaction in which E-PAP is omitted can also be performed if one wants to ultimately determine the efficiency of the biochemical enrichment steps for allowing increased relative levels of circRNAs compared to their levels in total RNA. 6. Acid-phenol chloroform (pH 4.5) extraction followed by ethanol precipitation can alternatively be used to clean up and purify RNA from E-PAP and RNase R reactions. In this case, 1 μL glycogen (20 mg/mL) should be used as a coprecipitant to help observe the RNA pellet. Microtubes with phase lock gel can help eliminate interphase contamination and enable greater RNA yield to be obtained. 7. A control reaction where water is added in place of the RNase R enzyme can also be performed if one wants to ultimately determine the efficiency of the biochemical enrichment steps. 8. We have noticed differences in efficiencies across lots of RNase R. If it is observed that circRNAs have also been degraded by the polyadenylation/RNase R treatments, we recommend reducing the amount of RNase R enzyme used and/or the length of RNase R reaction time. 9. After performing ribosomal RNA depletion, we recommend performing a Bioanalyzer or TapeStation assay again to ensure that minimal ribosomal RNA is remaining. Large amounts of ribosomal RNA contamination will cause most of the sequencing reads to be derived from ribosomal RNAs. A 1/100 dilution of the ERCC Spike-In Mix (Thermo Fisher Scientific) can be added to the samples after ribosomal RNA depletion to measure sensitivity and dynamic range of the experiment as well as aid in quantitation of differential gene expression patterns. 10. Alternative tools, including Cutadapt (https://cutadapt. readthedocs.io/en/stable/), can be used for trimming and filtering sequencing reads. 11. Further details on the output file can be obtained from the CIRI2 developer’s documentation: https://ciri-cookbook. readthedocs.io/en/latest/CIRI2.html#release-notes.

18

Mei-Sheng Xiao and Jeremy E. Wilusz

12. Other mapping tools such as TopHat2 can also be used but STAR is much faster. If you prefer to use TopHat2 for mapping, it is more efficient and flexible to align manually instead of using the CIRCexplorer2 align command. 13. Further detailed information regarding the output file can be obtained from the CIRCexplorer2 developer’s documentation: https://circexplorer2.readthedocs.io/en/2.2.7/modules/ annotate/.

Acknowledgments This work was supported by NIH grant R35-GM119735. J.E.W. is a CPRIT Scholar in Cancer Research (RR210031). J.E.W. serves as a consultant for Laronde. References 1. Nilsen TW, Graveley BR (2010) Expansion of the eukaryotic proteome by alternative splicing. Nature 463(7280):457–463. https://doi. org/10.1038/nature08909 2. Wahl MC, Will CL, Luhrmann R (2009) The spliceosome: design principles of a dynamic RNP machine. Cell 136(4):701–718. https:// doi.org/10.1016/j.cell.2009.02.009 3. Wilusz JE (2018) A 360 degrees view of circular RNAs: from biogenesis to functions. Wiley Interdiscip Rev RNA:e1478. https:// doi.org/10.1002/wrna.1478 4. Yang L, Wilusz JE, Chen LL (2022) Biogenesis and regulatory roles of circular RNAs. Annu Rev Cell Dev Biol 38:263–289. https://doi. org/10.1146/annurev-cellbio120420-125117 5. Kristensen LS, Andersen MS, Stagsted LVW, Ebbesen KK, Hansen TB, Kjems J (2019) The biogenesis, biology and characterization of circular RNAs. Nat Rev Genet 20(11):675–691. https://doi.org/10.1038/s41576-0190158-7 6. Patop IL, Wust S, Kadener S (2019) Past, present, and future of circRNAs. EMBO J 38(16): e100836. https://doi.org/10.15252/embj. 2018100836 7. Zhang Y, Xue W, Li X, Zhang J, Chen S, Zhang JL, Yang L, Chen LL (2016) The biogenesis of nascent circular RNAs. Cell Rep 15(3): 611–624. https://doi.org/10.1016/j.celrep. 2016.03.058 8. Jeck WR, Sorrentino JA, Wang K, Slevin MK, Burd CE, Liu J, Marzluff WF, Sharpless NE (2013) Circular RNAs are abundant,

conserved, and associated with ALU repeats. RNA 19(2):141–157. https://doi.org/10. 1261/rna.035667.112 9. Salzman J, Gawad C, Wang PL, Lacayo N, Brown PO (2012) Circular RNAs are the predominant transcript isoform from hundreds of human genes in diverse cell types. PLoS One 7(2):e30733. https://doi.org/10.1371/jour nal.pone.0030733 10. Memczak S, Jens M, Elefsinioti A, Torti F, Krueger J, Rybak A, Maier L, Mackowiak SD, Gregersen LH, Munschauer M, Loewer A, Ziebold U, Landthaler M, Kocks C, le Noble F, Rajewsky N (2013) Circular RNAs are a large class of animal RNAs with regulatory potency. Nature 495(7441):333–338. https:// doi.org/10.1038/nature11928 11. You X, Vlatkovic I, Babic A, Will T, Epstein I, Tushev G, Akbalik G, Wang M, Glock C, Quedenau C, Wang X, Hou J, Liu H, Sun W, Sambandan S, Chen T, Schuman EM, Chen W (2015) Neural circular RNAs are derived from synaptic genes and regulated by development and plasticity. Nat Neurosci 18(4):603–610. https://doi.org/10.1038/nn.3975 12. Szabo L, Salzman J (2016) Detecting circular RNAs: bioinformatic and experimental challenges. Nat Rev Genet 17(11):679–692. https://doi.org/10.1038/nrg.2016.114 13. Dodbele S, Mutlu N, Wilusz JE (2021) Best practices to ensure robust investigation of circular RNAs: pitfalls and tips. EMBO Rep 22(3):e52072. https://doi.org/10.15252/ embr.202052072

Biochemical Purification of Circular RNAs 14. Gao Y, Wang J, Zhao F (2015) CIRI: an efficient and unbiased algorithm for de novo circular RNA identification. Genome Biol 16(1): 4. https://doi.org/10.1186/s13059-0140571-3 15. Zhang XO, Dong R, Zhang Y, Zhang JL, Luo Z, Zhang J, Chen LL, Yang L (2016) Diverse alternative back-splicing and alternative splicing landscape of circular RNAs. Genome Res 26(9):1277–1287. https://doi. org/10.1101/gr.202895.115 16. Xiao MS, Wilusz JE (2019) An improved method for circular RNA purification using RNase R that efficiently removes linear RNAs containing G-quadruplexes or structured 3′ ends. Nucleic Acids Res 47(16):8755–8769. https://doi.org/10.1093/nar/gkz576 17. Panda AC, De S, Grammatikakis I, Munk R, Yang X, Piao Y, Dudekula DB, Abdelmohsen K, Gorospe M (2017) Highpurity circular RNA isolation method (RPAD) reveals vast collection of intronic circRNAs. Nucleic Acids Res 45(12):e116. https://doi. org/10.1093/nar/gkx297 18. Cheng ZF, Deutscher MP (2005) An important role for RNase R in mRNA decay. Mol Cell 17(2):313–318. https://doi.org/10.1016/j. molcel.2004.11.048

19

19. Vincent HA, Deutscher MP (2006) Substrate recognition and catalysis by the exoribonuclease RNase R. J Biol Chem 281(40): 29769–29775. https://doi.org/10.1074/jbc. M606744200 20. Hossain ST, Malhotra A, Deutscher MP (2016) How RNase R degrades structured RNA: ROLE OF THE HELICASE ACTIVITY AND THE S1 DOMAIN. J Biol Chem 291(15):7877–7887. https://doi.org/ 10.1074/jbc.M116.717991 21. Varshney D, Spiegel J, Zyner K, Tannahill D, Balasubramanian S (2020) The regulation and functions of DNA and RNA G-quadruplexes. Nat Rev Mol Cell Biol 21(8):459–474. https://doi.org/10.1038/s41580-0200236-x 22. Aranda PS, LaJoie DM, Jorcyk CL (2012) Bleach gel: a simple agarose gel for analyzing RNA quality. Electrophoresis 33(2):366–369. https://doi.org/10.1002/elps.201100335 23. Chen S, Zhou Y, Chen Y, Gu J (2018) fastp: an ultra-fast all-in-one FASTQ preprocessor. Bioinformatics 34(17):i884–i890. https://doi. org/10.1093/bioinformatics/bty560

Part II In Silico circRNA Profiling

Chapter 2 State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite Tobias Jakobi Abstract Circular RNAs (circRNAs) are types of RNA molecules that have been discovered relatively recently and have been found to be widely expressed in eukaryotic cells. Unlike canonical linear RNA molecules, circRNAs form a covalently closed continuous loop structure without a 5′ or 3′ end. They are generated by a process called back-splicing, in which a downstream splice donor site is joined to an upstream splice acceptor site. CircRNAs have been found to play important roles in various biological processes, including gene regulation, alternative splicing, and protein translation. They can act as sponges for microRNAs or RNA-binding proteins and can also encode peptides or proteins. Additionally, circRNAs have been implicated in several diseases, including cancer, neurological disorders, and cardiovascular diseases. This protocol provides all necessary steps to detect and analyze circRNAs in silico from RNA sequencing data using the circtools circRNA analytics software suite. The protocol starts from raw sequencing data with circRNA detection via back-splice events and includes statistical testing of circRNAs as well as primer design for follow-up wet lab experiments. Key words Bioinformatics, Circular RNA, CircSeq, Circular RNA detection, Circular RNA analysis, Circular RNA detection, Circtools, CircRNA primer design

1

Introduction In silico circRNA analysis is a relatively new field compared to other branches of bioinformatics. Although the first tools to detect circular RNAs from sequencing data were published in the early 2010s, most of the current tools used in the field are newer [1–8]. One common feature among many circRNA detection algorithms is the need for external programs that provide read mapping functions. Commonly used mapping tools, such as STAR [9], bowtie2 [10], and bwa [11], are suitable for paired-end sequencing and are splicesite aware, allowing them to properly align reads or read pairs that cover back-splice junctions (BSJ), which is particularly important for circRNA analysis.

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_2, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

23

24

Tobias Jakobi

CircRNAs can be detected from rRNA-depleted total RNA sequencing data. To improve the ratio of sequenced circRNAs, Jeck and colleagues developed the circSeq protocol [12], where sequencing libraries are treated with RNase R to deplete linear RNAs, yet enhancing representation of circRNAs in libraries. Subsequently, advanced protocols, such as RPAD [13] (RNase R treatment, polyadenylation, and poly(A) + RNA depletion) and a similar approach with the addition of LiCl buffer [14], further increase enrichment for circRNA in sequencing libraries. In this protocol, we present a circRNA-specific workflow starting with stringent quality control of the sequencing data to ensure reliable downstream analysis. The alignment step follows, producing BAM files consisting of compressed and aligned reads required for circRNA detection in follow-up steps. The protocol described in this manuscript employs the circtools software for circular RNA detection, enrichment testing, and design of circRNA-specific primers for qRT-PCR follow-up experiments.

2

Materials

2.1 Required Hardware

To perform RNA-seq read alignment on human genome scale using the STAR software, a 64-bit Linux or Mac OS operating system and an x86-64-based machine (recommended; newer Mac systems, i.e., M1, M2, and M3 architectures, are not yet supported by STAR at the time of writing) with at least 30GB of memory are necessary. The read mapping step is the most resource-intensive step in the workflow, and therefore, all subsequent steps will also perform well on the specified hardware. This protocol was tested on Debian Linux 11 (bullseye), but should run without issues on any other Linux distribution, such as Ubuntu 22.04 or higher as well as on x86-64-based Mac systems.

2.2 Required Software

To execute all aspects of the workflow, several external software packages are required: STAR, R, and Python are usually obtainable as prepackaged binary downloads and do not require compilation from source. Specifically, Python and R can be installed for Linux distributions using the package manager. On macOS systems, external software can be installed via brew or bioconda [15]. Circtools is Python-based and will be installed from source. Nonetheless, an automatic installer performs the compilation process. The following list contains all necessary software required to run the presented workflow. In general, the exact version of the software is not required to be identical, however, where applicable, minimum versions are listed. • The STAR RNA-seq read aligner [9], obtainable directly from GitHub via https://github.com/alexdobin/STAR.

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

25

• The Bowtie2 read aligner, hosted on Sourceforge: https:// sourceforge.net/projects/bowtie-bio/files/bowtie2. • SAMtools [16], the universal SAM/BAM file manipulation tools, obtainable via http://www.htslib.org/download/. • The R environment for statistical computing [18] (≥4.0.0), either via package manager of the employed Linux distribution or directly via https://cran.r-project.org/bin/linux/; R packages required by circtools are automatically installed during installation of circtools. • A working Python (≥3.7) environment [17], including the pip installer either via package manager of the employed Linux distribution or directly via https://www.python.org/ downloads/; required Python packages are automatically installed during the circtools installation. • Circtools [18], the all-in-one software solution for computational circular RNA research, obtainable from https://github. com/jakobilab/circtools. • FastQC [19], a quality control tool for sequencing reads, obtainable from https://www.bioinformatics.babraham.ac.uk/ projects/download.html#fastqc. • flexbar [20] (≥3.0.0), a read-trimming software, allowing to remove residual adapter sequences from sequencing reads, obtainable from https://github.com/seqan/flexbar. • parallel [21], a command line tool for easy parallelization of tasks. Within the context of this protocol, the tool is used to simulate a scheduling system typically found on large compute clusters. Parallel be installed via the distribution package manager or from source via https://www.gnu.org/software/ parallel/. 2.3

Required Data

This workflow employs publicly available sequencing data. Specifically, data from the following published study is analyzed: • CircSeq RNase R-enriched raw RNA-seq data (2 × 100 bp, Illumina HiSeq 2000) generated from whole heart tissue of C57/Bl6N mice [22]; downloadable via NCBI Sequence Read Archive (SRA), accession number SRP071584: https://trace. ncbi.nlm.nih.gov/Traces/study/?acc=SRP071584.

2.4

General Notes

The protocol outlined here is optimized for paired-end data. Although detection from single-end RNA-seq data is possible, experience shows that results are improving when employing paired-end data [23]. In order to perform this protocol on singleend data, the .sh bash scripts executing each step have to be modified. More information can be found in the comprehensive

26

Tobias Jakobi

online documentation of circtools available at https://docs.circ. tools. If this protocol is run on a high-performance computing (HPC) system, some of the aforementioned tools such as STAR and R might already be installed and are usually loaded through modules software system. In this case, the local HPC support team will be able to advise how to leverage the installed software. The bash scripts employed in this protocol are fully functional job submission scripts that can be used with the SLURM job scheduling system [24] found on many HPC systems to launch jobs. If no scheduler is available, the bash scripts also work on a single computer and parallel is used as surrogate for a job scheduling system. For longer code blocks (written in code font), a short link in the top of the listing contains a direct link to the code within the online circtools documentation. These links can be employed to minimize copy-and-paste errors. Within the code blocks, lines starting with # are comments. Only lines without # are executable and need to typed or copied into a terminal application of choice. 2.5 How to Get Support

The circtools team works hard to make the software a formidable tool empowering your research. However, software and environments and continuously changing, thus the team would love to receive feedback from users. Issues with circtools can be reported either via the GitHub platform, https://github.com/jakobilab/circtools/issues, or, alternatively, via email at through [email protected]. This protocol was written using the most recent version of circtools and software that circtools depends on. However, future software changes cannot be corrected in this written protocol. As such, we recommend checking http://book.circ.tools for most recent updates regarding the workflow presented here before starting to run the protocol. Software Installation Installation of circtools is performed via pip, docker, or Github. No sudo access is required if the installation is suffixed with --user which will install the package in a user-writeable folder. In this case, the binaries should be installed to /home/$USER/.local/bin/ (for Linux-based systems). Installation of Circtools via PyPi The default installation will install everything needed to run circtools except R or STAR. Circtools may be installed locally (first call) or globally (second call, root required): • Local/per user: user].

python3 -m pip install circtools [--

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

• Global

(sudo

required):

python3

-m

pip

27

install

circtools.

Installation of Circtools via GitHub The GitHub installation will install the most recent version of circtools directly from the source repository. This method always contains the latest fixes and features before a new version is released. git clone https://github.com/jakobilab/circtools. git cd circtools python3 -m pip install . [--user]

Installation of Circtools via Docker Docker can be used as a one-stop alternative to install circtools, taking care of all Python and R dependencies, thus making it well-suited for novice users. The only requirement is a working docker installation. The following command will install the latest stable version of circtools: docker pull ghcr.io/jakobilab/circtools/circtools: latest

Subsequently, circtools can be run via docker run circtools Other than the additional docker run prefix all commands remain unchanged.

[command line arguments].

Installation of Required R Packages The required R libraries will be installed in the default location in the user’s home directory—unless the environment variable $R_LIBS_USER is set differently. In case users want to install a dedicated venv installation, the --user option can be omitted. For several modules, such primer design, circtools requires certain R packages to be installed and accessible. However, no direct interaction with R is required, as the installation of all packages can be performed using a single command after circtools has been installed. However, compilation of packages might take some time depending on the system used. For the R packages to compile successfully, a number of system libraries are required. The complete list (with Ubuntu/Debian package names) can be found at https://links.jakobilab.org/circtools_r_deps. circtools_install_R_dependencies

Importantly, this step is not required if the docker installation method is employed, as the docker container contains all required R and Python dependencies.

28

Tobias Jakobi

Installation of the NCBI SRA Toolkit The NCBI SRA Toolkit is required to convert sequencing data downloaded from the SRA from the binary .sra format into compressed FASTQ files commonly used in bioinformatics. The software is available for all operating systems and can be freely downloaded via https://links.jakobilab.org/sratoolkit. The installation depends on the operating system; however, packages for CentOS and Ubuntu are available if sudo access is available. Moreover, as user-installable archives that do not require root privileges are available, too. 2.6 Download and Preparation of Raw Sequencing Data

The following steps download the example data from the NCBI SRA and prepare all raw data files for processing with circtools. These steps can be easily adapted to any other SRA project or own datasets. # short link: # https://links.jakobilab.org/circtools_step1 # create main folder and raw reads folder mkdir -p workflow/reads cd workflow/reads # place a copy of wonderdump.sh in this directory # also make it executable # wget is a download utility that the protocol # will employ # for all downloads (except the SRA raw data) wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)"

https://links.jakobilab.org/wonderdump.sh chmod 755 wonderdump.sh # get list of accession numbers (sequencing runs) to download # also get a mapping file from SRA accession to original # file name wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/jakobi2016_sra_list. txt wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/file_name_mapping.txt # downloading and rewriting the files as gzipped

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

29

# FASTQ files # will take some time; in the end, the process will # generate a set of 4 files (2 samples x 2 pairs) # this step requires an installed and configured # NCBI sratoolkit # wonderdump.sh is a wrapper for the # sratoolkit that simplifies # downloading multiple files # # start wonderdump with the accession list # and download data while IFS="" read -r p || [ -n "$p" ] do ./wonderdump.sh --split-files --gzip "$p" & done < jakobi2016_sra_list.txt # # rename files from SRA accession numbers to # file names used throughout this tutorial # # mate 1 ln -s SRR7881333_1.fastq.gz ALL_1654_S_R1.fastq.gz ln -s SRR7881334_1.fastq.gz ALL_1654_T_R1.fastq.gz # mate 2 ln -s SRR7881333_2.fastq.gz ALL_1654_S_R2.fastq.gz ln -s SRR7881334_2.fastq.gz ALL_1654_T_R2.fastq.gz

3

Methods Quality Control of Raw Sequencing Data with FastQC FastQC is a software tool designed to perform quality control checks on raw next-generation sequencing (NGS) data. The tool generates a comprehensive report that highlights various aspects of the sequencing data, including per-base sequence quality, GC content, read length distribution, overrepresented sequences, and sequencing adapter content (Fig. 1). FastQC is widely used in the NGS community for initial data assessment and quality control before proceeding with downstream analysis. A common command to analyze all sequencing files in the current directory using FastQC might look like this:

30

Tobias Jakobi Quality scores across all bases (Sanger / illumina 1.9 encoding) 40 38 36 34 32 30 28 26 24 22 20 18 16 14 12 10 8 6 4 2 0

1 2 3 4 5 6 7 8 9 12-13 18-19 24-25 30-31 36-37 42-43 48-49 54-55 60-61 66-67 72-73 78-79 84-85 90-91 96-97 Position in read (bp)

Fig. 1 FastQC report of one of the raw data files employed in the protocol showing the base quality for each position in the sequencing reads. This output is typical for a successful sequencing run with quality scores improving within the first ten bases, plateauing for the most of the read and finally slowly decreasing within the last ten bases of the read. The quality scores are segmented into green (good), yellow (acceptable), and red (low quality) # enter folder with reads if not still in same folder cd workflow/reads fastqc *.fastq.gz -t 4

This command assumes that all raw data files are compressed in the gzipped FASTQ format (hence *.fastq.gz) that was produced after retrieving data from the SRA and that up to four processors on the system are utilized for the analysis (-t 4). While warnings produced by FastQC are relatively cautious, users must pay attention to the details and ensure that the data does not exhibit any apparent issues, such as below-average read quality or extremely high sequence duplication levels. Removal of Remaining Sequencing Adapter Fragments with Flexbar The main objective of flexbar is to eliminate any leftover sequencing adapter fragments from reads post-sequencing. These remaining adapter sequences can potentially hamper the read mapping process, leading to a reduced mapping rate, which is the percentage of reads that can be aligned to the reference genome. It is crucial to consider the sequencing library setup during subsequent steps because the processing steps for single-end and paired-end datasets vary. In the example below, flexbar is run for a paired-end library:

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

31

# short link: # https://links.jakobilab.org/circtools_step2 # download the wrapper script for flexbar wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" # https://links.jakobilab.org/flexbar_paired.sh # add execute permission chmod 755 flexbar_paired.sh # go up one folder cd .. # create output folder mkdir flexbar # go back into reads folder cd reads # run flexbar wrapper, keep the mate pairs synchronized parallel -j1 --xapply ./flexbar_paired.sh {1} {2} ../ flexbar/ _R1 ::: *_R1.fastq.gz ::: *_R2.fastq.gz

The abovementioned command will run the downloaded flexbar wrapper script that internally employs eight processors (-n 8), providing both mates of the sequencing run via -r mate1 and p mate2 using a generic Illumina adapter sequence (-a AGATCGGAAGAG), latest Illumina quality scores (-qf sanger), and a conservative minimal quality score of 28 (-qt 28). The flexbar step above will be carried out for all samples in the library. After flexbar finished processing, the folder flexbar/ contains the trimmed, quality-filtered reads. Removal of rRNA with Bowtie2 In a next step, rRNA-mapping reads are removed. Usually, we are not interested in these reads, especially when performing circRNA analysis. Removing those reads will also slightly speed up subsequent steps due to the reduced computational load. In this tutorial, bowtie2 is used to discard reads that map against rRNA loci. This tutorial assumes that bowtie2 has already been installed and can be directly called (i.e., the executable is available through $PATH). A precompiled bowtie2 index of mouse rRNA loci has been uploaded for this purpose. Briefly, bowtie2 maps the reads against a reference of rRNA loci and only keeps reads that do not align and therefore are deemed rRNA-free.

32

Tobias Jakobi # short link: # https://links.jakobilab.org/circtools_step3 # leave the reads/ folder cd .. # download pre-built mouse rRNA index wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" # https://links.jakobilab.org/rrna_mouse.tar.bz2 # decompress and unpack index tar -xvf rrna_mouse.tar.bz2 # download bowtie2 script and make it executable wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" # https://links.jakobilab.org/bowtie2_filter.sh chmod 755 bowtie2_filter.sh mkdir rrna/ cd flexbar/ #

execute

bowtie2

for

all

samples

using

the

../bowtie2_filter.sh

../

wrapper script parallel

-j1

--xapply

rRNA_cluster/mus-musculus.rRNA {1} {2} ../rrna/ _R1 ::: *1.fastq.gz ::: *2.fastq.gz

After this step the rrna/ folder contains adapter-free, rRNAfree reads ready for final mapping to the actual reference genome. Principal Read Mapping with STAR To map the preprocessed reads against the reference genome and ultimately detect circRNAs, the STAR read mapping software is employed. Mapping reads to the reference genome is the most resource-intensive part of the workflow and as such reasonable hardware as suggested in the hardware section is recommended. STAR has shown to exhibit a good performance, is highly customizable and, most importantly, can directly export chimeric reads that are the basis for the circRNA detection process. Again, we are employing a wrapper script that simplifies the process of calling STAR for all samples. Essentially, the wrapper script for STAR performs the following tasks: • Map both reads of each pair against the reference genome. • Map the unmapped reads of mate 1 and mate 2 again against the reference genome without the corresponding paired partner.

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

33

• Perform several conversion and cleanup steps of the STAR output. Similar to bowtie2, STAR requires an index in order to align reads. Since building this index requires a significant amount of memory and time, a precomputed STAR index for the mouse ENSEMBL 90 build has been prepared for direct download (file size: 22GB). # short link: # https://links.jakobilab.org/circtools_step4 # go one folder up, leave the flexbar/ folder cd .. # download STAR index and unpack # this step will take some time due to the file size wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/star_index_chr10_m38_ 90.tar.bz2 tar -xvf star_index_chr10_m38_90.tar.bz2 # download and unpack the mouse genome annotation wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/mouse.gtf.gz gzip -d mouse.gtf.gz # download wrapper for STAR and make executable wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/star_mapping.sh chmod 755 star_mapping.sh # create output folder mkdir star/ # use rRNA-free reads from bowtie2 as input cd rrna/ # execute STAR for all samples # input are adapter-free, rRNA-cleaned reads in the # rrna/ folder parallel -j1 --xapply ../star_mapping.sh ../star_index/ {1} {2} ../star/ .1 ../mouse.gtf ::: *fastq.1. gz ::: *fastq.2.gz

34

Tobias Jakobi

The parameters employed in the star_mapping.sh script called in the listing above are optimized for usage with circtools and should only be changed by advanced users. We recommend that paired-end datasets are mapped three times, (1) using both mates together, (2) using only mate 1, and (3) using only mate 2. Since STAR is running three times for each sample, the above call will take a significant amount of time (i.e., several hours). The mapping process is set up to employ eight CPUs; if more are available, the star_mapping.sh file can be adapted appropriately. Detection of CircRNAs with Circtools Detect Circtools can detect and quantify circRNAs with high specificity [25]. The software employs STAR’s Chimeric.out.junction files that contain chimerical aligned reads, including circRNA BSJ-spanning reads. Before circRNA detection with circtools can be performed, several input files must be prepared in order for circtools to work correctly: • A sample sheet file, containing either the absolute or the relative paths to the Chimeric.out.junction produced by STAR for the paired-end run of each library. • A GTF-formatted annotation of repetitive regions which is used to filter out circRNA candidates from repetitive regions. • A list of all source BAM files containing the actual reads from the paired-end mapping step must be prepared. If not specified via the -B flag circtools will try to calculate the location based on the supplied Chimeric.out.junction paths. • The software additionally requires the SJ.out.tab files generated by STAR. The program assumes that these files are in the same folder as the BAM files supplied via -B. • For paired-end sequencing two files are required, e.g., mate1. txt and mate2.txt which contain the paths to the Chimeric.out.junction files originating from the separate mate mapping step. Circtools has the capability to support various sequencing protocols that are typically used for whole-transcriptome experiments. The user can choose between using stranded or unstranded input data by utilizing the -N flag. Stranded mode is the default option, and in the case of a stranded library, circtools can identify the location of a circRNA on either the sense or antisense strand. However, for an unstranded library, circtools assigns the strand of the circular based on the host gene’s strand because the aligned reads do not provide strand-specific information. Additionally, the -ss flag can be used to select the library preparation type based on the technique employed, either “firststrand” or “secondstrand.” The appropriate option depends on the sequencing protocol used

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

35

and cannot be determined in a general manner. Nonetheless, a list of common sequencing kits and their corresponding parameter choices is provided for reference. First-strand sequencing kits (default): • All dUTP methods, NSR, NNSR. • TruSeq Stranded Total RNA Sample Prep Kit. • TruSeq Stranded mRNA Sample Prep Kit. • NEB Ultra Directional RNA Library Prep Kit. • Agilent SureSelect Strand-Specific. ss

Second-strand sequencing kits (the second-strand parameter needs to be set):

• Directional Illumina (Ligation), Standard SOLiD. • ScriptSeq v2 RNA-Seq Library Preparation Kit. • SMARTer Stranded Total RNA. • Encore Complete RNA-Seq Library Systems. The following code will prepare all required files for the circRNA detection step performed by circtools: # short link: # https://links.jakobilab.org/circtools_step5 # go back to main folder cd .. # create new folder for circRNA detection mkdir -p circtools_detect/ # the following script automatically prepares all # required input files for the circtools detection step wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/prep_circtools.sh chmod 755 prep_circtools.sh # two parameters are required: # 1) STAR directory # 2) circtools directory ./prep_circtools.sh star/ circtools_detect/

Additionally, the FASTA reference genome (i.e., the genome sequence) of the corresponding species has to be downloaded. The version of the sequence has to match the sequence provided to

36

Tobias Jakobi

generate the STAR index. For this protocol, the ENSEMBL mouse genome GRCm38 version 90 was employed. The corresponding genome sequence together with a second file containing repetitive sequence regions in the mouse region can be downloaded using the code below: # download and unpack the mouse # reference genome sequence wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/mouse_genome_90.fa.gz gzip -d mouse_genome_90.fa.gz # download and unpack mouse repetitive regions wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/mouse_repeats.gtf.bz2 bzip2 -d mouse_repeats.gtf.bz2

Once all input files have been prepared, the circtools detect module can be started with the previously generated input files using a provided wrapper script. The script expects seven command line arguments, in order: 1. Sample sheet file: samplesheet file from previous listing. 2. GTF file: mouse.gtf from step 4. 3. Genome FASTA file: mouse_genome_90.fa. 4. Mate 1 file: mate1 from previous listing. 5. Mate 2 file: mate2 from previous listing. 6. BAM files: bam_files.txt from previous listing. 7. Repeat file: mouse_repeats.gtf. 8. Output directory, e.g., output/. # short link: https://links.jakobilab.org/circtools_step6 # download wrapper for circtools and make executable wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/circtools_detect.sh chmod 755 circtools_detect.sh # start circtools with the seven defined # parameters separated by space: ./circtools_detect.sh circtools_detect/samplesheet mouse.gtf mouse_genome_90.fa circtools_detect/

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

37

mate1 circtools_detect/mate2 circtools_detect/ bam_files.txt

mouse_repeats.gtf

circtools_de-

tect/output

The following output files are generated by circtools detect: •

A table containing read counts for detected circRNAs. The first three columns are chromosome, circRNA start position, and circRNA end position. From the fourth column on circRNA read counts are stored, one sample per column, shown in the order given in the sample sheet.

•

output/CircCoordinates: CircRNA annotations in BED format. The columns in order are chromosome, circRNA start position, circRNA end position, gene name, splice junction type (based on STAR: 0: non-canonical; 1: GT/AG, 2: CT/AC, 3: GC/AG, 4: CT/GC, 5: AT/AC, 6: GT/AT), circRNA strand, circRNA region for circRNA start and end (i.e., is the circRNA 5′ or 3′ junction within a gene, intergenic, exon, or intron), and overall regions (the genomic features circRNA the coordinates from start to stop cover).

•

output/LinearCount:

•

output/CircSkipJunctions:

output/CircRNACount:

Host gene expression count table, column order as in CircRNACount file.

CircRNA skipping junctions. The first three columns are the same as in LinearCount/CircRNACount, the following columns represent the circSkip junctions found for each sample. circSkip junctions are given as chr: start-end:count format, e.g., chr1:1787-6949:10. It is possible that for one circRNA multiple circSkip junctions are found since circRNAs may arise from different isoforms. In this case, multiple circSkip junctions are delimited by semicolon. A 0 implies that no circSkip junctions have been detected for this circRNA.

Performing Quality Assessment with Circtools Quickcheck The circtools quickcheck module is designed to equip users with a fast way of assessing the quality of the circRNA library preparation and the success of the mapping process. The quickcheck module requires that sequencing reads have been mapped with STAR, since internally the STAR log files are processed. CircRNA detection metrics are provided by circtools detect, which must be run prior to the call of the quickcheck module. For the following call of circtools quickcheck the circtools detect data located in the folder circtools_detect/output, the STAR mappings are stored in star/, the circSeq experiment has two conditions, listed via -l RNaseR_minus, RNaseR_plus; the samples in the circtools detection data files are in the order specified via -g 1,2 (i.e., sample 1: condition 1, sample 2: and condition 2).

38

Tobias Jakobi

Fig. 2 Circular versus linear read counts for all mapped libraries based on the performed circRNA detection workflow. RNase R-treated and untreated samples clearly separate in two clusters, with treated samples reaching higher circRNA counts and untreated samples reaching higher linear gene counts # For all command line parameters for quickcheck see # https://links.jakobilab.org/circtools_step7 circtools quickcheck -d circtools_detect/output/ -s star

-l

RNaseR_minus,RNaseR_plus

-g

1,2

-o

quickcheck/

The above program call generates a PDF report (Figs. 2, 3, and 4), which can be used to easily assess the outcome of the circRNA detection step and help to identify issues with the underlying sequencing library. Significance Testing of CircRNAs Using Circtools Circtest The circtest module of circtools allows to test the variation of circRNAs, e.g., between RNase R-treated and RNase R untreated sequencing libraries. It is recommended to supply the output of the circtools detect module, but the module can also be used on custom count tables. Required are one table with circular RNA counts and one table containing with host-gene counts. These tables have to have the same order, i.e., circRNA[i,j] and linear[i,j] are read counts for the same circRNA in the same sample. As for the other module tutorials, we use data from [22], processed with the detection module for the circtest module. Below is the sample call for the newly generated circtools detect data:

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

39

Fig. 3 Number of mapped reads versus number of detected circRNAs for all mapped libraries based on the performed circRNA detection workflow. Again, RNase R-treated and untreated samples separate into two groups, showing higher circRNA detection rates for the treated samples

Fig. 4 CircRNAs per million uniquely mapped reads based on the performed circRNA detection workflow. Specifically, samples P and R show high enrichment for circRNAs

40

Tobias Jakobi

Fig. 5 Example PDF report page showing the significant ( p < 0.05) enrichment of a circRNA of the Utrn gene in mouse in the RNase R-treated sequencing libraries compared to untreated libraries circtools circtest -d circtools_detect/output/ -p 0.05 -s 3 -r 4 -C 2 -g 1,2 -l RNaseR_minus,RNaseR_plus -c 4,5 -o circtest/

The circtest module generates an .xlsx Excel file that contains all circRNA candidates passing the statistical test with the given values, as well as the raw data files. Additionally, a PDF file is generated that contains a graphical representation of the top significant circRNAs (Fig. 5). Reconstruction of CircRNA Internal Sequences with Circtools Reconstruct The circtools reconstruct module is based on FUCHS (FUll circular RNA CHaracterization from RNA-Seq) [7], designed to fully characterize circular RNAs. It requires a list of circRNAs to reconstruct and reads spanning the back-splice junction as well as BAM files containing the mappings of all reads to perform this task. Generally, the longer the read length of the sequencing library, the better results can be expected from the reconstruction module. Specifically, long paired-end reads (>=150 bp) from Illumina MiSeq systems are known to perform well in this regard.

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

41

Reads from each circRNA are extracted by the reconstruction module and saved in individual BAM files. Based on these BAM files, circtools will detect alternative splicing within a circRNA, summarize different circular isoforms from the same host-gene, and generate coverage plots for each circRNA. It will also cluster circRNAs based on their coverage profile. These results can be used to identify potential false positive circRNAs. To characterize the sequence of circRNAs from RNA-seq data, the following preprocessing steps are necessary: • Mapping of RNA-seq data from quality checked FASTQ files with STAR (step 4) • Detection of circRNAs using circtools detect (step 5) We continue by using the same data set that also has been used as an example for the circtools detect module. The following parameters are used to reconstruct circRNAs in samples based on a previous circtools detect run: 1. Sample name to run reconstruction on 2.

../reconstruct/:

3.

A BED file containing exons of the target organism’s genome

4.

../circtools_detect/:

5.

../circtools_detect/output/:

Output directory for the reconstruct module, here the relative path

../mm10.ensembl.exons.bed:

Main directory of circtools detect Output directory of cir-

ctools detect # short link:

# https://links.jakobilab.org/circtools_step8 # download the wrapper scripts for the # reconstruct module wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

WOW64)" https://links.jakobilab.org/reconstruct.sh # add execute permission chmod 755 reconstruct.sh # create output directory mkdir reconstruct # download exon annotations required # by the reconstruct module # unpack with bunzip2 (archive is of type BZIP2) wget

--user-agent="Mozilla/5.0

WOW64)"

(Windows

NT

10.0;

42

Tobias Jakobi https://links.jakobilab.org/mm10.ensembl.exons. bed.bz2 bunzip2 mm10.ensembl.exons.bed.bz2 # enter star/ folder for samples cd star/ # circtools reconstruct is independently run # on all samples: parallel -j1 reconstruct.sh {} ../reconstruct/ ../ mm10.ensembl.exons.bed ../circtools_detect/ ../ circtools_detect/output/ :: : *

The convenience script guided_denovo_circle_struccan be employed to obtain a more refined circRNA reconstruction based on intron signals. The circRNA-separated BAM files are required for the script. If an annotation file is supplied, unsupported exons will be reported with a score of 0, if no annotation file is supplied, unsupported exons will not be reported. The following call will execute the reconstruction script and requires three mandatory arguments:

ture

1.

-N:

2.

A BED file containing exons of the target organism’s genome (from step 8)

3.

-I:

Name of the sample to perform the guided de novo reconstruction on

-A:

Input directory; directory with reconstruction data from step 8

# call from within the star/ folder: parallel -j1 guided_denovo_circle_structure -N {} -A ../mm10.ensembl.exons.bed -I ../reconstruct/ ::: *

The output generated by circtools reconstruct consists of several output files and folders which bundle information gathered during the reconstruction process. The following list describes files and their contents (a detailed, technical description together with additional reconstruct parameters can be found via the short link provided in the code block). The following data are generated by circtools reconstruct: • sample_name/: Project folder containing all files generated by circtools reconstruct. • sample_name.coverage_pictures & sample_name.coverage_profiles: Automatically generated PDF files containing a visual representation of all circRNAs.

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

43

• sample_name.alternative_splicing.txt: This file summarizes the relationship of different circRNAs derived from the same host gene. • sample_name.exon_counts.bed: A BED-formatted file that describes the circRNA exon structure and can be loaded into any genome browser. Each line corresponds to one circRNA. • sample_name.exon_counts.txt: This file contains similar information as the previous file, but adds more detailed information on the exons. Each line corresponds to one exon. • sample_name.mate_status.txt: This output file contains the results of analyzing the amount of how often each fragment spans a chimeric junction. A fragment can either span the chimeric junction once (single), only one end spans the junction, twice (double) both ends span the chimeric junction, or neither of both (undefined). • sample_name.skipped_exons.bed: A BED12-based file containing skipped exons. • sample_name.skipped_exons.txt: Identical to the BED-based skip file, but in addition contains read information and other details. • sample_name_exon_chain_inferred_12.bed: Output generated by the de novo reconstruction script in BED12 format, viewable with any BED-enabled visualization tool. • sample_name_exon_chain_inferred_6.bed: Output generated by the de novo reconstruction script in BED6 format, viewable with any BED-enabled visualization tool. Designing CircRNA-Specific Primers via Circtools Primex The circtools primex module is a highly specialized primer design tool, tailored specifically for circRNA experiments. The module generates primers that are specific to the BSJ of circRNAs, thus making it possible to amplify only selected circRNAs, but not the host gene as the primer pairs are divergent and will not generate a product if the target does not have the correct circular form. The module offers complex designs that can also target the forward or reverse primer directly to the BSJ if required, e.g., due to sequence constraints. Additionally, circtools primex can generate primers based on FASTA sequences, which is useful in cases of artificial or genetically altered circRNAs that are not corresponding to circRNAs originating from the species reference genome. Circtools primex designs primer pairs for specific circRNAs based on coordinates or gene names, but is also capable to generate batches of hundreds of circRNAs for larger panels. In this case, circRNAs detected with circtools detect are recommended as input, but circtools can also work on lists with specific circRNA isoforms

44

Tobias Jakobi

from other tools or even entirely without any preliminary data purely based on the FASTA sequence of the circRNA. A sample call to primex using data from [22] processed with circtools detect requires as only external parameter the FASTA sequence of the reference genome in order to obtain DNA sequences for the primer design process. The following parameters are used to generate a list of primers all circRNAs originating from the gene Utrn: 1.

file from circtools run containing positions and annotations of detected circRNAs, output/CircCoordinates

2. Genome FASTA file: mouse_genome_90.fa from step 5. 3. GTF file: mouse.gtf from step 4. 4. Species code: search.

mm

for mus musculus, used for BLAST [26]

5. Gene name, e.g., Utrn. 6. Name of the primer design project, e.g., “Utrn primers”. # short link: https://links.jakobilab.org/circtools_step9 # obtain reference sequence and annotation # (if not already downloaded) # download and unpack wget

--user-agent="Mozilla/5.0

(Windows

NT

10.0;

NT

10.0;

WOW64)" https://links.jakobilab.org/mouse.gtf.gz wget

--user-agent="Mozilla/5.0

(Windows

WOW64)" https://links.jakobilab.org/mouse_genome_90.fa.gz gzip -d mouse_genome_90.fa.gz gzip -d mouse.gtf.gz # run circtools primex and design # primers for Utrn circRNA circtools primex -d output/CircCoordinates -f mouse_genome_90.fa -g mouse.gtf -O mm -G Utrn -T " Utrn primers"

By default, the primer design process requires an internet connection to perform a BLAST search for potential off-site targets. This feature is currently available for four species: human, mouse, rat, and pig. The final output of the primer design module is an interactive HTML file that can be viewed in any browser (Fig. 6).

State-of-the-Art Circular RNA Analytics Using the Circtools Software Suite

45

Fig. 6 Interactive HTML output of the circtools primer module for circRNAs of the Utrn host gene in mouse. The columns with important parameters such as GC content, melting temperature, and potential off-site targets detected by BLAST are color coded for easy inspection of auspicious primer pairs

Moreover, the primer design module can be precisely configured to design custom primer, i.e., product size, GC content, melting temperature, and other feature can be user-controlled (see full documentation via the short link). By default, these parameters are set to values that we have confirmed to generate well-working qRT-PCR primers in different species.

4 Conclusion Circtools was designed to cover the whole circRNA computational analysis workflow, from RNA-seq raw data to primer design for follow-up wet lab experiments. We strive to continuously improve circtools and add new functionality. As such, visiting us at circ.tools is recommend to check for new releases of circtools and interesting new features designed to support and empower the circRNA research community. References 1. Hoffmann S et al (2014) A multi-split mapping algorithm for circular RNA, splicing, transsplicing and fusion detection. Genome Biol 15(2):R34. https://doi.org/10.1186/gb2014-15-2-r34 2. Wang K et al (2010) MapSplice: accurate mapping of RNA-seq reads for splice junction discovery. Nucleic Acids Res 38(18):1–14. https://doi.org/10.1093/nar/gkq622 3. Memczak S et al (2013) Circular RNAs are a large class of animal RNAs with regulatory potency. Nature 495(7441):333–338. https://doi.org/10.1038/nature11928

4. Chuang TJ, Wu CS, Chen CY, Hung LY, Chiang TW, Yang MY (2015) NCLscan: accurate identification of non-co-linear transcripts (fusion, trans-splicing and circular RNA) with a good balance between sensitivity and precision. Nucleic Acids Res 44(3):e29. https://doi.org/ 10.1093/nar/gkv1013 5. Westholm JO et al (2014) Genome-wide analysis of drosophila circular RNAs reveals their structural and sequence properties and age-dependent neural accumulation. Cell Rep 9(5):1966–1980. https://doi.org/10.1016/j. celrep.2014.10.062

46

Tobias Jakobi

6. Zhang XO, Wang HB, Zhang Y, Lu X, Chen LL, Yang L (2014) Complementary sequencemediated exon circularization. Cell 159(1): 134–147. https://doi.org/10.1016/j.cell. 2014.09.001 7. Cheng J, Metge F, Dieterich C (2016) Specific identification and quantification of circular RNAs from sequencing data. Bioinformatics 32(7):1094–1096. https://doi.org/10.1093/ bioinformatics/btv656 8. Szabo L et al (2015) Statistically based splicing detection reveals neural enrichment and tissuespecific induction of circular RNA during human fetal development. Genome Biol 16: 126. https://doi.org/10.1186/s13059-0150690-5 9. Dobin A et al (2013) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29(1):15–21. https://doi.org/10.1093/bioinformatics/ bts635 10. Langmead B, Salzberg SL (2012) Fast gappedread alignment with Bowtie 2. Nat Methods 9(4):357–359. https://doi.org/10.1038/ nmeth.1923 11. Li H (2013) Aligning sequence reads, clone sequences and assembly contigs with BWA-MEM. 00(00):1–3. arXiv:1303.3997 [q-bio.GN] 12. Jeck WR et al (2013) Circular RNAs are abundant, conserved, and associated with ALU repeats. RNA 19(2):141–157. https://doi. org/10.1261/rna.035667.112 13. Panda AC et al (2017) High-purity circular RNA isolation method (RPAD) reveals vast collection of intronic circRNAs. Nucleic Acids Res 45(12):e116. https://doi.org/10.1093/ nar/gkx297 14. Xiao M-S, Wilusz JE (2019) An improved method for circular RNA purification using RNase R that efficiently removes linear RNAs containing G-quadruplexes or structured 3′ ends. Nucleic Acids Res 47(16):8755–8769. https://doi.org/10.1093/nar/gkz576 15. Gru¨ning B et al (2018) Bioconda: sustainable and comprehensive software distribution for the life sciences. Nat Methods 15(7): 475–476. https://doi.org/10.1038/s41592018-0046-7

16. Li H et al (2009) The sequence alignment/ map format and SAMtools. Bioinformatics 25(16):2078–2079. https://doi.org/10. 1093/bioinformatics/btp352 17. Van Rossum G, Drake FL et al (1995) Python reference manual. Centrum voor Wiskunde en Informatica, Amsterdam 18. Jakobi T, Uvarovskii A, Dieterich C (2019) Circtools-a one-stop software solution for circular RNA research. Bioinformatics 35(13): 2326–2328. https://doi.org/10.1093/bioin formatics/bty948 19. Andrews S (2012) FastQC. A quality control tool for high throughput sequence data. http://www.bioinformatics.bbsrc.ac.uk/pro jects/fastqc/ 20. Roehr JT, Dieterich C, Reinert K (2017) Flexbar 3.0 – SIMD and multicore parallelization. Bioinformatics 33(18):2941–2942. https:// doi.org/10.1093/bioinformatics/btx330 21. Tange O et al (2011) Gnu parallel-the command-line power tool. The USENIX Magazine 36(1):42–47 22. Jakobi T, Czaja-Hasse LF, Reinhardt R, Dieterich C (2016) Profiling and validation of the circular RNA repertoire in adult murine hearts. Genomics Proteomics Bioinformatics 14(4): 216–223. https://doi.org/10.1016/j.gpb. 2016.02.003 23. Hansen TB, Venø MT, Damgaard CK, Kjems J (2016) Comparison of circular RNA prediction tools. Nucleic Acids Res 44(6):e58. https:// doi.org/10.1093/nar/gkv1458 24. Yoo AB, Jette MA, Grondona M (2003) Slurm: Simple linux utility for resource management. In: Job scheduling strategies for parallel processing: 9th international workshop, JSSPP 2003, Seattle, WA, USA, June 24, 2003. Revised Paper 9, pp 44–60 25. Vromman M et al (2022) Large-scale benchmarking of circRNA detection tools reveals large differences in sensitivity but not in precision. bioRxiv, p 2022.12.06.519083. https:// doi.org/10.1101/2022.12.06.519083 26. Altschul SF, Gish W, Miller W, Myers EW, Lipman DJ (1990) Basic local alignment search tool. J Mol Biol 215(3):403–410. https:// doi.org/10.1016/S0022-2836(05)80360-2

Chapter 3 Exploring Circular RNA Profile and Expression in Extracellular Vesicles Jingjing Zhao, Qiaojuan Li, and Shenglin Huang Abstract Extracellular vesicles (EVs) are small vesicles secreted by various cell types and are enriched in multiple body fluids. EVs containing RNA have the potential to modulate biological processes and are being investigated for their diagnostic and therapeutic applications. Circular RNAs (circRNAs), generated through backsplicing of exons, are enriched in EVs. Given their unique characteristics and diverse functions, EV-circRNAs are important players in disease pathology. This chapter describes a workflow for investigating the expression profile of EV-circRNAs, which includes EVs separation, library preparation, and bioinformatics analysis. This workflow can aid the investigation of EV-circRNAs and their potential role in disease pathology. Key words Extracellular vesicles, Circular RNA, RNA isolation, Bioinformatics, CIRI2, ASJA

1

Introduction Extracellular vesicles (EVs) are a heterogeneous class of nano- to micrometer-sized membrane-bound vesicles, mainly categorized as exosomes (derived from endosomal membrane) and microvesicles (derived from plasma membrane) [1–3]. These vesicles are abundant in body fluids such as blood, urine, saliva, bile, and cerebrospinal fluid (CSF) [4, 5]. EVs contain a wide range of molecular cargo, including proteins, lipids, DNA, and RNA, which can regulate various physiological and pathological processes [6–9]. While small RNAs were initially the focus of EV RNA studies, whole transcriptome analyses have enabled the characterization of diverse EV RNA species, such as protein-coding RNA (mRNA), long non-coding RNA (lncRNA), and circular RNA (circRNA) [10, 11]. CircRNAs, generated through back-splicing, exhibit a range of functions, including microRNA sponging and protein interactions [12], and can constitute up to 6% of the total RNA content of EVs, with a higher back-splicing ratio than that of cells

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_3, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

47

48

Jingjing Zhao et al.

Fig. 1 This schematic illustration outlines the process of identifying circRNAs in EVs, which includes the collection of EVs from body fluids, isolation of EV RNA, construction of RNA-seq libraries for sequencing, and the identification of circRNAs through bioinformatics analysis

[10]. Notably, EV-circRNAs have been detected in EVs derived from various cancer types, including breast cancer (BC), colorectal cancer (CRC), and hepatocellular carcinoma (HCC) [11], and have been implicated in cancer development, such as circATP2B4 promoting metastasis of epithelial ovarian cancer [13] and circPDK1 promoting glycolysis in pancreatic cancer [14]. Given their high stability, EV-circRNAs have potential as disease biomarkers [15– 17]. However, their low expression levels and specific structures make the detection and identification of EV-circRNAs challenging. In this chapter, we outline a comprehensive workflow for obtaining circRNA profiles in EVs, comprising of EVs separation, library preparation, and bioinformatics analysis. The methodology is depicted in Fig. 1 as a schematic illustration.

2 Materials 2.1 Biofluids Separation and Storage

1. Whole blood in tube containing EDTA for plasma separation. Other biofluids see Note 1.

2.2 Purification of EVs

1. exoRNeasy Serum/Plasma Kit (Qiagen, Germany) (see Note 2). 2. 400 ul XE elution buffer (Cat. No. 76214, Qiagen, Germany). 3. Amicon Ultra-0.5 Centrifugal Filter 10 kDa (Merck Millipore, Germany).

2.3

EV RNA Isolation

1. miRNeasy (RNeasy MinElute spin column) (included in exoRNeasy Serum/Plasma Kit). 2. QIAzol (Qiagen, Germany). 3. Chloroform. 4. 100% ethanol.

Identification of CircRNAs in Extracellular Vesicles

2.4 Library Preparation for RNASeq

49

1. DNase I (NEB, USA). 2. SMARTer® Stranded Total RNA-Seq Kit - Pico Input Mammalian (Clontech, USA) (see Note 3). 3. RNA Clean beads (Vazyme, China). 4. VAHTS DNA Clean beads (Vazyme, China). 5. Qubit (Thermo Fisher Scientific, USA).

2.5 Hardware and Environment for EVcircRNA Analysis

To align sequences to a genome with STAR and BWA, a Linux operation system with 30 GB of RAM is recommended.

2.6 Software and Environment for EVcircRNA Analysis

The following software programs are utilized in the RNA-seq analysis to identify EV-circRNAs. Trimmomatic and FastQC are used for quality control of the sequencing data. The identification of circRNAs using ASJA requires STAR, StringTie, and Sambamba. Additionally, BWA is utilized in the circRNA detection program, CIRI2. The EV-circRNA analysis program’s GitHub repository is loacted at https://github.com/HuangLab-Fudan/circRNA. 1. STAR (version 2.5.3a) [18], an aligner tool, obtainable via https://github.com/alexdobin/STAR. 2. BWA (version 0.7.17-r1188) [19], an aligner tool, obtainable via https://sourceforge.net/projects/bio-bwa/files/. 3. FastQC (version v0.12.1), a quality control tool, obtainable via http://www.bioinformatics.babraham.ac.uk/projects/down load.html#fastqc. 4. Trimmomatic (version 0.36) [20], a trimmer for Illumina sequence data, obtainable via https://github.com/usadellab/ Trimmomatic. 5. Sambamba (version 0.66) [21], SAM and BAM files manipulation tool, obtainable via https://github.com/biod/ sambamba. 6. StringTie (version 2.12) [22], a transcript assembly and quantitation tool, obtainable via https://github.com/gpertea/ stringtie. 7. R software environment (version 4.0.2), a statistical computing and graphics environment, obtainable via https://cran.rproject.org/bin/linux/. 8. ASJA (version master) [23], identification junction program, obtainable via https://github.com/HuangLab-Fudan/ASJA. 9. CIRI2 (version 2.06) [24], identification circRNA tool, obtainable via https://sourceforge.net/projects/ciri/files/ CIRI2/.

50

Jingjing Zhao et al.

10. BEDTools (version v2.30.0) [25], a potent toolkit for genomic mathematics, obtainable via https://github.com/arq5x/ bedtools2/releases/download/.

3

Method

3.1 Biofluids Separation and Storage

1. Collect 5–10 mL whole blood in a tube containing EDTA. Store the tube at room temperature (15–25 °C) and process within 2 h. 2. Blood samples in primary blood collection tubes are centrifuged for 10 min at 3000 rpm (room temperature) in a swinging bucket rotor. Carefully transfer the upper plasma phase to a new tube. 3. Plasma samples are centrifuged for 10 min at 13,000× g and 4 ° C; Carefully transfer cleared supernatant to a new tube. Store at 2–8 °C until further processing if plasma will be used for isolation of EVs on the same day. For longer storage, keep plasma frozen in aliquots at -80 °C.

3.2 Purification of EVs

1. EVs are purified using the exoRNeasy Serum/Plasma kit. Add 1 volume buffer XBP to 1 volume of plasma (about 1 mL). Mix well immediately by gently inverting the tube five times. 2. Add the sample/XBP mix onto the exoEasy spin column and centrifuge for 1 min at 500× g. Discard the flow-through and place the column back into the same collection tube. 3. Add 3.5 mL XWP and spin 5 min at 5000× g to wash the column and remove residual buffer. Discard the flow-through together with the collection tube.

3.3

EV-RNA Isolation

1. Add 700 μL QIAzol to the membrane, incubate at room temperature (15–25 °C) for 2–3 min. Spin for 5 min at 5000× g to collect the lysate and transfer completely to a 2 mL tube. If any spike-in controls are used, they should be added to the lysate at this point. 2. Add 90 μL chloroform to the tube containing the lysate. Shake vigorously for 15 s. Incubate at room temperature (15–25 °C) for 2–3 min. Centrifuge for 15 min at 12,000× g at 4 °C. The centrifugation separates the sample into three phases: an upper, colorless aqueous phase containing RNA; a thin, white interphase; and a lower, red organic phase. The RNA is present in the upper aqueous phase, which can be carefully removed using a pipette and transferred to new collection tube. 3. Add 2 volumes of 100% ethanol and mix thoroughly by pipetting up and down several times. Pipet up to 700 μL sample into a RNeasy MinElute spin column. Close the lid gently and

Identification of CircRNAs in Extracellular Vesicles

51

centrifuge at ≥8000× g (≥10,000 rpm) for 15 s at room temperature (15–25 °C). Discard the flow-through. Repeat this procedure until all samples have been centrifuged. 4. In order to further purify the RNA and remove any remaining contaminants, it is recommended to wash the RNeasy MinElute spin column with Buffer RWT and Buffer RPE. This step is described in detail in the exoRNeasy Midi/Maxi Handbook, which can be found at www.qiagen.com/HB-2630. 5. Place the RNeasy MinElute spin column into a new 2 mL collection tube. Open the lid of the spin column, and centrifuge at full speed for 5 min to dry the membrane. Discard the collection tube with the flow-through. 6. Place the RNeasy MinElute spin column in a new 1.5 mL collection tube. Add 14 μL RNase-free water directly to the center of the spin column membrane. Close the lid gently, let column stand for 1 min and then centrifuge for 1 min at full speed to elute the RNA (see Note 1). 3.4 Library Preparation for RNASeq

To ensure the highest quality RNA-seq libraries, we utilized the SMARTer® Stranded Total RNA-Seq Kit – Pico Input Mammalian (Clontech) and made several key optimizations to the standard procedures. This protocol accommodates up to 8 μL of input RNA ranging from 0.25 to 10 ng. 1. Input EV RNAs should be free from genomic or carrier DNA and other contaminants, as the subsequently random priming used may lead to amplification of any DNA present in the starting material. Samples should have been treated with DNase I. The EV RNAs were treated with 0.5 μl DNase I (NEB) for 15 min at 37 °C, and purified by RNA Clean beads (Vazyme, China). 2. Mix the following components: RNA, SMART Pico Oligos Mix, 5X First-Strand Buffer and Nuclease-Free Water for fragmentation and first-strand cDNA synthesis. The purified RNAs were fragmented at 85 °C for 4 min 30 s, then immediately place the samples at 4 °C for 2 min. 3. Add 7 μL of the First Strand Master Mix to each reaction tube. Notably, use six cycles for the first round PCR to addition of Illumina Adapters and Indexes. 4. Use 16 cycles for the second round PCR of final RNA-seq library amplification. 5. Purification of final RNA-seq library using VAHTS DNA Clean beads (Vazyme, China).

52

Jingjing Zhao et al.

6. Quantify libraries with Qubit (>1 ng/μL), evaluate library size distribution with a local maximum at ~300–400 bp. 7. The libraries were sequenced on Illumina sequencing system. 3.5 Quality Control for Sequencing Data

1. FastQC is a tool for testing the quality of raw sequencing data, including basic statistics sequence quality, GC and N content, length distribution, and adapter content module. The command line (see Note 4) is as follows: We use exoN06 (N6) as example, which is a blood EV healthy control sample that was downloaded from GSE100206. >fastqc -t 2 -o out_path exoN06_R1.fastq.gz exoN06_ R2. fastq.gz 2. Removal of adaptors from sequencing is necessary for RNA-seq analysis to ensure accuracy. The Trimmomatic can quickly and flexibly remove Illumina sequential data adaptors. The pairedend reference commands are as follows: >java -jar trimmomatic-0.33.jar PE -threads 10 exoN06_R1. fastq.gz exoN06_R2.fastq.gz exoN06_1P.fq exoN06_1U. fq exoN06_2P.fq exoN06_2U.fq ILLUMINACLIP: path_of_trimmomatic/adapters/TruSeq3-PE.fa:2:30:12: 1:true LEADING:3 TRAILING:3 SLIDINGWINDOW:4:15 MINLEN:36

3.6 EV CricRNA Detection with CIRI2

CIRI2 is a perl package to identify and quantify circRNA from transcriptome. CIRI2 works with BWA alignment with reference genome. The following steps can obtain the circRNA of exoN06 by CIRI2. 1. Both BWA and STAR (an aligner that works with ASJA) alignment require the reference genome and annotation files. They can be download from NCBI, Ensembl, or GENCODE. #Download human reference genome fasta and gene annotation file from GENCODE. #Create a new directory named referenceFile >mkdir referenceFile cd referenceFile >wget https://ftp.ebi.ac.uk/pub/databases/gencode/ Gencode_human/release_29/GRCh38.p12.genome. fa.gz >wget https://ftp.ebi.ac.uk/pub/databases/gencode/ Gencode_human/release_29/gencode.v29.annotation. gtf.gz # Decompress files >gunzip GRCh38.p12.genome.fa.gz

Identification of CircRNAs in Extracellular Vesicles

53

>gunzip gencode.v29.annotation.gtf.gz #Create a directory to save the result of BWA alignment mkdir exoN06 >cd exoN06 >bwa mem -T 19 -t 12 referenceFile/GRCh38.p12.genome. fa.gz exoN06_1P.fq.gz exoN06_2P.fq.gz > aln-pe.sam 2. CIRI2 source files can be obtained from sourceforge.net and utilized without further configuration. path_ciri2 directory is the installation directory of CIRI2. >cd exoN06 perl path_ciri2/CIRI2.pl -I aln-pe.sam -0 -O outfile -F referenceFile/GRCh38.p12.genome.fa.gz -A referenceFile / gencode.v29.annotation.gtf -T 12 Here, CIRI2 is run on 12 CPUs (-T 12), the input SAM file is specified via -I, and the output directory is specified via -O. Outfile is a table including the ID of circRNA, read counts, and their annotation. 3.7 EV CricRNA Detection with ASJA

The ASJA program can identify both known and novel back-spliced junctions, fusion, and linear junctions. The primary steps of ASJA include genomic alignment by STAR, transcript assembly by stringtie, and junction detection and quantification. Users can download the ASJA program from GitHub and run perl ASJA-all.pl [parameters] to obtain the junctions and their annotations, or follow the steps below to obtain the results step by step. 1. The initial step in STAR alignment is to create a STAR index with the reference genome and its annotation. >cd .. STAR --runThreadN 20 \ --runMode genomeGenerate --genomeDir outpath/STAR_index \ --genomeFastaFiles referenceFile/GRCh38.p12.genome.fa --sjdbGTFfile referenceFile/gencode.v29.annotation.gtf --sjdbOverhang 100 A STAR index is created using 20 threads (--runThreadN 20) with an input reference genome (--genomeFastaFiles) and its annotation (--sjdbGTFfile), and it can be found in the outpath/STAR_index directory. 2. Sequencing with quality control and removal of the adaptor is aligned to the reference genome by STAR, and Two-pass mapping alignment is recommended. The first pass alignment is run to obtain the junction file exoN06_x1SJ.out.tab.

54

Jingjing Zhao et al.

>cd exoN06 >STAR \ --genomeDir outpath/STAR_index \ --readFilesCommand zcat \ --readFilesIn exoN06_1P.fq.gz exoN06_2P.fq.gz \ --runThreadN 8 \ --outFilterMultimapNmax 20 \ --alignIntronMax 500000 \ --alignMatesGapMax 1000000 \ --alignSJDBoverhangMin 1 \ --outFilterMatchNminOverLread 0.33 \ --outFilterScoreMinOverLread 0.33 \ --outSAMstrandField intronMotif \ --outSAMtype None \ --outSAMmode None \ --outFileNamePrefix exoN06_x1 # The first pass alignment is run with eight CPUs for pairedend exoN06 (--readFilesIn). The established reference genome index should be inputted (--genomeDir). The file suffix is set to sample_x1 (--outFileNamePrefix) to distinguish it from the second alignment. The purpose of the first alignment is to get the junction file (sample_x1SJ.out.tab), the SAM file can be set to NONE to save time and memory. 3. A new index is created named “exoN06_Index2” with reference genome and junction file (--sjdbFileChrStartEnd), and the output directory is specified by --genomeDir. >mkdir exoN06_Index2 >STAR \ --runMode genomeGenerate \ --genomeDir exoN06_Index2 \ --genomeFastaFiles refenceFiles/GRCh38.p12.genome.fa \ --sjdbOverhang 100 \ --runThreadN 12 \ --sjdbFileChrStartEnd exoN06_x1SJ.out.tab 4. The second pass alignment is run to obtain the genomic positions of the junctions. > STAR \ --genomeDir exoN06_Index2 \

Identification of CircRNAs in Extracellular Vesicles

55

--readFilesCommand zcat \ --readFilesIn exoN06_1P.fq.gz exoN06_2P.fq.gz \ --runThreadN 8 \ --outFilterMultimapNmax 20 \ --alignIntronMax 500000 \ --alignMatesGapMax 1000000 \ --alignSJDBoverhangMin 1 \ --outFilterMatchNminOverLread 0.33 \ --outFilterScoreMinOverLread 0.33 \ --outSAMstrandField intronMotif \ --outSAMattributes NH HI NM MD AS XS \ --outSAMunmapped Within \ --outSAMtype BAM SortedByCoordinate \ --outSAMheaderHD @HD VN:1.4 Sergey \ --chimOutType WithinBAM \ --chimSegmentMin 20 The second pass alignment is run with exoN06_Index2 and obtain alignment bam file “Aligned.sortedByCoord.out. bam” and junction file “Chimeric.out.junction.” Most settings are the same as the first. Chimeric and circular alignments results are saved to identify the circRNAs with -chimSegmentMin and –chimOutType settings. BAM files can also be saved as sorted with setting of -outSAMtype. ## Create an index for BAM file for downstream analysis. >mv Aligned.sortedByCoord.out.bam reads.bam

exoN06_mapped_-

> sambamba index exoN06_mapped_reads.bam 5. Next, transcripts are assembled by StringTie with reference gene annotations. > stringtie exoN06_mapped_reads.bam -f 0.1 -o stringtie_assembly.gtf -p 12 -G referenceFile/gencode.v29.annotation.gtf 6. ASJA is designed to identify circRNA working with Chimeric. out.junction file generated by two-pass alignment of STAR. ASJA depends on perl and Shell. Source files can be obtained from GitHub and utilized without further configuration. path_asja directory is the installation directory of ASJA. The following command can obtain all junctions including circRNA.

56

Jingjing Zhao et al.

# Preparing reference file for annotation junctions >perl path_asja/ASJA.pl -I path_asja -G referenceFile/gencode.v29.annotation.gtf -setup #All junctions of exoN06 can be identified >mkdir exoN06_results >perl path_asja/ASJA-all.pl -I path_asja -G referenceFile/gencode.v29.annotation.gtf -SI exoN06/stringtie_assembly. gtf -CI exoN06 -O exoN06_results ASJA is run based on the result of STAR and stringtie, the input STAR result directory is specified via -CI, the input result file of stringtie and directory is specified via -SI, and the output directory of junction is specified via -O. The output of ASJA contains a “raw” and “fitter” directory that saves unfiltered and filtered junctions, respectively. The “raw” directory contains three subfolders, circ, fusion, and norm, which store circRNA, fusion, and linear junction, respectively. circRNA.txt is a table containing the ID of circRNA, read counts, and their annotation. The specific path is exoN06_results /raw/circ/circRNA.txt 3.8 Characterization of EV-CircRNA Profile

In order to ensure the accuracy of the identified circRNA in EVs, the intersection of the results of CIRI2 and ASJA is used to define detectable circRNAs. Using the following command in the R environment allows for obtaining of the host genes, splicing ratio, and genomic origin of circRNA. 1. Reading the result of ASJA and CIRI2. >setwd(’D:/exoN06) >ASJA.rescirc2.rescircIDcirc2.res$circIDfinal.resfinal.res.v2library(ggplo2) >bsj bsj bsj$Var1ggplot(bsj,aes(x=Var1,y=Freq))+geom_col(fill=’grey75’) +theme_classic()+ylab(c("The number of circRNAs")) +xlab(c(’Back-splicing junction reads’))+scale_x_continuous(breaks = seq(0,5000,100)) 3. Obtaining the genomic origin of EV-circRNAs (Fig. 2b). >orgorgcolnames(org)% mutate(lab.ypos = cumsum(Freq) - 0.4*Freq) >ggplot(org, aes(x = "", y = Freq, fill = type)) +geom_bar (stat = "identity") +coord_polar("y", start = 0)+geom_text(aes(y = lab.ypos, label = ratio,vjust=0.01,hjust=0.4), color = "black")+theme_void() 4. Obtaining the distribution of junction reads ratio in EV-circRNAs (Fig. 2c). >range num numggplot(num,aes(x=Var1,y=Freq))+geom_col(fill=’grey75’) +theme_classic()+ylab(c("The number of circRNAs")) +xlab(c(’The range of junction reads ratio’))

4

Notes 1. In this chapter, blood (plasma) samples were utilized for EV-RNA isolation. However, other biofluids such as urine, saliva, bile, cerebrospinal fluid (CSF), and cell supernatant can also be used for EV-RNA isolation. It is essential to collect the biofluids in a fresh and sterile manner to ensure the integrity of the EVs and the RNA (see Subheading 2.1). 2. For the isolation of EV RNA, we recommend using the exoRNeasy Serum/Plasma Kit, which is also suitable for the isolation of EV RNA from various biofluids (see Subheading 2.2). 3. The SMARTer® Stranded Total RNA-Seq Kit-Pico Input Mammalian is essential for EV RNA-seq library preparation. This kit enables pre-amplification of cDNA and utilizes CRISPR/Cas9 to deplete rRNA dsDNA (see Subheading 2.4). 4. The Linux environment is used for all commands, except for Subheading 3.8. (see Subheading 3.5).

Acknowledgments This work was supported by grants from the National Key Research and Development Project of China (2021YFA1300500), Shanghai Science and Technology Innovation Action Plan (20JC1419000), and the National Natural Science Foundation of China (82072694, 82272625). References 1. Thery C, Witwer KW, Aikawa E, Alcaraz MJ, Anderson JD, Andriantsitohaina R et al (2018) Minimal information for studies of extracellular vesicles 2018 (MISEV2018): a position statement of the International Society for Extracellular Vesicles and update of the MISEV2014 guidelines. J Extracell Vesicles 7(1):1535750 2. Bordanaba-Florit G, Royo F, Kruglik SG, Falcon-Perez JM (2021) Using single-vesicle technologies to unravel the heterogeneity of extracellular vesicles. Nat Protoc 16(7): 3163–3185 3. Hu W, Liu C, Bi ZY, Zhou Q, Zhang H, Li LL et al (2020) Comprehensive landscape of extracellular vesicle-derived RNAs in cancer

initiation, progression, metastasis and cancer immunology. Mol Cancer 19(1):102 4. Murillo OD, Thistlethwaite W, Rozowsky J, Subramanian SL, Lucero R, Shah N et al (2019) exRNA atlas analysis reveals distinct extracellular RNA cargo types and their carriers present across human biofluids. Cell 177(2): 463–477 e15 5. Hulstaert E, Morlion A, Avila Cobos F, Verniers K, Nuytens J, Vanden Eynde E et al (2020) Charting extracellular transcriptomes in the human biofluid RNA atlas. Cell Rep 33(13):108552 6. Pathan M, Fonseka P, Chitti SV, Kang T, Sanwlani R, Van Deun J et al (2019)

Identification of CircRNAs in Extracellular Vesicles Vesiclepedia 2019: a compendium of RNA, proteins, lipids and metabolites in extracellular vesicles. Nucleic Acids Res 47(D1):D516– D519 7. Yanez-Mo M, Siljander PR, Andreu Z, Zavec AB, Borras FE, Buzas EI et al (2015) Biological properties of extracellular vesicles and their physiological functions. J Extracell Vesicles 4: 27066 8. Russo F, Di Bella S, Vannini F, Berti G, Scoyni F, Cook HV et al (2018) miRandola 2017: a curated knowledge base of non-invasive biomarkers. Nucleic Acids Res 46(D1):D354–D359 9. Han QF, Li WJ, Hu KS, Gao J, Zhai WL, Yang JH et al (2022) Exosome biogenesis: machinery, regulation, and therapeutic implications in cancer. Mol Cancer 21(1):207 10. Li Y, Zhao J, Yu S, Wang Z, He X, Su Y et al (2019) Extracellular vesicles long RNA sequencing reveals abundant mRNA, circRNA, and lncRNA in human blood as potential biomarkers for cancer diagnosis. Clin Chem 65(6): 798–808 11. Lai H, Li Y, Zhang H, Hu J, Liao J, Su Y et al (2022) exoRBase 2.0: an atlas of mRNA, lncRNA and circRNA in extracellular vesicles from human biofluids. Nucleic Acids Res 50 (D1):D118–D128 12. Kristensen LS, Jakobsen T, Hager H, Kjems J (2022) The emerging roles of circRNAs in cancer and oncology. Nat Rev Clin Oncol 19(3): 188–206 13. Wang F, Niu Y, Chen K, Yuan X, Qin Y, Zheng F et al (2023) Extracellular vesicle-packaged circATP2B4 mediates M2 macrophage polarization via miR-532-3p/SREBF1 Axis to promote epithelial ovarian cancer metastasis. Cancer Immunol Res 11(2):199–216 14. Lin J, Wang X, Zhai S, Shi M, Peng C, Deng X et al (2022) Hypoxia-induced exosomal circPDK1 promotes pancreatic cancer glycolysis via c-myc activation by modulating miR-628-3p/BPTF axis and degrading BIN1. J Hematol Oncol 15(1):128 15. Stella M, Falzone L, Caponnetto A, Gattuso G, Barbagallo C, Battaglia R et al (2021) Serum

59

extracellular vesicle-derived circHIPK3 and circSMARCA5 are two novel diagnostic biomarkers for glioblastoma multiforme. Pharmaceuticals (Basel) 14(7):618 16. Femmino S, Penna C, Margarita S, Comita S, Brizzi MF, Pagliaro P (2020) Extracellular vesicles and cardiovascular system: biomarkers and cardioprotective effectors. Vasc Pharmacol 135:106790 17. Tao W, Luo ZH, He YD, Wang BY, Xia TL, Deng WM et al (2023) Plasma extracellular vesicle circRNA signature and resistance to abiraterone in metastatic castration-resistant prostate cancer. Br J Cancer 128:1320 18. Dobin A, Davis CA, Schlesinger F, Drenkow J, Zaleski C, Jha S et al (2013) STAR: ultrafast universal RNA-seq aligner. Bioinformatics 29(1):15–21 19. Li H, Durbin R (2009) Fast and accurate short read alignment with burrows-wheeler transform. Bioinformatics 25(14):1754–1760 20. Bolger AM, Lohse M, Usadel B (2014) Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30(15): 2114–2120 21. Tarasov A, Vilella AJ, Cuppen E, Nijman IJ, Prins P (2015) Sambamba: fast processing of NGS alignment formats. Bioinformatics 31(12):2032–2034 22. Pertea M, Pertea GM, Antonescu CM, Chang TC, Mendell JT, Salzberg SL (2015) StringTie enables improved reconstruction of a transcriptome from RNA-seq reads. Nat Biotechnol 33(3):290–295 23. Zhao J, Li Q, Li Y, He X, Zheng Q, Huang S (2019) ASJA: a program for assembling splice junctions analysis. Comput Struct Biotechnol J 17:171143–171150 24. Gao Y, Zhang J, Zhao F (2018) Circular RNA identification based on multiple seed matching. Brief Bioinform 19(5):803–810 25. Quinlan AR, Hall IM (2010) BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26(6):841–842

Part III circRNA Detection, Sequence Validation and Quantification

Chapter 4 In Situ Hybridization of circRNAs in Cells and Tissues through BaseScope™ Strategy Eleonora D’Ambra, Erika Vitiello, Tiziana Santini, and Irene Bozzoni Abstract Imaging-based approaches are powerful strategies that nowadays have been largely used to gain insight into the function of different types of macromolecules. As for RNA, it is becoming clear how important is its intracellular localization for the control of proper cell differentiation and development and how its perturbation can be linked to several pathological states. This aspect is even more important if one thinks of highly polarized cells such as neurons. In this chapter, we describe in detail an innovative RNA-FISH approach for the detection of circular RNAs (circRNAs), a recently discovered class of noncoding RNAs, which display different subcellular localizations and whose functions still largely remain to be elucidated. The detection of these molecules represents a great challenge, above all because they share most of their sequence with the corresponding linear counterparts, from which they differ only for the back-splicing junction (BSJ) originating from the circularization reaction. This implies the use of RNA-FISH probes capable of specifically binding the BSJ and avoiding the detection of the linear counterpart. This requirement imposes the design of probes on a very small region, which implies the risk of obtaining a low and undetectable signal. The BaseScope™ Assay RNA-FISH technology overpasses this problem since it is based on branched-DNA probes. With this approach it is possible to target a specific region of the RNA, even small such as a splicing junction, and at the same time to obtain a strong and well detectable signal. All this is possible thanks to subsequent series of probes that, starting from the first hybridization to the BSJ, build a branched tree of probes that greatly amplifies the signal. Here we provide a detailed step-by-step protocol of BaseScope™ RNA-FISH on circRNAs coupled with immunofluorescence, both in cells and tissues, and we address difficulties which may arise when using this methodology that depend on cell type, specific permeabilization, image acquisition, and post-acquisition analyses. Key words CircRNA, lncRNA, In situ hybridization, RNA-FISH, Immunofluorescence, BaseScope™, Neurons, Confocal microscopy, Super-resolution microscopy

Eleonora D’Ambra and Erika Vitiello contributed equally. Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_4, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

63

64

1

Eleonora D’Ambra et al.

Introduction Understanding how RNA undergoes complex and dynamic biochemical processes is still to this day a field requiring further exploration, even though most of the mechanisms controlling its metabolism such as processing, transport, degradation, and translation have been thoroughly characterized [1–3]. Notably, the full understanding of how RNA localization affects its function, especially that of non-coding RNAs (ncRNAs), is largely undefined. One of the major aims of the molecular biology is to define spatial-temporal profiles of nucleic acids and proteins in different cellular conditions. Indeed, molecular distribution inside the cells can be highly influenced by cellular environment (i.e., stimuli, signals, stress, differentiation) and thusly their interactome and the mode of action [4]. Subcellular distribution and context-dependent dynamics are crucial features for RNA function, especially for many classes of ncRNAs [5]. Nuclear ncRNAs can influence RNA transcription and processing and can also contribute to 3D chromatin organization [6, 7], while in the cytoplasm, they can be involved in many processes of RNA metabolism such as mRNA turnover, local translation, and protein decoy [8]. Notably, most of the cellular mechanisms in which ncRNAs are involved rely on biomolecular condensates: membrane-less compartments whose function is to concentrate proteins and nucleic acids and whose molecular composition is rapidly influenced in response to different stimuli [9]. Indeed, ncRNAs are required to respond promptly to cellular stimuli in different manners [8] and although less abundant than mRNAs, they can exert their function by participating in the formation of these molecular condensates and acting with precise spatial-temporal dynamics [10–12]. Nonetheless, RNA localization is particularly important for highly polarized cells such as neurons, where ncRNAs have a decisive impact on transport, post-transcriptional regulation, and local translation [13]. For all these reasons, RNA visualization through RNA Fluorescence In Situ Hybridization (RNA-FISH) provides a fundamental tool for the investigation of RNA metabolism [4] and, over the years, the techniques for RNA imaging have multiplied and are constantly evolving, with the additional aim to detect all the different classes of coding and non-coding RNAs [14]. The most used and powerful RNA-FISH approaches are single-molecule RNA-FISH, based on fluorophore-labeled multiple probe set [15], LNA probes [16], and branched-DNA probes [17]. The multiple probes set approach was firstly developed by Singer and co-workers [15] and involves the use of many (~50) dye-labeled probes that pair to the target RNA sequence. Even though the multiple probes approach works perfectly for highly

In Situ Hybridization of circRNAs in Cells and Tissues

65

abundant long RNAs (mRNA, long-ncRNA) [18, 19], it is rarely employed when addressing different types of RNA species and splicing isoforms. For short and highly abundant RNAs such as miRNAs, the use of 20 nt long locked nucleic acid (LNA)-modified DNA probe [20], implemented with fluorescent dye or hapten, turned out to be of great benefit. Indeed, the incorporation of LNA-modified nucleotides inside the DNA probes significantly increases the specificity and the stability of the hybridization. Nevertheless, all these approaches are not suitable for all the ncRNAs species. Indeed, some ncRNAs derive from alternative splicing events and share most of their sequence with their mRNA counterpart, while also being low abundant. Moreover, ncRNAs usually rely on their structure to exert their functions, and particularly hinged secondary structures or participation to ribonucleoparticles (RNPs) might limit probes accessibility. One example of ncRNA class facing these criticisms is the class of circRNAs. CircRNAs are a novel class of ncRNA characterized by a covalently closed structure. Thanks to a noncanonical splicing mechanism called back-splicing, the downstream 5′ donor splice site is covalently bound with an upstream 3′ acceptor splice site, giving rise to a unique splicing junction, called back-splicing junction (BSJ). The BSJ represents the only unique region that allows the circular molecule to be distinguished from its linear counterpart arising from the same precursor mRNA [21, 22] (Fig. 1a). This feature makes the BSJ the only targetable region for RNA-FISH, which should avoid detection of the linear counterpart. Such limitation, associated with the low abundance of circRNAs, stresses the need for the use of a system which allows signal amplification, such as branched-DNA probe sets. Here we describe a new highly efficient method for circRNA detection, the BaseScope™ assay, which employs two target-specific probes directed toward the two flanking regions of the BSJ that are subsequently hybridized to a cascade of signal amplification molecules, resulting in enhanced signal detection (Fig. 1b, c). We provide step-by-step instructions for the BaseScope™ Assay applied to murine neuroblastoma cell line (N2a), embryonic stem cells (mESC)-derived motor neurons (MNs), and brain tissues. We show the circ-Hdgfrp3 RNA-FISH coupled with immunofluorescence, which achieves sub-compartment resolution and provides suggestions on how to overcome common obstacles in relation to cell-type specific fixation and permeabilization, as well as and to image acquisition and post-acquisition analyses.

66

Eleonora D’Ambra et al.

A 5’Cap

pA

Canonical splicing 5’ss

5’

3’ss

1

2

3

3’ss

4

3’

5’ss BSJ

Back-splicing reaction

B

C Amplification probes

Specific double Z-probes BSJ

pA

5’Cap

Target circRNA

Linear mRNA Fig. 1 Biogenesis and BaseScope™ probe design of a circRNA. (a) The pre-mRNA can undergo canonical splicing, that will generate linear mature mRNA (top) or lower efficiency back-splicing, which will generate a circular RNA molecule (bottom). In back-splicing, the downstream 5′ splice site donor attacks an upstream 3′ splice acceptor. ss splice site, BSJ back-splicing junction. (b, c) BaseScope™ FISH can be employed to detect splicing isoforms, such as circRNAs. It relies on series of primary, secondary, and tertiary oligonucleotide probes hybridized sequentially to enhance a fluorescent signal. In this case, unique Z probes are designed mapping on the sequences flanking the BSJ (panel c). This design guarantees specificity for the circular isoform, as secondary probes are unable to bind the primary probes unless they are correctly aligned (panel b)

2 2.1

Materials Probe Design

CircRNAs and their linear counterpart normally share most of their sequence; therefore, the probes for circRNAs detection must be specifically custom designed mapping on the BSJ (https://acdbio. com/BaseScope™-assays-0 3D > 3D Viewer command. 3.12 Structured Illumination Microscopy (SIM) (Fig. 5)

Structured illumination microscopy is a widefield-based microscopy technique that exploit interference patterns to reach an improvement in lateral resolution, close to ~120 nm. To perform SIM, no special precaution is needed in sample preparation, making it easily achievable, as long as the sample yields a good signal-tonoise ratio. Indeed, SIM has already been used in literature to ameliorate BaseScope™ visualization in tissues and cells, coming in particularly useful when studying the sub-structural distribution of an RNA in specific subcellular compartments [25, 26]. 1. To perform SIM, place a drop of proper immersion oil on the sample, and mount it facing down on a N-SIM Super-Resolution Microscope, equipped with a 1.49 NA 100x objective (Apo TIRF 100x Oil, Nikon, Tokyo, Japan) and with a 3D EX V-R 100x/1.49 Grating Block (see Note 37). 2. Depending on the excitation wavelength of the secondary antibodies, sequential 488 or 640 nm laser units can be used

In Situ Hybridization of circRNAs in Cells and Tissues

85

Fig. 5 SIM of BaseScope™ FISH coupled with IF and DAPI staining. SK-N-BE cells expressing human mutant FUS (P525L) stained for circ-Hdgfrp3 (red), FUS (green), and the stress granules marker TIAR (gray) are acquired with structured illumination microscopy (SIM). Nuclei were stained with DAPI (cyan). Panels a and b provide a magnified view of all the channels, highlighting the sub-structural interaction between circHdgfrp3, mutant FUS, and stress granules (TIAR). Scale bars represent 10 μm

to image IF staining, while a 405 nm and a 561 laser unit can be respectively used to excite DAPI and BaseScope™ Fast RED fluorescent signals. 3. If collecting images with the NIS-Elements AR software (Nikon), the ND acquisition module can be used for large image acquisition and stitching and for 3D-SIM image collection. 4. Once the sample is imaged, the SIM reconstruction parameters “illumination modulation contrast,” “high-resolution noise suppression,” and “out of focus blur suppression” can be adopted and set accordingly to generate a consistent Optical Transfer Function (OTF) and obtain a super-resolved acquisition (see Notes 38 and 39).

86

4

Eleonora D’Ambra et al.

Notes 1. Warm RNAscope™ 50× Wash Buffer up to 40 °C for 10–20 min before preparation. 1× Wash Buffer may be prepared ahead of time and stored at room temperature for up to one month. 2. IBIDI chambered coverglasses (Ibidi, 80,841 or equivalents, i.e., Nunc™ Lab-Tek™ II Chamber Slide™ System, Thermo Scientific™ 154534PK) or multi-well plates can also be used. However, IBIDI chambers are more expensive and seeding cells in single wells might result in a nonhomogeneous sample, as cells show a tendency to crowd the edges of the chamber. 3. Sterilize coverslip with 100% Ethanol in a 50 mL tube. 4. Alternatively, 30 min at room temperature under UV light can be sufficient for fast collagen polymerization. 5. Bivalent cations, such as magnesium (Mg2+) and calcium (Ca2+), can work as ligands for proteins on cell membranes, promoting cell-to-cell and cell-to-substrate adhesion [27–29]. 6. Performing fixation steps at high temperature can overaccelerate the cross-linking process, along with increasing risk of inhaling formaldehyde fumes, which are dangerous for human health [30–32]. Moreover, reduced temperature was shown to have a protective effect on nucleic acids preservation throughout the fixation process and the following FISH protocol steps [33]. Similarly, using ethanol-based solutions at high temperatures can induce over-permeabilization and damaging of the sample [34]. 7. Poly-L-Ornithine (PLO) coated dishes can be stored at -20 °C up to a month, prior PLO removal. Alternatively, PLO and Laminin coating steps can be performed in a range of 2–16 h, incubating at least 2 h at 37 °C each. For shorter incubation times of PLO, we suggest to carry them out at 37 °C to obtain a better efficiency of the coating although cells adhesions could not be optimal. 8. Increasing concentrations of sucrose serve as cryo-protectant agent, preventing ice crystal formation and freeze damage for optimal fixed-sample freezing [35]. Moreover, sucrose was proven to decrease the swelling of cell, a phenomenon usually occurring in specimens placed in fixative or saline solutions [36]. Importantly, in order to prevent RNA degradation, RNase inhibitors can be added to the sucrose solution (e.g., VRC 2 mM). However, cryo-protection step is not recommended for non-fixed tissues due to possible RNA degradation and poor morphology retention.

In Situ Hybridization of circRNAs in Cells and Tissues

87

9. Fixation timing can be established depending on the tissue features: for example, for big organs such as mice adult brain, we suggest 24–48 h. Indeed, a short fixation time can affect tissue integrity and proper molecules fixation, while excessive long-lasting fixation time can compromise tissue permeabilization. 10. Iso-pentane (liquid-nitrogen) will cool down the temperature at about -150 °C to -80 °C, allowing the tissue to freeze fast enough that ice crystals do not form. In fact, ice crystals might damage the tissue and can produce morphology artifacts. 11. For cryo-embedded tissues, OCT compound can be used. It is a water-soluble blend of glycols and resins that protects the sample from mechanical cracking during cutting and prevents desiccation during long-term storage. Cryomolds are commercial disposable plastic cases. However, they can be efficiently replaced with home-made coatings prepared with an aluminum foil. 12. If performing the technique for the first time, take into account supplementary coverslips to use for positive and negative controls. For example, if studying a particular circRNA, performing the FISH experiment on a housekeeping transcript (such as GAPDH mRNA) and on a transcript expressed in an unrelated species (such as yeast or bacterial transcripts) can serve respectively as positive and negative controls. These controls will be crucial to check for target specificity, hence to determine possible off-target hybridization, to set a background corresponding to a specific signal and to determine signal quality. 13. If IBIDI chambered coverglasses are used, culture and fix the cells exactly as described in Subheading 3.1, adjusting the coating and the cell media volumes according to the area of the chambers. 14. When working with fixed neuronal cells, it is mandatory to be very careful during incubation and wash steps, in order to avoid neurites detaching and disintegration. Therefore, to aspire and replace solutions, it is recommended to use a p1000 pipette, with a gentle jet, never directed right on the coverslips nor on the cells, but streamed on the wall of the Petri dish. 15. In order to prevent RNA degradation, make sure to use sterile equipment. For example, filtered tips are suggested to avoid contaminations from the pipettes, and all the glassware, including object holder slides, should be carefully washed with ethanol and RNase decontamination solutions. 16. Keep on using this procedure for all the subsequent incubation and wash steps.

88

Eleonora D’Ambra et al.

17. The protease treatment provides an increase of probes accessibility in the cells, with subsequent enhancement of hybridization efficiency; however, it can impair the immune-detection of protein targets. For this reason, it is strongly recommended to perform immunofluorescence staining after protease incubation, and, in general, after FISH probes hybridization, scaling down the amount of enzyme until an optimal dilution factor for the protease mix is set. Yet, we found that this dilution works perfectly on N2a cells and on murine motor neurons, even when combining BaseScope™ FISH with immunofluorescence assay. 18. It is also possible to use adhesive frames (Gene Frame), which offer the advantage of not detaching during hybridization incubations and washings steps. 19. The tissue or cell culture samples should always be completely covered by buffer solutions, to avoid the sample from drying out. 20. Be careful not to let the slides dry out between steps. In fact, tissue desiccation can lead to an increase in autofluorescence and to morphology damage. 21. The buffer must reach the boiling point and maintain uniform boiling al 99–100 °C using alternatively: dry oven, hot plate, or microwave oven. Avoid boiling the buffer beyond 30 min before the use. 22. A prolonged processing at high temperature can impair tissue morphology and antigen’s structure. For this reason, the incubation time is a parameter that must be calibrated empirically, based on tissue type. 23. RNAscope™ Universal pretreatment kit contains different proteases (Protease Plus, Protease III and Protease IV) to be used based on the tissue type, fixation method, and slice thickness. As suggested in Note 17, empirically select the optimal reagent and incubation parameters in order to obtain the maximum staining efficiency. The general rule to keep in mind is that the Protease Plus provides a milder tissue permeability with respect to the Protease IV, while Protease III shows an intermediate digestion efficiency compared to the previous two. 24. Try not to use excessively wet paper; otherwise, as probes hybridization is performed at 40 °C, the condensed vapor could dilute the probes concentration. On the other hand, excessively dry paper could cause evaporation of probes mix throughout the incubation steps.

In Situ Hybridization of circRNAs in Cells and Tissues

89

25. At this point, it is possible to interrupt the protocol and store the samples in 5× SSC solution overnight. The following morning, proceed to Subheading 3.5. 26. At this point, turn off the HybEZ Oven, as all the next steps will be carried on at room temperature. 27. For a 0.75″ × 0.75″ barrier, use 120 μL of Red A into a tube; for 12 mm coverslips, use 80 μL. Moreover, BaseScope™ Fast RED-B dilution can be empirically adjusted. 28. As protease pretreatment was already performed, classical permeabilization step is not necessary, and both goat or donkey serum can be used for blocking nonspecific antibody binding. The concentration of the blocking agent depends from several aspects, among which: biological sample type (cell culture or tissue) and antigenic affinity of the primary antibody. For these reasons, parameters calibration is recommended. Generally, 1–2% of serum/PBS is enough for fixed cell cultures, while, for tissues, a stronger blocking is required. For example, 5% serum/1% BSA (Bovine Serum Albumin)/0.05% Triton X-100/PBS for 30 min at room temperature can be useful to obtain a good compromise between optimal staining quality and low nonspecific antibody-binding events. Moreover, mild tensioactive properties of diluted Triton X-100 detergent do not alter Fast Red precipitates. 29. Prepare the IF blocking buffer fresh every time and sterilize by filtering before use. Indeed, BSA or normal serum powders easily clump and form precipitates if not perfectly dissolved, thus compromising IF staining producing a noisy background. Moreover, from now on, it is strongly recommended to add VRC (2 mM) to all the IF blocking buffers. In fact, VRC is able to bind active sites of many RNases, inhibiting their activity. 30. Primary antibodies incubation can be shortened up to 1 h at room temperature, depending on the protein to stain. Nonetheless, longer incubations at low temperature guarantee a more efficient visualization of the target [37]. 31. When selecting the secondary antibodies, take into account that the Fast-Red substrate of alkaline phosphatase provides a fluorescent signal with excitation and emission spectra near to Cy3 dye. Therefore, choose secondary fluorescent antibodies accordingly, in order to prevent overlapping of spectra when acquiring the samples. Moreover, make sure that the microscope is equipped with the proper set of filters, to avoid crosstalk. 32. The excess of liquid in this step could create bubble in the mounting media.

90

Eleonora D’Ambra et al.

33. Mounting media should contain antifade compound, to protect the fluorophore from photo-oxidation, and a high refractive index (RI), to perform high-quality and highmagnification imaging. 34. A laser scanning or a spinning disk confocal microscope can be used. Fast-Red fluorescence of BaseScope™ staining shows high brightness and photostability; therefore, it can be acquired with very intense light source (e.g., laser) and for prolonged exposition without excessive decrease of excitation/emission efficiency. 35. Inverted high-numerical aperture objectives (NA = 1.4) should be used to obtain images with a high subcellular lateral resolution (≥0.2 μm) in accordance with Abbe’s diffraction formula (Resolutionx,y = λ/2NA; [38]). 36. The spatial resolution of an image depends on the optical characteristics of the microscope (camera and objectives) and from the ratio between the physical scanned area and the pixel number. In order to obtain a highly informative image, it is strongly recommended to set zoom, Z-step, and pixel area as indicated by Nyquist rate (https://svi.nl/NyquistCalculator). 37. Whichever widefield microscope equipped with a SIM module and a grating block can be used. In general, when mounting coverslips on an object holder for SIM visualization, it is recommended to place one single coverslip at the very center of the object holder. This precaution helps to minimize light scattering, therefore preventing artifacts during reconstruction of the image. For the same reason, it is particularly important that the coverslips thickness matches the refraction index of the immersion oil, in order to avoid light distortion, as well as using a compatible mounting media (i.e., Prolong Diamond Antifade mounting media). 38. When processing a SIM image, the reconstruction algorithm of NIS-Elements AR software usually generates a “reconstruction score,” indicating the quality of the reconstructed image. This score can have a value from 1 to 8, where only images achieving a score of 8 can be considered as high-quality images. A score = 99.8% Ethanol BioUltra, for molecular biology (Merck Life Science, 51976). 7. RiboLock RNase Inhibitor (Thermo Fisher, # EO0381). 8. 10 mM dNTP Mix (Euroclone, # EMR416001). 9. 100 μM Random Hexamers (Euroclone, # EMR428200). 10. Nuclease-free water (DNase and RNase -free water). 11. 0.2 mL PCR tubes, RNase-free. 12. Filter tips. 13. Magnetic separation rack for 0.2 mL tubes.

112

Linda Masante et al.

14. PCR thermocycler. 15. Shaker vortex. 16. Centrifuge Strip Spin. 2.3

ddPCR Reagents

1. cDNA already prepared. 2. Forward (fw) and reverse (rv) primers (100 μM stock). 3. EvaGreen SuperMix (Bio-Rad, #1864033). 4. Nuclease-free water (DNase and RNase -free water). 5. QX200™ ddPCR Droplet Generation Oil for EvaGreen (Bio-Rad, #1864005). 6. Automated Droplet Generation Oil for EvaGreen (Bio-Rad, #1864112). 7. ddPCR™ Droplet Reader Oil (Bio-Rad, #1863004).

2.4 ddPCR Equipment

1. 0.2 mL PCR tubes, RNase-free. 2. DG8™ Cartridges and Gaskets (Bio-Rad, #1864007). 3. Droplets generator cartridge holders (Bio-Rad, #1863051). 4. DG32 Automated Droplet Generator Cartridges (Bio-Rad, #1864108). 5. Pipet Tips for the AutoDG™ System (Bio-Rad, #1864120). 6. Pipet Tip Waste Bins for the AutoDG™ System (Bio-Rad, #1864125). 7. ddPCR™ 96-Well Plates (Bio-Rad, #12001925). 8. 96-Well Plates Heat Foil (Bio-Rad, #1814040). 9. Multi-channels pipettes. 10. Filter tips (Rainin or Eppendorf). 11. Strips vortex. 12. Strips spinner.

2.5 ddPCR Instruments

1. QX200 Droplet Generator Bio-Rad (Bio-Rad, #1864002). 2. Automated Droplet Generator (Bio-Rad, #1864101). 3. Shaker vortex. 4. Centrifuge Strip Spin. 5. PX1 PCR Plate Sealer (Bio-Rad). 6. C-1000 Thermocycler (Bio-Rad). 7. QX200 Droplet Reader Bio-Rad (Bio-Rad, #1864003).

2.6 ddPCR Data Analysis

1. QX Software (Bio-Rad, available from: https://www.bio-rad. com/it-it/life-science/digital-pcr/qx-software) or QuantaSoft™ Analysis Pro (Bio-Rad, available from: https://www.

ddPCR for Absolute Quantification of circRNAs

113

bio-rad.com/it-it/life-science/digital-pcr/qx200-droplet-dig ital-pcr-system/quantasoft-software-regulatory-edition?ID=1 864011).

3

Methods

3.1 Divergent Primers Design

1. From RNA-seq data, or from publicly available databases (such as circBase http://www.circbase.org/), determine the chromosome location and the start and end coordinates of the BSJ of the circRNA of your interest. 2. Determine which exons/introns are connected through the BSJ using online software and exploiting the coordinates retrieved in step 1. Depending on the species you are using as animal model, you can insert the whole back-splice site sequence in the “blat” tool under tools menu at https:// genome.ucsc.edu/cgi-bin/hgBlat?hgsid=1510163519_ cBLFBKujK6emrOTmqbvo7ou9vDfk&command=start. 3. Retrieve the sequence of the exons/introns composing the circRNA of your interest from publicly available websites or RNA-seq data, depending on the species. 4. Invert the order of the circularized exon(s) (paste the sequence of the latter exon in front of the one of the prior) and thus consider the inverted sequence composing the circle as a linear RNA (Fig. 3) (see Note 1). 5. Paste the resulting target sequence into https://www.ncbi.nlm. nih.gov/tools/primer-blast/ (or any other available online tool). 6. Since the target sequence is the BSJ, the limit of the exons needs to be specified for forward and reverse annealing, during the process of the design. Define the exons limit at https:// www.ncbi.nlm.nih.gov/tools/primer-blast/ (or any other available online tool); for example, see Fig. 4. 7. Similarly to the rules of qPCR primer design, set the amplicon length to 60–200 bp at https://www.ncbi.nlm.nih.gov/tools/ primer-blast/ (or any other available online tool). 8. Design primers that have a GC content of 50–60%. 9. Design primers with a Tm between 50 °C and 65 °C. 10. Avoid repeats of Gs and Cs longer than three bases. 11. Place Gs and Cs at the 3′ nucleotide of primers, when possible. This so-called GC clamp facilitates the correct binding thanks to the stronger hydrogen bonding of G and C bases. However, since BSJ is a very short sequence to target, it is not always possible to design primers with the GC clamp at 3′ end due to sequence constraints.

114

Linda Masante et al.

pre-mRNA

2

1

3

4

5

6

circRNA specific primers

4

2 3

linRNA specific primers exon 5-6

Fig. 3 Schematic of linear and circular primer pairs. Left) schematic of a pre-mRNA strand, with both linRNA-specific (red arrows) and circRNA-specific (green arrows) primers. Right) schematic of circRNA. Divergent primers are designed on the BSJ

2

3

2

3

linear RNA sequence

3

2

circRNA sequence

fw

rv

TGTCTCTGCAAGATATTAACATGCGGAAAGCGTTTAAGAGTTCAACCACGCAGGACCAGAAGGTCG TCTCGCAGAACAGCACCCCAAAGCCGGTCATTGATATATATAAGACGTGCGACAGACCTCCTCCTC TCCATATACTTTCACGATACAGGCAAGCACGCAGAAGACATATTTGGCGAGCTGTTTAATGA GGCGAACAATTTCTACATCAGAACAAATTCTCTTCAAGATCGAATTGATCGGCTCGCTGT TAAAGTCACCCAGCTGGATTCAACAGTGGAAGAGG

Fig. 4 Schematic of BSJ primer design. Schematic of circRNA-specific primer design. The BSJ, formed by the covalent joining of the 3′ of exon 3 with the 5′ of exon 2, is reported in red. The forward primer is designed on the blue sequence, corresponding to exon 3, while the reverse one is designed to target the green sequence, corresponding to exon 2. The primer pair designed in this way will amplify exactly the BSJ sequence. Abbreviations: fw forward, rv reverse

12. Since extensive primer-dimer formation can significantly decrease or prevent amplification, check forward and reverse primer sequences to ensure no 3′ complementarity (avoid primer-dimers) and self-complementarity. Exploit publicly online tools (for instance OligoCalc http://biotools.nubic. northwestern.edu/OligoCalc.html to check the sequences). 3.2 RNase R Treatment and cDNA Synthesis

1. All the following steps have to be prepared in RNase-free conditions. Thus, use RNaseZap (or equivalent) to clean the workbench and pipettes. Thaw all the reagents—except for

ddPCR for Absolute Quantification of circRNAs

115

enzymes—on ice. Once thawed, vortex and spin well all the tubes. Note that all the enzymes must be kept at -20 °C, quickly spinned down and added to the proper reaction at the end of its assembly, just prior to the incubation step(s). SPRI beads have to be kept at room temperature, protected from light (see Note 2). 2. Assemble a 20 μL reaction containing: total RNA, 10× RNase R buffer, RNase R (20 U/μL), NF-water. Digest X μL of RNA with 0.01 U RNase R enzyme/ng RNA diluted in NF-water (see Notes 3–5). 3. Vortex and spin down the reactions. 4. Incubate the reactions for 30 min at 37 °C (see Note 6). 5. During the incubation time, vortex well the SPRI beads used to purify the RNA after the treatment. 6. Add 2.2× volume of SPRI beads to each reaction (here, 44 μL) and resuspend by pipetting up and down to homogenize the solution (see Note 7). 7. Leave the tubes at RT for 10 min to allow the binding of the beads to the RNA. 8. Move the tubes to the 0.2 mL tubes magnet and leave them for 7 min. 9. Wash twice in EtOH 85% (see Note 8). 10. Carefully remove all the residual ethanol using a p10 pipette. 11. Dry the beads by opening the lid of each tube (see Note 9). 12. Add 11 μL of nuclease-free water and resuspend by pipetting up and down to homogenize the solution. Leave the tubes at RT for 3 min. 13. Move the tubes to the magnet and leave it for 3 min in order to separate RNA from beads. 14. Collect 10 μL of purified RNA (see Note 10). 15. Retrotranscribe the purified RNA with Maxima H minus Reverse Transcriptase (RT) using random hexamers primers, following manufacturer instructions (see Notes 11 and 12). Assemble a 20 μL reaction containing: 10 μL of purified RNA, 50 μM random hexamers, 10 mM dNTPs dissolved in NF-water. Incubate the reactions at 65 °C for 5 min in order to disrupt highly complexed RNA structures. Add 4 μL of 5× RT buffer, 0.5 μL of 40 U/μL RiboLock RNase Inhibitor, and 1 μL of 200 U/μL Maxima RT enzyme to reach the final volume. 16. Vortex and spin down the reactions. 17. Incubate the reactions at 25 °C for 10 min, 50 °C for 30 min, 85 °C for 5 min.

116

Linda Masante et al.

3.3 cDNA and Primer Mix Preparation

1. Dilute cDNA into NF-water to reach a final volume of 5 μL, based on the ng chosen as input for each ddPCR reaction (see Note 13). Vortex and briefly spin down. 2. Prepare the primer mix solution ready-to-use: NF-water with a mix of fw and rv primers on ice. Starting from 100 μM primers aliquots: 38 μL of NF-water +1 μL of fw primer +1 μL of rv primer. Vortex and briefly spin down.

3.4 ddPCR Mix Assembly Using EvaGreen-Based Assay

1. Assemble the ddPCR mix reaction (prepare a master mix, taking into consideration a final volume per reaction of 23 μL). For each single mix, add: 5.12 μL of NF-water, 1.38 μL of primer mix (to reach a final concentration of 150 nM, suggested by Bio-Rad), and 11.5 μL of 2× EvaGreen SuperMix (see Note 14). Vortex and spin down (see Note 15). 2. Dispense 18 μL of mix in the tubes. 3. Add 5 μL of cDNA (see Note 13). 4. Mix thoroughly by vortexing (possibly, use a strips vortex: 15 s at 2000× g) and quickly spin down (possibly use a strips spinner) the tubes. 5. Allow the reactions to equilibrate at room temperature for about 3 min. During this incubation time, remove the block from the PX1 PCR Plate Sealer, close it, and set the protocol: 5 s at 180 °C (see Note 16). 6. Proceed to Subheadings 3.5.1 or 3.5.2 according to the modality chosen to generate droplets.

3.5 ddPCR Droplet Generation and Transfer to PCR Plate

1. Load 20 μL of PCR reaction in the row of the cartridge marked as “Sample” and load 70 μL of oil in the row marked as “Oil” (Fig. 5) (see Note 17).

3.5.1 Manual Droplet Generation and Transfer

2. Close the cartridge, making sure that all the four “teeth” on the holder are inserted in the corresponding four holes of the gasket (see Note 18). Insert the cartridge into the Droplet Generator and start the droplets generation. This should take approximately 2 min (see Note 19). 3. When droplets generation is complete, remove the gasket. The upper part of the cartridge marked as “Droplets” contains the generated droplets in which the sample has been partitioned. 4. Aspirate droplets from the cartridge: slowly pipette 40 μL of generated droplets. This should take approximately 5 s. Note that a few μL of air is expected inside the tips (see Note 20). 5. Dispense droplets into a single column of a 96-well PCR plate: position the pipette tips along the side of the well and slowly dispense the droplets. Avoid touching the bottom of the wells.

ddPCR for Absolute Quantification of circRNAs

117

droplets sample (20uL) oil (70uL)

Fig. 5 ddPCR cartridge schematic. Schematic of ddPCR cartridge. “Droplets” indicates the wells containing the generated droplets in which the sample was partitioned; “sample” indicates the wells where 20 uL of cDNA has to be loaded; “oil” indicates the wells where 70 uL of oil has to be loaded

6. Cover the 96-well plate with a foil plate compatible with both PX1 PCR Plate Sealer and the needles in the QX200 droplet reader (see Note 21). 7. Set the protocol on the PX1 PCR Plate Sealer: 180 °C for 5 s. 8. Open the PX1 PCR Plate Sealer by pressing the proper arrow, lace the support block (brought previously at RT) in the tray and the 96-well PCR plate loaded with the generated droplets on top of it. 9. Close the PX1 PCR Plate Sealer by pressing the proper arrow. The sealing will automatically start. 10. Once the sealing is complete, the PX1 PCR Plate Sealer will open. The sealed plate is ready for the thermal cycling. 3.5.2 Automated Droplet Generation and Transfer

1. When using the Automated Droplet Generator (AutoDG), ensure to place the cooling block upside down at -20 °C for at least 2 h before the run (see Note 22). 2. Load 22 μL of the ddPCR mix reactions into a 96-well PCR plate. 3. Set the protocol on the PX1 PCR Plate Sealer: 180 °C for 5 s. 4. Place the support block in the tray and the 96-well PCR plate loaded with the generated droplets on top of it (see Note 21). 5. Once the sealing is complete, the PX1 PCR Plate Sealer will open. Remove the 96-well PCR plate from the block. Centrifuge for 30 s at 1000× g and spin down. Remove the foil. 6. Configure the AutoDG instrument and load all the required consumables (cartridges, gaskets, filter tips) into the Automated Droplet Generator (see Note 23). 7. Load the 96-well PCR plate prepared in step 2 of this section and the Droplet Plate into the proper location inside the instrument. 8. Check that the oil used to generate droplets is the one needed to the chemistry selected (EvaGreen or probes). 9. Once all the consumables as well as the 96-well PCR plate loaded with the samples are correctly loaded into the AutoDG

118

Linda Masante et al.

instrument, start the droplets generation. The time required will appear on the screen. 10. Remove the plate from the AutoDG instrument and seal it following the procedure previously described in step 3 of this Section. 11. Once the sealing is complete, the PX1 PCR Plate Sealer will open. Remove the 96-well PCR plate from the block. The sealed plate is ready for the thermal cycling. 3.6 PCR for EvaGreen-Based ddPCR Assays

1. Insert the 96-well PCR plate in the thermal cycler and set the following protocol (see Note 24) (Table 1). Leave the plate at 4 °C for 1 h, to allow droplets re-hydration (see Note 25).

3.7

1. Remove the 96-well PCR plate from the thermocycler and leave it at room temperature for 5 min to re-equilibrate the droplets temperature.

Droplets Reading

2. Insert into the QuantaSoft™ Analysis Pro Software all the information related to the ddPCR experiment: type of assay, chemistry used, target gene and primers ID (see Note 26). 3. Start the plate reading with the QX200 Droplet Reader Bio-Rad (see Note 27). 3.8 ddPCR Data Analysis

1. In the raw folder of files related to the run, open the file with a . qlp extension. 2. Check that (i) the number of droplets generated is sufficient to proceed with the analysis (i.e., 10,000 droplets, see Notes 28 and 29) and (ii) you obtained a good separation among the positive cloud of droplets (containing the target) and the negative one (without the target) (see Notes 30 and 31). 3. If both the requirements listed in step 2 are met, proceed with the analysis. For each different ddPCR experiment, a new threshold has to be manually drawn for each primer pair used. Two options are available in order to set the threshold: “Single Well” and “Multiple Wells” thresholding. “Multiple Wells” threshold is always suggested. Ideally, set the threshold exactly between the positive and negative clouds (see Note 32). 4. As anticipated in Note 15, it is generally a good practice to run the same reaction in duplicate to minimize the effect of variability that can lead to erroneous quantifications. The automatically computed merged values of the absolute copy number in 20 μL of reaction have to be considered for subsequent analysis. In the left Menu “Main Options” > “Graph Tools,” it is possible to select the options “Merged Wells” or “Individual Wells.” The first choice will show the average of the two wells with the same name both in the graph and in the data table

ddPCR for Absolute Quantification of circRNAs

119

Table 1 EvaGreen PCR protocol. Example of PCR protocol for Evagreen. The asterisk refers to parameters that have to be defined experimentally Cycling step

Temp, °C

Time

Ramp rate

# of cycles

Enzyme activation

95

5 min

2 °C/s

1

Denaturation

95

30 s

40*

Annealing/extension

58*

1 min

40*

Signal stabilization

4

5 min

1

90

5 min

1

4

Infinite

1

Hold (optional)

(Window “Well Data,” bottom-right of the screen). The second one will show individual data for each well read (see Note 33). 5. Export the table containing all the details of the run. The Excel file exported will contain all the information related to the run. The most relevant ones, aiming to quantify the circRNA expression, are the columns “Conc (copies/μL) reporting the circRNA copy number per μL and “Copies/20μLWell” reporting the copy number in the PCR reaction, thus in 20 μL (see Notes 34 and 35).

4

Notes 1. In contrast to the commonly used convergent primer pairs designed in order to amplify a linear transcript (red arrows, in Fig. 3), the primer pair required to detect the BSJ of circRNAs has to be divergent (green arrows, Fig. 3) and become inward facing producing discrete amplicons when a BSJ connects distant exons. 2. To avoid beads saturation, it is preferrable to use common phenol-chloroform procedure to purify the reaction for high input RNA (> 10 μg). In our hands—using SPRI beads—the yield of purified RNA obtained starting with 10 and 20 μg of input RNA was the same. 3. The optimal concentration of RNase R to use has to be experimentally optimized, considering that the general aim is to digest the majority of linear RNA without affecting circRNA expression. Normally, a titration of the RNase R is performed. Consider as an optimal enzyme concentration the one that leads to a >90% of degradation of the linear RNA without

120

Linda Masante et al.

affecting (or affecting minimally, i.e., < 5%) the circRNA isoform. 4. A reaction assembled without the RNase R enzyme has to be used as control (mock). Replace it with the same amount of nuclease-free water. 5. The range of input RNA tested successfully in our hands is 10–750 ng. 6. RNase R incubation time can vary from 10 min to 1 h [8, 24– 27]. This parameter has to be optimized experimentally, considering the general aim is to digest the majority of linear RNA without affecting circRNA expression. 7. When the beads (such as SPRI or RNAClean XP beads) are used in common reactions clean-up, a beads-to-sample ratio of 1.8X, according to manufacturer instructions (Beckman Coulter), is generally used in order to avoid the exclusion of small (or large) RNAs. 8. Fresh and ready-to-use EtOH has to be prepared each time to guarantee the integrity of RNA bound to the beads. The concentration of EtOH depends on the type of beads used for the purification. SPRI beads work efficiently with EtOH 85%. 9. Leaving the lid opened allows the EtOH left on the beads to evaporate quickly. Overdrying the beads must be avoided since it would reduce the yield and increase the RNA degradation. As soon as the color of the beads turns darker and the first small cracks appear, they are considered dry and the nuclease-free water needs to be added. 10. Be very careful! Avoid any trace of beads in the solution of RNA, by pipetting slowly the purified RNA. 11. Since circRNAs lack poly-A tails, reverse transcription has to be performed using random hexamers instead of oligo (dT) primers. 12. Always run two control reactions in parallel to the actual samples: a “no template” control, in which the components of the reaction are present except for the RNA, and a “no enzyme” control, in which the RT enzyme is the missing reagent. The first control allows the detection of any contamination (crosscontamination among samples, presence of traces of leftover cDNA from previous reactions) in the reagents, while the second investigates the presence of genomic DNA (gDNA) contamination in the sample, which is revealed in the PCR reaction. 13. The cDNA can be diluted, depending on the ng of cDNA chosen as input. The range of input cDNA we tested successfully was 1.2–1.5 ng, assuming a 100% RT efficiency.

ddPCR for Absolute Quantification of circRNAs

121

14. ddPCR experiment can support both hydrolysis probes (e.g., TaqMan probes) and DNA binding dye (e.g., EvaGreen) -based assays. EvaGreen is a double-stranded DNA (dsDNA) binding dye and it is the only dye chemistry optimized for ddPCR. It allows the detection of dsDNA after PCR amplification with accurate and sensitive quantification, and is well suited for low abundance targets due to its low detection limit [22]. It also simplifies the experiment, favoring highthroughput assays, and reduces running costs compared to probes. The only limitation is that it requires more time to optimize parameters when detecting multiple targets. Alternatively, TaqMan probes, containing a 5′ reporter dye and a 3′ quencher, can be designed specifically for a unique target. 15. While performing highly sensitive experiments involving many consecutive steps such as ddPCR, it is always suggested to run replicates of the same reaction, to have more reliable quantifications avoiding false, inaccurate readings and reducing the impact of variability. In ddPCR experiments, it is generally suggested to run the same PCR reactions in duplicates. 16. Remove the block from the PX1 PCR Plate Sealer instrument before setting the desired protocol to seal the plate. An excessively hot block could impair the generated droplets. 17. Pay attention to avoid bubbles formation. This can heavily impair droplet generation. 18. If even one “tooth” is not properly inserted in the gasket, pressure needed for droplet generation will not be achieved resulting in droplets number impairment. 19. If the instrument is running regularly, the droplet indicator light (on the right) should start flashing green. If the instrument encounters issues, an amber light will flash. 20. Position the pipet tips in each of the 8 wells at a 45° angle. Do not position it in a vertical orientation (90°) and avoid the tips to be flat against the bottom of the wells. Erroneous pipetting would lead to droplets breaking. 21. Ensure that the foil covers all the wells. If the foil does not cover the wells, the plate will not be correctly sealed impairing the reading step (at the end of the ddPCR process). In that case, the needle that “reads the fluorescent signal” cannot make a hole in the sealing leading to errors in the procedure (the reader stops the reading). 22. The cooling block is critical to prevent droplet evaporation. 23. If any of the consumables are positioned incorrectly, the lights indicator on the touch screen will remain yellow, instead of turning to green. In case of incorrect positioning, place again the consumable until the light appears green.

122

Linda Masante et al.

24. The setting of the PCR protocol depends on the selected chemistry for the ddPCR assay, EvaGreen and TaqManbased, due to the different composition of the SuperMix used. 25. It is possible to leave the plate at 4 °C overnight to increase the droplet rehydration. This can improve the droplet signal reading, leading to a higher number of accepted droplets. 26. The duplicated wells have to be renamed in the same way, to allow the Software to automatically compute the average copy numbers of the duplicates. 27. The instrument takes around 12 min to read one entire column of the plate, corresponding to 8 wells. 28. If less than 10,000 droplets are generated, the well has to be excluded from the analysis. The ideal number of droplets generated is 20,000, so obtaining half or less of droplets could indicate an erroneous partition of the sample in the few generated droplets, leading to possible mis-quantification of copy numbers. Indeed, from a statistical point of view, a minimum number of droplets is required to achieve a more robust and accurate quantification, as the higher the number of partitions, the higher the convergence to the Poisson equation [28]. 29. The following information refers to the analysis run using the QuantaSoft™ Analysis Pro Software. Recently, QX Software— another Software released by Bio-Rad—is available for free. Here, a quick step-by-step explanation to analyze a common ddPCR experiment output is following. A detailed UserGuide on the QuantaSoft™ Analysis Pro Software is available online: https://www.bio-rad.com/webroot/web/pdf/lsr/litera ture/QuantaSoft-Analysis-Pro-v1.0-Manual.pdf. 30. When running a ddPCR experiment, always keep in mind that the first aim is to obtain a good separation among the positive cloud of droplets and the negative one. This is crucial to have a reliable quantification of the targeted molecule. When the positive and the negative clouds are not well separated, a “rain” of droplets between the two clusters is visible. Several parameters including suboptimal PCR annealing temperature, the amount of input cDNA, number of cycles, template accessibility and/or degradation, and PCR inhibitors can strongly influence such separation and thus require a proper experimental optimization (Fig. 6a). Of note, when working with lowly expressed molecules, changing the number of PCR cycles (from 40 to 50) could positively impact the separation among positive and negative droplets. 31. In our hands, the most critical parameters to optimize to improve cloud separation are cDNA concentration and annealing temperature. If a clear separation of partitions is not observed (Fig. 5), perform a cDNA titration to select the best

ddPCR for Absolute Quantification of circRNAs

123

Fig. 6 Optimization of droplet clouds separation and threshold setting. (a) Example of the effect of parameter optimization on droplet clouds separation. The threshold is defined based on the separation between negative and positive droplets, using the negative control as a reference. (b) The threshold is set based on the highest amplitude droplet in the negative control. Examples of non-acceptable (red box, showing a cluster of droplets above the lower cloud) and acceptable (blue circle, showing a single droplet above the lower cloud) negative controls to use in threshold setting. The main parameter optimize was the temperature annealing in PCR reaction. Abbreviations: NC negative control

cDNA dilution to use for your ddPCR experiment. Furthermore, a general rule is to run a thermal gradient to test a broad range of annealing temperature, for each primer pair. 32. Multi-wells thresholding tool allows to set a unique threshold for multiple wells when cDNA has been amplified with the same primer pair. It is generally used and set according to the level of background fluorescent signal in the negative control (no template control). The threshold is drawn above the negative cluster. In the case where a few droplets with higher amplitude are present in the negative control, the threshold has to be drawn just above it (Fig. 6b, bottom). Note that approximately five droplets with higher amplitude are tolerated in the negative control yet still considered as negative droplets for thresholding purposes. If more droplets are present, it is considered a contamination or suboptimal ddPCR condition, such as wrong PCR annealing temperature (Fig. 6b, top panel) and further optimization is therefore necessary. If the separation of the clouds is good, the threshold position in the space among clusters has a minimum impact on the quantification (Fig. 6a). 33. Although ddPCR is a powerful technique, it cannot completely eliminate the inherent variability that exists between biological replicates. According to Bio-Rad guidelines, the theoretical

124

Linda Masante et al.

limit of detectable variation is 10%. In practice, however, it is important to consider the specific characteristics of the sample being tested, such as the model, target, and expression levels. Particularly for low copy number targets, the impact of variability is more significant. To mitigate this, it is crucial to achieve a clear separation of the droplet clouds and to respect the minimum number of droplets (i.e., not lower than 10,000). Indeed, for low abundance targets, guaranteeing a minimum number of droplets increases the likelihood of detecting rare events. 34. By keeping the amount of RNA and the cDNA dilution constant, there is already an intrinsic normalization of the copies of the target based on these parameters. The resulting copy number depends on the RNA input and the subsequent dilution after retrotranscription. Always keep in mind that the software computes the target copy number in 20 μL of reaction without taking into account the starting amount of total RNA and the following cDNA dilution. For instance, 500 copies/20 μL of reaction corresponds to 250 copies /ng total RNA, considering a starting amount of 10 ng total RNA with a subsequent 1: 5 cDNA dilution. Potentially, external (spike-in) or internal (canonical housekeeping genes) calibrators can be exploited [23]. External calibration might still be required to detect biases in preanalytical steps (for instance, in RNA extraction or RT efficiency). For example, an external calibrator could be an RNA either not expressed in the tissue investigated, extracted from another species or even a synthetic RNA [29, 30]. Adding such a RNA at a known concentration to the tested sample can give indications on the RNA extraction (if added to the tissue from which the RNA has to be extracted) or on the RT efficacy (if this is added after the RNA extraction), by comparing the ng of RNA in input to the copies of target detected through ddPCR. Given that a reference circRNA is still lacking, internal normalizers are commonly linear RNAs. They still can be somewhat useful if no RNase R treatment is provided and can be employed by expressing results a housekeeping gene with known concentration in the tissue under investigation or of a gene well expressed in a particular sample (for instance, beta actin in the axonal compartment). 35. The expected outcome upon RNase R treatment is a decrease in copy number of the linear RNA isoform (> 90%) but virtually no change in circRNA copy number.

ddPCR for Absolute Quantification of circRNAs

125

Acknowledgments This work was supported by grants from The G. Armenise-Harvard Foundation (Mid-Career Grant) and MUR (Ministero dell’Universita` e della ricerca; PRIN grant) to M.-L.B. References 1. Kristensen LS, Andersen MS, Stagsted LVW et al (2019) The biogenesis, biology and characterization of circular RNAs. Nat Rev Genet. https://doi.org/10.1038/s41576-0190158-7 2. Chen LL (2020) The expanding regulatory mechanisms and cellular functions of circular RNAs. Nat Rev Mol Cell Biol 21:475–490 3. Salzman J, Gawad C, Wang PL et al (2012) Circular RNAs are the predominant transcript isoform from hundreds of human genes in diverse cell types. PLoS One 7:e30733. https://doi.org/10.1371/journal.pone. 0030733 4. Memczak S, Jens M, Elefsinioti A et al (2013) Circular RNAs are a large class of animal RNAs with regulatory potency. Nature 495:333–338. https://doi.org/10.1038/nature11928 5. Jeck WR, Sharpless NE (2014) Detecting and characterizing circular RNAs. Nat Publ Gr 32: 453. https://doi.org/10.1038/nbt.2890 6. Wang PL, Bao Y, Yee M-C et al (2014) Circular RNA is expressed across the eukaryotic tree of life. PLoS One 9:e90859. https://doi.org/10. 1371/journal.pone.0090859 7. Venø MT, Hansen TB, Venø ST et al (2015) Spatio-temporal regulation of circular RNA expression during porcine embryonic brain development. Genome Biol 16. https://doi. org/10.1186/s13059-015-0801-3 8. Westholm JO, Miura P, Olson S et al (2014) Genome-wide analysis of drosophila circular RNAs reveals their structural and sequence properties and age-dependent neural accumulation. Cell Rep 9:1966–1980. https://doi. org/10.1016/j.celrep.2014.10.062 9. Hansen TB, Jensen TI, Clausen BH et al (2013) Natural RNA circles function as efficient microRNA sponges. Nature 495:384– 388. https://doi.org/10.1038/nature11993 10. Capel B, Swain A, Nicolis S et al (1993) Circular transcripts of the testis-determining gene Sry in adult mouse testis. Cell 73:1019–1030. https://doi.org/10.1016/0092-8674(93) 90279-Y 11. Salzman J, Chen RE, Olsen MN et al (2013) Cell-type specific features of circular RNA

expression. PLoS Genet 9:e1003777. https:// doi.org/10.1371/journal.pgen.1003777 12. Guo JU, Agarwal V, Guo H, Bartel DP (2014) Expanded identification and characterization of mammalian circular RNAs. Genome Biol 15:409. https://doi.org/10.1186/s13059014-0409-z 13. Panda AC, De S, Grammatikakis I et al (2017) High-purity circular RNA isolation method (RPAD) reveals vast collection of intronic circRNAs. Nucleic Acids Res 45. https://doi. org/10.1093/nar/gkx297 14. Xiao M-S, Wilusz JE (2019) An improved method for circular RNA purification using RNase R that efficiently removes linear RNAs containing G-quadruplexes or structured 3′ ends. Nucleic Acids Res 47:8755–8769. https://doi.org/10.1093/nar/gkz576 15. Hansen TB (2018) Improved circRNA identification by combining prediction algorithms. Front Cell Dev Biol 1:20. https://doi.org/ 1 0 . 3 3 8 9 / f c e l l . 2 0 1 8 . 0 0 0 2 0 , w w w. frontiersin.org 16. Suzuki H, Zuo Y, Wang J et al (2006) Characterization of RNase R-digested cellular RNA source that consists of lariat and circular RNAs from pre-mRNA splicing. Nucleic Acids Res 34:e63. https://doi.org/10.1093/ nar/gkl151 17. Cocquet J, Chong A, Zhang G, Veitia RA (2006) Reverse transcriptase template switching and false alternative transcripts. Genomics 88:127–131. https://doi.org/10.1016/J. YGENO.2005.12.013 18. Roy CK, Olson S, Graveley BR et al (2015) Assessing long-distance RNA sequence connectivity via RNA-templated DNA-DNA ligation. elife 2015. https://doi.org/10.7554/ ELIFE.03700 19. Nielsen AF, Bindereif A, Bozzoni I et al (2022) Best practice standards for circular RNA research. Nat Methods 2022:1–13. https:// doi.org/10.1038/s41592-022-01487-2 20. Panda AC, Gorospe M Detection and analysis of circular RNAs by RT-PCR. Bio Protoc. https://doi.org/10.21769/BioProtoc.2775

126

Linda Masante et al.

21. Taylor SC, Nadeau K, Abbasi M et al (2019) The ultimate qPCR experiment: producing publication quality, reproducible data the first time. Trends Biotechnol 37:761–774. https:// doi.org/10.1016/j.tibtech.2018.12.002 22. Taylor SC, Laperriere G, Germain H (2017) Droplet digital PCR versus qPCR for gene expression analysis with low abundant targets: from variable nonsense to publication quality data. Sci Rep 7:2409. https://doi.org/10. 1038/s41598-017-02217-x 23. Whale AS, De Spiegelaere W, Trypsteen W et al (2020) The digital MIQE guidelines update: minimum information for publication of quantitative digital PCR experiments for 2020. Clin Chem 66:1012–1029. https://doi.org/10. 1093/clinchem/hvaa125 24. Tan LL, Loganathan N, Agarwalla S et al (2022) Current commercial dPCR platforms: technology and market review. Crit Rev Biot e c h n o l . h t t p s : // d o i . o r g / 1 0 . 1 0 8 0 / 07388551.2022.2037503 25. You X, Vlatkovic I, Babic A et al (2015) Neural circular RNAs are derived from synaptic genes and regulated by development and plasticity. Nat Neurosci 18:603–610. https://doi.org/ 10.1038/nn.3975

26. Errichelli L, Modigliani SD, Laneve P et al (2017) ARTICLE FUS affects circular RNA expression in murine embryonic stem cellderived motor neurons. Nat Commun 8. https://doi.org/10.1038/ncomms14741 27. Legnini I, Di Timoteo G, Rossi F et al (2017) Circ-ZNF609 is a circular RNA that can be translated and functions in Myogenesis. Mol Cell. https://doi.org/10.1016/j.molcel. 2017.02.017 28. Tellinghuisen J (2020) dPCR vs. qPCR: the role of Poisson statistics at low concentrations. Anal Biochem 611:113946. https://doi.org/ 10.1016/J.AB.2020.113946 29. Rowlands V, Rutkowski AJ, Elena M et al Optimisation of robust singleplex and multiplex droplet digital PCR assays for high confidence mutation detection in circulating tumour DnA. Sci Rep. https://doi.org/10.1038/s41598019-49043-x 30. Mika T, Maghnouj A, Klein-Scory S et al (2020) Digital-droplet PCR for quantification of CD19-directed CAR T-cells. Front Mol Biosci 7:533540. https://doi.org/10.3389/ FMOLB.2020.00084/BIBTEX

Chapter 7 Targeted Sequencing of Circular RNAs for Illumina-Based Counting and Nanopore Structure Determination Isabel S. Naarmann-de Vries and Christoph Dieterich Abstract In the past years, circular RNAs (circRNAs) became a major focus of many studies in animals and plants. circRNAs are generated by backsplicing from the same linear transcripts that are canonically spliced to produce, for example, mature mRNAs. They exhibit tissue-specific expression pattern and are potentially involved in many diseases, among them cardiovascular diseases. However, despite the tremendous efforts to establish circRNA catalogues, much less is known about the biological function of the vast majority of circRNAs. We have previously introduced Lexo-circSeq, a targeted RNA sequencing approach that can profile up to 110 circRNAs and their corresponding linear transcripts in one experiment from low amounts of input material on the Illumina platform. Here, we present an improved protocol for Lexo-circSeq and now extend our approach to Nanopore sequencing, which allows the structural assessment of small- and medium-sized circRNAs. Employing human-induced pluripotent stem-cell-derived cardiomyocytes originating from healthy controls or patients suffering from hypertrophic cardiomyopathy, we identify deregulated circRNAs and alternative exon usage. Key words Circular RNA, Nanopore, Illumina

1

Introduction Circular RNAs (circRNAs) are generated mainly by the canonical splicing machinery in a process named backsplicing [16]. In contrast to most linear RNAs, circRNAs are usually not polyadenylated and resistant to exonuclease (e.g., RNase R) digestion due to the lack of free ends [11, 13]. Consequently, circRNAs are usually identified from RNA-seq studies employing rRNA depletion instead of polyA enrichment [1, 16]. To further enrich for circRNA species, the gold standard is RNase R treatment [15]. One major drawback of this approach is however the substantial batch-tobatch variability of RNase R cleavage that has been noted by us [9] and others [2, 11, 17]. Consequently, non-targeted RNA-seqbased studies usually identify a high number of circRNAs, but with

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_7, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

127

128

Isabel S. Naarmann-de Vries and Christoph Dieterich

little overlap between studies and even more between species. This hinders the identification of functionally relevant circRNAs that may be relevant for physiological and pathological processes. Nonetheless, circRNAs may contribute to the progression of various diseases, including cardiovascular diseases (CVDs). We provided recently a catalogue of high-confident and positionally conserved circRNAs that are expressed in human, pig, and mouse hearts. This resource of 63 conserved circRNAs derived from 50 host genes is a valuable resource to elucidate the function of circRNAs in cardiac development and CVD [4]. Furthermore, we identified hundreds of circRNAs with a host gene-independent expression pattern [12], indicating that some of these circRNAs have host-gene-independent functions as well as the importance of monitoring circular-to-linear ratios. In line with this research, we recently presented a targeted short-read circRNA sequencing approach (Lexo-circSeq [9]). This approach enables the panel analysis of circRNAs and their corresponding linear transcripts with low input requirements. The target-specific primers are designed to bind to the 3’ exon relative to the backsplice junction (BSJ). Thus, Illumina short reads (MiSeq, SR300) cover both the BSJ and the canonical (forward) splice junction (FSJ) [9]. Here, we present an updated version of this approach and show that Lexo-circSeq strategy can be adapted to Nanopore sequencing, enabling the study of circRNA architecture. Employing human-induced pluripotent stem-cell-derived cardiomyocytes (hiPSC-CM) derived from healthy donors and patients diagnosed with hypertrophic cardiomyopathy (HCM), we identify deregulated circRNAs and alternative exon usage events.

2

Materials

2.1 Lexogen Library Generation and Amplification

This section summarizes all necessary consumables to generate Lexo-circSeq libraries suitable for Illumina or Nanopore sequencing. The target-specific primers have been published previously [9]: 1. 250 ng (Illumina) to 500 ng (Nanopore) total RNA isolated by standard methods. Store at - 80∘ C. 2. Target-specific first strand synthesis primers [9]. Store at - 20∘C. 3. QuantSeq 3’ mRNA-Seq Library Prep Kit (FWD) for Illumina (Lexogen, #015.24/ 015.96). Store at - 20∘ C. 4. QuantSeq-Flex First Strand Synthesis Module V2 for Illumina (Lexogen, #166.96). Store at - 20∘ C. 5. PCR Add-on Kit for Illumina (Lexogen, #020.96). Store at - 20∘ C. 6. 80% ethanol, freshly prepared.

Targeted Sequencing of Circular RNAs for Illumina-Based Counting. . .

129

7. Qubit Fluorometer (Thermo Fisher Scientific) and Qubit dsDNA HS assay kit (Thermo Fisher Scientific, #Q32854) or equivalent. 8. 5200 Fragment Analyzer System (Agilent) and HS NGS Fragment Kit (1–6000 bp) (Agilent, #DNF-474-1000) or equivalent. 9. Thermocycler. 10. PCR tubes. 11. Magnetic stand for PCR tubes. 2.2 Illumina Sequencing

Lexogen libraries were sequenced at the Cologne Center for Genomicss (CCG) on the instrumentation outlined below: 1. MiSeq system (Illumina) 2. MiSeq Reagent Kit v2 (Illumina)

2.3 Nanopore cDNA Sequencing

This section summarizes all necessary consumables for Nanopore sequencing of Lexo-circSeq libraries on the MinION or similar device. As the libraries are already amplified, we use the direct cDNA approach here: 1. Lexo-circSeq libraries (see above). Use same amounts (determined by Qubit quantification) for Nanopore library preparation. For the data presented here, approximately 15 ng of every library were subjected to Nanopore sequencing. 2. Nuclease-free water. Store at room temperature. 3. Direct cDNA sequencing kit (SQK-DCS109, Oxford Nanopore Technologies). Store at - 20∘ C. 4. Native Barcoding Expansion 1–12 (PCR-free) (EXP-NBD104, Oxford Nanopore Technologies). Store at - 20∘ C. 5. NEBNext Ultra II End Repair/dA-Tailing Module (#E7546S, New England Biolabs). Store at - 20∘ C. 6. Blunt/TA Ligase Master Mix (#M0367S, New England Biolabs). Store at - 20∘ C. 7. NEBNext Quick Ligation Reaction Buffer (#B6058S, New England Biolabs). Store at - 20∘ C. 8. T4 DNA Ligase (#M0202S, New England Biolabs). Store at - 20∘ C. 9. AMPure XP beads (#A63880, Beckman Coulter). Store at 4 °C. 10. 70% ethanol, freshly prepared. 11. Qubit dsDNA HS assay kit (#Q32854) and Qubit Fluorometer (Thermo Fisher Scientific).

130

Isabel S. Naarmann-de Vries and Christoph Dieterich

12. Flow cell priming kit (EXP-FLP002, Oxford Nanopore Technologies). Store at - 20∘ C. 13. Thermocycler. 14. Gentle rotator mixer. 15. Magnetic stand for 1.5 ml tubes. 16. 1.5 ml DNA LoBind tubes (Eppendorf). 17. MinION or GridION sequencing device and MinION R9.4.1 Flow cells (FLO-MIN106D, Oxford Nanopore Technologies). Store Flow cells at 4∘ C. 2.4 Hardware Requirements

The workflow requires a multi-core processor system with minimal main memory of 8 GB RAM and several GBs of free disk space (depending on data set size). We have tested the workflow on Linux and Apple (Mac OS X) computers.

2.5 Software Dependencies and Installation

We employ the latest circtools software suite to analyze the Illumina data: https://github.com/dieterich-lab/circtools. This was version 1.2.2 at the time of writing. The Nanopore read analyses are performed with a custom Python script (https:// github.com/dieterich-lab/circtools/blob/master/scripts/get_cir cJunction_seq.py) and Minimap2 (v2.22) [5].

3

Methods In previous work, we identified a set of conserved circRNAs in the heart [4]. To elucidate the function and possible disease contribution, we reasoned that it is beneficial to focus in further studies on this circRNA set and established a targeted sequencing approach for Illumina short-read sequencing [9]. Here we provide guidelines for the design of first strand primers for targeted sequencing. Furthermore, we present an updated library preparation protocol that prevents sequencing of adapter dimers and is adapted to a new version of the library preparation kit. Finally, we transfer our targeted sequencing approach to long-read sequencing on the Nanopore platform. Doing so, we can derive quantitative information on circRNA expression and structural information from the same library with minimal input requirements.

3.1 Design of First Strand Synthesis Primers

All primers were designed to bind to the 3’ exon of the back splice junction (BSJ). Primers were designed according to the guidelines proposed by Lexogen: All primers were composed of a partial Illumina P7 adapter extension followed by the target-specific sequence. In general, primers should have a length of 45 to 50 nts. The length may

Targeted Sequencing of Circular RNAs for Illumina-Based Counting. . .

131

be adjusted to match the optimal melting temperature (Tm). The Tm should be close to the reaction temperature of 50∘C ( ± 2∘ C). Tms were calculated with an IDT online tool assuming 75 mM Na+, 3 mM Mg2+, 0.5 mM dNTPs, and 0.0125 μM oligo. The primers used in this study (designed for short-read Illumina sequencing) have a medium distance of 25 nts from the BSJ. All primers should be first tested in RT-qPCR analysis by standard methods to estimate abundance of the targeted circRNAs (see below) and assess functionality of the primers. 3.2 Library Generation

1. Mix first strand synthesis primers equimolar with a final concentration of 100 nM. If individual circRNA isoforms are highly abundant, the amount of the respective primers may be reduced. For our use case, the circSLC8A1 primer (targeting the most abundant circRNA in the heart) was used only at a mix concentration of 1 nM. 2. Adjust 250 to 500 ng total RNA and 5 μl Primer mix in a total volume of 14 μl with nuclease-free water in a PCR tube. 3. For targeted first strand synthesis, denature the RNA in a thermocycler (3 min, 85∘ C) and cool down to 50∘ C. Ensure that the temperature of the samples during first strand synthesis does not fall below 50∘ C. 4. Meanwhile, prepare the first strand master mix composed of 5.5 μl buffer FS and 0.5 μl E1. Pre-incubate 2–3 min at 50∘ C. Very briefly spin down the RNA and place back to the thermocycler. Mix the denatured RNA-primer mix with 6 μl first strand master mix by pipetting while leaving the reaction in the thermocycler. Incubate 60 min at 50∘ C. 5. Add 5 μl RNA removal solution and mix by pipetting. Do not remove the samples from the thermocycler. Incubate 10 min at 95∘ C and cool down to 25∘ C. 6. Thaw buffer SS1 at 37∘ C, as it is highly viscous. For second strand synthesis, add 10 μl to the sample and according to this protocol: 1 min at 98∘ C, cooled down to 25∘ C in 0.5∘ C/ s increments, and incubated 30 min at 25∘ C. 7. Prepare a master mix of 4 μl SS2 and 1 μl E2. Add to second strand synthesis reaction and mix by pipetting. Incubate 15 min at 25∘ C. 8. To purify the Lexo-circSeq libaries, equilibrate PB, PS, and EB at room temperature, and resuspend PB (clean up beads). Mix reaction with 16 μl PB. Incubate 5 min at room temperature, collect beads on magnet, and remove supernatant.

132

Isabel S. Naarmann-de Vries and Christoph Dieterich

9. Carefully resuspend in 40 μl EB and incubate 2 min at room temperature. Add 56 μl PS, incubate 5 min at room temperature, collect beads on magnet, and remove supernatant. 10. Wash two times with 120 μ l 80% ethanol (leave 30 s on magnet). Carefully remove all ethanol and dry beads for 5 to 10 min at room temperature. 11. To elute the libraries, resuspend beads in 18 μl EB, and incubate 2 min at room. Collect beads on magnet and transfer 17 μl eluate into a new tube. 3.3 Amplification of Libraries

Indices for Illumina sequencing should be introduced during this amplification step. Therefore, the PCR is performed with specific i7 index primers. For choice of compatible indices, the Lexogen Index Balance Checker (https://www.lexogen.com/support-tools/ index-balance-checker/) may be used. For Nanopore sequencing, barcoding takes place at a later step. Due to this, amplification of the libraries is performed with the PCR Add-on kit and the universal i7 primer (P7000). Despite these differences, PCR and subsequent purification follow identical protocols: 1. Mix in a PCR tube 17 μl library with 7 μl PCR mix, 1 μl E3, and 5 μl i7 index primer or universal primer, respectively. 2. Use the following program to amplify the Lexo-circSeq libraries: initial denaturation (98∘ C, 30 s), 20 amplification cycles (98∘C, 10 s; 65∘C, 20 s; 72∘C, 30 s), and final elongation (72∘ C, 1 min). 3. To purify the amplified Lexo-circSeq libaries, equilibrate PB, PS, and EB at room temperature and resuspend PB (clean up beads). Mix reaction with 27 μl (0.9x) PB. Incubate 5 min at room temperature, collect beads on magnet, and remove supernatant. 4. Carefully resuspend in 30 μl EB and incubate 2 min at room temperature. Add 30 μl PS, incubate 5 min at room temperature, collect beads on magnet, and remove supernatant. 5. Wash two times with 120 μ l 80% ethanol (leave 30 s on magnet). Carefully remove all ethanol and dry beads for 5 to 10 min at room temperature. 6. To elute the libraries, resuspend beads in 20 μl EB and incubate 2 min at room temperature. Collect beads on magnet, and transfer 17 μl eluate into a new tube. 7. Determine the concentration of the final libraries by Qubit measurement with the Qubit dsDNA HS kit according to the manufacturers protocol. The quality of the libraries should be assessed on a Fragment Analyzer, Tape Station or comparable.

Targeted Sequencing of Circular RNAs for Illumina-Based Counting. . .

3.4 Nanopore cDNA Sequencing

133

1. Identical amounts of cDNA should be subjected to Nanopore sequencing. Calculate this based on the library with the lowest concentration, and adjust the volume to 16 μl per sample. 2. In a PCR tube, set up the end repair & dA-tailing reaction: mix 16 μl diluted library with 34 μl nuclease-free water, 7 μl Ultra II End-prep reaction buffer, and 3 μl Ultra II End-prep enzyme mix. Incubate in a thermocycler 5 min at 20∘ C, followed by 5 min at 65∘ C. 3. Prepare reverse transcription master mix on ice during ligation: 9 μl nuclease-free water, 2 μl 10 mM dNTPs, 8 μl 5x SuperScript IV first strand buffer, 4 μl 0.1 mM DTT. 4. Let the AMPure XP beads come to room temperature. Carefully resuspend beads before use. Transfer reaction to a 1.5 ml DNA LoBind tube, and mix with 60 μ l AMPure XP beads. Incubate 5 min at room temperature on a gentle rotator mixer. 5. Collect beads on a magnetic stand and remove supernatant. Wash pelleted beads two times (30 s) with 200 μ l freshly prepared 70% ethanol. Remove supernatant. Spin sample down and place on magnet again. Remove any residual ethanol. 6. Resuspend beads in 22.5 μl nuclease-free water by gentle flicking, and incubate 2 min at room temperature on a gentle rotator mixer. Collect beads on a magnetic stand, and transfer 22.5 μl eluate in a fresh 1.5 ml DNA LoBind tube. 7. For barcode ligation, add the following to 22.5 μ l end-prepped DNA: 2.5 μ l Native Barcode (from EXP-NBD104) and 25 μl Blunt/TA Ligase master mix. Mix by pipetting, and incubate 10 min at room temperature. 8. Add 50 μ l carefully resuspended AMPure XP beads to the reaction and mix by pipetting. Incubate 5 min at room temperature on a gentle rotator mixer. 9. Collect beads on a magnetic stand and remove supernatant. Wash pelleted beads two times (30 s) with 200 μ l freshly prepared 70% ethanol. Remove supernatant. Spin sample down, and place on magnet again. Remove any residual ethanol. 10. Resuspend beads in 15 μl nuclease-free water by gentle flicking, and incubate 2 min at room temperature on a gentle rotator mixer. Collect beads on a magnetic stand, and transfer 15 μl eluate in a fresh 1.5 ml DNA LoBind tube. 11. Determine the concentration of the barcoded libraries by Qubit measurement with the Qubit dsDNA HS kit according to the manufacturers protocol. 12. Pool libraries in equal amounts, with the highest possible concentration in a total volume of 65 μl.

134

Isabel S. Naarmann-de Vries and Christoph Dieterich

13. Set up the adapter ligation reaction: 65 μl pooled and barcoded sample, 5 μl Adapter Mix II (from EXP-NBD104), 20 μl NEB Next Quick ligation buffer, 10μ l T4 DNA ligase. Incubate 10 min at room temperature. 14. Add 50 μ l carefully resuspended AMPure XP beads to the reaction and mix by pipetting. Incubate 5 min at room temperature on a gentle rotator mixer. 15. Collect beads on a magnetic stand and remove supernatant. Wash pelleted beads two times with 140 μl wash buffer (WSB, from SQK-DCS109). Resuspend beads by flicking, spin down, and return to magnetic stand. Remove supernatant from pelleted beads. 16. Resuspend beads in 13 μ l elution buffer (EB, from SQK-DCS109) by gentle flicking, and incubate 10 min at room temperature on a gentle rotator mixer. Pellet beads on a magnetic stand and transfer 13 μ l eluate in a fresh 1.5 ml DNA LoBind tube. 17. Quantify 1 μl of the library on a Qubit fluorometer with the Qubit dsDNA HS kit according to the manufacturer’s protocol. 18. Insert MinION R9.4.1 Flow cell in the MinION or GridION sequencing device, and perform Flow cell check in the MinKNOW software. 19. Prepare Priming Mix by adding 30 μl flush tether (FLT, from EXP-FLP002) to a vial of flush buffer (FB, from EXP-FLP002) and mix by pipetting. Open priming port. Remove air bubble from priming port by inserting the tip of a P1000 pipette into the priming port and slowly dialing up, until a small volume of storage buffer enters the pipette tip. Load 800 μl Priming Mix via the priming port and carefully avoid introduction of air bubbles. Close the priming port and wait for 5 min. 20. Mix 12 μl library with 37.5 μl sequencing buffer (SQB, from SQK-DCS109) and 25.5 μ l loading beads (LB, from SQK-DCS109), and mix by pipetting. Take care to carefully resuspend the beads before pipetting. Open the priming port and the sample port. Load 200 μl Priming Mix via the priming port. Mix library by pipetting just before loading and load dropwise via the sample port. Carefully avoid introduction of air bubbles. Close the sample port and the priming port. 21. Start sequencing for 48 to 72 h in the MinKNOW software. Choose direct cDNA sequencing kit and high-accuracy base calling as parameters.

Targeted Sequencing of Circular RNAs for Illumina-Based Counting. . .

3.5 Illumina Read Processing

135

1. The first step is read processing and mapping. All relevant details can be found online: https://docs.circ.tools/en/ latest/Detect.html. It is highly recommended to assess the library quality with the subcommand quickcheck—see: https://docs.circ.tools/en/latest/Quickcheck.html. 2. Counts for circular and linear RNAs were obtained with circtools 1.2.2 using the subcommand detect [3]. The respective output files are described here: https://docs.circ.tools/ en/latest/Detect.html#output-files. 3. From these counts, the circular-to-linear ratio (CLR, [circ]/ ([circ]+[lin]) and read per million (RPM) were derived. 4. The mean of ctrl. and HCM CLR or RPM was plotted against each other using GraphPad Prism, which was also used to calculate the Pearson correlation coefficient.

3.6 Nanopore Read Processing

1. The first step is to convert the ionic current signal trace, which is stored in FAST5 files, into base calls. Typically, users apply real-time base calling with the MinKNOW-embedded Guppy base caller in high-accuracy mode on their Nanopore sequencing device. For base calling post-sequencing, we recommend to use ONT’s Guppy base caller. For the latter, the base caller minimally requires the path to FAST5 files, the output folder, and the config file or the flowcell/kit combination. The output are FASTQ files that can be compressed using the option “– compress_fastq.” $

guppy_basecaller

--compress_fastq

-i

path_to_fast5 -s path_to_output -c config_file. cfg --cpu_threads_per_caller 14 --num_callers 1

Set the number of threads “cpu_threads_per_caller” and the number of parallel base callers “num_caller” according to your resources. The optimal choice of the config file depends on the employed kit chemistry and flow cell type. We have used dnar 9:4:14 50bpsh ac on our own data. Additional details can be found at https://nanoporetech.com/. 2. Generate a circRNA panel-specific reference sequence.: We supply a Python script get_circJunction_seq.py under https://github.com/dieterich-lab/circtools/blob/master/ scripts/get_circJunction_seq.py to prepare database records, which represent a linearized version of the BSJ sequence region. Briefly, we capture the genomic region where the first strand primer is positioned and add the sequence region in which the first strand synthesis extends into.

136

Isabel S. Naarmann-de Vries and Christoph Dieterich $ get_circJunction_seq.py

-bed BSJcoordinates.

bed -db genome.fa > circularRNAcontigs.fa $ cat genome.fa circularRNAcontigs.fa > reference. fa

3. Create Minimap2 database: Minimap2 is a versatile mapper for long reads (see https://lh3.github.io/minimap2/, 5). We recommend to pre-index the reference sequence (reference. fa) and store the index. This saves time in the read alignment step if the same circRNA primer panel is used repeatedly. Please save the index with the option “-d” before read mapping and use the index instead of the reference file in the minimap2 command line. $ minimap2 -d reference.mmi reference.fa

4. Align reads to the database: We use minimap 2.22 to map Nanopore cDNA reads and samtools 1.16 to convert alignments into BAM format. We recommend to output only primary alignments by adding -secondary=no to the command line call. We also include the option -MD to add reference sequence information to the read alignments. Please adjust the number of threads (option -t) according to your resources and check the Minimap2 manual for more details. To enable spliced alignments, use the setting -ax splice -junc-bed annotation.bed -junc-bonus where the latter allows to tune the bonus score for an annotated splice site and the BED file provides the splice junction annotation. $ minimap2 -t 5 --MD -ax splice --junc-bonus 1 -secondary=no --junc-bed final_annotation_Ensembl96.bed -ub reference.mmi Reads.fastq.gz | samtools view -bS > mapping.bam

The BED file can be generated from EnsEMBL GTF files using the following command, which is part of the minimap2 distribution: $paftools.js gff2bed annotation.gtf > annotation.bed

4

Results Lexo-circSeq was already applied to the analysis of biopsy samples derived from healthy controls and patients diagnosed with HCM and dilated cardiomyopathy (DCM) [9]. Based on the CLR, this

Targeted Sequencing of Circular RNAs for Illumina-Based Counting. . .

137

analysis identified circRBM33 and circSLC8A1 as candidate deregulated circRNAs in cardiomyopathies [9]. Due to the limited availability of human biopsy samples and occasionally non-optimal RNA quality, we moved here to a model system for cardiomyopathies: hiPSC-CM derived from healthy donors and HCM patients. CircRNAs are rare, compared to other RNA classes. Furthermore, the overall circRNA level may strongly differ between different types of tissues and cell lines. It is therefore highly recommended to carefully check the quality of the libraries and also the overall quality of the obtained sequencing data prior to detailed analyses. Prior to sequencing on the Illumina platform, libraries should be analyzed on a Fragment Analyzer or equivalent. The proposed Lexo-circSeq protocol generates relatively long inserts for a shortread sequencing approach (in the range of 300 bp). The resulting products can be easily separated from adapter dimers ( 150 bp). The optimized clean-up protocol efficiently removes these adapter dimers (Fig. 1a). If the proportion of adapter dimers is still high for samples processed according to the optimized protocol, the input samples may have been of low quality, and one should consider to identify samples of higher quality for sequencing. Furthermore, the relative proportion of chimeric reads should be checked for the Illumina results. In general, the fraction of circRNAs may vary strongly between different sample types; however, comparable samples should contain similar fractions of circRNAs. A fraction of chimeric reads strongly deviating from the mean may indicate library quality issues. In the samples analyzed here, we find between 3 and 10% of chimeric reads in the uniquely mapped reads and no significant differences (Fig. 1b). A

B 12

1000

opt.

800 600 400 200 0

% chimeric reads

std.

RFU

4.1 Library and Sequencing Quality Control

10 8 6 4 2 0

200

400

600

800

ctrl.

HCM

size [nt]

Fig. 1 Quality control of Lexo-circSeq libraries. (a) Fragment Analyzer run comparing the standard purification and optimized purification of amplified libraries for an exemplary sample. The peak at 150 bp represents adapter dimers. (b) Fraction of chimeric reads in the ctrl. and HCM samples analyzed. Box plot displaying the 5–95% percentile

138

Isabel S. Naarmann-de Vries and Christoph Dieterich

A

B

HCM (RPM)

10 4

0.8

circular linear

HCM (CLR)

10 5

10 3 10 2

ZNF148

r = 0.9690 r = 0.9072

10 1

10 2

10 3

10 4

UXS1

0.0 0.0

10 5

C

0.2

0.4

r = 0.8939

0.6

0.8

ctrl. (CLR)

ctrl. (RPM)

D 0.8

0.6

0.6

CLR UXS1

CLR ZNF148

SLC8A1

0.4 0.2

10 1 10 0 10 0

MTHFD2L

0.6

0.4

*

0.4

0.2

0.2 0.0

0.0 ctrl.

HCM

ctrl.

HCM

Fig. 2 Analysis of Illumina sequencing data. (a) Mean RPM for ctrl. vs. HCM. The Pearson correlation coefficient was calculated separately for circular and linear reads. (b) Circular-to-linear ratio (CLR) for ctrl. vs. HCM samples. Pearson correlation coefficient. The solid black line represents the linear regression line, and the red dotted lines the 95% confidence interval. (c) CLR for ZNF148. (d) CLR for UXS1. (Unpaired two-tailed t-test, p < 0:05 = )

4.2 Analysis of circRNA Levels by Illumina Sequencing

We recently introduced Lexo-circSeq as targeted circRNA sequencing approach on the Illumina platform and identified circRBM33 and circSLC8A1 as potentially deregulated circRNAs in cardiomyopathies [9]. Here, we applied our optimized Lexo-circSeq protocol to four samples each derived from hiPSC-CMs that were either WT (ctrl.) or had a HCM-related genotype (Fig. 2). While human samples may suffer from low RNA quality, cell cultures allow the isolation of high-quality RNA. Thus, these cell-culture-based model systems represent a valuable resource to identify basic disease mechanisms. We employed circtools [3] to identify read counts for circRNAs and their linear counterparts (Fig. 3a). As expected, the expression of most linear transcripts is higher than their circular counterparts (Fig. 2a). While we detect almost no difference in the expression of the targeted mRNAs between ctrl. and HCM samples, the analyzed circRNAs (especially the lowly expressed ones) show larger differences in expression levels between ctrl. and HCM (Fig. 2a).

Targeted Sequencing of Circular RNAs for Illumina-Based Counting. . .

139

Fig. 3 Analysis of Nanopore sequencing data. (a) Schematic representation of Nanopore sequencing of Lexo-circSeq libraries and alignment to an artificial reference that covers the BSJ (b) Analysis of circNSD2 by Nanopore sequencing (c) Analysis of circUXS1 by Nanopore sequencing

Besides the expression level, the circular-to-linear ratio (CLR) is important. While the functional impact of circRNAs, whose expression level changes along with the expression level of a corresponding mRNA, is expected to be limited, circRNAs with changes in the CLR may be of particular interest. The analysis of the CLR in ctrl. vs. HCM samples revealed only minor changes for most circRNAs (Fig. 2b). However, some circRNAs exhibit significant changes in the CLR: The relative fraction of circZNF148 (Fig. 2c) and circMTHFD2L is increased in HCM compared to the linear transcript, whereas circUXS1 (Fig. 2d) and circSLC8A1 show a decreased CLR in HCM (Fig. 2b). Interestingly, circSLC8A1 was previously identified as candidate deregulated circRNA in cardiomyopathies [9] and reported to be relevant in cardiac development and pathogenesis [6, 7, 14]. Similar to the analysis of biopsy samples, we detected a high variability of circRNA expression levels and CLR between the

140

Isabel S. Naarmann-de Vries and Christoph Dieterich

samples of each group, as in the case of circZNF148 (Fig. 2c) rendering it difficult to identify significantly different circRNAs as circUXS1 (Fig. 2d). Thus, whenever possible, a high number of replicates (min. n = 5) should be considered for analysis. 4.3 Identification of Full-Length circRNA Architecture by Nanopore Long-Read Sequencing

5

The architecture of circRNAs cannot be derived without uncertainty from short-read sequencing data, as only reads covering the BSJ can be uniquely assigned to the circRNA. Consequently, longread sequencing approaches are required to elucidate the fulllength sequence of circRNAs. Several approaches have been presented in the literature that are dedicated to long-read Nanopore sequencing [8, 10, 18]. We show here that Lexo-circSeq libraries can also be directly analyzed on the Nanopore, revealing structural information from the same library (Fig. 3). Amplified Lexo-circSeq libraries are barcoded and subjected to ONT direct cDNA sequencing. The obtained reads are then aligned to an artificial reference that contains the BSJ (Fig. 3a). The target-specific primers for Lexo-circSeq sequencing bind in the 3’ exon relatively close to the BSJ to ensure proper identification of circRNA reads by Illumina sequencing [9]. The use of the artificial reference enables correct mapping of these reads also in long-read sequencing. The median read length in Nanopore Lexo-circSeq is 470 nts. This allows the full-length reconstruction of small to medium-sized circRNAs. For larger circRNAs, at least a partial architecture can be defined. The analysis of circNSD2 (predicted size: 788 nts, located on the plus strand) interestingly revealed that alternative exons are used in ctrl. samples compared to HCM samples (Fig. 3b). Also, in case of circUXS1 (predicted size: 377 nts, located on the minus strand) (Fig. 3c), alternative exon usage was detected for a subset of circRNA reads. In summary, we present here an updated protocol of our targeted Lexo-circSeq protocol and demonstrate Nanopore long-read sequencing as add-on option to elucidate the circRNA structure as well.

Notes

1. To analyze gene expression, a control reaction with oligo (dT) primer may be run in parallel. Please note that during first strand synthesis, this reaction has to be incubated at 42∘ C instead of 50∘ C.

Targeted Sequencing of Circular RNAs for Illumina-Based Counting. . .

141

2. Lexogen suggests some protocol adjustments for low-quality RNA. For the targeted circRNA sequencing, we found these protocol adjustments beneficial also if starting from highquality RNA. While some adjustments were already included in the initially published protocol [9], we now reduced the beads volume in the purification after library amplification to 27 μ l (0.9x reaction volume). This drastically reduces the amount of co-purified adapter dimers. 3. The optimal cycle number for library amplification in this work has been determined by the qPCR Assay employing the PCR add-on kit as described by the manufacturer. We highly recommend this step when setting up new panels for targeted sequencing of circRNAs, as the required cycle number may differ dependent on the abundance of the targeted circRNAs. 4. The first strand synthesis primers used in this analysis were designed for Illumina sequencing. If Nanopore sequencing is considered already at the start of a project, we recommend to place the primers approximately 50–100 nts apart from the BSJ.

Acknowledgements This work was kindly supported by the Klaus-Tschira foundation through grant 00.013.2021. We would like to thank Shubhada Kulkarni for her support in Python scripting. We are grateful to Malte Tiburcy (University Medical Center Go¨ttingen) for providing the hiPSC-CM pellets. We would like to thank Christian Becker (Cologne Center for Genomics) for excellent sequencing support and Jessica Eschenbach for excellent technical assistance. References 1. Chen L-L (2020) The expanding regulatory mechanisms and cellular functions of circular RNAs. Nat Rev Mol Cell Biol 21(8):475–490 2. Dodbele S, Mutlu N, Wilusz JE (2021) Best practices to ensure robust investigation of circular RNAs: pitfalls and tips. EMBO Rep 22(3):e52072 3. Jakobi T, Uvarovskii A, Dieterich C (2019) Circtools–a one-stop software solution for circular RNA research. Bioinformatics 35(13): 2326–2328 4. Jakobi T, Siede D, Eschenbach J, Heumu¨ller AW, Busch M, Nietsch R, Meder B, Most P, Dimmeler S, Backs J, et al (2020) Deep characterization of circular RNAs from human cardiovascular cell models and cardiac tissue. Cells 9(7):1616

5. Li H (2021) New strategies to improve minimap2 alignment accuracy. Bioinformatics 37(23):4572–4574 6. Li M, Ding W, Tariq MA, Chang W, Zhang X, Xu W, Hou L, Wang Y, Wang J (2018) A circular transcript of NCX1 gene mediates ischemic myocardial injury by targeting miR-133a-3p. Theranostics 8(21):5855 7. Lim TB, Aliwarga E, Luu TDA, Li YP, Ng SL, Annadoray L, Sian S, Ackers-Johnson MA, Foo RS-Y (2019) Targeting the highly abundant circular RNA circslc8a1 in cardiomyocytes attenuates pressure overload induced hypertrophy. Cardiovasc. Res. 115(14):1998–2007 8. Liu Z, Tao C, Li S, Du M, Bai Y, Hu X, Li Y, Chen J, Yang E (2021) circfl-seq reveals fulllength circular RNAs with rolling circular

142

Isabel S. Naarmann-de Vries and Christoph Dieterich

reverse transcription and nanopore sequencing. elife 10:e69457 9. Naarmann-de Vries IS, Eschenbach J, Schudy S, Meder B, Dieterich C (2022) Targeted analysis of circRNA expression in patient samples by Lexo-circSeq. Front. Mol. Biosci. 9: 875805 10. Rahimi K, Venø MT, Dupont DM, Kjems J (2021) Nanopore sequencing of brain-derived full-length circRNAs reveals circRNA-specific exon usage, intron retention and microexons. Nat Commun 12(1):1–15 11. Salzman J, Gawad C, Wang PL, Lacayo N, Brown PO (2012) Circular RNAs are the predominant transcript isoform from hundreds of human genes in diverse cell types. PloS One 7(2):e30733 12. Siede D, Rapti K, Gorska AA, Katus HA, Altmu¨ller J, Boeckel JN, Meder B, Maack C, Vo¨lkers M, Mu¨ller OJ, et al (2017) Identification of circular RNAs with host geneindependent expression in human model systems for cardiac differentiation and disease. J Mol Cell Cardiol 109:48–56 13. Suzuki H, Zuo Y, Wang J, Zhang MQ, Malhotra A, Mayeda A (2006)

Characterization of RNase R-digested cellular RNA source that consists of lariat and circular RNAs from pre-mRNA splicing. Nucleic Acids Res 34(8):e63–e63 14. Tan WLW, Lim BTS, Anene-Nzelu CGO, Ackers-Johnson M, Dashi A, See K, Tiang Z, Lee DP, Chua WW, Luu TDA, et al (2017) A landscape of circular RNA expression in the human heart. Cardiovasc. Res. 113(3): 298–309 15. Xiao M-S, Wilusz JE (2019) An improved method for circular RNA purification using RNase R that efficiently removes linear RNAs containing G-quadruplexes or structured 3’ ends. Nucleic Acids Res 47(16):8755–8769 16. Xiao M-S, Ai Y, Wilusz JE (2020) Biogenesis and functions of circular RNAs come into focus. Trends Cell Biol 30(3):226–240 17. Zhang J, Chen S, Yang J, Zhao F (2020) Accurate quantification of circular RNAs identifies extensive circular isoform switching events. Nat Commun 11(1):1–14 18. Zhang J, Hou L, Zuo Z, Ji P, Zhang X, Xue Y, Zhao F (2021) Comprehensive profiling of circular RNAs with nanopore sequencing and CIRI-long. Nat. Biotechnol. 39(7):836–845

Chapter 8 Nanopore-Mediated Sequencing of Circular RNA Morten Trillingsgaard Venø, Junyi Su, Yan Yan, and Jørgen Kjems Abstract Circular RNAs (circRNAs) constitute a group of RNAs defined by a covalent bond between the 5′ and 3′ end formed by a unique back-splicing event. Most circRNAs are composed of more than one exon, which are spliced together in a linear fashion. This protocol describes methods to sequence full-length circRNA across the back-splicing junction, allowing unambiguous characterization of circRNA-specific exon-intron structures by long-read sequencing (LRS). Two different sequencing approaches are provided: (1) Global circRNA sequencing (the circNick-LRS strategy) relying on circRNA enrichment from total RNA followed by total circRNA long-read sequencing, and (2) targeted circRNA sequencing (the circPanel-LRS strategy) where a preselected panel of circRNA are sequenced without prior circRNA enrichment. Both methods were originally described in Karim et al. (Rahimi et al., Nat Commun 12: 4825, 2021) where they were applied to characterize the exon-intron structure of >10.000 circRNAs in mouse and human brains. Key words Circular RNA, Back-splicing, Long-read sequencing, Nanopore, Exon

1

Introduction Linear and circular RNAs are both derived from the same precursor RNA and, apart from the back-splice junction, share similar sequences [2]. The length of circular RNA (circRNA) ranges from about one hundred nucleotides to many thousands and it is therefore not possible to cover by short read sequencing protocols like Illumina. Long-read sequencing (LRS) is made possible by the invention of several new platform technologies including Oxford Nanopore Technologies (ONT; [3]) or Pacific Biosciences (PacBio; [4]). Another major challenge when studying circRNA is the low abundance of circRNA. It is estimated that in a given mammalian cell, less than 1 out 1000 of all RNA molecules is circular [5]. Considering the more modest read capacity of ONT sequencing, it is therefore necessary to enrich for circRNA prior to sequencing. The present protocol describes how to prepare libraries for ONT through several enrichment steps to deplete linear RNA from total RNA (described in Subheading 3.1; see Fig. 1a). The enriched

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_8, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

143

144

Morten Trillingsgaard Venø et al.

Fig. 1 CircRNA enrichment and adapting cDNA-PCR protocol for long-read circRNA sequencing. (a) Total RNA is subjected to several steps to enrich for circRNA species. Ribosomal RNA is depleted, followed by RNase R treatment to digest linear RNA. The remaining linear RNA is polyadenylated and depleted by oligo-dT beads. Enriched circRNA pool is linearized by limited hydrolysis and polyadenylated to be compatible with cDNA-PCR long-read sequencing kit provided by ONT. (b) Linearized and polyadenylated circRNA is reverse transcribed using oligo-dT primers (VNP primer) and strand switching is induced for second strand synthesis. PCR amplification is performed using primers complementary to two primer target regions introduced by extended oligo-dT primers and strand switching oligos. After PCR amplification, the library is size selected from 350 bp to 10 kb and subjected to a second round of PCR amplification. The library is cleaned up and the ONT sequencing adapter is ligated. Samples are now loaded into flowcells and sequenced using MinION

circRNA samples were subsequently linearized, poly-adenylated, reverse transcribed, amplified by PCR, and subjected to ligating the ONT sequencing adapters before loading for full-length sequencing (described in Subheading 3.2; see Fig. 1b). We have furthermore developed a method for long-read sequencing of a targeted panel of circRNAs called circPanel-LRS. This method is highly efficient and can be used for samples with low amount of RNA and without need for prior circRNA enrichment. It is possible to multiplex different samples of interest and sequence them together to reduce the running cost of the sequencing experiments. Both sequencing methods have been adapted from the ONT cDNA-PCR sequencing protocol SQK-PCB109 which is an approach for full-length poly(A) RNA sequencing, but with modifications to capture a panel of circRNAs of interest instead of linear RNA [1].

Nanopore-Mediated Sequencing of Circular RNA

145

Three papers have described four methods for nanoporedependent long-read sequencing of circRNAs [1, 6, 7]. See Rahimi et al. [8] for reviewing the key features in each method and for discussion of advantages and disadvantages.

2

Material

2.1 Enriching CircRNAs from Total RNAs

1. TRIzol. LS reagent (Life Technologies). 2. Zymo directzol RNA miniprep (R2072, Zymo Research). 3. DNase I (1 U/μL) (Thermo Fisher Scientific). 4. DNA LoBind tubes (Eppendorf). 5. Nuclease-free water (NEB). 6. RNA Clean and ConcentratorTM -5 (R1016, Zymo Research). 7. Qubit 4.0 fluorometer (Thermo Fisher Scientific). 8. Qubit HS RNA kit (Thermo Fisher Scientific). 9. Agilent 2100 bioanalyzer (Agilent Technologies). 10. Nanodrop (Thermo Fisher Scientific). 11. Ribo-Zero Illumina).

rRNA

removal

kit

(Human/Mouse/Rat,

12. RNase R (20 U/μL) (Epicentre, RNR07250). 13. Poly(A) polymerase (5 U/μL) (NEB, M0276). 14. NEBNext Poly(A) mRNA Magnetic Isolation Module (NEB, E7490S). 15. NEBNext Magnesium RNA Fragmentation Module (NEB, E6150). 16. T4 Polynucleotide Kinase (10 U/μL) (NEB, M0201S). 17. Superscript VILO cDNA Synthesis Kit (Thermo Fisher Scientific). 18. LightCycler 480 and LightCycler 480 SYBR Green I Master Kit (Roche). 2.2 Long-Read Sequencing of Enriched CircRNA Pool

1. SQK-PCS108 Oxford Nanopore Technology sequencing kit (ONT). 2. Nuclease-free water (NEB). 3. VNP primer (ONT). 4. 10 mM dNTP mix (2.5 mM of each nucleotoide). 5. Superscript IV reverse transcriptase (200 U/μL) (Thermo Fischer Scientific). 6. Superscript IV reverse buffer (Thermo Fischer Scientific).

146

Morten Trillingsgaard Venø et al.

7. 10 μM Strand-Switching Primer (SSP) (ONT). 8. RNaseOUT (40 U/μL) (Thermo Fischer Scientific) or RiboLock RNase inhibitor (40 U/μL) (Thermo Fischer Scientific). 9. PCR machine with hot lid up to 105 °C. 10. LongAmp HS Taq Master Mix (M0533L, NEB). 11. cDNA primer mix (ONT). 12. Exonuclease I (20 U/μL) (M0293, NEB). 13. 2% TBE agarose gel (UltraPure, Thermo Scientific). 14. GeneJet Gel Extraction Kit (Thermo Fischer Scientific). 15. DNA LoBind tubes (Eppendorf). 16. SPRISelect beads (Beckman Coulter). 17. Freshly prepared 80% Ethanol. 18. Magnetic racks for 1.5 mL Eppendorf tubes. 19. Rapid Annealing Buffer (RAB) (ONT). 20. cDNA adapter mix (cAMX) (ONT). 21. Adapter Bead Binding Buffer (ABB) (ONT). 22. Elution buffer (ELB) (ONT). 23. Qubit 4.0 fluorometer (Thermo Fisher Scientific). 24. Qubit HS dsDNA kits (Thermo Fisher Scientific). 25. Running Buffer with Fuel mix (RBF) (ONT). 26. Library Loading Beads (ONT). 27. Flow cell R9.4.1 (ONT). 28. MinION or GridION sequencer (ONT). 2.3 Full-Length Sequencing of a Targeted Panel of CircRNAs

1. 10 mM dNTP mix (2.5 mM of each nucleotide). 2. Nuclease-free water (NEB). 3. CircRNA Specific Primers for each circRNA (CSP, see Subheading 3.2, item 1 for design). 4. PCR machine with hot lid up to 105 °C. 5. Maxima H Minus Reverse Transcriptase (200 U/μL) (Thermo Fisher Scientific). 6. 5X Reverse Transcriptase buffer (Thermo Fisher Scientific). 7. RNaseOUT (40 U/μL) (Thermo Fischer Scientific) or RiboLock RNase inhibitor (40 U/μL) (Thermo Fischer Scientific). 8. RNase H (5 U/μL) (Thermo Fisher). 9. RNAse cocktail (Thermo Fisher). 10. LongAmp HS Taq Master Mix (M0533L, NEB). 11. Exonuclease I (20 U/μL) (M0293, NEB). 12. SPRISelect beads (Beckman Coulter).

Nanopore-Mediated Sequencing of Circular RNA

147

13. 1.5 mL low-bind Eppendorf tube. 14. Freshly prepared 80% Ethanol. 15. Magnetic racks for 1.5 mL Eppendorf tubes. 16. Qubit 4.0 fluorometer (Thermo Fisher Scientific). 17. Qubit HS dsDNA kit (Thermo Fisher Scientific). 18. Agilent 2100 bioanalyzer (Agilent Technologies). 19. Nanodrop (Thermo Fisher Scientific). 20. SQK-PCB109 Oxford Nanopore Technology sequencing kit (ONT). 21. SQK-PCS109 Oxford Nanopore Technology sequencing kit (ONT). 22. Barcode Primers (provided by ONT SQK-PCB109 kit, ONT). 23. cPRM primer (provided by ONT SQK-PCB/PCS109 kit, ONT). 24. Elution buffer (EB) (provided by ONT SQK-PCB/PCS109 kit, ONT). 25. DNA High Sensitivity kit (Agilent Technologies). 26. Rapid Adapter (RAP) (ONT). 27. Microfuge (Thermo Fisher Scientific). 28. Flow cell R9.4.1 (ONT). 29. MinION or GridION sequencer.

3

Methods

3.1 Enriching CircRNA from Total RNA (Outlined in Fig. 1a) (See Note 1)

1. Prepare total RNA from tissue or cultured cells (see Note 2). DNase I was used for removal of DNA contamination in the primary total RNA samples. RNA quantity was measured by Qubit 4.0 fluorometer along with the Qubit HS RNA kit and RNA quality was assessed by Agilent 2100 bioanalyzer. 2. To deplete rRNA from 20 μg of total RNA (RNA integrity number > 8), use Ribo-Zero rRNA removal kit according to the manufacture’s manual. 3. Treat rRNA-depleted samples with RNase R following manufacture’s instruction to remove as much linear RNA as possible. 4. Clean up RNA using RNA Clean and ConcentratorTM -5 (see Note 3). 5. To deplete remaining linear RNA, polyadenylate the RNase R-digested RNA using poly(A) polymerase according to the manufacture’s instruction. 6. Clean up the RNA using RNA Clean and ConcentratorTM -5 (see Note 4).

148

Morten Trillingsgaard Venø et al.

7. Deplete the polyadenylated linear RNA from the circRNA pool using oligo-dT beads provided in NEBNext Poly(A) mRNA Magnetic Isolation Module following the kit’s manual. The enriched circRNA pool is the supernatant obtained from the first step of the poly(A) purification while the polyadenylated linear RNA are annealed to the oligo-dT magnetic beads that are collected using the magnetic rack. 8. Take the supernatant. 9. Clean up the RNA using RNA Clean and ConcentratorTM -5. 10. Use the NEBNext Magnesium RNA Fragmentation Module to linearize the enriched circRNA according to the kit’s instruction. 11. Divide the RNA sample into three reactions and incubate them at 80 °C for 30, 60, and 90 s, respectively, before pooling them again. 12. Clean up the RNA using RNA Clean and ConcentratorTM 5. 13. Phosphorylate and de-phosphorylate the 5′ and 3′ ends of the linearized circRNAs, respectively, using T4 Polynucleotide Kinase according to the manual instructions. 14. Clean up RNA using RNA Clean and ConcentratorTM -5. 15. Polyadenylate the RNA using poly(A) polymerase according to the manufacturer’s instruction. 16. Clean up the polyadenylated RNA using RNA Clean and ConcentratorTM -5. 17. Continue with ONT sequencing (Subheading 3.2) or store RNA at -80 °C. 3.2 Long-Read Sequencing of Enriched CircRNA Pool (circNick-LRS Protocol; Outlined in Fig. 1b) (See Note 5)

1. Transfer 1–50 ng of the enriched, linearized, and polyadenylated circRNAs to a 0.2 mL PCR tube. Adjust the volume to 9 μL with nuclease-free water (see Note 6). Mix by flicking the tube to avoid unwanted shearing and spin down. 2. Add 1 μL 2 μM VNP primer. 3. Add 1 μL of 10 mM dNTPs (2.5 mM each). 4. Mix the tube by flicking and collect the sample by brief spin down. 5. Incubate for 5 min at 65 °C and snap-cool on a pre-chilled freezer block. 6. Prepare 8 μL of strand-switching buffer reaction by mixing 4 μL 5× Superscript IV RT buffer, 1 μL of RNaseOUT or RiboLock RNase inhibitor, 1 μL of 100 mM DTT, and 2 μL of 10 μM Strand-Switching Primer (SSP,). Mix gently by flicking the tube, and spin down.

Nanopore-Mediated Sequencing of Circular RNA

149

7. Add the strand-switching buffer reaction (from step 6) to the snap-cooled and annealed primer-mRNA reaction (from step 5). 8. Mix the total of 19 μL reaction by flicking the tube and spin down. 9. Incubate at 42 °C for 2 min. 10. Gently add 1 μL of Superscript IV reverse transcriptase, mix by flicking and quickly spin down and return to the PCR machine. 11. Incubate for reverse transcription and template-switching for 10 min at 50 °C and 10 min at 42 °C, followed by 10 min heat inactivation at 80 °C and hold at 4 °C. 12. To perform PCR amplification of the generated cDNA, prepare four PCR tubes, each containing only 5 μL of the synthesized cDNA at the previous step, 25 μL 2× LongAmp HS Taq Master Mix, 3 μL cDNA primer (cPRM), and 17 μL nucleasefree water, ending up with the final volume of 50 μL per each of four PCR tubes per sample. 13. PCR cDNA amplification should be applied as 1 cycle at 95 °C for 30 s, 18 cycles of incubation at 95 °C for 15 s, 62 °C for 15 s, and 65 °C for 6 min, and incubate for final extension for 6 min at 65 °C then cooling down and hold at 4 °C. 14. Incubate each of the 50 μL PCR reactions separately with 1 μL of Exonuclease I for 15 min at 37 °C and heat inactivate the enzyme by 20 min incubation at 80 °C. 15. To avoid very short DNA fragments, run the amplified double stranded DNA on a 2% TBE agarose gel for 2 h at 80 V. 16. Excise DNA fragments in the range of 350 bp to 10 kb and purify using GeneJet Gel Extraction Kit. 17. To generate enough input DNA for ONT MinION sequencing, use the gel purified library as template and perform 6 cycles of PCR enrichment as described in steps 12 and 13, except only use 2 ul cPRM ONT primer mix at this step. 18. Use 5 μL of the PCR product and analyze in a 2% TBE agarose gel to check the quality of the re-amplified library. 19. Pool the 200 μL PCR product within a 1.5 mL Eppendorf DNA LoBind tube and use SPRISelect beads for library cleanup. 20. Add 160 μL (0.8X) of well mixed SPRISelect beads to each sample and mix by ten times pipetting and flicking and follow by 5 min rotation at room temperature. 21. Pellet the beads on a magnetic rack for 5 min and wash twice using 500 μL freshly prepared 80% Ethanol without disturbing the pellets.

150

Morten Trillingsgaard Venø et al.

22. Spin down the samples briefly and place back on the magnetic rack. Remove all residual ethanol. Allow the pellet to dry for 1 min before eluting the library using 21 μL of Rapid Annealing Buffer (RAB) and following by 10 min rotation at room temperature. 23. To ligate the sequencing adapter, use 400 fmol of the eluted library and adjust to the final volume of 23 μL in RAB buffer and add 2 μL cDNA adapter mix (cAMX) and mix by brief pipetting. Rotate the adapter ligation mix for 10 min at room temperature. 24. To clean up the adapter ligated library and deplete unligated adapters, add 20 μL of SPRISelect beads and mix by pipetting ten times and rotate for 5 min at room temperature. 25. Pellet the beads by 2 min incubation on magnetic rack. 26. Wash the pellets twice using 140 μL of Adapter Bead Binding Buffer (ABB). 27. Elute the final and ready to sequence library using 13 μL of elution buffer (ELB) by rotating 10 min at room temperature. 28. Of this, use 1 μL for quantification using Qubit, and mix 12 μL library with 35 μL of Running Buffer with Fuel mix (RBF), 25.5 μL of Library Loading Beads (LBB) and 2.5 μL of nuclease-free water for loading on the flowcell. 29. Using MinION or GridION sequencer, check the quality of the flow cell and load the priming mix according to the ONT instruction (see Note 7). 30. Load 75 μL prepared library from step 28 into the SpotON port in a dropwise fashion. Ensure each drop flows into the port before adding the next. 31. Closing the SpotON and priming port, starting the sequencing. 32. Optionally, monitor the status of the sequencing and quality of the generated reads by enabling real-time basecalling in the minKNOW GUI sequencing software. In case of need, pause the sequencing and refuel the flow cell according to the ONT instruction (see Note 8). 33. For data analysis see Note 18. 3.3 Full-Length Sequencing of a Targeted Panel of CircRNAs (circPanelLRS) 3.3.1 CircRNA Specific Primer (CSP) Design

To make the first strand of the cDNA in circPanel-LRS, a mixed pool of CircRNA Specific Primers (CSP) can be used. A pool of 1–100 primers can be used and should target 30 nt downstream of the BSJ or downstream of the 5′ end of a conserved exon among different isoforms obtained from the same circRNA host gene. The second strand of the cDNA is synthesized using a matching pool of back-to-back and divergent primers (Fig. 2). Both primer types need to be extended at the 5′ end to be compatible with the

Nanopore-Mediated Sequencing of Circular RNA

151

Fig. 2 Targeting a selected panel of circRNAs for ONT long-read sequencing. (a) An overview of targeting, sample indexing and amplification of a selected panel circRNAs using pairs of divergent primers for each circRNA. (b) First strand of the cDNA is driven by CSP forward primer complementary to circRNA (blue) extended with ssDNA sequences at the 5′ end that has overlapping sequence with ONT-cPRM forward primer (red). Second strand cDNA is made using CSP reverse primer which is extended at the 5′ end with ssDNA sequences partially overlapped with ONT-cPRM reverse primer (red). The first round of PCR amplification is using the CSP forward and reverse primers. The second round of PCR amplification uses only ONT-cPRM forward and reverse primers from SQK-PCS/PCB109 kits with or without sample indexing respectively. Finally, the amplified library or pooled barcoded libraries are ready to be loaded and sequenced on MinION or GridION sequencer. Abbreviations: CircRNA Specific Primers (CSP); cDNA primer (cPRM)

ONT cDNA-PCR SQK-PCS/PCB109 kit as highlighted in red (Fig. 2b). The extended sequence at the 5′ end of the cDNA synthesis primer (CSP forward primer) is 5′ ACTTGCCTGTCG CTCTATCTTC , and at the 5′ end of the reverse primer (CSP reverse primer) is 5′ TTTCTGTTGGTGCTGATATTGC (see Note 9). 3.3.2 Full-Length CircRNA Sequencing

1. To perform the synthesis of the first cDNA strand using multiple preselected circRNAs as input, add 100 ng DNase I-treated total RNA (see Note 2), 1 μL primer mix (1 pmol RT primer for each circRNA; see Subheading 3.3.1. for design), 1 μL 10 mM dNTPs (2.5 mM each), and nuclease-free water up to a total of 14 μL in a 0.2 mL PCR tube and separately for each sample. Mix the tubes by flicking to avoid unwanted shearing of RNA and spin down briefly in a microfuge. 2. Incubate the RNA-Primer mix for 3 min at 70 °C in a PCR machine with hot lid set to 105 °C. 3. Snap cool the sample on ice. Add 4 μL 5× RT Buffer (Maxima H Minus Reverse Transcriptase) and 1 μL of RNaseOUT or RiboLock RNase inhibitor and mix by gently flicking the tube and spin down (see Note 10).

152

Morten Trillingsgaard Venø et al.

Table 1 PCR program used in 3.3.2, step 12 Cycle no.

Denature

Anneal

Extend

Hold

1

95 °C, 2 min

50 °C, 1 min

65 °C, 10 min

–

2–3

95 °C, 15 s

55 °C, 15 s

65 °C, 5 min

–

4–9

95 °C, 15 s

62 °C, 15 s

65 °C, 5 min

–

10

–

–

65 °C, 5 min

–

11

–

–

–

4 °C

Table 2 PCR program used in 3.3.2, step 27 Cycle no.

Denature

Anneal

Extend

Hold

1

95 °C, 2 min

–

–

–

2–17

95 °C, 15 sec

62 °C, 15 sec

65 °C, 5 min

–

18

–

–

65 °C, 5 min

–

19

–

–

–

4 °C

4. Incubate for 2 min at 42 °C. 5. Dilute Maxima H Minus RT enzyme to 100 U and gently add 1 μL (see Note 11). 6. Gently mix by flicking, briefly spin down and return to the PCR machine. 7. Incubate for 30 min at 50 °C (pre-warmed) for first strand cDNA synthesis, followed by 5 min at 85 °C and hold at 4 °C. 8. Add 1 μL RNase H and 1 μL RNase cocktail, mix by flicking and collect by spinning down. 9. Incubate for 20 min at 37 °C, followed by 20 min at 75 °C for RNase enzymes inactivation and hold at 4 °C. 10. To synthesize the second strand of the cDNA, add the following reagents to the final 22 μL cDNA generated in the previous steps: 50 μL 2× LongAmp Taq HS Master Mix, 1 μL 10 pmol second strand primer pool, 1 μL primer mix (1 pmol reverse primer for each circRNA), 27 μL nuclease free water (up to 100 μL) (see Note 12). 11. Mix the PCR reaction well by 5X pipetting and divide each sample into 2× 50 μL in 0.2 mL PCR tubes. 12. Return the 50 μL samples prepared in the previous step to the PCR machine and run the PCR with the program showed in Table 1) (see Note 13).

Nanopore-Mediated Sequencing of Circular RNA

153

13. Treat each PCR reaction tube with 1 μL of Exonuclease I by 15 min incubation at 37 °C and heat inactivate by 15 min incubation at 80 °C. 14. Cleanup the PCR products from previous steps using SPRIselect beads to remove all unused and undigested primers to not interfere with the second round of PCR amplification and sample indexing. 15. Pool both PCR product tubes of each sample of a total of 100 μL into a 1.5 mL LoBind Eppendorf tube and add 65 μL well mixed SPRISelect beads (0.65X) to each tube and mix by ten times pipetting and gentle flicking. 16. Rotate the samples for 10 min at room temperature. 17. After a gentle spin, pellet the beads on a magnetic rack for about 1–2 min and discard the supernatant after it appears clear. 18. Wash the pellet twice using 500 μL freshly prepared 80% ethanol without disturbing the pellets. 19. After a brief spin, remove any remaining ethanol from the tube. 20. Dry the pellet for maximum of 2 min and 30 s while kept on the magnetic rack (see Note 14). 21. Remove the tubes from magnetic rack and resuspend the pellet in 25 μL nuclease free water by pipetting. 22. Incubate the samples for 5 min at 37 °C and follow by 5 min rotation at room temperature. Incubation at 37 °C facilitates release of the long DNA oligos from the beads. 23. Pellet the beads by 5 min incubation on a magnetic rack at room temperature or until the eluate appears clear and colorless. 24. Remove and retain the 23 μL eluate and dispose the pellet. 25. For sample quality control, at this step, measure the sample concentration by Qubit, check the quality on bioanalyzer and check the purity of the sample, using nanodrop. Use 1 μL of the eluate obtained from previous step for each quality control. 26. Re-amplification the enriched and cleaned up circRNAs cDNA library is performed in 4 parallel reactions in PCR tubes each containing: 2 ng PCR product (obtained at step 24) and diluted in 10 μL nuclease free water, 25 μL 2× LongAmp Taq HS Master Mix, 1.5 μL cPRM primer provided in SQK-PCB109 kit, and 13.5 μL nuclease free water up to a total volume of 50 μL. The primers are tagged at the 5′ ends, which facilitate the ligase-free attachment of the rapid sequencing adapter in the next step. Additionally, each primer pair contains a barcode for sample multiplexing (see Note 15).

154

Morten Trillingsgaard Venø et al.

27. Mix by gently flicking the tube, spin by microfuge and start the PCR on a preheated PCR machine with the program showed in Table 2 (see Note 16). 28. Add 1 μL of Exonuclease I per PCR tube and incubate for 15 min at 37 °C and heat inactivate by 15 min incubation at 80 °C. 29. The four PCR tubes from each sample are pooled and cleaned up for sequencing (see Note 17). 30. Add 130 μL of 0.65 X pre-mixed SPRISelect beads at room temperature to the 200 μL pooled PCR products per sample in 1.5 mL low-bind Eppendorf tube and mix by 10 times pipetting followed by 10 min rotation at room temperature. 31. Cleanup and eluate the library in 15 μL elution buffer (EB, provided by ONT SQK-PCB/PCS109 kit) as explained at steps 15–24. 32. For the final library quality control, use 1 μL of the eluate for quantification using Qubit and 1 μL for qualification using Nanodrop. If needed, 1 μL can also be checked by bioanalyzer using the DNA High Sensitivity kit. 33. Mix 200 ng (equal to ~100 fmol) of all the separate cleaned-up libraries in a total of 12 μL EB buffer. 34. Add 1 μL of Rapid Adapter (RAP) to the pooled libraries in the previous step. 35. Mix by pipetting and spin down using microfuge. 36. Incubate for 5 min at room temperature and keep on ice until the final library mix for loading into the SponON port of the flow cell is ready. 37. Using MinION or GridION sequencer, quality check the flowcell and prepare and load both priming mix and the library (12 μL) exactly as instructed by ONT for the SQK-PCB109 kit (see Note 7). 38. Start the sequencing using MinKnow software and enable the real time basecalling (see Note 8). 39. For data analysis see Note 18.

4

Notes 1. Carry out all the procedures at room temperature unless otherwise specified. 2. Total RNA from tissue can be prepared using Trizol and following the manufacture instruction. To prepare total RNA from cultured cells we usually prefer to use Zymo direct-zol RNA miniprep.

Nanopore-Mediated Sequencing of Circular RNA

155

3. In order to keep track of circRNA abundance, it is recommended to quantify selected circRNA by qPCR, and compare to housekeeping linear RNA and 18S ribosomal RNA species after each step of circRNA enrichment. Use Superscript VILO cDNA Synthesis Kit according to the manual. Use the LightCycler 480 SYBR Green I Master Kit for the qPCR reactions. 4. Some linear RNAs are resistant to digestion by RNase R presumably due to their structure. The processivity of the enzyme can be enhanced by addition of LiCl [9]. 5. A modified version of the protocol SQK-PCS108 provided by Oxford Nanopore was used for library preparation. 6. In case a lower amount of enriched circRNA was obtained or the concentration was out of range to be measured by Qubit, the presence of the enriched circRNA pool should be confirmed by qPCR. To do this, 1 μL of the Enriched CircRNA Pool (ECP) can be used for cDNA synthesis using random primer and followed by qPCR assay for some of the known circRNAs. 7. Before loading the sample, the flow cell needs to be checked by performing a Flow Cell Check using MinKNOW (the MinION and GridION device software). The Flow Cell Check will report the number of nanopores available for sequencing. If the flow cell has less than 800 active pores, it means the flow cell has quality issue and you should contact Oxford Nanopore. 8. Nanopore basecalling is the process to decode the characteristic electrical signal of the molecules using basecalling algorithms to determine the DNA or RNA sequence. By enabling realtime basecalling, the status of the sequencing and quality of the generated reads can be monitored. If a drop in translocation speed occurs, pause the sequencing and refuel the flowcell according to the ONT-provided manual. For future versions of the cDNA-PCR sequencing kits, the oligo sequences, PCR steps, and final library preparation and sequencing can be adjusted accordingly. 9. We have optimized the protocol for primers of about 25 nt length and more than 40% GC equivalent to more than 60 degrees melting temperature. Since the second strand of the cDNA and the following PCR steps are optimized for LongAmp HS Taq 2X Master Mix, we suggest to use the online Tm calculator provided by NEB for the corresponding Taq polymerase (https://tmcalculator.neb.com/#!/main). Desalted primers can be used and there is no need to order highly pure HPLC grade primers. DO NOT use the VNP primer provided by ONT kit. 10. DO NOT use the Strand-Switching Primer (SSP) provided by ONT kit.

156

Morten Trillingsgaard Venø et al.

11. It is recommended to use less enzyme for generating longer cDNA. 12. DO NOT use the VNP primer provided by ONT kit. 13. In some rare cases, if the targeted circRNAs are expected to be more than 5 kb long, apply more extension time of 1 min per extra kb length. 14. Do not allow the pellet to reach the point of cracking. 15. Even if only one sample is going to be sequenced, we suggest to use the barcoded primers from SQK-PCB109 kit. The advantage of using barcoded primers is that the flowcell can be easily re-used for multiple sequencing rounds for different samples if different barcoded primers are used (to reduce crosscontamination). We are aware that DNase treatment reagents provided in the flowcell washing protocol will remove the library from a previous run; however, in our experience, this is not 100% efficient. 16. If the amount of template used per tube is less than 2 ng, increase the cycles to 18 or more. In special cases, where the targeted circRNAs are expected to be more than 5 kb long, apply more extension time of 1 min per extra kb length. 17. In case of multiplexing four or more samples, the material from two tubes is usually enough for loading into one flowcell. Therefore, at this step it might be decided to only clean up two tubes per sample and store the rest at -20 °C. 18. All data analysis details can be found in our previously published work (PMID: 34376658). The codes generated for circPanel-LRS data analysis are deposited on Github (https:// github.com/omiics-dk/long_read_circRNA). The raw longread sequencing data are deposited to the Gene Expression Omnibus (GEO) repository database with the accession number GSE127059.

Acknowledgments We acknowledge the support of this project by the Danish Council for Independent Research and Villum Foundation. We thank Karim Rahimi for valuable input to the protocol. References 1. Rahimi K, Venø MT, Dupont DM et al (2021) Nanopore sequencing of brain-derived fulllength circRNAs reveals circRNA-specific exon usage, intron retention and microexons. Nat Commun 12(1):4825

2. Yang L, Wilusz JE, Chen LL (2022) Biogenesis and regulatory roles of circular RNAs. Annu Rev Cell Dev Biol 38:263–289 3. Wang Y, Zhao Y, Bollas A et al (2021) Nanopore sequencing technology, bioinformatics and

Nanopore-Mediated Sequencing of Circular RNA applications. Nat Biotechnol 39(11): 1348–1365 4. Amarasinghe SL, Su S, Dong X et al (2020) Opportunities and challenges in long-read sequencing data analysis. Genome Biol 21(1):30 5. Nielsen AF, Bindereif A, Bozzoni I et al (2022) Best practice standards for circular RNA research. Nat Methods 19(10):1208–1220 6. Xin R, Gao Y, Gao Y et al (2021) isoCirc catalogs full-length circular RNA isoforms in human transcriptomes. Nat Commun 12(1):266

157

7. Zhang J, Hou L, Zuo Z et al (2021) Comprehensive profiling of circular RNAs with nanopore sequencing and CIRI-long. Nat Biotechnol 39:836–845 8. Rahimi K, Nielsen AF, Venø MT et al (2021) Nanopore long-read sequencing of circRNAs. Methods (San Diego, Calif) 196:23–29 9. Xiao M-S, Wilusz JA (2019) An improved method for circular RNA purification using RNase R that efficiently removes linear RNAs containing G-quadruplexes or structured 3′ ends. Nucleic Acids Research 47:8755–8769

Part IV Functional Study: Tools for Loss- and Gain-of Function Analysis

Chapter 9 In Vivo Tissue-Specific Knockdown of circRNAs Using shRNAs in Drosophila melanogaster Ines Lucia Patop, Michael Canori, and Sebastian Kadener Abstract Studying circular RNAs’ function in vivo has been challenging due to the lack of generic tools to manipulate their levels without affecting their linear counterparts. This is particularly challenging as the back-splice junction is the only sequence not shared between the linear and circular version. In this chapter, we describe a method to study circRNA function in vivo targeting shRNAs against the desired back-splice junction to achieve knockdown with tissue-specific resolution in flies. Key words Circular RNAs, CircRNA, shRNA, Drosophila melanogaster, Off-targets, In vivo knockdown

1 Introduction Circular RNAs (circRNAs) are stable, non-polyadenylated, covalently closed RNA molecules. CircRNAs are generated through back-splicing (see [1, 2], for review). They are mainly formed by exons of coding genes; and as such, they share most of their sequence identity with the linear RNA expressed from the host locus. However, the exception to this, is their unique back-splice junction. Because of their similarity to their linear counterparts, they are difficult to study and manipulate, and hence have been understudied until recent years. In recent years, significant advancements in computational and biochemical techniques have substantially expanded our understanding of these novel molecules. Multiple studies showed that circRNAs are conserved across the tree of life [3–9], that they can bind miRNAs [10, 11] and RNA-binding proteins [12–16], that are enriched in synaptoneurosomes [17], that are tissue-specific

Ines Lucia Patop and Michael Canori contributed equally. Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_9, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

161

162

Ines Lucia Patop et al.

[17–19], and that accumulate with age in the brain of flies, mice, humans and worms [17–20]. A few studies addressed the function of circRNAs in vivo by manipulating their levels with different approaches. Two independent studies used CRISPR-Cas9 approaches to fully remove the locus that generates the most abundant circRNA in mammals: CRD1as [21, 22]. They showed that in mice, CDR1as has functions in the brain [22] and forms part of a regulatory network with miRNAs miR-7, miR-671, and the long non-coding RNA cyrano [21]. However, CDR1as is not a typical circRNA, as there is no linear RNA generated from the locus. So, for most circRNAs, approaches other than removing the whole locus need to be used. These mainly include removing the signals responsible for exon circularization (a laborious and difficult approach to use in a global screen) [23–25] or targeting circRNAs after they are generated using short-hairpin RNAs (shRNAs) [26] or the Cas13 system [27]. Any approach that targets specific circRNA sequences has one shared challenge: the fact that most sequences are shared between the linear and circular RNA isoforms of a given locus. The only targetable sequence for a circRNA is, therefore, the back-splice junction. Nevertheless, there is a potential for an shRNA designed to target a specific circRNA to also affect the linear mRNA. This can occur either due to the limited complementarity being sufficient to impact mRNA levels or because of the presence of unidentified splicing variants that may be influenced by the expression of the shRNA. In addition, the shRNA could have other off-targets. In this work, we describe a technique in Drosophila to knockdown circRNAs with tissue-specific resolution using shRNAs designed against the circRNA back-splice junction. To knockdown circRNAs, we designed shRNAs that match perfectly the circRNA splice junction (Fig. 1(1)). To this end, we developed scripts that take any circular or linear splice junction and generate the oligonucleotides with the shRNA sequence in the context of the miR-1 backbone with the overhangs for cloning in pVALIUM20 plasmid. To speed up the process and reduce possible errors, we developed an R-package with all the required functions for this design: https://github.com/ipatop/shRNA_design. We use the VALIUM20 vector containing a vermillion marker that is used for easy selection of transgenic flies and the miR-1 backbone for shRNA expression under UAS promoter [28, 29]. By crossing flies expressing this construct with GAL4 flies, we can then control the expression of the shRNA in a tissue- and temporalspecific fashion [30]. To determine the effects of knocking down circRNAs in vivo, we perform qRT-PCR to confirm both the knockdown of the target circRNA and that expression of the linear mRNA counterpart remains unchanged. We also provide

In Vivo Knockdown of circRNAs Using shRNAs

163

circRNA KD3 (3)

CGUUUAUUGCGGUCCUCUUGCU UUGCGGUCCUCUUGCUGUUGU

circRNA KD2 (2) circRNA KD1 (1)

UUAUUGCGGUCCUCUUGCUGU ......GCAAAUAACGCCAG GAGAACGACAACA......

Back-splice junction circRNA UUAUUGCXXXX XXXXUGCUGU

8 mismatch control (4)

Fig. 1 An shRNA strategy for in vivo knockdown of circular RNAs. Circular RNAs are generated through backsplicing and share most of their sequence with their linear counterparts. To knock down circRNAs without affecting the host linear RNA, we employ shRNAs designed to specifically target the circRNA back-splicing junction (1). Additionally, to ensure the exclusion of off-target effects, we generate 3’ and 5’ shifted versions of the shRNA (2 and 3), ensuring circRNA knockdown while targeting different putative off-targets. As a control, we utilize an shRNA with 8 sequence mismatches (4). This enables us to establish an isogenic line without any knockdown but with an active shRNA pathway

bioinformatic and experimental approaches to detect and address possible off-target effects [26]. Bioinformatically, we determine globally two types of potential off-targets. Firstly, transcripts that have long stretches of complementarity with the shRNA can be potentially targeted by it. The levels of these putative off-targets can be assessed by qPCR and/or RNA sequencing. To tackle this, we create a list of all possible targets of the shRNA due to extensive base complementarity to other RNAs by using the BLAST algorithm for short sequences against the Drosophila transcriptome. Secondly, shRNAs can be potentially loaded into AGO1 and act as miRNAs [29]. We assess this by generating and sequencing libraries of flies in which the shRNA is expressed. We then inspect the 3’UTR regions of the downregulated mRNAs and determine if they are enriched of the potential seed of the shRNA or shRNA with the Sylamer algorithm. To further rule out off-target effects, we recommend generating additional fly strains expressing shRNAs which are still perfectly complementary to the circRNA junction, but in which the shRNA is shifted 3 bases in the 5′ or 3′ directions (Fig. 1(2 and 3)). This should still knock down the targeted circRNA but with a different population of putative off-targets. The R-package described above automatically designs these shRNA shifts. Moreover, this R-package can also be used to generate a mismatch control shRNA. This is an shRNA where the central eight bases around the splice junction are shuffled to ensure no endogenous transcript is targeted (Fig. 1(4)). Using a mismatch control as this one is recommended because while the circRNA is not targeted anymore, the shRNA pathway is active.

164

2

Ines Lucia Patop et al.

Materials

2.1 Design shRNA and 5’/3’ Shifts

1. Coordinates of the splice junction to target. 2. Genomic sequence. 3. shRNA design R package https://github.com/ipatop/ shRNA_design.

2.2

Cloning of shRNA

1. Custom-designed oligonucleotides.

top

and

bottom-strand

short-hairpin

2. The pVALIUM20 plasmid (https://dgrc.bio.indiana.edu/ Home). 3. EcoR1 and Nhe1 restriction enzymes with buffers. 4. 1.2× annealing buffer: 12 mM Tris–HCl pH 7.5, 120 mM NaCl, 1.2 mM EDTA. 5. T7 DNA Ligase with 2× ligation buffer (NEB). 6. Competent E. coli cells. 7. LB agar plates with 100 μg/mL ampicillin. 8. LB broth with 100 μg/mL ampicillin. 9. 1.2% Agarose gel. 10. TAE (Tris-acetate-EDTA) buffer: 40 mM Tris–HCl, 20 mM Acetate, 1 mM EDTA. 11. Gel purification kit. 12. Plasmid midi-prep kit. 13. Forward PCR primer: ACCAGCAACCAAGTAAATCAAC. 14. Reverse PCR primer: TAATCGTGTGTGATGCCTACC. 2.3 Expression of the shRNA

1. Transgenic fly stocks. 2. GAL4-driver fly stock. 3. Fly food. 4. Carbon dioxide with fly pad.

2.4

qPCR

1. NanoDrop machine. 2. TRIzol (invitrogen). 3. Chloroform. 4. 100% Isopropanol (molecular grade). 5. 75% Ethanol. 6. Nuclease-free water. 7. 5 mg/mL Glycogen. 8. DNase 1 (NEB). 9. 10× DNase 1 reaction buffer (NEB).

In Vivo Knockdown of circRNAs Using shRNAs

165

10. 3 M Sodium acetate, pH 5.2. 11. EDTA. 12. QuantiTect reverse transcription kit (QIAGEN). 13. QuantiTect SYBR green PCR kit (QIAGEN). 14. Nested sieves (710 μm and 425 μm mesh). 15. Fly homogenizer with pestles. 16. Dry ice. 17. Liquid nitrogen. 2.5 Identify Transcripts with Long Complementary Stretches to the shRNA

1. shRNA and shRNA* sequences.

2.6 3’UTR Sequence Enrichment Analysis

1. 3’UTR sequences in fasta format.

2. Online BLAST or script (https://github.com/ipatop/blast_ for_shRNA_design).

2. shRNA and shRNA* sequences. 3. Sorted list of differentially expressed genes in knockdown versus a mismatch control (https://github.com/ipatop/ SylamerTools). 4. Sylamer algorithm (https://github.com/micans/sylamer). 5. Script to plot Sylamer results (https://github.com/ipatop/ SylamerTools).

3

Methods

3.1 Design of shRNAs and 5’/3’ Shifts

Using our R package for shRNA design (https://github.com/ ipatop/shRNA_design): 1. Generate a tab-separated file with the coordinates of the circRNA junctions. 2. Extract the exact sequence of the circRNA junction using the file generated in step 1: (a) Take the circRNA back-splice junction coordinates. (b) Extract the sequence from the genomic reference. 3. Generate the 21-nucleotide shRNA and its reverse complement oligo ready to order for cloning: (a) Take 10 bases to one side and 11 to the other of the circRNA junction from step 2. (b) Construct the shRNA and shRNA* sequence. (c) Put the shRNA/shRNA* sequences in the context of the miR1 backbone with the required overhangs for Nhe1 and EcoR1 digestion for cloning into VALIUM20.

166

Ines Lucia Patop et al.

4. Output the final oligos in a tab-separated table: this script outputs two oligonucleotide sequences (top and bottom strands) that when annealed will create the duplexed molecules with Nhe1 and EcoR1 sites ready to clone into VALIUM vectors (see Notes 1 and 2). 3.2 Cloning the shRNA for circRNA Knockdown in Drosophila

1. Digest the pVALIUM20 plasmid with restriction enzymes EcoR1 and Nhe1. Purify the linearized plasmid on a 1.2% agarose gel using a gel extraction kit (expected band size ~7.6 kb). 2. Dilute the top- and bottom-strand oligonucleotides with nuclease-free water to make 20 mM stock solutions. 3. Add 10 μL of top-strand and 10 μL of bottom-strand oligonucleotides with 100 μL of 1.2× annealing buffer. 4. Anneal the oligos by incubating the solution at 95 °C for 5 min, and then allow the solution to come to room temperature slowly (see Note 3). 5. Ligate the annealed oligonucleotides with the linearized pVALIUM20 plasmid using the T7 ligase according to the manufacturer’s instructions and transform the ligated product into competent E. coli cells. 6. Streak the transformed E. coli on LB agar plates with 100 μg/ mL ampicillin and culture overnight at 37 °C. 7. Isolate individual colonies and screen each colony for the presence of a ~ 300 bp insert by PCR using the primers below. Then, sequence positive colonies using the forward primer (see Note 4). 8. Purify the sequence-confirmed construct using a midi-prep plasmid kit and perform injections into embryos to generate the transgenic UAS-shRNA Drosophila line (see Note 5).

3.3 Expression of the shRNA by Crossing the Transgenic UASshRNAs with GAL4Drivers

1. Collect female virgins from the GAL4-driver line and male flies from the transgenic UAS-shRNA line (see Note 6).

3.4 Verify the Specificity of the shRNA Knockdown of circRNAs by qPCR

For body parts that can be collected using sieves (heads and bodies) (see Note 9):

2. Place 8–10 GAL4-driver virgins in a bottle of fly food with an equal number of UAS-shRNA males (see Note 7). 3. Upon eclosion of the progeny, select flies against their marker for the correct genotype and place the flies in a new bottle of fly food (see Note 8).

1. Using a funnel, collect the flies in a conical tube on dry ice. Snap freeze the flies in liquid nitrogen and store frozen flies at -80 °C until RNA extraction.

In Vivo Knockdown of circRNAs Using shRNAs

167

2. Place a funnel and labeled 1.5 mL microfuge tubes on dry ice and the nested sieves at -80 °C for 30 min before starting. The use of these nested sieves will allow us to separate bodies, heads, and other appendages based on their size. Keep the frozen flies in liquid nitrogen until use. 3. Vortex the frozen flies for 10 s and rapidly shake up and down five to six times. Pour the frozen flies into the 710 μm mesh and vigorously shake back and forth. The bodies will remain on the top 710 μm mesh; the heads in the 425 μm mesh, and the appendages (such as legs or wings) will be in the sieve pan. Collect the body parts of your interest into the labeled 1.5 mL microfuge tube using a small funnel. 4. Add 100 μL of TRIzol to the body parts and place on dry ice for 3 min. Homogenize the sample well with the homogenizer and pestle, and then add an additional 100 μL of TRIzol and continue homogenizing. For body parts and tissues that cannot be collected by sieve (brain, fat body, etc.): 5. Dissect your tissue of interest and place immediately in a PCR tube with 200 μL Trizol (see Note 10). Once all the dissections are finished, homogenize the sample using a homogenizer and pestle. 6. Homogenize the sample well with the homogenizer and pestle, and then add an additional 100 μL of TRIzol and continue homogenizing. Continue with protocol: 7. Add 800 μL of TRIzol to the homogenized sample (for a total of 1 mL Trizol) and incubate at room temperature for 5 min. 8. Add 200 μL of chloroform to the sample and vortex sample for 30 s. Vortex two additional times and incubate the sample at room temperature for 10 min. 9. Centrifuge the sample at 12,000 × g for 15 min at 4 °C. Collect the aqueous phase (top) in a new 1.5 mL microfuge tube and add 1 μL of glycogen. 10. Add equal parts isopropanol to the aqueous phase and precipitate at -20 °C for a minimum of 3 h. 11. After precipitation, centrifuge the sample at 21,000 × g for 30 min at 4 °C, and discard the supernatant. 12. Add 200 μL of 75% ethanol and vortex the pellet loose. Centrifuge the sample at 21,000 × g for 10 min at 4 °C, and discard the supernatant. Allow the pellet to air-dry and then resuspend the pellet in nuclease-free water. Freeze the sample at -20 °C for 30 min.

168

Ines Lucia Patop et al.

13. Thaw the sample at 55 °C for 5 min and check RNA concentration and quality using a NanoDrop machine. 14. Perform DNase 1 treatment on 10 μg of RNA according to the manufacturer’s instructions. 15. To precipitate, add 10 μL of 3 M sodium acetate (pH 5.2), 1 μL of glycogen, and 100 μL of isopropanol to the sample and incubate at -20 °C for a minimum of 3 h. 16. Following precipitation, repeat steps 11–13. RNA can be stored at -20 °C until use. 17. Perform reverse transcription from 1 μg of RNA using the QuantiTect reverse transcription kit according to the manufacturer’s instructions. 18. Thaw QuantiTect SYBR green PCR kit and qPCR primers on ice (see Notes 11 and 12). For each sample prepare the master mix according to the manufacturer’s instructions. 19. Dispense the master mix into a PCR plate and then add the cDNA (30 nt do not provide further improvement in knockdown efficiency [16]. Notably, with the appropriate spacer length (i.e., a 26 nt spacer targeting the circRNA BSJ site), the overlapping sequence matching the cognate mRNA is as short as 13 nt, which, in theory, cannot affect the expression of the linear RNA [35, 39]. Together, these features make the RfxCas13d/BSJ-gRNA system ideal for functional circRNA screening. Here, we describe the procedure for the high-throughput circRNA functional screening, based on the RfxCas13d/BSJ-gRNA system. By constructing a gRNA library that targets sequences across BSJ sites of highly expressed human circRNAs and expressing it together with RfxCas13d in human cell lines, circRNAs required for cell proliferation can be identified [40]. We also applied this system in vivo, to screen circRNAs involved in mouse embryo preimplantation development, by simultaneously injecting RfxCas13d mRNA and BSJ-gRNA16. The detailed procedure is illustrated in Fig. 1. Theoretically, this system can be applied for circRNA screening in various contexts, either under natural conditions or with environmental stimulations, for example, virus

Fig. 1 The workflow for the functional circular RNA (circRNA) screening. After candidate circRNAs are chosen for screening, a guide RNA (gRNA) library is designed, constructed, and packaged into lentivirus. Simultaneously, a cell line stably expressing RfxCas13d is constructed. CircRNA screening is performed by infecting the gRNA library containing lentivirus into RfxCas13d cells with a multiplicity of infection (MOI) of 0.3 and selection for 30 days. Next, the genomic DNA is extracted, and gRNA sequences are amplified, indexed, and subjected to next-generation sequencing (NGS). CDCscreen is used to identify circRNAs involved in cell proliferation. The selected circRNAs are then individually knocked down and validated by calculating cell confluence and conducting MTT assays

176

Hao Wu and Ling-Ling Chen

infection or drug treatment. Notably, although we and others have found that this system has minimal or no off-target effects in circRNA screening using zebrafish embryos [41] or eukaryotic cell lines [16, 35, 42], a recent study reported off-target effects of this system in Drosophila cells [43]. In summary, the RfxCas13d/ BSJ-gRNA system can serve as a useful tool for studying functional circRNAs and their involvement in different physiological and pathological processes.

2

Materials

2.1 Biological Materials

The following biological materials are needed for this protocol. Please note that all animal experiments should be approved and performed according to the ethical guidelines of your institute, and the cell lines used in your research should be regularly checked to ensure that they are authentic and are not infected with mycoplasma. • Nude mice for in vivo screening. • 293FT cell line for packaging lentivirus, and cell lines used for functional circRNA screening (i.e., HT29 and HeLa). • Chemically competent cells and electrocompetent cells.

2.2

Reagents

The following reagents are required in this experiment, and gloves and masks should be worn before handling them. • Plasmids, including p23-NLS-RfxCas13d-mCherry-NLS-Flag, for generating stable cell lines expressing RfxCas13d; RfxCas13d-backbone, for gRNA library generation; psPAX2 and pMD2.G, for packaging lentivirus; and pcDNA3-GFP, for generating control lentivirus. • Pooled oligo library containing the gRNA sequences with HPLC purification quality (the way to design gRNA sequences is described in Subheading 3.1, step 1). • PrimeSTAR® Max DNA Polymerase. • Gibson Assembly Master Mix. • SYBR qPCR Mix. • Lipofectamine® 3000 Transfection Reagent. • CellTiter 96 Aqueous One Solution Cell Proliferation Assay. • BsmBI restriction enzyme. • Puromycin.

The Functional Circular RNA Screening via RfxCas13d/BSJ-gRNA System

2.3

Equipment

177

The following equipment is required in this study: • MicroPulser Electroporator. • Illumina NextSeq 500. • NanoDrop 2000. • Real-time PCR system. • PCR instrument. • Digital gel imaging system. • Flow cytometry system. • Cell counter. • Ensight® Multimode Plate Reader.

2.4 Buffer Preparations

Buffer recipes are listed below. Please note that, in this protocol, room temperature indicates a range from 20 °C to 25 °C. • 1× TAE buffer: 40 mM Tris–HCl, pH 8.0, 20 mM acetic acid, and 1 mM EDTA in Milli-Q water. Filter-sterilize using a 0.22 μm filter unit. • 1× TBE buffer: 89 mM Tris–HCl, pH 8.0, 89 mM boric acid, and 2 mM EDTA in Milli-Q water. Filter-sterilize using a 0.22 μm filter unit. • TE buffer: 10 mM Tris–HCl, pH 8.0, and 0.1 mM EDTA in Milli-Q water. • 293FT and HeLa growth medium: Add 50 mL FBS into 450 mL DMEM-high glucose. • HT29 growth medium: Add 50 mL FBS into 450 mL RPMI 1640. • 6% (vol/vol) native-polyacrylamide gel: For each gel, combine 2.4 mL Acryl/Bis 30% solution (29:1), 1.2 mL 10 × TBE buffer, 120 μL 10% (wt/vol) ammonium persulfate, and 12 μL TEMED, and add Milli-Q water to 12 mL.

2.5

3

Oligos

Methods

3.1 Construction of the BSJ-gRNA Library

1. Apply CIRCexplore2 [31], an integrative circRNA analysis program based on the CIRCpedia database, to identify circRNAs in cell lines of interest and obtain highly expressed circRNAs as targets, based on the FPBcirc value (a parameter indicating circRNA abundance) [44]. Design a gRNA library spanning the BSJ of each target circRNA by applying custom

178

Hao Wu and Ling-Ling Chen

Linux one-line commands in BEDtools and Perl. To achieve high and specific knockdown efficiency, the spacer length of the gRNA should be 22–30 nt. We recommend the design of seven BSJ site-targeting gRNAs per circRNA. Below is the command to design a 26 nt spacer gRNA spanning 13 nt on either side of the BSJ site (13 nt + 13 nt): bedtools getfasta – fi genome.fa – bed circRNAs.bed – fo circRNAs.fa – s – split perl – lane’ if ($.%2 eq 1) {$name = ’’$_;13 + 13’’} else{$x = substr $_,0,13;$y = substr $_,–13;print’’$name\t$y$x’’}’ circRNAs.fa

In addition to circRNA-targeting BSJ-gRNAs, several control gRNAs should be included in the gRNA library, including: (1) one control gRNA per circRNA, with half of its sequence replaced by a scrambled sequence (i.e., 13 nt matched +13 nt random), and (2) a set of unrelated gRNAs with no matching sequence. 2. Dissolve the pooled oligos containing the gRNA sequences in ultrapure water to approximately 50 ng/μL, ready for use. 3. Perform PCR amplification to clone the gRNA sequences from the oligo pool, using high-fidelity DNA polymerase. The recommended 50 μL reaction mixture contains 50 ng pooled oligo library template, 2.5 μL 10 μM Customarray F primer, 2.5 μL 10 μM Customarray R primer, and 25 μL 2× PrimeSTAR Mix. The PCR program is 98 °C for 5 min, 15 cycles of [98 °C for 10 s, 55 °C for 15 s, and 72 °C for 30 s], and 72 °C for 5 min (all primer sequences used are listed in Table 1). 4. Purify library fragments containing PCR product by slow electrophoresis on 2% (wt/vol) agarose gel (130 V, >30 min). Recover the purified DNA fragments with the gel extraction kit. Measure the concentration of purified DNA fragments using the Nanodrop 2000. Ideally, a total of 400–500 ng of gRNA library fragments is sufficient for the following experiments. 5. Linearize the gRNA-expressing backbone vector, RfxCas13d Backbone, used for cloning the pooled BSJ-gRNA library using the BsmBI enzyme and incubate the reaction at 55 °C for 1 h in a dry bath. Take particular care that: (1) the backbone vector contains a ccdB sequence, which expresses a toxic protein to kill the competent cells, and this ensures that all the colonies grown include the assembled BSJ-gRNA library vectors; and (2) there are two BsmBI restriction sites in the vector and a 1550 bp sequence is visible on the agarose gel, which should be excluded. The reaction mixture is then purified on

Sequence (5′–3′)

TATATATCTTGTGGAAAGGACGAAACACCGAACCCCTACCAACTGGTCGGGG TTTGAAAC

TAAAATTGTGGATGAATACTGCCATTTGTCTCAAGATCTAGTTACGCCAAGCTTAAAAAA

AATGGACTATCATATGCTTACCGTAACTTGAAAGTATTTCG

CTTTAGTTTGTATGTCTGTTGCTATTATGTCTACTATTCTTTCC

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCG ATCTGGCTTTATATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCG ATCTGCTTTATATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTCTTTATATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGA TCTTTTATATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTTTATATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTTATATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTATATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTTATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTATATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATC TTATCTTGTGGAAAGGACGAAACACCG

Oligo names

Customarray F

Customarray R

CRISPRi-Seq-F1

CRISPRi-Seq-R1

CRISPRi-Seq-F2-1

CRISPRi-Seq-F2-2

CRISPRi-Seq-F2-3

CRISPRi-Seq-F2-4

CRISPRi-Seq-F2-5

CRISPRi-Seq-F2-6

CRISPRi-Seq-F2-7

CRISPRi-Seq-F2-8

CRISPRi-Seq-F2-9

CRISPRi-Seq-F2-10

(continued)

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

First round of gRNA library amplification

First round of gRNA library amplification

Targeted gRNA library cloning

Targeted gRNA library cloning

Usage

Table 1 Primer sequences used to construct and amplify gRNA oligo libraries and generate next-generation sequencing (NGS) libraries (barcodes are marked in bold)

The Functional Circular RNA Screening via RfxCas13d/BSJ-gRNA System 179

Sequence (5′–3′)

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTATCTTGTGGAAAGGACGAAACACCG

AATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGAT CTTCTTGTGGAAAGGACGAAACACCG

CAAGCAGAAGACGGCATACGAGATCGTGATGTGACTGGAGTTCAGACGTGTGCT CTTCCGATCTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATACATCGGTGACTGGAGTTCAGACGTGTGCTCT TCCGATCGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATGCCTAAGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCTGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATTGGTCAGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATCACTGTGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCTATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATATTGGCGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCTTATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATGATCTGGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCATTATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATTCAAGTGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCTATTATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATCTGATCGTGACTGGAGTTCAGACGTGTGCTCTT CCGATCCTATTATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATAAGCTAGTGACTGGAGTTCAGACGTGTGCTCT TCCGATCGCTATTATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATGTAGCCGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCTGCTATTATGTCTACTATTCTTTCCCCTGCACTGT

CAAGCAGAAGACGGCATACGAGATTACAAGGTGACTGGAGTTCAGACGTGTGCTC TTCCGATCTTGCTATTATGTCTACTATTCTTTCCCCTGCACTGT

Oligo names

CRISPRi-Seq-F2-11

CRISPRi-Seq-F2-12

CRISPRi-Seq-R2-index1

CRISPRi-Seq-R2-index2

CRISPRi-Seq-R2-index3

CRISPRi-Seq-R2-index4

CRISPRi-Seq-R2-index5

CRISPRi-Seq-R2-index6

CRISPRi-Seq-R2-index7

CRISPRi-Seq-R2-index8

CRISPRi-Seq-R2-index9

CRISPRi-Seq-R2index10

CRISPRi-Seq-R2index11

CRISPRi-Seq-R2index12

Table 1 (continued)

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Second round of gRNA library amplification for NGS

Usage

180 Hao Wu and Ling-Ling Chen

The Functional Circular RNA Screening via RfxCas13d/BSJ-gRNA System

181

1% (wt/vol) agarose gel and recovered by a gel extraction kit. Record the concentration of the digested vector. 6. Insert the pooled oligo library fragments into the linearized RfxCas13d Backbone vector by Gibson Assembly. Up to 10 reactions are needed to ensure the final concentration of the gRNA library vector is >100 ng/μL. Purify the BSJ-gRNA library vector with ethanol and dissolve it in 20 μL ultrapure water. 7. Add 2 μL of 100 ng/μL assembled gRNA library vectors into 25 μL of electrocompetent cells on ice, and perform the transformation according to the manufacturer’s protocol. Next, add 1 mL recovery buffer to the cells and mix well, transfer the solution into a loosely capped 1.5 mL tube, and shake it at 250 rpm for 1 h at 37 °C. Up to six electroporations are needed for the subsequent experiments. 8. Take an appropriate amount of successfully electroporated cells and dilute 10-fold, 100-fold, and 1000-fold in recovery buffer. Plate 300 μL of each dilution of electroporated cells, as well as all the remaining transformation mixture, onto pre-warmed 15 cm Petri dishes (ampicillin) and culture at 32 °C for 14 h. Choose one of the dilution plates with a countable number of colonies, count the colonies, and estimate the total numbers of colonies on all plates. Proceed to the following experiments only if the number of colonies is ≥60-fold, compared with the number of gRNAs in the library. 9. Pipette 5 mL of LB onto each 15 cm petri dish and scrape colonies into a 50 mL tube to harvest colonies containing the gRNA library (see Note 1). Extract gRNA library plasmids from the colonies. 3.2 Quality Evaluation of the Assembled BSJ-gRNA Library

1. Use a two-round PCR program to enrich the gRNA sequences, which will help to minimize the errors induced by amplification. During the second round of PCR, add a universal adaptor at the 5′ end and a different barcode for each library at the 3′ end, to ensure that the library can be sequenced on a NextSeq instrument. To generate two biological replicates of the gRNA library for evaluation, two pairs of primers with different barcodes are used for amplification in the second round of PCR. The PCR conditions and programs are outlined below. All oligos used are listed in Table 1. 2. For the first round of PCR: The 50 μL reaction mixture contains 50 ng plasmid library, 2.5 μL 10 μM CRISPRi-seq-F1 primer, 2.5 μL 10 μM CRISPRi-seq-R1 primer, and 25 μL 2× PrimeSTAR Mix. The PCR program is 98 °C for 5 min, 15 cycles of [98 °C for 10 s, 55 °C for 15 s, and 72 °C for 30 s], and 72 °C for 5 min.

182

Hao Wu and Ling-Ling Chen

3. For the second round of PCR: The 50 μL reaction mixture contains 5 μL first PCR product, 2.5 μL 10 μM CRISPRi-seq-F2-(1–2) primer, 2.5 μL 10 μM 10 μM CRISPRi-seq-R2-index (1–2) primer (barcode), and 25 μL 2× PrimeSTAR Mix. The PCR program is 98 °C for 5 min, 15 cycles of [98 °C for 10 s, 55 °C for 15 s, and 72 °C for 30 s], and 72 °C for 5 min. 4. Separate the product of the second PCR by 6% native polyacrylamide gel electrophoresis (PAGE), collect the gel band corresponding to the library in a 1.5 mL tube, and crush using a 1 mL syringe in TE buffer. Release the library by shaking at 1400 rpm at 37 °C for 3 h, and then filter the solution using a Spin-X centrifuge tube at 15,000 g for 5 min at 4 °C. Finally, further purify the library with 2.5–3.0 volumes of ethanol. 5. Sequence the library using the Illumina NextSeq 500 system, and calculate gRNA enrichment. High representation and uniformity of gRNAs in the library are important for the successful screening of circRNAs. It is recommended that the number of reads per gRNA in the library is >100. An ideal gRNA library should have >70% perfectly matching gRNAs and < 0.5% undetected gRNAs. Plot the cumulative distribution of the number of reads per gRNA of constructed libraries using R (v 3.5.1). A good distribution of reads has 90th and 10th percentiles 500 cells containing each gRNA; for example, for a library with 10,000 gRNAs, at least five million cells should be harvested. • Continue to culture and passage the remaining cells for 30 days. Harvest an equal number of cells as collected on D1 and label the samples D30. B. For in vivo screening: • Obtain D1 samples as described in in cell screening protocol above and store them at -80 °C. • Continue to culture and passage an aliquot of infected cells for 7 days (D7) to achieve maximal knockdown efficiency and rule out harmful effects of the virus on cell growth. • Transfer the same number of D7 cells as in D1 samples into nude mice. Obtain xenografts 22 days after injection as D30 samples. Take an aliquot of D1 and D30 samples, and isolate RNA using Trizol Reagent. Select three different target circRNAs for RT-qPCR detection, to ensure that RfxCas13d/BSJ-gRNA system-mediated knockdown was successful.

The Functional Circular RNA Screening via RfxCas13d/BSJ-gRNA System

3.5 Next-Generation Sequencing (NGS) Library Preparation and Deep Sequencing

185

1. Extract genomic DNA from the remaining D1 and D30 samples, and ensure that cells are completely lysed (see Note 4). 2. Conduct two-round PCR to amplify gRNA sequences from genomic DNA, as described in Subheading 3.2. The detailed reaction mixture and programs for in-cell screening are outlined below, and the primers used are listed in Table 1. 3. For the first round of PCR: The 100 μL reaction mixture contains 10 μg genomic DNA, 5 μL 10 μM CRISPRi-seq-F1 primer, 5 μL 10 μM CRISPRi-seq-R1 primer, and 50 μL 2× PrimeSTAR Mix. The PCR reaction is 98 °C for 5 min, 15 cycles of [98 °C for 10 s, 55 °C for 15 s, and 72 °C for 30 s], and 72 °C for 5 min. 4. For the second round of PCR: The 100 μL reaction mixture contains 5 μL first PCR product, 2.5 μL 10 μM CRISPRi-seq-F2-(1–12) primer, 2.5 μL 10 μM CRISPRi-seq-R2-index (1–12) primer (barcode), and 50 μL 2× PrimeSTAR Mix. The PCR reaction is 98 °C for 5 min, 15 cycles of [98 °C for 10 s, 55 °C for 15 s, and 72 °C for 30 s], and 72 °C for 5 min. The condition and program for in vivo screening are the same as in cell screening, but to increase the reaction mixture volume to 300 μL. 5. Purify the product of the second round PCR by 6% (vol/vol) native PAGE separation, according to the methods described in Subheading 3.2. Perform deep sequencing of the purified PCR products containing gRNA sequences using the Illumina NextSeq 500 system, according to the Illumina user manual.

3.6 Data Analysis of Screening Results via CDCscreen

Here we use HT29 cell datasets, which have been deposited in NCBI GEO (GSE149692), as an example to illustrate deep sequencing data analysis. 1. Index total gRNA library (ref_gRNA_seq.fa) reference sequences with Bowtie (v1.1.2). Remove the 3′ end and 5′ end BSJ-gRNA adapter sequences from raw paired-end sequencing datasets using cutadapt (v1.16), using the following adapter sequences:

R1 : –a TTTTTTAAGCTTGGCGTAACTAGATCT–m 15 –g CCCTACCAACTGGTCGGGGTTTGAAAC–m 15 R2 : –a GTTTCAAACCCCGACCAGTTGGTAGGG–m 15 –g AGATCTAGTTACGCCAAGCTTAAAAAA–m 15 2. Use Bowtie (v1.1.2, parameters: -v 3 -m 1 -k 1) to align the gRNA library sequences to the reference gRNA library

186

Hao Wu and Ling-Ling Chen

sequences and obtain gRNA sequences enriched for uniquely mapped reads. Normalize the reads of each gRNA to the total number of reads for all gRNAs in the samples, according to the formula: Normalized reads per gRNA =

Reads per gRNA × 105 þ 1 Total reads for all gRNAs in samples

3. Analyze the correlation between biological replicates by calculating the Pearson correlation coefficient, based on normalized reads per gRNA. Use the mean normalized reads per gRNA value from the biological replicates for subsequent analyses. 4. Use MAGeCK (v0.5.9.2) [45] to calculate permutation test Pvalues of RfxCas13d/BSJ-gRNA-targeted circRNAs. Obtain the “mean gRNA fold-change (FC)” from normalized reads of negatively selected gRNAs targeting the same circRNAs between D1 and D30 treatments. Finally, evaluate the effects of targeted circRNAs by calculating the CDCscreen score using the following equation: CDCscreen score = scale – log 10 ðP value Þ þ scale ½jlog 2 ðMean of gRNA fold change Þj The automated CDCscreen computational pipeline, used to identify negatively selected functional circRNAs, is available at https://github.com/YangLab/CDCscreen. Meanwhile, calculate the false discovery rate (FDR) for each circRNA, to improve the reliability of the results, as reported previously [25]. 5. Identify candidate circRNAs based on CDCscreen score, FC, and FDR values. We recommend selection of candidates with CDCscreen score ≥ 2, ≥ 2 selected gRNAs with FC ≤ 0.667 or ≥ 1.5, and FDR for each circRNA 70% perfectly matching gRNAs, < 0.5% undetected gRNAs, and a good distribution of reads. Good-quality gRNA libraries were packaged in lentivirus and used for further screening. 2. Data analysis for functional circRNA screening: Another dataset was used as an example of a transcriptome-scale knockdown screen targeting circRNAs involved in cell proliferation in different cell lines [16]. After screening, the distribution of gRNAs was measured by NGS. Some gRNAs targeting circRNAs were enriched or depleted in the experimental, relative to control, groups (Fig. 3a) [16, 40]. gRNA depletion indicates that the targeted circRNA promotes cell proliferation, while gRNA enrichment indicates the opposite function for the circRNA. CDCscreen analysis of depleted gRNAs that target circRNAs relative to controls was performed as described in Subheading 3.6; the computational pipeline is outlined in Fig. 3b [16, 40]. Several candidate circRNAs were identified that are required for cell proliferation (Fig. 3c) [16, 40] and were further validated by MTT assay and assessing cell confluence.

188

Hao Wu and Ling-Ling Chen

Fig. 2 The guide RNA (gRNA) library design and preparation for screening circular RNAs (circRNAs) involved in colorectal cancer growth. (a) The strategy for selecting candidate circRNAs [16, 40]. (b) The schematic of back-splicing junction (BSJ)-gRNA library design. All gRNAs are designed to target sequences spanning BSJ sites, to discriminate circRNAs from cognate mRNAs. Control gRNAs, with half of their sequences replaced by the random scrambled sequences or with all random sequences, serve as negative controls. (c) The alignment of deep sequencing reads from two PCR replicates of the BSJ-gRNA library. (d) Composition of deep sequencing reads from the two library replicates. Reads aligned to corresponding gRNA reference sequences, without any mismatches, deletions, or insertions, are classified into the “perfect match” group; reads with at least one mismatch, without any deletions or insertions, are classified into the “mismatch” group; reads with at least one deletion, without any mismatch or insertion, are classified into the “deletion” group; reads with at least one insertion, without any mismatch or deletion, are classified into the “insertion” group; reads with at least one deletion and one insertion, without any mismatch are classified into the “both deletion and insertion” group; and reads not belonging to the above groups are classified into the “others” group. (e) Read distributions of two library replicates of the BSJ-gRNA library (top). Perfectly matched read counts and corresponding ratios of gRNAs are ranked in the 10th and 90th percentiles (bottom)

4 Notes 1. Use high-efficiency electrocompetent cells, increase the number of cells used for transformation, and use plasmids at >100 ng/μL for transformation, to ensure sufficient colonies containing amplified plasmid library in Subheading 3.1. The transformation solution should be kept on ice. 2. Before lentivirus preparation, conduct an antibiotic kill curve analysis of the cells to be used for screening. The goal is to

The Functional Circular RNA Screening via RfxCas13d/BSJ-gRNA System

189

Fig. 3 Results of circRNA screening with potential functions in cell proliferation using the RfxCas13d/backsplicing junction (BSJ)-guide RNA (gRNA) system in HT29, 293FT, and HeLa cells [16, 40]. (a) Scatter plots showing the fold-change of control gRNA and circRNA-gRNA between the control (D1) and experimental (D30) groups. Some circRNA-gRNAs are depleted relative to control gRNAs in each examined cell line [16, 40]. (b) A pipeline of CDCscreen analysis to identify candidate circRNAs [16, 40]. (c) The rank of candidate circRNAs effectively depleted by BSJ-gRNAs, according to CDCscreen scores. Candidate circRNAs for further validation are labeled [16, 39]

identify a range of antibiotic concentrations that achieve 293FT_pm_unmapped.bam bamToFastq -i 293FT_pm_unmapped.bam -fq 293FT_pm_unmapped.fq samtools view -Sb -F 4 293FT_pm.sam > 293FT_pm_mapped.bam samtools sort 293FT_pm_mapped.bam 293FT_pm_sorted samtools index 293FT_pm_sorted.bam To retrieve fragments aligned to BSJ sites, the HISAT2unmapped fragments (293FT_pm_unmapped.fq, 293FT_rm_unmapped.fq, and 293FT_rr_unmapped.fq) were re-aligned by TopHat-Fusion (Fig. 2a). Except HISAT2-unmapped fragment file, input files also include bowtie1 index file for reference genome (hg38). The output is a folder containing TopHat2 mapping results

200

Xu-Kai Ma et al.

Fig. 2 A diagram of circular RNA (circRNA) annotation, sgRNA design, and evaluation of circRNA knockout by base editors. (a) Profiling of circRNAs from poly(A)–, ribo–, or RNaseR 293FT RNA-seq datasets with the CIRCexplorer2 pipeline. See text for details. (b) Designing of sgRNAs for mutating splice sites of predominantly circularized exons. See text for details. (c) Workflow of base editing back-splice sites of predominantly circularized exons and evaluating corresponding circRNA knockout effect. See text for details

(293FT_pm_fusion, 293FT_rm_fusion, and 293FT_rr_fusion). Right after that, we use CIRCexplorer2 to get fragments mapped to BSJ sites. The input file is the TopHat-Fusion mapping result (accepted_hits.bam in 293FT_pm_fusion, 293FT_rm_fusion, and 293FT_rr_fusion). The output is a new folder (293FT_pm_circ, 293FT_rm_circ, and 293FT_rr_circ) containing fragments aligned to BSJ sites (bsj.bed). Command lines that use poly(A)– 293FT RNA-seq dataset as an example for the analysis are listed below:

tophat2 -o 293FT_pm_fusion -p 20 --fusion-search --keep-fasta-order --bowtie1 --no-coverage-search hg38 293FT_pm_unmapped.fq &> 293FT_pm_fusion.log mkdir 293FT_pm_circ CIRCexplorer2 parse -f -t TopHat-Fusion 293FT_pm_fusion/accepted_hits.bam -b 293FT_pm_circ/bsj.bed &> ! 293FT_pm_parse. log

Base-Editor-Mediated circRNA Knockout by Targeting Predominantly Back. . .

201

To obtain high-confidence circRNAs with predominantly circularized exons, additional parameters are used to identify highconfidence BSJ site annotation (Fig. 2a) (additional parameters: mapped fragments ≥3, containing GU/AG splice site motif with 3-nt offset, length between two splice sites ≤30,000 nt). These annotations are merged with GENCODE gene annotation (hg38_gencode_merged.txt) for identifying circRNAs (circularRNA.txt in 293FT_pm_circ, 293FT_rm_circ, and 293FT_rr_circ). Then circRNAs are annotated with gene annotation file hg38_gencode_merged.txt. Except for the gene annotation file, input files also include the reference genome file (hg38.fa), and the “bsj.bed” file with the aforementioned BSJ information. The output file is “circularRNA.txt” in an extended BED12 format (in the folders 293FT_pm_circ, 293FT_rm_circ, and 293FT_rr_circ), including circRNA genome position, exons, and many other information (see Note 1). Command line that uses poly(A)– 293FT RNA-seq dataset as an example for the analysis is listed below: CIRCexplorer2 annotate -r hg38_gencode_merged.txt -g hg38.fa -b 293FT_pm_circ/bsj.bed -o 293FT_pm_circ/circularRNA. txt &>! 293FT_pm_annotate.log 3.2 Selection of Predominantly Circularized Exons for Base Editing

To search predominantly circularized exons for base editing, HISAT2-mapped fragments from 293FT poly(A)+ RNA-seq are included in the analyses (Fig. 2b). We use featureCounts to calculate the number of HISAT2-mapped fragments of 293FT poly(A)+ RNA-seq at back-splice sites. If the number of HISAT2-mapped fragments ≤3 (see Note 2), we define the exon as a predominantly circularized exon. Those circRNAs with predominantly circularized exons are listed in file circ_predom.txt (in the folders 293FT_pm_circ, 293FT_rm_circ, and 293FT_rr_circ). To design specific sgRNAs for targeted BSJs, flanking regions of their back-splice sites are retrieved to find nearby PAM motifs that could fit the targeted bases at back-splice sites into the editing windows of available BEs, such as hA3A-eBE-Y130F, by previously reported BEable-GPS method [28] (Fig. 2b). Command lines that use 3′ back-splice sites of circRNAs identified from poly(A)– 293FT RNA-seq dataset as an example for the analysis are listed below: python extract_bss.py 293FT_pm_circ/circ_predom.txt python extract_sequences.py 293FT_pm_circ/3bss.seq

hg38.fa

293FT_pm_circ/3bss.bed

python extract_sequences.py hg38.fa 293FT_pm_circ/3bss2th.bed 293FT_pm_circ/3bss2th.seq python BEable_Cas9_CtoT.py 293FT_pm_circ/3bss.seq NGG 20 3 8 > 293FT_pm_circ/3bss_hA3A_Y130F.txt

202

Xu-Kai Ma et al.

python BEable_Cas9_CtoT.py 293FT_pm_circ/3bss2th.seq NGG 20 3 8 > 293FT_pm_circ/3bss2th_hA3A_Y130F.txt By combining editing information of 5′ and 3′ back-splice sites (3bss_hA3A_Y130F.txt, 3bss2th_hA3A_Y130F.txt, 5bss_hA3A_ Y130F.txt, and 5bss2th_hA3A_Y130F.txt), BE-editable circRNAs and their corresponding sgRNAs for predominantly circularized exons are obtained (circ_predom_edit.txt and circ_sgRNAs.txt in 293FT_pm_circ, 293FT_rm_circ, and 293FT_rr_circ). 3.3 Evaluation of Knockout Effect

Here, we select one circRNA, circRALY(NE,2) with predominantly circularized novel exon (NE and see Note 3), as an example to illustrate the design and the whole experiment setup for BE-mediated circRNA KO (Figs. 2c and 3). Of note, this circRNA, previously listed as circRALY-nov in Gao et al. [20], has now been renamed as circRALY(NE,2) according to the most recently published nomenclature for circRNA naming [1].

3.3.1 Plasmid Construction

Construction of the sgRNA plasmid was achieved by inserting the sgRNA sequence into the BsaI sites of the PGL3 vector (Fig. 2c). 1. Linearize the PGL3 vector by BsaI restriction enzyme and recover DNA with the StarPrep Gel Extraction Kit. 2. Anneal the sgRNA oligos by heating at 95 °C at a thermocycler for 5 min, and then gradually cool down to room temperature at a speed of -1 °C per 30 s. 3. Mix 100 ng of the linearized PGL3 vector with 5 ng of annealed oligos, and ligate the oligos to the vector by T4 DNA ligase. 4. Transform the ligation mixture into Trans1-T1 chemically competent cells following the standard molecular cloning protocol. 5. Check the positive clones by Sanger sequencing and prepare sgRNA plasmid by NucleoBond Xtra Midi kit.

3.3.2 Cell Culture and Transfection

1. 293FT cells (Thermo) were maintained in DMEM supplemented with 10% fetal bovine serum and 0.1% penicillin/streptomycin at 37 °C in a 5% CO2 cell culture incubator. 2. The day before transfection, seed 293FT cells in a 12-well plate at a density of 2 × 105 cells per well. 3. Carry out co-transfection of 1.7 μg BE expressing vector and 1.1 μg sgRNA expressing vector with Lipofectamine™ 3000 Reagent according to the manufacturer’s protocol. 4. 24 h after transfection, add puromycin into the culture medium at a final concentration of 1 μg/mL to select transfected cells.

Base-Editor-Mediated circRNA Knockout by Targeting Predominantly Back. . .

203

5. After 3 days, collect cells into a 1.5 mL tube by trypsinization and centrifugation. Wash the cell pellet twice with DPBS and then store it at -80 °C for further analyses. 3.3.3 Calculation of Base Editing Ratio

1. Extract genomic DNA from collected cells with the TIANamp Genomic DNA Kit according to the manufacturer’s protocol. 2. Amplify the genomic DNA fragments flanking target sites with primers spanning the base editing site (Fig. 3a, and see Note 4), and separate the products with 1.5% agarose gel electrophoresis (Fig. 3b). 3. Cut and recover the target bands with the StarPrep Gel Extraction Kit.

Fig. 3 Evaluation of circular RNA (circRNA) knockout effect by base editing back-splice sites. (a) Schematic of partial RALY gene organization (according to the ENST00000375114.7 transcript from GENCODE v31). A novel exon (NE, pink bar), is predominantly circularized for the production of circRALY(NE,2). Of note, this specific NE is rarely spliced for the expression of cognate linear RALY RNA. (b) Validation of circRALY(NE,2) with divergent primers [20]. Note that the circRALY(NE,2) remains stable after RNase R treatment, while the linear RALY RNA is largely degraded. * indicates amplified products with divergent primers for circRALY(NE,2), possibly due to a rolling PCR. (c) Calculation of base editing efficiency at the 3′ back-splice site of the specific predominantly circularized NE [20]. Context sequences of targeted 3′ (b)ss were shown by a, t, c, and g for intron or by A, T, C, and G for exon. (d) Expression of circRALY(NE,2) and linear RALY RNA after base editing. Note that circRALY(NE,2) is specifically repressed, while its linear RALY RNA is barely affected by RT-qPCR [20]

204

Xu-Kai Ma et al.

4. Analyze the base editing ratio by DNA sequencing (Fig. 2c). To calculate the C-to-T (/G-to-A) editing by CBEs or A-to-G (/T-to-C) editing by ABEs at each target site, heights of A, T, C, and G signals of Sanger sequencing were retrieved by Bioedit and processed by the following equation: Editing ratio = [Cheight/(Cheight + Theight) or Aheight/(Aheight + Gheight)] (Fig. 3c and see Note 5). 3.3.4 RNA Isolation and Detection

1. Add 1 mL TRIzol to collected cells and isolate total RNAs (see Note 6) according to the manufacturer’s protocol. The RNA concentration was determined by the OD260 using Nanodrop (Thermo). 2. Treat 10 μg total RNA with 1 μL DNase I at 37 °C for 30 min to remove genomic DNA contamination, and extract RNA with phenol–chloroform–isoamyl alcohol (see Note 7). 3. (Optional) Enrich circRNAs by RNase R digestion: (a) Dilute 4 μg total RNA to a total volume of 34 μL with RNase-free water in a 1.5 mL tube. (b) Denature the RNA at 68 °C for 5 min, and place the tube on ice immediately for 2 min. (c) Split the denatured RNA into two aliquots in two new 1.5 mL tubes, one for RNase R digestion and the other one for control. (d) For RNase R digestion, add 2 μL 10 RNase R Reaction buffer and 20 U RNase R. For control treatment, add 2 μL 10 RNase R Reaction buffer and 1 μL RNase-free water to a final volume of 20 μL. Mix thoroughly, and quickly spin the tubes for a few seconds. (e) Incubate at 37 °C for 30 min in a ThermoMixer with a thermoblock. (f) Inactive the enzyme at 70 °C for 5 min. (g) Add 80 μL RNase-free water to the reaction to a total volume of 100 μL and proceed with phenol/chloroform extraction, followed by ethanol precipitation and air drying of the RNA pellet. (h) Dissolve the RNA in 20 μL RNase-free water. 4. Synthesize first-strand cDNA: (a) Mix 1 μg total RNA, 1 μL random hexamers (100 μM), and 1 μL dNTP Mix(10 mM each) with RNase-free water to 13 μL. For the RNase R digestion assay, add an equal volume of RNase R enriched RNA or non-treated RNA. (b) Heat the mixture at 65 °C for 5 min, and immediately chill on ice for 2 min.

Base-Editor-Mediated circRNA Knockout by Targeting Predominantly Back. . .

205

(c) Spin the tube briefly. Add 4 μL 5 First-Strand Buffer, 1 μL 0.1 M DTT, 1 μL Rnasin, and 1 μL SuperScript™ III RT. Gently mix well and incubate at 25 °C for 5 min. (d) Incubate at 50 °C for 60 min. (e) Inactivate the reaction by heating at 70 °C for 15 min. (f) Store the cDNA at -20 °C until use. 5. Analyze circRNA expression with PCR: (a) Design primers for circRNA or cognate linear RNA detection. Use divergent primers across the back-splicing junction site or convergent primers spanning at least two exons flanking the predominant exons to evaluate circular and linear RNA splice/expression, respectively. (b) PCR amplification with appropriate cDNA templates and corresponding primers. (c) Run the PCR products on a 1.5% agarose gel. (d) Recovery of the target bands for sequencing. 6. Perform the real-time PCR using SYBR Green™ Real-time PCR Master Mix according to the manufacturer’s protocol (Fig. 3d and see Note 8). (a) Prepare each sample as follows: 5 μL SYBR Green Mix, 0.4 μL forward primer (10 μM), 0.4 μL reverse primer (10 μM), 1 μL cDNA, and 3.2 μL RNase-free water. (b) Then run the qPCR reaction on a QuantStudio™ 6 Flex Real-Time PCR System.

4

Discussion A large number of circRNAs have been profiled from different cell lines/tissues across eukaryotes, mainly by computational analysis of massive deep sequencing datasets [2–6]. Generally, circRNAs are co-expressed with their cognate linear RNAs from the same gene loci. That is why commonly used approaches that work well for linear RNAs may not discriminate circRNAs from their cognate linear RNAs. Possibly due to these technical obstacles, biological potentials of most circRNAs are still not clearly addressed [4, 6]. To solve this problem, new approaches with careful design have been applied to achieve efficient effects specifically on circRNAs but not linear ones, including using the state-of-the-art BEs for circRNA LOF studies [20]. Compared with the whole exon deletion by CRISPR-Cas9 for circRNA KO, BE-mediated circRNA depletion only requires the change of a few bases near back-splice sites.

206

Xu-Kai Ma et al.

However, as many circularized exons can be also co-linearly spliced for the production of linear RNAs, BE-mediated circRNA KO is inevitably challenged by the common limitation of simultaneous effects on cognate circular and linear RNAs with overlapping sequences (Fig. 1a). Thus, we suggest to screen/choose exons predominantly included in circRNAs but not in cognate linear RNAs for BE-targeting (Fig. 1b). With the development of applying other methods to achieve efficient and specific circRNA KD, such as by RfxCas13d targeting BSJ sites, the application of BEs for circRNA KO can be complementarily used for functional exploration of circRNAs.

5

Notes 1. Detailed information can be retrieved from https:// circexplorer2.readthedocs.io/en/latest/modules/annotate/. 2. Different numbers of HISAT2-mapped fragments can be used as the threshold for filtering, which is also dependent on sequence depth. Here, we used at least three HISAT2-mapped fragments (for 293FT RNase R sample has ~38 M total reads) for the selection of high-confidence circRNA prediction. 3. The label with NE, indicating novel exon, is not ideal, and it should be viewed as a temporary placeholder. Further efforts to annotate the transcriptome—especially by incorporating circRNAs into databases such as GENCODE and RefSeq— should allow more unambiguous names to be given in the future. 4. Here, we generally design DNA fragment size at about 400 bp for amplification and sequencing. 5. To evaluate base editing efficiency at the target site, nearby bases across the spacer region should also be calculated, to achieve a convincing knockout effect. 6. For RNA experiments, use nuclease-free reagents, tubes, and tips. 7. The inactivator can be used to inactivate the DNase to perform an immediate reverse transcription reaction. 8. If there is more than one circular or linear transcript produced at the target gene locus, expressions of all these circRNAs and linear RNAs need to be detected.

Base-Editor-Mediated circRNA Knockout by Targeting Predominantly Back. . .

207

Acknowledgments This work was supported by the National Science Foundation of China (NSFC, grant number 31925011) and the Ministry of Science and Technology of China (MoST, grant numbers 2021YFA1300503 and 2019YFA0802804). References 1. Chen LL, Bindereif A, Bozzoni I, Chang HY, Matera AG, Gorospe M, Hansen TB, Kjems J, Ma XK, Pek JW, Rajewsky N, Salzman J, Wilusz JE, Yang L, Zhao F (2023) A guide to naming eukaryotic circular RNAs. Nat Cell Biol 25(1):1–5 2. Kristensen LS, Andersen MS, Stagsted LVW, Ebbesen KK, Hansen TB, Kjems J (2019) The biogenesis, biology and characterization of circular RNAs. Nat Rev Genet 20(11):675–691 3. Li X, Yang L, Chen LL (2018) The biogenesis, functions, and challenges of circular RNAs. Mol Cell 71(3):428–442 4. Liu CX, Chen LL (2022) Circular RNAs: characterization, cellular roles, and applications. Cell 185(13):2390 5. Wilusz JE (2018) A 360 degrees view of circular RNAs: from biogenesis to functions, Wiley Interdiscip rev. RNA 9:e1478 6. Yang L, Wilusz JE, Chen LL (2022) Biogenesis and regulatory roles of circular RNAs. Annu Rev Cell Dev Biol 38:263–289 7. Li X, Liu CX, Xue W, Zhang Y, Jiang S, Yin QF, Wei J, Yao RW, Yang L, Chen LL (2017) Coordinated circRNA biogenesis and function with NF90/NF110 in viral infection. Mol Cell 67(2):214–227 e7 8. Liu CX, Li X, Nan F, Jiang S, Gao X, Guo SK, Xue W, Cui Y, Dong K, Ding H, Qu B, Zhou Z, Shen N, Yang L, Chen LL (2019) Structure and degradation of circular RNAs regulate PKR activation in innate immunity. Cell 177(4):865–880 e21 9. Kramer MC, Liang D, Tatomer DC, Gold B, March ZM, Cherry S, Wilusz JE (2015) Combinatorial control of drosophila circular RNA expression by intronic repeats, hnRNPs, and SR proteins. Genes Dev 29(20):2168–2182 10. Pamudurti NR, Bartok O, Jens M, AshwalFluss R, Stottmeister C, Ruhe L, Hanan M, Wyler E, Perez-Hernandez D, Ramberger E, Shenzis S, Samson M, Dittmar G, Landthaler M, Chekulaeva M, Rajewsky N, Kadener S, Translation of CircRNAs (2017) Mol Cell 66(1):9–21 e7

11. Liu CX, Guo SK, Nan F, Xu YF, Yang L, Chen LL (2022) RNA circles with minimized immunogenicity as potent PKR inhibitors. Mol Cell 82(2):420–434 e6 12. Ai Y, Liang D, Wilusz JE (2022) CRISPR/ Cas13 effectors have differing extents of off-target effects that limit their utility in eukaryotic cells. Nucleic Acids Res 50(11):e65 13. Li S, Li X, Xue W, Zhang L, Yang LZ, Cao SM, Lei YN, Liu CX, Guo SK, Shan L, Wu M, Tao X, Zhang JL, Gao X, Zhang J, Wei J, Li J, Yang L, Chen LL (2021) Screening for functional circular RNAs using the CRISPRCas13 system. Nat Methods 18(1):51–59 14. Zhang Y, Nguyen TM, Zhang XO, Wang L, Phan T, Clohessy JG, Pandolfi PP (2021) Optimized RNA-targeting CRISPR/Cas13d technology outperforms shRNA in identifying functional circRNAs. Genome Biol 22(1):41 15. Anderson EM, Birmingham A, Baskerville S, Reynolds A, Maksimova E, Leake D, Fedorov Y, Karpilow J, Khvorova A (2008) Experimental validation of the importance of seed complement frequency to siRNA specificity. RNA 14(5):853–861 16. Wolters NM, MacKeigan JP (2008) From sequence to function: using RNAi to elucidate mechanisms of human disease. Cell Death Differ 15(5):809–819 17. Zhang Y, Xue W, Li X, Zhang J, Chen S, Zhang JL, Yang L, Chen LL (2016) The biogenesis of nascent circular RNAs. Cell Rep 15(3): 611–624 18. Piwecka M, Glazar P, Hernandez-Miranda LR, Memczak S, Wolf SA, Rybak-Wolf A, Filipchyk A, Klironomos F, Cerda Jara CA, Fenske P, Trimbuch T, Zywitza V, Plass M, Schreyer L, Ayoub S, Kocks C, Kuhn R, Rosenmund C, Birchmeier C, Rajewsky N (2017) Loss of a mammalian circular RNA locus causes miRNA deregulation and affects brain function. Science 357(6357):eaam8526 19. Zhang XO, Dong R, Zhang Y, Zhang JL, Luo Z, Zhang J, Chen LL, Yang L (2016) Diverse alternative back-splicing and

208

Xu-Kai Ma et al.

alternative splicing landscape of circular RNAs. Genome Res 26(9):1277–1287 20. Gao X, Ma XK, Li X, Li GW, Liu CX, Zhang J, Wang Y, Wei J, Chen J, Chen LL, Yang L (2022) Knockout of circRNAs by base editing back-splice sites of circularized exons. Genome Biol 23(1):16 21. Chen J, Yang B, Yang L (2019) To BE or not to BE, that is the question. Nat Biotechnol 37(5): 520–522 22. Yang L, Chen J (2020) A tale of two moieties: rapidly evolving CRISPR/Cas-based genome editing. Trends Biochem Sci 45(10):874–888 23. Kim D, Langmead B, Salzberg SL (2015) HISAT: a fast spliced aligner with low memory requirements. Nat Methods 12(4):357–360 24. Kim D, Salzberg SL (2011) TopHat-fusion: an algorithm for discovery of novel fusion transcripts. Genome Biol 12(8):R72 25. Li H, Handsaker B, Wysoker A, Fennell T, Ruan J, Homer N, Marth G, Abecasis G,

Durbin R, S. (2009) Genome project data processing, the sequence alignment/map format and SAMtools. Bioinformatics 25(16): 2078–2079 26. Quinlan AR, Hall IM (2010) BEDTools: a flexible suite of utilities for comparing genomic features. Bioinformatics 26(6):841–842 27. Liao Y, Smyth GK, Shi W (2014) featureCounts: an efficient general purpose program for assigning sequence reads to genomic features. Bioinformatics 30(7):923–930 28. Wang Y, Gao R, Wu J, Xiong YC, Wei J, Zhang S, Yang B, Chen J, Yang L (2019) Comparison of cytosine base editors and development of the BEable-GPS database for targeting pathogenic SNVs. Genome Biol 20(1):218 29. Bolger AM, Lohse M, Usadel B (2014) Trimmomatic: a flexible trimmer for Illumina sequence data. Bioinformatics 30(15): 2114–2120

Chapter 12 Directed Circularization of a Short RNA Cathrin E. Hansen, Danilo Springstubbe, Sabine Mu¨ller, and Sonja Petkovic Abstract Basic research and functional analyses of circular RNA (circRNA) have been limited by challenges in circRNA formation of desired length and sequence in adequate yields. Nowadays, circular RNA can be obtained using enzymatic, “ribozymatic,” or modulated splice events. However, there are few records for the directed circularization of RNA. Here, we present a proof of principle for an affordable and efficient RNA-based system for the controlled synthesis of circRNA with a physiological 3′,5′-phosphodiester conjunction. The engineered hairpin ribozyme variant circular ribozyme 3 (CRZ-3) performs self-cleavage poorly. We designed an activator-polyamine complex to complete cleavage as a prerequisite for subsequent circularization. The developed protocol allows synthesizing circRNA without external enzymatic assistance and adds a controllable way of circularization to the existing methods. Key words Circular, RNA, Ribozyme, Activator, Circularization, Bioengineering, Spermine, Catalysis

1

Introduction

1.1 Biological Relevance of Circular RNA

Circular RNAs (circRNAs) are defined by their covalent 3′-5′-linkage, which increases their resistance to exoribonucleases and half-life [1– 5]. Their relatively stable conformations provide the potential to retain functional features even when metabolic or structural changes within the cell occur. In the last decade, the scientific community recognized the abundant appearance of circRNA and its pivotal role in eukaryotes and archaea, which carry out numerous tasks in cellular metabolism [6–15]. Some prominent functions are as follows: (a) The competition of circRNA with (other) endogenous RNAs to bind miRNAs, thus acting as RNA sponges [16–21], (b) As reservoirs for RNA binding proteins (protein sponges) [22–25], or (c) In vivo and in vitro circular RNA translation to proteins [26– 32].

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_12, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

209

210

Cathrin E. Hansen et al.

Jeck et al. and Piwecka et al. reported that circular RNAs contribute primarily to gene regulation forming a miRNA reservoir, wherein miRNAs can be absorbed, stored, and released, hence modulating the free cytoplasmic concentration of miRNA [33, 34]. Knockdown studies in mice brain tissue revealed that circRNAs might co-express with specific miRNAs [34]. Circularization and ring-opening events of certain RNAs can constantly preserve the order of RNA domains and facilitate the rearrangement of sequences within the single-stranded RNAs, increasing the biological diversity in vivo [29]. Several in vivo and in silico studies identified circRNAs in pathogenic tissues and suggested their potential as novel biomarkers for human diseases [35–41]. This might be especially attractive for the fields of oncology [16, 42–45] and neurodegenerative diseases [34, 46–49] but also acute instances like cardiovascular events [50–55]. CircRNAs are even tested as vaccines against SARS-CoV-2 and emerging variants using mice [56] and rhesus macaques models [26]. Due to their high biological relevance, studies on circRNAs are exceedingly required and observed [57] and call for a wealth of standardized methods for their synthesis. 1.2 Common Circularization Strategies

So far, a handful of chemical, enzymatic, and RNA-based methods for RNA circularization in vitro are available [58]. Chemical strategies involve the solid-phase supported cyclization [59, 60], enzymatic approaches including the usage of T4 RNA Ligase I [61] and II [62], and T4 DNA ligase [32] mediated ligation, the permuted intron-exon (PIE) method [63, 64], and strategies using modified ribozymes [65]. In the past, all approaches showed disadvantages such as low yields, the high potential of multimerization of the RNA, or the production of non-physiological linkages. In addition, chemical circularization occurs merely in vitro, and the application of enzymes increases the expense of the experimental setups but is widely used anyway. The PIE method can be applied in vitro as well as in vivo and was optimized and itemized for some applications [66, 67], but the formation of new constructs comprises several synthesis steps and is time-consuming. Recently, Weinberg et al. identified a wealth of hairpin ribozymes bioinformatically in metatranscriptomes, suggesting an underestimated biological relevance of these RNAs [68]. Hence, we engineered RNA for its circularization in a controlled fashion using a hairpin ribozyme.

Directed Circularization of a Short RNA

2

211

Materials Milipore-quality deionized water was used in the preparation of the buffers. The solutions were filtered using a filtration apparatus (sterile filter: 0.2 μm pore size, Nitrocellulose) and degassed. Unless otherwise stated, pH values were measured at room temperature using a pH meter or pH indicator paper.

2.1

DNA Analysis

DNA isolation or quality control after the Klenow reaction may require DNA agarose gel electrophoresis. Use 4% (w/v) agarose gels with an agarose powder in “DNA grade” or “DNA quality.” Solve the agarose in 1x TAE buffer: 40 mM Tris, 2 mM EDTA-Na2 (pH 8.0), and 20 mM acetic acid by cooking for 2 min and using TAE as a running buffer during electrophoresis. After cooling the hot agarose solution for 2 min, add (0.5 μg/mL) ethidium bromide. Mix DNA products with DNA loading buffer: 50% (v/v) 1x TBE: 89 mM Tris (pH 8.3), 89 mM boracic acid, 20 mM EDTA (pH 8.0), 50% (v/v) glycerol, and 0.025% (w/v) bromophenol blue, and 0.025% (w/v) xylene cyanol.

2.2 DNA-Dependent In Vitro Transcription of RNA

The in vitro RNA transcription (with or without GMP priming) requires 1x HEPES buffer: (pH 7.5) 50 mM Na-HEPES, 12 mM MgCl2, 2 mM spermidine, and 40 mM DTT.

2.3

For the separation of linear and circular RNAs, use polyacrylamide gels: 20% (v/v) acrylamide solution, denaturing: 40% (v/v), Rotiphorese Gel 40® (19:1), including 1x TBE and 7 M urea. Add 10% (w/v) APS in H2O and 5–15 μL TEMED. The prepared acrylamide solution was filtered with a solvent-resistant sterile filter (0.2 μm, regenerated cellulose). Use 2x loading buffer including 7 M urea: 50 mM EDTA, 1x TBE: 89 mM Tris (pH 8.3), 89 mM boracic acid, 20 mM EDTA (pH 8.0), 0.5% (w/v) Bromophenol, and 0.5% (w/v) Xylene cyanol. Stain the RNA 0.2x SYBR® Gold using 1 μL SYBR® Gold in 50 mL 1x TBE. If isolation of RNA from gel slices is necessary, use (0.1–0.3 M) sodium acetate buffer (pH = 7.0) as elution buffer.

2.3.1 Gels

RNA Analysis Polyacrylamide

2.3.2 Reverse Transcription PCR

Approximately 3 ng RNA, 0.1 μM of the reverse primer (RP 1: AGT RP2: CGC GAA GTA ATA TCT TTC TCC AGT TCA C),0.5 mM dNTP mix, and H2O are required to synthesize cDNA. The reaction buffer consists of 200 mM Tris– HCl (pH 8.4), 500 mM KCl, 5 mM DTT, RNase Inhibitor, and 200 U SuperScript® III Reverse Transcriptase were added to the reaction. For a fast and reliable polymerase chain reaction (PCR), we added a combination of Taq DNA polymerase and Pfu DNA TCA CTT CTC TCT TTG ACG GAA CCC,

212

Cathrin E. Hansen et al.

polymerase to less than 1:10 of the RT product. A buffer for the Taq polymerase is required: 20 mM Tris–HCl (pH 8.4), 50 mM KCl, 5 mM MgCl2, 2.5 mM dNTP mix, and forward and reverse primers to 1 μM each (forward primer 1: GGA GAA AGA TAT TAC TTC GCG CGT GGT and forward primer 2: CGT GGT ATA TTA CCT GGT CCC CC; for reverse primers, see above).

3

Methods Hairpin ribozymes are a versatile tool to engineer their activity toward a desired function, such as RNA recombination [69] and self-splicing [70]. The origin of the system used in this proof of principle experiment is a small satellite RNA from the (-) strand of the tobacco ringspot virus [71]. This RNA undergoes replication via a rolling circle mechanism, and the catalytic domains mediate the cleavage of the resulting concatemers to monomeric genomes. The cleavage results in 2′,3′-cyclic phosphates, and 5′-hydroxyl (OH) groups. Reversibility of the reaction is preserved due to the energy difference between the cyclic phosphate and the 3′,5′-phosphodiester being nearly neutral. In vitro studies of several variants of the catalytic RNA led to the development of hairpin ribozyme variants that favor either cleavage or ligation of their substrates [62, 65]. We aimed to develop a protocol for circularizing a short stable RNA. The synthesis of the RNA construct follows four steps (Fig. 1): 1. Klenow reaction 2. GMP-primed in vitro transcription 3. Two autocatalytic cleavages 4. Ribozyme ligation

3.1

Klenow Reaction

The double-stranded DNA (dsDNA) template was prepared by the Klenow reaction (Fig. 1, step 1). We checked the primer sequences below for their necessary ability to hybridize using the OligoAnalyzer™ tool. The forward Klenow primer contains the T7 RNA promotor (bold): CRZ-3 forward (86 nt): A ATA CGA CTC ACT ATA GGG AGA TCA CAG TCC TCT TTG ACG GGG TTC CGT CAA AGA GAG AAG TGA ACT GGA GAA AGA TAT TAC TTC G

CRZ-3 reverse (85 nt):

AAA GAG GAC TGT GAG GGG GAC CAG GTA ATA TAC CAC GCG CGA AGT AAT ATC TTT CTC CAG TTC ACT TCT CTC TTT GAC GGA ACC C

To check the length and quality of the Klenow primers, 15% denaturing polyacrylamide gels were used.

Directed Circularization of a Short RNA

213

Fig. 1 Synthesis of the RNA construct. A double-stranded DNA (dsDNA) is synthesized in the first step to prepare the RNA. The Klenow DNA serves as a template for in vitro transcription to single-stranded RNA needed in the second step. The binding site of the T7 RNA polymerase promoter sequence of the RNA polymerase is emphasized in gray. The third step involves cleavage of the free 3′- and 5′-termini followed by ligation of the RNA in the last step. The blue-highlighted end of the promotor is transcribed and removed with the rest of the 5′-cleavage substrate. The self-processing activity of the test hairpin ribozyme fulfills cleavage and ligation reactions in the presence of an activator and spermine. The red sequences illustrate the RNA segments (substrates) to be cleaved. CRZ-3: cyclic ribozyme 3

1. For a single Klenow reaction setup, 2 μM of each Klenow primer, 1x Klenow buffer (pH 8.0), 0.5 mM dNTP mix, and 0.04 U/μL Klenow fragment exo- were used. The final volume was 500 μL. 2. The Klenow primers were first denatured at 90 °C for 2 min to prevent the formation of secondary structures or mismatches of the primers. 3. The Klenow fragment exo- was not thermostable, so adding the polymerase and dNTP mix was done after cooling to 37 °C. 4. The polymerization reaction time was 30 min at 37 °C resulting in a dsDNA template.

214

Cathrin E. Hansen et al.

5. Precipitation was applied to purify nucleic acids from DNA reaction mixtures. For this purpose, first, by adding 2.5 times the absolute volume of ethanol, the solubility of the nucleic acid was minimized, and the RNA or DNA was isolated after 40 min of centrifugation at 4 °C, and 20,000 g precipitated as a solid. We discarded the supernatant, washed the DNA with 70% ethanol, and centrifuged for 10–15 min (4 °C, 20,000 g). The supernatant was removed, and the DNA pellet dried at room temperature. The samples were dissolved in 50–200 μL H2O (Millipore grade) for further use. 6. We checked the quality of the DNA products with 4% (w/v) agarose gels. 3.2 In Vitro Transcription with GMP Priming and RNA Isolation

In guanosine monophosphate (GMP) priming, an excess of GMP is added to the reaction mix and to the components required for in vitro transcription. This results in 70–90% incorporation of a GMP instead of a GTP as the first nucleotide of the transcript [72]. (See Note 5.1 for additional information.) The DNA template was transcribed into single-stranded RNA (ssRNA) by GMP-primed in vitro transcription. 1. A single reaction requires 1x HEPES buffer (pH 7.5), 1 μM DNA template, 0.625 mM GTP, 2 mM ATP, 2 mM CTP, 2 mM UTP, and 3 mM GMP (GMP:GTP 4.8:1). The final volume was 50 μL. To prevent hydrolysis of the synthesized RNA by RNases, 1 U/μL RiboLock RNase inhibitor was added to the reaction. 2. The synthesis of RNA was carried out at 37 °C for 3 h but can be run overnight, too. However, increasing the reaction time did not result in higher RNA yields. 3. Subsequently, the DNA template was hydrolyzed using 2 U DNase I by adding it directly to the mix. 4. RNA precipitation with ethanol separated the product from the hydrolyzed DNA fragments and residual nucleotides. 5. We used denaturing polyacrylamide gels (15% (v/v), 7 M urea) to isolate and purify the desired full-length ssRNA products. Samples were mixed 1:2 with 2x denaturing loading buffer (7 M urea, 50 mM EDTA, 1x TBE (pH 8.0)) and denatured at 90 °C for 2 min immediately before electrophoresis. The application volume was 50 μL per well. Gel electrophoresis was performed for 4 h and 30 min at 200 V at room temperature. (See Note 5.2 for additional information.) After electrophoresis, the RNAs were stained with SYBR® Gold, and the desired RNAs were cut out. The elution of RNA from the gel pieces was performed using two different concentrations of sodium acetate buffer (pH 7.0). We overlaid the pieces with

Directed Circularization of a Short RNA

215

0.3 M sodium acetate buffer first and shook them for 2 h at 4 °C. We repeated the procedure with 0.1 M sodium acetate for 2 h and did the final elution with 0.1 M sodium acetate overnight at 4 °C. All eluates were collected and combined according to their constructs, filtered, and stored at -20 °C. A lyophilization reduced the volume before RNA precipitation with ethanol. 3.3 RNA Purification by Precipitation

To purify RNA, 2.5 times the volume of absolute ice-cold ethanol was added to the nucleic acid solution (eluted and lyophilized in sodium acetate). RNA precipitation took place overnight at -25 ° C. Then, centrifugation was performed at 21,000 g for 40 min. The supernatant was decanted, and the pellet was washed with 70% ethanol and centrifugated again. Once more, we discarded the supernatant, and the pellet was dried at room temperature. The resuspension was done in DEPC-treated water.

3.4 Ribozyme Reaction

The optimized ribozyme reaction resulting in an RNA with covalently linked 5′ and 3′ termini comprises cleavage and ligation reactions that merge into each other. The RNA transcript is (partly) cleaved using 10 mM MgCl2, 50 pmol RNA, and 10 mM Tris–HCl pH 7.5, in a total volume of 20 μL, for 30 min at 37 °C. To obtain a fully cleaved RNA and ligated termini, adding an RNA oligomer “activator” in a tenfold excess to the linear RNA and combination with 5 mM spermine is necessary. The activator was 18 nt long and consisted of the following sequence: 5′-ACC AGA GAA ACA CAC GCG3′. We purchased the activator (purification: HPLC grade) from Biomers GmbH. The ligation takes place within 15–30 min at 37 ° C and is stopped by adding the same volume of loading buffer to the reaction. (See Note 5.3 for additional information.)

3.5 Reverse Transcription PCR

Approximately 3 pg RNA was heat denaturated for 2 min at 90 °C, and the addition of 0.1 μM of the reverse primer,0.5 mM dNTP mix, and H2O is required to synthesize cDNA. The reaction buffer consists of 200 mM Tris–HCl (pH 8.4), 500 mM KCl, 5 mM DTT, RNase Inhibitor, and 200 U SuperScript® III Reverse Transcriptase were added to the reaction. Reverse transcription was performed at 55 °C for 45 min, followed by heat inactivation of the polymerase at 70 °C for 15 min. For a fast and reliable polymerase chain reaction (PCR), we added a combination of Taq DNA polymerase and Pfu DNA polymerase to less than 1:10 of the RT product. A buffer for the Taq polymerase is required: 20 mM Tris–HCl (pH 8.4), 50 mM KCl, 5 mM MgCl2, 2.5 mM dNTP mix, and forward and reverse primers to 1 μM each. The final volume was 70 μL. The PCR was started with an initial denaturation for 4 min at 95 °C, followed by 30 cycles of 30 s of denaturation (94 °C), 25 s of annealing (54 °C), and 20 s of elongation (72 °C). The final elongation was done at 72 °C for 4 min.

216

4

Cathrin E. Hansen et al.

Results Our system, “cyclic ribozyme 3, CRZ-3,” contains most but not all crucial domains for the original hairpin ribozyme activity (Fig. 2, purple: unfavorable sequences). The mechanism of the reaction is composed of three steps (Fig. 2). The 109-nt-long full-length transcript (5′-GGG AGA UCA CAG UCC UCU UUG ACG GGG UUC CGU CAA AGA GAG AAG UGA ACU GGA GAA AGA UAU UAC UUC GCG CGU GGU AUA UUA CCU GGU CCC CCU CAC AGU CCU CUU U-3´)

comprises two substrate sequences of 9 and 11 nt length, one at each terminus (Fig. 2, red: substrates). Cleavage of both sequences removes a short patch at both the 5′- and 3′-end. It produces the functional groups required for ribozymatic ligation: a 5′-OH-group and a 2′-3′-cyclic phosphate (Fig. 2). If we included all crucial hairpin ribozyme sequences, we would observe ligation to some extent and would be unable to control ligation, unintended ring opening, or formation of linear or circular RNA concatemers (repeating sequences). The unregulated RNA circularization worked for our former catalytic RNA “cyclic ribozyme 2” [62]. Cleavage would be followed by intramolecular ligation of the engendered ends. The standard protocol for typical hairpin ribozyme reactions includes three steps: (a) RNA denaturation and folding (2–5 min 90 °C, folding at 20 ° C for 15 min of the reaction mixture) (b) Cleavage reactions (50 pmol RNA, 10 mM Tris–HCl pH 7.5, 10 mM MgCl2, total volume 20 μL, 2 h, 37 °C) (c) Ligation (50 pmol RNA, 10 mM Tris–HCl pH 7.5, 50 mM MgCl2, total volume 20 μL, 2 h, 37 °C) [62, 73]. The usage of this original protocol did not result in either complete cleavage of our test ribozyme or production of circularized RNA “CRZ-3.” The product yield following a hairpin ribozyme reaction depends on several variables, from which we investigated kinetics, MgCl2 concentration, activator efficacy, and adding the co-factor spermine. The first assays lacked the activator (RNA, 18 nt long, 5′-ACC AGA GAA ACA CAC GCG-3′) to prove that the full-length transcript cannot circularize itself in an uncontrolled manner. We analyzed the kinetics of the cleavage reaction from 0 to 120 min and monitored the results by polyacrylamide gel electrophoresis. The cleavage reactions may be performed within 2 h (“long protocol”) or in an even shorter period of 0, 15, 30 (“cut protocol”), 45, and 60 min (results of the cut protocol are exemplarily shown in Fig. 3a). The upper band is the residual 109mer, whereas the lower band consists of the cleavage product. We reduced the

Directed Circularization of a Short RNA

217

Fig. 2 Schematic of the controlled RNA circularization steps: (a) Upon activator binding to the linear 109 nt long RNA (l-109mer), two cleavage reactions take place, the order depending on which terminus complements the stem-loop structure (substrate sequences illustrated in red). Cleavages result in the formation of ligationcompetent termini: the 2′,3′-cyclic phosphate (cp) and 5′-OH of the final, linear cleaved product (l-89mer), which is subsequently ligated (b) to a circular RNA (c-89mer). *All steps are generally reversible, although a ring-opening and two substrate ligation steps are unlikely. The RNA ring is stable upon removal of the activator from the ribozyme, e.g., using denaturing polyacrylamide electrophoresis. The red arrow indicates the cleavage and ligation position. The purple region contains an unfavorable sequence that hinders ligation and may be substituted with an arbitrary sequence as long as the cleavage sites may be formed. The green dotted loop represents a variable sequence, which is replaceable with an RNA sequence of interest

Fig. 3 Results of protocol adjustments and circular product identification. (a) Kinetics without activator and spermine. Lanes: increase in reaction time 1–3: 0, 15, and 30 min, an increase in time 5–7: 45, 60, and 120 min; 4: single-stranded RNA (ssRNA) size standard (100 nt); 8: aliquot of the purified in vitro transcript. This is the full-length RNA. (b) Increasing concentrations of the applied activator (no spermine). Lanes: 1–4: 0, 5, 15, and 25 pmol; 6–8: 35, 50, and 100 pmol; 5: ssRNA size standard 100 nt. (c) Combined addition of activator (10× over RNA) and increasing concentrations of spermine. Lanes: 1: 0 mM spermine, loading of cleaved RNA product, 2: 2 mM spermine, 3: single-stranded RNA size standard 100 nt, 4–7: increase in spermine concentration (2.75, 3.5, 4.25, and 5 mM); RNA product: lane 1: 150 pmol: lanes 2 and 4–7: 50 pmol. (d) Verification of the RNA circularization via RT-PCR. After the ribozyme reaction, reverse transcription was applied with or without urea, followed by PCR. Lane 1: RT mixture contained 450 mM urea, 2: RT mixture without urea, 3: DNA size standard Gene Ruler ULR bottom up: 50, 75, 100, 150, and 200 bp, and 4: additional length control: 90 bp. The used gels consist of polyacrylamide (15% (v/v) containing 7 M urea). All nucleic acids are stained with SYBR®Gold

218

Cathrin E. Hansen et al.

typical reaction time for the following experiments from 2 h to 30 min. Notably, an increase in the cleavage duration, even overnight, did not change the amount of cleavage product. However, storing frozen RNA and thawing it may lead to complete cleavage of the in vitro transcript, so the following experiments were done with both full or partial cleavage products, resulting in the same outcome. The cleavage steps form the chemical prerequisite for the success of circularization: providing the required free functional groups, an OH-group, and a 2′,3′-cyclic phosphate for the following ligation step. Moreover, the principle of structural control of the ligation activity depends on the engineered CRZ-3 system and the corresponding oligonucleotide, the activator (Fig. 2, gray structure). The binding of the activator to the complementary RNA sequence in the inactive hairpin construct shifts the conformation to its active state. To improve the activity of the ribozyme, we analyzed the effect of different activator concentrations on the efficiency of total RNA cleavage and subsequent RNA ligation. Our results demonstrate that the cleavage reaction occurs but is incomplete. Hence, some cleavage product (l-89mer) becomes visible (lower band), but ligation did not occur (Fig. 3b). The yield of cleavage product did not increase with rising activator concentration. We defined a tenfold excess of the activator over the ribozyme for the following assays to saturate all ligationinactive RNAs. Considering the topology and natural electrostatic repulsion of the RNA and the activator, which is designed to form a second helix-loop-helix domain, it seems natural to test different cation concentrations. Interestingly, varying MgCl2 concentrations have been shown to stabilize the structural transitions but not facilitate the hairpin ribozyme reactions per se. Also, here, we tested a wide range of MgCl2 from 0.5 to 500 mM, not resulting in a ligated product. In contrast, different polycations may enhance proper Watson-Crick base pairing [74]. Hence, the reaction mix was supplemented with different concentrations of the non-toxic polycation spermine (Fig. 3c), which has been reported to aid structural transitions of RNA [75]. With rising spermine concentration, several effects became visible: • The amount of cleavage products decreased • Ligated (circular) RNA appears, as known from former experiments with CRZ-2 above the transcript and cleavage products in denaturing polyacrylamide gels. • The total amount of ribozyme signal intensity decreased concentration-dependent.

Directed Circularization of a Short RNA

219

A concentration of at least 2 mM spermine is required to obtain the circular RNA, which migrates slower than its linear forms. At 5 mM spermine as well as at higher concentrations, the circular RNA was detected exclusively. Comparing the total band intensity of the RNA species visualized in the polyacrylamide gel (Fig. 3c), the signal seems to decrease with increasing concentration of spermine. This may be an artifact due to spermine partially blocking the RNA from SYBR™Gold, which was used for visualization of the RNA species or the circular RNA may not bind the dye as well as linear RNA. Spermine acts as a weak, general base and may contribute to RNA hydrolysis prior to electrophoresis. A slower migration pattern features circular RNA in denaturing polyacrylamide gels compared with linear RNAs with the same length. To further differentiate between a linear oligomer and a circular RNA, we used RT-PCR (Fig. 3d). We designed two sets of primers (Fig. 4a), one for the detection of the circular RNA (c-89mer; forward: 5′-GGA GAA AGA TAT TAC TTC GCG CGT GGT-3′ and reverse: 5′-AGT TCA CTT CTC TCT TTG ACG GAA CCC-3′), which does not detect the l-89mer as the cDNA amplicon includes the ligation site (Fig. 4b left). The second primer set can bind cDNAs derived from RNA without dependency on the template’s topology and linkage state (forward: 5´-GAC GGG GTT CCG TCA AAG-3′ and reverse: 5′-AAA GAG GAC TGT GAG GGG-3′). This second primer set flanks the cDNA after reverse transcription at the natural termini (Fig. 4b right). We acknowledge that if the 3′-end and 5′-end of linear RNA are adjacent but not covalently closed (pseudo-circular), the reverse transcriptase may synthesize over the open ligation site. To avoid false-positive results, we disintegrate these non-covalent linkages while keeping the reverse primer hybridized with the RNA. We analyzed the primer-RNA annealing using varying concentrations of urea. We found a stable interaction up to a concentration of 1.8 M urea, whereas pseudo-circularization was already dissolved at 450 mM urea. Consequently, we added low urea concentrations (450 mM) to cDNA synthesis mixtures to avoid potential intramolecular interactions and remained with specific results for circularized RNA (Fig. 3d lane 1). Both the PCR product formation with primer set 1 per se and the product length strongly indicate successful circularization of the RNA (Fig. 4d). A minor second product (Fig. 3d, lane 1), which can be interpreted as an additional unspecific PCR product, was obtained due to multiple rounds of the reverse transcriptase or indicating a multimerization as described previously [62]. When loading a larger amount of the DNA onto the gel, we observed a dark smear that points toward a random variation than a defined DNA species. This also indicates that a circular RNA served as a template for cDNA synthesis as each reverse transcription reaction was stopped at an individual synthesis stage, leading to DNA concatemers. The

220

Cathrin E. Hansen et al.

Fig. 4 Product verification and validation using RT-PCR. The scheme shows the experimental setup if the circular product (gray box) is the template for the cDNA synthesis compared with linear products (b, right), which were cleaved but not ligated. (a) Identification of a circular RNA template: Reverse primer 1 (RP 1) binds in the center of the RNA, meaning not at the former termini “ligation site,” and therefore, cDNA synthesis results in full-length transcripts if the template is circular (89 nt). (b) Control reactions. Using Reverse primer 1 (RP 1) in a reaction mixture that contains a linear RNA product, the resulting cDNA is shorter (panel B, left). It lacks the binding site for the forward primer for a subsequent polymerase chain reaction (PCR) step, meaning if the template RNA was not circularized, the resulting cDNA could not be amplified. Reverse primer 2 (RP2) binds terminally to the RNA, resulting in the full cDNA transcript, independent of whether the template RNA was circularized

concatemer formation somehow compares with rolling circle replications in vivo. We developed a rapid, low-priced, and non-toxic technique for controlled RNA circularization resulting in physiological 3′,5′-phosphodiester linkage using an engineered hairpin ribozyme. To obtain the circular RNA in a controlled fashion, the transcript was designed to be incapable of ligating itself. An external oligonucleotide (activator) is required to form a sufficiently stable conformation mimicking the active hairpin ribozyme elements. The activator invades the RNA stretch without defined secondary structures. It has been shown in the past that hairpin ribozyme activity can be induced by external co-factors [76, 77]. Circularization

Directed Circularization of a Short RNA

221

becomes irreversible with high yields when the activator is separated from the circular RNA (Fig. 2), which is readily done using size exclusion chromatography. Our results serve as proof of principle, showing that our engineered system can controllably synthesize circular RNA by adding an external activator together with spermine. The intensity of the detected product is lower than the corresponding transcript due to either slight hydrolysis because of the mild alkaline pH or, most likely, due to a reduced binding capacity between dye and circRNA. The reaction relies on a non-enzymatic RNA-based system, which requires merely an RNA activator, spermine, and magnesium ions, but no proteins or non-physiological linkers. One can apply a short version or a long protocol if the reaction needs to be done overnight. The short-cut version provides a ligation product and a brief reaction time. By performing the cleavage and ligation reaction as separate steps, e.g., freezing and thawing the transcript, cleavage can be even faster and may contribute to an increased yield of circular RNA. Furthermore, the application in vivo by coding the desired sequence in a DNA plasmid is certainly possible. Another promising possibility to associate the ligation-inactive ribozyme with its activator and spermine is an encapsulation strategy. Peng et al. showed that joining RNAs in vesicles promotes RNA–RNA association and stabilizes tertiary folding [78]. Future analyses may include inserting various sequences into the variable range of the RNA (green dotted and purple regions in Fig. 2), including any desired sequence, as long as the catalytic domains remain operational.

5 5.1

Notes GMP Priming

5.2 Polyacrylamide Gel Electrophoresis

The incorporation of the GMP compared with the GTP is advantageous because, as experience taught us, the triphosphate’s charge interferes with the formation of catalytically active structures. In contrast, the monophosphate does not disturb the functionality of the ribozyme. Alternatively, an in vitro transcription can be carried out without GMP and using phosphatases to generate a phosphatefree 5′-terminus. 1. To keep the RNA stable at room temperatures above 27 °C, the gel electrophoresis was performed at 4–8 °C in a refrigerator room. The migration is slower at low temperatures and takes 7 h. 2. When isolating the products by denaturing polyacrylamide gel electrophoresis, we could differentiate the GMP from the GTP products by altering migration behavior. The low charge of the

222

Cathrin E. Hansen et al.

monophosphate at the 5′-terminus causes the GMP product to migrate slower than the GTP product. 3. All elution steps may elute the gel pieces in 0.3 mM sodium acetate buffer (pH = 7). However, we noted that the salt concentration is very high (including urea from the gel), inhibiting RNA precipitation in the next step. To have a series of three elution steps consisting of 0.3 mM sodium acetate first for a good buffer impact, and slightly keep the buffer effect in the two following steps using 0.1 mM sodium acetate buffer turned out to give the highest RNA yields. Water usage for elution was also possible, but we recommend using a buffer with pH = 7 as the 2′,3′-cyclic phosphates are sensitive even to minimal pH shifts. 5.3 Ribozyme Reaction

1. Instead of purchasing the activator from Biomers GmbH, it can also be synthesized using Gene Assembler Special from Phamarcia AG in its own lab. 2. The ribozyme reaction can be stopped in three ways: • Add loading buffer to the reaction mixture (vol 1:1) and freeze immediately. • Add loading buffer to the reaction mixture (vol 1:1) and perform gel electrophoresis immediately. • Add 2.5 volumes of ice-cold ethanol and precipitate the RNA overnight.

Acknowledgments We thank Georg Sczakiel for inspiring discussion of the manuscript and Sarah Graff for RT primer analyses. References 1. Harland R, Misher L (1988) Stability of RNA in developing Xenopus embryos and identification of a destabilizing sequence in TFIIIA messenger RNA. Development 102(4):837–852. https://doi.org/10.1242/dev.102.4.837 2. Puttaraju M, Been MD (1995) Generation of nuclease resistant circular RNA decoys for HIV-Tat and HIV-Rev by autocatalytic splicing. Nucleic Acids Symp Ser 33:152–155 3. Enuka Y, Lauriola M, Feldman ME, Sas-Chen A, Ulitsky I, Yarden Y (2016) Circular RNAs are long-lived and display only minimal early alterations in response to a growth factor. Nucleic Acids Res 44(3):1370–1383. https://doi.org/10.1093/nar/gkv1367

4. Cocquerelle C, Mascrez B, Hetuin D, Bailleul B (1993) Mis-splicing yields circular RNA molecules. FASEB J 7(1):155–160 5. Mackie GA (1998) Ribonuclease E is a 5′-end-dependent endonuclease. Nature 395(6703):720–723. https://doi.org/10. 1038/27246 6. Capel B, Swain A, Nicolis S, Hacker A, Walter M, Koopman P et al (1993) Circular transcripts of the testis-determining gene Sry in adult mouse testis. Cell 73(5):1019–1030 7. Kjems J, Garrett RA (1988) Novel splicing mechanism for the ribosomal RNA intron in the archaebacterium Desulfurococcus mobilis. Cell 54(5):693–703

Directed Circularization of a Short RNA 8. Lasda E, Parker R (2014) Circular RNAs: diversity of form and function. RNA (New York, NY). 20(12):1829–1842. https://doi.org/10.1261/rna.047126.114 9. Memczak S, Jens M, Elefsinioti A, Torti F, Krueger J, Rybak A et al (2013) Circular RNAs are a large class of animal RNAs with regulatory potency. Nature 495(7441): 3 3 3 – 3 3 8 . h t t p s : // d o i . o r g / 1 0 . 1 0 3 8 / nature11928 10. Molina-Sanchez MD, Martinez-Abarca F, Toro N (2006) Excision of the Sinorhizobium meliloti group II intron RmInt1 as circles in vivo. J Biol Chem 281(39):28737–28744. https://doi.org/10.1074/jbc.M602695200 11. Salzman J, Gawad C, Wang PL, Lacayo N, Brown PO (2012) Circular RNAs are the predominant transcript isoform from hundreds of human genes in diverse cell types. PLoS One 7(2):e30733. https://doi.org/10.1371/jour nal.pone.0030733 12. Wang PL, Bao Y, Yee MC, Barrett SP, Hogan GJ, Olsen MN et al (2014) Circular RNA is expressed across the eukaryotic tree of life. PLoS One 9(6):e90859. https://doi.org/10. 1371/journal.pone.0090859 13. Ramon YCS, Segura MF, Hummer S (2019) Interplay between ncRNAs and cellular communication: a proposal for understanding cellspecific signaling pathways. Front Genet 10: 281. https://doi.org/10.3389/fgene.2019. 00281 14. Danan M, Schwartz S, Edelheit S, Sorek R (2012) Transcriptome-wide discovery of circular RNAs in archaea. Nucleic Acids Res 40(7): 3131–3142. https://doi.org/10.1093/nar/ gkr1009 15. Liu CX, Chen LL (2022) Circular RNAs: characterization, cellular roles, and applications. Cell 185(12):2016–2034. https://doi.org/ 10.1016/j.cell.2022.04.021 16. Yang J, Qi M, Fei X, Wang X, Wang K (2022) Hsa_circRNA_0088036 acts as a ceRNA to promote bladder cancer progression by sponging miR-140-3p. Cell Death Dis 13(4):322. https://doi.org/10.1038/s41419-02204732-w 17. You J, Chen Y, Chen D, Li Y, Wang T, Zhu J et al (2023) Circular RNA 0001789 sponges miR-140-3p and regulates PAK2 to promote the progression of gastric cancer. J Transl Med 21(1):83. https://doi.org/10.1186/s12967022-03853-2 18. Li H, Zheng S, Wan T, Yang X, Ouyang Y, Xia H et al (2022) Circular RNA circ_0000212 accelerates cervical cancer progression by acting as a miR-625-5p sponge to upregulate

223

PTP4A1. Anti-Cancer Drugs 34:659. https:// doi.org/10.1097/cad.0000000000001435 19. Wu B, Xia L, Zhang S, Jin K, Li L, Sun C et al (2023) circRNA-SMO upregulates CEP85 to promote proliferation and migration of glioblastoma via sponging miR-326. Histol Histopathol 38:18587. https://doi.org/10.14670/ hh-18-587 20. Krishnan RH, Sadu L, Akshaya RL, Gomathi K, Saranya I, Das UR et al (2023) Circ_CUX1/miR-130b-5p/p300 axis for parathyroid hormone-stimulation of Runx2 activity in rat osteoblasts: a combined bioinformatic and experimental approach. Int J Biol Macromol 225:1152–1163. https://doi.org/ 10.1016/j.ijbiomac.2022.11.176 21. Liu C, Liang D (2023) Letter to the editor: there may be a misunderstanding about the function of circular RNA as miRNA sponges. Hepatology (Baltimore, Md). https://doi. org/10.1097/hep.0000000000000233 22. Ashwal-Fluss R, Meyer M, Pamudurti NR, Ivanov A, Bartok O, Hanan M et al (2014) circRNA biogenesis competes with pre-mRNA splicing. Mol Cell 56(1):55–66. https://doi.org/10.1016/j.molcel.2014. 08.019 23. Okholm TLH, Sathe S, Park SS, Kamstrup AB, Rasmussen AM, Shankar A et al (2020) Transcriptome-wide profiles of circular RNA and RNA-binding protein interactions reveal effects on circular RNA biogenesis and cancer pathway expression. Genome Med 12(1):112. https://doi.org/10.1186/s13073-02000812-8 24. Schreiner S, Didio A, Hung LH, Bindereif A (2020) Design and application of circular RNAs with protein-sponge function. Nucleic Acids Res 48(21):12326–12335. https://doi. org/10.1093/nar/gkaa1085 25. Barbagallo D, Caponnetto A, Brex D, Mirabella F, Barbagallo C, Lauretta G et al (2019) CircSMARCA5 regulates VEGFA mRNA splicing and angiogenesis in glioblastoma Multiforme through the binding of SRSF1. Cancers 11(2). https://doi.org/10. 3390/cancers11020194 26. Qu L, Yi Z, Shen Y, Lin L, Chen F, Xu Y et al (2022) Circular RNA vaccines against SARSCoV-2 and emerging variants. Cell 185(10): 1728–44.e16. https://doi.org/10.1016/j. cell.2022.03.044 27. Legnini I, Di Timoteo G, Rossi F, Morlando M, Briganti F, Sthandier O et al (2017) Circ-ZNF609 is a circular RNA that can be translated and functions in Myogenesis. Mol Cell 66(1):22–37.e9. https://doi.org/10. 1016/j.molcel.2017.02.017

224

Cathrin E. Hansen et al.

28. Gao X, Xia X, Li F, Zhang M, Zhou H, Wu X et al (2021) Circular RNA-encoded oncogenic E-cadherin variant promotes glioblastoma tumorigenicity through activation of EGFRSTAT3 signalling. Nat Cell Biol 23(3): 278–291. https://doi.org/10.1038/s41556021-00639-4 29. Li Y, Chen B, Zhao J, Li Q, Chen S, Guo T et al (2021) HNRNPL circularizes ARHGAP35 to produce an oncogenic protein. Adv Sci (Weinheim, Baden-Wurttemberg, Germany) 8(13): 2001701. https://doi.org/10.1002/advs. 202001701 30. Jiang T, Xia Y, Lv J, Li B, Li Y, Wang S et al (2021) A novel protein encoded by circMAPK1 inhibits progression of gastric cancer by suppressing activation of MAPK signaling. Mol Cancer 20(1):66. https://doi.org/10. 1186/s12943-021-01358-y 31. Zhang Y, Jiang J, Zhang J, Shen H, Wang M, Guo Z et al (2021) CircDIDO1 inhibits gastric cancer progression by encoding a novel DIDO1-529aa protein and regulating PRDX2 protein stability. Mol Cancer 20(1): 101. https://doi.org/10.1186/s12943-02101390-y 32. Chen CY, Sarnow P (1995) Initiation of protein synthesis by the eukaryotic translational apparatus on circular RNAs. Science (New York, NY) 268(5209):415–417. https://doi.org/10.1126/science.7536344 33. Jeck WR, Sorrentino JA, Wang K, Slevin MK, Burd CE, Liu J et al (2013) Circular RNAs are abundant, conserved, and associated with ALU repeats. RNA (New York, NY) 19(2):141–157. https://doi.org/10.1261/rna.035667.112 34. Piwecka M, Glazar P, Hernandez-Miranda LR, Memczak S, Wolf SA, Rybak-Wolf A et al (2017) Loss of a mammalian circular RNA locus causes miRNA deregulation and affects brain function. Science (New York, NY) 357(6357). https://doi.org/10.1126/sci ence.aam8526 35. Okholm TLH, Nielsen MM, Hamilton MP, Christensen LL, Vang S, Hedegaard J et al (2017) Circular RNA expression is abundant and correlated to aggressiveness in early-stage bladder cancer. NPJ Genom Med 2:36. https://doi.org/10.1038/s41525-0170038-z 36. Zhang X, Xu Y, Qian Z, Zheng W, Wu Q, Chen Y et al (2018) circRNA_104075 stimulates YAP-dependent tumorigenesis through the regulation of HNF4a and may serve as a diagnostic marker in hepatocellular carcinoma. Cell Death Dis 9(11):1091. https://doi.org/10. 1038/s41419-018-1132-6

37. Ning L, Long B, Zhang W, Yu M, Wang S, Cao D et al (2018) Circular RNA profiling reveals circEXOC6B and circN4BP2L2 as novel prognostic biomarkers in epithelial ovarian cancer. Int J Oncol 53(6):2637–2646. https://doi. org/10.3892/ijo.2018.4566 38. Weng W, Wei Q, Toden S, Yoshida K, Nagasaka T, Fujiwara T et al (2017) Circular RNA ciRS-7-a promising prognostic biomarker and a potential therapeutic target in colorectal cancer. Clin Cancer Res 23(14): 3918–3928. https://doi.org/10.1158/ 1078-0432.Ccr-16-2541 39. Ahmed I, Karedath T, Andrews SS, Al-Azwani IK, Mohamoud YA, Querleu D et al (2016) Altered expression pattern of circular RNAs in primary and metastatic sites of epithelial ovarian carcinoma. Oncotarget 7(24): 36366–36381. https://doi.org/10.18632/ oncotarget.8917 40. Zou T, Wang PL, Gao Y, Liang WT (2018) Circular RNA_LARP4 is lower expressed and serves as a potential biomarker of ovarian cancer prognosis. Eur Rev Med Pharmacol Sci 22(21):7178–7182. https://doi.org/10. 26355/eurrev_201811_16250 41. Lei X, Bian C (2020) Integrating random walk with restart and k-nearest neighbor to identify novel circRNA-disease association. Sci Rep 10(1):1943. https://doi.org/10.1038/ s41598-020-59040-0 42. Shi L, Liu B, Shen DD, Yan P, Zhang Y, Tian Y et al (2021) A tumor-suppressive circular RNA mediates uncanonical integrin degradation by the proteasome in liver cancer. Sci Adv 7(13). https://doi.org/10.1126/sciadv.abe5043 43. Xu L, Lyu M, Yang S, Zhang J, Yu D (2022) CircRNA expression profiles of breast cancer and construction of a circRNA-miRNAmRNA network. Sci Rep 12(1):17765. https://doi.org/10.1038/s41598-02221877-y 44. Tao W, Luo ZH, He YD, Wang BY, Xia TL, Deng WM et al (2023) Plasma extracellular vesicle circRNA signature and resistance to abiraterone in metastatic castration-resistant prostate cancer. Br J Cancer 128:1320. https://doi.org/10.1038/s41416-02302147-8 45. Qadir J, Wen SY, Yuan H, Yang BB (2022) CircRNAs regulate the crosstalk between inflammation and tumorigenesis: the bilateral association and molecular mechanisms. Mol Ther 31:1514. https://doi.org/10.1016/j. ymthe.2022.12.005 46. Rahimi K, Venø MT, Dupont DM, Kjems J (2021) Nanopore sequencing of brain-derived full-length circRNAs reveals circRNA-specific

Directed Circularization of a Short RNA exon usage, intron retention and microexons. Nat Commun 12(1):4825. https://doi.org/ 10.1038/s41467-021-24975-z 47. Wu DP, Zhao YD, Yan QQ, Liu LL, Wei YS, Huang JL (2023) Circular RNAs: emerging players in brain aging and neurodegenerative diseases. J Pathol 259(1):1–9. https://doi. org/10.1002/path.6021 48. Mehta SL, Chokkalla AK, Bathula S, Arruri V, Chelluboina B, Vemuganti R (2023) CDR1as regulates α-synuclein-mediated ischemic brain damage by controlling miR-7 availability. Mol Ther Nucleic Acids 31:57–67. https://doi. org/10.1016/j.omtn.2022.11.022 49. Zhang X, Yang S, Han S, Sun Y, Han M, Zheng X et al (2023) Differential methylation of circRNA m6A in an APP/PS1 Alzheimer’s disease mouse model. Mol Med Rep 27(2). https://doi.org/10.3892/mmr.2023.12942 50. Zheng H, Huang S, Wei G, Sun Y, Li C, Si X et al (2022) CircRNA Samd4 induces cardiac repair after myocardial infarction by blocking mitochondria-derived ROS output. Mol Ther 30(11):3477–3498. https://doi.org/10. 1016/j.ymthe.2022.06.016 51. Su J, Hu Y, Cheng J, Li Z, Li J, Zheng N et al (2023) Comprehensive analysis of the RNA transcriptome expression profiles and construction of the ceRNA network in heart failure patients with sacubitril/valsartan therapeutic heterogeneity after acute myocardial infarction. Eur J Pharmacol 944:175547. https://doi. org/10.1016/j.ejphar.2023.175547 52. Li G, Zhu S, Zeng J, Yu Z, Meng F, Tang Z et al (2022) Arterial Pulsatility augments microcirculatory perfusion and maintains the endothelial integrity during extracorporeal membrane oxygenation via hsa_circ_0007367 upregulation in a canine model with cardiac arrest. Oxidative Med Cell Longev 2022: 1630918. https://doi.org/10.1155/2022/ 1630918 53. Liu Y, Dong Y, Dong Z, Song J, Zhang Z, Liang L et al (2021) Expression profiles of circular RNA in aortic vascular tissues of spontaneously hypertensive rats. Front Cardiovasc Med 8:814402. https://doi.org/10.3389/ fcvm.2021.814402 54. Shi P, Ji H, Zhang H, Yang J, Guo R, Wang J (2020) circANRIL reduces vascular endothelial injury, oxidative stress and inflammation in rats with coronary atherosclerosis. Exp Ther Med 20(3):2245–2251. https://doi.org/10.3892/ etm.2020.8956 55. Wen ZJ, Xin H, Wang YC, Liu HW, Gao YY, Zhang YF (2021) Emerging roles of circRNAs in the pathological process of myocardial infarction. Mol Ther Nucleic Acids 26:828–

225

848. https://doi.org/10.1016/j.omtn.2021. 10.002 56. Seephetdee C, Bhukhai K, Buasri N, Leelukkanaveera P, Lerdwattanasombat P, Manopwisedjaroen S et al (2022) A circular mRNA vaccine prototype producing VFLIP-X spike confers a broad neutralization of SARSCoV-2 variants by mouse sera. Antivir Res 204: 105370. https://doi.org/10.1016/j.antiviral. 2022.105370 57. Chen X, Lu Y (2021) Circular RNA: biosynthesis in vitro. Front Bioeng Biotechnol 9: 787881. https://doi.org/10.3389/fbioe. 2021.787881 58. Mu¨ller S, Appel B (2017) In vitro circularization of RNA. RNA Biol 14(8):1018–1027. https://doi.org/10.1080/15476286.2016. 1239009 59. Warminski M, Kowalska J, Jemielity J (2020) Solid-phase synthesis of RNA 5′-Azides and their application for labeling, ligation, and cyclization via click chemistry. Curr Protoc Nucleic Acid Chem 82(1):e112. https://doi. org/10.1002/cpnc.112 60. Micura R (1999) Cyclic Oligoribonucleotides (RNA) by solid-phase synthesis. Chem Eur J 5(7):2077–2082. https://doi.org/10.1002/ (SICI)1521-3765(19990702)5:73.0.CO;2-U 61. Beaudry D, Perreault JP (1995) An efficient strategy for the synthesis of circular RNA molecules. Nucleic Acids Res 23(15):3064–3066. https://doi.org/10.1093/nar/23.15.3064 62. Petkovic S, Muller S (2013) RNA selfprocessing: formation of cyclic species and concatemers from a small engineered RNA. FEBS Lett 587(15):2435–2440. https://doi.org/ 10.1016/j.febslet.2013.06.013 63. Puttaraju M, Been MD (1992) Group I permuted intron-exon (PIE) sequences self-splice to produce circular exons. Nucleic Acids Res 20(20):5357–5364 64. Ford E, Ares M Jr (1994) Synthesis of circular RNA in bacteria and yeast using RNA cyclase ribozymes derived from a group I intron of phage T4. Proc Natl Acad Sci U S A 91(8): 3117–3121. https://doi.org/10.1073/pnas. 91.8.3117 65. Petkovic S, Badelt S, Block S, Flamm C, Delcea M, Hofacker I et al (2015) Sequencecontrolled RNA self-processing: computational design, biochemical analysis, and visualization by AFM. RNA (New York, NY) 21(7): 1249–1260. https://doi.org/10.1261/rna. 047670.114 66. Wesselhoeft RA, Kowalski PS, Anderson DG (2018) Engineering circular RNA for potent

226

Cathrin E. Hansen et al.

and stable translation in eukaryotic cells. Nat Commun 9(1):2629. https://doi.org/10. 1038/s41467-018-05096-6 67. Rausch JW, Heinz WF, Payea MJ, Sherpa C, Gorospe M, Le Grice SFJ (2021) Characterizing and circumventing sequence restrictions for synthesis of circular RNA in vitro. Nucleic Acids Res 49(6):e35. https://doi.org/10. 1093/nar/gkaa1256 68. Weinberg CE, Olzog VJ, Eckert I, Weinberg Z (2021) Identification of over 200-fold more hairpin ribozymes than previously known in diverse circular RNAs. Nucleic Acids Res 49(11):6375–6388. https://doi.org/10. 1093/nar/gkab454 69. Hieronymus R, Godehard SP, Balke D, Mu¨ller S (2016) Hairpin ribozyme mediated RNA recombination. Chem Commun (Camb) 52(23):4365–4368. https://doi.org/10. 1039/c6cc00383d 70. Hieronymus R, Zhu J, Mu¨ller S (2022) RNA self-splicing by engineered hairpin ribozyme variants. Nucleic Acids Res 50(1):368–377. https://doi.org/10.1093/nar/gkab1239 71. Hampel A, Tritz R (1989) RNA catalytic properties of the minimum (-)sTRSV sequence. Biochemistry 28(12):4929–4933 72. Harris ME, Christian EL (1999) Use of circular permutation and end modification to position photoaffinity probes for analysis of RNA structure. Methods (San Diego, Calif) 18(1):51–59. https://doi.org/10.1006/meth.1999.0756

73. Pieper S, Vauleon S, Muller S (2007) RNA selfprocessing towards changed topology and sequence oligomerization. Biol Chem 388(7): 743–746. https://doi.org/10.1515/bc. 2007.067 74. Nesbitt S, Hegg LA, Fedor MJ (1997) An unusual pH-independent and metal-ion-independent mechanism for hairpin ribozyme catalysis. Chem Biol 4(8):619–630. https://doi. org/10.1016/s1074-5521(97)90247-7 75. Nakano S, Kirihata T, Fujii S, Sakai H, Kuwahara M, Sawai H et al (2007) Influence of cationic molecules on the hairpin to duplex equilibria of self-complementary DNA and RNA oligonucleotides. Nucleic Acids Res 35(2):486–494. https://doi.org/10.1093/ nar/gkl1073 76. Komatsu Y, Nobuoka K, Karino-Abe N, Matsuda A, Ohtsuka E (2002) In vitro selection of hairpin ribozymes activated with short oligonucleotides. Biochemistry 41(29): 9090–9098 77. Vauleon S, Muller S (2003) External regulation of hairpin ribozyme activity by an oligonucleotide effector. Chembiochem 4(2–3):220–224. https://doi.org/10.1002/cbic.200390035 78. Peng H, Lelievre A, Landenfeld K, Mu¨ller S, Chen IA (2022) Vesicle encapsulation stabilizes intermolecular association and structure formation of functional RNA and DNA. Curr Biol 32(1):86–96.e6. https://doi.org/10. 1016/j.cub.2021.10.047

Chapter 13 Engineering Synthetic circRNAs for Efficient CNS Expression Katie N. Clements, Trevor J. Gonzalez, and Aravind Asokan Abstract Circular RNAs (circRNAs) have recently emerged as a promising modality for gene and RNA-based therapies. They are more stable than their linear counterpart and can be designed for efficient expression in different cell and tissue types. In this chapter, we developed different backsplicing circRNA cassettes that can enable efficient gene expression in various cell and tissue types. Furthermore, we packaged cassettes encoding circRNAs into adeno-associated viral (AAV) vectors that can be delivered via intracerebroventricular (ICV) injections to achieve expression in murine brain tissue. We provide detailed methods for the design of backsplicing circRNAs, circRNA detection, and generation of AAV-circRNA vectors for CNS dosing and expression in mice. Keywords Backsplicing, circRNA, Gene therapy, CNS, AAV vectors

1

Introduction Circular RNAs (circRNAs) have recently emerged as a promising modality for gene and RNA therapies [1–3]. Studies reveal endogenous circRNAs are involved in a variety of functions such as modulating transcription [4, 5], regulating splicing [6, 7], sponging RNAs and proteins [8, 9], and encoding translational products [10, 11]. CircRNAs offer several potential advantages over linear mRNA transcripts. For instance, circRNAs are covalently closed and lack a 5′ cap or polyadenylation (polyA) signal, making them less susceptible to degradation and significantly more stable than their linear counterpart [12]. Moreover, circRNA expression can be precisely controlled in specific tissues and cell types [13]. Thus, engineering highly efficient circRNAs for targeted therapeutic delivery is an exciting avenue to explore. Circularization of RNAs can arise from multiple mechanisms: self-splicing introns [14], tRNA intronic splicing [15], and through spliceosome-mediated backsplicing [16]. Our lab has engineered

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_13, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

227

228

Katie N. Clements et al.

Fig. 1 Designing backsplicing circular RNA (circRNA)-encoding transgene cassettes. (a) Transgene cassette consisting of a promoter (blue), backsplicing intronic sequences (beige) flanking a split open reading frame (ORF) (green) and Internal Ribosome Entry Site (IRES) (gray), followed by a polyadenylation signal (burgundy). Complementary Alu sequences (white boxes) within the backsplicing introns are essential for the splice donor and acceptor at the beginning and end of the intron, respectively, to undergo backsplicing. Once circularized, the IRES recruits ribosomal machinery for translation into a protein product. (b) Components that can be engineered to generate cell/tissue-specific backsplicing circRNAs at different efficiency. Efficiency of two promoters (cytomegalovirus; CMV and elongation factor 1 alpha; Ef1α) in two different cell types (HEK293 and HUH7). Backsplicing and translation efficiency can also be engineered using different introns and IRES/polyA elements as described previously [18]

efficient synthetic backsplicing circRNAs, which consist of backsplicing introns flanking either a non-coding or a coding sequence containing an Internal Ribosome Entry Site (IRES) for circRNAspecific translation [17] (Fig. 1a). Although circRNAs themselves lack a 5′ cap or polyA signal, circRNA-expressing plasmids can be engineered to include promoters and polyadenylation signals, which allow for proper transcription of the circRNA and provide additional levels of expression control. Here we use different assays to evaluate how altering the promoter and backsplicing introns can help achieve increased efficiency of circRNA formation in various tissue and cell types, providing specific examples for engineering efficient circRNA constructs [18]. Once engineered, there are several delivery methods for circRNA constructs. These include transfection of plasmids for in vitro studies; delivery of circRNAs through non-viral delivery methods, such as lipid nanoparticles [19]; and virus-based approaches such as lentivirus [20] and adeno-associated virus (AAV) [18]. The focus of this contribution is to enable circRNA expression using AAV vectors in the brain.

AAV-circRNAs for CNS expression

229

To generate an AAV-circRNA vector, it is necessary to clone the circRNA plasmid between inverted terminal repeats (ITRs) recognized by the AAV packaging system [18]. There are several methods by which AAV can be delivered to the brain in vivo, including intravenous, retro-orbital, or intracerebroventricular (ICV) injections [21–23]. Overall, the current report describes methods to engineer circRNAs with synthetic backsplicing introns, carry out assays to validate and quantify circRNA expression, and achieve circRNA expression in the brain using an AAV vector-based approach.

2

Materials Plasmids and Transfections 1. For transformation of the Gibson Assembled construct, make LB-agar + amp plates: Measure out 12.5 g LB Broth (25 g/L), 7.5 g Agar (12 g/L), and bring to 500 mL with deionized H2O (dH2O). Autoclave on liquid cycle and bring to 60 °C in a water bath or incubator. Once the temperature is around 60 ° C, add 1:1000 dilution of ampicillin (100 mg/mL concentration). Pour plates and allow to solidify before use. 2. Super Optimal Broth: Dissolve 18 g of glucose into 90 mL dH2O. After the sugar has dissolved, adjust the volume of the solution to 100 mL with dH2O and sterilize by passing through a 0.22 μm filter. 3. PEI Stock Solution (1 mg/mL): Combine 500 mg of PEI MAX (MW 40,000) (Polysciences) with 500 mL of 1× dPBS, adjust the pH to 6.9–7.1, then aliquot into 15 mL conical tubes, and store at 4 °C until use. Molecular Kits 1. IBI Miniprep kit (IBI scientific). 2. Turbo DNA-free Kit (Invitrogen). 3. RNase R kit (Abcam). 4. Trizol (ThermoFisher). 5. High Capacity RNA-to-DNA kit (ThermoFisher). Recombinant AAV Vector Production 1. 40% Polyethylene glycol (PEG) with 2.5 M NaCl stock solution: Pour 500 mL deionized water into a 1 L glass bottle with a stir bar. Place on a magnetic hot plate. Add 146.1 g NaCl and heat to 75 °C while stirring. Add 200 g PEG8000 and allow to dissolve. Add remaining 200 g PEG8000 and allow to dissolve. Bring to 1 L with deionized water. The final concentration of PEG stock is 40% PEG8000 and 2.5 M NaCl. Store at RT.

230

Katie N. Clements et al.

2. Iodixanol Density Gradient (17%, 25%, 40%, and 60%): 17%— 5 mL 10× PBS, 50 μL 1 M MgCl2, 50 μL 2.5 M KCl, 10 mL 5 M NaCl, 12.5 mL Optiprep, and 22.4 mL dH2O. 25%— 5 mL 10× PBS, 50 μL 1 M MgCl2, 50 μL 2.5 M KCl, 20 mL Optiprep, 100 μL 0.5% Phenol Red, and 24.8 mL dH2O. 40%—5 mL 10× PBS, 50 μL 1 M MgCl2, 50 μL 2.5 M KCl, 33.3 mL Optiprep, and 11.6 mL dH2O. 60%— 50 μL 1 M MgCl2, 50 μL 2.5 M KCl, 50 mL Optiprep, and 25 μL 0.5% Phenol Red. 3. OptiPrep Density Gradient Medium (Iodixanol; SigmaAldrich). 4. DNase I resuspension buffer + DNase I: 100 mg DNase I in 10 mL of DNase resuspension buffer (50% glycerol, 20 mM Tris–HCl, 1 mM MgCl2, and 2 mM CaCl2). The final concentration of DNase I is 10 mg/mL. Store at -20 °C. 5. DNase I buffer for AAV titration: 20 mM Tris–HCl pH 7.5, 10 mM MgCl2, and 2 mM CaCl2 in molecular biology grade (MBG) water. Add Resuspended DNase I when ready to digest. When titrating, mix 89 μL DNase digestion buffer with 1 μL DNase I per sample. 6. 10% Tween Solution: Mix 5 mL of Tween-20 and 45 mL MBG water. 7. AAV formulation buffer: 1× dPBS, 1 mM MgCl2, 0.001% F-68. 8. 1× dPBS (Gibco). 9. Molecular Biology Grade Water (VWR). 10. EDTA (FisherSci). 11. LightCycler 480 SYBR Green I Master Mix (Roche). In Vivo Administration of AAV 1. Avertin (2,2,2-tribromoethanol, 0.06 mL of a 1.25% solution per gram body weight, 125–200 mg/kg; FisherSci). 2. Sodium azide (Sigma-Aldrich). 3. 10% Neutral buffered formalin (VWR). 4. 1× dPBS (Gibco). 5. Dry ice. 6. Small pieced ice cubes. 7. Cobalt Ultra Powder-Free Nitrile Examination Gloves, Microflex (VWR). 8. Lab Standard Stereotaxic Apparatus with Mouse and Neonates Adaptor (Stoelting). 9. Microinjection Syringe Pump (World Precision Instruments). 10. 10 μL Syringe Model 701 (Hamilton).

AAV-circRNAs for CNS expression

231

11. 26 s Gauge Hamilton Small Hub Removable Needles (Hamilton). 12. Microwave Heatpad, Small Animal Version (SnuggleSafe). 13. Vannas Micro Dissecting Spring Scissors (Roboz). 14. Air-Tite All-Plastic Henke-Ject Syringes (FisherSci). 15. BD 25G Needle 5/8 in. (VWR). 16. Red Tail Vein Restrainer for Mice 1 1/4 in. (Braintree Scientific). 17. BD Insulin Syringes with BD Ultra-Fine Needle 30G (BD). 18. 60 mL syringe (VWR). 19. Braintree Scientific Mouse Dissection Surgical Kit with Storage Case (VWR). 20. Surshield Safety-Winged Infusion Set (VWR). 21. Incu-line Digital Incubator IL 10 With Transparent Window (VWR). 22. Applied Biosystems ProFlex Polymerase Chain Reaction (PCR) System (ThermoFisher). 23. NanoDrop One Microvolume UV-Vis Spectrophotometer (ThermoFisher). 24. Pico 21 Microcentrifuge (ThermoFisher). 25. 1.7 mL microcentrifuge tubes (VWR). 26. Scintillation vial (VWR). Tissue Histology 1. Blocking Buffer: 5% goat serum and 1% Triton-X 100 in 1× PBS. 2. Triton X-100 (FisherSci). 3. 1× dPBS (Gibco). 4. ProLong Gold Antifade Mountant. 5. Zeiss 880 Confocal Microscope or another similar confocal microscope (Zeiss). 6. ECHO Revolve Fluorescent Microscope (ECHO). 7. Heathrow Scientific Microscope Slide Boxes (VWR). 8. Micro Cover Glasses (VWR). 9. 2D Platform Rocker (Poseidon). 10. VECTASTAIN Elite ABC-HRP Peroxidase (Rabbit IgG) Kit (Vector Laboratories). 11. ImmPACT DAB Substrate Peroxidase (HRP) (Vector Laboratories). 12. 75 × 25 mm VistaVision HistoBond Adhesive Microscope Slides (VWR).

232

Katie N. Clements et al.

13. 15 mL conical tubes (VWR). 14. Image J (https://imagej.nih.gov). 15. Prism (https://www.graphpad.com/scientific-software/ prism/) or Microsoft Office Excel (https://www.microsoft. com/en-us/microsoft-365/excel).

3

Methods

3.1 Efficient Backsplicing CircRNA 3.1.1 CircRNA Cassette Design

For circRNA cassette design, consider the following: 1. circRNAs can be coding or non-coding and users should first decide which approach is desired for their specific application. 2. If generating a coding circRNA, include an Internal Ribosome Entry Site (IRES) before the open reading frame (ORF) for protein translation (see Fig. 1a) [18]. 3. Another consideration is desired tissue and cell-type expression. This will influence promoter and backsplicing introns selection (see Table 1 for a short list of promoters and Table 2 for introns used in this study and [18] for IRES efficiencies). The circRNA cassette example used in this chapter contains a split open reading frame of EGFP and an IRES, as previously evaluated [18]. Circularization results in the formation of an in-frame open reading frame that supports the translation of fluorescent EGFP, providing an internal control for circRNA formation (Fig. 1a). In theory, any such protein product can be designed in a split format, although coding circRNA cassettes do not require a split ORF system separated by an IRES. Any conventional coding (or non-coding) region can be substituted for GFP as the region of interest.

Table 1 List of promoter and intron expression Name Promoter CMV Ef1a hSyn Intron

Cell expression

Tissue expression

Citation

Ubiquitous Ubiquitous Cortex

Ubiquitous Ubiquitous Brain

[24] [25] [26]

Brain, colon, kidney, lung, stomach Brain Ubiquitous, strong expression in liver and cardiac Heart, endothelial, testis

[28, 30, 31] [18] [18, 27, 28] [28, 29]

HIPK3

Non-specific with enhanced expression in cardiomyocytes HIPK3ΔΔ HEK293, HUH7 ZKSCAN1 HEK293, HeLa, hESCs, HCC EPHB4

Cancer cell lines

gcctcagcctctcaaagtgctaggattacagggatctatacttttcttttgagggaaaatgttggcaccgtttctagggcatattggccatttcagcttctca gtaaatatttgttaagtaattaaatgcacttgattctttattcttagccttttaacgcaatactcagaatagctgaagcaccaattaactgaaatggagatatta taaagatagttatcttctccaagggaaaaaatcatcttcatggaaattaattacttttttacaaattgtgaatttgacccttaagagttttcttcctgatatttaaaa ttgaaaaaaaaattgttgacattaatatttcttctttccttttttttcttttcctttttttttttttttttgcag gtaggtaacaactccatactttttggttgtttattaatgtgaaatttctgctaaatgaaatacttttgtgtgtgtttgtggtagaagagaccacttcagttaaa taaggaaatcaagagaggatcaatttaggttcgttttaaagagattaaaaaaaatcaagacataaaatctacccaagcaggatagaaatctccactgcaaa gttc catgccaaagacatctggttatttttatttttaatggaagacttgaaggaatgataggtgattaataatgatcaaacagaagtctttaaatgttggaaagtatt tacattaatctttgtatatatcattgggcattttagcacttgagagaaatagtttattaaagatataatcaatcatatgtaactgaacatttagaaaaattatata caggtttgagtagcccttatctgaaacttttggggccagaagtgttttggattccagatttttccggattttggaatatttgcactgccaactagttaagcac ccccaaatttgaaaattcgtttcctttgagtgtcatgtcaatgcccaaaaagtttcagatatttggatttgagatgct caacctgtataaggattcagaaagtt attctgattaatgattttaagattcagatatacaataatcccagcaacttgggaggctgaggcaggagaatca cttgaacccaggagatggaggttgcag tgagccgagatcatgccattgcactcca gcctcagcctctcaaagtgctaggattacagggatctatactacaaattgtgaatttgacccttaagagttttcttcctgatatttaaaattgaaaaaaaaatt gttgacattaatatttcttctttccttttttttcttttcctttttttttttttttttgcag gtaggtaacccaactagttaagcacccccaaatttgaaaattcgtttcctttgagtgtcatgtcaatgcccaaaaagtttcagatatttggatttgagatgct caacctgtataaggattcagaaagttattctgattaatgattttaagattcagatatacaataatcccagcaacttgggaggctgaggcaggagaatca cttgaacccaggagatggaggttgcagtgagccgagatcatgccattgcactcca agtgacagtggagattgtacagttttttcctcgatttgtcaggatttttttttttttgacggagtttaacttcttgtctcccaggtaggaagtgcagtggcgt aatctcggctcactacaacctccacctcctgggttcaagcgtttctcctgcctcagctttccgagtagctgggattacaggcgcctgccaccatgccct gctgacttttgtatttttagtagagacggggtttcaccatgttggccaggctggtcttgaactcctgaccgcaggcgattggcctgcctcggcctccc aaagtgctgagattacaggcgtgagccaccacccccggcctcaggagcgttctgatagtgcctcgatgtgctgcctcctataaagtgttagcagca cagatcactttttgtaaaggtacgtactaatgactttttttttatacttcag gtaagaagcaaggtttcatttaggggaagggaaatgattcaggacgagagtctttgtgctgctgagtgcctgtgatgaagaagcatgttagtcctgg gcaacgtagcgagaccccatctctacaaaaaatagaaaaattagccaggtatagtggcgcacacctgtgattccagctacgcaggaggctgaggt gggaggattgcttgagcccaggaggttgaggctgcagtgagctgtaatcatgccactactccaacctgggcaacacagcaaggaccctgtctca aaagctacttacagaaaagaattaggctcggcacggtagctcacacctgtaatcccagcactttgggaggctgaggcgggcagatcacttgaggt caggagtttgagaccagcctggccaacatggtgaaaccttgtctctactaaaaatatgaaaattagccaggcatggtggcacattcctgtaatcccag ctactcgggaggctgaggcaggagaatcacttgaacccaggaggtggaggttgcagtaagccgagatcgtaccactgtgctctagccttggtga

HIPK3

HIPK3ΔΔ

ZKSCAN1

(continued)

Left and right intronic sequences (5′–3′)

Intron

Table 2 Intronic sequences used for circularization

AAV-circRNAs for CNS expression 233

EPHB4

Intron

Table 2 (continued)

ccagctactcaggaggctgaggcagaagaatcattttaacccgggaggcggagattgcagtgagccaagatcgcgccactgcgctccaggcctgggt gacaccacggagacaggggtttggggctaaaagctatgagccgagcctccgagtccagtgggagttaattcccagctgacggggccctgcctgat ttctcag gtgagcaccctccctggcttctgcggccacccggagttcccacttacacccagaggccacttgggttaagaagccaggacagacagtgggtcccaggt cacctcctccagccttttcctcttgggctaagccctggtcctctgccttttcttttttttaagacagagcctcgctctgtcgcccaggctggagtgcagtg gcgcgatctcggctcattgctgtctccacctccagggttcaagcgattctcctgcctcagtctcccaagtagctggtactataggcatgcaccaccatg ctgactaatttttgtatttttagtagacacagggtttcaccatgtaggccaggctggtatcaaactcctgacctcaagtgatctccccacctcagcctccc aaagtgctggtattacaggtgtgaggcaccacgcctggccagccctctgcctttaattttccctctgggaaaggctgggctcctgggaccttcctttc ccactgccccatacagctgaaggttgtc

cagagcgagactgtcttaaaaaaaaaaaaaaaaaaaaaagaattaattaaaaatttaaaaaaaaatgaaaaaaagctgcatgcttgttttttgtttttagt tattctacattgttgtcattattaccaaatattggggaaaatacaacttacagaccaatctcaggagttaaatgttactacgaaggcaaatgaactatgc gtaatgaacctggtaggcatta

Left and right intronic sequences (5′–3′)

234 Katie N. Clements et al.

AAV-circRNAs for CNS expression 3.1.2 Synthesizing Fragments for Cassette

235

For assembly of the circRNA cassette, order each component of the circRNA cassette (promoter, backsplicing intronic sequences, sequence of interest, and polyadenylation signal) from a vendor (e.g., Twist Bioscience, IDT, and GenScript) with the following considerations: 1. If synthesizing the entire cassette inside a general plasmid backbone (e.g., pcDNA3.1), order segmented gene blocks for each component of the cassette and assemble in-house, or order smaller single-strand oligonucleotides with a reverse primer for each (usually limited to 200 nucleotide [nt] fragments) (see Note 1). 2. If synthesizing a gene block or single-strand oligonucleotide, ensure there are 20 nt flanking the 5′ and 3′ ends of the insert that overlap in sequence with the backbone, known as homology arms. Without these homology arms, the circRNA insert will not be ligated into the plasmid backbone using the Gibson Assembly. 3. If using single-strand oligonucleotides, perform a Polymerase Chain Reaction(PCR) to synthesize double-stranded oligonucleotides with reaction conditions based on the polymerasespecific manufacturer’s protocol.

3.1.3 Linearization of Backbone

Using a pcDNA 3.1 backbone map, choose a unique restriction endonuclease site to linearize the backbone. The pcDNA 3.1 backbone includes multiple cloning sites (MCS) with a variety of pre-existing restriction endonuclease sites. This backbone also includes multiple promoter and polyadenylation terminator sequences. To use existing promoter and polyA sequences in the circRNA cassette, select a restriction endonuclease site between those two sequences for linearization. 1. To linearize, add 1 μL Backbone Template DNA at 1 μg/μL concentration, 2 μL 10× reaction buffer, and 1 μL of each restriction enzyme, and bring to 20 μL with H2O (Restriction Enzymes from NEB are preferred). 2. Incubate the reaction at the appropriate digestion temperature for the enzyme used for 1 h. 3. Heat inactivate at the appropriate temperature based on the enzyme used (see Note 2 for optional clean-up steps). 4. Store at -20 °C until ready for use or perform Gibson Assembly of inserts into backbone immediately.

3.1.4 Gibson Assembly of Backbone and Inserts

1. Perform Gibson Assembly of gene fragments using the reaction conditions listed in Table 3. 2. Incubate samples at 50 °C for 60 min. 3. Following incubation, store samples on ice and begin transformation or store at -20 °C.

236

Katie N. Clements et al.

Table 3 Gibson Assembly protocol 2–3 Fragment assembly

4–6 Fragment assembly

Positive control

Total amount of fragments

0.02–0.5 pmol

0.2–1 pmol

10 μL

Gibson assembly master mix

10 μL

10 μL

10 μL

Deionized H2O

Bring to 20 μL

Bring to 20 μL

0 μL

Total volume

20 μL

20 μL

20 μL

Adapted from New England Biolabs

3.1.5

Transformation

This transformation protocol requires chemically competent cells (ordered from the vendor). 1. Remove competent cells from -80 °C and thaw on ice for 10 min. 2. Remove LB-agar (+ampicillin or appropriate antibiotic) plates from 4 °C storage and place, upside down, in a 37 °C incubator to warm. 3. Once competent cells are thawed, aliquot 50 μL into 1.5 mL tubes. 4. Pipette 2–4 μL of Gibson Assembly reaction into the competent cells. 5. Mix by flicking a few times. 6. Incubate on ice for at least 10 min. 7. After incubation, heat shock the transformation tube by placing it in a 42 °C water bath for 30 s. 8. Immediately place back on ice for 5 min. 9. Add 300 μL S.O.C. media to the 1.5 mL tube and place in a 37 °C shaker for 30 min (250 rpm). 10. Streak 30 μL of colonies on the LB-agar + amp plates and store them upside down at 37 °C overnight. 11. Next day, pick two to four colonies per construct and grow overnight in 2 mL of LB + 1:1000 ampicillin (100 mg/mL). 12. To collect DNA from media, perform miniprep using an IBI miniprep kit on the samples based on the manufacturer’s protocol. 13. Confirm successful ligation via sequencing using a vendor (e.g., Plasmidsaurus, and Genewiz).

AAV-circRNAs for CNS expression

3.2

Cell Culture

3.2.1 Cell Maintenance and Preparation

237

1. Culture HEK293 cells in Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin. 2. Maintain cells at 37 °C and 5% CO2. 3. Seed cells at an appropriate volume, based on plate size, so that they are at ~70% confluency at the time of transfection (~24 h after seeding).

3.2.2 Transfection of Adherent HEK293 Cells

1. 24 h post-cell seeding (or when cells are at ~70% confluency), perform transfections using a ratio of 3.5 μL of 1 mg/mL PEI MAX 40,000 for every 1 μg of plasmid DNA per well. 2. 4 days post-transfection, harvest cells from the plate by pipetting with cold 1× PBS. 3. Separate cell suspension into two tubes and centrifuge at 300 × g for 4 min at 4 °C. 4. Resuspend cell pellets in either 1× Passive Lysis Buffer (for western blotting, see Subheading 3.5) or Trizol Reagent (for RNA extraction, see Subheading 3.3). 5. Place cells on a rocker for 10 min at 4 °C. 6. Centrifuge lysates in Passive Lysis Buffer for 5 min at 16,000 × g at 4 °C to remove cell debris. 7. Store resuspended cells at -80 °C until use.

3.3

RNA Extraction

1. Extract RNA using Trizol Reagent following the manufacturer’s protocol. 2. Resuspend RNA in nuclease-free water and store at -80 °C until use.

3.4 RNase R Treatment

To degrade linear RNA while preserving circRNA, follow the Abcam RNase R protocol: 1. Treat 5 μg of RNA with 2 μL RNase R, 2 μL RNase Buffer, and 0.5 μL ABM RNaseOFF, and bring to 20 μL with RNase-free H2O. 2. Incubate at 37–45 °C for 2–3 h. 3. Inactivate the enzyme at 95 °C for 5 min.

3.5 RT-PCR and qPCR

1. Convert RNA to cDNA using the High Capacity RNA-toDNA kit, following the manufacturer’s protocol. 2. Use products from this RT-PCR reaction as a template for quantitative PCR (qPCR) using gene-specific primers for GFP and GAPDH as a housekeeping gene (see Table 1).

238

Katie N. Clements et al.

Fig. 2 Testing circRNA backsplicing intron efficiency in the murine CNS via intracerebroventricular (ICV) injection of AAV-circRNA vectors. (a) Engineered circRNA cassettes containing various backsplicing introns (HIPK3, modified HIPK3 [ΔΔ], ZKSCAN1, and EphB4) are packaged into AAV capsids. (b) Experimental setup for ICV stereotactic injections. (c) Location of injection site in the murine brain. (d) Representative immunohistochemistry images from a coronal section of murine brain tissue demonstrating EGFP expression from circRNAs generated by different backsplicing introns

3.6 Recombinant AAV Vector Production 3.6.1 Insertion of circRNA Cassette Between ITRs

To deliver the circRNA cassette using AAV vectors, insert the entire cassette (promoter to polyadenylation signal) between AAV serotype-specific ITRs (Fig. 2a). Either synthesize the newly designed cassette between flanking ITRs using an outside vendor or follow the steps below to insert the circRNA cassette between ITRs. For this study, we used the AAV9 serotype. 1. Acquire a plasmid from a vendor (e.g., Addgene) that has desired ITRs based on the specific serotype used for the experiments. Identify restriction endonucleases that will remove unwanted genetic material between the two ITR regions and use restriction endonuclease cutting to linearize the ITR backbone and remove the unwanted material between the ITRs. 2. Run a gel with the linearized ITR backbone and perform gel extraction of the properly-sized linearized plasmid band using the IBI Scientific Gel Purification Kit and nanodrop for the final sample concentration.

AAV-circRNAs for CNS expression

239

3. In parallel, design primers to amplify the entire circRNA cassette (promoter to polyadenylation signal) with Gibson Assembly homology arms to the ITR backbone (similar to Subheading 3.1.1). 4. Perform Gibson Assembly of backbone and insert, transformation, and plasmid prep following protocols from Subheading 3.1.4, step 6, with the careful consideration that ITR plasmids need to be incubated in the 32 °C incubator for no longer than 12 h. Otherwise, the ITRs can recombine. 5. To confirm successful ligation of backbone and insert, transform the cassette following the protocol in Subheading 3.1.5. 6. To verify that the ITRs did not recombine, perform a SmaI digest. 3.6.2 Upstream AAV Production

Following the successful generation of the AAV-circRNA vector, proceed with the triple plasmid transfection to produce AAV in HEK293 producer cells. For in vitro and in vivo studies, we recommend preparing the virus at a titer of 1e12–1e13 vector genomes/ mL (vg/mL). For the AAV9 serotype, this can be accomplished with ten plates of adherent HEK293. Other serotypes may require more or less plates to achieve the desired titer. Alternatively, transfections can be performed using suspension HEK293 cells (see Note 3): 1. Seed cells in ten 15 cm plates 24 h prior to transfection. Cells should be ~70–80% confluent at the time of transfection. 2. At the time of transfection, prepare the transfection cocktail in two separate 50 mL conical tubes. 3. The first conical tube contains sterile 1× dPBS with the plasmids at a 12:10:6 μg ratio per 15 cm dish of pXX680, AAV9 Rep/Cap plasmid, and pITR-AAV-circRNA, respectively (see Note 4). 4. The second conical tube contains 1× dPBS and 1 mg/mL PEI Max transfection reagent (1:3.5 μg plasmid to μL PEI ratio). 5. Initially, mix each conical tube separately and then combine into a single tube; vortex thoroughly. 6. Allow the plasmid–PEI mixture to sit at RT for 5–10 min. 7. Add 1 mL of transfection cocktail to each of the ten 15 cm plates, dropwise. 8. Mix well on the plate and place in a 37 °C incubator. 9. Harvest DMEM containing AAV on days 4 and 6 posttransfection and place into a 500 mL polypropylene centrifuge tube. Replenish 15 cm plates with fresh DMEM after day 4 harvest (see Note 5).

240

Katie N. Clements et al.

10. Following the final media harvest, add 40% polyethylene glycol (PEG)-8000 to each media tube to a final concentration of 12% and mix well. 11. Precipitate AAV out of the media at 4 °C for 24 h. 12. After 24 h, centrifuge the polypropylene tubes containing AAV at 3000 × g for 45 min at 4 °C. 13. Discard the supernatant and place bottles upside down to drain remaining media for 5 min. 14. Resuspend the PEG pellet in each tube with 1 mL of AAV formulation buffer (1× dPBS with 1 mM MgCl and 0.001% pluronic F-68). 15. Incubate PEG resuspension in 10 mg/mL DNase (100×) and 1 mM MgCl2 and 2 mM CaCl2 for 1 h at 37 °C. 16. Prepare iodixanol density gradient by overlaying 3 mL 60%, 4 mL 40%, 3 mL 25%, and 3 mL 17% iodixanol in a 17 mL ultracentrifugation tube. 17. Add the 2–3 mL of DNase-treated PEG suspension on top of the 17%. 18. Spin each gradient in an ultracentrifuge at 30,000 rpm at 17 °C for at least 8 h or overnight. 19. Next day, remove the iodixanol gradient from the ultracentrifuge and begin fractioning the gradient into 550 μL aliquots starting with the 17% to the 60% (see Note 6). 3.6.3 Determination of Vector Genome Titers

1. Prepare samples for qPCR to quantify the number of viral genomes in each fraction aliquot. To generate a standard curve, use a viral preparation of a known quantity. 2. To remove extraneous DNA contaminants that are outside of the capsid before lysing, combine 10 μL viral fractions and standard fractions with 90 μL of DNase I buffer and 1 μL 10 mg/mL DNase. Incubate at 37 °C for 1 h. 3. Inactivate the DNase with 6 μL of 0.5 M EDTA. 4. Lyse the viral capsids by incubating DNased samples with 100 μL of 10% Tween. Incubate at RT for 20 min. 5. To generate the standard curve, dilute the standard samples in 1:100 molecular biology grade (MBG) water, followed by serial dilution of 1:2 for 8 points up the standard curve. 6. For samples, dilute the fractions 1:100 in MBG water. 7. Prepare qPCR master mix following qPCR SYBR Green kit. qPCR primers used for cassettes in this study can be found in Table 4 (see Note 7).

AAV-circRNAs for CNS expression

241

Table 4 Nucleotide sequences of primers used in qPCR Name

Primer sequence

GFP-F

5′-CTGCTTGTCGGCCATGATATAGACGTTGTGGC-3′

GFP-R

5′-CAAGCTGACCCTGAAGTTCATCTGCACCACC-3′

GAPDH-F

5′-CCACTCCTCCACCTTTGAC-3′

GAPDH-R

5′-ACCCTGTTGCTGTAGCC-3′

CMV-F

5′-CAAGTACGCCCCCTATTGAC- 3′

CMV-R

5′-AGGGCACCTCCATCTCGGAAAC-3′

3.7 Intracerebroventricular (ICV) Injections of AAVcircRNA Vectors into Mice

1. Breed C57/B6 mice in order to generate P0–P1 neonatal mice for intracerebroventricular (ICV) injection of AAV-circRNA cassette.

3.7.1 Administration of AAV-circRNA Vectors

3. Cool stereotaxic stage with two-coin-size dry ice pieces and isopropanol in the cooling basin. This ensures that the pups stay anesthetized during the injection. Ensure that the cooled stereotaxic stage is not too cold or too warm to the touch.

2. Prepare the stereotaxic apparatus with the microinjection syringe pump and Hamilton syringe (Fig. 2b).

4. Warm a heating pad and cover with a paper towel for recovery following injection. 5. Assemble the anesthetizing station: stack two latex gloves filled with small ice chips on one another and place them in a Styrofoam box filled with ice. 6. Anesthetize the P0–P1 pups by sandwiching one to two pups between the ice-filled latex gloves. Incubate for 3–5 min. They will undergo color loss and become unresponsive when ready to inject. 7. Load the AAV vector into the prepared Hamilton syringe while the mice are being anesthetized. The total dose injected is 1–2e10 vg in 2 μL. 8. Place the anesthetized pup on the stereotaxic stage and secure using the thumb and index finger. Ensure that the visible suture line between the lambda and bregma of the brain is parallel to the stage. 9. Calibrate the syringe needle so that the baseline X, Y coordinates (0,0), and reside directly above the lambda point of the neonate brain. 10. Move the needle either left or right 0.8 mm (X), followed by up 1.5 mm (Y) toward the coronal suture (Fig. 2c, left).

242

Katie N. Clements et al.

11. Lower the needle and penetrate through the skin, inserting the needle just until the bevel is in. 12. Raise the needle slightly until the injection site is no longer concave and the bevel is at the surface. 13. Lower needle again 1.5 mm (Z) (see Note 8). 14. Using the syringe pump, infuse 2 μL at a rate of 66.6 nL/s into the left lateral ventricle. Do not exceed 2 μL of volume injected into each lateral ventricle. Keep the needle inserted for 30 s after the vector is completely infused before retracting the needle from the injection site. 15. If desired, repeat the process on the opposite ventricle using the same coordinates. 16. Identify neonates via toe-clip and allow them to recover on the prepared heating pad until responsive and full of color. 17. Scruff the dam and gently rub the neonates against her vulva and belly before placing the neonates back in the cage. 18. Wean pups into new cages at P21/P22. 3.7.2

Harvesting Tissue

1. To sacrifice mice, first administer a high dose of Avertin (2,2,2tribromoethanol, 0.06 mL of a 1.25% solution per gram BW, 125–200 mg/kg) to anesthetize. Check for responsiveness by pinching all four limbs. 2. If unresponsive to pinches, create a window into the thoracic cavity by making a sub-xiphoid window. Create an exit route for the blood and perfusate to clear by cutting the inferior vena cava. Perform cardiac perfusion through the left ventricle with 25 mL saline to remove blood from the mouse. 3. Dissect out the brain tissue for histological analysis and fix in 10% neutral buffered formalin for 24 h. Post-24 h, store tissue in 1× dPBS with 0.01% sodium azide until ready to section.

3.7.3 Preparation of Tissue

1. If continuing steps immediately, wash formalin in 1× dPBS for 5 min 3×. Incubate fixed brains in 30% sucrose on a rotator at 4 °C for 24 h. The tissue should sink. 2. Remove the tissue from sucrose and pat dry using a paper towel or kimwipe. 3. Embed fixed tissue in 3% Agarose (low melting) and allow solidifying on ice. Mount the agarose embedded tissue on the stage using super glue. 4. Collect 25 μm coronal sections (Fig. 2d, right). Store in 1× dPBS in a 12-well dish at 4 °C and proceed to either (1) immunohistochemistry or immunofluorescence protocol or if looking to detect native fluorescence (2) immediately to Subheading 3.7.5 for mounting slides and imaging.

AAV-circRNAs for CNS expression 3.7.4 Immunohistochemistry

243

1. To detect the protein of interest, begin by incubating sections in blocking buffer for 1 h at 4 °C. Vibratome sections require being on a rocker at each step. 2. While in blocking buffer, prepare the primary antibody staining cocktail. This consists of blocking buffer with an appropriate concentration of primary antibody based on the protein of interest. (We used anti-rabbit GFP at a 1:500 concentration.) Calculate the total amount of cocktail based off of 1 mL per well. 3. As an additional blocking step, carefully aspirate blocking buffer and replace it with 3% hydrogen peroxide for 15 min. 4. Wash with 1× PBS 3× for 5 min. 5. Incubate tissues with 1 mL primary antibody staining cocktail overnight at 4 °C, on the rocker. 6. Next day, remove the primary cocktail and wash with 1× PBS 3× for 5 min. 7. To develop the stain, we suggest using a VECTASTAIN Elite ABC-HRP Peroxidase (Rabbit IgG) Kit following the manufacturer’s instructions starting at the point of biotinylated secondary antibody incubation. Use ImmPACT DAB substrate peroxidase (HRP) to develop a signal following the manufacturer’s instructions. 8. After developing, continue to mount slides for imaging.

3.7.5 Histological Quantification of Vector Performance

1. Mount tissue sections on 75 × 25 mm VistaVision HistoBond Adhesive Microscope Slides. 2. Place a dab of ProLong Gold Antifade Mountant over each piece of tissue sample and seal with a coverslip. 3. Allow media to set overnight. Protect from light. 4. Next day, seal the coverslip with nail polish. Allow to dry fully before imaging. 5. Use a confocal microscope or an ECHO revolve fluorescence microscope to capture images of the sample. 6. For brain tissue, separate imaging into different regions and capture a total of six images from each region of interest. 7. To quantify the total number of GFP+ cells, count each GFP+ cell in a given area of tissue from each injected animal. The counting of cells can be separated based on cell morphology if distinguishable. 8. Further analysis can be performed using the ImageJ software (see Subheading 2 for the link).

244

4

Katie N. Clements et al.

Notes 1. We suggest initially synthesizing the entire plasmid from a vendor or ordering a gBlock for each gene fragment of interest. In doing so, it is important to consider future cloning of constructs into this cassette. The most efficient way to set this up is to insert restriction endonuclease sites (that are unique to only that site) between each fragment (i.e., promoter, intronic sequences, and polyA site). This will allow for quick removal of fragments via restriction digestion and insertion of a new fragment through T4 ligation. 2. Optional clean-up for restriction digest of backbone before ligation: Run digested product on a 1% agarose gel and perform gel extraction on a band of correct size. Perform gel purification using the IBI Scientific Gel Purification Kit and nanodrop for the final sample concentration. 3. As an alternative approach for transfections, use a 1.6 L flask of suspension HEK293 cells. Seed the flask at a density of 5e5 cells for an expected transfection density of 1e6 cells 24 h postseeding. The remaining transfection steps are the same as adherent HEK293 cells. 4. The Rep/Cap plasmid that is used will be based on the AAV serotype used in the experiments. The specific AAV serotype that is used will determine the tissue expression of the cargo [32, 33]. 5. Different AAV serotypes can require different time points for media harvest [34]. To ensure AAV harvest at an appropriate time, determine the standard points of collection for the specific serotype used in the experiments. For example, AAV2/4 vectors require processing of the cell pellet on Day 6 in addition to the described media harvests. Following resuspension of your PEG pellet with AAV formulation buffer, add 1 mL formulation buffer to your cell pellet and lyse the cells via sonication. Following sonication, combine the processed cell pellet and PEG resuspension and proceed with the protocol as described. 6. The virus can be found primarily in the 40% iodixanol layer, with full capsids migrating to the bottom of the 40% and partial/empty capsids toward the top of the 40%. Make sure to collect fractions as slowly and precisely as possible to avoid cross contamination. 7. Primers can be designed against ITRs, promoters, exons/ introns, or polyA regions as desired. 8. If using a Hamilton syringe, position the bevel toward you to prevent the needle from slipping as it is inserted.

AAV-circRNAs for CNS expression

245

Acknowledgments This work was supported by the National Institutes of Health (R01NS099371 to A.A., W.F.M., and J.E.W. and R01HL089221, R01GM127708, and UG3AR075336 to A.A). References 1. Holdt LM, Kohlmaier A, Teupser D (2018) Circular RNAs as therapeutic agents and targets. Front Physiol 9:1262. https://doi.org/ 10.3389/fphys.2018.01262 2. Stoll L, Sobel J, Rodriguez-Trejo A, Guay C, Lee K, Venø MT et al (2018) Circular RNAs as novel regulators of β-cell functions in normal and disease conditions. Mol Metab 9:69–83 3. Bai Y, Zhang Y, Han B, Yang L, Chen X, Huang R et al (2018) Circular RNA DLGAP4 ameliorates ischemic stroke outcomes by targeting miR-143 to regulate endothelial-mesenchymal transition associated with blood–brain barrier integrity. J Neurosci 38(1):32–50 4. Li Z, Huang C, Bao C, Chen L, Lin M, Wang X, Zhong G, Yu B, Hu W, Dai L, Zhu P, Chang Z, Wu Q, Zhao Y, Jia Y, Xu P, Liu H, Shan G (2015) Exon-intron circular RNAs regulate transcription in the nucleus. Nat Struct Mol Biol 22(3):256–264 5. Liu B, Ye B, Zhu X, Yang L, Li H, Liu N, Zhu P, Lu T, He L, Tian Y, Fan Z (2020) An inducible circular RNA circKcnt2 inhibits ILC3 activation to facilitate colitis resolution. Nat Commun 11(1):4076 6. Ashwal-Fluss R, Meyer M, Pamudurti NR, Ivanov A, Bartok O, Hanan M, Evantal N, Memczak S, Rajewsky N, Kadener S (2014) circRNA biogenesis competes with pre-mRNA splicing. Mol Cell 56(1):55–66 7. Gong X, Tian M, Cao N, Yang P, Xu Z, Zheng S et al (2021) Circular RNA circEsyt2 regulates vascular smooth muscle cell remodeling via splicing regulation. J Clin Investig 131(24): e147031 8. Hansen TB, Jensen TI, Clausen BH, Bramsen JB, Finsen B, Damgaard CK, Kjems J (2013) Natural RNA circles function as efficient microRNA sponges. Nature 495(7441): 384–388 9. Man W, Cui Y, Li J, Li Y, Jin J, Jin Y, Wu X, Zhong R, Li X, Yao H, Lin Y, Jiang L, Wang Y (2022) circTAB2 inhibits lung cancer proliferation, migration and invasion by sponging miR-3142 to upregulate GLIS2. Apoptosis. https://doi.org/10.1007/s10495-02201805-1

10. Chen CY, Sarnow P (1995) Initiation of protein synthesis by the eukaryotic translational apparatus on circular RNAs. Science (New York, NY) 268(5209):415–417 11. Zhang Y, Zhang X, Shen Z, Qiu Q, Tong X, Pan J, Zhu M, Hu X, Gong C (2022) BmNPV circular RNA-encoded peptide VSP39 promotes viral replication. Int J Biol Macromol 228:299–310 12. Enuka Y, Lauriola M, Feldman ME, Sas-Chen A, Ulitsky I, Yarden Y (2016) Circular RNAs are long-lived and display only minimal early alterations in response to a growth factor. Nucleic Acids Res 44(3):1370–1383 13. Wilusz JE (2018) A 360° view of circular RNAs: from biogenesis to functions. Wiley Interdiscip Rev RNA 9(4):e1478. https://doi.org/10.1002/wrna.1478 14. Monat C, Cousineau B (2016) Circularization pathway of a bacterial group II intron. Nucleic Acids Res 44(4):1845–1853 15. Lu Z, Filonov GS, Noto JJ, Schmidt CA, Hatkevich TL, Wen Y et al (2015) Metazoan tRNA introns generate stable circular RNAs in vivo. RNA 21(9):1554–1565 16. Wang Y, Wang Z (2015) Efficient backsplicing produces translatable circular mRNAs. RNA (New York, NY) 21(2):172–179 17. Chen R, Wang SK, Belk JA, Amaya L, Li Z, Cardenas A, Abe BT, Chen CK, Wender PA, Chang HY (2022) Engineering circular RNA for enhanced protein production. Nat Biotechnol 41:262. https://doi.org/10.1038/ s41587-022-01393-0 18. Meganck RM, Liu J, Hale AE, Simon KE, Fanous MM, Vincent HA, Wilusz JE, Moorman NJ, Marzluff WF, Asokan A (2021) Engineering highly efficient backsplicing and translation of synthetic circRNAs. Mol Ther Nucleic Acids 23:821–834 19. Zhao Y, Huang L (2014) Lipid nanoparticles for gene delivery. Adv Genet 88:13–36 20. Topal J, Panchal N, Barroeta A, Roppelt A, Mudde A, Gaspar HB, Thrasher AJ, Houghton BC, Booth C (2023) Lentiviral Gene Transfer Corrects Immune Abnormalities in XIAP Deficiency. J Clin Immunol 43(2):440–451.

246

Katie N. Clements et al.

https://doi.org/10.1007/s10875-02201389-0gonz 21. Gonzalez TJ, Simon KE, Blondel LO, Fanous MM, Roger AL, Maysonet MS, Devlin GW, Smith TJ, Oh DK, Havlik LP, Castellanos Rivera RM, Piedrahita JA, ElMallah MK, Gersbach CA, Asokan A (2022) Cross-species evolution of a highly potent AAV variant for therapeutic gene transfer and genome editing. Nat Commun 13(1):5947 22. Stanton AC, Lagerborg KA, Tellez L, Krunnfusz A, King EM, Ye S, Solomon IH, Tabebordbar M, Sabeti PC (2023) Systemic administration of novel engineered AAV capsids facilitates enhanced transgene expression in the macaque CNS. Med (New York, NY) 4(1):31–50.e8 23. Meganck RM, Borchardt EK, Castellanos Rivera RM, Scalabrino ML, Wilusz JE, Marzluff WF, Asokan A (2018) Tissue-dependent expression and translation of circular RNAs with recombinant AAV vectors in vivo. Mol Ther Nucleic Acids 13:89–98 24. Nieuwenhuis B, Laperrousaz E, Tribble JR, Verhaagen J, Fawcett JW, Martin KR, Williams PA, Osborne A (2023) Improving adenoassociated viral (AAV) vector-mediated transgene expression in retinal ganglion cells: comparison of five promoters. Gene Ther 30:503. https://doi.org/10.1038/s41434-02200380-z 25. Qin JY, Zhang L, Clift KL, Hulur I, Xiang AP, Ren BZ, Lahn BT (2010) Systematic comparison of constitutive promoters and the doxycycline-inducible promoter. PLoS One 5(5):e10611 26. Kuhn B, Ozden I, Lampi Y, Hasan MT, Wang SS (2012) An amplified promoter system for targeted expression of calcium indicator

proteins in the cerebellar cortex. Front Neural Circuits 6:49 27. Yao Z, Luo J, Hu K, Lin J, Huang H, Wang Q, Zhang P, Xiong Z, He C, Huang Z, Liu B, Yang Y (2017) ZKSCAN1 gene and its related circular RNA (circZKSCAN1) both inhibit hepatocellular carcinoma cell growth, migration, and invasion but through different signaling pathways. Mol Oncol 11(4):422–437 28. Liang D, Wilusz JE (2014) Short intronic repeat sequences facilitate circular RNA production. Genes Dev 28(20):2233–2247 29. Noren NK, Pasquale EB (2007) Paradoxes of the EphB4 receptor in cancer. Cancer Res 67(9):3994–3997 30. Xu T, Wu J, Han P, Zhao Z, Song X (2017) Circular RNA expression profiles and features in human tissues: a study using RNA-seq data. BMC Genomics 18(Suppl 6):680 31. Meganck RM, Borchardt EK, Castellanos Rivera RM, Scalabrino ML, Wilusz JE, Marzluff WF, Asokan A (2018) Tissue-dependent expression and translation of circular RNAs with recombinant AAV vectors in vivo. Mol Therapy Nucleic acids 13:89–98 32. Wang D, Tai PWL, Gao G (2019) Adenoassociated virus vector as a platform for gene therapy delivery. Nat Rev Drug Discov 18(5): 358–378 33. Pupo A, Ferna´ndez A, Low SH, Franc¸ois A, Sua´rez-Amara´n L, Samulski RJ (2022) AAV vectors: the Rubik’s cube of human gene therapy. Mol Ther 30(12):3515–3541 34. Elmore ZC, Patrick Havlik L, Oh DK, Anderson L, Daaboul G, Devlin GW, Vincent HA, Asokan A (2021) The membrane associated accessory protein is an adeno-associated viral egress factor. Nat Commun 12(1):6239

Chapter 14 Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via Microfluidic Mixing Malte Juchem, Sarah Cushman, Dongchao Lu, Shambhabi Chatterjee, Christian B€ar, and Thomas Thum Abstract This chapter serves as a guide for researchers embarking on circular RNA-based translational studies. It provides a foundation for the successful encapsulation of circular RNA into lipid nanoparticles (LNPs) and facilitates progress in this emerging field. Crucial scientific methods and techniques involved in the formulation process, particle characterization, and downstream processing of circ-LNPs are covered. The production of in vitro transcribed circular RNA-containing LNPs based on a commercially available lipid mix is provided, in addition to the fundamentals for successful encapsulation based on lipid mixes composed of single components. Furthermore, the transfection and validation protocols for the identification of a functional and potentially therapeutic circRNA candidate for initial in vitro verification, before subsequent LNP studies, are explained. Key words Circular RNA, In vitro transcription (IVT), RNA therapeutics, Lipid nanoparticle (LNP), Cardiomyocytes

1

Introduction

1.1 Basic Aspects of Circular RNA Biology

In recent years, circular RNAs have emerged as key regulators in development and diseases and could prove essential as therapeutic interventions in addition to the standard quality of care [1– 6]. Their mechanisms of action are much broader than that of linear RNAs, conferring a wider range of molecular interactions in several different cell types, tissues, and organs. In addition, circular RNAs show a longer half-life as they are resistant to exonucleolytic degradation due to their lack of linear ends, making them excellent therapeutic targets [7–9].

Malte Juchem and Sarah Cushman contributed equally. Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_14, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

247

248

Malte Juchem et al.

Circular RNAs have the capability to both directly and indirectly mediate several important cellular processes including the regulation of cell proliferation, maintenance of pluripotency, and mediating apoptosis within the cells [8–13]. 1.2 Translational Potential of Circular RNAs

The wide range of functional diversity, in combination with a more stable therapeutic potential from reduced exonuclease degradation, allows circular RNAs to be thoroughly explored for clinical applications [14, 15]. Over 65 clinical trials currently involve the delivery of RNA as a therapy, either using naked RNA, which is often modified, or fully encapsulated RNA with the use of lipid nanoparticles (LNPs) [3]. However, only small interfering RNAs (siRNAs), messenger RNAs (mRNAs), and antisense oligonucleotides (ASOs) have been developed so far for clinical applications [3, 16, 17]. Currently, circular RNAs are also involved in clinical trials, but only as biomarkers for genetic and cancer-based studies [18], highlighting that their full potential remains to be exploited.

1.3 Circular RNAs for Clinical Use as a Therapeutic Intervention

Apart from their well-accepted role as biomarkers, the modulation of disease-associated circular RNAs can be attained through up- or down-regulation to achieve therapeutic effects, based on their mechanism of action [1, 14, 19]. In addition to circular RNA depletion techniques using CRISPR-Cas13 and standard RNA interference, the exploration of overexpression tools can help to enhance the protective mechanisms a given circular RNA may possess [14, 20, 21]. The overexpression of circular RNAs can be achieved using viral approaches, encompassing adenoviral or lentiviral vectors, as well as non-viral methods, such as in vitro transcribed (IVT) circular RNAs [14, 22, 23]. Abe and colleagues expertly detail several methods by which to prepare IVT circular RNA in the previous chapter, including both chemical and enzymatic synthesis [23]. Circular RNAs with a known beneficial effect in vitro or in vivo can then be transfected to cells after IVT production as either a preventative or a therapeutic strategy, to mediate biological cellular processes (Fig. 4) [24]. However, exogenous circular RNAs that are synthetically produced are known to have a higher immunogenicity than preexisting endogenous circular RNAs [25]. There are several approaches to diminish an immunogenic response in vivo, including chemical modifications, purification strategies, or alterations in the physical preparation during IVT production [26–28].

1.4 Lipid Nanoparticle (LNP)Based Delivery of RNA Therapeutics

A variety of delivery techniques are used for IVT circular RNA distribution that can be further modified to avoid, or even in certain cases, produce, an immune reaction. Recently, LNPs have emerged as an advanced non-viral drug delivery system for in vivo applications [29–31], which has been impressively demonstrated by novel

Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via. . .

249

mRNA vaccines [32–36]. In a recent study, an IVT circular RNA vaccine was synthesized to target SARS-CoV-2 using an LNP-based delivery system, which could induce a beneficial level of immunogenicity. These data showed that optimized and commercially available LNPs (which encapsulated the IVT circular RNA vaccine against SARS-CoV-2) were able to induce the same level of neutralizing antibodies, even when the vaccination dose of the IVT circular RNA was reduced [19]. Biophysical properties of LNPs can be altered to prevent premature clearance and avoid cargo degradation, and in the case of non-vaccine purposes, evade unwanted immune responses [23, 29, 30, 37]. Typically, components such as cholesterol, phospholipids, and polyethylene glycol (PEG)-coupled lipids are incorporated into LNPs, as they are known to enhance stability, increase blood circulation time, and alter the permeability of the outer membrane for improved overall performance in vivo [38, 39]. Most important are the ionizable lipids that not only complex the ribonucleotide cargo but are also essential for tolerability and efficient delivery in that their charge is pH dependent [40]. At a nearly neutral pH, as found in the circulation, only a small percentage of the ionizable lipids in an LNP will be protonated due to a pKa of approximately 6.5. This prevents toxicity concerns associated with permanently cationic particles in the bloodstream [41]. The LNP only becomes protonated after it has been taken up by the cell and exposed to a more acidic environment in late endosomes with a pH as low as 4.5 [42, 43]. The cationic ionizable lipids then interact with negatively charged lipids in the endosomal membrane, which is thought to allow for the release of the therapeutic RNA cargo into the cytoplasm, resulting in increased expression, as seen in several mRNA vaccine packaging techniques [40, 42, 44]. LNPs can be engineered to possess distinct biophysical properties such as specific size, dimension, lipid composition, and charge, which enables specific cell or tissue enrichment. For instance, in lymph nodes, 30 nm LNPs are taken up by dendritic cells, while LNPs that are more T-cell-specific are often larger, in a range between 80 and 130 nm, maintaining a lower PEG density [43, 45–48]. In addition, more positively charged particles have been found to primarily localize in the lung [49]. Moreover, celland organ-specific targeting can be achieved by including LNP surface ligands or through selective organ targeting (SORT) LNPs, aiding in a more efficient delivery system with fewer off-target effects or aggregations in unwanted organs [43, 50, 51]. In this chapter, we explore delivery mechanisms of IVT circular RNA in vitro using standard Lipofectamine-based transfection techniques to show the functionally protective capabilities of IVT circular RNAs, and more complex LNP packaging techniques for the delivery of IVT circular RNAs to form circular RNA-LNPs, which are applicable to both in vitro and in vivo experiments.

250

2

Malte Juchem et al.

Materials All solutions are to be prepared with ultrapure water and analyticalgrade reagents.

2.1

LNP Packaging

2.1.1 Pre-formulation Preparation

All consumables and reagents used are to be RNAse-free. Avoid using material that is also used for other applications, which could lead to their contamination with endotoxins. 1. UV spectrophotometer. 2. Circular RNA in RNase-free water at 1–1.5 μg/μL. 3. GenVoy-ILM™ (Precision NanoSystems Inc.; NWW0041/ NWW0042). 4. GenVoy formulation buffer 100 mM, pH 4 (Precision NanoSystems Inc.; NWW0043). 5. Absolute ethanol.

2.1.2

LNP Formulation

1. NanoAssemblr® Ignite™ (Precision NanoSystems Inc.). 2. NanoAssemblr® Ignite™ NxGen™ Cartridge (Precision NanoSystems Inc.; NIN0061/NIN0062). 3. Centered tip syringes 1–10 mL (refer to the GenVoy-ILM™ user manual for syringe compatibility). 4. Mg2+/Ca2+-free PBS.

2.2 Downstream Processing and Particle Characterization

1. Slide-a-Lyzer G3 Dialysis Cassettes, 20 K MWCO (A52977, Thermo Fisher Scientific).

2.2.1 Dialysis, Concentration, and Sterile Filtration

4. Mg2+/Ca2+-free PBS.

2. 5-L beakers. 3. Magnetic stir bar and stirring plate. 5. Syringe. 6. Amicon Ultra-15, PLGC Ultracel-PL Membrane, 10-kDa filters. 7. 0.2-μm sterile filters (e.g., Acrodisc®, 13 mm, Supor® membrane).

2.2.2 RiboGreen Assay to Quantify Encapsulation Efficiency

1. Quant-iT™ RiboGreen™ RNA Assay Kit (R11490, Thermo Fisher Scientific). 2. Triton X-100. 3. RNase-free Water. 4. 96-well plate: black. 5. Mg2+/Ca2+-free PBS. 6. Plate reader.

Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via. . . 2.2.3 Particle Sizing and PDI Via Dynamic Light Scattering

3

251

1. Dynamic light scattering particle analyzer. 2. Low-volume cuvettes.

Methods

3.1 In Vitro Transcribed Circular RNA LNP Packaging

The protocol below describes the packaging of IVT circular RNA into circular RNA-LNPs (circ-LNP) using the commercially available proprietary GenVoy ionizable lipid mix (GenVoy-ILM) on the Ignite NanoAssemblr platform (see Note 1). Furthermore, downstream processing will then be described, enabling the final circLNPs to be applied in in vivo or in vitro functional investigations of the packaged circular RNA.

3.1.1 Pre-formulation Preparation and RNA Dilution

1. Make sure that the lipids are dissolved by observing no visible precipitates and mixing thoroughly if necessary. Dilute 1:2 in anhydrous EtOH to 12.5 mM. 2. Dilute the circular RNA in the formulation buffer provided with the GenVoy-ILM to the concentration stated in the user manual, corresponding to a specific flow rate ratio (FRR) or GenVoy-ILM working concentration. For instance, for an FRR of 3:1 and a lipid mix working concentration of 12.5 mM, this would require 170–180 ng of circular RNA/μL. Assure the circular RNA concentration pre-dilution is high enough so that the concentration of the formulation buffer following the dilution is still 65 mM or more (100 mM stock concentration). Validate the RNA concentration—and adjust if required— using standard procedures on a UV spectrophotometer via A260.

3.1.2 LNP Packaging— Commercial

Carry out the formulation of circ-LNPs following the GenVoyILM user manual. Good initial settings are an FRR (aqueous: organic) of 3:1, a total flow rate (TFR) of 12 mL/min, an N/P ratio of 4 (see Note 2), and a circular RNA concentration of 170–180 ng/μL. These settings generally yield LNPs within a size range of 80–115 nm, a polydispersity index (PDI) < 0.2, and an encapsulation efficiency of >90% for circular RNAs of several hundred nucleotides in length. However, further optimization of individual parameters might be desired to adjust traits of the final circ-LNPs such as their size (refer to the GenVoy-ILM user manual [52] and see Note 1).

3.1.3 Downstream Processing—Dialysis, Concentration, and Sterilization

Downstream processing of circ-LNPs was largely derived from a previously published protocol for siRNA-LNPs [53]. In brief: 1. Following formulation, quickly dilute the formulated circLNPs in PBS (1:3 or more) to reduce the EtOH concentration

252

Malte Juchem et al.

and raise the pH of the mixture. Next, equilibrate the dialysis cassette in PBS and load the sample into the cassette. Then, dialyze for 2 h against a 500-times excess of PBS (compared with the original formulation volume) with constant slow stirring. Exchange the buffer, and continue dialysis over night at 4 °C. 2. Retrieve LNP solution and concentrate using Amicon Ultra15, PLGC Ultracel-PL Membrane, 10-kDa filters to a desired concentration. 3. Sterilize the concentrated sample in a sterile hood using a 0.2-μ m filter. Collect an aliquot of about 50 μL for particle characterization (see Subheading 3.1.3, step 2). Store LNP solution at 4 °C until further use. Prepare the LNPs as freshly as possible before the planned administration. 3.1.4 Particle Characterization

1. Perform a modified RiboGreen Assay to determine encapsulation efficiency and RNA concentration. Use either the RiboGreen Assay kit-provided RNA standard for standard curve preparation or prepare a standard using your circular RNA IVT product of known concentration. Refer to the GenVoyILM user manual [52] and Walsh et al. [53] for detailed information on how to perform the RiboGreen Assay (please see Note 3). 2. Determine polydispersity and average particle size by diluting an aliquot of formulated LNPs in PBS to an appropriate concentration, so that attenuation during measurement will be in the range of 6–9. If necessary, dilute further or add more sample to be within that range. Follow the manual of the particle sizer used.

3.2

Transfection

3.2.1 Transfection of In Vitro Transcribed Circular RNA in Cardiomyocytes

1. Neonatal rat cardiomyocytes (NRCMs) are isolated from day 0 to 3 rat pups via the Neonatal Heart Dissociation Kit (Miltenyi #130-105-420) according to the manufacturer’s protocol [54]. NRCMs are cultured with MEM medium containing 5% FBS, 1% Penicillin/Streptomycin, 100 nM BrdU, 2 μg/mL Vitamin B12 at 37 °C, and 1% CO2. 2. NRCMs are used to test the circular RNA overexpression achieved by Lipofectamine 2000 mediated transfection of IVT circular RNAs in vitro. NRCMs are seeded into a 12-well plate at a density of 500,000 cells/mL. 3. 48 h after cell seeding, IVT circular RNA (500 ng/500,000 cells) is transfected into NRCMs using Lipofectamine 2000 according to the manufacturer’s protocol [55]. As a control, NRCMs were treated with Lipofectamine 2000 without transfecting any IVT circular RNA (Mock) (see Notes 4 and 5).

Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via. . .

253

Fig. 1 Lipofectamine 2000 packaged IVT circular RNA transfection in neonatal rat cardiomyocytes (NRCMs). (a) Schematic depiction of the IVT Circ-INSR transfection timeline in NRCMs using Lipofectamine 2000 (Lipo-2000). (b) Plot represents human Circ-INSR (hsa-Circ-INSR) expression measured using quantitative PCR by a divergent primer pair in RNA isolated from NRCMs transfected with 500 ng IVT hsa-Circ-INSR (n = 3–4 wells per group) at different time points compared with Mock. Mock control: Lipofectamine 2000 only, without IVT circular RNA. Exp. expression, FC fold change. Error bars indicate data ± SEM. (Image adapted from Lu et al. 2022 [24])

4. At the transfection endpoint, perform RNA isolation from NRCMs using miRNeasy Mini Kit (Qiagen # 74104) [56] according to the manufacturer’s protocol. 5. Reverse transcribe 500 ng total RNA with random primer using iScript Select cDNA Synthesis kit (Bio-Rad #1708891) according to the manufacturer’s protocol [57]. 6. Perform quantitative polymerase chain reaction (PCR) with iQ SYBR Green mix (Bio-Rad #1708880) with QuantStudio7 (ABI) using target-specific divergent primer pairs to confirm the overexpression of the targeted circular RNA (example shown in Fig. 1 using the circular RNA derived from the insulin receptor host gene, named Circ-INSR) compared with the control sample (Mock). 7. Prepare a 2% agarose DNA gel (Roth #3810.4) using 1× TAE buffer. 8. Run the gel at a constant 100 V for 1 h. After the gel electrophoresis, collect the PCR product via QIAquick Gel Extraction Kit (Qiagen #28706) [58].

254

Malte Juchem et al.

9. Mix PCR samples (50 ng in 15 μL) with 2 μL forward or reverse divergent primer of circular RNA. Perform Sanger sequencing to check the back-splicing site of your targeted circular RNA (see additional Chapter 5 in this book [23]). 3.2.2 LNP-Based Transfection

1. NRCMs are isolated and maintained following the protocols mentioned above (see Subheading 3.2.1, step 1). To test the overexpression efficiency achieved by transfection of circLNPs, NRCMs are seeded into 12-well plates before transfection (as described in Subheading 3.2.1, step 2). 2. 48 h after cell seeding, add different doses of circ-LNPs to the medium. Particularly, if screening for biological effects of the circular RNA, it is highly recommended to apply various controls such as untreated, mock-treated, and IVT, non-circularized, i.e., linear RNA, controls (see Note 5). Furthermore, for many cell types, cellular uptake of LNPs might be more efficient when adding recombinant apolipoprotein-E4 at 0–5 μg/mL to the cell culture media. 3. At the transfection endpoint, RNA isolation is performed from the cells (see Subheading 3.2.1, step 4). 4. Reverse transcribe as described before in Subheading 3.2.1, step 5. 5. Perform quantitative PCR as described in Subheading 3.2.1, step 6. (Example shown in Fig. 2: circ-LNPs induce dose-

Fig. 2 Dose-dependent increase in circular RNA expression after circ-LNP transfection. Human (hsa)-Circ-INSR expression measured using quantitative PCR in neonatal rat cardiomyocytes transfected with different doses (100–500 ng) of IVT hsa-Circ-INSR encapsulated by LNPs compared with the mock control (n = 3, 48 h). Mock control: LNP only, without IVT circular RNA. exp. expression, FC fold change. Error bars indicate data ± SEM

Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via. . .

255

Fig. 3 Expression timeline of circular RNA after circ-LNP transfection. Human (hsa)-Circ-INSR expression was measured at different time points (24–96 h) in neonatal rat cardiomyocytes transfected with LNP-packaged IVT hsa-Circ-INSR (300 ng) compared with the control (Mock) using quantitative PCR (n = 4). Mock control: LNP only, without IVT circular RNA. exp. expression, FC fold change. Error bars indicate data ± SEM

dependent overexpression of the exogenous circular RNA derived from the insulin receptor host gene, named CircINSR in NRCMs compared to the Mock control. In Fig. 3: stable overexpression of the circular RNA at different time points achieved by circ-LNP transfection compared with Mock control in NRCMs). 3.3 Pre-clinical Application of IVT Circular RNA in NRCMs to Gain Cardioprotection

In addition to inducing the expression of the targeted circular RNA (via Lipofectamine 2000 or LNP-mediated transfection of the IVT circular RNA), the downstream biological function of the circular RNA can also be investigated in vitro or in vivo using these strategies. Here, we present an experiment where IVT Circ-INSR (a circular RNA derived from the insulin receptor host gene) is transfected in NRCMs using Lipofectamine 2000 to test for its biological function in enhancing cardiomyocyte cell survival when NRCMs are treated with doxorubicin (an anti-cancer drug with cardiotoxic effects). The therapeutic effect of the IVT circular RNA is investigated in NRCMs before doxorubicin stress (as a preventative therapeutic approach to represent treatment pre-chemotherapy in cancer patients; Fig. 4a) as well as after doxorubicin-induced cardiotoxicity (to represent a curative therapy; Fig. 4d). 1. NRCMs are isolated and maintained following the protocols mentioned above (see Subheading 3.2.1). To perform this cardioprotection functional assay, NRCMs are plated in a 48-well plate (125,000 cells/well) for analysis of DNA damage induced by doxorubicin, using immunostaining for gammaH2AX (γ-H2AX).

256

Malte Juchem et al.

Fig. 4 Application of IVT circular RNA to protect from doxorubicin-induced cardiotoxicity. (a) Schematic representation of a preventative therapeutic strategy to protect from doxorubicin-mediated cytotoxicity in neonatal rat cardiomyocytes transfected with IVT human (hsa) Circ-INSR using Lipofectamine 2000 followed by doxorubicin treatment. (b and c) γ-H2AX (red) staining in nuclei (DAPI, blue) of neonatal rat cardiomyocytes (cTNT, green) transfected with Lipofectamine 2000 packaged IVT Circ-INSR (500 ng/mL) and Circ-INSR (linear, 500 ng/mL) compared with control, under doxorubicin treatment. Yellow arrows indicate DNA damage. Representative images left; quantification right (0.5 μM doxorubicin 48 h, n = 3 per group, scale bar, 50 μm. (d) Schematic representation of therapeutic strategy to prevent cardiotoxicity using Lipofectamine 2000-based IVT human (hsa) Circ-INSR transfection in neonatal rat cardiomyocytes already under doxorubicin stress. (e and f) γ-H2AX (red) staining in nuclei (DAPI, blue) of neonatal rat cardiomyocytes (cTNT, green) transfected with Lipofectamine 2000 packaged IVT Circ-INSR (500 ng/mL) and Circ-INSR (linear, 500 ng/mL) compared with control, under doxorubicin treatment. Yellow arrows indicate DNA damage. Representative images left; quantification right (0.25 μM doxorubicin 48 h, n = 6–9 per group, from three independent experiments, scale bar, 50 μm, image adapted from Lu et al. 2022 [24])

2. Perform Lipofectamine 2000-based IVT circular RNA transfection in NRCMs at a concentration of 500 ng/500,000 cells (see Subheading 3.2.1). As this experiment is intended to screen for biological effects of the circular RNA, we include the following controls: no transfection with only doxorubicin treatment (Control), and IVT, non-circularized, i.e., linear RNA (IVT Circ-INSR linear) controls with doxorubicin treatment.

Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via. . .

257

3. After 48 h (preventative therapy) or 24 h (curative therapy) transfection, perform the γ-H2AX immunostaining to check DNA damage in cardiomyocytes. (Results are shown in Fig. 4.)

4

Notes 1. Instead of the proprietary GenVoy-ILM, self-made lipid mixes can also be prepared and used to formulate LNPs. Various lipid formulations are described in the literature that facilitate specific delivery of therapeutic ribonucleotide molecules. Ongoing progress in the field is continuously expanding the repertoire of lipids and formulations for effective LNP-based therapeutics while addressing one of the major challenges in systemic RNA delivery via LNPs: extrahepatic targeting [36, 38, 59, 60]. However, lipid mixes for the formulation of LNPs usually consist of at least the following components dissolved in absolute ethanol: (a) Ionizable cationic lipids (b) Phospholipids (c) Pegylated lipids (d) Cholesterol When formulating circ-LNPs with self-made lipid mixes, resuspend circular RNA in sodium citrate or sodium acetate buffer pH 4 in the two-digit mM range. As we have found that standard settings for formulating mRNA into LNPs usually also work for packaging circRNAs, we recommend initially following the instructions described in the literature for formulating mRNA into LNPs using the respective ionizable or cationic lipids. However, individual parameters might be adapted in the process to yield optimized circ-LNPs for individual applications. Namely, the molar ratios of the abovementioned four components, as well as the formulation settings (generally, higher FRR and TFR lead to smaller particles) can be fine-tuned to yield LNPs with varying biophysical properties and possibly altered in vivo tropism [61, 62]. Following formulation, proceed with downstream processing and characterization of circ-LNPs as described above. 2. The N/P ratio is calculated as follows: N =P =

moles of ionizable lipid
No:of protonatable amines in the ionizable lipid moles of circRNA
length of circRNA ðNo:of anionic phosphatesÞ 3. If the fluorescence intensity falls outside the RNA standard curve then, depending on how much the LNP solution has

258

Malte Juchem et al.

been concentrated, varying amounts of LNPs may need to be measured in the RiboGreen Assay so that they occur within the range of the RNA standard curve (0.1–2.5 μg/mL). 4. We suggest that it is necessary to pre-test the dose of Lipofectamine 2000 or LNP-packaged IVT circular RNA for your target cell types. 5. We suggest that different controls can additionally be used for the functional assays. We recommend including untreated control (cells without any transfection reagent), mock-treated control (only transfection reagents such as Lipofectamine 2000 or LNP without IVT circular RNA), and linear RNA controls (IVT products without circularization procedures).

Acknowledgments This work was supported by funding from the German Research Foundation, DFG (SFB/Transregio TRR267, to C.B. and T.T.) and the European Research Council, ERC (ERC Advanced Grant REVERSE, to T.T.). Disclosure TT is a founder and shareholder of Cardior Pharmaceuticals GmbH (outside of this book chapter). D.L., C.B., and T.T. have filed and partly licensed patents for non-coding RNAs including for Circ-INSR.

References 1. Ren S et al (2020) Circular RNAs: promising molecular biomarkers of human aging-related diseases via functioning as an miRNA sponge. Mol Ther Methods Clin Dev 18:215–229 2. Beermann J, Piccoli MT, Viereck J, Thum T (2016) Non-coding RNAs in development and disease: background, mechanisms, and therapeutic approaches. Physiol Rev 96:1297–1325 3. Kaczmarek JC, Kowalski PS, Anderson DG (2017) Advances in the delivery of RNA therapeutics: from concept to clinical reality. Genome Med 9:60 4. Neufeldt D, Cushman S, B€ar C, Thum T (2023) Circular RNAs at the intersection of cancer and heart disease: potential therapeutic targets in cardio-oncology. Cardiovasc Res cvad013. https://doi.org/10.1093/cvr/ cvad013 5. Santer L, Bar C, Thum T (2019) Circular RNAs: a novel class of functional RNA molecules with a therapeutic perspective. Mol Ther 27:1350–1363

6. Bar C, Chatterjee S, Thum T (2016) Long noncoding RNAs in cardiovascular pathology, diagnosis, and therapy. Circulation 134:1484– 1499 7. Chen L-L, Yang L (2015) Regulation of circRNA biogenesis. RNA Biol 12:381–388 8. Yu C-Y, Kuo H-C (2019) The emerging roles and functions of circular RNAs and their generation. J Biomed Sci 26:29 9. Jeck WR, Sharpless NE (2014) Detecting and characterizing circular RNAs. Nat Biotechnol 32:453–461 10. Shan K et al (2017) Circular noncoding RNA HIPK3 mediates retinal vascular dysfunction in diabetes mellitus. Circulation 136:1629–1642 11. Zhong Z, Lv M, Chen J (2016) Screening differential circular RNA expression profiles reveals the regulatory role of circTCF25-miR103a-3p/miR-107-CDK6 pathway in bladder carcinoma. Sci Rep 6:30919

Encapsulating In Vitro Transcribed circRNA into Lipid Nanoparticles Via. . . 12. Yu C-Y et al (2017) The circular RNA circBIRC6 participates in the molecular circuitry controlling human pluripotency. Nat Commun 8:1149 13. Tian F, Yu CT, Ye WD, Wang Q (2017) Cinnamaldehyde induces cell apoptosis mediated by a novel circular RNA hsa_circ_0043256 in non-small cell lung cancer. Biochem Biophys Res Commun 493:1260–1266 14. Holdt LM, Kohlmaier A, Teupser D (2018) Circular RNAs as therapeutic agents and targets. Front Physiol 9:1262 15. Chen X, Lu Y (2021) Circular RNA: biosynthesis in vitro. Front Bioeng Biotechnol 9: 787881 16. Huang C-K, Kafert-Kasting S, Thum T (2020) Preclinical and clinical development of noncoding RNA therapeutics for cardiovascular disease. Circ Res 126:663–678 17. Hunkler HJ, Groß S, Thum T, B€ar C (2022) Non-coding RNAs: key regulators of reprogramming, pluripotency, and cardiac cell specification with therapeutic perspective for heart regeneration. Cardiovasc Res 118:3071–3084 18. https://clinicaltrials.gov/ 19. Qu L et al. Circular RNA vaccines against SARS-CoV-2 and emerging variants. https:// doi.org/10.1101/2021.03.16.435594. 20. Fischer JW, Leung AKL (2017) CircRNAs: a regulator of cellular stress. Crit Rev Biochem Mol Biol 52:220–233 21. Zhang Y et al (2021) Optimized RNA-targeting CRISPR/Cas13d technology outperforms shRNA in identifying functional circRNAs. Genome Biol 22:41 22. Xia P et al (2018) A circular RNA protects dormant hematopoietic stem cells from DNA sensor cGAS-mediated exhaustion. Immunity 48:688–701.e7 23. Circular RNAs: methods and protocols. vol. 1724. Springer, New York, 2018. 24. Lu D et al (2022) A circular RNA derived from the insulin receptor locus protects against doxorubicin-induced cardiotoxicity. Eur Heart J. https://doi.org/10.1093/eurheartj/ ehac337 25. Obi P, Chen YG (2021) The design and synthesis of circular RNAs. Methods 196:85–103 26. Wesselhoeft RA et al (2019) RNA circularization diminishes immunogenicity and can extend translation duration in vivo. Mol Cell 74:508–520.e4 27. Chen YG et al (2019) N6-methyladenosine modification controls circular RNA immunity. Mol Cell 76:96–109

259

28. Chen YG et al (2017) Sensing self and foreign circular RNAs by intron identity. Mol Cell 67: 228–238.e5 29. Cheng CJ, Tietjen GT, Saucier-Sawyer JK, Saltzman WM (2015) A holistic approach to targeting disease with polymeric nanoparticles. Nat Rev Drug Discov 14:239–247 30. Evers MJW et al (2018) State-of-the-art design and rapid-mixing production techniques of lipid nanoparticles for nucleic acid delivery. Small Methods 2:1700375 31. Khalil IA, Younis MA, Kimura S, Harashima H (2020) Lipid nanoparticles for cell-specific in vivo targeted delivery of nucleic acids. Biol Pharm Bull 43:584–595 32. Polack FP et al (2020) Safety and efficacy of the BNT162b2 mRNA covid-19 vaccine. N Engl J Med 383:2603–2615 33. Schoenmaker L et al (2021) mRNA-lipid nanoparticle COVID-19 vaccines: structure and stability. Int J Pharm 601:120586 34. Anderson EJ et al (2020) Safety and immunogenicity of SARS-CoV-2 mRNA-1273 vaccine in older adults. N Engl J Med 383:2427–2438 35. Baden LR et al (2020) Efficacy and safety of the mRNA-1273 SARS-CoV-2 vaccine. N Engl J Med 384:403–416 36. Hou X, Zaks T, Langer R, Dong Y (2021) Lipid nanoparticles for mRNA delivery. Nat Rev Mater 6:1078–1094 37. Paunovska K, Loughrey D, Dahlman JE (2022) Drug delivery systems for RNA therapeutics. Nat Rev Genet 23:265–280 38. Ickenstein LM, Garidel P (2019) Lipid-based nanoparticle formulations for small molecules and RNA drugs. Expert Opin Drug Deliv 16: 1205–1226 39. Hald Albertsen C et al (2022) The role of lipid components in lipid nanoparticles for vaccines and gene therapy. Adv Drug Deliv Rev 188: 114416 40. Hassett KJ et al (2019) Optimization of lipid nanoparticles for intramuscular administration of mRNA vaccines. Mol Ther - Nucleic Acids 15:1–11 41. Stern ST, McNeil SE (2008) Nanotechnology safety concerns revisited. Toxicol Sci 101:4–21 42. Carrasco MJ et al (2021) Ionization and structural properties of mRNA lipid nanoparticles influence expression in intramuscular and intravascular administration. Commun Biol 4:956 43. Kularatne RN, Crist RM, Stern ST (2022) The future of tissue-targeted lipid nanoparticlemediated nucleic acid delivery. Pharmaceuticals 15:897

260

Malte Juchem et al.

44. Moderna Clinical Study Protocol (2020), Aug 20. https://covid19crc.org/wp-content/ uploads/2020/09/mRNA-1273-P301-Proto col-2020. 45. Rurik JG et al (2022) CAR T cells produced in vivo to treat cardiac injury. Science 375:91– 96 46. Tomba´cz I et al (2021) Highly efficient CD4+ T cell targeting and genetic recombination using engineered CD4+ cell-homing mRNALNPs. Mol Ther 29:3293–3304 47. Ramishetti S et al (2015) Systemic gene silencing in primary T lymphocytes using targeted lipid nanoparticles. ACS Nano 9:6706–6716 48. Nakamura T et al (2020) The effect of size and charge of lipid nanoparticles prepared by microfluidic mixing on their lymph node transitivity and distribution. Mol Pharm 17:944– 953 49. Cheng Q et al (2020) Selective organ targeting (SORT) nanoparticles for tissue-specific mRNA delivery and CRISPR–Cas gene editing. Nat Nanotechnol 15:313–320 50. Wang X et al (2023) Preparation of selective organ-targeting (SORT) lipid nanoparticles (LNPs) using multiple technical methods for tissue-specific mRNA delivery. Nat Protoc 18: 265–291 51. Johnson LT et al (2022) Lipid nanoparticle (LNP) chemistry can endow unique in vivo RNA delivery fates within the liver that alter therapeutic outcomes in a cancer model. Mol Pharm 19:3973–3986 52. Genvoy User Manual. ht tps://www. precisionnanosystems.com/docs/defaultsource/pni-files/user-guides/genvoy-userguide-2020-final-emailsize.pdf. 53. Walsh C et al (2014) In: Jain KK (ed) Microfluidic-based manufacture of

siRNA-lipid nanoparticles for therapeutic applications BT - drug delivery system. Springer, New York, pp 109–120. https:// doi.org/10.1007/978-1-4939-0363-4_6 54. Miltenyi- Neonatal Heart Dissociation Kit, mouse and rat. https://www.miltenyibiotec. com/DE-en/products/neonatal-heart-dissoci ation-kit-mouse-and-rat.html#gref. 55. Thermofischer- Lipofectamine 2000. https:// w ww. ther mofisher.com/de/de/home/ references/protocols/cell-culture/transfec tion-protocol/lipofectamine-2000.html. 56. Qiagen- RNeasy Mini Kit. https://www. qiagen.com/us/resources/resourcedetail? id=14e7cf6e-521a-4cf7-8cbc-bf9f6fa33e24& lang=en. 57. Biorad- iScript. https://www.bio-rad.com/ sites/default/files/webroot/web/pdf/lsr/ literature/10001023.pdf. 58. QIAquick® Gel Extraction Kit. https://www. qiagen.com/us/resources/resourcedetail? id=1426dbb4-da09-487c-ae01-c587c2be14 c3&lang=en. 59. Ni H et al (2022) Piperazine-derived lipid nanoparticles deliver mRNA to immune cells in vivo. Nat Commun 13:4766 60. Lorenzer C, Dirin M, Winkler A-M, Baumann V, Winkler J (2015) Going beyond the liver: progress and challenges of targeted delivery of siRNA therapeutics. J Control Release 203:1–15 61. Sago CD et al (2018) High-throughput in vivo screen of functional mRNA delivery identifies nanoparticles for endothelial cell gene editing. Proc Natl Acad Sci 115:E9944–E9952 62. Roces CB et al (2020) Manufacturing considerations for the development of lipid nanoparticles using microfluidics. Pharmaceutics 12: 1095

Part V Functional Study: Biogenesis and Translation

Chapter 15 Characterizing Post-transcriptional Modifications of circRNAs to Investigate Biogenesis and Translation Gaia Di Timoteo, Dario Dattilo, and Irene Bozzoni Abstract Recent studies have shown that circular RNAs (circRNAs) are decorated with N6-methyladenosine (m6A), a co-transcriptional modification known to participate in the regulation of many processes governing linear RNA metabolism. Nevertheless, the activity of this mark on circRNAs is still poorly understood. In order to facilitate the study of m6A-dependent regulation of these molecules, we provide protocols that enable circOme-wide detection of m6A as well as the perturbation of several components of the m6A machinery followed by assays useful to evaluate the impact of their depletion on the production and, when applicable, on the translation of circRNAs. Other modifications exist and can be explored following the same principles. Key words Circular RNA (circRNA), Epitranscriptomics, N6-metyhyladenosine (m6A), Biogenesis, cap-independent translation

1 Introduction For many years, research about gene expression regulation has focused on the modulation of transcription by epigenetic modifications of DNA or on the alteration of protein activity and stability by post-translational modifications. Nevertheless, emerging studies have discovered over 150 post-transcriptional modifications occurring also on the RNA, collectively referred to as “epitranscriptome” [1]. The proteins affecting the presence of RNA modification on a transcript can be divided into three functional categories, where “writers” are responsible for their deposition, “erasers” catalyze their removal, and “readers” mediate downstream functions through binding of the modified transcripts.

Gaia Di Timoteo and Dario Dattilo contributed equally. Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_15, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

263

264

Gaia Di Timoteo et al.

Fig. 1 A schematic representation of an adenosine containing m6A modification

One of the best characterized RNA modifications is N6-methyladenosine (m6A), which consists in the addition of a methyl group to the nitrogen-6 position of adenosine (Fig. 1). M6A was among the first modifications to be identified at high levels on mRNAs in several eukaryotic organisms and was subsequently detected in other RNA species such as ribosomal RNAs (rRNAs), transfer RNAs (tRNAs), small nuclear RNAs (snRNAs), small nucleolar RNAs (snoRNAs), microRNAs (miRNAs), and long noncoding RNAs (lncRNA) [2]. Despite its discovery dating back to the 1970s, functional studies on m6A did not begin until 2012, when the use of an antibody against m6A allowed researchers to immunoprecipitate methylated transcripts and start transcriptome-wide profiling of m6A distribution with highthroughput sequencing [3, 4]. Subsequent studies have shown that m6A is able to fine-tune gene expression at multiple stages, modulating both nuclear processes (e.g., alternative splicing, miRNA maturation, and export) during its co-transcriptional deposition, as well as cytoplasmic ones (RNA stability and cap-dependent and cap-independent translation) [5]. Recent studies extended the concept of epitranscriptome also to circular RNAs (circRNAs): numerous methylated circRNAs have been identified in different cell lines through m6A RIP-seq analysis, with m6A levels being dependent on cell type and the presence of METTL3 and METTL14 [6, 7]. Interestingly, circRNAs can exhibit patterns of m6A modifications that are distinct from those of mRNAs [7], suggesting that this mark might play a role in differentiating the production or function of circular versus linear isoforms. In consideration of the important role of m6A in backsplicing as a form of alternative splicing required for circRNA production, it is important to establish standard approaches to adopt when exploring the circRNA epitranscriptome. In this chapter, we provide protocols that are useful to study the impact of m6A on the production of circRNAs but also on one of their possible functions, namely, translation.

Characterizing Post-transcriptional Modifications of circRNAs

265

Fig. 2 The workflow of the RPAD pipeline. Total RNA is isolated from the sample of interest. Treatment with the exonuclease RNase R is performed in order to degrade the majority of linear species. RNase R-resistant linear RNAs are then subjected to a polyadenylation reaction and subsequently removed with oligo-dT magnetic beads, leading to the enrichment of circRNAs in the final sample

circRNAs are generally lowly-expressed species [8]; therefore, all protocols for m6A identification on these molecules require stringent purification from linear transcripts, such as rRNAs and messenger RNAs [9]. The RPAD protocol [10], described later, is an efficient strategy for obtaining high-purity circRNAs (Fig. 2). The enrichment of methylated circRNAs is performed through m6A immunoprecipitation followed by high-throughput sequencing. Several approaches have been developed for this purpose;

266

Gaia Di Timoteo et al.

Fig. 3 The workflow of meRIP-seq and miCLIP protocols. Both approaches involve the fragmentation of previously enriched circular RNA (circRNA) followed by immunoprecipitation of methylated regions, library preparation, and next-generation sequencing. In the miCLIP, specifically, a crosslinking step ensures the formation of a covalent bond between the anti-m6A antibody and the methylated residues (red dots in the figure), which is later used in the RNA-seq data analysis as a footprint to detect modifications at singlenucleotide resolution

nevertheless, we will mainly discuss two protocols, known as methylated RNA immunoprecipitation and sequencing (MeRIPseq) or m6A individual-nucleotide resolution crosslinking and immunoprecipitation (miCLIP) [3, 11] (Fig. 3). The first method only allows to identify methylated regions on RNAs with a resolution of 100–200 nucleotides, while the latter allows locating individual m6A sites. M6A detection on circRNAs from high-throughput sequencing data is performed through bioinformatic pipelines that are heavily dependent on the protocol used for the m6A

Characterizing Post-transcriptional Modifications of circRNAs

267

immunoprecipitation. Analyses deriving from MeRIP-seq allow us to evaluate the distribution of the so-called “peaks” (regions enriched for the presence of m6A) along the transcripts. The pipeline used for miCLIP, instead, relies on two mutational events arising during library preparation: after proteinase K treatment, the remaining amino acid of the anti-m6A antibody covalently attached to a methylated residue interferes during the conversion of the RNA fragments into cDNA, leading to the generation of either crosslinking-induced mutation sites (CIMS) or crosslinkinginduced truncation sites (CITS). Such events are later used as signatures for the identification of methylated sites. The methylation of circRNAs detected through highthroughput analyses can be further validated on individual candidates by performing an m6A CLIP assay. Once the presence of m6A is established, the down-regulation of writers/readers/erasers represents the first and easiest approach to evaluate the effect of a modification on circRNA metabolism. As an example, we will describe how to obtain the down-regulation of METTL3, the main m6A writer, and check the effects on circRNA levels. Nevertheless, the same protocol can be applied to other factors. As METTL3 works in a heterodimer with METTL14, if effects are observed upon its down-regulation, it is recommended to perform METTL14 down-regulation to consolidate the results obtained. As m6A (similarly to other RNA modifications) has been shown to have a role at different levels of RNA maturation and function, additional experiments should be performed to understand which specific step is affected when the accumulation of specific circRNAs is altered upon writers/readers/erasers down-regulation: we will describe how to check m6A involvement in circRNA localization and decay. If these processes are not disturbed, circRNA production may be responsible for the changes in circRNA levels observed. CircRNAs can be rarely translated, and their translation can be affected by RNA modification, as in the case of circZNF609 [12]. Antibodies recognizing the proteins produced from circRNA ORFs are usually unavailable; thus, it is useful to introduce a tag in the circRNA ORF. When the genomic context and the cellular system allow it, the tag can be directly inserted in the genome through genome editing approaches, and experiments on the endogenous protein can be conducted. Circular and linear RNAs arise from the same precursor transcript, and the tag can potentially be included in the frame in the ORFs of both isoforms. However, because usually circRNAs include less exons than their linear counterparts and have shorter ORFs, it is possible to discriminate the different proteins produced by size at least in electrophoretic approaches. This is not true for imaging techniques or methodologies based on immunoprecipitation. Therefore, it is recommended to individually consider the

268

Gaia Di Timoteo et al.

effect of the tag insertion on the linear transcript and for the design of downstream experiments. Case-by-case considerations should be made for the design of the tagging system. When the insertion of the tag in the circRNA genomic locus is unfeasible (but also when just looking for preliminary data), it is suggested to insert the tag into a vector able to overexpress the circRNA of interest. As a complementary approach, it is possible to further assess the function of single m6A sites on specific circRNAs by introducing site-specific mutations either on the endogenous transcript (through genome-editing approaches) or in a vector able to overexpress the circRNA. On one hand, when compared with the wild-type version, the mutant should recapitulate the effects observed upon alteration of the m6A machinery factors (e.g., METTL3 knock-down); on the other hand, the mutant should also be resistant to such deregulation. Also, mutations on the single transcript exclude all the indirect effects due to the lack of the depleted factors. Finally, we describe how to perform a general CLIP for writers/readers/erasers. However, even though it is possible to observe the interaction of the targets with writers and erasers, this binding is mostly transient and less stable than the readers one.

2 2.1

Materials circRNA Isolation

1. miRNeasy kit (QIAGEN). 2. Nanodrop spectrophotometer. 3. RNase R 20 U/μL (Lucigen, Epicentre, cat. RNR07250). 4. TURBO DNA-free™ Kit (Thermo Fisher Scientific, cat AM1907). 5. RiboLock RNase inhibitor (Thermo Fisher Scientific, cat. EO0382). 6. Thermomixer. 7. Vortexer. 8. TRIzol™ Reagent (Thermo Fisher Scientific). 9. Chloroform. 10. Isopropanol. 11. 75% Ethanol. 12. RNase-free glycogen. 13. Microcentrifuge. 14. Poly(A) tailing kit (Thermo Fisher Scientific, cat. AM1350). 15. Poly(A)Purist™ MAG kit (Thermo Fisher Scientific, cat. AM1922).

Characterizing Post-transcriptional Modifications of circRNAs

269

16. Magnetic separation stand. 17. Random primers (Thermo Fisher Scientific, cat. 48190011). 18. dNTP mix (Thermo Fisher Scientific, cat. 18427088). 19. Maxima Reverse Transcriptase (Thermo Fisher Scientific, cat. EP0742). 20. Mir-X™ miRNA First Strand Synthesis Kit (Clontech, Takara Bio, cat. 638315). 21. MicroAmp® Optical Adhesive Film.

96-Well

Reaction

Plate

and

22. MPS 1000 Mini Plate Spinner. 23. PowerUp SYBR Green polymerase chain reaction (PCR) Master Mix (Applied Biosystems). 24. 96-Well Thermal Cycler. 25. Real-Time PCR System (es. QuantStudio 6 Real-Time PCR System, Thermo Fisher Scientific). 2.2 m6A Immunoprecipitation 2.2.1

MeRIP-seq

1. High-purity circRNA isolated with the RPAD protocol. 2. RNA Fragmentation Reagents: 10× Fragmentation Reagent and Stop Solution (Thermo Fisher Scientific, cat. AM8740). 3. Thermomixer. 4. TRIzol™ Reagent (Thermo Fisher Scientific). 5. Chloroform. 6. Isopropanol. 7. 75% Ethanol. 8. RNase-free glycogen. 9. Centrifuge. 10. Agilent 2100 Bioanalyzer. 11. Agilent RNA 6000 Pico kit (cat. 5067-1513). 12. NanoDrop spectrophotometer. 13. Equipment and reagents for agarose gel electrophoresis. 14. QIAquick Gel Extraction Kit (Qiagen, cat. 28606X4). 15. Anti-m6A antibody (Abcam, cat. ab151230). 16. Rabbit IgG antibody (Merck, cat. 12-370). 17. RiboLock RNase inhibitor (Thermo Fisher Scientific, cat. EO0382). 18. IP buffer: 50 mM Tris–HCl pH 7.4, 100 mM NaCl, 0.05% (vol/vol) NP-40, 1 U/μL RiboLock. 19. Rotator. 20. Dynabeads™ Protein A/Protein G (Thermo Fisher Scientific, cat. 88803).

270

Gaia Di Timoteo et al.

21. Magnetic separation stand. 22. N6-methyladenosine (Sigma-Aldrich, cat. CS220007). 23. Total RNA Prep with Ribo-Zero Plus (Illumina, cat 20040529). 2.2.2

miCLIP

1. High-purity circRNA isolated from the step above. 2. RNA Fragmentation Reagents: 10× Fragmentation Reagent and Stop Solution (Thermo Fisher Scientific, cat. AM8740). 3. Thermomixer. 4. TRIzol™ Reagent (Thermo Fisher Scientific). 5. Chloroform. 6. Isopropanol. 7. 75% Ethanol. 8. RNase-free glycogen. 9. Centrifuge. 10. Agilent 2100 Bioanalyzer. 11. Agilent RNA 6000 Pico kit (cat. 5067-1513). 12. NanoDrop spectrophotometer. 13. Equipment and reagents for agarose gel electrophoresis. 14. Anti-m6A antibody (Synaptic Systems, cat 202 003; or Abcam, cat. ab151230). 15. Rabbit IgG antibody (Merck, cat. 12-370). 16. RiboLock RNase inhibitor (Thermo Fisher Scientific, cat. EO0382). 17. IP buffer: 50 mM Tris–HCl pH 7.4, 100 mM NaCl, 0.05% (vol/vol) NP-40, 1 U/μL RiboLock. 18. 35-mm Cell culture dish. 19. Stratalinker (Agilent). 20. Dynabeads™ Protein A/Protein G (Thermo Fisher Scientific, cat. 88803). 21. Magnetic separation stand. 22. 1 M Dithiothreitol (DTT) (Thermo Fisher Scientific, cat. R0861). 23. High-Salt Wash Buffer: 50 mM Tris–HCl pH 7.4, 1 M NaCl, 1 mM EDTA, 1% (vol/vol) NP-40, and 0.1% (vol/vol) SDS. 24. PNK Wash Buffer: 20 mM Tris–HCl pH 7.4, 10 mM MgCl2, and 0.2% (vol/vol) Tween-20. 25. 5× PNK Buffer pH 6.5: 350 mM Tris–HCl pH 7.4, 50 mM MgCl2, and 5 mM DTT. 26. T4 PNK 10 U/μL (NEB, cat. M0201L).

Characterizing Post-transcriptional Modifications of circRNAs

271

27. T4 RNA ligase I—high concentration 30 U/μL supplied with 10× Ligation Buffer (NEB, cat. M0437M). 28. Polyethylene glycol (PEG) 400 (Merck, cat. 81170). 29. Pre-adenylated DNA linker (L3 linker): /5rApp/ AGATCG GAAGAGCGGTTCAG /3ddC/(HPLC-purified) 20 μM (IDT). 30. 4–12% bis-tris protein gel (Thermo Fisher Scientific, cat. NP0321BOX). 31. 20× MES SDS Running Buffer (Thermo Fisher Scientific, cat. NP0002). 32. Electrophoretic gel chamber and transfer unit. 33. 4× LDS buffer (Biorad). 34. Prestained Protein Ladder (Thermo Fisher Scientific, cat. 26620). 35. Nitrocellulose membrane. 36. Odyssey CLx Imaging System (LI-COR). 37. Proteinase K Buffer: 100 mM Tris–HCl pH 7.4, 50 mM NaCl, 1 mM EDTA, and 0.2% (vol/vol) SDS. 38. Proteinase K (Sigma-Aldrich). 39. Phenol:Chloroform:Isoamyl Alcohol. 40. 2 mL Phase Lock Gel Heavy tube. 41. GlycoBlue (Thermo Fisher Scientific). 42. 3 M sodium acetate pH 5.5. 43. Ethanol 100%. 44. 1 pmol/μL RT primer. 45. dNTP mix (Thermo Fisher Scientific, cat. 18427088). 46. Superscript IV (Thermo Fisher Scientific, cat. 18090050). 47. Thermocycler. 48. RNase H (Thermo Fisher Scientific, cat. EN0202). 49. RNase A (Thermo Fisher Scientific, cat. EN0531). 50. AMPure XP beads (Beckman Coulter). 51. CircLigase II reagents: 100 U/μL CircLigase II ssDNA Ligase; 5 M Betaine; 50 mM MnCl2; 2 pmol/μL CircLigase ssDNA Control; Nuclease-Free Water, Sterile; CircLigase II 10× Reaction Buffer (Epicentre, cat. CL9021K). 52. P5/P3 Solexa PCR primer mix (10 μM each). 53. Phusion High-Fidelity PCR Master mix (Thermo Fisher Scientific, cat. F531L). 54. 6% TBE gel (Thermo Fisher Scientific, cat. EC62652BOX).

272

Gaia Di Timoteo et al.

55. SYBR gold (Thermo Fisher Scientific, cat. S11494). 56. UV transilluminator. 2.3

Peak Calling

2.3.1

MeRIP-seq

1. Computer running Linux or MAC OSX. 2. UNIX SHELL. 3. Python https://www.python.org/downloads/. 4. SRA toolkit https://github.com/ncbi/sra-tools. 5. FastQC https://www.bioinformatics.babraham.ac.uk/pro jects/fastqc/. 6. Cutadapt https://cutadapt.readthedocs.io/en/stable/. 7. BWA http://bio-bwa.sourceforge.net/. 8. Bowtie2 https://bowtie-bio.sourceforge.net/bowtie2/index. shtml. 9. TopHat https://ccb.jhu.edu/software/tophat/index.shtml. 10. STAR https://github.com/alexdobin/STAR. 11. MACS2 https://github.com/macs3-project/MACS. 12. ExomePeak http://compgenomics.utsa.edu/exomePeak/. 13. MeTPeak https://github.com/compgenomics/MeTPeak. 14. MeTDiff https://github.com/compgenomics/MeTDiff. 15. m6AViewer http://dna2.leeds.ac.uk/m6A. 16. R https://www.r-project.org/. 17. DESeq2 https://bioconductor.org/packages/release/bioc/ html/DESeq2.html. 18. EdgeR https://bioconductor.org/packages/release/bioc/ html/edgeR.html.

2.3.2

miCLIP

1. Computer running Linux or MAC OSX. 2. UNIX SHELL. 3. Python https://www.python.org/downloads/. 4. SRA toolkit https://github.com/ncbi/sra-tools. 5. FastQC https://www.bioinformatics.babraham.ac.uk/pro jects/fastqc/. 6. Flexbar https://github.com/seqan/flexbar. 7. pyCRAC tool suite https://git.ecdf.ed.ac.uk/sgrannem/ pycrac. 8. STAR https://github.com/alexdobin/STAR. 9. CLIP tool Kit https://zhanglab.c2b2.columbia.edu/index. php/CTK_Documentation. 10. Bedtools2 https://github.com/arq5x/bedtools2. 11. Samtools https://github.com/samtools/.

Characterizing Post-transcriptional Modifications of circRNAs

2.4 Validation of m6A Presence on Specific Targets

273

1. DNase-treated Total RNA. 2. RiboLock RNase inhibitor (Thermo Fisher Scientific, cat. EO0382). 3. IP buffer: 50 mM Tris–HCl pH 7.4, 100 mM NaCl, 0.05% (vol/vol) NP40, 1 U/μL RiboLock. 4. Anti-m6A antibody (Abcam, cat. ab151230). 5. Rabbit IgG antibody (Merck, cat. 12-370). 6. Rotator. 7. 35-mm Cell culture dish. 8. Stratalinker (Agilent). 9. Dynabeads™ Protein A/Protein G (Thermo Fisher Scientific, cat. 88803). 10. Magnetic separation stand. 11. High-Salt Wash Buffer: 50 mM Tris–HCl pH 7.4, 1 M NaCl, 1 mM EDTA, 1% (vol/vol) NP-40, 0.1% (vol/vol) SDS. 12. Proteinase K Buffer: 10 mM Tris–HCl pH 7.4, 100 mM NaCl, 1 mM EDTA, and 0.2% (vol/vol) SDS. 13. Proteinase K (Sigma-Aldrich). 14. TRIzol™ Reagent (Thermo Fisher Scientific). 15. Chloroform. 16. Isopropanol. 17. 75% Ethanol. 18. RNase-Free glycogen. 19. SuperScript™ VILO™ cDNA Synthesis Kit (Thermo Fisher Scientific, cat. 11754250). 20. Thermal Cycler. 21. Optical 96-Well Reaction Plate and Adhesive Film. 22. MPS 1000 Mini Plate Spinner. 23. PowerUp SYBR Green PCR Master Mix (Applied Biosystems). 24. Real-Time PCR System (es. QuantStudio 6 Real-Time PCR System, Thermo Fisher Scientific).

2.5 Effects of m6A Factors Downregulation on circRNAs 2.5.1 Writers/Readers/ Erasers Downregulation

1. Cells of interest. 2. Cell medium. 3. Appropriate tissue culture plates and supplies. 4. Tissue culture incubator. 5. Opti-MEM reduced serum media (Gibco). 6. Lipofectamine RNAiMAX Reagent (Thermo Fisher Scientific, cat. 13778150).

274

Gaia Di Timoteo et al.

7. siRNAs against GS56339).

METTL3

(Smart

pool,

Qiagen,

cat.

8. Control siRNA (Qiagen, cat. 1022076). 9. Real-Time PCR System. 2.5.2 Assessing circRNA Level Upon METTL3 Knockdown

1. Protein buffer: 100 mM Tris–HCl pH 7.5, 1 mM EDTA, 2% SDS, and 1× cOmplete™ Protease Inhibitor Cocktail (Roche, cat. 11697498001). 2. Rotator. 3. Cell lifter. 4. Refrigerated centrifuge. 5. Equipment and reagents for Western Blot analyses. 6. METTL3 antibody (Abcam, cat. ab195352). 7. TRIzol™ Reagent (Thermo Fisher Scientific). 8. Direct-zol RNA Miniprep kit (Zymo Research). 9. PrimeScript RT Master Mix (Takara Bio, cat. RR036B). 10. NanoDrop spectrophotometer. 11. PowerUp SYBR Green Master Mix reagent (Thermo Fisher Scientific). 12. Reverse transcription-quantitative polymerase chain reaction (RT-qPCR) primers for the mature circRNA, its linear counterpart, and their common precursor (see Note 1). 13. 96-Well Thermal Cycler. 14. Real-Time PCR System. 15. Optical 96-Well Reaction Plate and Adhesive Film. 16. MPS 1000 Mini Plate Spinner.

2.6 Identifying Which Step of RNA Metabolism Is Affected Upon Inhibition of RNA Modification 2.6.1 Effects on circRNA Localization

1. Trypsin. 2. Phosphate-buffered saline (PBS). 3. Centrifuge. 4. Protein buffer:100 mM Tris–HCl pH 7.5, 1 mM EDTA, 2% SDS, and 1× cOmplete™ Protease Inhibitor Cocktail (Roche, cat. 11697498001). 5. Rotator. 6. Cell lifter. 7. Refrigerated centrifuge. 8. Equipment and reagents for Western blot analyses. 9. METTL3 antibody (Abcam, cat. ab195352). 10. Ambion PARIS Kit (Life Technologies, cat. AM1921). 11. TRIzol™ Reagent (Thermo Fisher Scientific).

Characterizing Post-transcriptional Modifications of circRNAs

275

12. Chloroform. 13. Isopropanol. 14. 75% Ethanol. 15. RNase-free glycogen. 16. ERCC RNA Spike-In Mix (Thermo Fisher Scientific, cat 4456740). 17. PrimeScript RT Master Mix (Takara Bio, cat. RR036B). 18. NanoDrop spectrophotometer. 19. 96-Well Thermal Cycler. 20. q-PCR primers for cytoplasmic control and nuclear control. 21. q-PCR primers for the targets of interest (see Note 1). 22. Optical 96-Well Reaction Plate and Adhesive Film. 23. MPS 1000 Mini Plate Spinner. 24. PowerUp SYBR Green PCR Master Mix (Applied Biosystems). 25. Real-Time PCR System. 2.6.2 Effects on circRNA Decay

1. Cells of interest depleted for METTL3. 2. Centrifuge. 3. Protein buffer: 100 mM Tris–HCl pH 7.5, 1 mM EDTA, 2% SDS, and cOmplete™ Protease Inhibitor Cocktail 1× (Roche, cat. 11697498001). 4. Rotator. 5. Cell lifter. 6. Refrigerated centrifuge. 7. Equipment and reagents for Western blot analyses. 8. METTL3 antibody (Abcam, cat. ab195352). 9. Actinomycin D. 10. DMSO. 11. TRIzol™ Reagent (Thermo Fisher Scientific). 12. Chloroform. 13. Isopropanol. 14. 75% Ethanol. 15. RNase-free glycogen. 16. NanoDrop spectrophotometer. 17. PrimeScript RT Master Mix (Takara Bio, cat. RR036B). 18. 96-Well Thermal Cycler. 19. q-PCR primers for Actinomycin D treatment control. 20. q-PCR primers for the targets of interest (see Note 1).

276

Gaia Di Timoteo et al.

21. Optical 96-Well Reaction Plate and Adhesive Film. 22. MPS 1000 Mini Plate Spinner. 23. PowerUp SYBR Green PCR Master Mix (Applied Biosystems). 24. QuantStudio 6 Real-Time PCR System (Thermo Fisher Scientific). 2.6.3 Effects on circRNA Translation

1. Cells of interest depleted for METTL3. 2. Cell medium. 3. Tissue culture incubator. 4. Appropriate tissue culture plates and supplies. 5. Lipofectamine 2000 (Life Technologies, cat. 11668019). 6. Opti-MEM (Gibco). 7. DNA vector able to overexpress the coding circRNA in its tagged version (see Note 2). 8. Centrifuge. 9. Protein buffer: 100 mM Tris–HCl pH 7.5, 1 mM EDTA, 2% SDS, and 1× cOmplete™ Protease Inhibitor Cocktail (Roche, cat. 11697498001). 10. Rotator. 11. Cell lifter. 12. Refrigerated centrifuge. 13. Western blot supplies. 14. Anti-Tag antibody (e.g., anti-Flag, Sigma, cat. A8592-2MG).

2.7 RNA Modifications SiteSpecific Mutagenesis

1. DNA vector able to overexpress the coding circRNA in its wildtype version (see Note 2). 2. PCR primers incorporating the desired change for site-directed mutagenesis/deletion of the vector in the putative modified sites. 3. CloneAmp™ HiFi PCR Premix (Takara, cat. 639298). 4. Equipment and reagents for agarose gel electrophoresis. 5. 96-Well Thermal Cycler. 6. NucleoSpin™ Gel and PCR Clean-up kit (Macherey-Nagel™). 7. NanoDrop spectrophotometer. 8. FastDigest DpnI (Thermo Fisher Scientific, cat. FD1703). 9. T4 DNA ligase (Neb). 10. DH5α competent cells. 11. LB-agar medium supplemented with the proper antibiotic. 12. 37 °C incubator. 13. LB-medium supplemented with the proper antibiotic.

Characterizing Post-transcriptional Modifications of circRNAs

277

14. NucleoSpin Plasmid (Macherey-Nagel™). 15. NucleoBond Xtra Midi EF (Macherey-Nagel™). 16. Cells of interest. 17. Cell medium. 18. Tissue culture incubator. 19. Appropriate tissue culture plates and supplies. 20. Lipofectamine 2000 (Life Technologies). 21. Opti-MEM (Gibco). 22. Centrifuge. 23. Protein buffer: 100 mM Tris–HCl pH 7.5, 1 mM EDTA, 2% SDS, 1× cOmplete™ Protease Inhibitor Cocktail (Roche, cat. 11697498001). 24. Rotator. 25. Cell lifter. 26. Refrigerated centrifuge. 27. Equipment and reagents for Western blot analyses. 28. antibodies against writers/readers/erasers. 29. TRIzol™ Reagent (Thermo Fisher Scientific). 30. Direct-zol RNA Miniprep kit (Zymo Research). 31. PrimeScript RT Master Mix (Takara Bio, cat. RR036B). 32. NanoDrop spectrophotometer. 33. PowerUp SYBR Green Master Mix reagent (Thermo Fisher Scientific). 34. q-PCR primers for the mature circRNA, its linear counterpart, and their common precursor (see Note 1). 35. 96-Well Thermal Cycler. 36. Real-Time PCR System. 37. Optical 96-Well Reaction Plate and Adhesive Film. 38. MPS 1000 Mini Plate Spinner. 2.8 Assessing Writers/Readers/ Erasers Binding on circRNAs

1. Cells of interest. 2. Cell medium. 3. Tissue culture incubator. 4. Appropriate tissue culture plates and supplies. 5. PBS 1×. 6. Stratalinker (Agilent). 7. RIPA buffer: 20 mM Tris–HCl pH 8, 100 mM NaCl, 0.5 mM EDTA, 0.5% NP-40, 0.1% SDS,1× cOmplete™ Protease

278

Gaia Di Timoteo et al.

Inhibitor Cocktail (Roche, cat. 11697498001), and RiboLock RNase inhibitor 1× (Thermo Fisher Scientific, cat. EO0382). 8. 21G needle. 9. Refrigerated centrifuge. 10. Bradford reagent. 11. Rotator. 12. Specific antibody. 13. IgG antibody (Merck). 14. Protein A/G Dynabeads (Thermo Fisher Scientific). 15. Magnetic separation stand. 16. Wash Buffer: 50 mM Tris–HCl pH 7.4, 150 mM NaCl, 1 mM MgCl2, and 0.05% NP-40. 17. High-Salt Wash Buffer: 50 mM Tris–HCl pH 7.4, 500 mM NaCl, 1 mM MgCl2, 0.05% NP-40. 18. Proteinase K Buffer: 10 mM Tris–HCl pH 7.4, 100 mM NaCl, 1 mM EDTA, 0.2% (vol/vol) SDS. 19. Proteinase K (Sigma-Aldrich). 20. TRIzol™ Reagent (Thermo Fisher Scientific). 21. Chloroform. 22. Isopropanol. 23. 75% Ethanol. 24. RNase-free glycogen. 25. NanoDrop spectrophotometer. 26. SuperScript™ VILO™ cDNA Synthesis Kit (Thermo Fisher Scientific, cat. 11754250). 27. 96-Well Thermal Cycler. 28. Optical 96-Well Reaction Plate and Adhesive Film. 29. MPS 1000 Mini Plate Spinner. 30. PowerUp SYBR Green PCR Master Mix (Applied Biosystems). 31. Real-Time PCR System. 32. q-PCR primers positive and negative controls. 33. q-PCR primers for the targets of interest (see Note 1).

3 3.1

Methods circRNA Isolation

1. After cell lysis with the lysis reagent provided in the miRNeasy Kit, isolate total RNA following the manufacturer’s instructions. Quantify the extracted RNA with a NanoDrop spectrophotometer.

Characterizing Post-transcriptional Modifications of circRNAs

279

The subsequent steps have been optimized for 5 μg total RNA input. Considering that circRNAs represent a small fraction of total RNA, it is recommended to perform parallel reactions for a total of at least 100 μg total RNA input. 2. Digest 5 μg total RNA in a 20 μL reaction containing 2 μL 10× RNase R reaction buffer, 0.2 μL RIboLock, 1 μL TURBO DNase, and 1 μL RNase R (20 U/μL), a strong 3’–5’ exonuclease which enriches for circRNAs and intronic lariats. Prepare another reaction without RNase R (by adding 1 μL RNase-free water) to be used as a control for the treatment. Incubate the reactions for 30 min at 37 °C, then inactivate by adding 10 volumes (200 μL) of TRIzol™ Reagent, and extract the RNA following the manufacturer’s protocol. 3. Use a fraction of the extracted RNA (e.g., 200 ng of the control sample and the same volume of the RNase R-treated sample) to check for linear RNA degradation (e.g., through RT-qPCR on selected linear species). 4. As RNase R is not able to digest intronic lariats, doublestranded RNAs, or other species (e.g., snRNAs) with 3’ overhangs shorter than 7 nucleotides, a polyadenylation step is used to remove any residual linear RNAs from the previous treatment. Prepare a 40 μL reaction containing 20 μL RNase R-treated RNA, 4 μL 10× poly(A) polymerase buffer, 1 μL poly(A) polymerase (E-PAP), 4 μL 10 mM ATP solution, 4 μL 25 mM MnCl2, and 0.4 μL RiboLock. Prepare another reaction without E-PAP (by adding 1 μL RNase-free water) to be used as a control for the treatment. The reactions are incubated for 30 min at 37 °C. 5. Wash 10 μL oligo-dT magnetic beads from Poly(A)Purist™ MAG Kit three times in 100 μL 1× binding buffer provided in the kit and recover them on a magnetic stand. 6. Dissolve the washed beads in 40 μL 2× binding buffer provided with the kit and add them to the 40 μL polyadenylation reaction. Incubate both the polyadenylation reaction with magnetic beads and the control reaction for 5 min at 75 °C and then for 20 min at 25 °C with shaking. 7. After 1-min incubation on a magnetic stand, collect the supernatant and extract RNA from this fraction through the miRNeasy Mini Kit following the manufacturer’s instructions, from both the control and the RPAD samples. 8. Reverse-transcribe equal volumes of each sample using the Mir-X™ miRNA first strand synthesis kit (for small RNAs) or the Maxima reverse transcriptase, according to the manufacturer’s protocol.

280

Gaia Di Timoteo et al.

9. Validate the depletion of linear RNAs by RT-qPCR on a QuantStudio 6 Real-Time PCR System for a subset of circRNAs and linear RNAs, including snRNAs. Do this for control RNA, RNase R-treated RNA, and RPAD-prepared RNA. Calculate the recovery after the RPAD protocol for the circRNA or linear RNA of interest as the fold change (2–ΔCt) with respect to the control treatment or the RNase R treatment alone. 3.2 m6A Immunoprecipitation 3.2.1

MeRIP-seq

1. For each sample needed for the immunoprecipitation, dilute 2.5–5 μg high-purity circRNA (see Notes 3 and 4) in 18 μL nuclease-free water and chemically fragment it by adding 2 μL 10× Fragmentation Buffer and incubating the solution in a thermomixer for 15 min at 70 °C. 2. Put the samples immediately on ice and stop the reaction by adding 2.2 μL Stop Solution. 3. Purify the fragmented RNA by adding TRIzol™ Reagent and following the manufacturer’s protocol. 4. Optional: RNA post-fragmentation size distribution and concentration can be validated by Agilent 2100 Bioanalyzer with an Agilent RNA 6000 Pico kit, or by measuring with a NanoDrop spectrophotometer and by running 0.5 μg of RNA on 1.5% (wt/vol) agarose gel. This step produces fragments approximately 100 nt long. If a different size range is desired, then the temperature and reaction time of the fragmentation process should be adjusted concordantly. 5. Save a fraction (e.g., 10%) of the sample as “input” for sequencing, then dilute the leftover in 450 μL IP buffer with 2.5–5 μg anti-m6A antibody (1 μg for every μg of high-purity circRNA), or the same quantity of IgG. 6. Incubate the samples for 2 h at 4 °C rotating head over tail. 7. Wash 30 μL Dynabeads™ protein A/Protein G twice with 500 μL IP buffer and then resuspend them in 50 μL IP buffer. 8. Incubate the samples with pre-washed Dynabeads for 2 h at 4 ° C rotating head over tail. 9. Recover bead-bound antibody–RNA complexes on a magnetic stand. 10. Optional: the liquid phase can be saved and further processed as the “supernatant” of the m6A IP, representing non-methylated circRNA species. 11. Wash the Bead-bound antibody–RNA complexes three times by adding 500 μL IP buffer, incubating on a rotator for 2 min at RT, placing on a magnetic stand for 1 min, and removing the supernatant.

Characterizing Post-transcriptional Modifications of circRNAs

281

12. Elute bead-bound RNA by adding 100 μL IP buffer with 0.5 mg/mL N6-methyladenosine and incubating for 1 h at 4 °C in a thermomixer with continuous shaking at 1100 rpm. 13. Optional: Repeat the above step to ensure complete dissociation and combine all elutes from the same sample. 14. Purify the RNA from each sample by adding TRIzol™ Reagent and following the manufacturer’s protocol. 15. Use the purified RNA for library preparation with available kits (the amount will depend on the kit of choice), such as Stranded Total RNA Prep with Ribo-Zero Plus. As RNA fragmentation is already performed in the meRIP protocol, this step is skipped during library preparation. 3.2.2

miCLIP

1. For each sample needed for the immunoprecipitation, dilute 2.5–5 μg high-purity circRNA (see Note 3) isolated from the step above in 18 μL nuclease-free water and chemically fragment it by adding 2 μL 10× Fragmentation Buffer and by incubating the solution in a thermomixer for 15 min at 70 °C. 2. Put the samples immediately on ice and stop the reaction by adding 2.2 μL Stop Solution. 3. Purify the fragmented RNA by adding TRIzol™ Reagent and following the manufacturer’s protocol. 4. Optional: RNA post-fragmentation size distribution and concentration can be validated by Agilent 2100 Bioanalyzer with an Agilent RNA 6000 Pico kit, or by measuring with a NanoDrop spectrophotometer and by running 0.5 μg of RNA on 1.5% (wt/vol) agarose gel. Ideally, this step produces fragments in the range of 30–130 nt. If this does not happen, then the temperature and reaction time of the fragmentation process should be further optimized. 5. Save 10% of the sample as (non-crosslinked) input, then dilute the leftover in 450 μL IP buffer with 2.5–5 μg anti-m6A antibody (1 μg for every μg of high-purity circRNA), or the same quantity of IgG (see Note 5). 6. Incubate the reactions for 2 h at 4 °C rotating head over tail. 7. Wash 30 μL Dynabeads™ protein A/Protein G twice with 500 μL IP buffer and then resuspend them in 50 μL IP buffer. 8. Transfer the antibody-bound RNA into a 35-mm cell culture dish (make sure that the solution covers the entire well surface) and crosslink it twice with 0.15 J cm-2 UV light (254 nm) in a Stratalinker (see Note 6). The plate lid must be removed during crosslinking.

282

Gaia Di Timoteo et al.

9. Put the RNA into a new 2-mL Eppendorf tube on ice, and then incubate with pre-washed Dynabeads for 2 h at 4 °C rotating head over tail. 10. Recover bead-bound antibody-RNA complexes on a magnetic stand. 11. Wash the Bead-bound antibody–RNA complexes twice with 500 μL High-Salt Buffer (at least 1 min in the cold room), twice with 500 μL IP buffer, and twice with 500 μL PNK Wash Buffer. 12. Recover bead-bound antibody–RNA complexes on a magnetic stand, resuspend them in 500 μL PNK Wash Buffer, and transfer them to a new 2-mL Eppendorf tube to limit the carry-over of RNases, DNases, and salts to the next step. 13. Recover bead-bound antibody–RNA complexes on a magnetic stand and resuspend them in 4 μL 5× PNK Buffer, 0.5 μL PNK, 0.5 μL RiboLock, and 15 μL water. Incubate the 3′-end RNA dephosphorylation reaction for 30 min at 37 °C in a thermomixer with continuous shaking at 1100 rpm. 14. Put the reaction on a magnetic stand, remove the supernatant, and wash with 500 μL PNK Wash Buffer. 15. Resuspend the complexes in a 20 μL reaction containing 11 μL water, 2 μL 10× T4 RNA ligase buffer, 1 μL T4 RNA ligase I, 0.5 μL RiboLock, 1.5 μL L3 linker, and 4 μL PEG400. Incubate the reaction overnight at 16 °C in a thermomixer with continuous shaking at 1100 rpm (see Note 7). 16. Wash the bead-bound antibody–RNA complexes twice with 500 μL High-Salt Buffer (at least 1 min in the cold room) and twice with 500 μL PNK Wash Buffer (see Note 8). 17. Recover bead-bound antibody-RNA complexes on a magnetic stand and resuspend them in 20 μL mix containing 14 μL water, 5 μL 4× LDS buffer, and 1 μL 1 M DTT, and then incubate for 10 min at 70 °C in a thermomixer with continuous shaking at 1100 rpm. 18. Recover the beads on a magnetic stand and load the supernatant on a 4–12% bis-tris protein gel with 1× MES SDS Running buffer. Load a prestained protein ladder on a separate lane. The running is performed in a cold room at 180 V for 1 h. 19. Cut off the dye front and transfer the RNA–antibody complexes to a nitrocellulose membrane in a cold room by blotting for 2 h at 30 V. 20. Visualize the transferred RNA-antibody complexes (e.g., on an infrared imaging imager system such as the Odyssey CLx from LI-COR) and use the image printout as a cutting template mask, then cut the appropriate region of the nitrocellulose

Characterizing Post-transcriptional Modifications of circRNAs

283

membrane from each lane with a sterile razor, and place it in a new 2-mL Eppendorf tube. 21. Add 190 μL Proteinase K Buffer and 10 μL Proteinase K to each 2-mL tube (making sure that the nitrocellulose membrane piece is submerged), and then incubate the solution in a thermomixer for 1 h at 50 °C with continuous shaking at 1100 rpm. 22. Collect the solution and move it to a new 2-mL Eppendorf tube, then add 200 μL Phenol:Chloroform:Isoamyl alcohol, and transfer the solution to a pre-spun 2-mL Phase Lock Gel Heavy tube. 23. Incubate the solution in a thermomixer for 5 min at 30 °C with continuous shaking at 1100 rpm, and then separate the phases with centrifugation for 5 min at 16,000 × g at room temperature. 24. Transfer the supernatant to a new 2-mL Eppendorf tube, then add 1 μL GlycoBlue, 40 μL 3 M sodium acetate pH 5.5, and 1 ml 100% ethanol, and precipitate the RNA overnight at 20 °C. 25. Spin down the RNA fragments in a refrigerated centrifuge for 20 min at 20,000 × g, then discard the supernatant, add 800 μL 75% ethanol, and then spin down the solution in a refrigerated centrifuge for 5 min at 20,000 × g. 26. Discard the supernatant and allow the pellet to air dry for 5–10 min. 27. Resuspend the RNA pellet in 5 μL water, and then add 1 μL 1 pmol/μL RT primer and 1 μL 10 mM dNTPs for the reverse transcription, which is performed in a thermocycler at 70 °C for 5 min, then 25 °C hold, and then add a mix with 2 μL Superscript IV buffer, 0.25 μL Superscript IV, 0.25 μL RiboLock, and 0.5 μL 0.1 M DTT. Thermocycling is performed at 25 °C for 5 min, 50 °C for 5 min, 55 °C for 5 min, and 4 ° C hold. 28. Remove the RNA by adding 0.5 μL 5 U/μL RNase H and 0.5 μL 1 ng/μL RNase A and by incubating for 15 min at 37 ° C. 29. Purify cDNA with AMPure XP beads (using 3× RT volume), according to the manufacturer’s instructions, and then resuspend it in a 15 μL reaction mix containing 9 μL water, 1.5 μL CircLigase Buffer II, 0.75 μL CircLigase II, 3 μL 5 M betaine, and 0.75 μL 50 mM MnCl2. The circularization is performed in a thermocycler at 60 °C for 2 h and 4 °C hold. 30. Purify circularized cDNA with AMPure XP beads (using 3× circularization volume), according to the manufacturer’s instructions, and then resuspend it in 10 μL water.

284

Gaia Di Timoteo et al.

31. Use 1 μL of circularized cDNA to perform PCR test amplification reactions with 3.75 μL water, 0.25 μL P5/P3 Solexa PCR primer mix, 5 μL Phusion High-Fidelity PCR Master mix in a thermocycler at 98 °C for 30 s, then 15–25 cycles (98 °C for 15 s, 65 °C for 30 s, and 72 °C for 40 s), then 72 °C for 5 min, and hold at 25 °C. 32. Run test PCR reactions on a 6% TBE gel (at around 180 V for 30 min), with a ladder run in parallel. In order to determine the appropriate number of PCR cycles for the library amplification (see Note 9), visualize the products on a UV transilluminator after staining the gel with SYBR gold (1:10,000) in a gel-running buffer for 10 min. If the library was correctly amplified, a signal should be faintly visible in the range of 150–250 nt (also longer products may be observed, depending on the size distribution of the RNA after the initial fragmentation). If too many cycles were performed, overamplification (i.e., a smear with a nonspecific signal) will be observed. 33. With the same approach as the preparative PCR, use the entire circularized cDNA or half of it for the final library amplification (using a number of cycles close to the optimal one determined in the previous step, also considering that, for the final PCR, a higher amount of template will be used; thus, one or two cycles less should be performed) and run a fraction of the reaction on a 6% TBE gel (at around 200 V for 30 min) to check for proper amplification, which should give signal similar to the one of the test PCR. 34. Save a fraction of each PCR sample for potential re-sequencing, and then mix samples with different barcodes for multiplexing, adjusting the quantity in accordance with the intensity of their signal on the gel. Excise the cDNA libraries from the gel (usually the signal above 145 nucleotides, in order to exclude the artifacts containing P5/P3 primers + barcode, and below 300–400 nucleotides) (see Note 10) and extract them with commercially available kits as the QIAquick Gel Extraction Kit. 35. Quantify gel-purified cDNA libraries and send them for sequencing. As RNA fragmentation is already performed in the miCLIP protocol, this step is skipped during library preparation (see Note 11). 3.3

Peak Calling

3.3.1

MeRIP-seq

1. Use the SRA Toolkit to convert the SRA files from sequencing to .fastq format. 2. Use the FastQC tool to perform quality control checks—which include average good per base quality score, per sequence GC content around 50%, homogeneous sequence length

Characterizing Post-transcriptional Modifications of circRNAs

285

distribution, little to no adapter content (if the adapters have been removed by the sequencing facility upon demultiplexing of the samples), homogeneous per base sequence content, moderate sequence duplication levels—on the raw data, to exclude possible biases that could affect results. 3. Remove bad-quality reads with tools such as Cutadapt by setting a minimum read length to 20–25 bp and a minimum Phred-quality score of 20–25. 4. Align processed reads to the reference genome by using a short-read aligner such as BWA, Bowtie2, TopHat, or STAR. 5. Detect m6A peaks with tools such as MACS2, ExomePeak, MeTPeak, MeTDiff, and m6AViewer. 6. Quantify m6A peaks in the IP over the input with R package tools such as DESeq2 or edgeR, considering only significant results (e.g., adjusted p-value 95 °C) and then immediately place it on ice. Add a denatured, cold probe to the lysate.

CircRNA Pulldown for Profiling RNA and RNA Interactomics

305

12. Incubate on a rotating wheel at 4 °C for 4 h to allow the probe to hybridize to the circRNA. 13. Add 600 μL of prepared (cleaned and blocked) Dynabeads in 1× Native Binding buffer. 14. Incubate at room temperature on a rotating wheel for 1 h to allow streptavidin beads to bind to the Biotin-TEG tag on the probe. 15. Use a magnetic rack to pulldown the beads and wash them four times with 1 mL 1× B&W buffer. During each wash, ensure that the beads are dislodged from the side of the tube by removing the microcentrifuge tube from the magnetic rack and gently resuspend by pipetting up and down with a 1-mL pipette tip. 16. Perform a final wash in RNAse-free water to remove salts and proceed to either protein elution (Subheading 3.3) or RNA elution (Subheading 3.4). 3.3 Profiling RNA Interactions 3.3.1 Elution and Purification of RNA from Beads

1. Elute RNA (and proteins) by adding 500 μL TRIzol to the beads and keep them at room temperature for 5 min occasionally disturbing the beads by flick mixing. STOP POINT: You can now store this slurry at –20 °C to – 80 °C prior to RNA purification. 2. RNA can be purified by standard solvent-based purification, or through a column—Direct-zol™ RNA miniprep kit (Zymo Research) (see Note 8). 3. Purified RNA is commonly used for reverse transcription or library preparation (see Note 9). Dissolve in the required amount of water and proceed with your RT protocol or library preparation (see Note 10).

3.3.2 Validation of circRNA Pulldown with qRT-PCR

1. Design oligonucleotide primers to detect both your target circRNA and its cognate parental linear mRNA (see Note 11). As a reference gene, use any common and abundant mRNA target gene (e.g., GAPDH). For circRNA reference gene, use, e.g., circHIPK3 [2] or circSMARCA5 [17, 18]. 2. Design and perform a control experiment on cells that do not overexpress the circRNA of interest to measure pulldown efficiency specificity—circRNA and parental mRNA (as the most likely RNA transcript that may be pulled down) (see Notes 12 and 13).

3.4 Profiling Protein Interactions

1. Perform protein elution from beads in 150 μL of 5 mM Tris– HCl, pH 8.0 with 1 μg of RNAse A to digest the bound RNA. Incubate for 18–24 h at 37 °C on a rotation wheel.

306

Marta M. Gabryelska et al.

2. Concentrate pulldown material with Centricon spin columns (10 kDa MWCO), and calculate protein concentrations using standard BCA protocol (e.g., Bio-RAD Protein Assay Dye reagent concentrate). 3. Subject 1 μg of total protein to in-solution trypsin digestion where the protein sample is incubated 30 min at 56 °C in the following conditions: 111 mM Tris–HCl pH 8.0, 1.1 mM CaCl2, and 5.6 mM DTT in 18 μL reduction reaction. 4. Subsequently, add 1 μL of 200 mM 2-Chloroacetamide (Sigma) to the above reaction after it has cooled to room temperature and incubate in the dark at room temperature for 30 min. 5. Digest protein samples in a 1:50 Enzyme:Protein ratio reaction, containing 19 μL of the above reaction mixture and Trypsin (Pierce) solution to the final concentration of 1 ng/μL. 6. Vortex samples and incubate at 37 °C for 18–24 h, transfer to mass spectrometry (MS) vials, and analyze by MS within 24 h. 7. Three biological replicates were prepared for both the positive (antisense) and negative (sense) probes. 8. Establish a filtered candidate list of interacting proteins with the following stringent criteria: (a) Reject protein hits without high levels of confidence, including where only a single peptide is identified. (b) Subsequently, eliminate high-confidence proteins found in negative control pulldowns from those found in positive pulldowns. (c) Remove BSA (used to block Dynabeads) and known biotinylated proteins from the candidate list as these would bind Streptavidin-Dynabeads. (d) Compare three replicates. (e) Protein hits meeting the aforementioned criteria and found in at least two replicates are consolidated into the final list of interacting proteins. 9. Use pulldown material to verify the protein of interest, including verification and quantification via Western blot. 10. Perform a reverse immunoprecipitation by using antisera specific to the candidate protein to verify the interaction with confirmation of the bound circRNA by qRT-PCR.

4

Notes 1. Biotin-TEG label can be at either the 3’ or 5’ end, but the advantage of utilizing the 3’ label is that it eliminates the ability

CircRNA Pulldown for Profiling RNA and RNA Interactomics

307

of the reverse transcriptase to utilize the labeled probe as a primer, which would confound the downstream analysis. 2. Add 1/10 of the final volume of 3 M sodium acetate, pH 5.2, and extract the solution twice with phenol:chloroform:isoamyl alcohol (25:24:1, pH 6.7). Precipitate the RNA with 2.5 volumes of ethanol at room temperature and recover the RNA by centrifugation at 13,000 g for 15 min at 4 °C. Dissolve the pellet of RNA at a concentration of 10 mg/mL in sterile TE (pH 7.6). Store the carrier RNA in small aliquots at –20 °C. 3. Native Binding buffer and B&W buffer need to be filtersterilized through a 0.2-μm filter and can be stored at room temperature. 4. To utilize the Co-IP buffer you need to add protease inhibitors (cOmplete™, Mini, EDTA-free Protease Inhibitor Cocktail (Roche), at 1 tablet/10 mL solution) just prior to the experiment. It is possible to store the co-IP buffer with protease inhibitors at –20 °C for up to 12 weeks with no appreciable loss of RNA or protein recovery following pulldown. 5. The required input cell number is largely dependent on the abundance of the circRNA target. This should be assessed in your target cells by quantitative PCR. We have successfully enriched for circRNA targets that range in cycle threshold (Ct) values of 16–35 (cDNA diluted 1:5 following reverse transcription on 1 μg RNA from cell culture, 24 h posttransfection of circRNA overexpression construct). If the target molecule is below this level of abundance, it is advisable to overexpress the circRNA and harvest the cells 18–48 h posttransfection. 6. This protocol (Fig. 2) lyses the cells but does not overcome native interactions between the circRNA and its RNA or protein cargo. After the pulldown, the analysis is a comparison of the presence and abundance of RNA transcripts (by qRT-PCR) or proteins (by peptide digestion followed by mass spectrometry) compared with the control pulldown and input material. 7. One of the most crucial factors is to optimize the ratio of the 100-μM probe: Dynabead volume—our optimum conditions are a ratio of 1:24. Once optimized, this allows for a reduction in the total amount of each reagent with successful pulldown routinely achieved with half the recommended starting amounts. 8. If choosing TriZOL purification, it is optional to retain the interphase and lower solvent (pink) phase after chloroform extraction which contains interacting proteins. Proteins can be purified using standard guanidine hydrochloride protocol and can be digested and analyzed by mass spectrometry.

308

Marta M. Gabryelska et al.

9. RNA yield can vary greatly but commonly ranges from 100 pg to 50 ng total recovered RNA which may contain traces of yeast tRNA, which has previously been shown to cause no detectable distortion of global RNA amplification [22]. 10. Due to the low RNA yield, ensure you select a library preparation protocol to accommodate this ultralow RNA input, including NEBNext® Single Cell/Low Input RNA Library Prep Kit for Illumina® (E6420). 11. Ensure the parental mRNA oligonucleotide primers target regions outside the circRNA region to avoid confounding. 12. Optional: prepare a negative control pulldown experiment including blocked Dynabeads™ without biotinylated probes to assess non-specific binding levels. 13. Compared with the control, we see enrichment of our target circRNA by 5–50-fold compared with the input, while no circRNA was detected in either of the negative controls. The cognate mRNA is commonly found at very low levels in both the sense and antisense pulldowns. Target RNA molecules are enriched two–ten-fold.

Acknowledgments Research reported in this publication was supported by the National Health and Medical Research Council (NHMRC) project grant funding to S.J.C. (GNT1144250). Fellowship support was provided by the Australian Research Council Future Fellowship to S.J.C. (FT160100318) and the NHMRC Investigator Leadership Grant to S.J.C. (GNT1198014). Fellowship support for B.W.S. was provided by the Flinders Foundation Brain Cancer Fellowship. We would like to acknowledge Flinders Proteomics for their guidance and expertise in regards to mass spectrometry methodologies. References 1. Conn S, Pillman K, Toubia J et al (2015) The RNA binding protein quaking regulates formation of circRNAs. Cell 160:1125–1134 2. Jeck WR, Sharpless NE (2014) Detecting and characterizing circular RNAs. Nat Biotechnol 32:453–461. https://doi.org/10.1038/nbt. 2890 3. Rybak-Wolf A, Stottmeister C, Glazˇar P et al (2015) Circular RNAs in the mammalian brain are highly abundant, conserved, and dynamically expressed. Mol Cell 58:870–885. https:// doi.org/10.1016/j.molcel.2015.03.027

4. Jeck WR, Sorrentino JA, Wang K et al (2013) Circular RNAs are abundant, conserved, and associated with ALU repeats. RNA 19:141– 1 5 7 . h t t p s : // d o i . o r g / 1 0 . 1 2 6 1 / r n a . 035667.112 5. Panda AC (2018) Circular RNAs act as miRNA sponges. Adv Exp Med Biol 1087:67–79. https://doi.org/10.1007/978-981-131426-1_6 6. Hansen TB, Jensen TI, Clausen BH et al (2013) Natural RNA circles function as

CircRNA Pulldown for Profiling RNA and RNA Interactomics efficient microRNA sponges. Nature 495:384– 388. https://doi.org/10.1038/nature11993 7. Memczak S, Jens M, Elefsinioti A et al (2013) Circular RNAs are a large class of animal RNAs with regulatory potency. Nature 495:333–338. https://doi.org/10.1038/nature11928 8. Lasda E, Parker R (2014) Circular RNAs: diversity of form and function. RNA 20: 1829–1842. https://doi.org/10.1261/rna. 047126.114 9. You X, Vlatkovic I, Babic A et al (2015) Neural circular RNAs are derived from synaptic genes and regulated by development and plasticity. Nat Neurosci 18:603–610. https://doi.org/ 10.1038/nn.3975 10. Du WW, Yang W, Liu E et al (2016) Foxo3 circular RNA retards cell cycle progression via forming ternary complexes with p21 and CDK2. Nucleic Acids Res 44:2846–2858. https://doi.org/10.1093/nar/gkw027 11. Guarnerio J, Bezzi M, Jeong JC et al (2016) Oncogenic role of fusion-circRNAs derived from cancer-associated chromosomal translocations. Cell 165:289–302. https://doi.org/ 10.1016/j.cell.2016.03.020 12. Liu X-Y, Zhang Q, Guo J et al (2022) The role of circular RNAs in the drug resistance of cancers. Front Oncol 11 13. Gabryelska M, Conn SJ (2023) The RNA Interactome in the hallmarks of cancer, vol 14. WIREs RNA In press 14. Conn VM, Gabryelska M, Toubia J, Kirk K, Gantley L, Powell JA, Cildir G, Marri S, Liu R, Stringer BW, Townley S, Webb ST, LinH, Samaraweera SE, Bailey S, Moore AS, Maybury M, Liu D, Colella AD, Chataway T, Wallington-Gates CT, Walters L, Sibbons J, Selth LA, Tergaonkar V, D’Andrea RJ, Pitson SM, Goodall GJ, Conn SJ (2023) Circular RNAs drive oncogenic chromosomaltranslocations within the MLL recombinome in leukemia. Cancer Cell 41(7):1309–1326

309

15. Conn VM, Hugouvieux V, Nayak A, Conos S, Capovilla G, Cildir G, Jourdain A, Tergaonkar V, Schmid M, Zubieta C, Conn SJ (2017) A CircRNA from SEPALLATA3 Regulates Splicing of its Cognate mRNA Through R-loop Formation. Nature Plants 3:17053 16. Du WW, Zhang C, Yang W et al (2017) Identifying and characterizing circRNA-protein interaction. Theranostics 7:4183–4191. https://doi.org/10.7150/thno.21299 17. Huang A, Zheng H, Wu Z et al (2020) Circular RNA-protein interactions: functions, mechanisms, and identification. Theranostics 10: 3503–3517. https://doi.org/10.7150/thno. 42174 18. Ulsho¨fer CJ, Pfafenrot C, Bindereif A, Schneider T (2021) Methods to study circRNAprotein interactions. Methods 196:36–46. https://doi.org/10.1016/j.ymeth.2021. 04.014 19. Stringer BW, Gabryelska M, Marri S, Clark L, Lin H, Gantley L, Liu R, Wilusz JE, Conn VM, Conn SJ (2023) Versatiletoolkit for highlyefficient and scarless overexpression of circular RNAs. bioRxiv.11.21.568171. https://doi. org/10.1101/2023.11.21.568171 20. Liu D, Conn V, Goodall GJ, Conn SJ (2018) A highly efficient strategy for overexpressing circRNAs. Methods Mol Biol 1724:97–105. https://doi.org/10.1007/978-1-49397562-4_8 21. Conn VM, Gabryelska M, Marri S et al (2020) SRRM4 expands the repertoire of circular RNAs by regulating microexon inclusion. Cell 9 : 2 4 8 8 . h t t p s : // d o i . o r g / 1 0 . 3 3 9 0 / cells9112488 22. Wang QT, Xiao W, Mindrinos M, Davis RW (2002) Yeast tRNA as carrier in the isolation of microscale RNA for global amplification and expression profiling. Biotechniques 33:788, 790, 792. https://doi.org/10.2144/ 02334st02

Chapter 17 Screening and Characterization of Functional circRNAs in Neuronal Cultures Darren Kelly and Gerhard Schratt Abstract This chapter describes a methodology for the screening and characterization of functional circRNAs, particularly in the context of neural circuit development. Taking advantage of a primary rat neuron culture model of synaptogenesis, we propose a means of selecting from the plethora of circRNA species based on their expression levels, dendritic localization, conservation, and activity regulation. These candidates are then knocked down with RNAi approaches in a functional screen for their potential role in the formation and maturation of excitatory synapses. Upon identification of top candidates regulating synaptogenesis, we tie together different “Omics” approaches to explore the molecular mechanisms underlying the phenotypes observed upon circRNA knockdown. We utilized our EnrichMir algorithm to identify overrepresented miRNA binding sites in differentially expressed genes from polyA-RNA-seq following circRNA knockdown. Furthermore, our ScanMiR web tool allows for the miRNA binding prediction of reconstructed internal circular RNA sequences. Small-RNA sequencing is used to monitor changes in miRNA levels in the circRNA knockdown to complement results obtained from EnrichMiR. Finally, the experimental validation of promising miRNA–circRNA pairs sets the stage for in-depth biochemical exploration of the circRNA interactome and mechanism of action. Key words Circular RNA, Neurons, Dendritic processes, Synapses, Enrichment, Activity regulation, RNA interference, Functional screen, microRNA, Binding site discovery, ScanMir, EnrichMir, Molecular mechanism, Small-RNA-seq, PolyA-seq, Compartment-enriched circRNAs

1

Introduction The circular RNA (circRNA) transcriptome in the nervous system is particularly diverse and abundant [1]. CircRNAs are highly expressed in developing neurons, often localized to the synaptodendritic compartment, and many are derived from genes with synaptic functions [2]. The loss of specific circRNAs (e.g., Cdr1-as) has been shown to affect synaptic transmission, specifically through the regulation of

Christoph Dieterich and Marie-Laure Baudet (eds.), Circular RNAs, Methods in Molecular Biology, vol. 2765, https://doi.org/10.1007/978-1-0716-3678-7_17, © The Author(s), under exclusive license to Springer Science+Business Media, LLC, part of Springer Nature 2024

311

312

Darren Kelly and Gerhard Schratt

microRNA (miRNA) stability in complex ncRNA regulatory networks [3, 4]. The role of miRNAs in synaptic plasticity through local regulation of protein synthesis has been well established, and miRNA activity can be regulated by neuronal activity/depolarization [5]. circRNAs are thus implicated in the regulation of synaptic plasticity and control of local protein synthesis with relevance in the pathogenesis of neurodevelopmental disorders. This prompts the systematic investigation of circRNAs enriched in specific neuronal processes as well as their regulation by neuronal activity. However, the vast numbers, tissue specificity, and splicing isoform variability of circRNAs make it difficult to distinguish those with a functional role from a majority produced as presumably non-functional splice variants. circRNAs are also a diverse group of transcripts, and it is probable that they could affect neuronal function through a variety of alternate mechanisms of action, such as an association with cytoplasmic RNA-binding proteins (RBPs) or the nuclear splicing machinery [6]. To address this complexity, we describe a method for the screening of functional circRNAs in specific neuronal tissue types. This approach utilizes high-throughput sequencing techniques performed on ribosomal RNA (rRNA)-depleted cellular RNA fractions from compartmentalized primary neurons cultured from rat hippocampus as a model. We also describe a paradigm to induce in these cultures homeostatic downscaling, a specific form of synaptic plasticity that is associated with highly compartment-specific changes in the neuronal transcriptome [7]. This allows for the detection of numerous circRNAs enriched in the dendritic processes over the soma, with additional filtering steps to identify highly expressed, conserved, and activity-regulated circRNA candidates. We then knock down these top circRNA candidates through RNA interference (RNAi) approaches to identify their potential role in excitatory synapse formation and maturation. We describe a screen for the impact of circRNA knockdown on excitatory synapse density as determined by the colocalization of presynaptic (Synapsin) and postsynaptic (PSD-95) markers within transfected neurons. The functional screen can be replaced depending upon the cell type, scalability, and particular cellular process of interest, making this an approach widely applicable to studying cellular functions of circRNAs. We outline essential validation steps in this process to confirm the fidelity of the screen reflecting the accepted standards in the field [8]. Confirmation of knockdown efficiency and specificity is achieved through highly penetrant electroporation of the small interfering RNA (siRNA)/short-hairpin RNAs (shRNAs) targeting the circRNA BSJ and circRNA detection through qPCR with divergent primers.

Screening Neuronal circRNAs

313

Upon identification of circRNAs with a functional role in a specific cellular process, characterization of the molecular mechanism by which the circRNA impacts the phenotype is still challenging. We describe downstream “Omics” approaches to identify the molecular consequences of the circRNA knockdown agnostic to the sequence of the circRNA itself. The identification of differentially expressed genes (DEGs) by poly-A RNA-sequencing upon circRNA knockdown is used to provide additional support for morphological phenotypes observed in the screen. Our EnrichMir web tool [9] then allows for the identification of an enrichment or depletion of predicted miRNA binding sites within DEGs, which provides miRNA candidates mediating the effects of circRNA knockdown. The sequence of the circRNA itself can also be independently inspected to deduce specific interacting miRNAs. For example, our ScanMir web tool [10] can be used to identify potential miRNA binding sites within the reconstructed internal circRNA sequence. In conjunction with small-RNA sequencing, changes in the levels of these miRNAs can then be verified. Published AGO-IP or AGO-CLIP datasets can also support the interaction between this sequence and the miRNA of interest. Predictive analysis of circRNA function early on can be advantageous, but an unbiased approach can also allow for novel discoveries. In conjunction, bias and unbiased approaches coinciding with shared candidates increase our confidence in a mechanism of action. This pipeline sets the stage for further biochemical validations of circRNA interactions and the understanding of intricate ncRNA regulatory networks in neurons and other cell types.

2

Materials

2.1

Hardware

Confocal laser scanning microscope (Zeiss, CLSM 880). Nanodrop Spectrophotometer (Thermo Fisher Scientific). Amaxa™-4D Nucleofector™ (Lonza). HiSeq2500™ (Illumina).

2.2

Software

DCC, circRNA detection tool (https://github.com/dieterich-lab/ DCC) [11]. CircTest, circRNA testing package (https://github.com/ dieterich-lab/CircTest) [11]. topGO 2.52.0: Enrichment Analysis for Gene Ontology (https://bioconductor.org/packages/release/bioc/html/topGO. html) [12]. Fiji image analysis software (https://imagej.net/software/fiji/ downloads) [13]. Co-Cluster Analysis Fiji Plugin v3.0 (https://github.com/ dcolam/Cluster-Analysis-Plugin) [14].

314

Darren Kelly and Gerhard Schratt

2.3

Online Resources

Ensembl genome browser Release 109 (https://www.ensembl. org/index.html) [15]. circBase (http://www.circbase.org/) [16]. circAtlas V3.0 (https://ngdc.cncb.ac.cn/circatlas/) [17]. EnrichMir version 0.99.28 (https://ethz-ins.org/enrichMiR/ ) [9]. ScanMir version 1.5.2 (https://ethz-ins.org/scanMiR/) [10]. siRNA design tool, e.g., GeneAssist™ Custom siRNA Builder (Thermo Fisher Scientific; https://www.thermofisher.com/order/ custom-genomic-products/tools/sirna/).

2.4

Reagents

Picrotoxin (Sigma Aldrich P1675). Sprague–Dawley Rat (Charles River Laboratories, Sulzfeld, Germany). 1 μM pore PET membrane filter inserts (Millipore). Cell scraper 1.35 cm (Sarstedt). Neurobasal+ media (Thermo Fisher Scientific).

2.4.1 Cell Culture and Stimulation

2.4.2 RNA Extraction and Preparation

mirVana™ total RNA Isolation Kit. rRNA depletion (Ribo-Zero rRNA removal Kit [Illumina]). NEBNext Ultra™ Directional RNA Library Prep Kit for Illumina (New England Biolabs).

2.4.3

RNAi Approaches

Silencer Select siRNA Custom RNAi Screen (Thermo Fisher Scientific). pSUPER shRNA-expressing plasmid (Oligoengine).

2.4.4

Transfection

Nucleofection of primary neurons: Rat Neuron Nucleofector™ kit (Lonza, Basel, Switzerland). Amaxa™-4D Nucleofector™. Transfection of primary neurons: Lipofectamine 2000 (Invitrogen).

2.4.5 Immunocytochemistry

3

mPSD95 Primary Antibody 1:200 (BioLegend). rSynapsin1 Primary Antibody 1:1000 (AbCam). Anti-mouse Alexa Fluor™ 546 Secondary Antibody 1:1000 (Thermo Fisher Scientific). Anti-rabbit Alexa Fluor™ 647 Secondary Antibody 1:1000 (Thermo Fisher Scientific). Hoechst 33342 (Thermo Fisher Scientific).

Methods 1. Extract primary hippocampal neurons from embryonic p19 rat brain and plate on porous membrane cell culture inserts (aperture diameter 1 μM) for up to 20 days in neurobasal+ media as described in Colameo et al. [7] (see Note 1).

Screening Neuronal circRNAs

3.1 circRNA Candidate Discovery: Compartment and Activity-Dependent Enrichment of circRNAs

315

2. For a homeostatic downscaling paradigm in compartmentalized cultures, cultures are stimulated with Picrotoxin (PTX) at 100 μM (or vehicle 70% EtOH) for 48 h at DIV18–20 (see Note 2). 3. Extract RNA in a compartment specific manner using miRvana total RNA isolation kit™ as per instructions, scraping the top and bottom of the microporous membrane with sterile fine cell scrapers into separate Eppendorf tubes of cell lysis buffer. 4. Measure the RNA concentration of each sample using spectrophotometry (Nanodrop) and deplete rRNA from at least 1 μg of each sample using a Ribo-Zero rRNA removal kit (Illumina). 5. We suggest the NEBNext Ultra™ directional RNA library Prep Kit (Illumina) for library preparation. Sequencing can be performed as 100-bp single-read sequencing on HiSeq2500™ (Illumina) (see Note 3). 6. circRNAs can be reconstructed from the RNA sequencing results through the Deep Computational Circular (DCC) RNA Analytics approach [11] (see Note 4). 7. Determination of significant differences of circular over linear ratios between the compartments and treatments requires appropriate statistical tests such as a beta-binomial test as described by Cheng et al. [18]. 8. This approach results in the identification of thousands of unique circRNAs detected in a model system, with potentially hundreds of circRNAs significantly enriched in a specific compartment with or without activity paradigms.

3.2 Screening Criteria for circRNA Screen Candidates

1. To filter the thousands of circRNA candidates for a technically manageable functional screen, we suggest the following screening criteria in order of preference. Depending upon the throughput of the selected functional screen and model system, the criteria can be loosened or refined (Fig. 1):

3.2.1 Filtering of Process-Enriched circRNAs

1. Top 5% highest expressed circRNAs in the process compartment (average read counts) (see Note 5). 2. Dendritic enrichment (significantly different (FDR