The Bacteriocins: Current Knowledge and Future Prospects [1 ed.] 9781910190388, 9781910190371

Citation preview

The Bacteriocins Current Knowledge and Future Prospects

Edited by Robert Dorit Sandra M. Roy and Margaret A. Riley

Caister Academic Press

The Bacteriocins

Current Knowledge and Future Prospects

Edited by Robert Dorit,1 Sandra M. Roy2 and Margaret A. Riley3 1Department

of Biological Sciences Smith College Northampton, MA USA 2Department

of Veterinary and Animal Sciences University of Massachusetts Amherst Amherst, MA USA

Caister Academic Press

3Department

of Biological Sciences University of Massachusetts Amherst Amherst, MA USA

Copyright © 2016 Caister Academic Press Norfolk, UK www.caister.com British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-1-910190-37-1 (paperback) ISBN: 978-1-910190-38-8 (ebook) Description or mention of instrumentation, software, or other products in this book does not imply endorsement by the author or publisher. The author and publisher do not assume responsibility for the validity of any products or procedures mentioned or described in this book or for the consequences of their use. All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without the prior permission of the publisher. No claim to original U.S. Government works. Cover design adapted from the following images: image of 4LE7 (McCaughey, L.C., Grinter, R., Josts, I., et al. (2014). Lectin-like bacteriocins from Pseudomonas spp. utilise d-rhamnose containing lipopolysaccharide as a cellular receptor. Tran Van Nhieu, G., ed. PLoS Pathog. 10, e1003898); image of 4FZL (Grinter, R., Roszak, A.W., Cogdell, R.J., Milner, J.J., and Walker, D. (2012). The crystal structure of the lipid II-degrading bacteriocin syringacin M suggests unexpected evolutionary relationships between colicin M-like bacteriocins. J. Biol. Chem. 287, 38876–38888); image of 2MWR (Acedo, J.Z., van Belkum, M.J., Lohans, C.T., McKay, R.T., Miskolzie, M., and Vederas, J.C. (2015). Solution structure of acidocin B, a circular bacteriocin produced by Lactobacillus acidophilus M46. Appl. Environ. Microbiol. 81, 2910–2918); image of 4AQN (Patzer, S.I., Albrecht, R., Braun, V., and Zeth, K. (2012). Structure and mechanistic studies of pesticin, a bacterial homolog of phage lysozymes. J. Biol. Chem. 287, 23381); and image of 1CII (Wiener, M., Freymann, D., Ghosh, P., and Stroud, R.M. (1997). Crystal structure of colicin Ia. Nature 385, 461–464). Created with Jmol, an open-source Java viewer for chemical structures in 3D: http://www.jmol.org/ Ebooks Ebooks supplied to individuals are single-user only and must not be reproduced, copied, stored in a retrieval system, or distributed by any means, electronic, mechanical, photocopying, email, internet or otherwise. Ebooks supplied to academic libraries, corporations, government organizations, public libraries, and school libraries are subject to the terms and conditions specified by the supplier.

Contents Contributorsv Prefaceix Forewordxi 1

The Natural History of Bacteriocins

2

Microcins and Other Bacteriocins: Bridging the Gaps Between Killing Strategies, Ecology and Applications

David M. Gordon

Sylvie Rebuffat

3

1

11

Nuclease Colicins: Mode of Action, Immunity and Mechanism of Import into Escherichia coli35 Justyna A. Wojdyla, Grigorios Papadakos and Colin Kleanthous

4

Capturing the Power of van der Waals Zone in the Creation of a Novel Family of Bacteriocin-based Antibiotics65 Xiao-Qing Qiu and Margaret A. Riley

5

The Use of Pyocins in Treating Pseudomonas aeruginosa Infections81 Suphan Bakkal

6

Streptococcal Bacteriocin-producing Strains as Oral Probiotic Agents

103

Treating Bovine Mastitis with Nisin: A Model for the Use of Protein Antimicrobials in Veterinary Medicine

127

John D.F. Hale, Philip A. Wescombe, John R. Tagg and Nicholas C.K. Heng

7

Sandra M. Roy, Margaret A. Riley and Joseph H. Crabb

iv  | Contents

8

The Phenotypic and Genotypic Landscape of Colicin Resistance141 Adrienne Kicza, Christine Pureka, Diana Proctor, Margaret A. Riley and Robert Dorit

Index155

Contributors Suphan Bakkal Sabanci University Foundations Development Directorate Istanbul Turkey [email protected] Joseph H. Crabb ImmuCell Corporation Portland, ME USA [email protected] Robert Dorit Department of Biological Sciences Smith College Northampton, MA USA [email protected] David M. Gordon Research School of Biology ANU College of Medicine, Biology & Environment The Australian National University Acton, ACT Australia [email protected] John D.F. Hale BLIS Technologies Ltd., Centre for Innovation Dunedin New Zealand [email protected]

Nicholas C.K. Heng Sir John Walsh Research Institute, Faculty of Dentistry University of Otago Dunedin New Zealand [email protected] Adrienne Kicza Department of Biological Sciences Smith College Northampton, MA; and Albany Medical College Albany, NY USA [email protected] Colin Kleanthous Department of Biochemistry University of Oxford Oxford UK [email protected] Grigorios Papadakos Department of Biochemistry University of Oxford Oxford UK [email protected]

vi  | Contributors

Diana Proctor Department of Biological Sciences Smith College Northampton, MA; and Department of Biology Stanford University Stanford, CA USA [email protected] Christine Pureka Department of Biological Sciences Smith College Northampton, MA USA [email protected] Xiao-Qing Qiu Laboratory of Biomembrane and Membrane Protein West China Hospital Wuhou District, Chengdu Sichuan China [email protected] Sylvie Rebuffat Muséum National d’Histoire Naturelle Centre National de la Recherche Scientifique Sorbonne Universités Communication Molecules and Adaptation of Microorganisms group (MCAM UMR 7245 CNRS-MNHN) Paris France [email protected]

Margaret A. Riley Department of Biological Sciences University of Massachusetts Amherst Amherst, MA USA [email protected] Sandra M. Roy Department of Veterinary and Animal Sciences University of Massachusetts Amherst Amherst, MA USA [email protected] John R. Tagg BLIS Technologies Ltd., Centre for Innovation Dunedin New Zealand [email protected] Philip A. Wescombe BLIS Technologies Ltd., Centre for Innovation Dunedin New Zealand [email protected] Justyna A. Wojdyla Department of Biochemistry University of Oxford Oxford UK [email protected]

Current Books of Interest Chloroplasts: Current Research and Future Trends2016 Microbial Biodegradation: From Omics to Function and Application2016 Influenza: Current Research2016 MALDI-TOF Mass Spectrometry in Microbiology2016 Aspergillus and Penicillium in the Post-genomic Era2016 Omics in Plant Disease Resistance2016 Acidophiles: Life in Extremely Acidic Environments2016 Climate Change and Microbial Ecology: Current Research and Future Trends2016 Biofilms in Bioremediation: Current Research and Emerging Technologies2016 Microalgae: Current Research and Applications2016 Gas Plasma Sterilization in Microbiology: Theory, Applications, Pitfalls and New Perspectives2016 Virus Evolution: Current Research and Future Directions2016 Arboviruses: Molecular Biology, Evolution and Control2016 Shigella: Molecular and Cellular Biology2016 Aquatic Biofilms: Ecology, Water Quality and Wastewater Treatment2016 Alphaviruses: Current Biology2016 Thermophilic Microorganisms2015 Flow Cytometry in Microbiology: Technology and Applications2015 Probiotics and Prebiotics: Current Research and Future Trends2015 Epigenetics: Current Research and Emerging Trends2015 Corynebacterium glutamicum: From Systems Biology to Biotechnological Applications2015 Advanced Vaccine Research Methods for the Decade of Vaccines2015 Antifungals: From Genomics to Resistance and the Development of Novel Agents2015 Bacteria-Plant Interactions: Advanced Research and Future Trends2015 Aeromonas2015 Antibiotics: Current Innovations and Future Trends2015 Leishmania: Current Biology and Control2015 Acanthamoeba: Biology and Pathogenesis (2nd edition)2015 Microarrays: Current Technology, Innovations and Applications2014 Full details at www.caister.com

Preface

From the fine-scale biophysics of protein folding to the mathematical modelling of antagonism and altruism in microbial communities, the bacteriocins have proven an unexpectedly rich entryway into multiple subdisciplines within the life sciences. This structurally, functionally, and phylogenetically diverse class of proteins, first described nearly a century ago, has steadily attracted the attention of talented investigators around the world. As this volume makes clear, these proteins continue to generate unexpected insights and rich hypotheses at multiple scales of biological organization. At the molecular level, every aspect of bacteriocin function – from their synthesis, to their translocation across the membrane of target bacteria, to their rapid killing activity – justly elicits our amazement. While we have learned a great deal about structure–function relationships from the bacteriocins, their stability, folding kinetics and possible interactions with other proteins are not yet fully elucidated. At the cellular level, the regulation of bacteriocin synthesis and activity has attracted increasing attention. We have known for some time that bacteriocin production is often a component of the stress response in a number of bacterial species, but we are only now beginning to fully understand the regulatory networks in which bacteriocins are embedded. Conversely, while the bactericidal effects of these proteins are well known, the potential functional consequences of sublethal bacteriocin concentrations remain unexplored. Finally, these proteins raise rich ecological and evolutionary riddles: What role do bacteriocins play in the emerging portrait of the microbiome? Why do virtually all Bacteria and Archaea examined to date produce some kind of bacteriocin-like protein? Why do some, but not all, bacteriocins exhibit a narrow killing spectrum? These and other questions are tantalizing in their own right, but understanding the biology of bacteriocins has taken on a new urgency. Although their potential as therapeutic molecules was foreshadowed decades ago, it is only recently that they have emerged as serious lead compounds for drug design. The renaissance of interest in the applied biology of bacteriocins has been driven, at least in part, by the growing realization that we are reaching the end of the conventional antibiotic era. As the clinical and economic costs of antibiotic resistance continue to mount, as the pipeline for new antibiotic classes slows to a trickle, and as the importance of healthy microbiome snaps into focus, interest in the therapeutic potential of the bacteriocins has grown. As this volume makes clear, evolution has been confronting and solving both the challenges of evolving novel antibiotics and of outpacing emerging resistance. As our species, Homo sapiens, confronts those very same challenges, it behoves us to consider that potential answers may be hidden in plain sight – if only we are astute enough to see them. This volume, we hope, represents a step in that direction, and

x  | Preface

honours a colleague whose career has been dedicated to understanding bacteriocins in all their complex and messy beauty. We dedicate this book to our colleague Richard James and to the generation of researchers and students that he has inspired. Robert Dorit, Margaret Riley and Sandra Roy

Foreword

It was a very pleasant surprise when Peg Riley and Rob Dorit informed me that this Caister Academic Press volume on bacteriocins was to be dedicated in my honour. This accolade follows a Biochemical Society Focussed Meeting1 in Nottingham in 2012 organized by Colin Kleanthous, Chris Penfold and Dan Walker in honour of my retirement. I am very grateful for the recognition of the work of my research group with bacteriocins conferred by my peers by both events. In this preface I will provide some background on my 34 years of working with bacteriocins and my perspective on the bacteriocin research field. When I was a final-year undergraduate student in London in 1970 I gave a short talk on antibiotics to my fellow students at the graduation dinner of my BSc (Hons) Applied Biology course in which I had specialized in microbiology. Antibiotics was a subject that had fascinated me during my undergraduate degree; little did I know that I would spend much of the rest of my academic career working in this field. My PhD was carried out at the Middlesex Hospital Medical School in London in a small lab working on the radiation biology of E. coli bacteria, led by Dr Neil Gillies. This work introduced me to mutations such as lon which make E. coli cells exquisitely sensitive to low doses of radiation by inhibiting cell division and inducing the formation of long filamentous cells. In 1973, my interest in bacterial cell division then took me to work as a Postdoctoral Fellow with Professor Arthur Pardee at Princeton University, New Jersey, USA, for 2 years. One of the major lines of research in the lab was the use of the new beta-lactam antibiotic FL1060 (mecillinam) which caused the rounding up of cells. Another English postdoc, Brian Spratt, and I speculated that mecillinam may inhibit murein synthesis required for cell elongation as a rod, in contrast to other beta-lactams such as cephalexin which caused filament formation by inhibiting murein synthesis required for the formation of cross walls during septation. Brian and I separately published a number of papers which explained the mechanism of inhibition of cell division by mecillinam. On my return to England I was offered a Lectureship in the School of Biological Sciences at the University of East Anglia (UEA) teaching mostly medical microbiology. I was happy to offer a postdoctoral position to work in my laboratory to Dr Pearl Cooper, a microbial geneticist, and this was serendipitously to lead to my interest in colicins, bacteriocins produced by E. coli. Pearl, who had taught microbiology at UEA previously, was helping to run a practical class which she had designed in which students isolated bacteria from the intestines 1 How bugs kill bugs: progress and challenges in bacteriocin research. Biochem. Soc. Trans. 40, 1433–1565 (2012).

xii  | Foreword

of chickens and then characterized them in terms of aerobic/anaerobic growth, antibiotic sensitivity and bacteriocin production. Bacteriocin-producing strains were further characterized by the students in an attempt to identify the group of bacteriocin being produced. One pair of students identified a J strain which exhibited an unusual pattern of production of a colicin. In subsequent work back in the lab Pearl determined that J contained seven different plasmids and produced both an E colicin and colicin M. Using transformation and conjugation methods, Pearl demonstrated that two of the small plasmids encoded different E colicins, and a large plasmid encoded colicin M production and colicin B immunity. This single strain turned out to be a treasure trove of science which kept researchers in my laboratory busy for the next 30 years. I always used the story of this strain as an example to academics of the potential value to them of teaching students. E. coli transformed with one of the two E colicin encoding plasmids was then tested for immunity to the seven known E colicin-encoding strains (ColE1 to ColE7). The basis of immunity testing is that each of the seven strains are immune to killing by strains producing the same E colicin but are sensitive to the other six E colicins. This immunity testing confirmed that J encoded two new E colicins (ColE8 and ColE9). The immunity testing also highlighted an unusual non-reciprocal immunity in which a ColE3-producing strain was immune to both ColE3 and ColE8, whilst a ColE9-producing strain was immune to both ColE9 and ColE5. I speculated that the observed non-reciprocal immunity could be due to a high homology between ColE3 and ColE8 (or between ColE5 and ColE9), such that the ColE3 immunity gene protected against both ColE3 and ColE8. An alternative explanation was that the ColE3 plasmid encoded two different immunity genes, one of which gave protection against ColE3 and the other against ColE8. In order to answer this question we had to learn the new techniques of gene cloning and transposon mutagenesis in the early 1980s so that we could clone and analyse fragments of E colicin-encoding plasmids. Kin Chak in my laboratory showed unequivocally that the ColE3-CA38 plasmid carried a separate, tandem immunity gene that confers immunity to ColE8. A follow-on paper by Kin Chak in 1986 confirmed that the ColE9 plasmid isolated from J carried two tandem immunity genes, one providing protection to ColE9 and the other to ColE5. Subsequent DNA sequence analysis suggested a close evolutionary relationship between the ColE5 and ColE9 plasmids in which a segment of DNA which encodes the last 73 amino acids of the colicin E9 structural gene, the E9 immunity gene and the E9 lysis gene has been transferred into the colicin E5 structural gene at a site 48 amino acids from the C-terminus. Pearl Cooper and I then turned our attention to the bacteriocins produced by Klebsiella pneumoniae strains of the Edmondson and Cooke klebicin typing panel. Many younger researchers may not know that in the 1970s bacteriocin typing was a widely used technique for distinguishing strains of important pathogens. As a result several bacteriocin typing panels of strains were available. We demonstrated that the pP5a plasmid that encodes klebicin A1 and immunity to klebicin A1, also encodes immunity to colicin E6, with both immunities being encoded by the same gene. In contrast, the pP5b plasmid that encodes klebicin A2 and immunity to klebicin A2 has a tandem immunity gene to colicin E3, and the pP3 plasmid that encodes klebicin A3 and immunity to klebicin A3 also has a tandem immunity gene to colicin E3. K. pneumoniae strains are not killed by E colicins, as they lack the BtuB outer-membrane receptor, and strains are not killed by klebicins as they lack the aerobactin receptor. It is therefore difficult to envisage how either a single or tandem

Foreword |  xiii

immunity genes encoding immunity to both a klebicin and an E colicin could evolve in the absence of selection pressure. Having had the good fortune of isolating two new members of the E colicin family (ColE8 and ColE9), we now had four homologous DNase type E colicins (ColE2, ColE7, ColE8 and ColE9) along with their immunity genes (Im2, Im7, Im8 and Im9). I realized that this would be an excellent model system for a research grant call to study the emerging field of the molecular recognition of protein–protein interactions. Understanding the basis of molecular recognition between an E colicin and its immunity protein required knowledge of (1) the amino-acids which act as specificity determinants both in the DNase domain of the colicin and in the Im protein; (2) the kinetics of binding between each DNase domain and the four different Im proteins; and (3) the structures of the four DNase/Im protein complexes. I setup a long term collaboration at UEA with two talented colleagues, Geoff Moore and Colin Kleanthous, which was able to contribute important data in all three areas, and has also provided significant insights into the basis of molecular recognition of proteins in other biological systems. Studying colicin biology is like a gift that never stops giving as it keeps throwing up other important properties that are of wider relevance in biology. Examples include (1) Im9 is one of the fastest folding proteins in nature; (2) the complex of ColE9 DNase with Im9 is one of the highest affinity interactions observed in nature; (3) Im9 does not bind at the active site of the DNase; (4) the DNase domain of E colicins form channels in lipid bilayers; (5) the similarity of the HNH domain of the DNase to Caspase-activated DNase (CAD) in eukaryotes; and (6) the mechanism of translocation of E colicins. In order to kill, the folded, functional DNase domain of E colicins has to be delivered from the outside to the inside of an cell, a process that involves interactions with a surface receptor and Tol proteins. This translocation process is a unique event in prokaryotic biology. An understanding of this process requires knowledge of (l) the receptor binding process; (2) how colicin domains cross membranes; (3) the specificity determinants of colicins and Tol proteins; (4) the kinetics of binding of Tol binding domains and Tol proteins; and (5) the structures of colicin/Tol protein complexes. Though Colin Kleanthous moved to York (and now Oxford), I moved to Nottingham, and Geoff Moore remained at UEA, we managed to maintain a very productive collaboration, also involving Chris Penfold and Mireille Vankemmelbeke in Nottingham, that has provided valuable information on the translocation process. One other aspect of our research with colicins to mention is that the study of any biological system requires the use of novel tools. I was very keen on developing novel tools or adapting existing tools to facilitate the study of bacteriocins in the lab. Examples of this include: promoter mapping vectors and transposon mutagenesis (with Kin Chak); chemical mutagenesis, PCR mutagenesis and alanine scanning mutagenesis (with Carole Garinot-Schneider); chimeric proteins with GFP and introducing disulphide locks into colicin domains to study their flexibility (with Chris Penfold); SOS promoter–lux fusions to monitor DNA damage (with Mireille Vankemmelbeke); Surface Plasmon resonance to study the affinity of protein–protein interactions (with Sarah Hands); introducing enterokinase cleavage sites to study surface accessibility of colicin molecules (with Ying Zhang); and Alexa Fluor labelled Im proteins to study their release from colicin-Im protein complexes (with Mireille Vankemmelbeke).

xiv  | Foreword

It was 1991 that Claude Lazdunski, Franc Pattus and I organized a meeting on the Île de Bendor, an island close to Marseille in France, to bring scientists from around the globe to discuss bacteriocins. Many of the participants at this meeting, including Volkmar Braun, Bill Cramer, Colin Hill, Karen Jakes, Roland Lloubes and Bauke Oudega, attended a follow up bacteriocin meeting that I organized in Nottingham 7 years later, and my retirement meeting in 2012. I would like to thank all members of the bacteriocin community for making possible the exciting progress that has happened in the last 24 years in so many aspects of the biology of these fascinating protein antibiotics. I have been privileged to make so many good friends who share my love for bacteriocins. During my academic career at both UEA and Nottingham I have taught medical microbiology to undergraduates as well as masters and medical students. In all cases I provided lecture courses on both the mechanism of action of antibiotics and the variety of mechanisms by which bacteria acquire resistance to antibiotics. With each year that passed I was spending more time talking about the latter topic, concentrating a lot on what could be done to counter the threat to healthcare systems of antibiotic resistance. In January 2007, I founded the Centre for Healthcare-associated Infections (CHAI) which involved researchers from seven Schools of the University of Nottingham together with colleagues in Nottingham University Hospital. The aim of CHAI was to carry out research on the treatment and prevention of healthcare-associated infections, many of which were due to antibiotic-resistant, bacterial pathogens. On the day of the CHAI launch Symposium I was accused of being a ‘sensationalist and scaremonger’ about the problems of antibiotic resistance by the Chief Nursing Officer at the Department of Health, UK. As I expected, antibiotic resistance was a very important topic and I was asked to give over 300 media interviews over the next three years, and was awarded the Communications Award of the Society for Applied Microbiology in 2008 for helping to inform the public of the scale of the problem. When the Chief Medical Officer (England) stated in 2013 that ‘antimicrobial resistance poses a catastrophic threat that is as serious as global warming’ I was satisfied that at last the scale of the threat had been recognized. With the increasing awareness of the paucity of new antibiotics combined with increasing incidence of antibiotic resistance to current antibiotics, it would be hoped that the large number of diverse bacteriocins that have been characterized could help provide new options for treating serious bacterial infections. Nisin, a lantibiotic produced by lactic acid bacteria, is widely used as a preservative in the food industry in processed cheese, dairy and canned products. Research in Colin Hill’s laboratory has shown exciting prospects for the bioengineering of nisin to enhance its antibacterial activity, especially against Gram-negative pathogens. The potential immunological problems of using protein antibiotics like bacteriocins to treat systemic infections is seen as a significant impediment to their development as new antibiotics, but there are some examples of the treatment of topical infections such as UTIs, which have been shown by Peg Riley to have some promise. It would be so exciting to those who have worked with bacteriocins over many years if they became useful treatment options for a wide range of antibiotic-resistant infections. Richard James

The Natural History of Bacteriocins David M. Gordon

1

Abstract A variety of empirical and theoretical studies have demonstrated the significant role that the production of allelopathic compounds known as bacteriocins play in mediating inter- and intra-specific interactions among bacteria, and hence in shaping bacterial community diversity. There is also increasing recognition that bacteriocins may provide viable alternatives to traditional antibiotics and are likely to be a key characteristic of probiotic strains used to prevent or limit the establishment of diarrhoeal pathogens. The goal of this chapter is to highlight some potentially important factors, genetic and environmental, that influence the likelihood that bacteriocin production will actually confer a fitness advantage on the producing strains and that also influence the type of bacteriocin being produced. The applied use of bacteriocins requires understanding why the production of multiple bacteriocins by a single strain is such a common phenomenon in species like Escherichia coli. Finally, much of our understanding of the ecological role of bacteriocins comes from studies of the colicins released as a consequence of cell lysis, but all microcins and many colicins are secreted from the cell, and some experimental evidence is presented that would suggest we have a very incomplete understanding of the dynamics of the secreted bacteriocins. Introduction The production of compounds, by bacteria, that are toxic to other bacteria is a very common phenomenon. Further, there is an extensive body of in vitro and in vivo evidence, as well as mathematical theory, that demonstrates that these compounds have an important role in mediating among-strain interactions in bacteria. The strongest evidence for these assertions comes from the protein antibiotics known as bacteriocins, in particular the bacteriocins produced by Escherichia coli and related bacteria: the colicins and the microcins. Given the rise in antibiotic resistance there is a great deal of interest in exploiting bacteriocins as therapeutic agents and as food preservatives (Gillor et al., 2004). The use of naturally occurring or synthetic bacteriocins requires a good understanding of the bacteriocin’s mode of action, mechanisms of resistance, pharmokinetics and eukaryotic cell toxicity. However, bacteriocins are also of interest in the probiotics industry, as many researchers believe bacteriocin production is a critical characteristic of a successful probiotic bacterium. If this is true, then it is essential there be a thorough understanding of all aspects of the ecology of bacteriocin production and the role of bacteriocin production in mediating among-strain interactions. The primary goal of this chapter will be to highlight aspects of the ecology

2  | Gordon

and evolution of bacteriocin production where this knowledge is lacking. Consequently this chapter will explore several aspects of bacteriocin production where enhancing our understanding of bacteriocin ecology is likely to be important if we are to exploit bacteriocinogeny in order to create or select probiotic strains. Unfortunately, but inevitably, much of our knowledge concerning bacteriocins and their role in bacterial ecology comes from studies with E. coli and there is a real need to expand such studies to other species. The frequency of bacteriocin production In E. coli the frequency of bacteriocin production differs depending on the source of the isolates, and may vary from 10% to over 70% (Riley and Gordon, 1996; Gordon et al., 1998; Šmajs et al., 2010). Theoretical and empirical studies have shown that environmental conditions will, in part determine the frequency of bacteriocin production. Mathematical models developed by Frank (1994) predict that colicin-producing cells will have an advantage in benign habitats, whereas harsher habitats will favour non-producers. Empirical evidence supports this prediction (Gordon et al., 2007). E. coli population densities and growth rates are significantly lower in aquatic environments than in the lower intestines of mammals and, in Australia, the frequency of colicin-producing E. coli recovered from water samples was 9%, significantly lower than the 30% production observed for faecal isolates from Australia. Of much relevance to the field of probiotics are the results of recent study using mathematical models to examine the dynamics of bacteriocinogeny in the vertebrate gut (Barnes et al., 2007). The models predict that the advantage accruing to bacteriocin production will decline as the transit time of material through the gut decreases. The empirical evidence accords with these predictions; the frequency of colicinogenic strains is lower in isolates from carnivorous mammals (short gut transit times) than in herbivorous or omnivorous mammals (longer gut transit times). Additional evidence for the importance of gut transit comes from data collected on the frequency of bacteriocin production in E. coli isolated from the intestinal tract of pigs (Abraham et al., 2012). This study reported that the frequency of strains isolated from the duodenum that encoded one or more colicin genes was 13%, significantly less than the 51% of the isolates encoding a colicin gene that were recovered from the lower gastrointestinal tract. In Hafnia paralvei the frequency of bacteriocin production varies depending on if the isolates were recovered from fish, where 4% of the isolates were bacteriocin producers, as compared to mammals or reptiles where, 29% and 64%, respectively, of the isolates were found to be bacteriocin producers (Gordon et al., 2007). H. paralvei provides another example of the potential importance of host effects on the nature of bacteriocin production. The species produces at least two bacteriocins, alvecin A and B (Wertz and Riley, 2003). These bacteriocins can be detected in isolates from mammals but not in isolates from reptiles. The bacteriocinogenic H. paralvei isolates from reptiles produce several uncharacterized bacteriocins, which appear to be absent from H. paralvei isolates from mammals. Genotype effects Strains of E. coli belong to four main genetic groups, or phylo-groups, designated A, B1, B2 and D (Gordon et al., 2008). Strains of the phylo-groups differ in their phenotypic and genotypic characteristics, their ecological niche and in their propensity to cause disease

The Natural History of Bacteriocins |  3

(Alm et al., 2011). Although all colicins in E. coli are plasmid encoded, as is microcin V, the distribution of these bacteriocins among strains is not independent of the genetic background of the host. For example, in isolates from native Australian mammals, colicin Ia is detected significantly more frequently among phylo-group B2 strains than in the strains of the other phylo-groups ( Jeziorowski and Gordon, 2007) Microcins H47 and M are also more common among phylo-group B2 isolates from humans (Šmajs et al., 2010). More specific associations have also been observed. The presence of the genes for colicin E1 is more likely in strains that encoded the K1 capsule type (Gordon et al., 2007) than those that do not have this capsule type. Among phylo-group B2 isolates from humans, microcin V was found to be absent from strains encoding focG, hylD, iha or she ( Jeziorowski and Gordon, 2007). Overall, two-thirds of phylo-group B2 strains encode at least one of these four traits, yet microcin V is absent from all of these strains ( Jeziorowski and Gordon, 2007). The reasons for these non-random associations between bacteriocins and chromosomal genotype are unknown. Bacteriocin release: lysis versus excretion Bacteriocins make their way from within the cell to the external environment in two ways: they are released as a consequence of cell lysis or they are actively secreted from the cell. Much of the mathematical theory that has been developed concerning the dynamics of bacteriocin production – as well as most of the empirical evidence – has concerned those bacteriocins released as a consequence of cell lysis. The costs associated with bacteriocin release through cell lysis are obvious: soon after the environmental conditions inducing bacteriocin production occur, a certain fraction of the producing population will lyse, and this inevitably results in a reduced rate of population growth. The fraction of the population lysing depends on the exact nature of the conditions leading to induction, and varies with the bacteriocin being produced and the genotype of the producing strain (Gordon and Riley, 1999; Gordon et al., 1998). Although the E group colicins and colicin K are some of the colicins released as a consequence of cell lysis, colicins M, B, Ia and Ib, for example, are actively secreted from the cell and death is not an inevitable consequence of colicin production. All of the microcins are also secreted from the producing cell. We have little understanding of the costs associated with the production of bacteriocins that are not released via cell lysis. It has been argued that these costs may be substantially less than those associated with release via cell lysis (Dykes and Hastings, 1997). However, there are few good data to support these assertions. If the assumption of single hit kinetics is correct for all bacteriocins, then bacteriocin-killing dynamics does not depend on the manner in which the bacteriocin moves from the cell to the external environment. The number of sensitive cells killed will depend on their density, the density of bacteriocin molecules, and the rate at which bacteriocin molecules and sensitive cells contact one another. All else being equal, the frequency at which a bacteriocin-producing strain can invade a sensitive cell population decreases as the number of bacteriocin molecules released per producing cell increases. It has been estimated that an induced cell harbouring colicin E1 will release about 100,000 colicin molecules and the net production of other colicin types may be many times greater (Gordon and Riley, 1999b). If levels of bacteriocin production are similar for those bacteriocins secreted from the cell then production may impose a significant metabolic cost to the producing cells and substantially reduce the rate of population growth.

4  | Gordon

There is some indirect evidence to suggest that the costs associated with the production of secreted bacteriocins may be substantial. As previously discussed, mathematical models examining the dynamics of bacteriocinogeny in the vertebrate gut predict that the advantage accruing to bacteriocin production declines as the transit time of material through the gut decreases (Barnes et al., 2007b). The empirical evidence accords with these predictions; the frequency of colicinogenic strains is lower in isolates from carnivorous mammals (short gut transit times) than herbivorous or omnivorous mammals (longer gut transit times). This outcome holds true even when the colicins released via cell lysis are excluded from the analysis (Gordon et al., 2007). This in turn suggests that the costs associated with the production of secreted bacteriocins may be comparable to the cost resulting from colicin release through lysis. Experiments with the aim of determining what fraction of cells produce and secrete bacteriocins following induction and what impact bacteriocin production and release have on the population growth rate are required. Later in this chapter some preliminary data concerning the invasion dynamics of bacteriocins secreted from the cells will be presented. The results of these experiments further emphasize the need for more experimental work with this group of bacteriocins. Multiple bacteriocin production The available data would suggest that, in E. coli, about 50% of bacteriocin producers produce more than one bacteriocin, with the production of three or more different bacteriocins being quite common (Gordon and O’Brien, 2006; Gordon et al., 2007; Šmajs et al., 2010). The bacteriocins co-occurring in a strain are not simply random combinations of those bacteriocins present in the population. Colicins B and M, Ia and microcin V, Ia and E1, in particular co-occur significantly more often than would be expected by chance (Gordon and O’Brien, 2006; Šmajs et al., 2010). Whilst microcins H47 and M commonly co-occur in a strain (Braun et al., 2002), microcins V and H47/M are seldom observed in the same strain (Gordon et al., 2007). In E. coli, all of the bacteriocins that are released via cell lysis appear to be encoded on small non-conjugative plasmids (e.g. colicin K), while those secreted from the cell are encoded on large conjugative plasmids (e.g. colicin Ia, microcin V) or on the chromosome (e.g. microcin H47). The co-occurrence in a cell of two large plasmids each encoding a different colicin is seldom observed. When two bacteriocins normally encoded on large plasmids are observed in the same cell both bacteriocins are on the same plasmid, indicating that the colicin genes have moved from one plasmid backbone onto another (Christenson and Gordon, 2009; Jeziorowski and Gordon, 2007). The co-association of colicins Ia and E1 in cell is often observed; however in this case, the evidence suggests that the two colicin operons are being maintained on separate plasmids (Šmajs et al., 2010). Few strains appear to encode for two colicins which are also associated with lysis genes. Those that do – for example, colicins E2 and E7 – may represent chimeras. At least two examples of E2 and E7 chimeras have been reported which resulted from recombination of portions of the E2 and E7 colicin operons in a single plasmid and, in both cases, there is a single copy of the lysis gene (Nandiwada et al., 2004; Tan and Riley, 1997). It may be that the co-occurrence of two colicins released via cell lysis imposes too high a cost to the cell due to the expression of two lysis genes.

The Natural History of Bacteriocins |  5

We have little understanding of the reasons why multiple bacteriocin production is so common. There are several explanations based on theoretical expectations, but none of these have been tested empirically. To understand one of the potential benefits arising from the production of multiple bacteriocin types, consider a community initially consisting of a sensitive cell population and two populations of producing cells, each encoding a single bacteriocin type. If one of the producing cells acquires, through recombination, the genes for the other bacteriocin type, then this multiple producer can kill sensitive cells and those cells encoding only a single bacteriocin type. Another likely explanation for the high frequency of multiple bacteriocin production concerns the evolution of resistance to bacteriocins. Resistance to bacteriocins evolves quite readily and most E. coli cells are resistant to most co-occurring colicins (Feldgarden and Riley, 1999; Gordon et al., 1998). Many of the bacteriocins that are significantly coassociated in a cell, for example colicins Ia and E1 (Gordon et al., 2007; Šmajs et al., 2010), exploit different receptors on sensitive cells: Cir and BtuB, respectively. A mutation in one receptor is far more likely than the simultaneous occurrence of mutations in two different receptors. Thus, harbouring multiple bacteriocins exploiting different surface receptors may slow the evolution of resistance in populations where the dominant bacteriocinogenic strain produces multiple bacteriocins, compared with populations where the dominant bacteriocinogenic strain produces a single bacteriocin. Microcins H47 and M are thought to exploit the same receptors (Cir, Fiu, IroN and FepA), whilst colicin Ia and microcin V both exploit the Cir receptor (Braun et al., 2002). The evolution of resistance may be slower or more costly, even for those co-associations where both bacteriocins exploit the same receptors, if the bacteriocins bind to different regions of the receptor (Smarda and Macholan, 2000). Evolving simultaneous resistance to bacteriocins that target different parts of the same receptor may greatly increase the likelihood that receptor function is completely lost. Multiple bacteriocin production may also confer a selective advantage in different environments (Gordon, 2009). Microcins are induced under iron limited conditions and target receptors involved in iron uptake. Some colicins, such as colicins Ia, Ib and B, target iron receptors and may be induced under iron limited conditions. Other bacteriocins, such as colicins E1, E2 or K are induced under non-specific nutrient limitation conditions and target receptors involved in the uptake of other essential nutrients such as vitamins or nucleosides. A recent study of bacteriocin production in E. coli strains isolates from different regions of the porcine intestinal tract provides some support for this conjecture (Abraham et al., 2012). Isolates from the ileum were more likely to produce the colicins Ia/Ib or colicins B and M while faecal isolates were more likely to produce the colicins E2 or E7. Is iron a limited resource in the ileum, while more generalized nutrient limitation is occurring in the colon? It may also be that much of the observed multiple bacteriocin production in E. coli is a co-incidental by-product of the evolution of virulence. The co-association of colicin Ia with microcin V and the co-occurrence of colicins B and M with microcin V in E. coli strains are frequently observed (Christenson and Gordon, 2009; Jeziorowski and Gordon, 2007). These co-associations are thought to have arisen by the movement of microcin V genes together with a number of known or suspected virulence factors such as, salmochelin (iroN), increased serum survival (iss), haemolysin production (hylF), and aerobactin (iutA). In the case of the co-association of microcin V with colicins B and M, the co-association has, in some cases, resulted in the loss of colicin B activity and immunity (Christenson and

6  | Gordon

Gordon, 2009). This suggests that the evolution of multiple bacteriocin production was not the primary advantage resulting from the co-association of microcin V and colicins M and B. Invasion dynamics of secreted bacteriocins There have been few studies investigating the invasion dynamics of bacterial strains encoding bacteriocins that are secreted from the cell or of the invasion dynamics of multiple bacteriocin producers. A series of preliminary experiments were undertaken using two wildtype E. coli strains: one producing colicin Ia and the other producing colicins Ia and M, as well as microcin V. Both strains were also naturally resistant to one or more antibiotics, and antibiotic selection could be used to distinguish the two strains. Serial transfer-type invasion experiments where the Ia producer or the IaVM producer was placed in competition with a sensitive K12 strains in Luria broth resulted in the expected outcomes (Gordon and Riley, 1999). The bacteriocin-producing strain displaced the sensitive strain and the time required for the sensitive strain to be driven to extinction decreased as the initial frequency of the producing strain increased. A second series of serial transfer invasion experiments in Luria broth were undertaken. In one experiment the initial frequency of the Ia producer was varied and the IaVM producer was the dominant strain. The complementary experiment was also undertaken: the initial frequency of the IaVM producer was varied and the Ia producer was the dominant strain. The outcomes of these experiments were unexpected (Fig. 1.1). The first unexpected outcome was that the strain producing only colicin Ia was able to increase in frequency in the presence of a numerically dominant strain that produced both colicin Ia and M as well as microcin V, although the probability of invasion did depended on the initial frequency of the colicin Ia producer. The observation that the frequency of the colicin Ia producer during the first transfers either did not change or declined slightly was also unexpected. When the colicin Ia-producing strain was able to increase its frequency, it did not drive the multiple bacteriocin producers to extinction, rather, and unexpectedly, both strains reached a steady state where they were at roughly equal frequencies. The strain producing multiple bacteriocins (IaVM) was always able to invade in the experiment where its initial frequency was varied and the colicin Ia producer was the dominant strain. However, the strain producing multiple bacteriocins did not drive the single producer to extinction; rather the two strains maintained roughly equal densities for several transfers. The experiments depicted in Fig. 1.1 are far from perfect. Although the strain encoding multiple bacteriocins did produce colicin Ia there was no direct evidence that it also produced colicin M or microcin V. Neither the plasmids nor the chromosomes of these two strains could be considered ‘isogenic’. However, the growth rates of the two strains were similar, and it seems unlikely that growth rate differences could explain the observed outcomes. Finally, microcin V is induced under conditions of iron-limitation and Luria broth is not considered to be an iron-limiting medium. Despite the limitations of the experimental design, the results do raise some interesting questions that seem to highlight our lack of knowledge concerning the excreted bacteriocins and the significance of multiple bacteriocin production. When the colicin Ia producer attempted to invade the bacteriocin IaVM producer why did the Ia producer only increase in frequency after the first transfer? Was this delay due to a delay in bacteriocin induction? Both the single and multiple producers were able to invade a population of the other producer, yet

The Natural History of Bacteriocins |  7

Figure 1.1  Outcome of serial transfer invasion experiments in Luria broth of strains producing either colicin Ia alone or colicins Ia and M in addition to microcin V. The graphs on the left depict three replicates of the experiment where the multiple producer (IaVM) is attempting to invade a population of a strain producing a single bacteriocin (Ia), whilst the graphs on the right show three replicates of the experiment where the single producer (Ia) is attempting to invade a population of a strain producing multiple bacteriocins (IaVM).

failed to drive the dominant strain to extinction. Did the invading strains cease bacteriocin production once they achieved a high frequency? Conclusions There is much to be learnt concerning the role of bacteriocins in the ecology of bacteria. The understanding that has been gleaned from mathematical theory, in vitro experiments and

8  | Gordon

retrospective studies does highlight some important issues concerning the potential role and utility of bacteriocin production in probiotic strains. If there are strains of bacteria whose presence in a host will reduce the likelihood of the host contracting a diarrheal disease, then the recent theoretical and empirical evidence leads to some predictions concerning the bacteriocinogenic properties of such probiotic strains. For diarrhoeal pathogens targeting the small intestine, such as the entero-invasive and entero-haemorrhagic E. coli, the probiotic strain should perhaps encode bacteriocins with a lower cost of production and which mediate among-strain competition for iron. In contrast a probiotic strain developed to prevent colon infection by enterotoxigenic or enteropathogenic E. coli may require the production of bacteriocins released through cell lysis and which mediate among strain carbon source competition. References

Abraham, S., Gordon, D.M., Chin, J., Brouwers, H.J.M., Njuguna, P., Groves, M.D., Zhang, R., and Chapman, T.A. (2012). Molecular characterization of commensal escherichia coli adapted to different compartments of the porcine gastrointestinal tract. Appl. Environ. Microbiol. 78, 6799–6803. Alm, E.W., Walk, S.T., and Gordon, D.M. (2011). The niche of Escherichia coli. In Population Genetics of Bacteria: A Tribute to Thomas S. Whittam, Walk, S.T., and Feng, P.C.H., eds. (ASM Press, Washington, DC), pp. 69–90. Barnes, B., Sidhu, H., and Gordon, D.M. (2007). Host gastro-intestinal dynamics and the frequency of colicin production by Escherichia coli. Microbiology 153, 2823–2827. Braun, V., Patzer, S.I., and Hantke, K. (2002). Ton-dependent colicins and microcins: modular design and evolution. Biochimie 84, 365–380. Christenson, J.K., and Gordon, D.M. (2009). Evolution of colicin BM plasmids: the loss of the colicin B activity gene. Microbiology 155, 1645–1655. Dykes, G.A., and Hastings, J.W. (1997). Selection and fitness in bacteriocin-producing bacteria. Proceedings of the Royal Society of London B: Biological Sciences 264, 683–687. Feldgarden, M., and Riley, M.A. (1999). The phenotypic and fitness effects of colicin resistance in Escherichia coli K-12. Evolution 53, 1019–1027. Frank, S. (1994). Spatial polymorphism of bacteriocins and other allelopathic traits. Evol. Ecol. 8, 369–386. Gillor, O., Kirkup, B., and Riley, M. (2004). Colicins and microcins: the next generation antimicrobials. Advances in Applied Microbiology 54, 129–146. Gordon, D.M. (2009). The potential of bacteriocin-producing probiotics and associated caveats. Future Microbiol. 4, 941–943. Gordon, D.M., and Riley, M.A. (1999). A theoretical and empirical investigation of the invasion dynamics of colicinogeny. Microbiology 145, 655–661. Gordon, D.M., and Brien, C.L. (2006). Bacteriocin diversity and the frequency of multiple bacteriocin production in Escherichia coli. Microbiology 152, 3239–3244. Gordon, D., Oliver, E., and Littlefield-Wyer, J. (2007). The diversity of bacteriocins in Gram-negative bacteria. In Bacteriocins: Ecology and Evolution, Riley, M.A., and Chavan, M.A., eds. (Springer, Berlin, Heidelberg, Germany), pp. 5–18. Gordon, D., Riley, M., and Pinou, T. (1998). Temporal changes in the frequency of colicinogeny in Escherichia coli from house mice. Microbiology 144, 2233–2240. Gordon, D.M., Clermont, O., Tolley, H., and Denamur, E. (2008). Assigning Escherichia coli strains to phylogenetic groups: multi-locus sequence typing versus the PCR triplex method. Environ. Microbiol. 10, 2484–2496. Jeziorowski, A., and Gordon, D. (2007). Evolution of microcin V and colicin Ia plasmids in Escherichia coli. J. Bacteriol. 189, 7045–7052. Nandiwada, L.S., Schamberger, G.P., Schafer, H.W., and Diez-Gonzalez, F. (2004). Characterization of an E2-type colicin and its application to treat alfalfa seeds to reduce Escherichia coli O157:H7. Int. J. Food Microbiol. 93, 267–279. Riley, M.A., and Gordon, D.M. (1996). The ecology and evolution of bacteriocins. J. Ind. Microbiol. 17, 151–158.

The Natural History of Bacteriocins |  9

Šmajs, D., Micenková, L., Šmarda, J., Vrba, M., Ševčíková, A., Vališová, Z., and Woznicová, V. (2010). Bacteriocin synthesis in uropathogenic and commensal Escherichia coli: colicin E1 is a potential virulence factor. BMC Microbiol. 10, 288–288. Smarda, J., and Macholan, L. (2000). Binding domains of colicins E1, E2 and E3 in the receptor protein BtuB of Escherichia coli. Folia Microbiol. (Praha) 45, 379–385. Tan, Y., and Riley, M.A. (1997). Nucleotide polymorphism in colicin E2 gene clusters: evidence for nonneutral evolution. Mol. Biol. Evol. 14, 666–673. Wertz, J.E., and Riley, M.A. (2003). Chimeric nature of two plasmids of Hafnia alvei encoding the bacteriocins alveicins A and B. J. Bacteriol. 186, 1598–1605.

Microcins and Other Bacteriocins: Bridging the Gaps Between Killing Strategies, Ecology and Applications

2

Sylvie Rebuffat

Abstract Bacteriocins are ribosomally synthesized antimicrobial peptides or proteins produced by bacteria, which use these potent weapons to thrive in the microbial wars. To complete this arsenal, bacteriocin-producing strains are endowed with efficient strategies to evade being killed by their own toxins. Most bacteriocins are active in the pico- or nanomolar range and target bacterial species that are phylogenetically close to the producing strain, although some exhibit broader spectra of activity. Bacteriocins have been widely studied in Gram-positive (lantibiotics, pediocin-like bacteriocins) and Gram-negative bacteria (colicins, microcins). However, it is becoming increasingly apparent that bacteriocin production is widespread in nature, including in the archaea, which produce similar defence proteins, the archaeocins. Bacteriocins can differ significantly both in size and in chemical properties, ranging from small peptides (the Gram-positive bacteriocins and microcins are peptides below 10 kDa) to large proteins (colicins are 30–80 kDa proteins). Many are post-translationally modified using dedicated enzymes, leading to highly complex peptide-derived structures. This structural diversity is associated with complex and refined killing strategies, which contribute to the ecological roles of bacteriocins. Here we review these different aspects, bridging the gaps between biosynthesis, killing strategies, ecology and potential future applications. Introduction This chapter discusses the role played by the production of antimicrobial peptides, the bacteriocins, in the competitive strategies of Gram-negative and Gram-positive bacteria. These strategies involve both the biosynthetic pathways that they develop in order to generate and maintain mature active structures, and the uptake/killing strategies they deploy to fight against their competitors. Gram-negative bacteria produce both higher-molecular-mass (30–80 kDa) and lower-molecular-mass ( 90%, with all impurities characterized). ImmuCell achieved this using commercially viable methods, and has received a US Patent for their method (Brigham, 2011). In addition, formulation of the drug substance into the final drug product utilizes proprietary excipients (pH and antioxidants) to stabilize the formulation for a 2-year shelf life.

136  | Roy et al.

Use in food-producing animals A second challenge in drug production for use in food-producing animals involves the milk and meat withhold requirements – the timeframe over which milk and meat from treated animals cannot be sold. The FDA evaluates new drugs using data submitted in the human food safety technical section of a new animal drug application (NADA). Safety of the drug must be demonstrated, as must the safety of any other products that are produced from the treated animal, including milk or meat. This determination involves assessing the level of the drug that may be present in tissues or products, possible sites of bioaccumulation, and testing for any presence that could affect the food chain. In addition, levels of resistance to the drug must also be evaluated. ImmuCell was able to show that their new drug was safe and were granted a zero milk discard and zero meat withhold designation by the FDA (Brigham, 2011). To get this designation, ImmuCell had to perform several toxicology studies including chronic, acute, two-generation reproduction, and teratology. This allowed them to establish an acceptable daily intake (ADI) for humans. Metabolism studies were then done with radioactive-drug to ascertain what tissues the drug infiltrated when administered in the animal. Where the drug is stored in the animal, effects the ADI for each tissue or product produced by the treated animal. Residue Depletion studies are conducted to determine the concentrations of drug in the milk during and after infusion into the udder. These data are used to determine a withdrawal period for the milk, essentially the time it takes for the drug residues in the milk to decline to a safe concentration (permitted concentration). The concentrations of Nisin in milk after infusion of Mast Out® are within the safe range at all time points, and thus is granted a zero withdrawal designation Efficacy In order to get US FDA approval, a new drug must also be significantly better than a placebo and clinically relevant to the treatment of the target condition. A double-blind, placebocontrolled pivotal field study in several commercial dairy herds throughout the USA assessed the efficacy of Mast Out®. Two primary laboratory-based studies, two non-pivotal field trials and a final double-blind, placebo-controlled pivotal field study of Mast Out®, which included 295 cows with subclinical mastitis at 16 sites around the USA were conducted by ImmuCell. Mast Out® was significantly more effective at curing mastitis than the placebo control (P