The genes and factors that drive the conversion of the Pseudomonas aeruginosa Pf4 prophage into the superinfective form

Janice Gee Kay Hui

A thesis submitted in fulfilment of the requirements for the degree of

Doctor of Philosophy

School of Biotechnology and Biomolecular Sciences

Faculty of Science

The University of New South Wales

Sydney, Australia

December 2013

!

Originality Statement

‘I hereby declare that this submission is my own work and to the best of my knowledge it contains no materials previously published or written by another person, or substantial proportions of material which have been accepted for the award of any other degree or diploma at UNSW or any other educational institution, except where due acknowledgement is made in the thesis. Any contribution made to the research by others, with whom I have worked at UNSW or elsewhere, is explicitly acknowledged in the thesis. I also declare that the intellectual content of this thesis is the product of my own work, except to the extent that assistance from others in the project’s design and conception or in style, presentation and linguistic expression is acknowledged.’

Signed ……………………………………

Date ……………………………………

! Table of Contents

Table of Contents ...... I!

Acknowledgements ...... IV!

Abstract ...... VI!

List of Publications ...... VIII!

List of Figures ...... IX!

List of Tables ...... XI!

Abbreviations ...... XII!

Chapter 1 - Literature Review ...... 1!

1.1! Introduction to biofilms ...... 1!

1.2! Pseudomonas aeruginosa biofilms ...... 2! 1.2.1! Chronic infections and biofilms ...... 4! 1.2.2! Variant formation during biofilm development ...... 6! 1.2.3! DNA repair mutants and mutator phenotypes ...... 8! 1.2.4! The development of variants during biofilm development ...... 12! 1.2.5! Role of bacteriophage in biofilm development, dispersal and variant formation ...... 14!

1.3! Bacteriophage ...... 16! 1.3.1! Phage therapy ...... 19! 1.3.2! The lytic and lysogenic switch in lambda phage ...... 22! 1.3.3! Filamentous bacteriophage ...... 26! 1.3.3.1! Lifecycle of filamentous phage ...... 28! 1.3.4! Pseudomonas Pf phage ...... 30!

1.4! Aims of this study ...... 31!

Chapter 2 - The isolation, characterisation and biofilm formation of PAO1 biofilm isolates that carry superinfective Pf4 prophage ...... 32!

2.1 Introduction ...... 32!

I 2.2 Materials and Methods ...... 36! 2.2.1 Bacterial strains and culture conditions ...... 36! 2.2.2! Biofilm experiments ...... 36! 2.2.2.1! Continuous-culture flow cell biofilm ...... 36! 2.2.2.2! Identifying colony morphological variants from the biofilm effluent ...... 38! 2.2.2.3! Determination of plaque forming units (PFU) – Phage assay ...... 38! 2.2.2.4! Swimming and swarming motility assays ...... 38! 2.2.2.5! Attachment and biofilm formation assays ...... 39! 2.2.2.6! Confocal imagining and analysis ...... 39!

2.3! Results ...... 41! 2.3.1! Isolation and characterisation of phage resistant variants ...... 41! 2.3.2! Confocal microscopy of variant biofilms ...... 48!

2.4! Discussion ...... 53!

Chapter 3 – Mutations from superinfective variants ...... 59!

3.1! Introduction ...... 59!

3.2! Materials and Methods ...... 62! 3.2.1! Bacterial strains and culture conditions ...... 62! 3.2.2! Superinfection and resistance screening ...... 63! 3.2.2.1! Phage assay ...... 63! 3.2.3! Molecular characterisation of superinfective variant PAO1 SCV2 and S4 ...... 63! 3.2.3.1! Primers for genome walking ...... 63! 3.2.3.2! Genomic DNA extraction ...... 65! 3.2.3.3! Phage DNA extraction ...... 65! 3.2.3.4! Amplification for PAO1 SCV2 Pf4 phage genome ...... 66! 3.2.4! Biofilm experiments ...... 67! 3.2.4.1! Continuous-culture biofilm ...... 67! 3.2.4.2! Genomic DNA extraction from biofilm effluent ...... 68! 3.2.4.3! Phage DNA extraction from biofilm effluent ...... 69! 3.2.4.4! Determination of Colony Forming Units (CFU) ...... 69! 3.2.4.5! Determination of Plaque Forming Units (PFU) – Phage assay ...... 69! 3.2.4.6! Deep sequencing analysis ...... 70!

3.3! Results ...... 71! 3.3.1! Mutations within the PAO1 biofilm variants ...... 71! 3.3.2! Deep sequencing analysis of pre- and post- dispersal PAO1 biofilm ... 75!

3.4! Discussion ...... 81!

II Chapter 4 – Environmental cues and genes involved in establishment of the superinfective Pf4 phage ...... 86!

4.1! Introduction ...... 86!

4.2! Materials and Methods ...... 89! 4.2.1! Bacterial strains and culture conditions ...... 89! 4.2.2! Biofilm experiments ...... 89! 4.2.2.1! Planktonic cultures ...... 89! 4.2.2.2! Batch biofilms ...... 90! 4.2.2.3! Continuous-culture biofilm ...... 90! 4.2.2.4! CFU counts and morphological variants from the biofilm effluent 91! 4.2.2.5! Phage assay ...... 91!

4.3! Results ...... 92!

4.4! Discussion ...... 101!

Chapter 5 – General Discussion ...... 108!

5.1! Phenotypic and genotypic changes influenced by superinfective Pf4 phage ...... 109!

5.2! Proposed models ...... 114!

5.3! Conclusions ...... 116!

Appendix ...... 119!

References ...... 120!

III Acknowledgements

The completion of my thesis would not have been possible without the endless support from my supervisors, friends, family and colleagues. I would like to thank and show my appreciation to:

Associate Professor Scott Rice, for being the best of the best supervisor. I couldn’t have done it without your strong support. Your passion for science is inspirational. Your guidance and enthusiasm has motivated me to work harder and to explore the world of science. I am extremely thankful and extremely honoured to be your student.

Professor Staffan Kjelleberg, for giving me the opportunity to join the CMB. Thank you for your guidance and encouragement throughout the years. I’ve really enjoyed my time here and thank you for always believing in me.

Associate Professor Ruiting Lan and Dr Diane McDougald, for being my reviewer throughout my PhD. Thanks for the support and encouragement towards my project and for the great discussions and ideas.

Dr Torsten Thomas and Dr Kerensa McElroy, for the collaboration work and discussion. Thank you for your assistance and expertise with meta-genomics and sequencing analysis.

Dr Anne Mai-Prochnow, Dr Nicolas Barraud, Dr Janosch Klebensberger and Dr Krager Koh, for your help and support as a friend and a mentor. Thank you for your patience and your input towards my PhD. I’ve always looked up to you. Your attitude to life and work has been inspiring and has encouraged me to be more optimistic. Thank you for your support and encouragement.

Kirsty Collard, Adam Abdool, Leena Koop, Sharon Longford, Kylie Jones, Penny Hamilton for your help with all administrative matters and for making my PhD a smooth process.

Dr Anne Galea and Dr Rebecca Lebard, for giving me the opportunity to gain teaching experience. I truly admire you both as great course co-ordinators and thank you for your help and support.

IV Martin, Maria, Debra, Jerry, Leesia, Lan, Tran, Yuki, Erina, Anne, Amy, Gee, Martina, Carla, Mel, Adam, Tamsin, Becca, Nigel, Maisarah, Budoor, Brendan, Vanessa, Pejhman, Andrew, Bernard, Vivian, Hanae, Laetitia, Jeong, Lai, Lakshami, Menuk, Garfy, Nidhi, Ana-Maria, Hazlin, Chris, Valentina, Raymond, Vipra, Melani, Min Low and everyone in CMB. I’ve really enjoyed working with you all, and I really appreciate the endless love and support. I believe the friendship we’ve built throughout the years will continue forever. Thank you for watching out for me and for cheering me on til the end.

My friends in Sydney and all around the world, for all the support you’ve given me over the years. Sorry for missing out on dinners and events. Thank you for understanding and thanks for your endless love and support.

My family in Sydney, thank you for all your patience, encouragement and understanding. Thank you Fonda for all the amazing cooking and for your motherly love. Thanks to the Twins (Leanne and Julian) and AJ for hearing me out and being my brothers and sisters. Thanks for loving me as the nerdy Scientist cousin.

My family in Hong Kong, who allowed me to pursue my interests and have never given up on me. Mum and dad, thank you for loving and caring for me unconditionally, I love you both dearly. To Marco and Polo, sorry I have been away for so long, I will be back soon. To my dearest brother Ben, sorry I did not invent immortality potion for you during my PhD. Thank you for your support during my PhD, I love you.

Chao, thank you for loving and caring for me throughout these years, your endless support and patience has given me the strength to carry on. Thank you for giving me the reassurance I needed from time to time, and for believing in me and supporting my passion for Science. I love you.

V Abstract

Prophage have been identified in most sequenced bacterial genomes, however the effects of prophage genes on the host are poorly understood. Bacteriophage have been shown to play a role in the cell death phenomenon observed during biofilm development of Pseudomonas aeruginosa PAO1. The filamentous Pf4 prophage is important for mediating cell death within microcolonies, dispersal events and variant formation during biofilm development. Further, these effects were shown to result from the establishment of a superinfective phage phenotype. These observations have demonstrated the importance of phage activity in bacterial biofilms and their effects leading to biofilm variant formation and bacterial adaptation. However, little is known about the genetic mechanisms and triggers that lead to the conversion from lysogenic to the superinfective Pf4 phage. The aim of this study was to determine the factors and genes that induce the conversion of the lysogenic Pf4 phage into its lytic, superinfective form during P. aeruginosa PAO1 biofilm development.

Here, the mechanisms in the induction into the superinfective phage and the gene responsible for variant formation in the biofilm were demonstrated. Two morphotypic variants were isolated during the cell death and dispersal phase of the biofilm. These variants, the small colony variant SCV2 and spreader variant S4, carry a superinfective Pf4 phage and exhibited changes in motility and biofilm formation. Genetic analysis of these variants identified mutations within the immunity region of the Pf4 prophage genome. The mutations were identified to lie within the putative repressor c gene of the prophage and the putative promoter of the gene. Moreover, meta-genomic sequencing of pre- and post- dispersal biofilm populations identified the same mutations. The majority of the mutations were within the prophage genome at a frequency of up to 79% in contrast to mutations at less than 7% within the PAO1 genome. Collectively, these results suggest a role of the repressor C in superinfective phage conversion and strong selection for mutations within the immunity region of the phage to facilitate adaptation to biofilm lifecycle in the presence of phage infection.

VI Variant formation is a common trait during biofilm development, as a result of genetic changes in the biofilm community. Furthermore, it has been observed that the appearance of variants from the biofilm correlated with the occurrence of the conversion into the superinfective phage. Environmental stresses have previously shown to cause phenotypic variants and were here tested for induction of the superinfective phage of PAO1 biofilm. Reactive oxygen and nitrogen species have been shown to accumulate within the biofilm microcolonies. Biofilms exposed to oxidative stress induced the conversion of the superinfective phage. Phenotypic variants are a commonly isolated cystic fibrosis patients suffering from lung infection, and have shown to exhibit mutator phenotypes with lost of mut genes of the mismatch repair system. DNA damaging agent mitomycin C and mutS mutant biofilms have also shown to induce superinfective phage. These results have indicated that the oxidative stress and DNA damage are triggers to induce mutations within the biofilm resulting in the conversion into the superinfective form. Interestingly, the OxyR oxidative stress response regulator binds within the repressor c gene, suggesting a potential role of OxyR in the conversion of the superinfective phage.

Biofilms have been estimated to be associated with 80% of chronic bacterial infections and have been a challenge for treatment and therapy. While the mechanisms contributing to the conversion into the superinfective Pf4 phage remains elusive, this study has demonstrated the complexity involved in the study of phage infection in biofilms. Phage genes play a major role in the ability of the phage to cause infection against the host and provide resistance for host against infection. The proposed model involves oxidative stress induced mutations within the repressor C region of the prophage and lead to selection for variants that are resistant against superinfective phage. Thus, biofilm variants carrying the superinfective Pf4 phage persist within the biofilm. As P. aeruginosa is known to be a pathogenic bacterium and is involved in several biofilm-related diseases such as chronic infections and cystic fibrosis, it is important to understand the mechanism of biofilm development and to further improve the current treatments conducted.

VII List of Publications

Kerensa McElroy, Janice G. K. Hui, Jerry K. K. Woo, Alison Luk, Jeremy S. Webb, Staffan Kjelleberg, Scott A. Rice and Torsten Thomas. Strain-specific, parallel evolution drives short-term diversification during Pseudomonas aeruginosa biofilm formation. (2014) Proceedings of the National Academy of Sciences of the United States of America. Vol. 11, E1419-1427.

Anne Mai-Prochnow, Janice G. K. Hui, Staffan Kjelleberg, Diane McDougald, and Scott A. Rice. Big things in small packages: The genetics of filamentous phage and effects on fitness of their host. FEMS Microbiology Reviews.

Janice G. K. Hui, Anne Mai-Prochnow, Janosch Klebensberger, Diane McDougald, Kerensa McElroy, Torsten Thomas, Staffan Kjelleberg and Scott A. Rice. Mechanisms leading to establishment of superinfective Pf4 phage in Pseudomonas aeruginosa PAO1 biofilms. 14th International Symposium for Microbial Ecology Conference Copenhagen, Denmark 2012. Best Poster Award.

Submitted for publication

Janice G. K. Hui, Anne Mai-Prochnow, Diane McDougald, Staffan Kjelleberg and Scott A. Rice. Oxidative stress involved in induction of superinfection in Pseudomonas aeruginosa biofilm.

In preparation

Janice G. K. Hui, Anne Mai-Prochnow, Janosch Klebensberger, Diane McDougald, Kerensa McElroy, Torsten Thomas, Staffan Kjelleberg and Scott A. Rice. The role of prophage repressor in variant formation and superinfection Pseudomonas aeruginosa biofilm.

VIII List of Figures

Figure 1-1. The five stages of the development of P. aeruginosa biofilm...... 3 Figure 1-2. The MMR system mechanism in E. coli...... 10 Figure 1-3. The functions of RecA in bacteria...... 11 Figure 1-4. Confocal micrographs of biofilm dispersal resulting in the formation of hollow microcolonies and programmed cell death in P. aeruginosa biofilms. 16 Figure 1-5. The lysogenic and lytic lifecycle of E. coli lambda phage...... 24 Figure 1-6. The mechanism of the lysogenic lytic switch of lambda phage...... 24 Figure 1-7. The replication of filamentous phage DNA...... 29 Figure 2-1. Diagram of a biofilm flow cell system...... 37 Figure 2-2. Dispersal and phage production during biofilm development by P. aeruginosa PAO1...... 42 Figure 2-3. Colony morphology of biofilm-derived variants (white arrows) isolated from the PAO1 WT biofilm compared to the PAO1 WT colony morphology (red arrows)...... 43 Figure 2-4. Plaque formation determined using the soft agar lawn overlay method...... 45 Figure 2-5. Growth curves for the PAO1 WT (triangles), PAO1 SCV2 (circles) and PAO1 S4 (squares) in M9 medium...... 46 Figure 2-6. Comparison of A) swimming and B) swarming for the PAO1 WT, SCV2 and S4 variants...... 47 Figure 2-7. Confocal images of PAO1 WT, PAO1 S4, and PAO1 SCV2 biofilm development on days 3, 5, and 7 using an Olympus FV1000 confocal laser microscope...... 49 Figure 2-8. Quantitative comparison of biofilms formed by the WT PAO1 and the biofilm-derived variants S4 and SCV2...... 50 Figure 2-9. The percentage of SCV colonies isolated from the effluent of the PAO1 WT biofilm...... 51 Figure 2-10. The percentage of WT-like colonies isolated from the biofilm effluent of the PAO1 SCV2 variant biofilm...... 52 Figure 3-1. Alignment of filamentous phage M13, Pf1, Pf5 and Pf4 genomes. ... 73 Figure 3-2. Overview of the mutations identified in the variants carrying the superinfective form of the Pf4 phage...... 74

IX Figure 3-3. Dispersal and phage production of P. aeruginosa PAO1 biofilm submitted for deep sequencing analysis of pre- (white arrow) and post (black arrow) dispersal cells...... 76 Figure 3-4. Percentage of SCVs from the P. aeruginosa PAO1 biofilm submitted for deep sequencing analysis of pre- (white arrow) and post (black arrow) dispersal cells...... 76 Figure 4-1. The appearance of superinfective Pf4 phage during biofilm development of PAO1 WT (open triangle with dotted line), PAO1 treated with mitomycin C (hexagon), PAO1 treated with H2O2 (diamond) and PAO1 treated with SNP (inverted triangle)...... 95 Figure 4-2. The percentage of SCVs from the PAO1 biofilm untreated (black bars), treated with 150 µM of mitomycin C (grey bars), treated with 10 mM of H2O2 (white bars) and treated with 1 mM of SNP (stripes bars)...... 97 Figure 4-3. The role of OxyR in the development of superinfection...... 97 Figure 4-4. The percentage of SCVs from the dispersal population of the PAO1 WT biofilm (black bars) and the PAO1 oxyR mutant biofilm (white bars)...... 99 Figure 4-5. The appearance of the superinfective Pf4 phage during biofilm development for PAO1 (closed triangle), the mutS mutant (inverted triangle) and the recA mutant (hexagon)...... 99 Figure 4-6. The percentage of SCVs in the biofilm dispersal population for PAO1 WT (black bars), the mutS mutant (dark grey) and the recA mutant (light grey)...... 100 Figure 4-7. The location of the OxyR binding site within the ORF of the repressor c gene of Pf4 phage...... 103 Figure 5-1. The proposed mechanisms of the activation of superinfective Pf4 phage in P. aeruginosa PAO1...... 117 Appendix Figure 1. The SNPs (grey box) within and upstream of the repressor c gene of the Pf4 phage genome. ………………………………………..……… 120

X List of Tables

Table 2-1. List of P. aeruginosa strains used in this study...... 36 Table 2-2. Cross infection and resistance of variants and biofilm sampled from different days of the biofilm...... 44 Table 2-3. The percentage of morphotypes from a 5 d biofilm effluent of P. aeruginosa PAO1 biofilm that have a superinfective phenotype...... 45 Table 3-1. The mutations identified within the Pf4 phage genome from the variants of the P. aeruginosa PAO1 biofilm and dispersal population...... 79 Table 4-1. List of P. aeruginosa strains used in this study...... 89 Table 4-2. The phage assay using the PAO1 WT bacterial lawn to detect the presence of the superinfective phage in planktonic and batch biofilm cultures. .. 93

XI Abbreviations oC Degree Celsius Δ Delta: deletion/knockout % Percentage λ Lambda A Absorbance ANOVA Analysis of variance ATP Adenosine triphosphate bp Base pair c-di-GMP Cyclic-di-GMP

CaCl2 Calcium chloride CF Cystic fibrosis CFTR Cystic fibrosis transmembrane conductance regulatory CFU Colony forming units CFU/ml Colony forming units per millilitre cm Centimetre CRISPR Clustered Regular Interspaced Palindromic Repeats DNA Deoxyribonucleic acid dNTPs Deoxyribonucleotides dsDNA Double-stranded deoxyribonucleic acid dsRNA Double-stranded ribonucleic acid EDTA Ethylenediaminetetraacetic acid EPS Extracellular polymeric substances g Relative centrifugal force Gmr Gentamycin resistance h Hours

H2O2 Hydrogen peroxide HGT Horizontal gene transfer

KH2PO4 Monopotassium phosphate LB10 Luria Bertani medium with 10 % (w/v) sodium chloride LTTR LysR-type transcriptional regulators

MgSO4 Magnesium sulphate min Minute

XII ml Millilitre ml/h Millilitre per hour mm Millimetre mM Millimolar MMR Mismatch repair NA Non applicable NaCl Sodium chloride

Na2HPO4 Disodium phosphate

NaN3 Sodium azide ng Nanogram

NH4C2H3O2 Ammonium acetate

NH4Cl2 Ammonium chloride nm Nanometre nM Nanomolar NO Nitric oxide O2- Superoxide anion OD Optical density OH- Hydroxyl radical

OR1 Right operator site 1

OR2 Right operator site 2

OR3 Right operator site 3 PBS Phosphate buffered saline PCR Polymerase chain reaction PFU Plaque forming units PFU/ml Plaque forming units per millilitre pmol Picomolar

PR cro promoter

PRM cI promoter PS Packaging signal RF Replicative form RNA Ribonucleic acid ROS Reactive oxygen species RONS Reactive oxygen and nitrogen species rpm Revolutions per minute

XIII rRNA Ribosomal ribonucleic acid RSCV Rugose small colony variant s Second S Spreader variant SCFM Synthetic cystic fibrosis sputum medium SCV Small colony variant SDS Sodium dodecyl sulphate SEM Standard error of the mean SNP Sodium nitroprusside SNPs Single Nucleotide Polymorphism ssDNA Single-stranded deoxyribonucleic acid ssRNA Single-stranded ribonucleic acid TA Toxin-antitoxin Tris-HCl Tris-hydrochloride tRNA Transfer ribonucleic acid U Unit UV Ultraviolet µg Microgram µg /ml Microgram per millilitre µl Microlitre µm Micrometre µm3 Micrometre cube µM Micromolar UNSW University of New South Wales v/v Volume per volume WT Wild-type w/v Weight per volume

XIV Chapter 1

Chapter 1 - Literature Review

1.1 Introduction to biofilms

Bacteria in the environment are mostly found in multicellular aggregates such as slimes, also known as a biofilm. Biofilms are composed of microorganisms adhered onto surfaces enclosed within a self-produced extracellular polymeric substances (EPS) matrix. The EPS matrix contains proteins, polysaccharides and other substances and has been shown to provide stability and protection for microorganisms, which enables their survival in their natural environment (Flemming & Wingender, 2010). This enhanced resistance of the biofilm is one likely reason why biofilm formation is so widely observed for bacteria.

Biofilm formation is an important bacterial survival strategy and from a human perspective, this form of bacterial growth can be either beneficial or detrimental. For example, bacterial biofilms can be used as biocontrol agents based on their ability to suppress ascidia and algal growth on ship hulls (Zapata, et al., 2007) and to suppress the growth of phytopathogens on the roots of plant (Bais, et al., 2004). In contrast, in shipping as well as oil and gas industries, biofouling increases corrosion, decreases efficiency of pipelines and results in increased fuel consumption. These effects result in increased costs for repairing ships and also result in environmental pollution when corrosion of pipelines leads to leakage of oil into the environment. In the food industry, bacteria are commonly known to form biofilms and cause food spoilage (Carpentier & Cerf, 1993). Furthermore, contamination in food production plants can cause outbreaks of bacterial infections and food poisoning in the general public. In humans, biofilms contribute to a high percentage of infectious disease and chronic wound infections. Indeed, biofilms are reportedly responsible for up to 60% of all infections (Lewis, 2001). Biofilms are highly tolerant to antimicrobial drugs therefore they become hard to remove and prevent the healing processes. Additionally, biofilms cause concerns in the use of medical devices, where biofilm-associated bacteria can develop on implants or catheters (Carpentier &

1 ! Chapter 1

Cerf, 1993, Donlan, 2002, prevention, 2002, Curtin & Donlan, 2006, James, et al., 2007, Rybtke, et al., 2011). Overall, biofilms have been shown to represent financial burdens in a range of industries and thus, there is significant interest in finding novel, cost-effective strategies to disinfect or remove biofilm-associated problems. The understanding of biofilm formation is essential for the investigation of novel strategies for removal, prevention and treatment of biofilms. Pseudomonas aeruginosa has been the best studied biofilm forming bacteria and biofilms formed by P. aeruginosa have significant impacts in the environment and human diseases.

1.2 Pseudomonas aeruginosa biofilms

Biofilm development by P. aeruginosa has been extensively studied (Tolker- Nielsen, et al., 2000, Whiteley, et al., 2001, Sauer, et al., 2002, Haussler, et al., 2003, Klausen, et al., 2003, Webb, et al., 2003, Webb, et al., 2004, Williams & Camara, 2009, Hoiby, et al., 2010). P. aeruginosa is a Gram-negative, opportunistic pathogen that is ubiquitous in the environment (Haussler, et al., 1999). This organism is a highly diverse and ecologically significant organism as it exists in different environments including in association with various industrial systems such as those associated with water purifications and food handling and production (Grobe, et al., 1995, Thanomsub, et al., 2007), domestic homes (Finch, et al., 1978), plants and soil (Elrod & Braun, 1942, Green, et al., 1974), swimming pools (Rice, et al., 2012) as well as from contaminated sites (Wu, et al., 2008). It represents significant problems in the medical industry as it is related to a range of biofilm-related infections (Sadikot, et al., 2005) and is best known for its association with the lungs of cystic fibrosis patients (Bjarnsholt, et al., 2009). As a result, P. aeruginosa is one of the most commonly studied biofilm model organisms.

Bacteria develop into biofilm communities, which enhances their survival in their niche. In the planktonic state, free-living cells, bacteria are more exposed to environmental stress and are less protected. On the other hand, in the biofilm state, the cells are embedded in an EPS matrix, which shields the cells from 2 ! Chapter 1

external factors. Cells within a biofilm interact and communicate to form a communal community through signal molecules and regulatory mechanisms known as quorum sensing (Davies, et al., 1998, De Kievit, et al., 2001). Quorum sensing is a well understood mechanism of how cells communicate within a biofilm, and it has been reported that the lack of quorum sensing factors can lead to thin and dense biofilm structures in P. aeruginosa biofilms (Davies, et al., 1998).

P. aeruginosa biofilm formation generally undergoes a series of stages in a step- wise manner (Figure 1-1). These are typically described as reversible and irreversible attachment of free-living cells to surfaces, cell aggregation, formation of microcolonies and biofilm maturation (Sauer, et al., 2002). The mature biofilms are structurally complex, comprised of mushroom- and tower-like structures interspersed with open channels (Sauer, et al., 2002, Parsek & Fuqua, 2004, Rybtke, et al., 2011). This stage is followed by the dispersal of physiologically differentiated free-living cells from the biofilm, which completes the life cycle of the P. aeruginosa biofilm (Tolker-Nielsen, et al., 2000, Sauer, et al., 2002).

Figure 1-1. The five stages of the development of P. aeruginosa biofilm. Stages: 1) reversible surface attachment of planktonic cells; 2) irreversible surface attachment of planktonic cells; 3) maturation of cells; 4) microcolony formation; and 5) dispersal of biofilm cells from cell clusters. This imagine was adapted from (Monroe, 2007).

3 ! Chapter 1

1.2.1 Chronic infections and biofilms Bacterial biofilms are responsible for many medical infections such as chronic wound infections (James, et al., 2007) and lung infections in cystic fibrosis sufferers (Haussler, et al., 1999, Haussler, et al., 2003, Bjarnsholt, et al., 2009). Chronic wound infections are characterised as open skin wounds colonised with replicating microbial organisms that prevent wound healing and contribute to subsequent host tissue damage (Dow, et al., 1999, Edwards & Harding, 2004). In contrast to acute wound infections which are typically caused by free-living bacteria and are characterised as primary inflammatory response, bacteria in wounds have been shown to primarily be present as biofilms (James, et al., 2007), making them more difficult to treat as biofilm cells are more resistant against antimicrobial treatment than planktonic cells (Costerton, et al., 1999, Donlan & Costerton, 2002, Parsek & Singh, 2003). Indeed, biofilms have been shown to be 100 to 1,000 fold more resistant to antibiotic treatments than planktonic cells (Anwar & Costerton, 1990, Moskowitz, et al., 2004). Therefore, biofilm development is important for the development of chronic infection as the pulmonary functions decline and the lack of lung defense against inflammation allows the development of biofilms.

The bacteria most commonly found in chronic infections include Gram-positive bacteria such as Staphylococci and Enterococci, and Gram-negative bacteria such as P. aeruginosa, Escherichia coli, Klebsiella and Serratia species (Church, et al., 2006, Hammond, et al., 2011). There are a large range of biofilm-related chronic infections including but not limited to burn wounds (Trafny, 1998, Hammond, et al., 2011), diabetic wounds (James, et al., 2007, Bjarnsholt, et al., 2008), urinary catheter and biomaterials infections (Saint & Chenoweth, 2003, Ramage, et al., 2006) and ear infections (Hall-Stoodley, et al., 2006).

Cystic Fibrosis (CF) is one of the best studied biofilm-related diseases affecting humans, and is the leading cause of morbidity and mortality in patients diagnosed with CF. CF is an autosomal recessive disease that is caused by mutations in the cystic fibrosis transmembrane conductance regulatory (CFTR) gene. This results in a dysfunctional electrolyte secretion and absorption causing dehydrated airway surface liquid where the primary site of disease is in the respiratory system. The

4 ! Chapter 1

dysfunctional electrolyte secretion and absorption leads to a viscous mucosal layers lining the lungs, and creates a habitat favourable for bacterial colonisation (Haussler, et al., 1999, Haussler, et al., 2003, Bjarnsholt, et al., 2009, Rubin, 2009).

CF mostly affects the Caucasian population, with as many as one affected individual in 2,500 births (Rowntree & Harris, 2003). There are over 1,000 different mutations that have been identified in the CFTR gene and these mutations impact on pancreatic secretions, gastrointestinal function, fertility, lung and sinus functions (Rowntree & Harris, 2003). A dysfunctional electrolyte secretion and absorption causes increased mucus built up in the lungs and creates a favourable surface for bacterial colonisation and provides a binding site for bacterial adhesion. Mucus clearance defends the lungs from bacterial infection however CF patients develop dense mucus lining, which slows down clearance and facilitates the infection (Smith, et al., 1996, Donaldson, et al., 2006). As the infection progresses, the dense mucus and failure of mucus clearance affects the levels of airway fluids causing lack of secretion (Chmiel & Davis, 2003). This leads to chronic airway infections that can lead to respiratory failure. With changes in diet and modern medical treatments, most CF patients can survive the impaired organ function and consequently survive into adulthood. As a consequence of the increased life span, recent efforts have focused on controlling the effects of severe chronic lung infection which currently is the primary cause of mortality in CF patients (Kerem, et al., 1992, Mila & Warwick, 1998).

While the CF lung environment is conducive to infection, the infecting bacteria have also developed an additional strategy that facilitates chronic infection. It is understood that lung infection by P. aeruginosa originates from a single clone. Over time, this leads to the emergence of a population of variants that are less virulent than the parental strain but are more resistant to antibiotic treatment due to the selective pressure imposed by antibiotic treatment (Smith, et al., 2006, Mena, et al., 2008, Bragonzi, et al., 2009, Workentine & Surette, 2011). Such phenotypic changes that occur during adaptation to the lung include loss of motility, changes in mucoidy and antibiotic resistance (Thomassen, et al., 1979, Hancock, et al., 1983, Luzar, et al., 1985, Deretic, et al., 1995, Oliver, et al.,

5 ! Chapter 1

2000, Thomas, et al., 2000, Drenkard & Ausubel, 2002, Smith, et al., 2006, Starkey, et al., 2009, Ratjen, et al., 2010). Two distinctive morphology phenotypes are evident and correlate with chronic lung infections. They are described as a mucoid variant (Kirov, et al., 2007) and a rugose small colony variant (Starkey, et al., 2009). Comparison of clonally related P. aeruginosa rugose small colony variants (RSCV) strains isolated from CF sputum and laboratory biofilm cultures showed that RSCV display increased levels of the intracellular signalling molecule cyclic-di-GMP (c-di-GMP), as well as increased expression of the pel and psl polysaccharide genes (Starkey, et al., 2009). Both increased c-di-GMP and polysaccharide production are indicators of enhanced biofilm growth, suggesting that long term adaptation to the lung environment results in the evolution of strong biofilm forming variants. These traits of RSCV may contribute to the persistence in biofilms in the airways of CF lungs. Kirov et al. (2007) showed similarities between mucoid P. aeruginosa isolates from CF patients and laboratory derived mucoid P. aeruginosa PAO1 isolates that were collected from biofilms. This further supports the hypothesis that long-term chronic infection is connected to biofilm formation in vivo.

Variants have been shown to contribute to the persistence of biofilm cells in CF lungs. The changes in the environment such as lack of nutrient or starvation can lead to cells mutating to adapt to these environmental cues (Hunt, et al., 2004). Alternatively, variant formation may also be a consequence of exposure to stress, such as oxidative stress. Ultimately, most variants are the result of mutations within the bacterial population and these mutations increase the persistence of the bacteria in the CF airway environment where they may confer a selective advantage relative to the parental strain (Deretic, et al., 1995, Drenkard & Ausubel, 2002, Kirov, et al., 2007). It is interesting to note that variant formation is not limited to human clinical infections but has also been observed in food microbiology (Karatzas, et al., 2007) and in the rhizosphere of plants (Achouak, et al., 2004) suggesting that variant formation may be a more general biofilm phenomenon that is important for adaptation in multiple habitats.

1.2.2 Variant formation during biofilm development

The establishment of phenotypic variants during biofilm development has been 6 ! Chapter 1

widely reported in a range of bacteria, including Pseudomonas fluorescens (Workentine, et al., 2010), Vibrio vulnificus (Grau, et al., 2005), Streptococcus pneumoniae (Allegrucci & Sauer, 2007), Staphylococcus epidermidis (Conlon, et al., 2004). Some common phenotypes of such variants include increased antibiotic resistance, morphotypic variation such as small or mucoid colonies, and the increased expression of genes involved in environmental adaptation (such as psl genes, polysaccharide expression for surface adhesion) (Watnick, et al., 2001, Ma, et al., 2006, Karatzas, et al., 2007, Singh, et al., 2009). Interestingly, planktonic cultures do not show the same degree of variant formation and hence this may be a biofilm specific adaptation (Woo, et al., 2012).

In most instances, variants are isolated and characterised based on their changes in colony morphologies such as mucoid, small or mini, colour and wrinkly variants (Haussler, et al., 2003, Boles, et al., 2004, Webb, et al., 2004). When these isolates are further characterised based on their biochemical and physiological characteristics, they tend to also display a range of distinct phenotypes such as altered substrate utilisation and resistance (Besier, et al., 2008, Woo, et al., 2012). It has been hypothesised that these changes result in adaptation of the bacteria to the environment from which they were isolated. For example, isolates from CF patients, tend to develop a mucoid colony morphology due to the overproduction of the exopolysaccharide alginate which has been linked to increased protection from the host immune response (Doggett, et al., 1966, Gilligan, 1991). Additionally, these CF isolates have often have reduced motility and lack flagella (Luzar, et al., 1985, Mahenthiralingam, et al., 1994). Similarly, small colony variants isolated from chronically infected CF patients have increased biofilm forming ability as a result of expression of the chaperone usher pathway, cup, genes encoding putative fimbrial adhesins (Haussler, et al., 2003). Hypermutators are also a common phenotype of CF lung isolates that exhibit increased mutation rates due to a defect in the mismatch repair (MMR) system. Mutations in the MMR genes result in an increase in the spontaneous mutation rate and further leads to the development of adaptive phenotypes which are maintained due to a selective advantage in the niche from which they were isolated (Oliver, et al., 2002).

7 ! Chapter 1

In P. aeruginosa, several phenotypic variants have been described and these consist of both genetic and phenotypic variations with changes to their ability to reside in or form a biofilm, virulence factor expression, stress resistance and motility (Thomassen, et al., 1979, Hancock, et al., 1983, Luzar, et al., 1985, Deretic, et al., 1995, Oliver, et al., 2000, Thomas, et al., 2000, Drenkard & Ausubel, 2002, Smith, et al., 2006, Starkey, et al., 2009, Ratjen, et al., 2010). Variant formation has been suggested to be a mechanism by which biofilm cells generate genetic diversity to protect the biofilm community under changing environments and this phenomenon has been called the ‘insurance effect’. The Insurance hypothesis, which was originally described in relation to ecosystems, is a process whereby self-generated diversity provides a buffering effect by increasing genetic diversity, which is protective against future stressors. This means the increased diversity increases the odds that at last some species within the community can respond to variable conditions and replacement with functionally capable species (Yachi & Loreau, 1999, McCann, 2000). In terms of biofilms, the insurance effect relates to the diversity of the community composed of functionally diverse populations that allows the biofilm as a whole to thrive in adverse conditions. For example, mini and wrinkly variants isolated from P. aeruginosa PAO1 biofilms were identified to exhibit specialised biofilm functions. Furthermore, biofilms that produced variant subpopulations were more resistance against oxidative stress treatments supporting the hypothesis of variants produce insurance effect in communities under stress (Boles, et al., 2004). Similarly, mixed biofilms of S. marcescens variants were better protected from protozoa grazing than single variant biofilms. This demonstrates the importance of variant formation with mixed variation in different niche to provide greater protection for the biofilm as a whole, and also on the genetic level for enhanced protection to susceptible genotypes (Koh, et al., 2012). This also demonstrates that bacteria undergo genetic changes to survive in different niches, even in unfamiliar conditions and develop traits specialised for adaptation purposes.

1.2.3 DNA repair mutants and mutator phenotypes

P. aeruginosa is one of the most common bacteria associated with chronic pulmonary infections in CF patients. It is also the primary cause of mortality in

8 ! Chapter 1

these patients, which has been linked to its high adaptability to the CF lung and diverse phenotypes that are generated such as mucoid variants (Govan & Deretic, 1996, Lyczak, et al., 2002, Gibson, et al., 2003). P. aeruginosa generates numerous variants suggesting that this is an important trait for enhancing the ability to adapt to changing environment. Previous studies have shown that variants confer selective advantage in nature and leading variants dominating the population (Mao, et al., 1997, LeClerc, et al., 1998). Many of the P. aeruginosa strains isolated from CF patients also exhibit increased spontaneous mutation rates, termed mutators, and these strains typically carry mutations within the MMR system which is involved in DNA repair (Rayssiguier, et al., 1989, Matic, et al., 1995, Oliver, et al., 2000, Oliver, et al., 2002, Macia, et al., 2005, Hogardt, et al., 2006).

The MMR system is one of the most important DNA repair mechanism in bacteria and in higher organisms (LeClerc, et al., 1996, Matic, et al., 1997). The MMR system was originally described in E. coli, and consists of the proteins MutS, MutL and MutH, which are required for initiation of DNA repair (Figure 1-2). MutS functions by locating mismatched basepairs or small insertions or deletions that arise during DNA replication. MutL is recruited to form a complex with MutS to activate the endonuclease, MutH, to repair the mismatched basepairs (Modrich, 1991). Mutants defective in components of the MMR system have been shown to result in increased rates of mutation and recombination activity (Modrich & Lahue, 1996). Several CF isolates and non-CF isolates from chronic respiratory diseases have found to consist of mutS mutations resulting in a hypermutator phenotype, which are characterised as having up to a 1,000 fold increase in spontaneous mutation rates (Miller, 1996, Horst, et al., 1999). It has been reported that up to 37% of CF patients were colonised by hypermutable strains (Oliver, et al., 2000). Interestingly, hypermutators may be specific for chronic, long term colonisation where it was reported that of 75 acutely infected patients, none were colonised with hypermutable strains (Gutierrez, et al., 2004). The hypermutation phenotype may facilitate the establishment of successful phenotypes such as antibiotic resistance and resistance to host defenses (Oliver, et al., 2002, Macia, et al., 2005).

9 ! Chapter 1

Figure 1-2. The MMR system mechanism in E. coli. a) The MMR system initiates through the recognition of mismatch basepair in DNA by the MutS protein. b) In the presence of ATP, MutL binds to MutS to forms a complex. The binding of the MutS-MutL complex signals the activation and binding of the endonuclease MutH. c) The MutH endonuclease searches for the discrimination signal to create a nick in the strand to allow for resynthesis of basepairs. Figure reproduced from (Sixma, 2001).

Recombination repair is another important DNA repair system in bacteria. RecA plays an important role in DNA recombination, to align and pair DNA molecules and promote strand switching for DNA repair. RecA also plays a role in induction of the SOS response, which it achieves by facilitating the autocatalytic cleavage of the LexA repressor (Figure 1-3). The SOS response is activated in bacteria when exposed to high levels of DNA damage. RecA also activates the DNA polymerase V responsible for SOS mutagenesis, that replicates across DNA lesions and introduces mutations (Cox, 2007, Patel, et al., 2010). The SOS response of E. coli involves up to 40 genes that are induced for error-prone DNA- damage repair and regulation of cell division and this response is primarily regulated by the LexA repressor protein (Radman, 1974, Courcelle, et al., 2004). Functionally, RecA forms filaments on single-stranded DNA to mediate strand

10 ! Chapter 1

exchange for homologous recombination repair. RecA filaments also recognise DNA lesions that were bypassed during replication, left behind by other DNA repair systems or where replication forks are stalled (Courcelle, et al., 2001, Patel, et al., 2010). Because of its role in DNA repair, loss of a functional RecA can result in increased mutations or genetic diversity in the bacterial population. For example, Boles et al. (2004) demonstrated that loss of RecA resulted in less genetic diversity amongst the P. aeruginosa biofilm population where no morphological variants were detected. Since the formation of small colony variants was linked to stress resistance, loss of the RecA function, and hence reduced genetic diversity, resulted in a concomitant loss of stress resistance in the biofilm. Thus, variant formation, as driven by altered function of either the MMR system and RecA function is important for the resilience of biofilms and it is therefore important to understand the environmental factors that induce such mutations in the biofilm.

Figure 1-3. The functions of RecA in bacteria. During the SOS response, RecA forms filaments on ssDNA, which forms as a result of DNA damage caused by stress. Once bound to the ssDNA, RecA mediates recombination repair. Additionally, RecA induces cleavage of LexA, leading to derepression of the error-prone DNA polymerase II, IV and V. The error-prone DNA synthesis is activated by RecA to introduce mutation. Stress and SOS induction also induces phage (bacteriophage) expression and mobilisation. Figure adapted from (Smith & Romesberg, 2007). 11 ! Chapter 1

1.2.4 The development of variants during biofilm development

To understand the environmental factors that lead to the development of mutations and the establishment of variants within biofilms, it is important to first understand where and when such variants appear in the biofilm lifecycle. Variant formation is commonly observed during the dispersal phase in the biofilm (Webb, et al., 2004, Mai-Prochnow, et al., 2006). This suggests that there may be dispersal stage specific physiology that influences variant formation. Bacteria can switch between growth in the planktonic and biofilm states and do so based on the prevailing environmental conditions. In particular, there are a range of well studied cues and signals to which bacteria within the biofilm respond leading to their dispersal back into planktonic, free living cells (McDougald, et al., 2012). Biofilm dispersal is a naturally occurring process and may represent a mechanism for survival when the local biofilm conditions become unfavourable. Some of the environmental cues that induce biofilm dispersal include changes in nutrient availability (Sauer, et al., 2004), changes in oxygen concentration (Thormann, et al., 2005) and the production of nitric oxide, a by-product of anaerobic cell metabolism (Barraud, et al., 2006).

Several dispersal mechanisms have been described for a broad range of species such as E. coli (Jackson, et al., 2002), Xanthomonas campestris (Dow, et al., 2003), Actinobacillus actinomycetemcomitans (Kaplan, et al., 2003) and Staphylococcus aureus (Boles & Horswill, 2008). Signalling regulated processes have been shown to induce dispersal in bacterial biofilms such as the small diffusible molecule (DSF) which has been shown to control biofilm dispersal in X. campestris biofilms (Dow, et al., 2003) and the agr quorum sensing system that mediates dispersal in S. aureus biofilms (Boles & Horswill, 2008). A lysin oxidase protein, AlpP, has been shown to cause cell death and dispersal in biofilms through the production of hydrogen peroxide in several Gram-negative bacteria including Pseudoalteromonas tunicata, Chromobacterium violaceum, Caulobacter crescentus and Marinomonas mediterranea (Lucas-Elio, et al., 2005, Mai-Prochnow, et al., 2008). It was further suggested that the hydrogen peroxide production was an important mechanism for generating genetic diversity and variant formation. In this way, the self-production of a reactive oxygen species,

12 ! Chapter 1

hydrogen peroxide, may lead to DNA damage and hence the formation of phenotypic variants during biofilm dispersal.

In P. aeruginosa, a chemotaxis transducer protein, BdlA, was shown to mediate biofilm dispersal in response to environmental cues (Morgan, et al., 2006). The bdlA mutant showed reduced dispersal when exposed to dispersal cues such as succinate, and heavy metal exposure. It was also suggested that BdlA is part of a signalling cascade within the cells that in part controls the intracellular concentration of c-di-GMP, a key regulator of biofilm development and dispersal. Exposure to EDTA, a metal chelator, also induces dispersal of P. aeruginosa biofilms as well as cell death within biofilm microcolonies (Banin, et al., 2006). Similarly, byproducts of anaerobic metabolism such as nitric oxide (NO) can cause P. aeruginosa biofilm dispersal at low, sublethal concentrations. The loss of the nitrite reductase gene (nirS), responsible for generating metabolic NO through anaerobic respiration, resulted in no dispersal and the loss of the NO reductase gene (norCB) which removes NO, was correlated with enhanced dispersal and biofilm cell death (Barraud, et al., 2006). Davies and Marques (2009) identified a fatty acid messenger produced by P. aeruginosa, cis-2-decenoic acid, that induces biofilm dispersal within microcolonies. Furthermore, the same compound can induce dispersal of a variety of additional bacterial biofilms, including E. coli, Klebsiella pneumoniae, Proteus mirabilis, Streptococcus pyogenes, Bacillus subtilis and S. aureus. Cis-2-decenoic acid also acts as a cell-to-cell communication molecule and has the induce dispersal in Candida albicans biofilms showing its ability to cross-kingdom functional activity (Davies & Marques, 2009).

The formation of the EPS matrix may be an important factor in the accumulation of signals or metabolic by-products. The self-produced EPS matrix provides a cohesive matrix of polysaccharides, proteins, nucleic acids and lipids that physically bind bacterial cells together in the biofilm and that provides a protective barrier (Flemming & Wingender, 2010). In addition to preventing external compounds from reaching the underlying cells, the EPS matrix can also trap the by-products of metabolic reactions and built up waste products within the biofilm communities. This can lead to an accumulation of toxic compounds that

13 ! Chapter 1

can cause DNA damage or degradation to biofilm cells. As discussed above, aerobic metabolism generates a range of reactive oxygen species (ROS) and the diffusion of these away from the cells can be reduced within the enclosed matrix causing damage to the biofilm (Storz & Imlay, 1999, Vinckx, et al., 2010).

Nutrient limitation and starvation in the biofilm is a common environment biofilm cells are exposed to. It has been shown that P. aeruginosa biofilms dispersed upon exposure to increased concentrations of carbon substrates (Sauer, et al., 2004). Up to 80% reduction in surface-associated biofilm biomass was observed after exposure to glutamate as a carbon substrate. Dispersal cells showed the induction of a number of genes including flagellar motility, ribosomal proteins, kinases, and phage-associated genes in dispersal cells. In comparison, genes involved in denitrification pathways and pilus biosynthesis genes were upregulated in biofilm cells relative to the dispersed cells. Nutrient depletion can also induce biofilm dispersal (Hunt, et al., 2004, Schleheck, et al., 2009, Huynh, et al., 2012). These observations would indicate that natural biofilm dispersal may be due to a gradient of nutrient concentration within the microcolonies that results in cells in the biofilm interior experiencing starvation conditions. Similarly, changes in oxygen and carbon substrate concentration, pH, or other chemical parameters have been reported to induce dispersal of mature biofilms of P. aeruginosa, Pseudomonas putida, and Shewanella oneidensis (Applegate & Bryers, 1991, Tolker-Nielsen, et al., 2000, Sauer, et al., 2004, Gjermansen, et al., 2005, Thormann, et al., 2005). This supports observations that nutrient limitation within the biofilms is a general environmental cue for biofilm dispersal.

1.2.5 Role of bacteriophage in biofilm development, dispersal and variant formation

In addition to nutritional cues and some environmental stresses that have been shown to induce biofilm dispersal in P. aeruginosa PAO1, it was recently demonstrated that the filamentous phage, Pf4, also plays an important role in dispersal and variant formation. This phage is present in the genome of P. aeruginosa PAO1 as a prophage known as Pf4 phage, and consistent with the established model of replication of filamentous phage, the Pf4 phage is continually produced without killing the PAO1 host (Webb, et al., 2003). Further, 14 ! Chapter 1

the presence of the prophage protects the PAO1 host from reinfection by the phage. During biofilm development, the Pf4 phage has been shown to convert into a ‘superinfective’ form, where the phage becomes infectious towards the parental host (Rice, et al., 2009). Interestingly, the appearance of the superinfective Pf4 phage was correlated with two key events during biofilm development, cell death and dispersal as well as the evolution of genetic variants in the biofilm population.

When the Pf4 prophage was deleted from the host genome, it was demonstrated that biofilms of the Pf4 phage mutant no longer underwent cell death and hollow colony formation (Figure 1-4). Further, biofilms formed by the Pf4 deletion mutant were less stable than the wild-type (WT) and the Pf4 mutant strain was less virulent than the parental strain when tested in a mouse model of acute lung infection. The mutant biofilm also failed to generate morphotype variants, such as the small colony variants (SCVs), which were common in the WT biofilms and that appeared at the time when the biofilms were undergoing cell death. The role of the Pf4 phage in the formation of variants was further demonstrated when planktonic cells of PAO1 were infected with the superinfective phage, resulting in approximately 20% of the population being represented by SCVs. Thus, the prophage plays an integral role in the biofilm development of PAO1 (Rice, et al., 2009). The specific mechanism by which the Pf4 mediates these effects is currently unknown. However, it has been shown that specific phage products can have a significant impact on the host (Canchaya, et al., 2004). For example the bor and lom genes of E. coli lambda prophage confer serum resistance protection to the lysogen during growth (Barondess & Beckwith, 1990). Lambdoid phage carrying Shiga-like toxin genes of E. coli O157 induce damage in the gut during iron deficiency for iron supply (Wagner, et al., 2002). Therefore, elucidation of how the phage conversion occurs and the molecular mechanisms by which the Pf4 phage contribute to biofilm development will improve the overall understanding of its role modifying host behaviours, including biofilm development and virulence.

15 ! Chapter 1

Figure 1-4. Confocal micrographs of biofilm dispersal resulting in the formation of hollow microcolonies and programmed cell death in P. aeruginosa biofilms. (A) Seven day-old PAO1 WT biofilm with enhanced hollow microcolony formation. (B) Five day-old PAO1 WT biofilm exhibiting cell death in localised regions in hollow-microcolonies of the biofilm and a subpopulation of cells that were not killed inside microcolony (arrow). P. aeruginosa PAO1 biofilm was established in glass flow cells and visualised by using the BacLight LIVE/DEAD viability stain. Green fluorescence represents viable cells, while red fluorescent cells are dead cells. Left scale bar = 50 µm, and right scale bar = 25 µm. These images were adopted from (Webb, et al., 2003).

1.3 Bacteriophage

Bacteriophage are viruses that infect bacteria and were first described by Frederick William Twort in 1915 and Felix Hubert d’Herelle in 1917 (d'Hérelle, 1917, Ackermann & d'Hérelle, 1997). They are found in all habitats such as the ocean, soil, water and food. It has been estimated that there are on average 2.6 prophage per free-living bacterial cell (Lawrence, et al., 2002), and approximately 1031 bacteriophage particles on earth (Hatfull, 2008), making them the most widely distributed and abundant biological entity on the planet (Hendrix, 2003, Kutter & Sulakvelidze, 2005, Suttle, 2005, McAuliffe, et al., 2007, Clasen, et al., 2008). The importance of phage research has become more evident in the past three decades as phage play important roles in microbial ecology and evolution. Therefore, phage have significant impacts on ecosystem function, where bacteria are the drivers of all the major biogeochemical cycles, as well as potential applications for the control of infectious disease (Brockhurst, et al., 2005, Kutter & Sulakvelidze, 2005).

16 ! Chapter 1

Phage include viruses with double-stranded DNA (dsDNA), single-stranded DNA (ssDNA), double-stranded RNA (dsRNA) and single-stranded RNA (ssRNA). The majority of known phage are tailed, accounting for over 95% of all phage, although phage have been shown to display a broad range of morphologies including some with cubic, filamentous or pleomorphic shapes. They can exist in a lysogenic state (temperate) where the phage is integrated into the host genome or as extrachromosomal elements, analogous to plasmids, without causing lysis to the host. Alternatively, phage can exhibit a lytic form of replication (virulent) resulting in host cell death (Ackermann, 2003, McAuliffe, et al., 2007). Typically, phage display relatively narrow host ranges limited to single bacterial species. However, some phage are polyvalent and have broader host ranges, infecting multiple bacterial species and even spanning to other genera. The broadening of host range also influences the bacteriophage genome and its properties, which may suggest that infection of other species is important for supporting the viral population and evolution (Coetzee, 1987, Hendrix, et al., 1999, Hambly & Suttle, 2005, Hyman & Adbedon, 2010).

Initial studies on bacteriophage were focused on understanding the replication and infection cycles as well as the development of phage as agents for molecular techniques or for phage therapy to control infectious diseases (Ackermann & d'Hérelle, 1997, Ackermann, 2003). Bacteriophage have been shown to provide adaptive advantages for their host as well as driving microbial diversity. Bacteriophage in Bacillus anthracis confer resistance to the antibiotic fosfomycin (Schuch & Fischetti, 2006). Furthermore, it has also been shown that phage in B. anthracis provide the host with phenotypic advantages for ecological adaptation such as promotion of biofilm formation and EPS production and sporulation, and play important roles in colonisation in soils as well as in invertebrate intestines (Schuch & Fischetti, 2009). The major driving forces of bacterial evolution include the vertical transfer of spontaneous mutations, from mother cell to daughter cell during replication, as well as horizontal gene transfer (HGT). With respect to HGT, phage have been shown to be one of the important vehicles of gene transfer between bacterial species and thus can change the genetic capacity of the host by introducing new genetic material. The potential for phage to perform HGT was also exploited to develop some of the first molecular tools for 17 ! Chapter 1

introducing foreign DNA as a pharmaceutical agent (gene therapy) and were some of the first cloning vectors used in the development of molecular biology (Hendrix, et al., 1999, Canchaya, et al., 2003, Canchaya, et al., 2004). For example, the role of phage mediated HGT was harnessed in 1977 when the M13 coliphage was developed as one of the first molecular cloning tools. This transformed molecular microbiology, paving the way for modern molecular biology and genomic biology where manipulation of genome is dependent on cloning of DNA fragments (Messing, et al., 1977).

Bacteriophage have been suggested to play important roles in the marine environment because of their ability to infect and kill bacteria, resulting in the release of dissolved organic matter and nutrient recycling, and hence contributing to the marine microbial food chain. It has been estimated that approximately 20% of bacterial biomass in the sea is killed by phage daily (Suttle, 2007) and up to 50% of the overall microbial mortality is attributed to phage mediated lysis (Fuhrman & Suttle, 1993, Suttle, 1994, Fuhrman & Noble, 1995, Sadikot, et al., 2005). Lysed bacterial cells release their cell contents and this biomass is subsequently recycled as nutrients for other microorganisms. This process, also known as viral shunt, releases up to 0.63 gigatonnes of carbon per year making phage an essential player in the marine ecosystem and for prokaryotic metabolism (Middelboe, et al., 1996, Fuhrman, 1999, Wommack & Colwell, 2000, Danovaro, et al., 2008).

The role of phage as agents of HGT and the impact of HGT on bacterial evolution has been known for some time, however, the full significance of this process is only recently appreciated as a consequence of increased genome and metagenome sequencing activities. Bacterial genomes carry on average 2.6 prophage-like genomes and in some cases, carry up to 18 prophage elements, equivalent to 16% of the host’s genome content (Hayashi, et al., 2001, Canchaya, et al., 2003). Prophage are important for bacterial genomes because prophage are mobile genetic elements that can be transferred amongst bacteria contributing to pathogenic properties of the host, protection for the host, and adaptations to their environment. In some cases, bacteria gain virulence traits through phage mediated HGT. For example, the ability of V. cholerae to cause disease depends on the

18 ! Chapter 1

expression of cholera toxin and toxin regulated pili, which are encoded in the CTX phage (Kovach, et al., 1996, Karaolis, et al., 1999, Mukhopadhyay, et al., 2001). Moreover, it has been shown that the CTX phage can be exchanged between V. cholerae and V. mimicus, which demonstrates the lateral transfer between bacterial species (Boyd, et al., 2000). Thus, phage directly contributes to the emergence and evolution of new pathogenic bacterial strains (Mead & Griffin, 1998, Perna, et al., 1998, Canchaya, et al., 2004). Besides providing virulence factors to the host, bacteriophage can also transfer genes to a naive host providing new metabolic capacities. For example, the bacteriophage S-PM2 of Synechococcus encodes photosynthetic proteins as a result of HGT through cyanophage infection in the marine environment (Bailey, et al., 2004, Hambly & Suttle, 2005, Sullivan, et al., 2005).

1.3.1 Phage therapy and biocontrol Bacteriophage were classified as antibacterial agents in vitro because of the capacity of bacteriophage to cause cell lysis and there was much effort put into developing bacteriophage for use as therapeutic agents. Phage therapy was one of the leading approaches for the treatment of bacterial infections prior to the discovery and subsequent development of antibiotics in the 1940s. Both phage therapy and antibiotics were heavily used during the World War II. However, due to their broad host range, the ability to chemically synthesise pure compounds and the generally poor efficacy of bacteriophage, mainly due to a lack of understanding of the underlying biology, the ‘West’ ceased to pursue this path. In contrast, research on bacteriophage therapy continued in Eastern Europe. Due to the over use of antibiotics and the development of resistance against antibiotic drugs and improved understanding of phage and their hosts, phage therapy is once again gaining interest as a viable treatment option for infectious diseases (Carlton, 1999, Inal, 2003). Phage therapy poses several advantages over antibiotic treatments, as phage have the ability to evolve to reinfect a resistant host and can out replicate the bacterial host whilst killing the bacteria (Carlton, 1999, Inal, 2003). Furthermore, phage are highly specific for bacteria and do not infect mammalian cells. Indeed the specificity of the phage, normally at the species or even at the strain level, means that such treatments are less likely to affect the

19 ! Chapter 1

normal flora when used, unlike broad spectrum antibiotics (Debarbieux, et al., 2010).

Aside from phage therapy, phage have also been applied in the food industry as biocontrol measures for microbial safety and antimicrobial activities. Phage can contribute to various processes of the food industries from prevention of colonisation and disease, to disinfection and preservation of products (Gracia, et al., 2008). For example, a cocktail of three bacteriophage was used for prevention of colonisation of E. coli O157:H7 on meat surfaces and showed a 5 log reduction in pathogen numbers within an hour of treatment (O'Flynn, et al., 2004). More notably, the use of bacteriophage as biocontrol agents is common in diary industries including milk and cheese production. S. aureus is a major concern in the dairy industry where it is the primary pathogen-associated Mastitis in cows and is a contaminant in the milk product causing spoilage. A cocktail of bacteriophage has been used as a biopreservative to reduce the viable counts of S. aureus during curd manufacturing processes (Gracia, et al., 2007). Moreover, a combination of phage therapy and the bacteriocin nisin was also shown to be a useful biocontrol tool for the inhibition of S. aureus in milk (Gracia, et al., 2010). In the cheese industry, phage biocontrol was used to control the food pathogen Salmonella through the use of Salmonella phage SJ2 (Modi, et al., 2001). Phage biocontrol has also been used in wastewater treatment systems as a cost effective method for the control and removal of pathogenic bacteria from activated sludge (Bura, et al., 1998). Phage mediated lysis has applications in improved sludge dewatering to remove of excess water built up from microbes that interfere with activated sludge flocs (Kang, et al., 1989). Thus, phage-induced bacterial lysis can be utilised in various industries where bacterial activity is critical and/or problematic.

Given that biofilms are responsible for a significant proportion of infections, it is not surprising that there is interest in demonstrating that phage can also kill or control microbial biofilms. And indeed, there are reports of the use of phage for biofilm-related treatment of infectious disease in animals (Loc Carrilo, et al., 2005, McVay, et al., 2007) and in food applications (Sillankorva, et al., 2008, Pires, et al., 2011). Sillankorva et al. (2008) showed the phage phiIBB-PF7A

20 ! Chapter 1

could be used to remove P. fluorescens biofilms, a common food spoilage organism. The phage phiIBB-PF7A treatment removed up to 91% of matured biofilms and it was noted that the biofilm killing was more effective under dynamic conditions as compared to static conditions. Phage have the ability to access neighbouring cells in the biofilm and the lack of nutrient may cause the host to enter starvation state and influencing the phage lifecycle (Sillankorva, et al., 2008). Debarbieux et al. (2010) suggested the use of phage therapy for preventing and treating P. aeruginosa lung infections. Mice were infected with bioluminescent P. aeruginosa PAK strain for 2 hours followed by infection with P. aeruginosa PAK-P1 bacteriophage at 10:1 bacteriophage to bacteria ratio. Four hours after the bacteriophage was administered, the phage-treated mice had significantly reduced amount pathogen in their lungs relative to the non-phage- treated mice. Furthermore, 22 hours after bacteriophage was administered, phage- treated mice showed no or low light emission, an indication that the luminescent pathogen had been eliminated, while the non-phage-treated mice were dead or highly luminescent. It was concluded that the bacteriophage rapidly killed the bacteria in the lungs. Similarly, prevention studies, where mice were pre-exposed to phage followed by bacterial infection, showed prevention of lung infection (Debarbieux, et al., 2010).

Phage therapy also has applications in controlling biomaterials related infections, where biofilms play a key role in the infection process. Catheters that were pre- treated with P. aeruginosa M4 phage showed reduced attachment and biofilm formation by P. aeruginosa M4 in laboratory conditions. In another study, catheters were treated with a cocktail of phage isolated from biofilm variants with a broader host range. Results showed that the catheters pre-treated with the phage cocktail were resistant to the accumulation of biofilm compared to catheters pre- treated with P. aeruginosa M4 phage alone. These results suggest that the use of phage cocktails can reduce attachment of bacteria and subsequently biofilm formation and that their efficacy is greater than the use of single phage isolates (Fu, et al., 2010). Curtin and Donlan (2006) also demonstrated that phage can be use to reduce biofilm formation by S. epidermidis by pre-treating catheters with phage. Thus, it is clear that there are a number of applications for bacteriophage in the control of bacterial pathogens in medical applications. This is especially 21 ! Chapter 1

important for pathogens such as P. aeruginosa that establish recalcitrant biofilm infections, which are highly antibiotic resistant. Furthermore, bacteriophage represents low toxicity, high biocompatibility treatments that are biologically compatible with the human or animal host.

1.3.2 The lytic and lysogenic switch in lambda phage

Given that the Pf4 phage activity and superinfection have been shown to contribute to P. aeruginosa biofilm development and variant formation, it is important to understand how bacteriophage infect the host and replicate, and how bacteria become resistant to infection. Infection by a phage classically has been demonstrated to have two possible outcomes in relation to the phage replication cycle, the lysogenic and lytic cycles. Lysogeny is where a bacteriophage can exist within a bacteria host as a prophage or an extrachromosomal element, without causing cell death to the host (Brussow, et al., 2004). In contrast, lytic replication results in host cell lysis during phage release and hence, cell death. It is important to note that a lysogenic phage can ultimately activate the lytic replication cycle to escape the host. Phage lifecycle allows the phage to be passively replicated and inherited during cell division of the host. While this protects the host from lysis, it may also benefit the host by protecting it from reinfection by similar phage and the phage may additionally carry accessory genes that improve the fitness of the host (Boyd & Brussow, 2002). For example, E. coli carrying bacteriophage P2 encodes a bacteriophage resistance gene to protect the host from other phage infections (Mirold, et al., 2001). In the CTX phage of P. aeruginosa, the ctx gene encodes the cytotoxin and provides the host with bactericidal activities (Nakayama, et al., 1999). The phage-encoded superoxide dismutase gene confers Salmonella enterica with bacterial defense against macrophage oxidative burst (Figueroa-Bossi & Bossi, 1999, Figueroa-Bossi, et al., 2001, Uzzau, et al., 2001).

Phage lambda is the best studied temperate phage (Roberts & Roberts, 1975, Oppenheim, et al., 2005, Lesic & Rahme, 2008), which exists as a lysogenic prophage in E. coli. In its lysogenic state, the phage can replicate passively in the host without causing cell lysis. In its lytic state, the phage replicates and progeny phage particles are released out of the cell through lysis of the cell wall thereby killing the host (Figure 1-5). The mechanism of the lysogenic lytic switch of 22 ! Chapter 1

lambda phage is understood to involve a bistable genetic switch (Figure 1-6). Under lysogenic lifecycle, the repressor CI protein (also referred to as repressor C protein) binds to the operator sites of the cro promoter (PR) and regulates the cI repressor gene expression by a positive autoregulatory loop. The operator sites are stretches of palindromic sequences flanking the cI repressor gene at which the repressor C protein binds to prevent RNA polymerases from initiating transcription. Thus, the repressor C protein represses expression of the lytic phage cro genes and prophage is not activated, therefore no phage particles are expressed. The lysogenic lambda phage is stable as a prophage but can be induced in response to DNA damage and the activated SOS response. During the SOS response, RecA becomes a highly specific protease, which cleaves the repressor C protein (Roberts & Roberts, 1975, Roberts, et al., 1978). This inactivates the repressor C protein therefore allows transcription of lytic phage cro genes (Ptashne, 2004, Oppenheim, et al., 2005). The lytic phage replicates and assembles progeny phage that are released out of the cell through lysis of the host.

The decision to enter lysogenic or lytic lifecycle depends on the environment of the host. Oxidative stress (Los, et al., 2010), UV induced DNA damage (Boyce & Howard-Flanders, 1964), or exposure to antibiotics (Matsushiro, et al., 1999), can induce the lysogenic lytic switch in phage and activate the SOS response and RecA (Roberts, et al., 1978, Love & Yasbin, 1986, Martin, et al., 1995).

Mutations within the cI promoter (PRM) can also result in loss of lysogenic to lytic conversion. This demonstrates the importance of the Cro repression of the PRM promoter. Other mutations can also influence the lysogenic lytic switch (Harris, et al., 1967). The loss of the repressor protein in various bacteriophage of bacteria such as X. campestris pv. citri and Lactobacillus can lead to induction of the lytic lifecycle (Ladero, et al., 1998, Cheng, et al., 1999). Note that there are other regulatory proteins that play a role in the phage genetic switch. The mechanism described here, is a brief overview of the function of the switch. For detailed reviews, refer to (Ptashne, 2004, Oppenheim, et al., 2005).

23 ! Chapter 1

Figure 1-5. The lysogenic and lytic lifecycle of E. coli lambda phage. In lambda phage lifecycle, the viral DNA is integrated into the bacterial chromosome and replicates along with the host DNA. When the repressor responsible for repression of lytic phage genes fails to bind to the operator site, the phage lifecycle switches to lytic cycle. The release of new phage particles causes lysis to the lysogen. Figure reproduced from (Todar, 2012).

Figure 1-6. The mechanism of the lysogenic lytic switch of lambda phage. During lysogenic cycle (green arrows), cI gene encodes for the repressor C protein (represented as λ), which represses transcription of the Cro protein (represented as C) by binding to operator site OR1 and OR2 of the cro promoter (PR). During lytic cycle (red arrows), the cro gene encodes the Cro protein which blocks transcription of the repressor C protein, by binding to operator site OR3 of the cI promoter (PRM). This figure is adopted from (Hasty, et al., 2001).

24 ! Chapter 1

The repressor C protein also confers immunity to the host against additional phage infections. When a second, related phage enters the lysogenised host, the repressor C protein binds to the lytic operator site, suppressing expression of lytic phage genes thus conferring immunity to superinfection (lytic phage) by the new phage particles. (Hendrix, et al., 1983, Ptashne, 2004). Homologues of the repressor C protein have been identified in a number of different phage, including the P2 phage of E. coli (Ljungquist, et al., 1984) and the Pf4 phage of P. aeruginosa (Webb, et al., 2003). In contrast to lambda phage, the Pf4 prophage continually produces phage particles and this process does not result in host killing. Thus, Pf4 mediated cell death appears to be specifically associated with superinfection. However, the mechanism of killing is currently unknown and the role of the repressor C protein in controlling superinfection is also currently unknown (Rice, et al., 2009).

The incorporation of prophage into the host genome is not the only mechanism by which bacteria can protect themselves from infection. It has been recently shown that the clustered, regularly interspaced short palindromic repeat (CRISPR) system also confers phage resistance in bacteria and Archaea. CRISPRs were first discovered in 1987 in the genome of E. coli (Ishino, et al., 1987) and consist of repeats of 24 – 48 basepairs (bp) that are separated by non-repeating spacer sequences of 26-72 bp. Approximately 40% of bacterial and 90% of Archaeal genomes contain CRISPR elements suggesting that this system is widespread in microorganisms (Grissa, et al., 2007, Kunin, et al., 2007). Interestingly, the spacer elements were shown to have homology to phage sequences and after phage infection, new spacers are often incorporated into the CRISPR elements. In this way, the CRISPR represents a type of adaptive immune system of bacteria and Archaea and serves as a molecular history of phage infection for each strain. Phage infection poses a threat to the dairy industry, which relies heavily on bacteria for production of fermented products such as yoghurt. One application of the CRISPR system has been to engineer phage resistance into the starter culture bacteria by modifying the existing CRISPR elements or to introduce CRISPRs into those strains that do not naturally encode this resistance mechanism (Sturino & Klaenhammer, 2006, Sorek, et al., 2008). While CRISPR elements are present in a broad range of bacteria, including some P. aeruginosa strains, they are not 25 ! Chapter 1

present in P. aeruginosa PAO1 and thus, this strain must rely on more classic mechanisms of defense and immunity from phage.

1.3.3 Filamentous bacteriophage

Given that PAO1 does not encode the CRISPR phage immunity systems, resistance of PAO1 to infection by the filamentous phage Pf4 must be mediated by another mechanism, such as the repressor C system described above. As noted above, even this repressor protein is likely to function differently from what has been described in lambda phage. Therefore it is essential to understand the biology of filamentous phage in relation to their host. Filamentous bacteriophage are generally different from the lambda phage and other lytic phage in terms of replication mechanisms as well as effects on their hosts. Filamentous bacteriophage are viruses that have circular, single-stranded DNA genomes that target Gram-negative bacteria bearing retractile pili as receptors for infection. Filamentous phage can produce progeny phage particles within the host without killing the host. The Ff phage (f1, fd and M13) that infect E. coli are well characterised and have played important roles as cloning tools (Russel, 1995, Marvin, 1998). Based on the Ff phage genomes, there are approximately ten open reading frames that are responsible for replication, structure and secretion/integration. Replication is normally initiated from an extrachromosomal element, termed the replicative form (RF) of the phage. This double stranded, circular copy of the phage generated new copies of the ssDNA phage genome via rolling circle replication. Three of the phage-encoded proteins are responsible for replication (pII, pV, and pX). These proteins facilitate the replication of the RF DNA to synthesis newly single-stranded DNA for phage assembly. The ssDNA genome is coated with thousands of copies of the major coat protein (pVIII) that forms the tubular phage structure. Four smaller proteins (pIII, pVI, pVII and pIX) are minor capsid proteins and lack of these proteins has shown to prevent subsequent infection and failure of release of mature phage particles from the host. Two additional phage-encoded proteins (pI and pIV) are required for phage secretion and mutants of these phage genes have been shown to alter phage particle production, and in some cases these mutants are lethal to the host cell (Model & Russel, Pratt, et al., 1966, Marvin & Hohn, 1969, Russel, 1995,

26 ! Chapter 1

Marvin, 1998). For example in E. coli f1 and fd infection, mutants defective for all except replication gene II, resulted in killing of the host as the release of progeny particles is prevented (Hohn, et al., 1971, Woolford, et al., 1974, Horabin & Websster, 1986).

Filamentous phage have been described for a range of Gram-negative bacteria such as the Pseudomonas Pf phage (Minamishima, et al., 1968, Hill, et al., 1991, Luiten, et al., 1995, Webb, et al., 2004, Tan, 2006, Mooij, et al., 2007, Rice, et al., 2009, Klockgether, et al., 2010), Xanthomonas Cf phage (Kuo, et al., 1994, Cheng, et al., 1999), Vibrio CTX phage (Davis & Waldor, 2003) and the Ralstonia RSM phage (Yamada, et al., 2007). The Pseudomonas Pf phage have similar genomes to the Ff phage in E. coli but differ in the overall particle morphology where the Ff phage have five fold screw axis helical symmetry while Pf phage do not (Model & Russel). Six Pf phage have been described to date (Minamishima, et al., 1968, Bradley, 1973, Bradley, 1973, Hill, et al., 1991, Luiten, et al., 1995, Tan, 2006, Mooij, et al., 2007, Sillankorva, et al., 2008), and for Pf4 and Pf5, studies focused on their role in biofilm development. It was demonstrated that the Pf4 and Pf5 phage contributed to variant formation during biofilm development (Webb, et al., 2003, Mooij, et al., 2007, Rice, et al., 2009). Additionally, the Pf4 phage was shown to contribute to biofilm development of PAO1 (Rice, et al., 2009). Similarly, Vibrio harbours a CTX filamentous phage, which has also been linked to the virulence of Vibrio cholerae. V. cholerae carries two key virulence factors known as the toxin co-regulated pilus and exotoxin cholera toxin, which were introduced into Vibrio through HGT of CTX phage. (Davis & Waldor, 2000, Davis & Waldor, 2003). Other examples of filamentous phage shown to contribute to the pathogenicity of their host and influencing the behaviour of the host include increased EPS synthesis and enhanced PilA and type IV pili production for improved twitching motility identified in Ralstonia solanacearum that harbours RSSI prophage. Similarly, mice infected with Yersinia pestis harbouring Ypf prophage were more virulent compared to the host lacking a prophage (Derbise, et al., 2007, Addy, et al., 2012).

27 ! Chapter 1

1.3.3.1 Lifecycle of filamentous phage

The lifecycle of filamentous bacteriophage is different from the lambda phage lifecycle described above (Figure 1-6). The lifecycle of filamentous bacteriophage was studied using Ff phage and begins with the adsorption of the phage to the pilus of the host cell (Figure 1-7). Once the phage is adsorbed to the surface of the cell membrane, the phage penetrates through the membrane, leaving coat proteins in the inner membrane and the DNA injected into the cytoplasm. The host enzymes synthesise the phage DNA to form the dsDNA RF, which then acts as a template for gene expression (Russel, 1995, Marvin, 1998). Gene expression results in the production of structural proteins for the assembly of the phage particle. The newly synthesised and assembled progeny phage are released through the host cell wall, a process called budding (Marvin & Hohn, 1969, Day, et al., 1988). The release of new filamentous phage particles from the host does not result in cell death, unlike lambda phage where the lysogen is killed once the phage enters the lytic cycle. (Hendrix, et al., 1983, Oppenheim, et al., 2005)

The replication genes play an important role in the regulation of copy number of RF and ssDNA in the cell to prevent host cell death. The replication protein pII initiates the replication of the RF for the production of new viral DNA (ssDNA) (Higashitani, et al., 1993). The ssDNA-binding protein pV is important for regulating the translation of replication proteins pII and pX as well as the formation of the pV-ssDNA complex, required as the primary structure for the assembly of progeny phage at the membrane (Michel & Zinder, 1989). The level of RF determines the rate of phage protein synthesis whilst the ssDNA level determines that rate at which the phage proteins expressed are released from the cell. It has been hypothesised that these proteins are important to maintain the balance between the replication process and assembly of progeny phage (Russel, 1995). Furthermore, filamentous phage require the facilitation of host enzymes during replication and assembly of progeny phage. In E. coli, filamentous phage assembly also requires the assistance of a host protein, thioredoxin, which may act as stabilising proteins during phage assembly (Russel, 1995, Marvin, 1998).

The Ff phage such as fd, f1 and M13 phage exist as a plasmid in the host whereas the temperate or lysogenic phage integrates into the host chromosome. Integration 28 ! Chapter 1

of filamentous phage can occur via its own integrase and transposase genes, or via host-encoded genes such as the XerCD recombinases. The XerCD recombinases of the host mediate integration of CTX phage into a host tRNA site of V. cholerae (Hubor & Waldor, 2002). In Ralstonia solanacearum, the RSM1 filamentous prophage catalyse the site-specific recombination between the phage and host attachment sites via a phage-encoded integrase for integration (Askora, et al., 2011). Two of the Pseudomonas Pf phage have been shown to integrate into the host tRNA site and this may have been facilitated through its own integrase genes (Webb, et al., 2004, Mooij, et al., 2007).

Figure 1-7. The replication of filamentous phage DNA. The phage DNA is injected into the host cell and is then converted from single-stranded phage DNA (+) into supercoiled, double-stranded replicative form (+/-) facilitated by host enzymes. The pII phage protein facilitates the replication of the viral DNA using the replicative form as a template and pV phage protein prepares the phage DNA for assembly. The pV complex is transferred to the cell membrane where the phage particle is assembled with the structural proteins pIII, pVI, pVII, pVIII and pIX. The membrane proteins pI, pXI and pIV facilitate the release of the phage particle through the cell membrane. Figure adopted from (Rakonjac, et al., 2011).

29 ! Chapter 1

1.3.4 Pseudomonas Pf phage

There are six known Pf phage (Minamishima, et al., 1968, Hill, et al., 1991, Luiten, et al., 1995, Webb, et al., 2004, Tan, 2006, Mooij, et al., 2007, Rice, et al., 2009, Klockgether, et al., 2010) and similar to the Ff phage, the Pf phage do not cause cell lysis. However, less is known of the Pf phage lifecycle and the replication and infection mechanism. Most Pf phage have almost no homology with Ff phage at the nucleotide or amino acid level, however the organisation of the genome resembles the Ff phage in respect to the overall pattern of genes based on size and locations of major and minor coat proteins. For example, the genome of the Pf3 phage, which is 5,833 bp has a similar gene organisation to M13 phage of 6,407 bp (Luiten, et al., 1995). Similarly, the Pf2 phage was isolated from P. aeruginosa P28 and has no homology besides genome organisation with Ff phage genomes (Minamishima, et al., 1968). In the MPAO1 and PAO1-DSM sublines of P. aeruginosa, a secondary Pseudomonas filamentous phage was identified (Klockgether, et al., 2010). Its genome resembles the Pf phage, however the physiology of this phage (designated Pf6 here (Tay, 2013)) has not been investigated.

Many of the Pf phage have been identified based on morphology or genome content, however the effect of these phage on the host has largely been understudied. In contrast, it has been shown that the Pf4 phage of P. aeruginosa PAO1 contributes significantly to biofilm development, where it specifically causes cell death in microcolonies as well as colony expansion and was shown to play a role in dispersal and a phage deletion mutant was less virulent during lung infection (Webb, et al., 2003, Webb, et al., 2004, Rice, et al., 2009). The phage also was linked to the formation of morphological variants, specifically SCVs during biofilm development. In contrast, the Pf5 phage of P. aeruginosa PA14 (Mooij, et al., 2007) did not appear to play a role in the appearance of SCVs from P. aeruginosa PA14. These differences in effects could either be due to intrinsic differences in the phage, e.g. slight differences in genome content or in the host or could be due to differences in experimental conditions. For example, the Pf4 specifically induces SCV formation when the phage is superinfective and the WT phage does not induce SCV formation in the PAO1 WT. Additionally, SCVs are

30 ! Chapter 1

typically only observed during biofilm development, at the time when the superinfective phage can be detected in the biofilm effluent, which corresponds to the timing of dispersal. Thus, many of the effects of the Pf4 phage on biofilm development of PAO1 are closely linked to the establishment of the superinfective form of the phage. It was suggested that the type IV pili act as a receptor for the Pf4 phage as cells that cannot produce type IV pili were shown to confer resistance against phage infection (Castang & Dove, 2012). It was also shown that histone-like nucleoid-structuring (H-NS) proteins MvaT and MvaU represses gene expression of Pf4 phage. The loss of functional MvaT and MvaU proteins results in increase in Pf4 production leading to cell death and inhibition in cell growth. To date, no mechanism at the molecular level has shown how phage can reinfect the prophage carrying host and it is unclear what conditions drive this change during biofilm development.

1.4 Aims of this study

The Pf4 prophage genes are some of the most highly upregulated in the PAO1 genome during biofilm development and the Pf4 phage significantly contributes to multiple aspects of biofilm development, dispersal and virulence of the host. Therefore, it is important to understand how the phage, can mediate these effects. Further, many of the effects of the phage are linked to the conversion of the phage into a superinfective form that can reinfect the WT PAO1 host. For example, exogenous addition of the superinfective phage to a biofilm, leads to cell death, which is an essential stage of biofilm development. Given the dependence of these biofilm phenotypes on the formation of the superinfective form of the phage, it is essential to understand this process. However, to date, no data has been published demonstrating a mechanism that allows for Pf phage to become superinfective. Therefore the aims of this study were:

1. To isolate and characterise superinfective mutants. 2. To determine the genes involved in the control of superinfection. 3. To determine the physiological conditions with then the biofilm that may trigger superinfection.

31 ! Chapter 2

Chapter 2 - The isolation, characterisation and biofilm formation of PAO1 biofilm isolates that carry superinfective Pf4 prophage

2.1 Introduction

The Pf4 prophage in P. aeruginosa PAO1 has been shown to play a crucial role in biofilm development where it influences maturation, cell death, dispersal and variant formation during biofilm development. Further, these effects were manifested in the biofilm when the Pf4 phage acquired a superinfective phenotype. The Pf4 mutant biofilm undergoes limited microcolony formation and there was no cell death within the biofilm, which are normally observed in the WT PAO1 biofilm. This demonstrates that the prophage plays an important role in the environmentally relevant phenotype of biofilm formation (Rice et al., 2009).

In addition to contributing to cell death and dispersal, the appearance of the superinfective form of the Pf4 phage in the biofilm effluent coincides with the appearance of phenotypic variants in the dispersal population. The effluent from biofilms typically contained 10% - 20% of variants, which were dominated by small colony variants (SCVs) as well as spreader and wrinkly variants in lower proportions. The addition of superinfective Pf4 phage to planktonic cultures of P. aeruginosa was also shown to result in the formation of SCVs, which do not appear when identical cultures were infected with the WT phage or in the uninfected controls (Rice, et al., 2009). It was also shown that addition of the superinfective phage to the biofilm resulted in cell death within the WT biofilm indicating that the superinfective filamentous phage could cause cell death during biofilm development. These data suggest that the superinfective phage is responsible for the formation or selection for morphotypic variants, which may play important roles in biofilm development.

32 ! Chapter 2

Phenotypic variation is a common trait observed in Gram-negative bacteria such as Salmonella enterica serovar Typhimurium (Zogaj, et al., 2001), P. fluorescens (Spiers, et al., 2002), and S. marcescens MG1 (Koh, et al., 2012) and they represent an important survival strategy (Henderson, et al., 1999). The formation of phenotypic variants from bacterial biofilms have been extensively studied. Starkey et al. (2009) showed that in the sputum from CF patients and in laboratory biofilms, P. aeruginosa biofilm variants had increased production of intracellular signalling molecules and higher expression of polysaccharide genes for biofilm matrix formation. In V. cholerae O139, rugose variants were identified as flagella mutants with better biofilm formation capability due to the induction of exopolysaccharide production. Exopolysaccharide production is important for V. cholerae survival and adaptation in various habitats and hence serogroup strain O139 is a better biofilm former on abiotic surfaces compared to the ancestor strain serogroup O1 (Watnick, et al., 2001). Similarly, various phenotypic variants isolated from S. marcescens were better biofilm formers than the WT biofilm which may suggest that biofilm-derived variants are beneficial for long term survival in the natural habitat (Koh, et al., 2012). In the natural environment, P fluorescens biofilms that were exposed to heavy metals generated small colony and wrinkly variants that had increased resistance to copper ions and silver ions, respectively (Workentine, et al., 2010). These adaptive features are important for persistence and survival for bacteria to evolve in the wild.

There are numerous P. aeruginosa biofilm-derived morphological variants that have been previously identified, such as SCVs, sticky and wrinkly variants (Drenkard & Ausubel, 2002, Boles, et al., 2004, Webb, et al., 2004, Kirisits, et al., 2005). The hypothesis is that such phenotypic variants arise randomly in the population, where they are maintained or tolerated, and represent a pool of variants that may have a selective advantage which would be beneficial at a later time when the local environmental conditions change. Furthermore, mutation and selection is responsible for the appearance and maintenance of these variants (Haussler, et al., 2003). For example, it has been shown that the wrinkly variant, identified from a collection of P. aeruginosa PAO1 transposon mutants, has a dysfunctional regulator responsible for fimbrial adhesin production, which is required for biofilm formation (D'Argenio, et al., 2003). The wrinkly variant has a 33 ! Chapter 2

faster growth rate over the WT P. aeruginosa at low temperature. P. aeruginosa SCVs isolated from the respiratory tract of CF patients, show autoaggregative properties in liquid culture and improved capacity to form biofilms (Haussler, et al., 2003). Drenkard and Ausubel (2002) also showed that not only were P. aeruginosa PA14 phenotypic variants, isolated from biofilms in the lungs of a CF patient, more antibiotic resistant, but they also displayed enhanced biofilm formation. Furthermore, these variants also had improved resistance to a range of other antibiotic treatments. Genetic diversity in biofilms has been shown to require the RecA mediated recombination function. Phenotypic variants were observed from 5 day old P. aeruginosa PAO1 WT biofilm with up to 48% of mini and wrinkly variants and no variants from the recA mutant biofilm. Furthermore, other specialised biofilm phenotypes including swimming motility, pyomelanin production and auxotrophic phenotypes were shown to be RecA dependent in the WT biofilm. The mini and wrinkly variant had specialised biofilm phenotypes of accelerated detachment and hyper-biofilm formation, respectively. Moreover, the wrinkly variant was responsible for the increased resistance of the WT biofilm when exposed to oxidative stress and tobramycin. This demonstrated that genetic diversity generated by biofilms is dependent on RecA and that the selected variants confer biofilm specific phenotypes that are more resistant to stress. (Boles, et al., 2004)

The infection by phage has been shown to select for infection resistant variants and this has been linked to changes in lipopolysaccharide production or surface receptors as attachment of phage to bacteria is the first step to phage infection (Levin & Bull, 2004). As a consequence of changes to the cell surface, such phage infection mutants also appear as morphotypic variants. In P. aeruginosa, phage induced mucoid variants were observed from patients diagnosed with chronic diseases in the respiratory tract and it has been acknowledged that the appearance of mucoid strains in the lungs of CF patients is associated with a poor prognosis (Martin, 1973, Mathee, et al., 1999). Similarly, phage induced smooth variants from Myxococcus xanthus showed poor biofilm development and are defective in motility (Ruiz-Vazquez & Murillo, 1984). Thus it is clear that phage infection has strong influence on the evolution of bacteria and more importantly for their survival, how they become resistant to subsequent phage infections. The majority 34 ! Chapter 2

of these studies have focused on the roles of lytic phage in the formation of morphotypic variants and few studies have investigated the role of filamentous phage, which do not normally lyse the host, in the process of variant formation in the evolution of infection resistance. The goal of the work presented in this chapter was to determine the role of superinfective filamentous phage in the establishment of variants and how those variants may alter biofilm development. Further, variants were isolated and characterised after infection to investigate the mechanisms of resistance to filamentous phage infection.

35 ! Chapter 2

2.2 Materials and Methods

2.2.1 Bacterial strains and culture conditions

The P. aeruginosa strains used in this study are listed in Table 2-1. All P. aeruginosa strains were cultured in Luria-Bertani (Bertani, 1951) medium supplemented with 1% (w/v) NaCl (LB10) or in M9 minimal medium containing:

48 mM Na2HPO4, 22 mM KH2PO4, 9 mM NaCl, 19 mM NH4Cl, 2 mM MgSO4,

100 µM CaCl2, supplemented with 15 mM glucose (M9 complete medium). Strains were maintained on LB10 agar (1.5% w/v agar) plates and incubated overnight at 37oC. Liquid bacterial cultures were incubated overnight at 37oC with constant agitation on a shaker at 200 rpm.

Table 2-1. List of P. aeruginosa strains used in this study.

Strain Colony morphotype - form, elevation, Reference margin, diameter

PAO1 WT Circular, convex, entire, diameter 2 mm Laboratory stock

PAO1 ΔPf4 Circular, convex, entire, diameter 2 mm Rice et al., 2009

PAO1 SCV2 Circular, convex, entire, diameter 0.5 mm This study

PAO1 S4 Irregular, raised, undulate, diameter 3 mm This study

2.2.2 Biofilm experiments

2.2.2.1 Continuous-culture flow cell biofilm Biofilms were cultivated in flow cells (Figure 2-1), as described (Moller, et al., 1998), with some modifications. A glass coverslip (6 x 2.4 cm) was adhered onto the flow cell with silicone glue (Plastic Putty Selleys Pty Ltd.) and allowed to set overnight. The flow cells were sterilised by soaking in 10% (v/v) bleach for 90 min and connected to sterile silicon tubing (inner diameter 2.64 ± 0.28 mm and outer diameter 4.88 ± 0.28 mm) (Silastic® laboratory tubing) attached to the biofilm flow cell system (Figure 2-1). Prior to inoculation, the flow cell system 36 ! Chapter 2

was flushed thoroughly with M9 minimal medium for 4 h at a flow rate of 6 ml/h to remove bleach residues within the flow cells.

Figure 2-1. Diagram of a biofilm flow cell system. The M9 minimal medium supplemented with glucose was pumped through the biofilm system using a peristaltic pump at a constant speed of 6 ml/h to supply the flow cell chamber with continuous fresh medium. The effluent was collected at the waste end into a large flask. Prior to inoculation, clamps are used (indicated in arrows) to restrict the inoculum from flowing towards the pump and medium reservoir and to allow bacteria to colonise within the flow cell chamber. Figure adapted from (Tay, 2008).

Inoculation of the flow cells was performed while the pump was switched off and the inlet end of the flow cell was clamped to prevent back flow of the inoculum towards the medium end. Two millilitres of overnight cultures of PAO1 WT, PAO1 SCV2 or PAO1 S4 were injected, using a syringe and a 26G needle (Becton-Dickinson Precision Glide Needle), into the inlet end of the flow cell. The opening formed by the needle was sealed using silicone glue. The outlet ends of the flow cells were clamped and the flow cells were inverted (glass cover slip face down) for 1 h without flow to allow the bacterial culture to attach to the surface of the flow cell. Medium flow was resumed at a flow rate of 6 ml/h. All biofilm flow cell experiments were performed at room temperature with three replicates.

37 ! Chapter 2

2.2.2.2 Identifying colony morphological variants from the biofilm effluent

To observe colony morphological variants from the biofilm, 5 ml of the biofilm effluent was collected from each flow cell on days 1, 3, 5, 7, 9 and 11. The samples were serially diluted in M9 minimal medium without glucose and 100 µl from the dilution tubes was spread plated onto LB10 agar. The plates were incubated overnight at 37oC. The plates were incubated for additional 12 h at room temperature to facilitate observation of morphotypic variants.

2.2.2.3 Determination of plaque forming units (PFU) – Phage assay

The phage titre assay was performed to detect and quantify the superinfective form of the phage using a modified version of the top-layer agar method previously described by (Eisenstark, 1967). Briefly, the biofilm effluent was centrifuged at 13,000 x g for 5 min and filtered through a 0.22 µm filter (Millipore Millex GP) to obtain cell-free supernatant. The supernatant was serially diluted and two, 10 µl drops were spotted onto LB10 agar plates containing an overlay of the bacterial lawn, i.e., PAO1 WT, PAO1 SCV2, PAO1 S4 or PAO1 ΔPf4.

The overlay of the bacterial lawn was prepared by mixing 500 µl of an overnight culture, grown in M9 complete medium, with 5 ml of 0.8% (w/v) molten LB10 agar that had been cooled to 55oC in a water bath. The mixture was poured onto a LB10 agar and was allowed to dry. The biofilm supernatant samples were spotted onto the plate, and air-dried before incubation overnight at 37oC to observe plaque formation. Plaque formation on the parental strain (PAO1 WT) target lawn indicates the presence of the superinfective Pf4 phage.

2.2.2.4 Swimming and swarming motility assays The variants were tested for swimming and swarming motility. M9 agar plates

(47.8 mM Na2HPO4, 22 mM KH2PO4, 8.6 mM NaCl, 18.7 mM NH4Cl, 2 mM

MgSO4, 0.1 mM CaCl2, supplemented with 5.5 mM glucose (pH 6.8) (Univar Australia Pty. Ltd) with 0.3% or 0.7% w/v Bacto® agar (Difco Laboratories) for swimming and swarming motility, respectively. The overnight cultures of variants and the parental strain were inoculated in the middle of the agar plate, and

38 ! Chapter 2

incubated overnight at 37oC. The distance of swimming and swarming motility was measured as the radius of the colony.

2.2.2.5 Attachment and biofilm formation assays The attachment and biofilm formation of variants were tested using 24 well tissue culture treated polystyrene microtitre plates (Falcon). Briefly, 10 µl of overnight cultures of variants and the parental strain were cultivated in 1 ml of M9 medium supplemented with glucose. One millilitre of diluted culture was inoculated into individual wells. The plates were incubated under static conditions at 37oC for 6 and 24 h to quantify attachment and biofilm formation, respectively.

After incubation, the optical density of the cultures was measured at 600 nm using a plate reader (Wallac Victor2 1420, Perkin Elmer). The supernatants were carefully removed from each well and rinsed thrice with 1 ml phosphate-buffered saline (PBS; 7 mM NaCl, 10 mM Na2HPO4, 2.7 mM KH2PO4 and 2.7 mM KCl, pH 7.2). The wells were stained for 20 min with 0.1 % (w/v) filtered crystal violet further and were then rinsed thrice with PBS to remove excess crystal violet. One millilitre of 100% ethanol was added to each well to solubilise the crystal violet for 5 min at room temperature. The absorbance at 490 nm was measured using a plate reader (Wallac Victor2 1420, Perkin Elmer).

2.2.2.6 Confocal imagining and analysis

The biofilm continuous flow system was clamped (indicated in Figure 2-1) to stop the flow for biofilm staining. The biofilm flow cells were stained using the LIVE/DEAD BacLight viability kit (Molecular Probes Inc. Eugene. Oregon). The stock solution of the two stains (SYTO 9 and propidium iodine) was mixed with 1 ml of M9 medium to a final concentration of 2.5 nM of SYTO 9 and propidium iodide. The staining solution was injected into the flow cells while the pump was switched off, using a syringe and a 26G needle (Becton-Dickinson Precision Glide Needle) into the inlet end of the flow cell. The opening formed by the needle was sealed using silicone glue. The outlet ends of the flow cells were clamped and incubation for 20 min in the dark. Live (SYTO-9 stain) and dead (propidium iodide stain) cells were visualised using an Olympus FluoView FV1000 laser microscope (Olympus, Germany) at 40X magnification. Images from five different positions in the biofilm chamber were used as replicates. These images 39 ! Chapter 2

were used to reconstruct three dimensional images with IMARIS® software (Bitplane, Zurich, Switzerland). Each set of images consisted of a series of xy images and z stack of the biofilm for data analysis with IMARIS® software. The parameters for IMARIS® software data analysis used are 3.1 µm for surface area and 1 µm for estimated diameter of bacterial cell size, at an absolute intensity of 700.

40 ! Chapter 2

2.3 Results

2.3.1 Isolation and characterisation of phage resistant variants

To isolate variants that appear as a consequence of superinfective phage production during biofilm development, biofilms were formed and effluent was collected. The colony forming units (CFU) were monitored throughout the biofilm development to identify the time point at which dispersal occurred and to correlate dispersal with the appearance of morphotypic variants. During biofilm development, the number of CFUs in the effluent continually increased from day 1 to 7 (Figure 2-2). After day 7, the number of CFUs decreased dramatically by 2 fold and would indicate that the biofilm has dispersed at this point. The phage titre assay was used as a quantitative analysis of phage production from the biofilm and plaque formation on the agar lawns seeded with bacteria was identified as a zone of clearing (Figure 2-2). The PAO1 ΔPf4 lawn was used to detect the Pf4 phage in the biofilm effluent since it has been previously shown to be sensitive to Pf4 phage infection (Rice et al., 2009). The Pf4 phage was detected in the wild type PAO1 biofilm effluent from day 1 at 4.98 x 107 PFU/ml and increased throughout the biofilm development, reaching a maximum of 4.02 x 109 PFU/ml after 11 d. The PAO1 WT lawn was used to detect the superinfective Pf4 phage which was observed starting from day 5 of the biofilm at 9.82 x 107 PFU/ml and increased to 3.51 x 109 PFU/ml on day 7 at which point it remained similar for the remainder of the experiment. Thus, while the superinfective phage did not appear until day 5, the amounts of superinfective phage and total phage were similar from day 7 onwards suggesting that the total phage counts, based on PFUs on the Pf4 mutant lawn, were likely a reflection of the amount of superinfective phage as quantified on the WT lawn. It is interesting to note that the peak in phage production occurred on day 7, corresponding to the peak in CFUs in the effluent. This may reflect the dispersal of the biofilm on day 7, as the hollowing of microcolonies has been shown to correlate with dispersal events in the PAO1 WT biofilm and dispersal variant formation (Webb, et al., 2003).

41 ! Chapter 2

Figure 2-2. Dispersal and phage production during biofilm development by P. aeruginosa PAO1. The CFU counts (open triangles) indicate the viable cells per ml of biofilm effluent. Phage production was quantified as plaque forming units per ml of biofilm effluent on two types of bacterial lawns, PAO1 ΔPf4 (squares) and PAO1 WT (circles). Data are means of three experiments and error bars show SEM.

Biofilm variants were collected from the dispersal population of the PAO1 WT biofilm and these were represented by small colony variants (SCV) (Figure 2-3A) and spreader variants (S) (Figure 2-3B). Up to 27% of the biofilm effluent was comprised of SCVs and S variants (Figure 2-3C). One representative of each of the variants was isolated and characterised, where the SCV variant was designated as SCV2, and the spreader (S) variant was designated as S4. While the variants were observed in the biofilm from days 2 - 4, they were present at less than 1% of the population. Interestingly, on day 5, the percentage of variants increased dramatically to 27.14% and this corresponded to the appearance of the superinfective phage.

Because the appearance of the variants was strongly correlated with the appearance of the superinfective phage, it was hypothesised that the variants were phage resistant. Therefore, the sensitivity profiles of the variants were tested using the agar overlay method. Supernatant from overnight cultures of the parental strain PAO1 WT was dropped onto bacterial lawns seeded with the variants. Under these conditions, no plaques were observed (Table 2-2, Figure 2-4), 42 ! Chapter 2

indicating the variants were phage resistant. When the supernatant of overnight cultures from the variants was spotted onto the PAO1 WT lawn, the formation of plaques was observed which indicated cell lysis as a result of phage infection

Figure 2-3.! Colony morphology of biofilm-derived variants (white arrows) isolated from the PAO1 WT biofilm compared to the PAO1 WT colony morphology (red arrows). A) Small Colony Variant (SCV). B) Spreader variant (S). Scale bar 2 mm. C) The percentage of variants from the PAO1 WT biofilm effluent. Data represent the means of three experiments and error bars show SEM.!

Supernatant from the Pf4 mutant did not form plaques on any bacterial lawn as it does not produce the Pf4 phage, and the PAO1 WT only formed plaques on the Pf4 mutant lawn as expected. Effluent samples from the biofilm collected on day 2 could only form plaques on the Pf4 mutant lawn. On day 5, corresponding to late biofilm development, conversion of the Pf4 phage into the superinfective form was observed since plaques were detected on the PAO1 WT lawn. Interestingly, similar patterns of plaque formation observed for the 5 d biofilm effluent were also observed for two biofilm variants, SCV and S4, where they 43 ! Chapter 2

could infect both the PAO1 Pf4 mutant and WT lawns. The variants were resistant to phage present in their own supernatants as well as in the supernatant from the WT and showed no cross infection between the two variants. These data suggest that the variants carry a superinfective phage and were immune to their self- produced phage. Given that the variants produce superinfective phage, these variants produce the same plaque phenotypes as the 5 d biofilm effluent and they appeared in the biofilm effluent on day 5, suggests that the superinfective biofilm phenotype is due to the development of variants in the biofilm that produce the superinfective phage. Hence, there is a relationship between the biofilm-derived variants and the superinfective phenotype. There were 24.29% and 2.85% of SCV and S variants from the 5 d PAO1 biofilm effluent, and 92% and 100% of these variants expressed superinfective phenotype, respectively (Table 2-3). On the other hand, only 27.03% of the 72.86% WT-like colony from the biofilm expressed superinfective phenotype. The majority of superinfective phage variants had distinct colony morphology, and this may be a consequence of exposure to the superinfective phage.

Table 2-2. Cross infection and resistance of variants and biofilm sampled from different days of the biofilm.

Cell-free supernatant

Target ΔPf4 WT WT biofilm WT biofilm S4 SCV2 lawns effluent day 2 effluent day 5 ΔPf4 - + + + + + WT - - - + + + SCV2 ------S4 ------(+) plaque formation; (-) no plaque formation.

44 ! Chapter 2

A B

C !

!

!

Figure 2-4. Plaque formation determined using the soft agar lawn overlay method. A) The WT PAO1 bacterial lawn on an agar plate where 10 µl of sample has been spotted onto the lawns at different serial dilutions. Clear zones indicate plaque formation. Insets B) and C) represent undiluted supernatant from day 11 biofilm effluent and a dilution of the same supernatant sample with countable plaque forming units (PFU), respectively. Scale bar for B is 5 mm and 100 µm for C.

!

Table 2-3. The percentage of morphotypes from a 5 d biofilm effluent of P. aeruginosa PAO1 biofilm that have a superinfective phenotype.!

Colony Percentage of colonies Percentage of the morphotype that Morphology from 5 d PAO1 biofilm exhibits the superinfective phenotype

WT-like 72.86 % 27.03%

Small Colony 24.29 % 92% Variant (SCV)

Spreader (S) 2.85 % 100%

!

!

45 ! Chapter 2

To determine if the morphotypic variants have altered biofilm development phenotypes relative to their PAO1 parental strain, motility, attachment, growth rate and biofilm development in flow cells were quantified. The comparison of the growth rate between the WT and variants did not show significant differences (Figure 2-5). The SCV2 and S4 variants did not differ from the WT strain in swimming motility (Figure 2-6A). In contrast, the S4 variant lost the ability to swarm completely and the SCV2 had a reduced swarming motility, which was approximately 60% of the WT (Figure 2-6B). While defective in swarming motility, the SCV2 variant had up to 4 fold increased attachment (p value < 0.05), while the S4 variant did not differ from the WT (Figure 2-6C). The SCV2 variant also showed an increase in biofilm formation approximately 2 fold (p value < 0.05), while the S4 variant did not show a difference in biofilm formation compared to the WT (Figure 2-6D).

Figure 2-5. Growth curves for the PAO1 WT (triangles), PAO1 SCV2 (circles) and PAO1 S4 (squares) in M9 medium. The growth curve was constructed using a spectrophotometer to measure the optical density (at A600nm) at various time points. Data represent the means of three experiments and error bars indicate SEM.

46 ! Chapter 2

Figure 2-6. Comparison of A) swimming and B) swarming for the PAO1 WT, SCV2 and S4 variants. The results are expressed as the radius of the motility zone from the site of inoculation at the centre of the agar plate. Data are the means of three experiments and error bars indicate SEM. The C) attachment and D) biofilm formation of PAO1 WT, SCV2 and S4 variants. The crystal violet absorbance value (A490nm) was normalised with the culture density (A600nm) for attachment. Data are the means of three experiments and error bars indicate SEM. (*) indicates a statistically significant difference when compared to PAO1 WT at 95% confidence interval with One-way ANOVA.

47 ! Chapter 2

2.3.2 Confocal microscopy of variant biofilms

The biofilm development of PAO1 WT was compared with biofilms formed by variants carrying the superinfective Pf4 phage in flow cell biofilm chambers. Analysis of confocal images of live and dead stained PAO1 WT biofilms showed that microcolony formation occurred on days 3 and 5, followed by cell death within the microcolonies on day 7 (Figure 2-7). In contrast, the variant biofilms formed different biofilm structures and with an earlier event of cell death occurring on day 3 of the variant biofilms. The SCV2 variant biofilm had zones of dead cells in the biofilm on day 3 and aggregation between cells and microcolony formation was not observed until day 7. There were fewer clusters of cells and the diameter of microcolonies from the SCV2 biofilm were half the size in width and height compared to microcolonies formed by the PAO1 WT biofilm (approximately 120 µm). In contrast, biofilms formed by the S4 variant had smaller and irregularly shaped microcolonies on day 3 compared to the WT biofilm. The variant biofilms were less dense and consisted of dead cells in the biofilm from day 3 onwards compared to the WT biofilm where dead cells were primarily only observed on day 7.

The differences between the PAO1 WT biofilm and the variant biofilms were also evident from the quantitative analysis of the biofilm biovolumes. The total volume of biomass in the biofilms (live and dead cells) indicated that there was no difference between the biofilms formed by the three strains (Figure 2-8A). However, the number of dead cells (Figure 2-8B), increased from day 3 and peaked at day 5 for variant biofilms, followed by a gradual decrease as the variant biofilms start to develop. The PAO1 WT biofilm showed a steady increase in number of dead cells during the biofilm lifecycle where it reaches the stage of cell death within microcolonies. Furthermore, the PAO1 WT biofilm was double the height at an average of 45.90 µm from day 7, whilst the average height of PAO1 SCV2 and S4 biofilms were 24.48 µm and 23.97 µm, respectively (Figure 2-8C).

48 ! Chapter 2

Figure 2-7. Confocal images of PAO1 WT, PAO1 S4, and PAO1 SCV2 biofilm development on days 3, 5, and 7 using an Olympus FV1000 confocal laser microscope. The scale bar represents 30 µm. Biofilms were stained with the BacLight LIVE/DEAD stain (Molecular Probes Inc. Eugene. Oregon) where green represents live cells, red represents dead cells, and yellow represents an overlap of green and red cells. !

49 !

Chapter 2

A B C

Figure 2-8. Quantitative comparison of biofilms formed by the WT PAO1 and the biofilm-derived variants S4 and SCV2. A) Total biofilm biovolume, B) Total number of dead cells, and C) total biofilm height during biofilm development of PAO1 WT (triangle), PAO1 S4 (square), and PAO1 SCV2 (circle) on days 3, 4, 5, 6 and 7. IMARIS® software was used to analyse the three dimensional biofilm images which were reconstructed from the x-y image stacks taken using the Olympus FV1000 confocal laser microscope. The parameters used for the IMARIS® biovolume analysis were 3.1 µm for surface area and 1 µm for estimated diameter of bacterial cell size, at an absolute intensity of 700. Data represent the means of three experiments and error bars indicate SEM. (*) indicates that both variants were statistically significant difference when compared to PAO1 WT at 95% confidence interval with Two-way ANOVA with a Tukey’s post test.

50 ! Chapter 2

Morphotypic variants are commonly observed during biofilm formation. For example, dispersal of the PAO1 WT biofilm occurred on day 5 (Figure 2-2), and up to 25% of the morphological variants from the dispersal population were SCVs (Figure 2-9). Therefore, the SCV2 biofilm was sampled for CFU counts as well as for variant formation. Interestingly, biofilm effluent collected from the SCV2 biofilm had colonies that no longer exhibited the small colony morphology. From day 2 onwards, some variants detected from the biofilm had reverted to the WT- like colony morphology and the overall frequency of WT morphotypes was on average 22.69% on day 2 and increased to 30.49% on day 6 (Figure 2-10). When these revertant variants were tested for their superinfective phenotypes using the agar overlay method, many of these isolates no longer formed plaques on the WT lawn, indicating that they had lost the superinfective phage phenotype. Similarly, these WT appearing isolates were subsequently sensitive to infection by the phage produced by the SCV2 isolate. Similarly, morphotypic variants were also observed from the S4 variant biofilm consisting of WT-like colony morphology and SCVs although they were not characterised further here (data not shown).

Figure 2-9. The percentage of SCV colonies isolated from the effluent of the PAO1 WT biofilm. The biofilm was sampled from days 2 - 7, and biofilm effluent was collected for CFU counts. Data are the means of three experiments and error bars indicate SEM.

!

51 ! Chapter 2

!

Figure 2-10. The percentage of WT-like colonies isolated from the biofilm effluent of the PAO1 SCV2 variant biofilm. The biofilm was sampled from days 2 to 7, and biofilm effluent was collected for CFU counts. Data are the means of three experiments and error bars indicate SEM.!

52 ! Chapter 2

2.4 Discussion

As the biofilm cells experience stresses in their environment, this can result in a selection for genetic and phenotypic changes known as variants. In some cases, the stress may also directly induce the underlying mutation by causing DNA damage, as is the case for oxidative stress. Within the biofilm it has been shown that reactive oxygen and nitrogen species (RONS) accumulate in the P. aeruginosa microcolonies, as a consequence of anaerobic respiration and this has been linked to the formation of variants (Barraud, et al., 2006). Variant formation is a common trait during biofilm development, as a result of genetic changes in the biofilm community. During phage infection within the PAO1 biofilm, microcolonies undergo cell death followed by dispersal events. The appearance of SCVs coincides with the dispersal event in the biofilm. Interestingly, during this phase, the Pf4 lysogenic phage has been observed to convert into the superinfective form of the phage. The variants that are resistant to the phage superinfection have been isolated in this study, which may in part explain why the variants appear in high proportions at the time of superinfection. This may suggest that the SCVs from the dispersal population are the result of exposure to phage superinfection and subsequent lysis of the sensitive host. Therefore, the aim of experiments presented in this chapter was to isolate and characterise variants to better understand their influence on biofilm development traits.

It has been reported that mucoid and sticky variants, a consequence of over- production of alginate, are commonly isolated from biofilms formed in cystic fibrosis patient’s lungs and that such mucoid variants are more resistant to oxidative stress than the WT parental strains. In contrast, SCVs, wrinkly and rugose variants are more frequently isolated from laboratory biofilms (Haussler, et al., 1999, Haussler, et al., 2003, Starkey, et al., 2009). This may indicate that there is different selective pressure in different biofilm environments. In this study, small colony variants were the most commonly observed variants and these were accompanied at a low frequency by spreader colony variants.

The variants have undergone genotypic and phenotypic changes during the biofilm development that has resulted in mutations that lead to the establishment of the superinfective Pf4 phage. The superinfection event in the PAO1 biofilm 53 ! Chapter 2

may select for variants that carry the superinfective Pf4 phage allowing the variants to be resistant against superinfection. Variant formation within the biofilm generates subpopulations of variants that exhibit biofilm specific phenotypes. However, it is unclear whether these morphotypic variants dominate and persist within the community. There were approximately 25% of morphotypic variants from the effluent of the PAO1 biofilm and up to 92% of these morphotypic variants showed superinfective phage and resistance against superinfection. These results suggest that the variants carrying superinfective phage are selected for in the biofilm. This is supported by the observation of superinfective phage added to a planktonic WT culture resulted in SCV formation, hence phage infection selects for variant formation in P. aeruginosa (Rice, et al., 2009). The majority of the colonies isolated from the PAO1 biofilm effluent were WT-like colony morphology with 27% of these variants carrying the superinfective phage resulting in resistance against superinfection. This means that only a subpopulation of the variants undergo morphological changes and colony morphology is not an accurate indication of superinfective phage variants but can be a good indication, at least in the case of SCVs, that the variants carry the superinfective phenotype. The WT-like variants that carry the superinfective phage and are resistant against infection may harbour other mutations that lead to changes in surface receptors or that produce inhibitors against phage infection that are yet to be identified. For example, the lack of a specific lipopolysaccharide has been shown to be responsible for phage resistance (Mutoh, et al., 1978). Additionally, a lack of tetragonally arrayed T-layer protein that make up the cell wall of Bacillus sphaericus P-1 has also been shown to prevent phage absorption (Howard & Tipper, 1973).

Bacteriophage are the most abundant organisms on this planet and outnumber bacteria in most environments (Brussow & Hendrix, 2002). Therefore, phage resistance is a crucial survival phenotype for bacteria. Phage infect bacteria through cell surface receptors as absorption sites, and changes in surface receptors are commonly detected in phage resistant hosts. There are several mechanisms for the host to alter surface receptors to prevent phage infection such as blocking phage receptors and competition for binding by inhibitors. For example, E. coli

54 ! Chapter 2

phage T5 produces a lipoprotein that blocks the ferrichrome-iron phage receptors during infection to prevent superinfection against the host (Pedruzzo, et al., 1998). Competition for binding by a microcin J25 molecule produced by E. coli, can outcompete the phage T5 for binding at the iron transporter FhuA of the cell, which is also the point of entry for the phage (Destoumieux-Garzon, et al., 2005). Furthermore, mutations that result in changes in receptor binding or loss of receptor function, can also lead to phage resistance as the point of infection has been compromised. The type IV pili were shown to act as a receptor for Pf4 phage (Castang & Dove, 2012), therefore this may suggest a role in phage resistance. Conformational changes in binding site or change in pilus-related gene expression may influence phage resistance of the host. It is important to identify the role of type IV pili in phage infection and resistance, thus its role in lysogenic and lytic phage cycle.

The host defense system consists of genes that are responsible for phage resistance. The CRISPR-cas system has recently been extensively studied as it confers phage resistance against phage infection and is found in a large range of bacterial and Archael genomes (Godde & Bickerton, 2006, Sorek, et al., 2008). The CRISPR-cas system is an immunity system of the host that targets foreign nucleic acids. In P. aeruginosa, the CRISPR-cas system was shown to be involved in lysogeny-dependent inhibition of biofilm formation and swarming motility (Zegans, et al., 2009). However, CRISPR loci have not been identified in the P. aeruginosa PAO1 strain used here.

Another system also known for prevention of phage infection is the superinfection exclusion systems. For example, in coliphage T4, phage-encoded proteins prevent penetration of the phage through the cell wall and change the conformation of the phage DNA injection site (Lu & Henning, 1994). In addition to immunity systems of the host, there are other genes that can also influence phage resistance such as mutations in flagella and type IV pili for phage infection. In Caulobacter crescentus, the phage infects the cell with the aid of flagellum which facilitates the accumulation of phage particles at the cell surface receptors thereby increasing the likelihood of infection (Guerrero-Ferreira, et al., 2011). In Ff phage of E. coli, the F-pilus serves as the primary phage receptor and the TolA protein as the

55 ! Chapter 2

secondary receptor. The SCV2 and S4 variants are both immune to superinfection therefore they may harbor mutations within the host genome that confers immunity against superinfection. While not tested here, previous data indicated that the double mutant of PilA (type IV pili subunit) and FilM (flagella), of P. aeruginosa PAO1, were resistant to superinfection (Webb, et al., 2003). Therefore, it is possible that these variants carry mutations that prevent type IV pili production and/or flagella production. However, it was noted that revertants that regained the WT morphology, lost the superinfective phenotype and could subsequently be infected with superinfective phage. If the initial resistance of this variant was due to altered phage receptors, then this must be reversible to subsequently allow reinfection.

The variants also showed adaptation in other biofilm relevant phenotypes. The SCV2 variant had reduced swarming ability and a substantial increase in attachment capability. These changes contribute to the ability to colonise surfaces, an essential role in biofilm formation, and suggests that SCV2 variant is a better coloniser than PAO1 WT. It was suggested that the copious interwoven phage filaments that surround the SCV cells may contribute to the enhanced attachment of the SCV variants (Webb, et al., 2004). Similar biofilm phenotypes were observed from the SCV variant of P. aeruginosa 57RP and CF isolates that have been shown to exhibit strongly adherent biofilms, defective swarming motility, increased expression of intracellular signalling molecules and polysaccharide production (Deziel, et al., 2001, Haussler, et al., 2003, Kirisits, et al., 2005, Starkey, et al., 2009). Subsequently, the SCV variant was selected for under biofilm conditions. Similarly, the S4 variant lost the ability to swarm. The EPS matrix of the biofilm allows cells to require less motility as they are protected within the matrix and exist as loosely associated cells (Watnick & Kolter, 2000).

The biofilm development of variants was compared with the PAO1 WT and differences were evident at the various stages of biofilm development. Variant biofilms showed a different pattern of biofilm development relative to the WT with large volumes of dead cells that emerged early during biofilm formation. This observation of early cell death on day 3 may be explained by the existence of superinfective phage resulting in selective pressure for cells to form biofilm and

56 ! Chapter 2

as a trade off, lose the resistance phenotype for biofilm formation. The population of cells that survive the cell death events on days 3 to 5, proliferate and form the mature biofilm. The PAO1 SCV2 biofilm formed smaller microcolonies (approximately half the size of the WT microcolonies) whilst there were no microcolonies formed from the S4 variant biofilm. P. aeruginosa mutants defective in type IV pili and flagella motility have previously been shown to play a role in biofilm structure and microcolony formation (O'Toole & Kolter, 1998). Mutants defective in type IV pili did not develop microcolonies and mutants defective in flagella were associated with attachment and initial cell-to-surface interactions. The reduction or loss of swarming ability in variants may explain the effect on microcolony formation. There are many genes that have been associated with defective swarming ability in P. aeruginosa (Overhage, et al., 2007), and mutations driven by stress in the biofilm environment can lead to the loss of motility. Hence, the loss of microcolony formation may also be reflected in the observed changes in motility.

The results presented in this chapter suggest that superinfective phage is selected for in the biofilm and variant formation is a consequence of this selection. In the SCV2 biofilm, WT-like colony morphotypes were detected in the biofilm effluent. The percentage of WT-like variants from the SCV2 biofilm increased during biofilm development and when tested for superinfectivity, they no longer exhibited this property. Thus, these WT-like variants have reverted back to the PAO1 WT colony morphology and lost the ability to superinfect the PAO1 WT host. Sequencing of superinfective variants and non-superinfective variants may suggest genes that are responsible for the superinfective phenotype (refer to Chapter 3).

The superinfective variants were stable upon passage on plates and in liquid culture indicating that they were not the consequence of transient changes in gene expression and are therefore more likely due to genetic mutations. The SCV2 variant cultures were passaged over four generations and the morphology of the variants remain as small colonies. On the other hand, revertants from the SCV2 biofilm resembled WT-like colonies therefore these data suggest that revertants are biofilm-related. Mutations as a result of nutrient limitation and starvation in

57 ! Chapter 2

the biofilm can also give rise to changes in the phage genome, resulting in activation of lethal phage genes and leading to death of the host. Mutations in most of the phage genes of f1 or fd filamentous phage, except for gene II, which encodes a replication initiation protein, can lead to cell lysis and death of the host (Hohn, et al., 1971, Woolford, et al., 1974, Horabin & Websster, 1986). These phage genes mainly consist of attachment, assembly, immunity and coat protein genes. This observation was also noted in other phage including X174, MS2, QB and lambda phage. Membrane proteins of the phage can interact with host proteins to promote cellular autolysin and activate mechanisms that result in cell wall damage (Harris, et al., 1967, Young, 1992, Bernhardt, et al., 2002, Young & N., 2006). Immunity genes such as the repressor c gene responsible for suppressing lytic phage genes, can cause loss of host immunity against superinfection (Salmon, et al., 2000, Lynch, et al., 2010). In order to determine the cause of the conversion from lysogenic to lytic form of the Pf4 phage, genetic analysis of the variants can give an insight to the mechanism of conversion in phage life cycle.

Variants that survive superinfective phage infection in the biofilm were shown to carry the superinfective phage and were resistant against the infection. Therefore, superinfective phage is likely selected for in the biofilm and as a result induces variant formation. These variants are altered in their biofilm development as compared to the PAO1 WT, and display biofilm specific phenotypes including higher adherence to surfaces and lack swarming motility. In this chapter, the phenotypic traits of variants that carry superinfective phage have been isolated and characterised. Given that the formation of biofilm-derived variants has been linked to increased biofilm resilience under stress conditions and it was shown here that superinfection drives variant formation, it is important to understand the mechanism leading to the conversion into the superinfective phage during biofilm development. Further, the genetic changes may also contribute to the phenotypes observed from the biofilm. Genome sequencing may identify important genes that play a role in the conversion into the superinfective phage or identify genes that undergo genetic changes during biofilm development of P. aeruginosa PAO1.

58 ! Chapter 3

Chapter 3 – Mutations from superinfective variants

3.1 Introduction

When bacteriophage were first discovered, they were viewed as parasites of bacteria that largely mediated bacterial lysis, although some phage were developed as genetic tools for exploring genetic regulation in bacteria, e.g. as cloning and expression vectors. However, as prophage have increasingly been identified within bacterial genomes, this view of bacteriophage has changed and it is now appreciated that bacteriophage contribute a number of advantages to the host as well as playing an important role in the diversification of bacterial species. For example, it was demonstrated in Salmonella that a prophage-encoded superoxide dismutase gene contributed to the bacterial host defense against oxidative stress (Farrant, et al., 1997). Moreover, bacteriophage are versatile carriers of genetic information, including virulence factors, and as such are important vectors in Horizontal Gene Transfer (HGT) (Boyd & Brussow, 2002). In V. cholerae, E. coli O157 and Clostridium diphtheriae, prophage have been shown to encode disease-related genes such as toxins that contribute to the virulence of the bacterium (Pappenheimer, 1977, Cook, et al., 1984, Huang, et al., 1987, Nakayama, et al., 1999, Wagner, et al., 2001). Thus, it has been postulated that carrying a prophage has contributed to the fitness of bacteria (Brussow & Hendrix, 2002). This may be reflected in the observation that bacterial genomes carry on average 2.6 prophage-like elements (Canchaya, et al., 2003, Canchaya, et al., 2004).

As a consequence of strong selective pressure, bacteria have evolved several mechanisms to avoid phage infection. There are several mechanisms of resistance to phage infection, such as changes to cell surface receptors, protein inhibitors and triggers of cell lysis. It is important to understand the role of phage adaptation in bacterial evolution and how it contributes to the host range expansion. Lysogens that carry bacteriophage possess immunity against phage infection by carrying the repressor protein synthesised by the prophage. The loss of the repressor protein in

59 ! Chapter 3

the immunity region of the prophage genome can result in cell death of the host. It has been reported that mutations in the immunity region of the prophage genome are also responsible for superinfection in bacteria. In Xanthomonas campestris pv. citri, bacteriophage cf carries a mutation within the immunity region of the phage that results in superinfection against lysogenic host strain (Cheng, et al., 1999). In Lactobacillus, prophage A2 carries a repressor gene that has DNA binding activity and cleavage domains which mimic the repressor C in lambda phage and has been identified to play a role in host immunity against superinfection (Ladero, et al., 1998). Similarly, deletions in the immunity region of Salmonella phage P22 have also been shown to be responsible for host immunity against superinfection with loss of either immI or immC gene resulting in superinfection. Furthermore, the mutations in the host gene (pro) were accompanied with the loss of a immC gene in the phage genome (Chan & Botstein, 1972). However, it has been demonstrated in E. coli K12, another, as yet unidentified, component besides the repressor gene of prophage P1 is required for immunity against superinfection by phage P1 (Scott, 1975, West & Scott, 1977). This suggests that there can be more than one mechanism of superinfection or immunity defense of the host against infection.

The results shown in Chapter 2 demonstrated that a large percentage of the PAO1 biofilm-derived morphotypic variants constitutively produced superinfective phage that would form plaques on the PAO1 parent, but that were resistant to reinfection by their self-produced phage. Interestingly, revertant variants of the SCV2 variant biofilm showed the loss of superinfective phage and became sensitive to superinfection. Similar morphotypic variants have been isolated from other bacterial hosts, such as Cf1tv filamentous phage carrying infectious phage derived from host community that evolved and adapted the resistance gene (Kuo, et al., 1994), suggesting that this is a common phenomenon amongst bacteria with prophage. Collectively, these data may suggest that a switch between lysogenic and lytic phage exists in the Pf4 phage, however this process may differ substantially from the lambda phage system.

60 ! Chapter 3

The aim of this chapter was to identify mutations within the prophage genomes of the biofilm variants isolated from PAO1 WT (Chapter 2). PAO1 SCV2 and S4 variants have both been identified to confer resistance against superinfection because they carry the superinfective form of the phage. Therefore, the prophage genomes of the variants were sequenced to identify differences in those constitutive superinfective phage from the phage of the PAO1 parent strain. Phage genome sequencing demonstrated that there either of the S4 or SCV2 variants only carried nonsynonymous substitutions between PA0716 and PA0717 gene and had large deletion in the intergenic region between PA0716 and PA0717 in the Pf4 genome, respectively.

Analysis of this region indicated that it contained an open reading frame that shared 42% homology with the repressor C protein of phage P2 in the intergenic region between PA0716 and PA0717 genes encoded on the complementary strand (Webb, et al., 2004). Furthermore, deep sequencing of the biofilm dispersal population showed similar results, where there were limited or no mutations in the phage genome except in the repressor c gene or in the putative promoter for this open reading frame. This would suggest that mutations of repressor c, resulting in conformational changes in the protein, alter its function as an immunity protein and thus allow for the mutant phage to superinfect a host carrying a related prophage. Alternatively, mutation of the putative promoter alters the binding ability of the repressor C protein to the binding site of the promoter, thus leads to loss of immunity protein function and allows for phage to superinfect the host.

61 ! Chapter 3

3.2 Materials and Methods

3.2.1 Bacterial strains and culture conditions

The P. aeruginosa strains used in this study are listed in Table 3-1. All P. aeruginosa strains were cultured as described in section 2.2.1 in Chapter 2.

Table 3-1. The P. aeruginosa strains used in this study.

Strain Reference

PAO1 WT Laboratory stock PAO1 ΔPf4 Rice, et al., 2009 PAO1 SCV2 This study

PAO1 ΔPA0716 Jacobs, et al., 2003 (Jacobs, et al., 2003) PAO1 ΔPA0717 Jacobs, et al., 2003 PAO1 ΔPA0718 Jacobs, et al., 2003 PAO1 ΔPA0719 Jacobs, et al., 2003 PAO1 ΔPA0720 Jacobs, et al., 2003 PAO1 ΔPA0721 Jacobs, et al., 2003 PAO1 ΔPA0722 Jacobs, et al., 2003 PAO1 ΔPA0723 Jacobs, et al., 2003 PAO1 ΔPA0724 Jacobs, et al., 2003 PAO1 ΔPA0725 Jacobs, et al., 2003 PAO1 ΔPA0726 Jacobs, et al., 2003 PAO1 ΔPA0727 Jacobs, et al., 2003 PAO1 ΔPA0728 Jacobs, et al., 2003 PAO1 ΔPA0729 Jacobs, et al., 2003

!

62 ! Chapter 3

3.2.2 Superinfection and resistance screening

3.2.2.1 Phage assay The phage assay was performed using a modified version of the top-layer agar method previously described (Eisenstark, 1967). Briefly, the overnight cultures of the knockout mutants was centrifuged at 13,000 x g for 5 min and filtered through a 0.22 µm filter (Millipore Millex GP) to obtain cell-free supernatant. Ten microlitre drops of the supernatant were spotted onto LB10 agar plates containing an overlay of the bacterial lawn with PAO1 ΔPf4 or PAO1 WT.

The overlay of the bacterial lawn was prepared by mixing 500 µl of an overnight culture, grown in M9 complete medium, with 5 ml of 0.8% (w/v) molten LB10 agar, which had been cooled to 55oC in a water bath. The mixture was poured onto a LB10 agar and was allowed to dry. The biofilm supernatant samples were spotted onto the plate, and air-dried before incubation overnight at 37oC to observe plaque formation.

!

3.2.3 Molecular characterisation of superinfective variant PAO1 SCV2 and S4

3.2.3.1 Primers for genome walking

The oligonucleotides used to sequence the Pf4 genome of P. aeruginosa superinfective variants SCV2 and S4 are listed in Table 3-2.

63 ! Chapter 3

Table 3-2. List of oligonucleotides used in this study.

Primer Name Sequence (5’ – 3’) Reference

16S 27F AGAGTTTGATCMTGGCTCAG Weisburg et al.,

1991(Weisburg, et al., 1991) 16S 1497R ACGGTTACCTTGTTACGACTT Weisburg et al.,

1991(Weisburg, et al., 1991) rpCorfF ATGAGCACGTCAGCCGATAGAGC This study rpCorfR CTATCCCGCGTTTTGATTGGACA This study rpC2F CTCGGGGCAGCATGTGTTTGTC This study rpC2R GGGGCTAAGCTCTTCCAGTTCC This study rpC3F AGCTTCATTTCTTGCCTTCCATCC This study rpC3R GCTCCGCCCACCGTTCG This study coaAF ACCCGGCGAAAGAGAACTGC (Tan, 2006) coaAR CGAGGTTGATGATTTCCGCCG (Tan, 2006)

SCV2 walk GATCGACGTTGGCCTTCACCTT This study coaA1 SCV2 walk GGGATTGCCGCCATACTCACAG This study coaA2 SCV2 walk ACTTGGCGATGCGTTGCTTCA This study coaA3 SCV2 walk CAGCGGTGTTGGGTGAAAGTGAT This study coaA4 SCV2 walk CCAGCGGCCACCTTCCAGA This study coaA5 phdF TAAACCATGGTAATGCGAGTCGAG (Tan, 2006) ACAATTAGT phdR TATGTTTAAACTTATTCTGGCTGAG (Tan, 2006) CGAACCT SCV2 walk AGCAGCACCAGGCGGAACAC This study phd1 SCV2 walk CGCTTGCCCTTGATCGGTTCTA This study phd2 SCV2 walk CCGAGCCTGGAGGGTCTTGTTAT This study phd3 SCV2 walk CGCCGTCATCACCGTCCAC This study phd4 SCV2 walk CGCCTTGGTGTCCTGGGTCTT This study phd5

64 ! Chapter 3

3.2.3.2 Genomic DNA extraction

Genomic DNA was extracted using the XS buffer extraction method (Tillett & Neilan, 2000) with slight modifications. Briefly, 2 ml of overnight bacterial culture was resuspended with XS buffer containing 1% potassium ethyl xanthogenate (Fluka, Buchs), 100 mM Tris-HCl (pH 7.4), 20 mM EDTA (pH o 8.0), 1% SDS, and 800 mM NH4C2H3O2, and incubated at 65 C for 2 h. The samples were vortexed and incubated on ice for 30 min. Cell debris was removed by centrifugation at 14,000 x g for 10 min. The supernatant was removed and equal volume of isopropanol was added to precipitate the genomic DNA. The genomic DNA was subsequently washed with 70% (v/v) ethanol and the pellet was air-dried. The genomic DNA was resuspended in 100 µl of sterile, nuclease- free water (Millipore). DNA concentration and purity was determine using a NanoDrop® ND-100 spectrophotometer (NanoDrop Technologies) at 260 nm and 280 nm.

3.2.3.3 Phage DNA extraction

Phage DNA was extracted from 10 ml overnight culture of PAO1 SCV2, using the QIAGEN® lambda extraction kit (QIAGEN Pty Ltd.), and performed according to manufacturer’s instructions. Briefly, 10 ml of overnight culture was centrifuged at 13,000 x g for 10 min to remove bacterial cell debris and the supernatant was retained. Thirty microlitres of L1 buffer containing RNase A and DNase I were added to the supernatant and incubated for 30 min at 37oC to digest extracellular RNA and DNA without damaging the protein coated, single stranded genome of the Pf4 phage particles. The samples were filtered using a 0.22 µm filter (Millipore Millex GP) and 2 ml of ice cold L2 buffer was added and incubated on ice overnight to precipitate the phage particles. The samples were then centrifuged at 13,000 x g for 10 min and the supernatant was removed. One millilitre of L3 buffer was used to resuspend the precipitated phage particles. The phage proteins were denatured to release the phage DNA by adding 1 ml of L4 buffer. The mixture was gently mixed, and incubated for 10 min at 70oC, and subsequently chilled on ice. One millilitre of L5 buffer was added and mixed by inverting the tubes several times. The precipitated phage proteins were removed by centrifugation at 15,000 x g for 30 min at 4oC. The supernatant containing the 65 ! Chapter 3

phage DNA was transferred into a clean tube and was centrifuged again at 15,000 x g for 10 min at 4oC to ensure that the precipitated phage proteins were completely removed.

After the crude phage DNA was obtained, anion-exchange chromatography (using the QIAGEN-tip 20) was performed to purify the phage DNA. The QIAGEN-tip 20 was equilibrated with 1 ml of QBT buffer prior to the addition of the crude phage DNA sample. Crude phage DNA samples were added to the QIAGEN-tip 20 column and the bound phage DNA was washed with 2 ml of QC buffer. The phage DNA was eluted with 1.5 ml of QF buffer and collected in a sterile tube. The phage DNA was concentrated by adding 1 ml of isopropanol and precipitated DNA was centrifuged at 15,000 x g for 30 min at 4oC. The supernatant was removed and the pellet containing the precipitated phage DNA was washed with 70% (v/v) ethanol and then centrifuged at 15,000 x g for 10 min. The pellet was air-dried for 15 min and the DNA was resuspended in 50 µl of sterile nuclease- free water (Millipore). The concentration and purity of the phage DNA was determined using a NanoDrop® ND-100 spectrophotometer (NanoDrop Technologies) at 260 nm and 280 nm.

3.2.3.4 Amplification for PAO1 SCV2 Pf4 phage genome

The purity of the phage DNA was first determined using with Nanodrop spectrophotometer as described above, followed by PCR amplification with 16S rRNA gene (16S 27F and 16S 1497R) primers to control for genomic DNA contamination. Approximately 50 ng of template DNA was used in a 20 µl PCR reaction containing 2.5 mM MgCl2, 1x AmpliTaq PCR buffer, 20 pmol of each primer pair, 0.4 µM dNTPs and 1 U of AmpliTaq Polymerase (Applied Biosystems). The PCR cycling conditions for the 16S primer pair were: 25 cycles of denaturation at 98oC for 30 s, annealing at 55oC for 30 s, and elongation at 72oC for 2 min, with a final extension step at 72oC for 10 min. The PCR products were analysed by gel electrophoresis on a 1% (w/v) agarose gel.

PCR amplification of the Pf4 phage genomes of the PAO1 SCV2 and PAO1 S4 superinfective variants was performed using the primer sets shown in Table 3-2. Approximately 100 ng of template DNA was used in a 20 µl PCR reaction

66 ! Chapter 3

containing 2.5 mM MgCl2, 1x AmpliTaq PCR buffer, 20 pmol of each primer pair, 0.4 µM dNTPs and 1 U of AmpliTaq Polymerase (Applied Biosystems). The PCR cycling conditions for phage genome primer pairs (rpCorf, rpC2, rpC3, phd, coaA, walk coaA 1-5, walk phd 1-5) were: 25 cycles of denaturation at 98oC for 30 s, annealing at 55oC for 30 s, and elongation at 72oC for 1 min, with a final extension step at 72oC for 10 min. The PCR products were analysed by gel electrophoresis on a 1% (w/v) agarose gel. The PCR products were purified using DNA Clean & ConcentratorTM -5 Kit (Zymo Research) according to manufacturer’s instructions. The concentration of purified PCR products was determined using a Nanodrop® ND-100 spectrophotometer (NanoDrop Technologies) at 260 nm and 280 nm.

Sequencing was performed using BigDye TerminatorTM sequencing mix. Briefly, a 20 µl sequencing reaction mixture containing 100 ng of phage or genomic DNA, 3.2 pmol of primer, 1 x BigDye TerminatorTM buffer and the BigDye TerminatorTM (version 3.1 Applied Biosystems), was prepared. The sequencing cycling conditions were: 1 cycle of denaturation at 96oC for 1 min followed by 25 cycles of 96oC for 10 s, 50oC for 5 s, and 60oC for 4 min. The sequencing products were purified using the butan-1-ol purification protocol (Argyropoulos, et al., 1999). Briefly, the sequencing reaction mixture was adjusted to 100 µl with sterile nuclease-free water (Millipore). The mixture was transferred into a 0.5 ml tube containing 100 µl of phenol (pH 8.0), vortexed for 10 s and centrifuged at 17,900 x g for 4 min. After centrifugation, the top-layer aqueous phase was transferred into a 1.5 ml tube containing 900 µl of butan-1-ol. The mixture was vortexed for 10 s and centrifuged at 17,900 x g for 10 min. The supernatant was discarded and the pellet was air-dried for at least 15 min (in the dark). The air- dried purified sequencing products were analysed by the Ramaciotti Centre for Gene Function Analysis, UNSW.

3.2.4 Biofilm experiments

3.2.4.1 Continuous-culture biofilm Biofilms were set up as described in section 2.2.2.1 in Chapter 2, with slight modification. The biofilms were grown in silicon tubing (inner diameter 2.64 x

67 ! Chapter 3

0.28 mm and outer diameter 4.88 x 0.28 mm) (Silastic® laboratory tubing) in replacement of the flow cell attachment as shown in Figure 2-1. Prior to inoculation, the flow cell system was flushed thoroughly with M9 minimal medium for 1 h. Inoculation of the flow cells was performed while the pump was switched off and the inlet end of the flow cell was clamped to prevent back flow of the inoculum towards the medium end. Two millilitres of overnight culture were injected using a syringe and a 26G needle (Becton-Dickinson Precision Glide Needle), into the inlet end of the tubing. The opening formed by the needle was sealed using the silicone glue (Plastic Putty Selleys Pty Ltd.). The outlet ends of the tubing were clamped and left for 1 h without flow to allow the bacterial culture to attach to the surface of the tubing. Medium flow was resumed at a flow rate of 6 ml/h. All biofilm tubing experiments were performed at room temperature with three replicates.

3.2.4.2 Genomic DNA extraction from biofilm effluent On day 4 of biofilm development, 20 cm of tubing containing the biofilm cells was harvested and flushed with 2 ml of M9 medium without glucose to retrieve the biofilm cells for genomic DNA extraction using the GenElute Bacterial Genomic Kit (Sigma). On day 11, biofilm effluent was collected over a 12 h period and simultaneously, M9 medium supplemented with 20 mM NaN3 was added to the biofilm effluent collected at a rate of 9 ml/h. Briefly, the biofilm cells and biofilm effluent collected were centrifuged at 9,000 x g to obtain the cell pellet. The pellet was resuspended in 180 µl of lysis solution T and 20 µl RNase A solution was added and incubated at room temperature for 2 min. Twenty microlitres of Proteinase K solution was added to the mixture and incubated for 30 min at 55oC. After incubation, 200 µl of lysis solution C was added and vortex briefly for 15 s. The mixture was incubated for 10 min at 55oC.

The GenElute Miniprep Binding Column was prepared by the addition of 500 µl of Column Preparation Solution to the column placed in a 2 ml collection tube. The column was centrifuged at 12,000 x g for 1 min and the eluate was discarded. Two hundred microlitres of 100% ethanol was added to the mixture and vortexed for 10 sec. The mixture was transferred into the column and centrifuged at 9,000 x g for 1 min and the eluate was discarded. Five hundred microlitres of Wash

68 ! Chapter 3

Solution 1 was added to the column and centrifuge at 9,000 x g for 1 min and the eluate collected was again discarded. A second wash of 500 µl of Wash Solution was added to the column and centrifuge at 9,000 x g for 3 min and the eluate was discarded. The column was centrifuged at 9,000 x g for an additional 1 min and the eluate was discarded. Genomic DNA was eluted by adding 200 µl of Elution Solution to the column and incubated for 5 min at room temperature. The column was centrifuged at 9,000 x g for 1 minute. A second elution was eluted by adding 200 µl of Elution Solution to the column and incubated for 5 min at room temperature, followed by centrifugation at 9,000 x g for 1 min. The concentration and purity of the genomic DNA was determined using a Nanodrop® ND-100 spectrophotometer (NanoDrop Technologies) at 260 nm and 280 nm.

3.2.4.3 Phage DNA extraction from biofilm effluent The biofilm effluent collected was centrifuged to obtain the supernatant to extract phage DNA from the biofilm effluent. Phage DNA was extracted as described in section 3.2.2.1. The concentration and purity of the phage DNA was determined using a Nanodrop® ND-100 spectrophotometer (NanoDrop Technologies) at 260 nm and 280 nm.

3.2.4.4 Determination of Colony Forming Units (CFU) To identify when dispersal occurred, colony forming units were determined by collecting 5 ml of biofilm effluent at 12 h intervals from day 4 onwards. Serial dilutions with M9 medium (without glucose supplement) were performed and 100 µl of diluted samples were spread plated onto LB10 agar. The plates were incubated overnight for at least 12 h at 37oC. At the same time, 500 ml of biofilm effluent was collected and M9 medium supplemented with 20 mM NaN3 was added simultaneously at a rate of 6 ml/h. These samples were centrifuged at 9,000 g for 30 min and cell pellets were stored at -80oC for subsequent genomic DNA extraction for deep sequencing of the dispersal population.

3.2.4.5 Determination of Plaque Forming Units (PFU) – Phage assay The plaque forming units were determined as described in section 2.2.2.3 in Chapter 2. Briefly, the cell-free supernatant biofilm effluent was serially diluted and spotted onto LB10 agar plates containing an overlay of the bacterial lawn

69 ! Chapter 3

with PAO1 ΔPf4 or PAO1 WT. Plaque formation on PAO1 WT lawn represents the presence of superinfective phage.

3.2.4.6 Deep sequencing analysis The deep sequencing analysis was done in collaboration with McElroy K. and Thomas, T. (UNSW). The genomic and phage DNA samples extracted from the PAO1 WT biofilm, were sequenced by the Ramaciotti Centre for Gene Function Analysis, UNSW. The deep sequencing data was analysed for single nucleotide polymorphisms by K. McElroy.

70 ! Chapter 3

3.3 Results

3.3.1 Mutations within the PAO1 biofilm variants

The superinfective phenotypes for SCV2 and S4 were stable and heritable and therefore it is likely that the superinfective variants harbour mutations within the phage genome. To test this hypothesis, PCR and sequencing were performed on genomic and phage DNA of the variants S4 and SCV2 and their genomes were compared with the Pf4 phage from the PAO1 parent strain. The WT Pf4 genome is 12,437 bp, located between positions 785311 to 797747 in the PAO1 genome (Tan, 2006, Klockgether, et al., 2010) and consists of seventeen open reading frames (Table 3-3). In comparison to the M13 filamentous phage and the Pf1 phage, the majority of the genes are core genes that are highly conserved and have previously been shown to play a role in replication, assembly and secretion of the phage (Figure 3-1). The Pf4 phage was first identified based on its homology to the Pf1 Pseudomonas phage (Webb, et al., 2004). In comparison to Pf1 and Pf5 phage, the Pf4 phage has three unique open reading frames at the 3’ end. Two of these encode a toxin-antitoxin (TA) system, the parE-like and phd-like genes, and one gene with homology to an integrase. The Pf4 also has two unique genes at the 5’ end, PA0715, a putative reverse transcriptase, and PA0716 with homology to an ATPase component of ABC transporter.

The initial hypothesis was that mutations in a gene within the phage genome were responsible for the superinfective phage conversion. Therefore, transposon mutants that separately inactivated each of the prophage genes from the Washington University transposon library (Jacobs, et al., 2003) were tested for superinfective phage using the phage assay (section 3.2.4.5). However, none of the 13 knockout mutants tested displayed superinfective phage. It should be noted that transposon insertions in PA0715, PA0723 and the non-annotated putative phd gene upstream of PA0729 were not available from the Washington transposon mutant library and hence could not be tested here (Table 3-3). Interestingly, six knockout mutants did not form plaques on the Pf4 mutant lawn that is sensitive to reinfection by the Pf4 phage, which may suggest those genes are essential for the production of phage particles.

71 !

Chapter 3

Table 3-3. The genes of the Pf4 prophage and the effect of transposon insertional deletion on superinfection.

Position Gene Name Gene Function Phage Function Phage assay Phage assay Pf4 lawn WT lawn 785969-786925 PA0715 Reverse transcriptase Replication NA NA 786928-788253 PA0716 ATPase binding protein Replication + - 788542-788808 Repressor NA NA NA 789144-789356 PA0717 Hypothetical protein Structural - - 789360-789650 PA0718 Hypothetical protein Structural - - 789654-790031 PA0719 Hypothetical protein Structural - - 790166-790600 PA0720 Hypothetical protein Structural + - 790617-790709 PA0721 Hypothetical protein Structural - - 790722-790973 PA0722 Hypothetical protein Structural - - 790986-791234 PA0723 Coat B protein Structural NA NA 791370-792632 PA0724 Coat A protein Structural + - 792637-792993 PA0725 Hypothetical protein Structural + - 792997-794271 PA0726 Zonular occludens toxin (zot) like protein Structural + - 794501-795793 PA0727 Hypothetical protein Structural - - 795793-796776 PA0728 Integrase Assembly + - 796990-797241 Prevent Host Death protein phd Toxin anti-toxin system NA NA 797251-797598 PA0729 Toxin protein parE Toxin anti-toxin system + -

The function of the genes was inferred based on similarity to the M13 filamentous phage and the Pf1 Pseudomonas filamentous phage. The phage assay was performed 13 knockout mutants to test for lysogenic Pf4 phage and superinfective Pf4 phage using detection bacterial lawns seeded with PAO1 ΔPf4 and PAO1 WT, respectively. Abbreviations: (+) represents plaque formation on target lawn, (-) represents no plaque formation on target lawn. (Winsor, et al., 2011)

72 ! Chapter 3

Figure 3-1. Alignment of filamentous phage M13, Pf1, Pf5 and Pf4 genomes. White arrows represent replication genes, grey arrows represent structural genes, black arrows represent assembly genes, and patterned arrows represent genes unique to that phage genome. All sequences are presented from 5’ to 3’ (left to right) and the direction of the arrow represents the direction of transcription.

Since none of the transposon mutants were observed to produce superinfective phage, the next approach was to sequence the phage genome of the variants that carry the superinfective phage, SCV2 and S4 variants isolated in Chapter 2. The putative repressor c gene in the Pf4 genome, located at position 788542 to 788808 and is encoded on the complementary strand, was first targeted. Sequencing results identified single nucleotide polymorphisms (SNPs) within and upstream of the repressor c gene (Figure 3-2 and Table 3-4) for the S4 variant. In contrast, the repressor c gene of variant SCV2 could neither be amplified nor sequenced. Therefore, genome walking of the Pf4 prophage genome of variant SCV2 was performed. The sequencing results showed a large deletion of 3,410 bp from the Pf4 genome of the SCV2 variant, resulting in a Pf4 genome of 9,027 bp and the loss of two open reading frames at the 5’ end of the Pf4 genome, PA0715 and PA0716, as well as the repressor c gene located on the complementary strand of the intergenic region between PA0716 and PA0717 (Figure 3-2). The large deletion of 3,410 bp is located at position 785529 to 788911 in the PAO1 genome. Other mechanisms can lead to superinfection such as changes to the phage receptor of the host or proteins that inhibit phage infection. Surprisingly, sequencing of WT-like revertants derived from SCV2 biofilm indicated that they no longer carried the large deletion of the repressor c gene as was observed for the SCV2 variant. The mechanism that allows for it to recover the deletion region is unknown.

73 ! Chapter 3

! Figure 3-2. Overview of the mutations identified in the variants carrying the superinfective form of the Pf4 phage. The Spreader Variant S4 had single nucleotide mutations within and upstream of the repressor c gene. The small colony variant SCV2 had a 3,410 bp deletion in the 5’ end of the Pf4 genome, which contained the repressor c gene (black arrow), and the first two open reading frames at the 5’ end of the Pf4 genome, PA0715 gene (reverse transcriptase gene) and PA0716 gene (ATPase binding protein gene). The direction of the arrow represents the transcription orientation in the genome. The white arrows represent the annotated genes in the Pf4 phage genome and the red markers indicate the SNPs within and upstream of the repressor c gene of the S4 variant.

74 ! Chapter 3

Table 3-4. The mutations upstream and within the repressor c gene of PAO1 S4 variant.

Position Mutation on Within/upstream of Codon change leading strand repressor c gene1

788570 C ! A within the repressor CGA! CTA c gene Arginine ! Leucine

788799 A ! G within the repressor TCA ! CCA c gene Serine ! Proline

788826 G ! T upstream of the non-coding region repressor c gene

788857 A ! G upstream of the non-coding region repressor c gene

1, it should be noted that the repressor c gene is encoded on the complementary strand.

3.3.2 Deep sequencing analysis of pre- and post- dispersal PAO1 biofilm To determine the overall frequency of mutations in the Pf4 phage during P. aeruginosa PAO1 biofilm development and the correlation with the establishment of superinfective variants, biofilm effluent from the pre- and post- dispersal stages were sampled and genomic as well as phage DNA was collected for deep sequencing analysis. The deep sequencing of the pre- and post- dispersal population of the biofilm may also give insight to the types of mutation and mutation frequencies in the host or phage genome that may result in superinfection. Dispersal was indicated by the peak of CFU counts during P. aeruginosa PAO1 biofilm development (Figure 3-3). The pre- and post- dispersal event of the biofilm was determined to be on day 4 and day 11, respectively. Phenotypic variants were also observed from the biofilm and peaked on day 11 with an average of 24.24% of SCVs (Figure 3-4). The deep sequencing analysis of the dispersal populations was performed by K. McElroy (McElroy et al. submitted).

75 ! Chapter 3

Figure 3-3. Dispersal and phage production of P. aeruginosa PAO1 biofilm submitted for deep sequencing analysis of pre- (white arrow) and post (black arrow) dispersal cells. The CFU counts (right axis, open triangles) indicate the viable cells per ml of biofilm effluent. Phage production was quantified as plaque forming units per ml of biofilm effluent on two types of bacterial lawns, PAO1 ΔPf4 (left axis, squares) and PAO1 WT (circles). Data are the means of three experiment and error bars show SEM.

Figure 3-4. Percentage of SCVs from the P. aeruginosa PAO1 biofilm submitted for deep sequencing analysis of pre- (white arrow) and post (black arrow) dispersal cells. Data are the means of three experiments and error bars show SEM.

76 ! Chapter 3

Bacteria within the biofilm at day 4, representing the pre-dispersal phase, were observed to have no mutations in the PAO1 host genome for both replicate experiments (Table 3-5). There were five mutations observed in the 1st experiment within the Pf4 phage genome and these were determined to be Single Nucleotide Polymorphisms (SNPs) in the putative promoter region upstream of the repressor c gene of the Pf4 phage genome (Table 3-6). There were no mutations observed in the 2nd experiment within the Pf4 phage genome.

The deep sequencing analysis of the post-dispersal cells detected a combination of SNPs and deletions. There were 3 mutations from the 1st experiment and 4 mutations in the 2nd experiment detected from the PAO1 host genome (Table 3-5). The mutations lie within the pilT, pilY1, and wzy gene, which play roles in twitching motility, LPS production and type IV pili biosynthesis. The frequencies of these mutations ranged from 0.5% to 7.3%. There were more mutations detected from the phage genome of the post-dispersal cells. The 5 mutations (of the phage genome) observed on day 4 of the 1st experiment were also present, along with mutations at an additional 8 positions therefore a total of 12 mutations were detected from the phage genome (Table 3-6). Amongst these 12 mutations, five were within the repressor c gene and 8 were in the putative promoter region upstream of the repressor c gene, and the mutation frequencies approached up to 68.6%. There were 16 mutations detected from the 2nd experiment, with 8 in the repressor c gene and an additional 8 upstream of the repressor c gene in the putative promoter region, and the mutation frequencies were up to 70.6%. For the 2nd experiment, free phage particles were also collected and sequenced. A total of 16 SNPs were detected and most of the mutations detected from the free phage samples corresponded to the genomic DNA samples for the 2nd experiment. In stark contrast, the mutation frequencies detected in the host genome, 7.3%, was much lower than the mutation frequencies detected in the phage genome, 79.8%. It was of interest to note that there were no mutations detected in any of the other phage genes suggesting the repressor gene represents a mutational hot-spot.

77 !

Chapter 3

Table 3-5. The mutations identified within the host genome from the variants of the P. aeruginosa PAO1 biofilm dispersal population.

PA011 (%) PA012 (%)

Position2 Variants Gene name Function Effect day 41 day 111 day 41 day 111

570 -12 bp pilT Type IV pili Frameshift - 7.3 - 3.1

959 -1 bp pilT Type IV pili Frameshift - - - 0.7

970 -15 bp pilT Type IV pili Deletion ‘KGLVA’ - 1.0 - 0.5

3274 C→T pilY1 Type IV pili Stop codon - - - 4.3

620 -1 bp wzy Lipopolysaccharide Frameshift - 0.5 - -

1The day 4 (pre-dispersal cells) and day 11 (post-dispersal cells) samples were collected and mutations identified include Single Nucleotide Polymorphisms (SNPs) and deletions. 2Position represents the position within the relevant gene.

st nd Abbreviations: PAO11 represents the 1 experimental replicate, PAO12 represents the 2 experimental replicate and a dash represents the variant with no mutation detected within the ~0.5% limit of detection. (McElroy, K. submitted)

78 !

Chapter 3

Table 3-1. The mutations identified within the Pf4 phage genome from the variants of the P. aeruginosa PAO1 biofilm and dispersal population.

PA011 (%) PA012 (%) Position2 Variants Gene Effect day 41 day 111 day 41 day 111 free3 787237 G→T PA0716 D→Y - - - 2.9 8.1 788699 G→T repressor c A→E - - - 31.9 52.5 788706 T→C repressor c K→E - 1.0 - - - 788714 C→A repressor c S→K - - - 32.7 20.3 788727 C→T repressor c E→K - - - 0.7 0.5 788739 -1 bp repressor c Frameshift - 24.8 - - - 788760 G→T repressor c P→T - - - 0.4 - 788762 C→T repressor c G→D - 0.3 - - - 788786 G→T repressor c A→E - 49.9 - 0.3 - 788799 A→G repressor c S→P - 18.9 - 5.9 3.5 788806 C→T repressor c Alternative start - - - 16.8 16.3 788823 C→A upstream of repressor c NA 16.2 - - - - 788825 G→C upstream of repressor c NA - - - 12.4 12.6 788826 G→T upstream of repressor c NA - 6.0 - - - 788827 G→A upstream of repressor c NA 2.0 47.2 - 0.9 1.1 788832 A→G upstream of repressor c NA 1.4 1.2 - - - 788837 C→A upstream of repressor c NA 13.0 16.0 - 70.6 79.8 788839 C→T upstream of repressor c NA - 18.3 - - - 788852 G→T upstream of repressor c NA - - - 12.7 19.7

79 !

Chapter 3

continued from previous page

PA011 PA012 Position2 Variants Gene Effect day(%) 41 day(%) 111 day 41 day 111 free3 788854 G→T upstream of repressor c NA - - - 2.7 5.2 788857 A→G upstream of repressor c NA 16.4 68.6 - 21.1 20.0 788865 C→A upstream of repressor c NA - 18.3 - 1.0 1.3 788865 C→T upstream of repressor c NA - - - 0.5 1.5 788867 C→A upstream of repressor c NA - - - - 0.3 788868 C→A upstream of repressor c NA - - - - 0.3

1The day 4 (pre-dispersal cells) and day 11 (post-dispersal cells) samples were collected and mutations identified include Single Nucleotide Polymorphisms (SNPs) and deletions. 2Position represents the position within the relevant gene. 3Free represents the Pf4 phage particles sequenced.

st nd Abbreviations: PAO11 represents the 1 experimental replicate, PAO12 represents the 2 experimental replicate and a dash represents the variant with no mutation detected within the ~0.5% limit of detection. (McElroy, K. submitted)

80 ! Chapter 3

3.4 Discussion

The phage replication and production genes are highly conserved amongst filamentous phage genomes. In M13, the 9 genes involved are named gene I to IX, and represent replication, structural and assembly genes of the phage. The two filamentous phage (fd and f1 phage) have up to 97% similarity to M13 (van Wezenbeek, et al., 1980), and mutations in all genes except gene II, have been shown to be lethal to bacterial host resulting in host cell death via inhibition of growth or affect host outer membrane (Hohn, et al., 1971, Woolford, et al., 1974, Horabin & Websster, 1986). Therefore, to determine if inactivation of any of the Pf4 genes would lead to the expression of the superinfective Pf4 phage, all available transposon insertions in the phage genome were tested for the production of superinfective phage. However none of the 13 knockout mutants exhibited the superinfective phage against the PAO1 WT and all of these transposon mutants were viable.

While the genes responsible for filamentous phage production are highly conserved amongst filamentous phage, the novel genes found in Pf4 could potentially play a role in superinfection as they are not found in the related phage Pf5 (Figure 3-1) which was reported to not play a role in SCV formation (Mooij, et al., 2007). Reverse transcriptase in retroviruses is responsible for synthesising a DNA copy of the viral genome of retroviruses and this step precedes integration into the host genome (Goff, 1990). In bacteria, reverse transcriptase has been shown to be encoded in retron elements that are involved in the synthesis of the unusual extragenetic elements, multi-copy, ssDNA (Rice & Lampson, 1996). Interestingly, some retron elements have been shown to be carried in phage (Inouye, et al., 1991). The ATPase binding protein is known to play a role in a large range of cellular functions in bacteria, such as DNA replication, protein degradation, membrane fusion, signal transduction pathways and chemoreceptors (Tam & Saier, 1993, Ogura & Wilkinson, 2001). The phd-like and parE-like genes at the 3’ end of the Pf4 phage genome are postulated to make up the toxin antitoxin (TA) system of the phage. TA systems are commonly found in bacterial genomes and have been suggested to function as ‘addiction modules’ (Yarmolinsky, 1995). There are roughly 8 defined classes based on mechanism of 81 ! Chapter 3

action. It is well established that the toxin and the antitoxin proteins or RNAs have different half-lives due to their susceptibility to proteases or nucleases, where the toxin has a longer half-life than the antitoxin. Some of the toxins of TA systems have been shown to target DNA gyrase, responsible for the supercoiling mechanism in DNA and mRNA, affecting protein synthesis and inducing RNase activity. When the antitoxin is bound to the toxin, this causes a conformational change, and inhibits the toxin activity. TA systems in bacteria are also considered as a programmed cell death system and have been shown to be activated by stressful conditions such as amino acid starvation and DNA damage, leading to cell death (Couturier, et al., 1998, Jiang, et al., 2002, Buts, et al., 2005, van Melderen & de Bast, 2009). These unique genes found in the Pf4 phage may play important roles in phage conversion into lytic phage, and variant formation, as Pf5 phage that do not carry these genes, have been shown to form biofilms with no conversion to the lytic filamentous phage and no SCV formation (Mooij, et al., 2007). Sequencing of these genes from the S4 variants showed no mutations in the PA0715, PA0716 and TA system genes. On the other hand, there were difficulties amplifying the PA0715 and PA0716 gene from the SCV2 variant. These results, which are consistent with the transposon insertion mutant data above, are significant as this indicates that the majority of the core genes within the phage genome are not responsible for the superinfection phenotype. Hence the replication, structural, assembly and toxin genes are not involved in the conversion of the lysogenic phage into the lytic superinfective phage.

It has been suggested that the immunity region of the phage plays a role in host resistance to superinfection in temperate and lysogenic phage (Cheng, et al., 1999, Canchaya, et al., 2003, Canchaya, et al., 2004). Here it has been shown that both superinfective variants isolated from PAO1 biofilms, carry defective or mutated repressor c genes. The PAO1 SCV2 variant had a 3,410 bp deletion in the Pf4 genome with the loss of two genes (PA0715 and PA0716). Furthermore, the intergenic region between PA0716 and PA0717 was also deleted, which contains the repressor c gene and the possible promoter region upstream of the repressor. The PAO1 S4 variant carries mutations within and upstream of the repressor c gene. When the complete Pf4 phage genome of the PAO1 SCV2 variant was sequenced, there were no mutations found in the remaining 9,027 bp of the 82 ! Chapter 3

genome. Collectively these data suggest that mutations involving the repressor c gene or its promoter are likely to be involved in the establishment of superinfection. The intergenic region between PA0716 and PA0717 was checked for possible promoter sites using detection software however nothing was detected. There was no homology to any existing known repressor promoter sites. The two SNPs identified within the repressor c gene result in amino acid substitutions of arginine to leucine and serine to proline (Table 3-4). The change from a hydrophilic arginine to a hydrophobic leucine can influence the protein folding. Hydrophobic amino acids tend to be buried within the core of the protein whilst hydrophilic amino acids cover the surface of the protein to interact with solvents. Proline consists of a cyclic structured side chain therefore it tends to disrupt secondary structures by inhibiting backbone structure to alpha-helix or beta-sheet conformations. These major changes to the secondary structure of the repressor could alter the affinity of the protein for binding. Similarly, mutations within the putative promoter of the repressor could also influence the binding between the promoter and the repressor C protein, which may control the lysogeny of the phage and the host. The deep sequencing results also supported this conclusion, where the same mutations (within the intergenic region of PA0716 and PA0717 gene) observed by manual sequencing were also observed in the dispersal population. Similarly, the deep sequencing data showed no mutations in the rest of the phage genome and only a limited number of mutations outside of the phage genome. The sequencing data of WT-like revertant variants derived from the PAO1 SCV2 variant that lost the superinfection phenotype showed that these variants no longer carry the deletion of the repressor c gene as was observed for the SCV2 variant. The mechanism of the restoration of the complete Pf4 phage is unknown. This restoration of lost genes may suggest that a RF carries a copy of the complete Pf4 genome or the genes were somehow gained through recombination after infection with WT phage.

Similar observations support the importance of the repressor C protein in other bacteriophage. It was identified that two mutations in the Salmonella phage P22 affected immunity to superinfection (Chan & Botstein, 1972). Both mutations were located within the phage immunity gene (Levine & Curtiss, 1961, Chan & Botstein, 1972). Furthermore, the loss of the phage repressor in the lysogenic 83 ! Chapter 3

phage KS9 of Burkholderia cepacia LMG 21824 triggers conversion into its lytic form and results in loss of its ability to undergo lysogeny (Lynch, et al., 2010).

The deep sequencing analysis also yielded results that support the hypothesis that the repressor c gene is responsible for the immunity to superinfection and the superinfection phenotype of the PAO1 SCV2 and S4 variants. As shown in Table 3.6, most of the mutations identified from the pre- and post- dispersal cells lie within or upstream of the repressor c gene. The mutation frequencies were as high as 79.8%. It was shown that the global transcriptional regulator OxyR, binds to the Pf4 genome between PA0716 to PA0719 (Wei, et al., 2012). Based on the binding motif, the OxyR protein, which plays a role in oxidative stress response, binds within the repressor c gene. However, none of the significant mutations identified in the deep sequencing analysis lie within the OxyR binding site in the repressor c gene or the possible helix-turn-helix domain promoter site for the PA0717 to PA0719 operon. This supports the hypothesis here that the repressor c gene or repressor C binding site upstream of the gene plays a crucial role in immunity against superinfection and/or superinfection.

The results presented in this chapter have highlighted the importance of the mutations concentrated within the intergenic region of PA0716 and PA0717 genes of the Pf4 phage genome. Biofilm communities have been shown to undergo mutations during biofilm development, and many variants isolated have shown mutations in biofilm specific genes, such as quorum sensing, motility, attachment and more (Deretic, et al., 1995, Davies, et al., 1998, De Kievit, et al., 2001, Starkey, et al., 2009). From the deep sequencing analysis, four mutations at low mutation frequencies were found in twitching and motility related genes, as well as one mutation involved in LPS production. The low number of mutations was unexpected, as was the observation that the mutations detected were concentrated within the phage genome and even then, within one specific, short open reading frame and its putative promoter. The mutations were also associated with the hypothesised immunity region of the phage. This may suggest that the limited genetic changes are associated with the phage conversion into the superinfective form, and which plays a role in mediating cell death and dispersal in the biofilm, and variant formation. Furthermore, the immunity region of the phage may be a

84 ! Chapter 3

hot spot for mutations leading to the establishment of the superinfective phage. If this were the case, it would suggest strong evolutionary pressure to evolve a mechanism that can allow or facilitate the formation of superinfective phage. It will be particularly interesting to define the mechanism that leads to these highly specific mutations and to ultimately understand whether the selection is at the level of the host or the phage. A mutant knockout of the repressor c gene may potentially identify the effects of loss of its function on biofilm development and superinfective phage conversion. Confirmation of the promoter region of the repressor can be achieved by binding assays with expressed repressor C proteins to the putative promoter region. Thus, binding assays of mutated repressor promoter region with WT repressor or mutated repressor with WT promoter region may identify the specific SNPs within the S4 that affects the function of the WT repressor and induce superinfective phage conversion.

85 ! Chapter 4

Chapter 4 – Environmental cues and genes involved in establishment of the superinfective Pf4 phage

4.1 Introduction

The data presented in Chapter 3 suggests that the establishment of a superinfective variant of the Pf4 phage is a consequence of mutations in a putative phage encoded repressor protein gene. Because the formation of the superinfective phage is an integral part of the biofilm development life-cycle, which does not commonly occur in planktonic cultures, it is likely that the mutations observed are the result of biofilm specific physiology and gene expression. However, the specific conditions that lead to the mutations responsible for superinfection are currently not known.

Biofilms consist of a stratified population, where cells in different parts of the biofilm exhibit varied physiologies based on differences in nutrient gradients, oxygen gradients, signalling molecules and the accumulation of metabolic products. For example, an oxygen gradient can be formed by the failure of oxygen to penetrate through the biofilm as oxygen is respired by the upper layer of cells of the biofilm (Stewart & Franklin, 2008). Similar observations have demonstrated a gradient of nutrients decreases from the nutrient source, normally the bulk phase, to deeper into the biofilm (Zhang, et al., 1995, Stewart, 2003). This can result in the reduction of nutrient availability and causes stress to the biofilm cells leading to nutrient starvation in the biofilm’s interior. Conversely, there is higher concentration of metabolic byproducts in the interior of the biofilm compared to the cells in the upper layer of the biofilm, where such waste products may freely diffuse away from the biofilm into the bulk solution. The accumulation of waste products such as reactive oxygen and nitrogen species (RONS) has been observed in biofilms leading to oxidative and nitroxidative stress (Webb, et al., 2003, Barraud, et al., 2006). For example, aerobic bacteria generate high concentrations of electrons during redox reactions, especially during respiration. These reactions also are partly responsible for the release of different species of 86 ! Chapter 4

oxygen. This results in the build up of ROS and can lead to DNA damage, protein carbonylation, cofactor degradation, and lipid peroxidation. Therefore, bacteria are dependent on oxidative stress defense systems to mitigate the build up of oxygen and oxygen derivatives such as hydrogen peroxide (H2O2), superoxide anions (O2-), and hydroxyl radicals (OH-). Bacteria counteract oxidative stress by expressing enzymes to detoxify reactive oxygen species, repair damages and express adaptive response to changing environmental conditions (Storz & Imlay, 1999, Vinckx, et al., 2010). This again supports the observation that bacteria in the biofilm interior differ from those on the exterior surface in terms of not only general physiology, but also in terms of gene expression. Therefore, it is anticipated that stress responses play important roles in host defense in constantly changing environments, similar to that observed in biofilm conditions.

In Pseudomonas aeruginosa, the OxyR transcriptional regulator is the key global regulator of the oxidative stress response and regulates 56 genes (Wei, et al., 2012), including katA, kayB, ahpB, ahpCF, sodA and sodB, which are involved in oxidative stress defense. The alternative sigma factor RpoS also plays a role in the regulation of host defense mechanisms in response to heat stress, osmotic stress, oxidative stress and nutrient starvation (Jorgensen, et al., 1999, Suh, et al., 1999). Based on DNA microarray analysis, up to 800 genes have been shown to be regulated by RpoS in P. aeruginosa and RpoS is expressed at a higher level under biofilm conditions compared to planktonic cultures (Schuster, et al., 2004).

When these stress response systems of P. aeruginosa fail to detoxify RONS, these reactive molecules can damage DNA, potentially leading to deleterious genetic changes. Bacteria have also evolved a number of DNA repair mechanisms such as the methyl directed mismatch repair (MMR) and the RecA recombination repair systems. The MMR corrects errors that occur during DNA replication and is a key factor in minimising mutations during replication (Modrich, 1991). RecA is regulated at many different levels in the cell and is activated under a range of different conditions. It acts as a recombinase and facilitates translesion synthesis during DNA repair as well as facilitating the cleavage of the LexA repressor during the SOS response (Cox, 2007). These repair systems are important when cells encounter stresses and spontaneous mutations that occur during replication

87 ! Chapter 4

and therefore, may play important roles in the genetic changes that have been observed to be associated with the formation of the superinfective phage. Here, the physiological triggers that lead to mutations and hence superinfection and variant formation and to determine the role of DNA repair mechanisms in formation of the superinfective Pf4 phage were investigated. Multiple inducers were tested, including starvation for different key nutrients, oxidative stress, exposure to H2O2 and mitomycin C. The induced DNA damage from H2O2 and mitomycin C resulted in significantly increased superinfection. Similarly, mutational inactivation of the oxidative response regulator, OxyR as well as the inactivation of the MMR response via mutation of mutS gene, resulted in earlier phage production and a higher titer of the superinfective Pf4 phage. Interestingly, loss of RecA resulted in a decrease in the formation of superinfective phage suggesting that recombination may be important in this process.

88 ! Chapter 4

4.2 Materials and Methods

4.2.1 Bacterial strains and culture conditions The P. aeruginosa strains used in this study are listed in Table 4-1. All P. aeruginosa strains were cultured as described in section 2.2.1 in Chapter 2.

Table 4-1. List of P. aeruginosa strains used in this study.

Strain Reference

PAO1 WT Laboratory stock

PAO1 ΔPf4 Rice et al., 2009

PAO1 SCV2 This study

PAO1 ΔmutS Jacobs, et al., 2003

PAO1 ΔrecA Jacobs, et al., 2003

PAO1 ΔoxyR Wei, et al., 2012

4.2.2 Biofilm experiments

4.2.2.1 Planktonic cultures Planktonic cultures were cultivated in 15 ml centrifuge tubes (Falcon). Briefly, strains were cultured in M9 medium supplemented with 15 mM glucose and incubated overnight at 37oC with constant agitation on a shaker at 200 rpm. One millilitre of culture was removed for phage titre assay. The cultures were centrifuged at 10,000 x g for 3 min and the supernatant was discarded. The pellet was resuspended in 9 ml of M9 complete medium and were incubated at 37oC at 200 rpm for 24 h. Starvation was induced by replacing M9 complete medium with a solution of M9 salts without glucose (carbon starvation) or without ammonium chloride (nitrogen starvation) and this solution was added to the culture tubes for 3 d. Cultures were exposed to nitric oxide (NO) by supplementing the M9 complete medium with the NO donor SNP (Sigma Aldrich) at 10 µM, 100 µM, and 1 mM for 3 d. Oxidative stress was induced by supplementing M9 complete medium with 100 µM, 1 mM and 10 mM H2O2 (Biorad) for 3 d. Mitomycin C (Sigma Aldrich) was added to M9 complete medium at 3 µM, 30 µM and 150 µM 89 ! Chapter 4

to induce DNA damage and cultures were treated for 3 d. All treatments were performed as biological triplicates and were compared to cultures maintained in M9 complete medium. The samples were collected daily to determine the phage titre.

4.2.2.2 Batch biofilms Batch biofilms were cultivated in tissue culture treated 24 well microtitre plates (Falcon). Briefly, overnight cultures grown in M9 medium supplemented with 15 mM glucose were diluted 1:100 and 1 ml of the diluted culture was inoculated into each well. Biofilms were allowed to form for 1 d before being treated. Starvation was induced by replacing M9 complete medium with a solution of M9 salts without glucose (carbon starvation) or without ammonium chloride (nitrogen starvation) and this solution was added to the wells for 3 d. Biofilms were exposed to nitric oxide (NO) by supplementing the M9 complete medium with the NO donor SNP (Sigma Aldrich) at 10 µM, 100 µM, and 1 mM for 3 d. Oxidative stress was induced by supplementing M9 complete medium with 100 µM, 1 mM and 10 mM H2O2 (Biorad) for 3 d. Mitomycin C (Sigma Aldrich) was added to M9 complete medium at 3 µM, 30 µM and 150 µM to induce DNA damage and biofilms were exposed for 3 d. The plates were incubated for 72 h, at 37oC with shaking at 80 rpm. At 24 h intervals, samples were centrifuged at 13,000 x g for 3 min to obtain cell-free supernatant. One millilitre of M9 medium supplemented with 15 mM glucose was added to each well. The plates were further incubated for 24 h at 37oC with shaking at 80 rpm. All treatments were performed as biological triplicates and were compared to biofilms maintained in M9 complete medium. The samples were collected daily to quantify the phage titre (see below).

4.2.2.3 Continuous-culture biofilm Biofilms were set up as described in section 2.2.2.1, with slight modifications. Biofilms were allowed to form for 2 d before being treated. Starvation was induced by replacing M9 compete medium with a solution of M9 salts without glucose (carbon starvation) or without ammonium chloride (nitrogen starvation) and this solution was flowed across the biofilm for 5 d. Biofilms were exposed to nitric oxide (NO) by supplementing the M9 complete medium with the NO donor SNP at 1 mM for 5 d. Oxidative stress was induced by supplementing M9

90 ! Chapter 4

complete medium with 10 mM H2O2 for 5 d. Mitomycin C was added to M9 complete medium at 150 µM to induce DNA damage and biofilms were exposed for 5 d. All treatments were performed as biological triplicates and were compared to biofilms maintained in M9 complete medium. The samples were collected daily for CFU counts and quantification of phage titre.

4.2.2.4 CFU counts and morphological variants from the biofilm effluent To determine the CFU counts from the biofilm and observe morphological variants, 5 ml of the biofilm effluent was collected from each flow cell on days 2 – 7 and sampled as described in section 2.2.2.2.

4.2.2.5 Phage assay The supernatant was tested for superinfective phage using the phage assay described in Chapter 2 section 2.2.2.3, a modified version of the top-layer agar method previously described (Eisenstark, 1967). The cell free supernatant samples were spotted onto LB10 agar plates containing an overlay of the bacterial lawn of PAO1 WT or PAO1 ΔPf4.

91 ! Chapter 4

4.3 Results

In P. aeruginosa PAO1 biofilms, the lysogenic Pf4 prophage converts into its superinfective form during the dispersal phase and this is accompanied with the appearance of SCVs in the dispersal population. Dispersal has previously been linked to nutrient starvation and the accumulation of nitric oxide (Sauer, et al., 2004, Barraud, et al., 2006), however the specific metabolic or stress conditions that lead to the establishment of the superinfective phage has not been determined. Therefore, planktonic cultures, batch biofilms and continuous culture biofilms were used to investigate different stress conditions that may be involved in the development of superinfection.

Several strains including ΔmutS, ΔrecA and ΔoxyR mutants, as well as chemical treatments including H2O2, mitomycin C, and NO donor SNP were tested for a possible induction of the conversion of the superinfective phage. Planktonic cultures showed no consistent appearance of the superinfective phage over a period of 3 d (Table 4-2). Furthermore, the untreated control cultures were similarly inconsistent in that the appearance of the superinfective phage was highly variable and appeared to be random. When the same treatments were tested on batch grown biofilms, no superinfection was observed, even after three days of cultivation. Most surprisingly, even the control biofilm showed no superinfection, suggesting that batch biofilms in microtitre plates, over the three days tested, do not replicate flow cell conditions sufficiently to allow for the formation of the superinfective phage. Therefore, continuous culture biofilms were subsequently used to test the various treatments for their impact on the development of superinfection.

92 ! Chapter 4

Table 4-2. The phage assay using the PAO1 WT bacterial lawn to detect the presence of the superinfective phage in planktonic and batch biofilm cultures.

Planktonic culture Batch biofilm culture Treatment replicate 1 d 2 d 3 d 1 d 2 d 3 d 1 - - + - - - WT control 2 - - + - - - 3 - + + - - - Carbon 1 - - + - - - Starvation 2 ------3 ------Nitrogen 1 ------Starvation 2 - + + - - - 3 ------1 - - + - - - SNP 10 µM 2 - - + - - - 3 - - + - - - 1 - - + - - - SNP 100 µM 2 - + + - - - 3 ------1 - - + - - - SNP 1 mM 2 - - + - - - 3 ------H O 100 1 - + + - - - 2 2 2 - + + - - - µM 3 - - + - - - 1 - - + - - - H2O2 1 mM 2 - + + - - - 3 - + + - - - 1 - + + - - - H2O2 10 mM 2 - - + - - - 3 - + + - - - Mitomycin C 1 - - + - - - 2 - - + - - - 3 µM 3 - + + - - - Mitomycin C 1 - - + - - - 30 µM 2 - - + - - - 3 - + + - - - Mitomycin C 1 - + + - - - 2 - + + - - - 150 µM 3 - + + - - - ΔmutS 1 - + + - - - mutant 2 - + - - - - 3 - - + - - - ΔrecA 1 - - + - - - 2 - - + - - - mutant 3 - - + - - - ΔoxyR 1 - - + - - - 2 - - + - - - mutant 3 - + + - - - Data shown here is the first experimental data set of the phage assay. Abbreviations: d represents days, (+) represents plaque formation, (-) represents no plaque formation. 93 ! Chapter 4

In continuous culture biofilms, the control biofilm produced no superinfective phage until day 4 at 9.72 x 104 PFU/ml (Figure 4-1). The number of superinfective phage increased exponentially to day 6, reaching a maximum of 7.03 x 109 PFU/ml. When the nutrient starved (carbon or nitrogen) biofilms were examined, it was observed that these biofilms dispersed within 24 h of treatment. This was accompanied by a dramatic decrease in the overall phage titre and this was most likely due to the loss of biomass as a consequence of dispersal (data not shown). These results were in agreement with previous work showing that starvation induces biofilm dispersal (Huynh, et al., 2012). Therefore, these experiments were not repeated. When the continuous culture biofilms were exposed to 10 mM of H2O2 starting from day 2, the superinfective phage appeared on day 4 at 4.68 x 107 PFU/ml and increased throughout the duration of the experiment reaching a maximum of 3.65 x 1010 PFU/ml on day 7. Similarly, PAO1 biofilms treated with the NO donor SNP, showed 9.12 x 107 PFU/ml in the biofilm effluent on day 4. The number of PFU/ml increased on day 5 and 6, reaching a maximum of 8.10 x 109 PFU/ml on day 6 and decreased on day 7 to 2.14 x 109 PFU/ml. This pattern of change in PFU numbers observed from the SNP treated biofilms was also observed when the biofilm was exposed to DNA damaging agent, mitomycin C (Figure 4-1). There was an exponential increase in PFUs starting on day 4 of 3.36 x 107 PFU/ml, then reached a plateau on day 6 of 1.86 x 1010 PFU/ml and then declined on day 7 to 5.55x109 PFU/ml. The treated biofilms compared to the untreated PAO1 biofilm did not show significant differences. The 2 log difference on day 4 that was evident in three independent experiments may suggest that these treatments do have a biological effect on the appearance of the superinfective phage even though they did not induce the earlier production of superinfective phage.

It has previously been shown that superinfective Pf4 phage can induce the formation of SCVs (Rice et al., 2009). Therefore, the biofilm effluents from the biofilms treated with mitomycin C, H2O2 and SNP were collected, diluted and spread plated to quantify the number of SCVs present relative to the untreated control biofilms (Figure 4-2). The percentage of SCVs observed from the mitomycin C treated biofilms were higher than the percentage of SCVs from the untreated PAO1 biofilm. When the biofilms were exposed to mitomycin C on 94 ! Chapter 4

Figure 4-1. The appearance of superinfective Pf4 phage during biofilm development of PAO1 WT (open triangle with dotted line), PAO1 treated with mitomycin C (hexagon), PAO1 treated with H2O2 (diamond) and PAO1 treated with SNP (inverted triangle). The superinfective Pf4 phage was detected via the plaque assay with the PAO1 WT as the target lawn. Plaque formation on PAO1 WT lawn shows the presence of superinfective Pf4 phage. Data represent the means of three independent experiments and error bars show SEM.

! day 3 of the biofilm, 1.11% of SCVs were observed from the biofilm effluent and peaked on day 5 with 6.09% of SCVs, and decreased on day 7 to 2.7% of SCVs. In comparison, the untreated PAO1 biofilm effluent consisted of less than 1% of SCVs on day 3 and 2.49% and 2.36% of SCVs from day 5 and 7, respectively. The DNA damaging agent induced the production of more SCV variants in the biofilm and the effect was observed within 24 h of exposure. In contrast, PAO1 biofilms treated with H2O2 and SNP had a lesser effect on the biofilm. There were less than 1% of SCVs on day 3 from the biofilm exposed to H2O2, and this increased on days 5 and 7, reaching 5.13% of SCVs on day 7 in the biofilm effluent. Similarly, the biofilm exposed to SNP had less than 1% of SCVs from day 3, reaching its maximum of 9.06% of SCVs on day 7. In comparison to the untreated PAO1 biofilm, there was a 2 and 4 fold increase of SCVs from the biofilm effluent on day 7 for the H2O2 and SNP treated biofilms, respectively.

These results suggest that mitomycin C, H2O2 and SNP, exert their effects on the 95 ! Chapter 4

PAO1 biofilm at different stages of development. The treated biofilms compared to the untreated PAO1 biofilm did not show significant differences in terms of SCV production. However, the trend was observed in three independent experiments, which may suggest these treatments induce the appearance of the SCVs.

The data presented above (Figures 4-1 and 4-2) may suggest that oxidative stress and DNA damage may lead to the development of the superinfective phage and induce the appearance of SCVs. Oxidative stress is primarily perceived in P. aeruginosa to be controlled by the transcriptional regulator OxyR (Storz & Imlay, 1999). This protein changes in conformation upon exposure to oxidative stress, allowing it to bind to specific promoters to control their expression. Therefore, the oxyR mutant biofilm was compared with the PAO1 WT biofilm to investigate the role of the global transcriptional regulator OxyR in the conversion of the Pf4 phage into its superinfective form during biofilm development. The oxyR mutant biofilm showed early conversion into the superinfective Pf4 phage on day 4 as compared to the PAO1 WT biofilm, which produced superinfective phage on day 5 in these experiments (Figure 4-3). Statistical analyses were performed on the data for figures 4-1, 4-2 and 4-3 and no statistically significant differences were observed. On day 4 for the oxyR mutant biofilm, there was 1.23 x 103 PFU/ml detected from the biofilm effluent, and reached 8.08 x 109 PFU/ml on day 5, and plateaued thereafter. The oxyR mutant biofilm was also observed to have a higher percentage of SCVs during the late stages of biofilm development (Figure 4-4). The PAO1 WT biofilm effluent had 1.29% SCVs on day 6 and 1.92% on day 7 in comparison to 16.73% and 7.58% of SCVs from the oxyR mutant biofilm on days 6 and 7, respectively. The difference between the two biofilms was statistically significant on day 6, with more than 10 fold difference in SCVs detected from the biofilms. The same trend was observed amongst the three independent experiments therefore it is highly likely that OxyR plays an important role in the conversion to the superinfective phage and the appearance of SCVs.

96 ! Chapter 4

Figure 4-2. The percentage of SCVs from the PAO1 biofilm untreated (black bars), treated with 150 µM of mitomycin C (grey bars), treated with 10 mM of H2O2 (white bars) and treated with 1 mM of SNP (stripes bars). Data represent the means of three experiments and error bars show SEM.

!

Figure 4-3. The role of OxyR in the development of superinfection. Biofilm effluents from the PAO1 WT (triangle) and oxyR mutant (diamond) were screened for the appearance of the superinfective Pf4 phage, using the soft agar overlay method with the PAO1 WT as the target lawn. Data are the means of three independent experiments and error bars show SEM.

97 ! Chapter 4

The early and increased number of superinfective phage observed for the oxyR mutant further supports the hypothesis that oxidative stress and DNA damage are associated with this conversion. Bacteria have evolved a number of mechanisms to repair damaged DNA and two of the most important mechanisms are the MMR system and the RecA recombination system. To mimic the loss of a functional repair system and the effects it has on the phage lifecycle, a mutS mutant biofilm was compared to the PAO1 WT biofilm. It was observed that the mutS mutant biofilm produced 6.50 x 104 PFU/ml of superinfective phage on day 3 compared to the PAO1 WT biofilm when conversion occurred two days later on day 5 (Figure 4-5). Additionally, on day 5 the mutS mutant produced 4.29 x 1012 PFU/ml, which was 4 log higher than the WT at 9.82 x 107 PFU/ml. The number of PFU from the mutS mutant biofilm decreased to 4.57 x 109 PFU/ml on days 9 and 11. The number of PFUs were statistically significant different on days 5 and 7 when compared to the WT biofilm (P < 0.01). In contrast, the recA mutant showed a significant reduction in the number of superinfective phage, 4.69 x 106 PFU/ml compared to the PAO1 WT biofilm, with Pf4 phage conversion occurring on day 5 of biofilm development (Figure 4-5). Indeed, the number of superinfective phage observed in the biofilm effluent for the recA mutant was approximately 1 log lower than the WT at all time points tested.

When comparing the dispersal variants from the biofilms, SCVs were observed from the mutS mutant biofilm from day 1 onwards and peaked on day 5 at 6.09% SCVs, which was statistically significant (P < 0.05) compared to the WT biofilm with 2.49% of SCVs on day 5 (Figure 4-6). While the number of SCV’s was higher for the mutS mutant on days 1-5, the percentage of SCVs produced by the mutS mutant was similar to the WT from day 7 onward. In contrast, the percentage of SCVs observed in the effluent of the recA mutant biofilm was not statistically different from the WT for days 1 - 3. However, from day 7 onwards, the recA mutant generated significantly more SCVs than the WT and the mutS mutant. For example, on day 11 of the biofilm, the recA mutant biofilm consisted of 9.03% SCVs from the dispersal population, with a statistically significant difference compared to 3.22% of SCVs for the WT (P < 0.01). These data suggest that the appearance of the superinfective phage correlates with the timing of appearance of the SCV’s but does not correlate with the titre of the superinfective 98 ! Chapter 4

Figure 4-4. The percentage of SCVs from the dispersal population of the PAO1 WT biofilm (black bars) and the PAO1 oxyR mutant biofilm (white bars). Colony forming units were determined from biofilm effluents for phenotypic variants from the biofilms. Data represent the means of three experiment and error bars show SEM. (*) indicates a statistically significant difference when compared to PAO1 WT at 95% confidence interval with Two-way ANOVA with a Sidak’s post test.

!

Figure 4-5. The appearance of the superinfective Pf4 phage during biofilm development for PAO1 (closed triangle), the mutS mutant (inverted triangle) and the recA mutant (hexagon). Data are the means of three independent experiments and the error bars show SEM. (**) indicates a statistically significant difference compared to PAO1 WT at the 99% confidence interval as determined using a One-way ANOVA. 99 ! Chapter 4

phage. For example, the recA mutant produced less superinfective phage, but significantly more SCVs than either the WT or the mutS mutant (Figure 4-6), which produced greater than 4 log more phage than the recA mutant from days 7 to 11 (Figure 4-5).

!

Figure 4-6 The percentage of SCVs in the biofilm dispersal population for PAO1 WT (black bars), the mutS mutant (dark grey) and the recA mutant (light grey). Data are means of three experiments and the error bars indicate SEM. (*) and (**) indicate a statistically significant difference when compared to PAO1 WT at 95% and 99% confidence intervals respectively as determined by using a One-way ANOVA.

!

100 ! Chapter 4

4.4 Discussion

It was previously shown (Chapters 2 and 3) that growth in a biofilm leads to the formation of a superinfective variant of the Pf4 phage and that this conversion is the consequence of a heritable genetic change in a specific gene that is putatively involved in infection immunity. The results presented in this chapter strongly suggest that the superinfection conversion is linked to a dysfunctional oxidative stress response and MMR system. This is also supported by the results showing that chemical treatments that are linked with either DNA damage (mitomycin C) or oxidative stress (H2O2 and SNP) also result in increased or early development of superinfection. Oxidative stress is a consequence of the build up of RONS as a result of endogenous by products that accumulate during aerobic metabolism and or upon external exposure to ROS, such as the oxidative burst of immune cells. Aerobic bacteria naturally generate high concentrations of electrons and multiple oxygen species through oxidative phosphorylation and respiration. This results in 2- the build up of ROS such as H2O2, superoxide anion (O ) and hydroxyl radical (OH-) (Henle & Linn, 1997, Storz & Imlay, 1999). It has been shown that ROS and RONS accumulate within microcolonies of the biofilm leading to cell death (Barraud, et al., 2006). While bacterial cells are constantly exposed to intracellularly generated ROS, they are also exposed to exogeneous ROS generated by other microorganisms which can either be used to eliminate competitors or can be released by immune cells to kill invading pathogens (Klotz & Hutcheson, 1992, Brunder, et al., 1996). For example, the release of ROS is a host defense mechanism used by macrophages to kill pathogens by producing superoxides by the phagocyte NADPH oxidase (Miller & Britigan, 1997, Babior, 1999, Janssen, et al., 2003).

Hydrogen peroxide can freely diffuse through cellular membranes, making it a lethal antimicrobial where it causes DNA damage through DNA strand breakages (Schweitz, 1969), deoxynucleotide base damage (Rhaese & Freese, 1968), deoxynucleotide base release (Ward & Kuo, 1976) and DNA cross-linking (Massie, et al., 1972). It has previously been shown that the rate of damage caused by hydroxyl radicals, the product of H2O2 degradation via the Fenton reaction, is greater than H2O2, where hydroxyl radicals have a greater capability to bind to 101 ! Chapter 4

DNA (Schweitz, 1969, Ward, et al., 1987). This could explain the observation that

H2O2 had a relatively minor effect on the biofilm compared to the loss of the OxyR regulator (oxyR mutant biofilm), as OxyR acts in responses to combinations of RONS. Exposure of H2O2 gives rise to oxidative stress, however bacterial cells harbour DNA repair mechanisms to correct DNA damages and produce enzymes to scavenge and remove H2O2 (Ma & Eaton, 1992). Therefore, H2O2 may induce the expression of superinfective phage by a few hours because it has limited capacity to damage the cell by comparison to the effects of deletion of OxyR. In the case of the OxyR mutant, the superinfective phage was induced on day 4 and the PAO1 WT biofilm on day 5 which may reflect the broader role of OxyR in controlling the global oxidative stress response.

OxyR acts as a sensor to oxidative stress, and ROS activates the binding of OxyR to activate the expression of stress response genes. During H2O2 exposure, OxyR undergoes a conformational change to convert into its oxidised state and actively binds and activates transcription of antioxidant genes (Hassett, et al., 2000, Ochsner, et al., 2000, Wei, et al., 2012). When the oxidative stress response fails to respond, DNA damage occurs leading to more SNPs and mutations in the genome. It has been suggested that for P. aeruginosa, the hypermutable CF isolates observed from the lungs of CF patients are induced as a result of chronic oxidative stress from the immune cells, which promotes genetic changes (Ciofu, et al., 2005). These data suggest that oxidative stress promotes genetic changes in the PAO1 biofilm population, leading to SNPs in the genome (especially within the phage genome which may be a hot-spot for mutation) leading to the superinfective phage conversion.

It was determined that OxyR of P. aeurginosa binds to multiple sites in the P. aeruginosa PAO1 genome (Wei, et al., 2012). Interestingly, one of those sites is the intergenic region between PA0716 and PA0717 gene of the Pf4 genome. The binding region lies within the ORF of the repressor c gene (Figure 4-7). This may suggest interactions of the OxyR protein with the Pf4 phage genome are important in the superinfection conversion and overall control of phage production. One possible mechanism is that the OxyR normally binds to the repressor C promoter and represses gene expression. When the repressor C acquires mutations, these

102 ! Chapter 4

may prevent binding of the OxyR to the phage genome leading to overproduction of the phage particles. Binding assays of the OxyR to the repressor c gene and/or competition binding between OxyR and repressor C may elucidate the role of OxyR superinfective phage conversion and the interactions between OxyR regulator and the prophage.

Figure 4-7. The location of the OxyR binding site within the ORF of the repressor c gene of Pf4 phage. Number in brackets is the position within the P. aeruginosa PAO1 genome. Red bars represent mutations identified in the PAO1 S4 variant (Chapter 3).

In E. coli, the key proteins involved in the oxidative stress response have been previously shown to prevent DNA damage. For example, superoxide dismutase mutants, sodA and sodB mutants, were shown to have increased mutations in aerobic systems (Carlioz & Touati, 1986, Farr, et al., 1986). Similarly, hpx- mutants, lacking enzymes to scavenge hydrogen peroxide, exhibit high rates of mutagenesis relating ROS to DNA damage. Furthermore, significantly higher rates of mutation were observed in a hpx- oxyR double mutant indicating that OxyR plays a role in suppressing DNA damage (Park, et al., 2005). The MMR system is also an essential component of the DNA damage repair system.

103 ! Chapter 4

The MMR system detects and repairs mismatched nucleotides during DNA replication and is important for minimising mutations during replication. Defective MMR systems have been shown to stimulate the production of spontaneous mutations in the cell. Inactivation of the mutS gene of MMR in E. coli has been shown to accelerate the acquisition of resistance up to 900 fold relative to the WT when exposed to antibiotics such as rifampin and ciprofloxacin (Cirz & Romesberg, 2006). Indeed, loss of the MMR system is associated with the highly mutable state, called hypermutation. For example, P. aeruginosa CF isolates have been shown to lack the mutS and mutY gene of the MMR system leading to the hypermutatable phenotype (Oliver, et al., 2000). The mutS protein plays an important role in recognising mismatches to initiate the repair in association with other Mut proteins, the loss of this protein will completely arrest the MMR system. Therefore, the loss of a functional mutS was investigated here to determine the involvement of the mismatch repair system in generating variants and superinfective phage conversion. The results suggest that loss of a functional MMR system leads to early conversion of the superinfective phage and higher SCV formation. This supports the hypothesis that SNPs in the genome lead to the conversion of the superinfective phage and that this process is linked to DNA damage via oxidative stress and requires active MMR functions to reduce conversion to the superinfective phage.

The loss of RecA leads to a decrease of the superinfective Pf4 phage in the biofilm. This would imply that RecA is required for the conversion to the superinfective phage. RecA plays a significant role during the SOS response including facilitating recombinational repair and induction of autocleavage of the SOS repressor LexA, to activate transcription of the repressor gene. Given its multiple effects, RecA may either be required for its functions via the SOS response or for its role in recombination mediated repair which may also give rise to other mutations in the genome (Schlacher & Goodman, 2007). RecA plays an important role in the switch between lysogenic and lytic life cycle in lambda phage. Upon DNA damage, RecA becomes a highly specific protease, which cleaves the repressor C protein that has significant homology to LexA, and inactivates the repressor so it can no longer bind to the promoter region to prevent transcription of lytic phage. Therefore, RecA controls the expression of lytic 104 ! Chapter 4

phage (Ptashne, 2004). While there are some similarities in the regulation of phage activity via RecA, the observation that superinfective phage eventually appear in the PAO1 biofilms in this study suggest that the mechanism of superinfective Pf4 phage conversion is distinct from the induction of lytic lambda phage. However, RecA may play a minor role in the production of superinfective phage as a reduction in the superinfective phage production was shown in the recA mutant biofilm. The LexA repressor plays a crucial role in the phage lysogenic switch and interactions with RecA, therefore it would be of interest to investigate its role in superinfective phage conversion in P. aeruginosa.

P. aeruginosa generates more variants and superinfective phage when grown as a biofilm relative to planktonic cultures, therefore the physiological conditions encountered during biofilm growth and the associated mutants are an important feature of this life-style. The selection for mutations is a result of bacterial adaptation responding to environmental stress or changes in the environment. By introducing genetic changes, some cells may gain functions that are beneficial for survival such as antibiotic resistance. The loss of the MMR system and mitomycin C treatment were shown here to generate more variants at the early stages of the biofilm, on day 5 relative to the WT or untreated control. The effect of DNA damage and the acceleration of mutation rates, may be responsible for induction of the conversion into the superinfective phage. The formation of biofilm variants has been linked to the superinfective phage, where exposure of planktonic cultures to superinfective phage, but not WT phage directly led to the induction of SCVs (Rice, et al., 2009). Variant formation and bacteriophage activity have been observed in CF isolates which reflects on the importance of interconnection of these events in the biofilm. Similarly, variant formation was also induced by oxidative stress and nitroxidative stress in the biofilm. Genetic variants were observed from P. aeruginosa biofilms as a result of exposure to endogenous oxidative stress generated from dsDNA breaks repaired by mutagenic mechanisms involving recombinase DNA repair activity (Boles & Horswill, 2008). The endogenous oxidative stress promotes more DNA damage leading to more genetic diversity within the phage genome. And as a result, the selective pressure for variants carrying superinfective phage is selected for under phage infection and variants that are resistant against infection survive. 105 ! Chapter 4

P. aeruginosa biofilms have been shown to generate genetic diversity by producing phenotypic variants with a variety of functions such as the production of pyomelanin for protection against oxidative stress (Boles, et al., 2004), loss of flagellar and twitching motility for enhanced adherence to surfaces (Deziel, et al., 2001), and increased tolerance against antibiotic treatment (Drenkard & Ausubel, 2002). These adaptive phenotypes are important for the survival and the fitness of bacteria. However, the deep sequencing data, shown in Chapter 3, suggest that limited mutations occur in the host genome. Minimal mutations outside the phage genome imply that the superinfective phage conversion regulates the phenotypic variants observed from the biofilm through an as yet unknown mechanism. It is interesting to note that revertant variants from the PAO1 SCV2 biofilm, variants that lose the SCV colony morphology, also no longer carry the superinfective form of the phage. Variants are capable of switching phenotypes at a high frequency to adapt to changes in the environment, which has been shown amongst variants with reversible phenotypic variations when grown as a biofilm or in planktonic culture (Deziel, et al., 2001).

Less variant formation was observed at the early stages of the recA mutant biofilm (day 3 and 5), however during the late stages of the biofilm (day 7, 9 and 11), more variants were observed. These results agree with previous published results for the RecA mutant biofilms that showed they RecA mutant generated less genetic diversity and was less fit when exposed to stress (Boles, et al., 2004). In contrast, the results presented here showed that the number of variants in the recA mutant increased in the late stages of biofilm development (Boles, et al., 2004). While the primary differences between the work presented in this thesis and previously published work are related to the age of the biofilm sampled (11 d in this study compared to up to 5 d in previous reports), differences in the experimental system may have also contributed (e.g. drip-flow biofilms grown in trypticase soy broth at 37˚C compared to room temperature flow cell biofilms using M9 medium with glucose).

The conversion into the superinfective Pf4 phage coincides with the appearance of SCVs from the dispersal population of the PAO1 biofilm. In the work presented here, this process appears to be primarily a function of a functional MMR system

106 ! Chapter 4

and the oxidative stress response mediated by OxyR. Further, the results presented here suggest that high levels of either DNA damaging agents or reactive oxygen species can induce the superinfective phentoypes. Therefore, it is likely that during biofilm maturation, high concentrations of RONS accumulate due to endogenous metabolism, which overwhelms the ability of the host to detoxify those DNA damaging molecules. This leads to the accumulation of SNPs. The nucleotide composition of the repressor c gene may predispose it to acquire SNPs and such mutations disrupt or change the immunity function of the repressor C protein allowing the mutant phage to subsequently reinfect hosts with WT immunity functions. Alternatively, accumulation of RONS can cause cellular stress to signalling pathways, resulting in changes in the genome such as deletions as a consequence of recombination which may also result in the development of superinfective variants. This leads to strong selection pressure in the biofilm for variants that are resistant to the superinfection (and additionally carries the superinfective Pf4 phage) that subsequently persist.

To identify whether the repressor C or OxyR protein plays a role in Pf4 phage conversion, the location of the repressor c gene and promoter must be identified. More importantly, the connection between OxyR and Pf4 phage activity and what occurs at the molecular level that controls expression of superinfective Pf4 phage. The superinfective phage, through an unknown mechanism, also drives changes in the morphology of the superinfective host. Therefore, it is of interest to understand how superinfection results in morphotypic variation, the involvement of the repressor c gene and the mechanism by which OxyR controls superinfection.

107 ! Chapter 5

Chapter 5 – General Discussion

It has been estimated that up to 80% of chronic bacterial infections are biofilm- associated, and the challenge with biofilm-associated diseases is the resistance to host immune responses and chemotherapies (Davies, 2003). Therefore an improved understanding of the molecular mechanism of biofilm development could facilitate the discovery of novel treatments. Biofilms are highly resistant against a range of antimicrobial treatments and therefore recent studies have focused on the use of bacteriophage as strategy to specifically target and kill pathogenic bacteria. In addition to their use in treating infections, bacteriophage have been used for many applications including, phage display for the generation of antibodies, food decontamination and bio-detection (Gracia, et al., 2008).

Pf4 phage, a filamentous phage found in P. aeruginosa PAO1 has been shown to mediate cell death within microcolonies and dispersal events associated with biofilm development. In addition, the phage is associated with increased resilience of the biofilm when exposed to surfactant stress and is also important for virulence during lung infection (Rice, et al., 2009). Interestingly, the production of morphotypic variants and biofilm cell death also correlates with an increased phage titre in the biofilm effluent. Further, the increased phage titre appears to be specifically correlated with the sudden appearance of a form of the Pf4 phage that can reinfect the PAO1 host, which has been termed here as superinfection. Interestingly, it has been reported that the Pf4 phage genes were the most highly induced genes in the PAO1 biofilm relative to planktonic cells (Whiteley, et al., 2001). In those experiments, the biofilms were sampled after 5 days of development, a time at which superinfection is consistently observed in the experiments reported here. Thus, the high level of induced Pf4 genes observed to be expressed may contribute to genetic transfer and subsequently link to reinfection by the superinfective phage. This is supported by the sequence analysis of biofilm population reported in Chapter 3, where there was an increase in the number of phage particles relative to the PAO1 genome (from a ratio of prophage and free phage 1:0.2 on day 0 to a ratio of up to 1:6 on day 11 of the biofilm). The production of the superinfective phage has not been observed in

108 ! Chapter 5

planktonic cultures and therefore appears to be dependent on growth as a biofilm.

5.1 Phenotypic and genotypic changes influenced by superinfective Pf4 phage

The appearance of the superinfective phage appears to be restricted to the maturation stages of development (during microcolony formation or dispersal), as the superinfective phage does not appear before these events. This therefore suggests that genetic or physiological conditions that are specific to this stage of development are involved in the conversion to the superinfective phage. For example, comparison of planktonic cells and matured P. aeruginosa biofilm cells, showed up to 50% difference in proteome with up to 6 fold differences in gene expressions (Sauer, et al., 2002). Genetic and physiological changes of E. coli biofilm cells compared to planktonic cells also identified genes upregulated to have biofilm-specific functions such as adhesion, aggregation and stress-related resistance (Schembri, et al., 2003). Sequencing of the dispersal population did not reveal any mutations in such genes associated with biofilm development stages. In contrast, what was discovered, by both sequencing individual isolates that constitutively produced superinfective phage and the dispersal population, was a mutation in a small open reading frame with homology to repressor c gene. Interestingly, there were no mutations in the Pf4 phage genome.

The majority of genetic variation from the post-dispersal population of the PAO1 biofilm was within and upstream of the repressor c gene of the Pf4 phage genome. The deep sequencing results showed that the repressor c gene region had a high mutation frequency, up to 79% (Chapter 3). There were a total of 25 mutations detected in the host and phage genome, with 20 of those mutations in the repressor c gene. This strongly suggests that this gene is a hotspot for mutation during biofilm development and superinfection. This leads to the proposed idea that stress-induced mutations result in genotypic variants that harbour mutations around the repressor c gene, which leads to the superinfective phage conversion. It is interesting to note that no mutations were detected in oxidative stress genes or DNA repair genes. In combination with the general lack of mutations in the host genome, this suggests that the high frequency of variants and superinfective phage

109 ! Chapter 5

is not due to a hypermutator phenotype. These results strongly suggest that this genetic change observed in late biofilm cells are critical in understanding the variant formation from the biofilm and the impact on the phage.

Unlike filamentous phage such as fd and f1, the Pf4 phage exists as a lysogenic prophage and replicative form phage. As a lysogen, the secretion of the filamentous phage particles out of the host occurs through a process termed ‘budding’ and does not result in host cell death (Delbrock, 1946). The budding process of lytic phage was identified in Xanthomonas filamentous cf phage and it was noted that this is a new observation to the lifecycle of filamentous lytic phage (Kuo, et al., 1994). Budding occurs at the membrane surface, where major and minor coat proteins pIII, pVI, pVIII, pVII, and pIX facilitate the assembly of the phage particle. The phage-encoded pI/pXI inner membrane complex and the secretin protein pIV assess the release of the phage particle through the membrane. Since none of these phage genes were mutated in the superinfective phage, it is unlikely that this process of phage assembly is disrupted and hence cannot account for the observed cell death. Therefore, it was concluded that the core phage genes are not responsible for the appearance of superinfective phage in Xanthomonas spp.

Based on the discovery of mutations concentrated within the repressor c gene, it is hypothesised here that mutation of the repressor C protein will alter its RNA or DNA binding affinity. The loss of the repressor C protein may cause the loss of host immunity against phage infection leading to increased phage production and cell lysis. Mutations in the repressor C protein can interfere with RNA-RNA interaction and DNA-protein interactions that have been previously shown to contribute to the control of lysogenic and lytic conversion (Simons & Kleckner, 1988, Cheng, et al., 1999). Two repressor genes have been previously described in filamentous phage, the RstR repressor of V. cholerae CTX phage (McLeod, et al., 2005) and the repressor in Cf phage in X. campestris pv. citri (Cheng, et al., 1999). The phage-encoded RstR repressor plays a role in the control of replication and structure of CTX phage genes. SNPs in the immunity region of filamentous bacteriophage cf of Xanthomonas, were shown to be responsible for the loss of host immunity against phage infection (Cheng, et al., 1999). Moreover, mutations

110 ! Chapter 5

in the immunity region of the phage disrupted the RNA-RNA interactions also contributing to the loss of control over the immunity against infection mechanism (Simons & Kleckner, 1988, Cheng, et al., 1999). Other bacteriophage have also been shown to require the RNA-RNA interaction for controlling the process of selection between lysogenic and lytic pathways (Forti, et al., 1995, Sabbattini, et al., 1995). The cf bacteriophage and Pf4 phage share similar features with both lysogenic phage constitutively producing superinfective progeny phage which are released from the host leading to cell death. The mechanism of superinfective phage conversion for those phage was not identified, however the repressor c gene, and specific mutations with this gene, appears to play an important role in the conversion of the superinfection phenotype and the immunity of the host against phage infection. Interestingly, the promoter of the repressor C was also observed to carry mutations. It is possible that the changes in the conformation of the repressor C protein, due to nonsynonymous substitutions, will alter its promoter binding. Thus, by carrying a compensatory mutation in the promoter region, the protein-promoter pair still functions to control phage gene expression. This may explain why the variant SCV2 is resistant to reinfection by its own phage, while the phage that it produces can infect the wild type PAO1 host. Therefore, it is likely that a phage carrying a mutated repressor C protein or promoter region, evades the immunity of the host and causes superinfection.

Given that superinfection only occurs during maturation stages of biofilm development, it may be that the accumulation of specific metabolic by-products or nutrient gradients may be the trigger for the induction of superinfection. It has been shown that biofilms are stratified with respect to nutrient and oxygen gradients (De Beer, et al., 1994). Further, it has been shown that reactive oxygen or nitrogen species accumulates within microcolonies (Barraud, et al., 2006), suggesting that such compounds could induce biofilm specific mutations, such as those observed in the repressor c gene leading to superinfection. This is also supported by the observation that a lysine oxidase, which produced hydrogen peroxide as a bioproduct, was responsible for cell death within microcolonies and variant formation in Marinomonas mediterranea and Pseudoalteromonas tunicate (Mai-Prochnow, et al., 2008). Interestingly, Conibear et al. (2009) showed that mutations occur in the centre of the microcolonies of P. aeruginosa biofilms and 111 ! Chapter 5

this correlates with the cell death event of the biofilm (Conibear, et al., 2009). Collectively, these reports would suggest that oxidative stress compounds may accumulate within biofilms and microcolonies, leading to increased mutation in a subset of the biofilm population. Experiments preformed here, where addition of hydrogen peroxide to the biofilm, as well as the DNA damaging agent mitomycin C, resulted in early and elevated production of superinfective Pf4 phage. Stress- induced mutations have been commonly seen in a range of bacteria including E. coli (Bjedov, et al., 2003), P. putida (Karunakaran & Davies, 2000) and Mycobacterium smegmatis (Saumaa, et al., 2002). As the biofilm communities develop and proliferate, the by-products of cellular processes are trapped within the EPS matrix of the biofilm. This results in the build up of reactive oxygen species and reactive nitrogen species, and activates the oxidative stress response system. The activation of the oxidative stress response is initiated by OxyR, the global transcriptional regulatory that promotes the activation of defensive genes responsible for DNA repair and DNA mutagenesis (Storz & Imlay, 1999). Oxidative stress has shown to play a role in inducing the appearance of Pf4 superinfective phage and in the oxyR mutant biofilm, the number of superinfective phage is significantly increased relative to the WT biofilm (Chapter 4). Here, both internal and external factors of oxidative stress were examined and the loss of the OxyR protein function and hydrogen peroxide treatment also showed early induction of the conversion into the superinfective phage. These observations suggest that oxidative stress response plays a role in the conversion of the superinfective phage. Additionally, significantly more SCVs were observed from the oxyR mutant biofilm as compared to the WT biofilm. Previous work has demonstrated that endogenous oxidative stress is linked to generating diversity and variant formation in biofilms (Boles & Horswill, 2008). OxyR was identified to bind within the Pf4 phage genome specifically within the repressor C region, therefore it is possible that an interaction between OxyR binding with Pf4 phage may induce superinfection.

The infection of PAO1 with superinfective phage results in the production of morphotypic variants and deletion of the prophage results in biofilms that are deficient in variant formation, linking the phage to the biofilm-specific production of variants (Rice, et al., 2009). This may have significant implications for the 112 ! Chapter 5

resilience of the biofilm community, as noted above, the production of variants in a biofilm population has been linked to stress resistance (Boles & Horswill, 2008). Two morphotypic variants were characterised here, SCV2 and S4 variant, which carry the superinfective phage and are resistant against reinfection. These variants had altered colony morphologies as well as other phenotypes such as motility and biofilm formation. These altered phenotypes are thought to represent trade offs to allow them to persist during chronic infection. The formation of such biofilm- derived variants are clinically relevant as variants have been commonly isolated from the lungs of cystic fibrosis (CF) patients and these variants have been shown to exhibit higher tolerance against antimicrobial treatments (Deretic, et al., 1995, Haussler, et al., 2003, Kirov, et al., 2007). Phenotypic variants such as rough small colony variant of P. aeruginosa PA14 have altered biofilm formation and increased antibiotic resistance (Rahme, et al., 1995, Drenkard & Ausubel, 2002). Biofilm-derived variants also exhibit a range of specific traits such as swarming, attachment and biofilm formation (Koh, et al., 2007).

The emergence of variants from biofilms or from the lungs of CF patients has been linked to specific mutations. For example, mucoid variant formation in CF isolates of P. aeruginosa has been linked to mutations in the mucA gene (Boucher, et al., 1997) and SCV variant formation is associated with mutL, part of the mismatch repair system, mutations (Besier, et al., 2008). The mismatch repair system is very important for the correction of mismatched basepairs during DNA replication. The lack of a MutS protein in MMR system resulted in induced superinfective phage in the biofilm. RecA on the other hand, did not induce the conversion into the superinfective phage, which suggests a mechanism of repair that is involved in mutation processes. DNA damages to the biofilm induced by mitomycin C treatment also caused an early induction of superinfective phage. In results shown here, mutations were not widely detected in the host genome, this may rule out the possibility that superinfective phage induction is a result of mutator phenotypes (Chapter 3). Therefore, it is more likely that there is a stress that generates mutations specifically in the repressor C region and fails to be repaired by host repair systems. However, it is not known how mutation of the repressor C leads to altered host morphology. The biofilm community sequencing data indicated that while there were not many mutations outside the phage 113 ! Chapter 5

genome, mutations were observed in PilT and these have been associated with the retraction of pili and motility of the host. It remains to be determined if there is a link between mutation in the repressor C and the PilT. Alternatively, the repressor C may regulate other genes outside of the phage genome, where mutations of the repressor C, changing its DNA binding affinity, could alter gene expression in the host genome.

It is hypothesised that the activation of the conversion into the superinfective Pf4 phage during biofilm development is initiated by stress response. The primary response is to repair DNA damage that has occurred as a result of the stress encountered followed by the secondary response, to induce mutation for adaptation as a survival strategy. These mutations will be maintained if they are not immediately deleterious and could even have a selective advantage if the variants have an increased fitness such as increased oxidative stress resistance. Mutations identified amongst the biofilm population lie within the repressor C region of the Pf4 phage genome, signifying the hotspot for mutation driven by stress response. Therefore, the presence of superinfective phage within the biofilm population may drive selection of resistance against superinfection. Variants that are selected for in the presence of superinfective phage confer immunity against superinfection. In combination with the observation that phage have been found in the CF patients’ sputum (Hoiby, et al., 2001), the presence of phage plays an important role in selection in the biofilm.

5.2 Proposed models

Based on the observations reported here and elsewhere, changes in environmental conditions, stimulate mutations in the repressor c gene of Pf4 phage. In this scenario, variants that carry the superinfective phage are selected for in the PAO1 biofilm and these variants have a selective advantage, leading to their persistence in the biofilm. Here, three models are proposed for the control of the repressor C mutations and the associated superinfection (Figure 5-1 A-C). The first model is based on stress-induced mutations that result in SNPs within the repressor c gene. This can disrupt the binding between the repressor C protein and its promoter, which is involved in the regulation of phage production (Figure 5-1A). This results in no expression of the repressor c gene and subsequently over production 114 ! Chapter 5

of superinfective phage in the absence of a functional immunity system. Similarly, SNPs within the promoter site, which disrupts the binding between the repressor C protein to the promoter, can lead to over production of superinfective phage in the absence of the immunity system of the host (Figure 5-1B). For both models, mutations are recognised and repaired by a functional MMR system. It is also possible that superinfective variants consist of both mutations, thereby the mutation in the promoter region compensates or counteracts for the mutation within the repressor c gene leading to a binding of the protein to promoter that can still function to control phage gene expression. This can explain the dual characteristic of superinfective variants isolated here, that also exhibits superinfective phage and resistance against superinfection.

The third model incorporates the role of the OxyR transcriptional regulator of oxidative stress response. The stress-induced mutations also result in SNPs within the repressor c gene that affects the binding of the OxyR protein to the binding site within the repressor c gene (Figure 5-1C). The proposed model incorporates the activation of binding of the repressor C to the promoter via OxyR binding. When OxyR fails to bind to the binding site located within the repressor c gene, the repressor C fails to bind to the promoter thereby preventing expression of repressor C and resulting in the over production of superinfective phage. In all models, the prevention of expression of the repressor causes the loss of the host immunity system against phage infection. However, the function of the repressor in superinfection has not been described in filamentous phage.

The mechanism of superinfective phage infection leading to cell death is unclear. The Pf4 encoded toxin-antitoxin system may play a role in host cell death, however the deep sequencing results shown no mutations in the TA system of the Pf4 phage during biofilm development. It was noted that infected Xanthomonas cells release lytic cf phage during infection and the cell death is a result of an unknown mechanism (Kuo, et al., 1994). A proposed mechanism by (Russel, 1995) of the imbalance between the levels of the RF and ssDNA in the infected cell may influence the cell’s ability to control superinfective events. As previously noted, the regulation of balance between the level of RF and ssDNA is crucial for preventing host cell death. A lack of replication protein pII can prevent RF

115 ! Chapter 5

replication (Higashitani, et al., 1993) and similarly, a lack of the ssDNA-binding protein pV can prevent the assembly of new progeny phage (Michel & Zinder, 1989). When superinfective phage infects the cell, host enzymes are recruited to replicate the superinfective ssDNA into RF and produce newly synthesised superinfective ssDNA to be packaged into progeny phage. As the level of parental and superinfective RF increases in the cell, this may disrupt the balance between the level of RF and synthesised ssDNA. From the deep sequencing data in Chapter 3, the prophage to free phage ratio (of 1:0.2 on day 0) increased to 1:6 on day 11 of the biofilm. The increased number of phage particles during superinfection of the biofilm may be a result of higher copy number of RF per cell producing more phage particles. The increased copy numbers of RF caused by superinfective phage may disrupt the balance between RF and ssDNA-binding protein pV leading to cell death of the host.

5.3 Conclusions

Viruses are the most abundant organisms on the planet and have contributed significantly to bacterial adaptation and evolution through horizontal and vertical gene transfer as well as through direct selection effects. Amongst the roles of bacteriophage, the transmission of virulence genes is the best studied. For example, cholera toxin of V. cholerae CTX phage (Mukhopadhyay, et al., 2001) and Shiga-toxin of E. coli (O'Brien, et al., 1984) contribute to host pathogenicity causing diseases in humans. The role of bacteriophage during biofilm development has recently been discovered and more importantly their roles vary in different hosts. Pf4 filamentous phage mediates cell death in microcolonies of P. aeruginosa biofilms and generates variant formation that is observed in many bacterial biofilms (Rice, et al., 2009). On the other hand, Pf5 filamentous phage with high homology to Pf4 phage does not generate variants from its biofilm (Mooij, et al., 2007). The discrepancies in the role of phage between Pseudomonas Pf phage could be related to the genetic difference. For example, Pf4 carries a toxin-antitoxin system and putative reverse transcriptase gene that are unique to the Pf4 genome and may play a role in the host. There are many phage genes that encode proteins with unknown functions, although they may have homology to core phage genes or metabolic functions (Suttle, 2005). For

116 ! Chapter 5

example, in Pf4 phage genome, out of fifteen annotated open reading frames, two characterised proteins are the coaA and coaB coat proteins and the remaining proteins are uncharacterised. The majority of phage genomes are described as sharing homology to core genes of the Ff phage but the function of the putative proteins are yet to be identified. It is important to study their role in the host in order to further understand how the prophage may influence the host or contribute to its virulence. Consequently, this will provide more knowledge in phage infection and phage-host interactions in the environment.

Many phage genes have been shown to contribute to the pathogenicity and protection of the host. However, a large number of novel phage proteins and their involvement in the biology of its host remain unknown. For example, it is unknown which of the Pf4 phage genes is important for acute lung infection, but it is clear that the phage mutant is less virulent. Hence in recent years, significant research has been focused on the role of bacteriophage in host related infections. It is imperative to identify the mechanics behind phage driven biofilm formation in order to manipulate the process to modern medical needs. It is also clear that despite having been intensely studied for 50 years, there is yet much to be learned about these very filamentous phage and how they influence the behaviours of their host.

!

!

Figure'caption'for'the'following'page:'

Figure 5-1. The proposed mechanisms of the activation of superinfective Pf4 phage in P. aeruginosa PAO1. A) The activation is stress induced causing SNPs within the repressor c gene preventing the repressor C to bind to its promoter region. This results in the loss of immunity defense and over production of superinfective phage. B) The activation is stress induced causing SNPs within the promoter region, preventing the repressor C to bind to its promoter region. This results in the loss of immunity defense and over production of superinfective phage. C) The activation is stress induced causing SNPs within the repressor c gene disrupting the binding between OxyR and its binding site location within the repressor c gene resulting in the loss of immunity defense and over production of superinfective phage.

!

117 !

Chapter 5

118 !

Appendix

Appendix

!

Appendix Figure 1 The SNPs (grey box) within and upstream of the repressor c gene of the Pf4 phage genome. The sequence represents position 788542 to 788901 of the P. aeruginosa PAO1 genome. (*) indicates the start of the repressor c gene (267 bp) and ends at position 1. The arrow indicates the direction of the transcription.

119 ! References

References

Achouak W, Conrod S, Cohen V & Heulin T (2004) Phenotypic variation of Pseudomonas brassicacearum as a plant root-colonization strategy. Molecular Plant-Microbe Interactions 17: 872-879.

Ackermann HW (2003) Bacteriophage observations and evolution. Research in Microbiology 154: 245-251.

Ackermann HW & d'Hérelle F (1997) Découvreur des bactériophages. Medical Science 8: 3-6.

Addy HS, Askora A, Kawasaki T, Fujie M & Yamada T (2012) The filamentous phage ϕRSS1 enhances virulence of phytopathogenic Ralstonia solanacearum on tomato. Phytopathology 102: 244-251.

Allegrucci M & Sauer K (2007) Characterization of colony morphology variants isolated from Streptococcus pneumoniae biofilms. Journal of Bacteriology 189: 2030-2038.

Anwar H & Costerton JW (1990) Enhanced activity of combination of tobramycin and piperacillin for eradication of sessile biofilm cells of Pseudomonas aeruginosa. Antimicrobial Agents and Chemotherapy 34: 1666-1671.

Applegate DH & Bryers JD (1991) Effects on carbon and oxygen limitations and calcium concentrations on biofilm removal processes. Biotechnology and Bioengineering 37: 17-25.

Argyropoulos G, Brown AM & Garvey WT (1999) n-Butanol purification of dye terminator sequencing reactions. Biotechniques 26: 606-610.

Askora A, Kawasaki T, Fujie M & Yamada T (2011) Resolvase-like serine recombinase mediates integration/excision in the bacteriophage RSM. Journal of Bioscience and Bioengineering 111: 109-116.

Babior BM (1999) NADPH oxidase. Blood 93: 1464-1476.

Bailey S, Clokie MRJ, Millard A & Mann NH (2004) Cyanophage infection and photoinhibition in marine cyanobacteria. Research in Microbiology 155: 720-725.

Bais HB, Fall R & Vivianco JM (2004) Biocontrol of Bacillus subtillus against infection of Arabidopsis root by Pseudomonas syringae is facilated by biofilm formation and surfactin production. Plant Physiology 134: 307-319.

Banin E, Brady KM & Greenerg EP (2006) Chelator-induced dispersal and killing of Pseudomonas aeruginosa cells in a biofilm. Applied and Environmental Microbiology 72: 2064-2069.

Barondess JJ & Beckwith J (1990) A bacterial virulence determinant encoded by lysogenic coliphage lambda. Nature 346: 871-874. 120 ! References

Barraud N, Hassett DJ, Hwang SH, Rice SA, Kjelleberg S & Webb JS (2006) Involvement of nitric oxide in biofilm dispersal of Pseudomonas aeruginosa. Journal of Bacteriology 188: 7344-7353.

Bernhardt TG, Wang IN, Struck DK & Young R (2002) Breaking free: "protein antibiotics" and phage lysis. Research in Microbiology 153: 493-501.

Bertani G (1951) Studies on lysogenesis I: the mode of phage liberation by lysogenic Escherichia coli. Journal of Bacteriology 62: 293-300.

Besier S, Zander J, Kahl BC, Kraiczy P, Brade V & Wichelhaus TA (2008) The thymidine-dependent small colony variant phenotype is associated with hypermutability and antibiotic resistance in clinical Staphylococcus aureus isolates. Antimicrobial Agents and Chemotherapy 53: 2813-2819.

Bjarnsholt T, Kirketerp-Moller K, Jensen PO, Madsen KG, Phipps R, Krogfelt K, Hoiby N & Givskov M (2008) Why chronic wounds will not heal: a novel hypothesis. Wound Repair and Regeneration 16: 2-10.

Bjarnsholt T, Jensen PO, Fiandaca MJ, Pdersen J, Hansen CR, Anderson CB, Pressler T, Givskov M & Hoiby N (2009) Pseudomonas aeruginosa biofilms in the respiratory tract of cystic fibrosis patients. Pediatric Pulmonology 44: 547- 558.

Bjedov I, Tenaillon O, Gerard B, Souza V, Denamur E, Radman M, Taddei F & Matic I (2003) Stress-induced mutagenesis in bacteria. Science 300: 1404-1409.

Boles BR & Horswill AR (2008) Agr-mediated dispersal of Staphylococcus aureus biofilms. PLoS Pathogens 4: 1-13.

Boles BR, Thoendel M & SIngh PK (2004) Self-generated diversity produces "insurance effects" in biofilm communities. Proceedings of the National Academy of Sciences of the United States of America 101: 16630-16635.

Boucher RC, Yu H, Mudd MH & Deretic V (1997) Mucoid Pseudomonas aeruginosa in cystic fibrosis: characterization of muc mutations in clnical isolates and analysis of clearance in a mouse model of respiratory infection. Infection and Immunity 65: 3838-3846.

Boyce RP & Howard-Flanders P (1964) Release of ultraviolet light-induced thymine dimers from DNA in E. coli K-12. Biochemistry 51: 293-300.

Boyd EF & Brussow H (2002) Common themes among bacteriophage-encoded virulence factors and diversity among the bacteriophages involved. TRENDS in Microbiology 10: 521-529.

Boyd EF, Moyer KE, Shi L & Waldor MK (2000) Infectious CTX phage and the Vibrio pathogenicity island prophage in Vibrio mimicus: evidence for recent horizontal transfer between V. mimicus and V. cholerae. Infection and Immunity 68: 1507-1513.

121 ! References

Bradley DE (1973) The adsorption of the Pseudomonas aeruginosa filamentous bacteriophage Pf to its host. Journal of Microbiology 19: 623-631.

Bradley DE (1973) The length of the filamentous Pseudomonas aeruginosa bacteriophage Pf. Journal of General Virology 20: 249-252.

Bragonzi A, Paroni M, Nonis A, Cramer N, Montanari S, Rejman J, Serio CD, Doring G & Tummler B (2009) Pseudomonas aeruginosa microevolution during cystic fibrosis lung infection establishes clones with adapted virulence. American Journal of Respiratory and Critical Care Medicine 180: 138-145.

Brockhurst MA, Buckling A & Rainey PB (2005) The effect of a bacteriophage on diversification of the opportunistic bacterial pathogen, Pseudomonas aeruginosa. Proceeding of the Royal Society Biological Science 272: 1385-1391.

Brunder W, Schmidt H & Karch H (1996) Kat P, a novel catalase-peroxidase encoded by the large plasmid of enterohaemorrhagic Escherichia coli O157:H7. Microbiology 142: 3305-3315.

Brussow H & Hendrix R (2002) Phage genomics: small is beautiful. Cell 108: 13- 16.

Brussow H, Canchaya C & Hardt WD (2004) Phages and the evolution of bacterial pathogens: from genomic rearrangements to lysogenic conversion. Microbiology and Molecular Biology Reviews 68: 560-602.

Bura R, Cheung M, Liao B, Finlayson J, C. LB, Droppo IG, Leppard GG & Liss SN (1998) Composition of extracellular polymeric substances in the activated sludge floc matrix. Water Science and Technology 37: 325-333.

Buts L, Lah J, Dao-Thi MH, Wyns L & Loris R (2005) Toxin-antitoxin modules as bacteria metabolic stress managers. TRENDS in Biochemical Sciences 30: 672- 679.

Canchaya C, Fournous G & Brussow H (2004) The impact of prophages on bacterial chromosomes. Molecular Microbiology 53: 9-18.

Canchaya C, Proux C, Fournous G, Bruttin A & Brussow H (2003) Prophage genomics. Microbiology and Molecular Biology Reviews 67: 238-276.

Carlioz A & Touati D (1986) Isolation of superoxide dismutase mutants in Escherichia coli: is superoxide dismutase necessary for aerobic life? The EMBO Journal 5: 623-630.

Carlton RM (1999) Phage therapy: past history and future prospects. Archivum Immunologiae et Therapiae Experimentalis 47: 267-274.

Carpentier B & Cerf O (1993) Biofilms and their consequences, with particular reference to hygiene in the food industry. Journal of Applied Bacteriology 75: 499-511.

122 ! References

Castang S & Dove SL (2012) Basis for the essentiality of H-NS family members in Pseudomonas aeruginosa. Journal of Bacteriology 194: 5101-5109.

Centers for disease control and prevention (2002) Guidelines for prevention of intravascular catheter-related infections. MMWR 51: 1-32.

Chan RK & Botstein D (1972) Genetic of bacteriophage P22 I. Isolation of prophage deletions which affect immunity to superinfection. Virology 49: 257- 267.

Cheng CM, Wang HJ, Bau HJ & Kuo TT (1999) The primary immunity determinant in modulating the lysogenic immunity of the filamentous bacteriophage cf. Journal of Molecular Biology 287: 867-876.

Chmiel JF & Davis PB (2003) State of the Art: Why do the lungs of patients with cystic fibrosis become infected and why can't they clear the infection? Respiratory Research 4: 8.

Church D, Elsayed S, Reid O, Winston B & Linday R (2006) Burn wound infections. Clinical Microbiology Reviews 19: 403-434.

Ciofu O, Riis B, Pressler T, Poulsen HE & Hoiby N (2005) Occurrence of hypermutable Pseudomonas aeruginosa in cystic fibrosis patients is associated with the oxidative stress caused by chronic lung inflammation. Antimicrobial Agents and Chemotherapy 49: 2276-2282.

Cirz RT & Romesberg FE (2006) Induction and inhibition of ciprofloxacin resistance-conferring mutations in hypermutator bacteria. Antimicrobial Agents and Chemotherapy 50: 220-225.

Clasen JL, Brigden SM, Payet JP & Suttle CA (2008) Evidence that viral abundance across oceans and lakes is driven by different biological factors. Freshwater Biology 53: 1090-1100.

Coetzee J (1987) Phage Ecology. Wiley, New York.

Conibear TCR, Collins SL & Webb JS (2009) Role of mutation in Pseudomonas aeruginosa biofilm development. PLoS ONE 14: 6289.

Conlon KM, Humphreys H & O'Gara JP (2004) Inactivations of rsbU and sarA by IS256 represent novel mechanisms of biofilm phenotypic variation in Staphylococcus epidermidis. Journal of Bacteriology 186: 6208-6219.

Cook WL, Wachsmuth K, Johnson SR, Birkness KA & Samadi AR (1984) Persistence of plasmids, cholera toxin genes, and prophage DNA in classical Vibrio cholerae O1. Infection and Immunity 45: 222-226.

Costerton JW, Stewart PS & Greenberg EP (1999) Bacterial biofilms: a common cause of persistent infections. Science 284: 1318-1322.

Courcelle J, Belle JJ & Courcelle CT (2004) When replication travels on damaged templates: bumps and blocks in the road. Research in Microbiology 155: 231-237. 123 ! References

Courcelle J, Khodursky A, Peter B, Brown PO & Hanawalt PC (2001) Comparative gene expression profiles following UV exposure in wild-type and SOS-deficient Escherichia coli. Genetics 158: 41-64.

Couturier M, Bahassi EM & van Melderen L (1998) Bacterial death by DNA gyrase poisoning. TRENDS in Microbiology 6: 269-275.

Cox MM (2007) Regulation of bacterial RecA protein function. Critical Reviews in Biochemistry and Molecular Biology 42: 41-63.

Curtin JJ & Donlan RM (2006) Using bacteriophages to reduce formation of catheter-associated biofilms by Staphylococcus epidermidis. Antimicrobial Agents and Chemotherapy 50: 1268-1275.

D'Argenio DA, Calfee MW, Rainey PB & Pesci EC (2003) Autolysis and autoaggregation in Pseudomonas aeruginosa colony morphology variants. Journal of Bacteriology 184: 6481-6489. d'Hérelle F (1917) Sur un microbe invisible antagoniste des bacilles dysenteriques. Comptes rendus Academic Science Paris 165: 373-375.

Danovaro R, Dell'Anno A, Corinaldesi C, Magagnini M, Noble RT, Tambuini C & Winbauer M (2008) Major viral impact on the funtioning of benthic deep-sea ecosystems. Nature 454: 1084-1087.

Davies DG (2003) Understanding biofilm resistance to antibacterial agents. Nature Reviews Drug Discovery 2: 114-122.

Davies DG & Marques CNH (2009) A fatty acid messenger is responsible for inducing dispersion in microbial biofilms. Journal of Bacteriology 191: 1393- 1403.

Davies DG, Parsek MR, Pearson JP, Iglewski BH, Costerton JW & Greenberg EP (1998) The involvement of cell-to-cell signals in the development of a bacterial biofilm. Science 280: 295-298.

Davis BM & Waldor MK (2000) CTXphage contains a hybrid genome derived from tandemly integrated elements. Proceedings of the National Academy of Sciences of the United States of America 97: 8572-8577.

Davis BM & Waldor MK (2003) Filamentous phages linked to virulence of Vibrio cholerae. Current Opinion in Microbiology 6: 35-42.

Day LA, Marzec CJ, Reisberg SA & Casadevall A (1988) DNA packing in filamentous bacteriophages. Annual Review of Biophysics and Biophysical Chemistry 17: 509-539.

De Beer D, Stoodley P, Roe F & Lewandowski Z (1994) Effects of biofilm structures on oxygen distribution and mass transport. Biotechnology and Bioengineering 43: 1131-1138.

124 ! References

De Kievit TR, Gillis R, Marx S, Brown C & Iglewski BH (2001) Quorum-sensing genes in Pseudomonas aeruginosa biofilms: their role and expression patterns. Applied and Environmental Microbiology 67: 1865-1873.

Debarbieux L, Leduc D, Maura D, Morello E, Criscuolo A, Grossi O, Balloy V & Touqui L (2010) Bacteriophages can treat and prevent Pseudomonas aeruginosa lung infections. The Journal of Infectious Diseases 201: 1096-1104.

Delbrock M (1946) Bacterial viruses or bacteriophages. Biological Reviews 21: 30-40.

Derbise A, Chenal-Francisque V, Pouillot F, Fayolle C, Prevost MC, Medigue C, Hinnebusch BJ & Carniel E (2007) A horizontally acquired filamentous phage contributes to the pathogenicity of the plague bacillus. Molecular Microbiology 63: 1145-1157.

Deretic V, Schurr MJ & Yu H (1995) Pseudomonas aeruginosa, mucoidy and the chronic infection phenotype in cystic fibrosis. Microbiology 3: 351-356.

Destoumieux-Garzon D, Duquesne S, Peduzzi J, Goulard C, Desmadril M, Letellier L, Rebuffat S & Boulanger P (2005) The iron-siderophore transporter FhuA is the receptor for the antimicrobial peptide microcin J25: role of the microcin Val11-Pro16 β-hairpin region in the recognition mechanism. Biochemical Journal 389: 869-876.

Deziel E, Comeau Y & Vilemur R (2001) Initiation of biofilm formation by Pseudomonas aeruginosa 57RP correlates with emergence of hyperpiliated and highly adherent phenotypic variants deficient in swimming, swarming and twitching motilities. Journal of Bacteriology 183: 1195-1204.

Doggett RG, Harrison GM, Stillwell RN & Wallis ES (1966) An atypical Pseudomonas aeruginosa associated with cystic fibrosis of the pancreas. The Journal of Pediatrics 68: 215-221.

Donaldson SH, Bennett WD, Zeman KL, Knowles MR, Tarran R & Boucher RC (2006) Mucus clearance and lung function in cystic fibrosis with hypertonic saline. The New England Journal of Medicine 354: 241-250.

Donlan RM (2002) Microbial life on surfaces. Emerging Infectious Disease Journal 8: 881-890.

Donlan RM & Costerton JW (2002) Biofilms: survival mechanisms of clinically relevant microorganisms. Clinical Microbiology Reviews 15: 167-193.

Dow G, Browne A & SIbbald RG (1999) Infection in chronic wounds: controversies in diagnosis and treatment. Ostomy/Wound Manage 45: 23-40.

Dow JM, Crossman L, Fondlay K, He YQ, Feng JX & Tang JL (2003) Biofilm dispersal in Xanthomonas campestris is controlled by cell–cell signaling and is required for full virulence to plants. Proceedings of the National Academy of Sciences of the United States of America 100: 10995-11000.

125 ! References

Drenkard E & Ausubel FM (2002) Pseudomonas biofilm formation and antibiotic resistance are linked to phenotypic variation. Nature 416: 740-743.

Edwards R & Harding KG (2004) Bacteria and wound healing. Current Opinion in Infectious Disease 17: 91-96.

Eisenstark A (1967) Bacteriophage techniques: Methods in virology. Academic Press, New York, N. Y.

Elrod RP & Braun AC (1942) Pseudomonas aeruginosa: its role as a plant pathogen. Journal of Bacteriology 44: 633-644.

Farr SB, D'Ari R & Touati D (1986) Oxygen-dependent mutagenesis in Escherichia coli lacking superoxide dismutase. Proceedings of the National Academy of Sciences of the United States of America 83: 8268-8272.

Farrant JL, Sansone A, Canvin JR, Pallen MJ, Langford PR, Wallis TS, Dougan G & Kroll JS (1997) Bacterial copper- and zinc-cofactored superoxide dismutase contributes to the pathogenesis of system salmonellosis. Molecular Microbiology 25: 785-796.

Figueroa-Bossi N & Bossi L (1999) Inducible prophages contribute to Salmonella virulence in mice. Molecular Microbiology 33: 167-176.

Figueroa-Bossi N, Uzzau S, Maloriol D & Bossi L (2001) Variable assortment of prophages provides a transferable repertoire of pathogenic determinants in Salmonella. Molecular Microbiology 39: 260-272.

Finch JE, Prince J & Hawksworth M (1978) A bacteriological survey of the domestic environment. Journal of Applied Bacteriology 5: 357-364.

Flemming HC & Wingender J (2010) The biofilm matrix. Nature Reviews 8: 623- 633.

Forti F, Sabbattini P, Sironi G, Zangrossi S, Deho G & Ghisotti D (1995) Immunity determinant of phage-plasmid P4 is a short processed RNA. Journal of Molecular Biology 249: 869-878.

Fu WL, Forster T, Mayer O, Curtin JJ, Lehman SM & Donlan RM (2010) Bacteriophage cocktail for the prevention of biofilm formation by Pseudomonas aeruginosa on catheters in an in vitro model system. Antimicrobial Agents and Chemotherapy 54: 397-404.

Fuhrman JA (1999) Marine viruses and their biogeochemical and ecological effects. Nature 399: 541-548.

Fuhrman JA & Suttle CA (1993) Viruses in marine planktonic systems. Oceanography 6: 51-63.

Fuhrman JA & Noble RT (1995) Viruses and protists cause similar bacterial mortality in coastal seawater. Limnology and Oceanography 40: 1236-1242.

126 ! References

Gibson RL, Burns JL & Ramsey BW (2003) Pathophysiology and management of pulmonary infections in cystic fibrosis. American Journal of Respiratory and Critical Care Medicine 168: 918-951.

Gilligan PH (1991) Microbiology of airway disease in patients with cystic fibrosis. Clinical Microbiology Reviews 4: 35-51.

Gjermansen MP, Ragas PC, Sternberg C, Molin S & Tolker-Nielsen T (2005) Characterization of starvation-induced dispersion in Pseudomonas putida biofilms. Environmental Microbiology 7: 894-906.

Godde JS & Bickerton AJ (2006) The repetitive DNA elements called CRISPRs and their associated genes: evidence of horizontal transfer among prokaryotes. Journal of Molecular Evolution 62: 718-729.

Goff SP (1990) Retroviral reverse transcriptase: synthesis, structure and function. Journal of Acquired Immune Deficiency Syndromes 3: 817-831.

Govan JRW & Deretic V (1996) Microbial pathogenesis in cystic fibrosis: mucoid Pseudomonas aeruginosa and Burkholderia cepacia. Microbiology 60: 539-574.

Gracia P, Madera C, Martinez B & Rodriguez A (2007) Biocontrol of Staphylococcus aureus in curd manufacturing processes using bacteriophages. International Dairy Journal 17: 1232-1239.

Gracia P, Martinez B, Obeso JM & Rodriguez A (2008) Bacteriophages and their application in food safety. Letters in Applied Microbiology 47: 479-485.

Gracia P, Martinez B, Rodriguez L & Rodriguez A (2010) Synergy between the phage endolysin LysH5 and nisin to kill Staphylococcus aureus in pasteurized milk. International Journal of Food Microbiology 141: 151-155.

Grau BL, Henk MC & Pettis GS (2005) High-frequency phase variation of Vibrio vulnificus 1003: isolation and characterisation of a rugose phenotypic variant. Journal of Bacteriology 187: 2519-2525.

Green SK, Schoroth MN, Cho JJ, Kominos SD & Vitanza-Jack VB (1974) Agricultural plants and soil as a reservoir for Pseudomonas aeruginosa. Applied and Environmental Microbiology 28: 987-991.

Grissa I, Vergnaud G & Pourcel C (2007) The CRISPRdb database and tools to display CRISPRs and to generate dictionaries of spacers and repeats. BMC Bioinformatics 8: 172.

Grobe S, Wingender J & Truper HG (1995) Characterization of mucoid Pseudomonas aeruginosa strains isolated from technical water systems. Journal of Applied Bacteriology 79: 94-102.

Guerrero-Ferreira RC, Viollierb PH, Elyc B, Poindexterd S, Georgievae M, Jensenf GJ & Wrighta ER (2011) Alternative mechanism for bacteriophage adsorption to the motile bacterium Caulobacter crescentus. Proceedings of the National Academy of Sciences of the United States of America 108: 9963-9968. 127 ! References

Gutierrez O, Juan C, Perez JL & Oilver A (2004) Lack of association between hypermutation and antibiotic resistance development in Pseudomoans aeruginosa isolates from intensive care unit patients. Antimicrobial Agents and Chemotherapy 48: 3573-3575.

Hall-Stoodley L, Hu FZ, Gieseke A, Nistico L, Nguyen D, Hayes J, Forbes M, Greenberg DP, Dice B, Burrows A, Wackym A, Stoodley P, Post JC, Ehrlich GD & Kerschner JE (2006) Direct detection of bacterial biofilms on the middle-ear mucosa of children with chronic otitis media. The Journal of the American Medical Association 296: 202-211.

Hambly E & Suttle CA (2005) The viriosphere, diversity and genetic exchange within phage communities. Current Opinion in Microbiology 8: 444-450.

Hammond AA, Miller KG, Kruczek CJ, Dertien J, Colmer-Hamood JA, Griswold JA, Horswill AR & Hamood AN (2011) An in vitro biofilm model to examine the effect of antibiotic ointments on biofilms produced by burn wound bacterial isolates. Burns 37: 312-321.

Hancock RE, Mutharia LM, Chan L, Darveau RP, Speert DP & Pier GB (1983) Pseudomonas aeruginosa isolates from patients with cystic fibrosis: a class of serum-sensitive, nontypable strains deficient in lipopolysaccharide O side chains. Infection and Immunity 42: 170-177.

Harris AW, Mount DWA, Fuerst CR & Siminovitch L (1967) Mutations in bacteriophage lambda affecting host cell lysis. Virology 32: 553-569.

Hassett DJ, Alsabbagh E, Parvatiyar K, Howell ML, Wilmott RW & Ochsner UA (2000) A protease-resistant catalase, kat A, released upon cell lysis during stationary phase is essential for aerobic survival of Pseudomonas aeruginosa oxyR mutant at low cell densities. Journal of Bacteriology 182: 4557-4563.

Hasty J, McMillen D, Isaacs F & Collins J (2001) Computational studies of gene regulatory networks: in numero molecular biology. Nature Reviews in Genetics 2: 268-279.

Hatfull GF (2008) Bacteriophage genomics. Current Opinion in Microbiology 11: 447-453.

Haussler S, Tummler B, Weissbrodt H, Rohde M & Steinmetz I (1999) Small- colony variants of Pseudomonas aeruginosa in cystic fibrosis. Clinical Infectious Diseases 29: 621-625.

Haussler S, Ziegler I, Lottel A, Gotz F, Rohde M, Wehmhohner D, Saravanamuthu S, Tummler B & Steinmetz I (2003) Highly adherent small- colony variants of Pseudomonas aeruginosa in cystic fibrosis lung infection. Journal of Medical Microbiology 52: 295-301.

128 ! References

Hayashi T, Makino K, Ohnishi Y, Kurokawa K, Ishii K, Yokoyama K, Han CG, Ohysubo E, Nakayama K, Murata T, Tanaka M, Tobe T, Iida T, Takami H, Honda T, Sasakawa C, Ogasawara N, Yasunaga T, Kuhara S, Shiba T, Hattori M & Shinagama H (2001) Complete genome sequence of Enterohemorrhagic Escherichia coli O157:H7 and genomic comparison with a laboratory strain. DNA Research 8: 11-22.

Henderson IR, Owen P & Nataro JP (1999) Molecular switches - the on and off of bacterial phase variation. Molecular Microbiology 33: 919-932.

Hendrix R, Roberts J, Stahl F & Weisberg R (1983) Lambda II. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.

Hendrix R, Smith MCM, Burns RN, Ford ME & Hatfull GF (1999) Evolutionary relationships among diverse bacteriophages and prophages: all the world's a phage. Proceedings of the National Academy of Sciences of the United States of America 96: 2192-2197.

Hendrix RW (2003) Bacteriophage genomics. Current Opinion in Microbiology 6: 506-511.

Henle ES & Linn S (1997) Formation, prevention, and repair of DNA damage by iron/hydrogen peroxide. Journal of Biological Chemistry 272: 19095-19098.

Higashitani N, Higashitani A & Horiuchi K (1993) Nucleotide sequence of the primer RNA for DNA replication of filamentous bacteriophages. Journal of Virology 67: 2175-2182.

Hill DF, Short NJ, Perham RN & Petersen GB (1991) DNA sequence of the filamentous bacteriophage Pf1. Journal of Molecular Biology 218: 349-364.

Hogardt M, Schubert S, Adler K, Gotzfriend M & Heesemann J (2006) Sequence variability and functional amalysis of MutS of hypermutable Pseudomonas aeruginosa cystic fibrosis isolates. International Journal of Medical Microbiology 296: 313-320.

Hohn B, von Schutz H & Marvin DA (1971) Filamentous bacterial virus II. Killing of bacteria by absorptive infection with fd. Journal of Molecular Biology 56: 155-165.

Hoiby N, Bjarnsholt T, Givskov M, Molin S & Ciofu O (2010) Antibiotic resistance of bacterial biofilms. International Journal of Antimicrobial Agents 35: 322-332.

Hoiby N, Johansen HK, Moser C, Song ZJ, Ciofu O & Kharazmi A (2001) Pseudomonas aeruginosa and the in vitro and in vivo biofilm mode of growth. Microbes and Infection 3: 23-35.

Horabin J & Websster R (1986) Morphogenesis of fl filamentous phage Increased expression of bacteriophage gene 1 inhibits bacterial growth. Journal of Molecular Biology 188: 403-413.

129 ! References

Horst JP, Wu TH & Marinus MG (1999) Escherichia coli mutator genes. TRENDS in Microbiology 7: 29-36.

Howard L & Tipper DJ (1973) A polypeptide bacteriophage receptor: modified cell wall protein subunits in bacteriophage-resistant mutants of Bacillus sphaericus strain P-1. Journal of Bacteriology 113: 1491-1504.

Huang A, Friesen J & Brunton JL (1987) Characterization of a bacteriophage that carries the genes for production of Shiga-like toxin 1 in Escherichia coli. Journal of Bacteriology 169: 4308-4312.

Hubor KE & Waldor MK (2002) Filamentous phage integration requires the host recombinases XerC and XerD. Nature 417: 656-659.

Hunt SM, Werner EM, Huang B, Hamilton MA & Stewart PS (2004) Hypothesis for the role of nutrient starvation in biofilm detachment. Applied and Environmental Microbiology 70: 7418-7425.

Huynh TT, McDougald D, Klebensberger J, Al Qarni B, Barraud N, Rice SA, Kjelleberg S & Schleneck D (2012) Glucose starvation-induced dispersal of Pseudomonas aeruginosa biofilms is cAMP and energy dependent. PLoS ONE 7: e42874.

Hyman P & Adbedon ST (2010) Bacteriophage host range and bacterial resistance. Advanced Applied Microbiology 70: 217-248.

Inal JM (2003) Phage therapy: a reappraisal of bacteriophages as antibiotics. Archivum Immunologiae et Therapiae Experimentalis 51: 237-244.

Inouye S, Sunshone MG, Six EW & Inouye M (1991) Retronphage phi R73: an E. coli phage that contains a retroelement and integrates into a tRNA gene. Science 252: 969-971.

Ishino Y, Shinagawa H, Makino K, Amemura M & Nakata A (1987) Nucleotide sequence of the iap gene, responsible for alkaline phosphatase isozyme conversion in Escherichia coli, and identification of the gene product. Journal of Bacteriology 169: 5429-5433.

Jackson DW, Suzuki K, Oakford L, Simecha JW, Hart ME & Romeo T (2002) Biofilm formation and dispersal under the influence of the global regulator CsrA of Escherichia coli. Journal of Bacteriology 184: 290-301.

Jacobs MA, Alwood A, Thaipisuttikui I, Spencers D, Haugen E, Ernst S, Bovee D, Olson MV & Manoil C (2003) Comprehensive transposon mutant library of Pseudomonas aeruginosa. Proceedings of the National Academy of Sciences of the United States of America 100: 14339-14344.

James GA, Swogger EBS, Wolcott R, deLancery Pulcini E, Secor PBS, Sestrich JBS, Costerton JW & Stewart PS (2007) Biofilms in chronic wounds. Wound Repair and Regeneration 16: 37-44.

130 ! References

Janssen RT, van der Straaten T, van Diepen A & van Dissel JT (2003) Responses to reactive oxygen intermediates and virulence of Salmonella typhimurium. Microbes and Infection 5: 527-534.

Jiang Y, Pogliano J, Helinski DR & Konieczny I (2002) ParE toxin encoded by the broad-host-range plasmid RK2 is an inhibitor of Escherichia coli gyrase. Molecular Microbiology 44: 971-979.

Jorgensen F, Bally M, Chapon-Herve V, Michel G, Lazdunski A, WIlliams P & Stewart GS (1999) RpoS-dependent stress tolerance in Pseudomonas aeruginosa. Microbiology 145: 835-844.

Kang SM, Kishimoto M, Shioya S, Yoshida T, Suga KI & Taguchi H (1989) Dewatering characteristics of activated sludges and effect of extracellular polymer. Journal of Fermentation and Bioengineering 68: 117-122.

Kaplan JB, Ragunath C, Ramasubbu N & Fine DH (2003) Detachment of Actinobacillus actinomycetemcomitans biofilm cells by an endogenous β- hexosaminidase activity. Journal of Bacteriology 185: 4693-4698.

Karaolis DKR, Somara S, Maneval Jr DR, Johnson JA & Kaper JB (1999) A bacteriophage encoding a pathogenicity island, a type-IV pilus and a phage receptor in cholera bacteria. Nature 399: 375-379.

Karatzas KA, Zervos A, Tassou CC, Mallidis CG & Humphrey TJ (2007) Piezotolerant small-colony variants with increased thermotolerance, antibiotic susceptibility, and low invasiveness in a clonal Staphylococcus aureus population. Applied and Environmental Microbiology 73: 1873-1881.

Karunakaran P & Davies J (2000) Genetic antagonism and hypermutability in Mycobacterium smegmatis. Journal of Bacteriology 182: 3331-3335.

Kerem E, Reisman J, Corey M, Canny GJ & Levison H (1992) Prediction of mortality in patients with cystic fibrosis. The New England Journal of Medicine 326: 1187-1191.

Kirisits MJ, Prost L, Starkey M & Parsek MR (2005) Characterisation of colony morphology variants isolated from Pseudomonas aeruginosa biofilms. Applied and Environmental Microbiology 71: 4809-4821.

Kirov SM, Webb JS, O'May CY, Reid DW, Woo JKK, Rice SA & Kjelleberg S (2007) Biofilm differentiation and dispersal in mucoid Pseudomonas aeruginosa isolates from patients with cystic fibrosis. Microbiology 153: 3264-3274.

Klausen M, Heydorn A, Ragas P, Lambertsen L, Aaes-Jorgensen A, Molin S & Tolker-Nielsen T (2003) Biofilm formation by Pseudomonas aeruginosa wild type, flagella and type IV pili mutants. Molecular Ecology 48: 1511-1524.

131 ! References

Klockgether J, Munder A, Neugebauer J, Davenport CF, Stanke F, Larbig KD, Heeb S, Schock U, Pohl TM, Wiehlmann L & Tummler B (2010) Genome diversity of Pseudomonas aeruginosa PAO1 laboratory strains. Journal of Bacteriology 192: 1113-1121.

Klotz MG & Hutcheson SW (1992) Multiple periplasmic catalases in phytopathogenic strains of Pseudomonas syringae. Applied and Environmental Microbiology 58: 2468-2473.

Koh KS, Lam KW, Alhede M, Queck SY, Labbate M, Kjelleberg S & Rice SA (2007) Phenotypic diversification and adaptation of Serratia marcescens MG1 biofilm-derived morphotypes. Journal of Bacteriology 189: 119-130.

Koh KS, Matz C, Tan CH, Le HL, Rice SA, Marshall DJ, Steinberg PD & Kjelleberg S (2012) Minimal increase in genetic diversity enhances predation resistance. Molecular Ecology 21: 1741-1753.

Kovach ME, Shaffer MD & Peterson KM (1996) A putative integrase gene defines the distal end of a large cluster of ToxR-regulated colonization genes in Vibrio cholerae. Microbiology 142: 2165-2174.

Kunin V, Sorek R & Hugenholtz P (2007) Evolutionary conservation of sequence and secondary structures in CRISPR repeats. Genome Biology 8: R61.

Kuo TT, Chiang CC, Chen SY, Lin JH & Kuo JL (1994) A long lytic cycle in filamentous phage Cf1tv infecting Xanthomonas campestris pv. citri. Archives of Virology 135: 253-264.

Kutter E & Sulakvelidze A (2005) Bacteriophage biology and applications. CRC Press, UK.

Ladero V, Garcia P, Bascaran V, Herrero M, Alvarez MA & Suarez JE (1998) Identification of the repressor-encoding gene of the Lactobacillus bacteriophage A2. Journal of Bacteriology 180: 3474-3476.

Lawrence JG, Hatfull GF & Hendrix RW (2002) Imbroglios of viral taxonomy: genetic exchange and failings of phenetic approaches. Journal of Bacteriology 184: 4891-4905.

LeClerc JE, Li B, Payne WL & Cebula TA (1996) High mutation frequencies among Escherichia coli and Salmonella pathogens. Science 274: 1208-1211.

LeClerc JE, Payne WL, Kupchelle E & Cebula TA (1998) Detection of mutator subpopulations in Salmonella typhimurium LT2 by reversion of his alleles. Mutation Research 400: 89-97.

Lesic B & Rahme LG (2008) Use of the lambda Red recombinase system to rapidly generate mutants in Pseudomonas aeruginosa. BMC Molecular Biology 9: 1-9.

132 ! References

Levin BR & Bull JJ (2004) Population and evolutionary dynamics of phage therapy. Nature Reviews Microbiology 2: 166-173.

Levine M & Curtiss R (1961) Genetic fine structure of the C region and the linkage map of phage P22. Genetics 46: 1573-1580.

Lewis K (2001) Riddle of biofilm resistance. Antimicrobial Agents and Chemotherapy 45: 999-1007.

Ljungquist E, Kockum K & Bertani E (1984) DNA sequences of the repressor gene and operator region of bacteriophage P2. Proceedings of the National Academy of Sciences of the United States of America 81: 3988-3992.

Loc Carrilo C, Atterbury RJ, El-Shibiny A, Connerton PL, Dillon E, Scott A & Connerton IF (2005) Bacteriophage therapy to reduce Campylobacter jejuni colonization of broiler chickens. Applied and Environmental Microbiology 71: 6554-6563.

Los JM, Los M, Wegrzyn A & Wegrzyn G (2010) Hydrogen peroxide-mediated induction of the Shiga toxin-converting lambdoid prophage ST2-8624 in Escherichia coli O157:H7. FEMS Immunology & Medical Microbiology 58: 322- 329.

Love PE & Yasbin RE (1986) Induction of the Bacillus subtilis SOS-like response by Escherichia coli RecA protein. Proceedings of the National Academy of Sciences of the United States of America 83: 5204-5208.

Lu MJ & Henning U (1994) Superinfection exclusion by T-even-type coliphages. TRENDS in Microbiology 2: 137-139.

Lucas-Elio P, Hernandez P, Sanchez-Amat A & Solano F (2005) Purification and partial characterization of marinocine, a new broad-spectrum antibacterial protein produced by Marinomonas mediterranea. Biochimica et Biophysica Acta 1721: 193-203.

Luiten RGM, Putterman DG, Schoenmakers JGG, Konings RNH & Day LA (1995) Nucleotide sequence of the genome of Pf3, an incP-1 plasmid-specific filamentous bacteriophage of Pseudomonas aeruginosa. Journal of Virology 56: 268-276.

Luzar MA, Thomassen MJ & Montie TC (1985) Flagella and motility alterations in Pseudomonas aeruginosa strains from patients with cystic fibrosis: relationship to patient clinical condition. Infection and Immunity 50: 577-582.

Lyczak JB, Cannon CL & Pier GB (2002) Lung infections associated with cystic fibrosis. Clinical Microbiology Reviews 15: 194-222.

Lynch KH, Seed KD, Stothard P & Dennis JJ (2010) Inactivation of Burkholderia cepacia complex phage KS9 gp41 identifies the phage repressor and generates lytic virions. Journal of Virology 84: 1276-1288.

133 ! References

Ma L, Jackson KD, Landry RM, Parsek MR & Wozniak DJ (2006) Analysis of Pseudomonas aeruginosa conditional psl variants reveals roles for the psl polysaccharide in adhesion and maintaining biofilm structure postattachment. Journal of Bacteriology 188: 8213-8221.

Ma M & Eaton JW (1992) Multicellular oxidant defense in unicellular organisms. Proceedings of the National Academy of Sciences of the United States of America 89: 7924-7928.

Macia MD, Blanquer D, Togores B, Sauleda J, Perez JL & Oilver A (2005) Hypermutation is a key factor in development of multiple-antimicrobial resistance in Pseudomonas aeruginosa strains causing chronic lung infection. Antimicrobial Agents and Chemotherapy 49: 3382-3386.

Mahenthiralingam E, Campbell ME & Speert DP (1994) Nonmotility and phagocytic resistance of Pseudomonas aeruginosa isolates from chronically colonized patients with cystic fibrosis. Infection and Immunity 62: 596-605.

Mai-Prochnow A, Webb JS, Ferrari BC & Kjelleberg S (2006) Ecological advantages of autolysis during the development and dispersal of Pseudoalteromonas tunicata biofilms. Applied and Environmental Microbiology 72: 5414-5420.

Mai-Prochnow A, Lucas-Elio P, Egan S, Thomas T, Webb JS, Sanchez-Amat A & Kjelleberg S (2008) Hydrogen peroxide linked to lysine oxidise activity facilitates biofilm differentiation and dispersal in several Gram-negative bacteria. Journal of Bacteriology 190: 5493-5501.

Mao EF, Lane L, Lee JW & Miller JH (1997) Proliferation of mutators in a cell population. Journal of Bacteriology 179: 417-422.

Martin B, Garcia P, Castanie MP & Claverys JP (1995) The recA gene of Streptococcus pneumoniae is part of a competence-induced operon and controls lysogenic induction. Molecular Microbiology 15: 367-379.

Martin DR (1973) Mucoid variation in Pseudomonas aeruginosa induced by the action of phage. Journal of Medical Microbiology 6: 111-118.

Marvin DA (1998) Filamentous phage structure, infection and assembly. Current Opinion in Structural Biology 8: 150-158.

Marvin DA & Hohn B (1969) Filamentous bacterial viruses. Bacteriological Reviews 33: 172-209.

Massie HR, Samis HV & Baird MB (1972) The kinetics of degradation of DNA and RNA by hydrogen peroxide. Biochimica et Biophysica Acta (BBA) - Nucleic Acids and Protein Synthesis 272: 539-548.

134 ! References

Mathee K, Ciofu O, Sternberg C, Lindum PW, Campbell JIA, Jensen PO, Johnsen AH, Givskov M, Ohman DE, Soren M, Hoiby N & Kharazmi A (1999) Mucoid conversion of Pseudomonas aeruginosa by hydrogen peroxide: a mechanism for virulence activation in the cystic fibrosis lung. Microbiology 145: 1349-1357.

Matic I, Rayssiguier C & Radman M (1995) Interspecies gene exchange in bacteria: the role of SOS and mismatch repair systems in evolution of species. Cell 80: 507-515.

Matic I, Radman M, Taddei F, Picard B, Doit C, Bingen E, Denamur E & Elion J (1997) High variable mutation rates in commensal and pathogenic Escherichia coli. Science 277: 1833-1834.

Matsushiro A, Sato K, Miyamoto H, Yamamura T & Honda T (1999) Induction of prophages of Enterohemorrhagic Escherichia coli O157:H7 with norfloxacin. Journal of Bacteriology 181: 2257-2260.

McAuliffe O, Ross RP & Fitzgerald GF (2007) Chapter 1 The new phage biology: from genomic to applications. Caister Academic Press, Norfolk, UK.

McCann KS (2000) The diversity-stability debate. Nature 405: 228-233.

McDougald D, Rice SA, Barraud N, Steinberg PD & Kjelleberg S (2012) Should we stay or should we go: mechanisms and ecological consequences for biofilm dispersal. Nature Reviews Microbiology 10: 39-50.

McLeod SM, Kimsey HH, Davis BM & Waldor MK (2005) CTX and Vibrio cholerae: exploring a newly recognized type of phage-host cell relationship. Molecular Microbiology 57: 347-356.

McVay CS, Velasquez M & Fralick JA (2007) Phage therapy of Pseudomonas aeruginosa infection in a mouse burn wound model. Antimicrobial Agents and Chemotherapy 51: 1934-1938.

Mead PS & Griffin PM (1998) Escherichia coli O157:H7. Lancet 352: 1207- 1212.

Mena A, Smith EE, Burns JL, Speert DP, Moskowitz SM, Perez JL & Oilver A (2008) Genetic adaptation of Pseudomonas aeruginosa to the airways of cystic fibrosis patients is catalyzed by hypermutation. Journal of Bacteriology 190: 7910-7917.

Messing J, Gronenborn B, Muller-Hill B & Hofschneider PH (1977) Filamentous coliphage M13 as a cloning vehicle: insertion of a HindII fragment of the lac regulatory region in M13 replicative form in vitro. Proceedings of the National Academy of Sciences of the United States of America 74: 3642-3646.

Michel B & Zinder ND (1989) In vitro binding of the bacteriophage f1 gene V protein to the gene II RNA-operator and its DNA analog. Nucleic Acids Research 17: 7333-7344.

135 ! References

Middelboe M, Jorgensen N & Kroer N (1996) Effects of viruses on nutrient turnover and growth efficiency of noninfected marine bacterioplankton. Applied and Environmental Microbiology 62: 1991-1997.

Mila CE & Warwick WJ (1998) Risk of death in cystic fibrosis patients with severely compromised lung function. CHEST 113: 1230-1234.

Miller JH (1996) Spontaneous mutators in bacteria: insights into pathways of mutagenesis and repair. Annual Review in Microbiology 50: 625-643.

Miller RA & Britigan BE (1997) Role of oxidants in microbial pathophysiology. Clinical Microbiology Reviews 10: 1-18.

Minamishima Y, Takeya K, Ohnishi Y & Amako K (1968) Physicochemical and biological properties of fibrous Pseudomonas bacteriophages. Journal of Virology 2: 208-213.

Mirold S, Rabsch W, Tschape H & Hardt WD (2001) Transfer of the Salmonella type III effector sopE between unrelated phage families. Journal of Molecular Biology 312: 7-16.

Model P & Russel M Chapter 6: Filamentous bacteriophage. The Rockefeller University, New York.

Modi R, Hirvi Y, Hill A & Griffiths MW (2001) Effect of phage on survival of Salmonella enteritidis during manufacture and storage of Cheddar cheese made from raw and pasteurized milk. Journal of Food Protection 64: 927-933.

Modrich P (1991) Mechanisms and biological effects of mismatch repair. Annual Review in Genetics 25: 229-253.

Modrich P & Lahue R (1996) Mismatch repair in replication fidelity, genetic recombination, and cancer biology. Annual Review of Biochemistry 65: 101-133.

Moller SC, Sternberg C, Andersen JB, Christensen BB, Ramos JL, Givskov M & Molin S (1998) In situ gene expression in mixed culture biofilms: evidence of metabolic interactions between community members. Applied and Environmental Microbiology 64: 721-732.

Monroe D (2007) Looking for chinks in the armor of bacterial biofilms. . PLoS Biology 5: e307.

Mooij MJ, Drenkard E, Llamas A, Vandenbroucke-Grauls CMJE, Savelkoul PHM, Ausubel FM & Bitter W (2007) Characterization of the integrated filamentous phage Pf5 and its involvement in small-colony formation. Microbiology 153: 1790-1798.

Morgan R, Kohn S, Hwang SH, Hassett DJ & Sauer K (2006) BdlA, a chemotaxis regulator essential for biofilm dispersion in Pseudomonas aeruginosa. Journal of Bacteriology 188: 7335-7343.

136 ! References

Moskowitz SM, Foster JM, Emerson J & Burns JL (2004) Clinically feasible biofilm susceptibility assay for isolates of Pseudomonas aeruginosa from patients with cystic fibrosis. Journal of Clinical Microbiology 42: 1915-1922.

Mukhopadhyay AK, Chakraborty S, Takeda Y, Nair B & Berg DE (2001) Characterization of VPI pathogenicity island and CTX prophage in environmental strains of Vibrio cholerae. Journal of Bacteriology 183: 4737 - 4746.

Mutoh N, Furukawa H & Mizushima S (1978) Role of lipopolysaccharide and outer membrane protein of Escherichia coli K-12 in the receptor activity for bacteriophage T4. Journal of Bacteriology 136: 693-699.

Nakayama K, Kanaya S, Ohnishi Y, Terawaki Y & Hayashi T (1999) The complete nucleotide sequence of φCTX, a cytotoxin-converting phage of Pseudomonas aeruginosa: implications for phage evolution and horizontal gene transfer via bacteriophages. Molecular Microbiology 31: 399-419.

O'Brien AD, Newland JW, Miller SF, Holmes RK, Smith HW & Formal SB (1984) Shiga-like toxin-converting phages from Escherichia coli strains that cause hemorrhagic colitis or infantile diarrhea. Science 226: 694-696.

O'Flynn G, Ross RP, Fitzgerald GF & Coffey A (2004) Evaluation of a cocktail of three bacteriophages for biocontrol of Escherichia coli O157:H7. Applied and Environmental Microbiology 70: 3417-3421.

O'Toole GA & Kolter R (1998) Flagella and twitching motility are necessary for Pseudomonas aeruginosa biofilm development. Molecular Ecology 30: 295-304.

Ochsner UA, Vasil ML, Alsabbagh E, Parvatiyar K & Hassett DJ (2000) Role of the Pseudomonas aeruginosa oxyR-recG operon in oxidative stress defense and DNA repair: oxyR-dependent regulation of katB-ankB, ahpB, and ahpC-ahpF. Journal of Bacteriology 182: 4533-4544.

Ogura T & Wilkinson AJ (2001) AAA+ superfamily ATPase: common structure- diverse function. Genes to Cells 6: 575-597.

Oliver A, Baquero F & Blazquez J (2002) The mismatch repair system (mutS, mutL and uvrD genes) in Pseudomonas aeruginosa: molecular characterization of naturally occurring mutants. Molecular Microbiology 43: 1641-1650.

Oliver A, Canton R, Campo P, Baquero F & Blazquez J (2000) High frequency of hypermutable Pseudomonas aeruginosa in cystic fibrosis lung infection. Science 288: 1251-1254.

Oppenheim AB, Kobiler O, Stavans J, Court DL & Adhya S (2005) Switches in bacteriophage lambda development. Annual Review of Genetics 39: 409-429.

Overhage J, Lewenza S, Marr AK & Hancock REW (2007) Identification of genes involved in swarming motility using a Pseudomonas aeruginosa PAO1 mini-Tn5-lux mutant library. Journal of Bacteriology 189: 2164-2169.

137 ! References

Pappenheimer AMJ (1977) Diphtheria toxin. Annual Review in Biochemistry 46: 69-94.

Park S, You XJ & Imlay JA (2005) Substantial DNA damage from submicromolar intracellular hydrogen peroxide detected in Hpx- mutants of Escherichia coli. Proceedings of the National Academy of Sciences of the United States of America 102: 9317-9322.

Parsek MR & Singh PK (2003) Bacterial biofilms: an emerging link to disease pathogenesis. Annual Review in Microbiology 57: 677-701.

Parsek MR & Fuqua C (2004) Biofilm 2003: Emerging themes and challenges in studies of surface-associated microbial life. Journal of Bacteriology 186.

Patel M, Jiang Q, Woodgate R, Cox MM & Goodman MF (2010) A new model for SOS-induced mutagenesis: how RecA protein activates DNA polymerase V. Critical Reviews in Biochemistry and Molecular Biology 45: 171-184.

Pedruzzo I, Rosenbusch JP & Locher KP (1998) Inactivation in vitro of the Escherichia coli outer membrane protein FhuA by a phage T5-encoded lipoprotein. FEMS Microbiology Letters 168: 119-125.

Perna NT, Mayhew GF, Posfai G, Elliott S, Donnenberg MS, Kaper JB & Blattner FR (1998) Molecular evolution of a pathogenicity island from enterohemorrhagic Escherichia coli O157:H7. Infection and Immunity 66: 3810-3817.

Pires D, Sillankorva S, Faustino A & Azeredo J (2011) Use of newly isolated phages for control of Pseudomonas aeruginosa PAO1 and ATCC 10145 biofilms. Research in Microbiology 162: 1-9.

Pratt D, Tzagologg H & Erdahl WS (1966) Conditional lethal mutants of the same filamentous coliphage M13. I. Isolation, complementation, cell killing, time of cistron action. Virology 30: 397-410.

Ptashne M (2004) A genetic switch. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York.

Radman M (1974) Molecular and environmental aspects of mutagenesis. Thomas, Springfield.

Rahme LG, Stevens EJ, Wolfort SF, Shao J, Tompkins RG & Ausubel FM (1995) Common virulence factors for bacterial pathogenicity in plants and animals. Science 268: 1899-1902.

Rakonjac K, Bennett NJ, Spagnuolo J, Gagic D & Russel M (2011) Filamentous bacteriophage: biology, phage display and nanotechnology applications. Current Issues in Molecular Biology 13: 51-76.

Ramage G, Martinez JP & Lopez-Ribot JL (2006) Candida biofilms on implanted biomaterials: a clinically significant problem. FEMS Yeast Research 6: 979-986.

138 ! References

Ratjen F, Munck A, Kho P & Angyakosi G (2010) Treatment of early Pseudomonas aeruginosa infection in patients with cystic fibrosis: the ELITE trial. Thorax 65: 286-291.

Rayssiguier C, Thaler DS & Radman M (1989) The barrier to recombination between Escherichia coli and Salmonella typhimurium disrupted in mismatch- repair mutants. Nature 342: 396-401.

Rhaese HJ & Freese E (1968) Chemical analysis of DNA alterations I. Base liberation and backbone breakage of DNA and oligodeoxyadenylic acid by hydrogen peroxide and hydroxylamine. Biochimica et Biophysica Acta 155: 476- 490.

Rice SA & Lampson BC (1996) Bacterial reverse transcriptase and msDNA. Virus Genes 11: 95-104.

Rice SA, van den Akker B, Pomati F & Roser D (2012) Pseudomonas aeruginosa contamination in public swimming pools. Journal of Water and Health 10: 181- 196.

Rice SA, Tan CH, Mikkelsen PJ, Kung V, Woo J, Tay M, Hauser A, McDougald D, Webb JS & Kjelleberg S (2009) The biofilm life cycle and virulence of Pseudomonas aeruginosa are dependent on a filamentous prophage. The ISME Journal 3: 271-282.

Roberts JW & Roberts CW (1975) Proteolytic cleavage of bacteriophage lambda repressor in induction. Proceedings of the National Academy of Sciences of the United States of America 72: 147-151.

Roberts JW, Roberts CW & Craig NL (1978) Escherichia coli recA gene product inactivates phage lambda repressor. Proceedings of the National Academy of Sciences of the United States of America 75: 4714-4718.

Rowntree RK & Harris A (2003) The phenotypic consequences of CFTR mutations. Annals of Human Genetics 67: 471-485.

Rubin BK (2009) Mucus, phlegm, and sputum in cystic fibrosis. Respiratory Care 54: 726-732.

Ruiz-Vazquez R & Murillo FJ (1984) Abnormal motility and fruiting behaviour of Myxococcus xanthus bacteriophage-resistant strains induced by a clear-plaque mutant of bacteriophage Mx8. Journal of Bacteriology 160: 818-821.

Russel M (1995) Moving through the membrane with filamentous phages. TRENDS in Microbiology 3: 223-228.

Rybtke MT, Jensen PO, Hoiby N, Givskov M, Tolker-Nielsen T & Bjarnsholt T (2011) The implication of Pseudomonas aeruginosa bioflms in infections. Inflammation and Allergy - drug targets 10: 141-157.

139 ! References

Sabbattini P, Forti F, Ghisotti D & Deho G (1995) Control of transcription termination by an RNA factor in bacteriophage P4 immunity: identification of the target sites. Journal of Bacteriology 177: 1425-1434.

Sadikot RT, Blackwell TS, Christman JW & Prince AS (2005) Pathogen-host interactions in Pseudomonas aeruginosa pneumonia. American Journal of Respiratory and Critical Care Medicine 171: 1209-1223.

Saint S & Chenoweth CE (2003) Biofilms and catheter-associated urinary tract infections. Infectious Disease Clinics of North America 17: 411-432.

Salmon KA, Freedman O, Ritchings BW & DuBow MS (2000) Characterization of the lysogenic repressor C gene of the Pseudomonas aeruginosa transposable bacteriophage D3112. Virology 272: 85-97.

Sauer K, Camper AK, Ehrlich GD, Costerton JW & Davies DG (2002) Pseudmonas aeruginosa displays multiple phenotypes during development as a biofilm. Journal of Bacteriology 184: 1140-1154.

Sauer K, Cullen MC, Rickard AH, Zeef LA, Davies DG & Gilbert P (2004) Characterization of nutrient-induced dispersion in Pseudomonas aeruginosa PAO1 biofilm. Journal of Bacteriology 186: 7312-7326.

Saumaa S, Tover A, Kasak L & Kivisaar M (2002) Different spectra of stationary- phase mutations in early-arising versus late-arising mutants of Pseudomonas putida: involvement of the DNA repair enzyme MutY and the stationary-phase sigma factor RpoS. Journal of Bacteriology 184: 6957-6965.

Schembri MA, Kjaergaard K & Klemm P (2003) Global gene expression in Escherichia coli biofilms. Molecular Microbiology 48: 253-267.

Schlacher K & Goodman MF (2007) Lessons from 50 years of SOS DNA- damage-induced mutagenesis. Nature Reviews 8: 587-594.

Schleheck D, Barraud N, Klebensberger J, Webb JS, McDougald D, Rice SA & Kjelleberg S (2009) Pseudomonas aeruginosa PAO1 preferentially grows as aggregates in liquid batch cultures and disperses upon starvation. PLoS ONE 4: e5513.

Schuch R & Fischetti VA (2006) Detailed genomic analysis of the Wβ and γ phages infecting Bacillus anthracis: implications for evolution of environmental fitness and antibiotic resistance. Journal of Bacteriology 188: 3037-3051.

Schuch R & Fischetti VA (2009) The secret life of the anthrax agent Bacillus anthracis: bacteriophage-mediated ecological adaptations. PLoS ONE 4: 1-23.

Schuster M, Hawkins AC, Harwood CS & Greenberg EP (2004) The Pseudomonas aeruginosa RpoS regulon and its relationship to quorum sensing. Molecular Microbiology 51: 973-985.

Schweitz H (1969) Degradation du DNA par hydrogen peroxide en presenced d'ions Cu++, Fe++ et Fe+++. Biopolymers 8: 101-119. 140 ! References

Scott JR (1975) Superinfection immunity and prophage repression in phage P1. Virology 65: 173-178.

Sillankorva S, Neubauer P & Azeredo J (2008) Pseudomonas fluorescens biofilms subjected to phage phiIBB-PF7A. BMC Biotechnology 9: 79-90.

Simons RW & Kleckner N (1988) Biological regulation by antisense RNA in prokaryotes. Annual Review in Genetics 22: 576-600.

Singh R, Ray P, Das A & Sharma M (2009) Role of persisters and small-colony variants in antibiotic resistance of planktonic and biofilm-associated Staphylococcus aureus: an in vitro study. Journal of Medical Microbiology 58: 1067-1073.

Sixma TK (2001) DNA mismatch repair: MutS structures bound to mismatches. Current Opinion in Structural Biology 11: 47-52.

Smith EE, Buckley DG, Wu Z, Saenphimmachak C, Hoffman LR, D’Argenio DA, Miller SI, Ramsey BW, Speert DP, Moskowitz SM, Burns JL, Kaul R & Olson MV (2006) Genetic adaptation by Pseudomonas aeruginosa to the airways of cystic fibrosis patients. Proceedings of the National Academy of Sciences of the United States of America 103: 8487-8492.

Smith JJ, Travis SM, Greenberg EP & Welsh MJ (1996) Cystic fibrosis airway epithelia fail to kill bacteria because of abnormal airway surface fluid. Cell 85: 229-236.

Smith PA & Romesberg FE (2007) Combating bacteria and drug resistance by inhibiting mechanisms of persistence and adaptation. Nature Chemical Biology 3: 549-556.

Sorek R, Kunin V & Hugenholtz P (2008) CRISPR - a wide spread system that provides acquired resistance against phages in bacteria and archaea. Nature 6: 181-186.

Spiers AJ, Kahn SG, Bohannon J, Travisano M & Rainey PB (2002) Adaptive divergence in experimental populations of Pseudomonas fluorescens. I. Genetic and phenotypic bases of wrinkly spreader fitness. Genetics 161: 33-46.

Starkey M, Hickman JH, Ma L, Zhang N, De Long S, Hinz A, Palacios S, Manoil C, Kirisits MJ, Starner TD, Wozniak DJ, Harwood CS & Parsek MR (2009) Pseudomonas aeruginosa rugose small-colony variants have adaptations that likely promote persistence in the cystic fibrosis lung. Journal of Bacteriology 191: 3482-3503.

Stewart PS (2003) Diffusion in biofilms. Journal of Bacteriology 185: 1485-1491.

Stewart PS & Franklin MJ (2008) Physiological heterogeneity in biofilms. Nature Reviews Microbiology 6: 199-210.

Storz G & Imlay JA (1999) Oxidative stress. Current Opinion in Microbiology 2: 188-194. 141 ! References

Sturino JM & Klaenhammer TR (2006) Engineered bacteriophage-defence systems in bioprocessing. Nature Reviews Microbiology 4: 395-404.

Suh SJ, Silo-Suh L, Woods DE, Hassett DJ, West SE & Ohman DE (1999) Effect of rpoS mutation on the stress response and expression of virulence factors in Pseudomonas aeruginosa. Journal of Bacteriology 181: 3890-3897.

Sullivan MB, Coleman ML, Weigele P & Chisholm SW (2005) Three Procholorococcus cyanophage genomes: signature features and ecological interpretations. PLoS Biology 3: 790-806.

Suttle CA (1994) The significance of viruses to mortality in aquatic microbial communities. Microbial Ecology 28: 237-243.

Suttle CA (2005) Viruses in the sea. Nature 437: 356-361.

Suttle CA (2007) Marine viruses - major players in the global ecosystem. Nature Reviews Microbiology 5: 801-812.

Tam R & Saier MHJ (1993) Structural, functional and evolutionary relationships among extracellular solute-binding receptors of bacteria. Microbiological Reviews 57: 320-346.

Tan CH (2006) The involvement of Pf4 filamentous phage in Pseudomonas aeruginosa biofilm development. Thesis, University of New South Wales, Sydney, Australia.

Tan CH (2006) The involvement of Pf4 filamentous phage in Pseudomonas aeruginosa biofilm development. Thesis. University of New South Wales, Sydney, Australia.

Tay M (2008) Identification of the secondary phage in Pseudomonas aeruginosa and determination of its role in biofilm development. Honours thesis. University of New South Wales, Sydney, Australia.

Tay MQX (2013) The role of bacteriophage in granulation. Thesis, Nanyang Technological University, Singapore.

Thanomsub B, Pumeechockchai W, Limtrakul A, Arunrattiyakom P, Petchleelaha W, Nitoda T & Kanzaki H (2007) Chemical structures and biological activities of rhamnolipids produced by Pseudomonas aeruginosa B189 isolated from milk factory waste. Bioresource Technology 98: 1149-1153.

Thomas SR, Ray A, Hodson ME & Pitt TL (2000) Increased sputum amino acid concentrations and auxotrophy of Pseudomonas aeruginosa in severe cystic fibrosis lung disease. Thorax 55: 795-797.

Thomassen MJ, Demko CA, Boxerbaum B, Stern RC & Kuchenbrod PJ (1979) Multiple of isolates of Pseudomonas aeruginosa with differing anti-microbial susceptibility patterns from patients with cystic fibrosis. The Journal of Infectious Diseases 140: 873-880.

142 ! References

Thormann KM, Saville RM, Shukla S & Spormann AM (2005) Induction of rapid detachment in Shewanella oneidensis MR-1 biofilms. Journal of Bacteriology 187: 1014-1021.

Tillett D & Neilan BA (2000) Xanthogenate nucleic acid isolated from cultured and environmental cyanobacteria. Journal of Phycology 36: 251-258.

Todar K (2012) Bacteriophage. ed.^eds.), p.^pp. Madison, Wisconsin.

Tolker-Nielsen T, Brinch UC, Ragas PC, Anderson JB, Jacobsen CS & Molin S (2000) Development and dynamics of Pseudomonas sp. biofilms. Journal of Bacteriology 182: 6482-6489.

Trafny EA (1998) Susceptibility of adherent organisms from Pseudomonas aeruginosa and Staphylococcus aureus strains isolated from burn wounds to antimicrobial agents. International Journal of Antimicrobial Agents 10: 223-228.

Uzzau S, Figueroa-Bossi N, Rubino S & Bossi L (2001) Epitope tagging of chromosomal genes in Salmonella. Proceedings of the National Academy of Sciences of the United States of America 98: 15264-15269. van Melderen L & de Bast MS (2009) Bacterial toxin-antitoxin systems: more than selfish entities? PLoS Genetics 5: e1000437. van Wezenbeek PMGF, Hulsebos TJM & Schoenmakers JGG (1980) Nucleotide sequence of the filamentous bacteriophage M13 DNA genome: comparison with phage fd. Gene 11: 129-148.

Vinckx T, Wei Q, Matthijs S & Cornelis P (2010) The Pseudomonas aeruginosa oxidative stress regulator oxyR influences production of pyocyanin and rhamnolipids: protective role of pyocyanin. Microbiology 156: 678-686.

Wagner PL, Neely MN, Zhang XP, Acheson DWK, Waldor MK & Friedman DI (2001) Role of a phage promoter in Shiga toxin 2 expression from a pathogenic Escherichia coli strain. Journal of Bacteriology 183: 2081-2085.

Wagner PL, Livny J, Neely MN, Acheson DWK, Friedman DI & Waldor MK (2002) Bacteriophage control of Shiga toxin 1 production and release of Escherichia coli. Molecular Microbiology 44: 957-970.

Ward JF & Kuo I (1976) Strand breaks, base release and post-irradiation changes in DNA gamma irradiated in dilute oxygen saturated aqueous solution. Radiation Research 66: 485.

Ward JF, Evans JW, Limoli CL & Calalbro-Jones PM (1987) Radiation and hydrogen peroxide induced free radical damage to DNA. British Journal of Cancer 8: 105-112.

Watnick P & Kolter R (2000) Biofilm, city of microbes. Journal of Bacteriology 182: 2675-2679.

143 ! References

Watnick PI, Lauriano CM, Klose KE, Croal L & Kolter R (2001) The absence of a flagellum leads to altered colony morphology, biofilm development and virulence in Vibrio cholerae O139. Molecular Microbiology 39: 223-235.

Webb JS, Lau M & Kjelleberg S (2004) Bacteriophage and phenotypic variation in Pseudomonas aeruginosa biofilm development. Journal of Bacteriology 186: 8066-8073.

Webb JS, Thompson LS, James S, Charlton T, Tolker-Nielsen T, Koch B, Givskov M & Kjelleberg S (2003) Cell death in Pseudomonas aeruginosa biofilm development. Journal of Bacteriology 185: 4585-4592.

Wei Q, Minh PNL, Dotsch A, Hildebrand F, Panmanee W, Elfarash A, Schulz S, Plaisance S, Charlier D, Hassett D, Haussler S & Cornelis P (2012) Global regulation of gene expression by OxyR in an important human opportunistic pathogen. Nucleic Acids Research 40: 1-14.

Weisburg WG, Barns SM, Pelletier DA & Lane DJ (1991) 16s ribosomal DNA amplification for phylogenetic study. Journal of Bacteriology 173: 697-703.

West BW & Scott JR (1977) Superinfection immunity and prophage repression in phage P1 and P7. Virology 78: 267-276.

Whiteley M, Bangera MG, Bumgarner RE, Parsek MR, Teitzel GM, Lory S & Greenberg EP (2001) Gene expression in Pseudomonas aeruginosa biofilms. Nature 413: 860-864.

Williams P & Camara M (2009) Quorum sensing and environmental adapation in Pseudomonas aeruginosa: a take of regulatory networks and multifunctional signal molecules. Current Opinion in Microbiology 12: 182-191.

Winsor GL, Lam DK, Fleming L, Lo R, Whiteside MD, Yu NY, Hancock RE & Brinkman FS (2011) Pseudomonas Genome Database: improved comparative analysis and population genomics capability for Pseudmonas genomes. Nucleic Acids Research 39: 596-600.

Wommack KE & Colwell RR (2000) Virioplankton: viruses in aquatic ecosystems. Microbiology and Molecular Biology Reviews 64: 69-114.

Woo JKK, Webb JS, Kirov SM, Kjelleberg S & Rice SA (2012) Biofilm dispersal cells of a cystic fibrosis Pseudomonas aeruginosa isolate exhibit variability in functional traits likely to contribute to persistent infection. FEMS Pathogens and Disease 66: 251-264.

Woolford J, Cashman J & Webster R (1974) F1 coat protein synthesis and altered phospholipid metabolism in fl infected E. coli. Virology 588: 544-560.

Workentine M & Surette MG (2011) Complex Pseudomonas population structure in cystic fibrosis airway infections. American Journal of Respiratory and Critical Care Medicine. 183: 1581-1583.

144 ! References

Workentine ML, Harrison JJ, Weljie AM, Tran VA, Stenroos PU, Tremaroli V, Vogel HJ, Ceri H & Turner RJ (2010) Phenotypic and metabolic profiling of colony morphology variants evolved from Pseudomonas fluorescens biofilms. Environmental Microbiology 12: 1565-1577.

Wu JY, Yeh KL, Lu WB, Lin CL & Chang JS (2008) Rhamnolipid production with indigenous Pseudomonas aeruginosa EM1 isolated from oil-contaminated site. Bioresource Technology 99: 1157-1164.

Yachi S & Loreau M (1999) Biodiversity and ecosystem productivity in a fluctuating environment: the insurance hypothesis. Proceedings of the National Academy of Sciences of the United States of America 96: 1463-1468.

Yamada T, Kawasaki T, Nagata S, Fujiwara A, Usami S & Fujie M (2007) New bacteriophages that infect the phytopathogen Ralstonia solanacearum. Microbiology 153: 2630-2639.

Yarmolinsky MB (1995) Programmed cell death in bacterial populations. Science 267: 836-837.

Young R (1992) Bacteriophage lysis: mechanism and regulation. Microbiological Reviews 56: 430-481.

Young R & N. WI (2006) Phage Lysis. In Bacteriophages. Oxford University press, New York.

Zapata M, Silva F, Luza F, Wilkens M & Riqueline C (2007) The inhibitory effect of biofilms produced by wild bacterial isolates to the larval settlement of the fouling ascidia Ciona intestinalis and Pyura praeputialis. Biofilms 10: 1-4.

Zegans ME, Wagner JC, Cady KC, Murphy DM, Hammond JH & O'Toole GA (2009) Interaction between bacteriophage DMS3 and host CRISPR region inhibits group behaviors of Pseudomonas aeruginosa. Journal of Bacteriology 191: 210- 219.

Zhang T, Fu Y & Biship P (1995) Competition for substrate and space in biofilms. Water Environment Research 67: 992-1003.

Zogaj X, Nimtz M, Rohde W, Bokranz W & Romling U (2001) The multicellular morphotypes of Salmonella typhimurium and Escherichia coli produce cellulose as the second component of the extracellular matrix. Molecular Microbiology 39: 1452-1463.

145 !