Elucidating Potential Virulence Factors in Fusobacterium Nucleatum by Kyla L.S Cochrane

Elucidating potential virulence factors in Fusobacterium nucleatum by Kyla L.S Cochrane

A Thesis presented to The University of Guelph

In partial fulfilment of requirements for the degree of Doctor of Philosophy in Molecular and Cellular Biology

Guelph, Ontario, Canada © Kyla L.S. Cochrane, February, 2016

ABSTRACT

ELUCIDATING POTENTIAL VIRULENCE FACTORS IN FUSOBACTERIUM NUCLTEAUM

Kyla L.S. Cochrane Advisor: University of Guelph, 2016 Dr. Emma Allen-Vercoe

Fusobacterium nucleatum (Fn) is a strictly anaerobic, Gram negative bacterial species that has been implicated in multiple clinical pathologies including periodontal disease, inflammatory bowel disease and colorectal cancer (CRC). The species is unusual in its phenotypic and genotypic heterogeneity, with some strains demonstrating a more virulent phenotype than others; however, the genetic basis for these differences and the association of any particular subspecies of Fn with disease has not yet been proven.

With the recent confirmation of the Fusobacterium sequence enrichment in colorectal carcinomas, the characterization of cultured Fn isolates from CRC biopsies was warranted. Fn strains were isolated from human intestinal mucosal biopsies and were phenotypically and genetically analyzed. The invasive, inflammatory potential of Fn in vitro was also assessed.

There was extensive variability between the phenotypic, invasive and inflammatory potentials of

8 different CRC-derived Fn strains. Further genetic characterization showed the prevalence of a class of genes of unknown function containing MORN2-domains that were expanded in

Fusobacterium spp. known to be highly invasive. Proteins containing MORN2-domains are predicted to be surface-associated and as such may play a role in enabling host cell adherence and invasion by fusobacterial species. In addition, the presence of bacteriophages in Fn strains

ii were also predicted to explain the genomic heterogeneity between Fn strains especially in relation to adherence and invasion.

These predictions prompted the exploration of the Fn transcriptome for the purpose of determining the proteins that govern adherence and subsequent invasion of the bacteria into host cells. Going forward, RNA-seq will be used as a tool to help unravel the mechanisms of

Fusobacterium pathogenesis. Continuing research in this area will enable the development of diagnostic and therapeutic strategies for the detection and treatment of Fusobacterium-associated diseases.

iii

Acknowledgments

Firstly, I would like to express my sincere gratitude to my advisor Dr. Emma Allen-

Vercoe for not only convincing me to return to the research world but also for her continuous support, patience, motivation and immense knowledge. Most of all, I want to thank her for having the faith in me that was necessary to handle this particular project. It was not always easy but her guidance and unwavering support allowed to me to succeed.

Besides my advisor, I would like to thank the rest of my advisory committee: Dr. Nina

Jones, Dr. Lucy Mutharia and Dr. Katrina Gee for their insightful comments, encouragement and also for their hard questions, which incented me to widen my research from various perspectives.

Additional thanks go to Dr. Abigail Mason McGuire, Dr. Ashlee Earl, Dr. Reggie Lo and Dr.

Robert Holt, without their bacterial, genomic and technical expertise I would have been lost. I would also like to acknowledge the government of Ontario for providing financial support throughout my graduate career in the form of scholarships, fellowships and bursaries.

Additionally, I would like to recognize the Crohn's and Colitis Canada Foundation and the

Canadian Cancer Society Research Institute for providing funding to the Allen-Vercoe lab.

Special thanks go out to the members of the Allen-Vercoe lab, past and present, for all of your help in the lab, coffee breaks to vent and shoulders to cry or laugh hysterically on whenever it was a bad science day. In particular I would like to thank the friends who became family: you know who you are. Completing this work would have been all the more difficult were it not for the support provided by my wonderful family, my closest friends, and of course, my best friend and partner, Michael Beswick. It is always impossible to personally thank everyone who has facilitated successful completion of a project. To those of you who I did not specifically name, I give my thanks for moving me towards my goal. I will forever be indebted to you all.

Author’s Declaration of Works Performed

I declare that this dissertation is my own account, based upon work actually carried out by me and that all sources of material, not resulting from my own investigation, including observational information, have been clearly indicated in their representative sections.

For the work reported in Section 2.1.1 B, isolation of bacterial strains from colorectal cancer biopsy samples from Cohort One were processed by Dr. Emma Allen-Vercoe and Dr.

Jaclyn Strauss. Additional help from both Michelle Daigneault and Dr. Emma Allen-Vercoe was provided occasionally during the isolation of bacterial strains from Cohort Two and Three as well. Sequencing analysis of all bacterial strains isolated from the colorectal cancer biopsies was performed by Sindy Babinsky at the BC Cancer Agency.

In Section 2.3.1 A, the Pseudomonas MORN2 recombinant protein was cloned and purified by Dr. Bryan Berger at Lehigh University in Bethlehem, Pennsylvania.

Whole genome sequencing and full annotation of the two Fn 7-1 bacteriophage, from

Section 2.4.1 C, were performed by Dr. Abigail Manson McGuire and colleagues at the Broad

Institute of MIT in Cambridge, Massachusetts.

Table of Contents ABSTRACT ...... ii Acknowledgments...... iv Author’s Declaration of Works Performed ...... v List of Figures ...... ix List of Tables ...... xi List of Abbreviations ...... xii Chapter 1. Literature Review and Research Objectives ...... 1 1.1 The Gut Microbiota ...... 1 1.2 Commensals, Health and Immunity ...... 2 1.3 Inflammatory Diseases in the Gut ...... 3 1.4 Inflammatory Bowel Disease and the associated risk of Colorectal Cancer ...... 6 1.5 Fusobacterium nucleatum ...... 7 1.6 Fusobacterium nucleatum and Virulence Factors ...... 9 1.7 Fusobacterium nucleatum and Inflammatory Bowel Disease ...... 16 1.8 Fusobacterium nucleatum and Colorectal Cancer ...... 18 1.9 Conclusion ...... 21 1.10 Research Objectives ...... 22 Chapter 2. Materials and Methods ...... 23 2.1 Isolation of bacterial species from colorectal cancer biopsies and characterization of isolated Fusobacterium nucleatum (Fn) species...... 23 2.1.1 Bacterial Strains ...... 23 2.1.2 Identification of CRC-derived strains and the characterization of Fn strains ...... 40 2.2 Fn interactions with human colonic epithelial cells ...... 51 2.2.1 Assessing the ability of CRC-derived Fn isolates to invade human epithelial cells in vitro ...... 51 2.2.2 Host Immune Response to Fn and CRC-derived Strains ...... 63 2.2.3 Assessing the effects of human cytokines on Fn invasion in vitro ...... 68 2.3 Determining the role of MORN2 domain-associated proteins in Fusobacterium adhesion/invasion ...... 71 2.3.1 Characterization of MORN2 domains in adhesion and invasion of Fn strains ...... 71

2.4 Determining the presence and role of bacteriophage in invasive Fn strains ...... 74 2.4.1 Induction and characterization of the invasive Fn subspecies animalis 7-1 bacteriophage ɸFunu1 and ɸFunu2 ...... 74 2.4.2 Induction and Characterization of the invasive Fn subspecies nucleatum 13-3C bacteriophage ...... 76 Chapter 3. Isolation and characterization of bacterial species from CRC biopsies ...... 80 Results ...... 84 3.1 Isolation of bacterial species from CRC biopsies ...... 84 3.2 Identification of the CRC bacterial strains ...... 84 3.3 Characterization of the isolated Fn strains ...... 93 3.4 Isolation and characterization of Campylobacter spp...... 106 3.5 Co-aggregation of Fn gut isolates with other bacterial isolates from human CRC biopsies ...... 115 Discussion ...... 119 Chapter 4. Determining the invasive potential of CRC-derived Fn isolates and the effects on host cytokine secretion in response to Fn invasion ...... 128 Results ...... 130 4.1 Assessing the ability of CRC-derived Fn isolates to invade human intestinal epithelial cells in vitro ...... 130 4.2 Host Cell Immune Response to Fn and CRC-derived strains ...... 137 4.3 Assessing the effects of human cytokines on Fn invasion in vitro ...... 146 Discussion ...... 150 Chapter 5. Determining the role of MORN2 domain-associated proteins in Fusobacterium adhesion/invasion ...... 161 Results ...... 167 5.1 Characterization of MORN2 domains in adhesion and invasion of Fn strains ...... 167 Discussion ...... 173 Chapter 6. Determining the presence and role of bacteriophage in invasive Fn strains ...... 177 Results ...... 181 6.1 Induction and characterization of the invasive Fn 7-1 bacteriophage ɸFunu1 and ɸFunu2 ...... 181 6.2 Induction and characterization of the bacteriophage from Fn 13-3C ...... 186 Discussion ...... 193

vii

Chapter 7. Summary and Significance...... 197 Literature Cited ...... 202 Appendix I- Method Development for Fn transcriptome studies (RNA-Seq) ...... 222 Materials and Methods ...... 225 A. Isolating total RNA from invaded Fn 7-1 bacterial cells ...... 225 B. Ribodepletion of total RNA samples ...... 227 Results ...... 231 A. Isolation and purification of high quality mRNA from Fn 7-1 and Caco-2 cells from an adhesion/invasion model ...... 231 Discussion and Future Directions ...... 234 Appendix II - Media and Buffers ...... 237 Appendix III- Supplementary Data ...... 242 Antibiotic exclusion assay pre-screen to determine the effects of IL-27 addition to Cac02 cells on Fn invasion ...... 242 Bicinchroninic acid (BCA) assay for P. aeruginosa MORN2 peptide quantification ...... 244

viii

List of Figures

Figure 1.1 Multispecies Biofilms…...... 11

Figure 1.2 Dental Plaque Diagram...... 13

Figure 2.1 Screen captures for Pixel Quantification ...... 57

Figure 3.1 Co-occurance of Microbes Network ...... 82

Figure 3.2 fadA gene amplicons...... 97

Figure 3.3 Phase contrast micrographs of CRC-derived Fn strains ...... 100

Figure 3.4 Fn Phylogenetic tree ...... 105

Figure 3.5 Campylobacter spp. Phylogenetic tree ...... 112

Figure 3.6 Phase microscopy of CC57C and CC53 aggregation ...... 118

Figure 4.1 Western Blot analysis of Fn antibodies ...... 132

Figure 4.2 Fn Invasion assay analysis ...... 133

Figure 4.3 Mircographs of low resolution Fn strains ...... 134

Figure 4.4 Cytokine secretion by Caco2 cells exposed to CRC-derived Fn ...... 139

Figure 4.5 Cytokine secretion by CRC and IBD-derived Fn strains ...... 143

Figure 4.6 Cytokine secretion by bacterial consortia...... 144

Figure 4.7 Cytokine secretion summary ...... 145

Figure 4.8 Data representation of Caco-2 cells pretreated with cytokines prior to infection .....149

Figure 5.1 Phylogenetic tree of 26 Fusobacterium strains ...... 164

Figure 5.2 Gene categories in active and passive invaders ...... 165

Figure 5.3 MORN2 orthogroups ...... 166

Figure 5.4 MORN2 alignments...... 168

Figure 5.5 MORN2 competition assay ...... 172

Figure 6.1 Fusobacterium bacteriophage phylogeny ...... 180

Figure 6.2 Funu1 and Funu2 electron micrographs ...... 184

Figure 6.3 Funu1 and Funu2 alignment to the Fn 7-1 genome ...... 185

Figure 6.4 NuPh1 electron micrograph ...... 189

Figure 6.5 Linear NuPh1 genome diagram ...... 190

Figure 6.6 NuPhu1 DNA gene annotation ...... 191

Figure 6.7 Bacteriophage specific gene amplicons...... 192

Figure AIa Schematic of Fn 7-1 RNA isolation ...... 228

Figure AIb Representative Fn 7-1 RNA chromotography ...... 233

Figure AIIIa Invasion of Caco-2 cells (pre-treated with IL-27) with Fn 7-1 ...... 243

List of Tables Table 2.1 Bacterial strains used in this thesis……………………………………………… …...24 Table 2.2 Cohort One Metadata ...... 29 Table 2.3 Cohort Two Metadata …...... 30

Table 2.4 Cohort Three Metadata ….…...... 33

Table 2.5 Media types …...... 37

Table 2.6 Primers used in this study …...... 44

Table 2.7 Bacterial strains used in the cytokine secretion assays …...... 65

Table 2.8 Bacteriophage specific primers ...... 79

Table 3.1 List of bacterial species isolated from each CRC patient …...... 85

Table 3.2 Details of Fn strain isolates from CRC ...... 94

Table 3.3 16s rRNA gene-based identification of CRC-derived Fn strains…...... 95

Table 3.4 Fn Biochemical Profiles...... 101

Table 3.5 Fn Antibiotic Resistance Profiles...... 102

Table 3.6 Campylobacter spp. isolated from CRC biopsies...... 107

Table 3.7 Campylobacter spp. morphological table…...... 109

Table 3.8 Antibiotic Resistance profiles for CC57C and 7/5 D55...... 110

Table 3.9 Key predicted genes from CC57C……...... 114

Table 3.10 Co-aggregation Assay Data……...... 117

Table 4.1 Adherance statistics……...... 136

Table 4.2 Bacterial consortia...... 142

Table 4.3 Cytokine information...... 148

Table 5.1 MORN2 Aggregation Assay…...... 170

List of Abbreviations

AIEC Adherent-invasive E. coli ANOVA Analysis of variance Ab Antibody BBE Bacter oides Bile Esculin BCA Bicinchoninic Acid BCCA_TRR British Columbia Cancer Agency_ Tissue Repository BCIP/NBT 5-Bromo-4-chloro-3-indolyl phosphate/ BSA Bovine Serum Albumin CC Colon Cancer CD Crohn’s Disease CRC Colorectal Cancer CXCL8 CX Chemokine Ligand 8 DMEM Dulbecco’s minimal essential media DMSO Dimethyl sulfoxide EM Electron Microscopy FAA Fastidious Anaerobic Agar FBS Fetal Bovine Serum FadA Fusobacterium adhesin A Fn Fusobacterium nucleatum GI Gastrointestinal HRP Horseradish Peroxidase IBD Inflammatory Bowel Disease IL-6 Interleukin-6 IL-10 Interleukin-10 IL-12 Interleukin-12 IL-27 Interleukin-27

xii

JVN Josamycin Vancomycin Norfloxacin MAMP Microbial Associated Molecular Patterns MAP Mycobacterium avium subspecies paratuberculosis MBP Maltose Binding Protein MOI Multiplicity of Infection MORN2 Membrane Occupation Recognition Nexus 2 OMP Outer Membrane Protein ORF Open Reading Frame PAI Pathogenicity Island PBS Phosphate Buffered Saline PRR Pattern Recognition Receptor PVDF Polyvinylidene Fluoride SFM Serum Free Media sp. Species spp. Species (plural) TGF-β Transforming Growth Factor beta TLR Toll-like Receptor TNF-α Tumour Necrosis factor alpha

TSBsupp Tryptic Soy Broth (supplememnted) TSE Trypticase soy esculin agar UC Ulcerative colitis

xiii

Chapter 1. Literature Review and Research Objectives

1.1 The Gut Microbiota

The intestinal microbiota, the collection of microorganisms that live in the gut, is a dynamic ecosystem of emerging importance to human health. The composition and distribution of this ecosystem varies through the gastrointestinal (GI) tract from the mouth to anus and between the lumen and the mucosa, with the highest concentration of bacteria being in the colon.

The colon houses 1011-1012 bacterial cells per gram of content and between 500-1000 different species, thereby outnumbering human cells 10:1 (Kamada et al., 2013). It has been suggested that the genetic potential of this microbiota gives rise to the metabolic activity of an organ such as a liver (Kelly et al., 2005, O’Hara & Shanahan, 2006). The microbiome is often separated into two groups: [A] the autochthonous (ubiquitous) community; and [B] the allochthonous

(transient) community (Ley et al., 2006, Nava & Stappenbeck, 2011). The microbiota of each individual host is distinct and unique, and develops from the time of birth to become a relatively stable community by the time of weaning (Hope et al., 2005). Host selection of microbes relies on environmental factors, such as maternal microbiota, diet and clinical intervention (Palmer et al., 2007), as well as genetic factors, such as cell surface receptors (Ley et al., 2006).

The intestinal microbiota plays a pivotal role in local and systemic host physiology by accomplishing various functions the host cannot achieve alone, for example: [A] salvaging energy and absorbing nutrients by metabolizing dietary polysaccharides and carbohydrates, and synthesizing vitamins (Fujimura et al., 2010; Thompson-Chagoyan et al., 2005); [B] providing

1 colonisation resistance against potential pathogens through competition for host adherence sites and nutrient availability, altering physiological conditions, and producing anti-microbial peptides

(Thompson-Chagoyan et al., 2005); [C] development of the mucosal immune system by inducing tolerance mechanisms (Meddings, 2007); and [D] maintenance of the physiological gut environment, especially epithelial differentiation (Yang & Pei, 2006; Thompson-Chagoyan et al.,

2005). Relationships occur not only between the host and its microbiota, but also between the microbes themselves. In order for a dynamic ecosystem to develop, bacteria co-aggregate to take advantage of optimal food webs between species as well as exploitation of host epithelial adherence sites, and may also undergo genomic adaptations (most likely between the closely connected autochthonous species) to allow for selection, adaptation and persistence of microbes in their niche environments (Nava & Stappenbeck, 2011).

Despite a wealth of knowledge now available through “omic” technologies of increasing sophistication, it is still not understood how ecosystem homeostasis is achieved although it is thought that the presence of a core microbiota (microbial species shared between different individuals) may drive this (Sekelja et al., 2011). Additionally, while the abundance of various species within the gut microbiota population can now be accurately determined using molecular techniques, the role of abundance in the overall dynamics of the community is not known; microbes within the numerical minority may have disproportionately large roles in microbiota- host physiology (Marchesi, 2011).

1.2 Commensals, Health and Immunity

As home to a vast diversity of autochthonous microbes, the gut epithelium requires varied and intricate mechanisms to allow it to maintain a balance between immune system tolerance and immune system activation. These mechanisms can be broadly categorized into

2 barrier production and innate immune defences. The barrier in this case refers to the presence of tight junctions between epithelial cells, and a thick mucus layer across the luminal side of the epithelium, acting as a protective layer (Corazziari, 2009). On the other hand, innate immune defences are much more diverse, ranging from sampling of luminal antigens through Microfold

(M) cells or dendritic cells, to the release of antimicrobial peptides (Kagnoff & Eckmann, 1997;

Shanahan, 2002). The innate immune system is known to be able to discern resident ‘harmless’ commensals from pathogens through interpretation of Microbe Associated Molecular Patterns, or

MAMPs, a process that confers immune tolerance in a healthy host. Specialist host cell receptors for MAMPS, such as Toll Like Receptors (TLRs) or Nucleotide-binding Oligomerization

Domains (NODs), act as Pattern Recognition Receptors (PRRs). PRR’s are an interface between the microbial environment and the host immune response and can both stimulate the activation of the immune system or promote immunogenic tolerance depending on prior host exposure to the microbe (Hooper & Gordon, 2001; O’Hara & Shanahan, 2006). Innate immune system activation generally leads to clearance of the offending antigen and its pathogen source through mechanisms such as inflammatory mediator release and apoptosis of damaged cells (Macpherson

& Harris, 2004; Kelly & Conway, 2005). Aberrant immune responses to commensal antigens, the failure of tolerance, and subsequent inappropriate activation of the mucosal immunity underlies manifestations of various inflammatory diseases (O’Hara & Shanahan, 2006).

1.3 Inflammatory Diseases in the Gut

It is generally believed that the pathogenesis of inflammatory diseases is related to a combination of the genetics of the host, the microbes with which the host is colonized, and the environment in which both exist. One theory of inflammatory disease suggests that genetic mutations in pivotal genes involved in the cell cycle, apoptosis and/or DNA repair can cause

3 subsequent abnormalities in immune system function (Rowland, 2009). Such abnormalities may manifest as a ‘loss of tolerance’ response. A second theory suggests that the presence of a specific pathogen that uses virulence determinants to override the host immune response could result in a general inflammatory response (Kelly et al., 2005). Finally, changes in the genetic make-up of the commensal microbial population through such processes as e.g. horizontal gene transfer from allochthonous (transient) microbes could result in functional changes in the microbes that skew the balance between pathogens and commensals. Such an imbalance could result in a hypersensitive immune response and chronic inflammation because of the enhanced ability of such opportunistic pathogens to outcompete, adhere, invade, release toxins and otherwise stimulate an immune response (Ley et al., 2006). When the balance of microbes is skewed in such a way, this is referred to as dysbiosis.

Inflammatory bowel disease (IBD) is an umbrella term used for a group of chronic, relapsing-remitting GI disorders, generally divided clinically into two diseases: ulcerative colitis and Crohn’s disease. Both of these forms of IBD result in extensive inflammation and ulceration of host tissue but they differ in histology and lesion range. The inflammation of ulcerative colitis is limited to the mucosal lining of the colon, whereas in Crohn’s disease inflammatory lesions can span the entire GI tract, from mouth to anus, and ulcerations can penetrate beneath the mucosal lining to the serosa below (Hendrickson et al., 2002).

The development of IBD is known to be intimately connected to the gut microbiota.

Swidsinski et al., (2002) showed that there were higher amounts of bacterial attachment to epithelial cells in IBD patients compared to healthy individuals indicating that changes in the microbiota of patients with IBD is not secondary to inflammation but a consequence of a specific

4 host response. Several studies have been done to demonstrate the involvement of various pathogens in IBD, and examples are briefly given below:

• Mycobacterium avium paratuberculosis (MAP) was the first organism implicated

as a pathogen in IBD. It has been argued that this pathogen is the causative

organism underlying Crohn’s disease with respect to Koch’s postulates; cultured

human MAP administered orally to various animals in vivo results in intestinal

and mesenteric inflammation characteristic of Crohn’s disease (Greenstein, 2003).

• Adherent-Invasive E. coli (AIEC) is a suspect causative organism in ileal Crohn’s

disease (Darfeuille-Michaud et al., 2004). This bacterium has the ability to invade

intestinal epithelial cells and reside within human macrophages. In this case, the

macrophages respond by releasing high levels of tumour necrosis factor alpha

(TNF-α), a key cytokine in intestinal inflammation recognized as being released

in large amounts by Crohn’s patients (Murch et al., 1993).

• Fusobacterium varium has been implicated in ulcerative colitis. F. varium was

isolated from ulcerative colitis lesions and proved to be cytotoxic when tested in

vitro with Vero cells (a kidney epithelial cell lineage). The culture supernatant

was analysed and butyric acid was found in large quantities which could be

potentially toxic in high concentrations, and was able to cause ulcerative colitis-

like lesions in mice (Ohkusa et al., 2003).

Although several interesting observations and correlations of these microbes and inflammation in IBD have been clearly demonstrated, what is less understood is how these

5 pathogens initiate and perpetuate inflammation in IBD, and whether these microbes are responsible for inflammation in all forms of IBD or are restricted to a certain disease subtype.

1.4 Inflammatory Bowel Disease and the associated risk of Colorectal Cancer

Although few infectious agents have been explicitly linked to cancer, it is estimated that

15% or more of the worldwide cancer burden is attributed to mostly unidentified infectious agents (Castellarin et al., 2011). It has been shown that individuals with IBD have an increased associated risk of developing colorectal cancer that is thought to be related to the duration and severity of inflammation in the GI tract (Hope et al., 2005). Many factors are at play in the development of colon carcinogenesis, especially host genetic mutations; however, since microbes provide a major stimulus in activation and development of intestinal inflammation, it is possible that microbes may also play a large role in the initiation and/or perpetuation of colorectal cancer. Sellon et al., used a well characterized IL-10 deficient rodent model (lacking in the ability to develop mature hematopoietic immune cells) to prove that the microbiota is a pre-requisite for the development of colonic inflammation. Only IL-10 knockout mice exposed to microbes developed spontaneous colitis, unlike germ free IL-10 knockout mice which developed no inflammation. Swidsinski et al. (1998) determined the presence of intracellular bacteria in colorectal cancer biopsies and not in the asymptomatic controls indicating a possible association between bacteria and carcinogenesis. Kado et al., went on to use the TCRβ/p53 double knockout mouse colitis model (mimics the development of an adenocarcinoma in ulcerative colitis) and clearly demonstrated a connection between bacteria, inflammation and development of colon cancer. Such chronic inflammation in the intestines, caused by microbes, may actually provide a desirable microenvironment for altered molecular events during the adenoma-carcinoma progression in CRC (Sansone & Bromberg, 2011).

The gut microbiota is capable of inducing a wide range of host processes leading to various neoplastic effects in the colon including: [A] induction of mucosal inflammation by triggering the mucosal immune response; [B] producing mutagens and reactive metabolites, including reactive oxygen intermediates that can cause oxidative DNA damage, and hydrogen sulfide that can cause impaired function in the colonic epithelial barrier; and [C] alteration in carbohydrate expression resulting in changes in the mucous layer and therefore changes in proliferative, metastatic and invasive potentials of tumor cells (Hope et al., 2005). Overall, abundant evidence exists implicating the gut microbiota as a key player in the development of

IBD and more recently, colorectal cancer.

1.5 Fusobacterium nucleatum

The family Fusobacteriaceae can be divided into nine genera including Fusobacterium.

Fusobacterium can be further divided into fourteen species, several of which are known pathogens. Fusobacterium nucleatum (Fn) is the most frequently isolated species from humans and contains five subspecies: animalis, polymorphum, fusiforme, vincentii and nucleatum. Recent advances in molecular techniques continue to redefine bacterial taxonomy making

Fusobacterium a more concise genus and suggesting the Fn subspecies fusiforme belongs to the subspecies vincentii (Citron 2002). Fn is a non-spore forming, non-motile, anaerobic, Gram negative microbe with an approximate genome size of 2.4 X 106bp and a 27% G+C content, based on sequenced genomes of various strains (Bolstad et al., 1996). It is a fermentative anaerobe that uses peptides as energy sources (Mira et al., 2004). Fn has been implicated in a wide variety of inflammatory conditions including, but not limited to, endocarditis, septic arthritis, bacteraemia, pleuropulmonary infections and, due to it ubiquitous presence in the human oral cavity, gingivitis and peridontitis (Bolstad et al., 1996). It has also been associated

7 with idiopathic preterm labour, often being found exclusively in the amniotic fluid of women delivering preterm (Citron, 2002).

The association of any particular subspecies of Fn with disease has not yet been proven straightforward. Fn is a highly heterogeneous species genomically, phenotypically, phylogenetically and biochemically, and even antigenic determinants are extremely heterogeneous within subspecies (Bolstad et al., 1996). Analysis of the Fn genomes sequenced to date indicates that the apparent high genetic heterogeneity between strains may be due to mechanisms such as genome plasticity and an enhanced ability to take up and incorporate foreign

DNA. Despite heterogeneity between Fn strains, tools for genetic manipulation exist and have been used successfully to produce: [A] the successful use of an E. coli shuttle plasmid for the transformation of Fn strains (Kapalan et al., 2005); [B] the first demonstration of targeted mutagenesis in Fn using a suicide plasmid (Haake et al., 2006); and [C] the first successful double cross-over allelic exchange mutant to determine the adhesion properties of FadA (Han et al., 2005). Apparent differences in genetic tractability between Fn strains may be due to e.g. different restriction systems, and may be overcome through modifications to transformation techniques (Lui et al., 1979). Sonoporation makes use of sound, typically ultrasonic frequencies, to modify the permeability of cell plasma membranes by employing the acoustic cavitations of microbubbles to enhance delivery of large molecules, such as DNA. The use of sonoporation has demonstrated potential for bacterial gene transfection in bacterial species where conventional electrical/chemical induction or conjugation has proven futile, including Fn (Han et al., 2005;

Han et al., 2007).

1.6 Fusobacterium nucleatum and Virulence Factors

Fn has a plethora of characteristics that may be described as virulence determinants, including the ability [1] to alter the host immune system (through e.g. pro-inflammatory cytokine and human beta-defensin release) (Han et al., 2000; Santosh et al., 2011), [2] to release tissue irritants and proteases (Leonel & Alvarez-Leite, 2012; Bachrach et al., 2004), [3] to adhere to and invade host tissues (Han et al.,, 2000), and [4] to co-aggregate with itself and other bacterial species (Kolenbrander et al., 1989) to form biofilms.

Within ecosystems, the functionality of the resident microbiota relies on the ability of member species to attach, persist and multiply. To achieve this, the microbiota can exist in a state known as a biofilm. Biofilms are highly organized, dynamic and complex aggregative communities of adherent microbial cells growing on a solid substrate. Multispecies biofilms possess a hierarchal structure whereby primary colonizer species adhere first to surface substrata and then secondary colonizers can associate with the primary species (Rickard et al., 2003).

Biofilms possess a complex and layered structure (comprised of various microbial species and their array of exopolymeric proteins) which can provide resistance to both physical and chemical stresses (Figure 1.1). This can include the resistance to antibiotics and other conventional antimicrobial agents, as well as a capacity for evading host defences (Costerton et al., 1999). The development of a biofilm depends on both environmental factors such as temperature, pH, osmolarity and nutrient availability, as well as microbial factors such as bacterial attachment via pili, fimbriae or flagellar structures (O`Toole et al., 2000). While biofilms are not required by bacteria for survival, the biofilm state is an adaption that enhances survival, metabolism and propagation of bacteria, especially under adverse conditions (Kolenbrander et al., 2002). Since biofilms can persist within environments, the ability of microorganisms to form biofilms is often

9 considered a virulence-associated trait. Indeed, biofilms have been found to be involved in a wide variety of microbial infections within the human body (Kolenbrander et al., 1989; Rickard et al., 2003).

Figure 1.1 Multispecies biofilms are structured in a predictable hierarchy where a) the surface substrate provides binding sites for primary colonizer species followed by b) the proliferation of primary colonizers and the secretion of exopolymeric proteins that allow the biofilm to develop into a cohesive and protective community. The c) addition of subsequent secondary colonizers, which associate with primary colonizers, as well as the continued secretion of exopolymeric proteins allows for the biofilm to development into a microcolony that when d) conditions remain favourable, continues to mature. Microbes are continuously added or sloughed off over time (Adapted from Rickard et al., 2003).

a) Primary colonizers b) Secreted molecules

Cell division

Substratum Substratum

c) Secondary colonizers d) Mature multispecies biofilm

Substratum Substratum

In the human body, the majority of biofilm research has been associated with the oral cavity. As such, the hierarchal biofilm formation, as well as the role of Fn, within the oral cavity is well documented. The bacterial relationships between Fn and unrelated bacterial species in such communities are highly specific; Fn acts as an important bridge organism between primary and secondary colonizers in the formation of biofilms (Kolenbrander & London, 1993) (Figure

1.2). Since Fn can coaggregate to both primary and secondary colonizers it is able to link two distinct bacterial populations that would otherwise be unable to coaggregate to one another; late colonizers appear to aggregate exclusively with Fn (Kolenbrander et al., 1989; Rickard et al.,

2003). Different Fn strains show predilections for binding to different ranges of partner species

(Kolenbrander & London, 1993; unpublished observations) and may therefore exert a significant influence on biofilm development in distinct ecological niches by dictating the presence of particular colonizers (Kapalan et al., 2009). Since Fn can be isolated from both diseased and healthy sites throughout the human body, it has been suggested that different subgroups of Fn vary in virulence capacity as a consequence of both their own putative virulence gene potentials, and also of the virulence determinants of their binding partners (Allen-Vercoe et al., 2011). This can occur either directly through the bacteria bound to Fn or indirectly through Fn’s acquisition of genes from these binding partners through e.g. horizontal gene transfer (Bolstad et al., 1996;

Kolenbrander & London, 1993).

Figure 1.2 A diagrammatic representation of the proposed bacterial accretion found within dental plaque. Here Fn is depicted as a bridge between early colonizers and late colonizers.

Research is currently underway to determine if a similar bacterial accretion is present within the

GI tract and what species are involved (Kolenbrander et al. 2002).

Fn has been shown to be an invasive bacterial species (Strauss et al., 2011); first, adherence to host cells takes place, involving adhesins such as FadA (Han et al., 2005).

Associated with both attachment and subsequent invasion, FadA is an autonomous protein that uses its secreted and non-secreted forms to anchor, oligomerize, attach and interact with host cells (Xu et al., 2007; Ikegami et al., 2009). FadA adheres to vascular-endothelial or epithelial cadherins (Fardini et al., 2011; Rubinstein et al., 2013). FadA binding to the EC5 domain of these calcium-dependant host glycoproteins triggers the actions of the β-catenin/Wnt signalling transduction pathway, which, in turn, internalizes the bacterium into the host cell (Gumbiner,

2005; Rubinstein et al., 2013). Invasive Fn cells appear to use a unique mechanism to enter host epithelial cells pole first, using a zipper mechanism mediated by adhesins localized on the bacterial surface (such as FadA) (Han et al., 2000, Strauss et al., 2011). It appears that in vitro invasion depends on the host’s actin cytoskeleton, as this has been shown using both HaKaT

(human skin keratinocytes) and human gingival epithelial cell lines as model cell lines for invasion assays (Gursoy et al., 2008; Han et al., 2000). Based on in vitro immunofluorescence studies it has been shown that once internalized within host cells, Fn appears to proliferate within membrane-bound vacuoles as early as four hours post infection. Fn remains viable in the host for up to twenty-four hours or more (depending on cell line and Fn strain used), and seems to localize in the peri-nuclear region (Strauss 2011). As an obligate anaerobe it is unclear how Fn is able to survive in this niche for extended periods of time, but it has been shown that bacterial aerotolerance genes are up-regulated during the infection/invasion process (Silva et al., 2005).

Surprisingly, Fn does not appear to have cytopathic effects on host cells even after extended periods of incubation (Ji et al., 2010).

Interestingly Edwards et al., (2006) have shown that Fn is not only able to coaggregate with non-invasive bacteria, such as Streptococcus cristatus (a common oral bacterial species), but also to facilitate the internalization of this normally non-invasive streptococcal species into epithelial cells in vitro. Akin to the study done by Edwards et al.,, in her doctoral work Strauss

(2011) used differential fluorescence microscopy with a Caco-2 cell model and showed that Fn

7/1, a strain isolated from the inflamed Sigmoid colon tissue from a male Crohn’s disease patient, was indeed able to transport the non-invasive strains, Streptococcus gordonii [5/1/36

FAA], Weeksella sp. [2/1/11] and Pediococcus acidilactici [7/4 A], into these epithelial cells in vitro. The ability of a bacterium to invade host cells is considered a virulence factor. It is possible that the ability of a commensal bacterial species to coaggregate with an invasive species and become internalized may contribute to the pathogenesis of an inflammatory disease.

Fn has the ability to directly affect the host’s immune system. Fn is pro-inflammatory in human oral epithelial cells, eliciting the secretion of the various chemokines including CXCL8

(previously known as IL-8), IL-6, IL-1β and TNF-α (Han et al., 2000; Stathopoulou et al., 2011;

Allen-Vercoe et al., 2011). Fn has also been implicated in the direct destruction of lymphatic cells by the outer membrane protein RadD. RadD shares homologous regions to type Va secretion autotransporters and is believed to play a role not only in facilitating adherence between bacteria and host, but also in depleting the host immune response by triggering an apoptotic cascade in host lymphocytes (Kapalan et al., 2010). In addition to the pathogenic ability of Fn to coaggregate, invade and alter the host immune system, Fn may also play a role in pathogenicity through the release of tissue irritants and proteases. Fn produces and releases copious amounts of butyrate, proprionate and ammonium ions, which are common tissue irritants that can inhibit the proliferation of fibroblasts and compromise wound healing (Ohkusa et al.,

2003). Butyrate, although known to be anti-tumorgenic and anti-inflammatory when properly metabolized by aerobic respiration, can also act as a histone deacetylase inhibitor in circumstances, such as in a biofilm or a tumour, where anaerobic respiration is inefficient at fully metabolizing butyrate. The metabolic intermediate, butyric acid, can enter the host cell’s nucleus where it acts as a histone deacetylase inhibitor which results in epigenetic regulation. Butyric acid can result in hyperacetylation and maintenance of less condensed chromatin structure, suppresses NF-kB (a pro-inflammatory transcription factor that plays a key role in eliciting the host’s immune response to non-self antigens) and can also aid in the activation of genes implicated in apoptosis, cell differentiation and cell cycle arrest in cancer cells (Leonel &

Alvarez-Leite, 2012). The degradation of host proteins by Fn enzymes may also contribute to the tissue damage and inflammation seen in many of the diseased sites inhabited by Fn. Through the release of effector proteins such as the 65kDa serine protease, Fn is capable of degrading native host proteins, including Immunoglobulin A, a key player in the immediate response of the host innate and adaptive immune system (Bachrach et al., 2004). Furthermore Fn produces a large amount of the toxic metabolite hydrogen sulfide (H2S). It has been demonstrated that H2S damages epithelial cells, increases the permeability of mucosa and is associated with the modification and release of haemoglobin in erythrocytes, endotoxin-induced inflammation and the apoptosis of smooth muscles cells and fibroblasts (Yoshida et al., 2010).

1.7 Fusobacterium nucleatum and Inflammatory Bowel Disease

A key question yet to be answered concerning the etiology of IBD is whether dysbiosis and functional alterations in the commensal microbiota cause the inflammatory disease, or if the introductions of one or multiple pathogenic species are to blame. The complexity and diversity

16 of microbes in the gut make it hard to pinpoint exact species and strains associated with the development of IBD.

Loss of mucosal homeostasis in IBD may be due to damage and loss of integrity of the mucosal barrier. Fn is a recognized opportunistic pathogen that previously may have been overlooked as a gut mucosal-associated microbe. Since a subset of Fn strains have been associated with attachment and invasion capabilities in epithelial and endothelial cells both in vitro and in vivo (Han et al., 2000; Han et al., 2004), it is thought that this bacterium may be a perpetrator of barrier dysfunction.

As well as potential invasive capabilities, Fn may also have specific effects on host barrier function. Recently it was shown that invasive strains isolated from IBD patients evoked a greater mucin gene and TNF-α gene expression than minimally invasive strains from healthy controls. Also MUC2 (a component of human mucin) was shown to be more readily secreted from Fn-infected epithelial cells (Dharmani et al., 2011). Alterations in mucin gene expression and protein secretion may alter the physical barrier between luminal antigen and the epithelial cells (Dharmani & Chadee, 2008). Bacterial disruption of the mucus barrier could be one way in which Fn gains a foothold in its host, and encourages the colonization of the epithelium by other microbial species. Further pro-inflammatory effects involving upregulation of pro-inflammatory cytokine genes may act synergistically with such a mucin-depleting mechanism, perpetuating inflammation (Dharmani et al., 2011).

A study by Strauss et al., (2011) enumerated the presence of Fn from biopsy material from IBD patients and healthy patients undergoing colon cancer screening. The study demonstrated a significant recovery of Fn strains from mucosal tissue, and furthermore strains of

Fn isolated from the GI mucosa of IBD (Crohn’s Disease) patients were invariably more invasive in a Caco-2 cell line model (representative of the intestinal epithelium) (Sambuy et al.,

2005) compared to strains isolated from healthy controls.

Despite its presence in the biopsies of IBD patients, ultimately more studies are required in order to establish a role for Fn in the etiology of IBD.

1.8 Fusobacterium nucleatum and Colorectal Cancer

Colorectal cancer is the second leading cause of cancer deaths world-wide and is the best characterized tumour type in terms of somatic mutations causing the progression to invasive carcinomas (Vogelstein et al., 1988). A colorectal tumour is essentially a complex community made up of genomically altered cells, non-neoplastic and immune cells, and in some cases a diverse array of microorganisms (Kostic et al., 2011). There is an emerging theory implicating the role of so-called `alpha-bugs` in epithelial transformations that can lead to cancer. ‘Alpha- bugs’ consist of select members of the microbiota that provide key virulence and pro- carcinogenic factors capable of remodelling the diversity and abundance of the microbiota to drive inflammatory responses (Sears & Pardoll 2011). So far, few infectious agents have been directly linked to cancer, but examples such as Helicobacter pylori and its role in gastric cancer have sparked interest in the search for a bacterial ‘alpha-bug’ relating to colorectal cancer.

In 2011, two independent studies (Kostic et al., 2011; Castellarin et al., 2011) implicated an association between the gut microbiota and colorectal carcinoma. Both studies used metagenomics to probe and detect infectious causative agents, ‘alpha bugs’, associated with colorectal cancer. Both groups extracted DNA from colon cancer tumours and matched normal tissues. Using high-throughput sequencing technologies and enhanced bioinformatic algorithms

18 they were able to subtract human gene sequences allowing for in-depth evaluation of remaining microbial gene fragments. Both groups subsequently used targeted quantitative PCR assays and

16S rRNA gene sequence analysis to interrogate, verify and confirm their findings; in both studies, a marked and highly statistically significant over-representation of Fn gene sequences in tumour biopsies compared to normal tissue was found. Kostic et al. detected Fn in situ in the tumour tissues of 95 paired samples. They also commented on shifts in the microbiome within tumours including the depletion of Bacteroidetes and Firmicutes, two ubiquitous phyla of the human gut. It was proposed that contribution of Fn to tumorgenesis may stem from its pro- inflammatory mechanisms and ability to invade. The second study by Castellarin et al., (2012) further reported on a recovered, invasive strain of Fn, CC53, which was isolated from a frozen biopsy sample. It must be noted that although CC53 appears to be invasive, whether the invasion mechanism is similar to that of other invasive isolates is yet to be elucidated.

Over the last four years research on the association between Fn and CRC has been studied in-depth. Various articles have been released verifying the over-abundance of Fn in tumour biopsies from CRC patients (compared to match normals) in North America and across the globe (Chen et al., 2012; Kostic et al., 2013; McCoy et al., 2013; Tahara et al., 2014; Ito et al., 2015; Allali et al., 2015)). In addition, a study by Warren et al., (2013) extended this work with deeper sequencing of a much larger number of colorectal carcinoma and matched normal control tissues. They confirmed tumour over-representation of Fusobacterium species and observed significant co-occurrence within individual tumours of Gram-negative (previously recognized as constituents of the oral microbiome) Fusobacterium, Leptotrichia and

Campylobacter species.

Other pivotal studies on Fn and its association with CRC include the study by Tahara et al. (2014) where they reported that CRC patients with a high abundance of Fusobacterium spp. in their tumours have a specific type of cancer at the molecular level: they observed a high rate of somatic mutations in KRAS, BRAF, CDH7 and CHD8 genes and CpG island methylation.

This suggests that Fn should be treated as a risk modifier bug in populations with pre-existing environmental and genetic predispositions. Flanagan et al. (2014) looked at survivability of CRC patients with Fusobacterium spp. present in their tumours. Interestingly those with a high fold increase of Fusobacterium spp. had a lower survivability rate. This suggests that the changes in tissues during colorectal tumourgenesis may provide an environment within which Fn is better able to multiply. Interestingly, the study by Rubinstein et al. (2013) was able to prove that the unique FadA receptor proteins on Fn were able to attach and invade CRC cells and induce an inflammatory and oncogenic response. This adherence, and subsequent invasion, is accomplished through the binding of FadA to the E-cadherin host receptors, thereby initiating the

β-catenin/Wnt signalling pathway. Finally, Kostic et al. (2013) released a comprehensive study that used mouse models to show that Fn can actually accelerate the onset and frequency of colorectal tumorgenesis. They were able to show that Fn exerts its tumorigenic effect by acting downstream of the loss of tumour suppressor adenomatous polyposis coli (APC) in the adenoma- carcinoma sequence. They also demonstrated that a previous downstream mutation in the APC precedes adenoma enrichment of Fn as these mutations contribute to the improper development of the epithelial barrier and mucus layer. Kostic et al. (2013) further suggest through the Fn release of chemoattractants (e.g. SCFAs) the direct recruitment of tumour infiltrating immune cells such as myeloid-derived suppressor cells, tumour-associated neutrophils, granulocytes, conventional dendritic cells, classical myeloid dendritic cells and tumour-associated

20 macrophages Fn can actually generate a microenvironment that is pro-inflammatory in nature which is conducive to neoplastic progression.

Ultimately, the association of Fn with the colonic mucosa of colon carcinomas is not indicative of a causal relationship but may in fact be representative of an opportunistic infection at an immunocompromised host site. The role of Fn in carcinogenesis deserves further study to properly define the microbe’s role (if any) in colorectal cancer pathogenesis. In addition, the microbes to which Fn aggregates should be considered in any emerging model of carcinogenesis.

1.9 Conclusion

Fn is an enigmatic species. Comparative analysis of multiple Fn genomes has so far given no clues to the existence of pathogenicity islands/loci likely virulence genes. In addition, more direct approaches to virulence gene identity are required. Fn is associated with a vast number of human disease manifestations, and its most recent association with colon cancer underlines the need to understand the pathogen more completely. Further research in the coaggregation and coinvasion of Fn with other bacterial species may give clues about Fn’s role in disease. It will also be important to understand the reasons behind the observed differences between Fn strains including, for example, determining the species able to coaggregate, the degree to which coaggregation results in coinvasion, the fate of internalized bacteria, the resultant response by the host immune system and the genes responsible for aggregation and invasion into the host. Further research will help to understand the association of Fn colonization with human health.

1.10 Research Objectives

The overall objective of my doctoral research was to elucidate potential virulence determinants from Fusobacterium nucleatum isolates from the human gut, in particular those isolated from patients with inflammatory bowel disease and/or colorectal cancer. My specific aims were to [1] isolate Fn from human intestinal mucosal biopsies of colorectal cancer patients and characterize these Fn CRC gut isolates phenotypically and genotypically, [2] to compare genomes of more and less invasive Fusobacterium isolates to identify potential virulence determinants, [3] to elucidate the involvement of bacteriophage in gene acquisition in Fn and [4] to elucidate the genes responsible for both adhesion and invasion of Fn into host epithelial cells.

Chapter 2. Materials and Methods

Recipes for bacterial growth media and buffers used throughout this thesis can be found in Appendix I.

2.1 Isolation of bacterial species from colorectal cancer biopsies and characterization of isolated Fusobacterium nucleatum (Fn) species.

2.1.1 Bacterial Strains

Table 1 lists bacterial strains used in this thesis and their sources. On initial characterization, some strains isolated from clinical samples were found to be likely novel species. Full characterization of these strains for appropriate description as novel species was beyond the scope of the work described in this thesis, and thus potential novel species were identified only to the clearest taxonomic level based on 16S rRNA gene based sequencing.

Table 2.1 Bacterial strains used in this thesis

Name Species Source

ATCC 25586 Fn subsp. nucleatum (NC_003454.1) ATCC 49256 (NZ_AABF02000298.1) Fn subsp. vincentii

ATCC 51190 Dzink et al., 1990 (NZ_AKXI01000190.1)

ATCC 10953 Fn subsp. polymorphum (NZ_AARG00000000.1)

ATCC 25563 Fn subsp. gonidiaformans Skerman et al., 1980 (NZ_KQ235756.1)

NCTC 12276 Fn subsp. animalis Gharbia and Shah, 1992 (NR_026084.1)

7-1 (EAVG_002) Fn subsp. animalis (HMP ID: 410)

4-8 (EAVG_001) (HMP ID: 409)

3-1-33B (EAVG_016) (HMP ID: 406)

21-1A (EAVG_006)

(HMP ID: 404)

2-1-50A (EAVG_027) Strauss 2011

3-2-4 (EAVG_009)

11-3-2 (EAVG_003) Fn subsp. nucleatum (HMP ID: 40)

13-3C (EAVG_005) (HMP ID: 2085)

4-1-13 (EAVG_012) Fn subsp. vincentii (HMP ID: 408)

3-1-36A2 (EAVG_018) (HMP ID: 946)

3-1-27 (EAVG_014) (HMP ID: 405)

CC53 Fn subsp. vincentii Castellarin et al., 2012

CC 7/4 Fmu3 Fn subsp. animalis

CC 7/3 JVN 3C1

CC 7/5 JVN 1A4

CC 2/6 JVN3

CC 2/3 Fmu1 Fn subsp. nucleatum

CC 2/3 FmuA

CC 2/1 JVN3

CC 57C Campylobacter showae Human intestinal CRC biopsy isolates CC 7/5 D55 Campylobacter concisus (this study)

CC 57F Gemella haemolysans

CC 7/3 CNA4

CC 57A Parvimonas micra

CC 7/3 MET1

CC 92I Eikenella corrodens

CC 94A Streptococcus oralis

CC 3/4 CNA1 Clostridium perfringens Gift from the Wadsworth WAL14542 Anaerobe lab

2.1.1 A. Study population and clinical samples

Colorectal cancer (CRC) biopsy specimens were handled with very close attention to maintaining integrity and isolation:

 Collection time (time from biopsy removal from patient to cryopreservation in

liquid nitrogen) for all CRC cases in the BC Cancer Agency Tumor Tissue

Repository (BCCA-TTR) was 30 min

 Biospecimens were held briefly at 20°C during frozen sectioning, using 100%

ethanol to clean the blade between all samples

 Once sectioned, biopsies were immediately placed in pre-reduced freezing media

containing DMSO and glycerol (Appendix I) and snap-frozen in liquid nitrogen

before storage at -80ºC until shipment

 Samples were shipped on dry ice from the BCCA to the University of Guelph in

three separate cohorts on dry ice and stored again at -80ºC at Guelph until

processed (Section 2.1.1 C)

 Cohort One: 12 patients, all CRC tumour sites (Table 2)

 Cohort Two: 30 patients; 2 biopsies from each patient (one from the CRC tumour site,

and one from nearby healthy biopsy) (Table 3)

Note: Cohort One biopsy samples were received and processed prior to my tenure

in the Allen-Vercoe lab

 Cohort Three: 8 patients; 6 biopsies from each patient (two from the CRC tumour site,

and four from nearby healthy tissue) (Table 4)

All received biopsy tissue samples were assigned unique identification numbers by the study coordinator so that microbiology work at Guelph could be done blinded (and thus without bias).

At the culmination of the culture and characterization work, the status of the biopsy (i.e. the nature of its source - tumour or attached healthy tissue) was revealed.

Table 2.2 Cohort One Metadata

Patient Tumour TTR Patient Patient Disease State of Patient No. ID Age Gender (Project-case- specimen)

14 0005-014-001 74 F Infiltrating undifferentiated adenocarcinoma 31 0005-031-001 50 M Poorly differentiated adenocarcinoma 42 0005-042-001 68 F Adenocarcinoma of sigmoid colon 53 0005-053-001 81 F Invasive mucinous adenocarcinoma 57 0005-057-001 58 F Poorly differentiates/mucinous adenocarcinoma 67 0005-067-001 64 F Invasive colonic adenocarcinoma 70 0005-070-001 73 M Moderately differentiated adenocarcinoma 87 0005-087-001 71 M Invasive adenocarcinoma of proximal ascending colon 88 0005-088-001 72 M Invasive adenocarcinoma of sigmoid colon 92 0005-092-001 63 M Synchronous mucinous adenocarcinoma 94 0005-094-001 80 F Poorly differentiated invasive mucinous adenocarcinoma of the rectosigmoid colon 98 0005-098-001 68 F Poorly differentiated adenocarcinoma

Table 2.3 Cohort Two Metadata

Patient Tumour Normal Patient Patient Disease State of No. TTR ID TTR ID Age Gender Patient (Project-case- (Project-case- specimen) specimen)

13 0005-013- 0005-013- 77 F Adenocarcinoma of 001 002 ascending colon 16 0005-016- 0005-016- 62 M Invasive 001 002 adenocarcinoma of lower rectum 19 0005-019- 0005-019- 50 F Adenocarcinoma of 001 002 colon 21 0005-021- 0005-021- 74 M Synchronous colonic 001 002 adenocarcinoma of ascending colon and splenic flexure 32 0005-032- 0005-032- 78 F Adenocarcinoma of 001 002 rectum 33 0005-033- 0005-033- 75 F Mucinous 001 002 adenocarcinoma 40 0005-040- 0005-040- 74 M Adenocarcinoma of 001 002 ascending colon 43 0005-043- 0005-043- 54 F Invasive 001 002 adenocarcinoma of cecum 44 0005-044- 0005-044- 59 M Invasive mucinous 001 002 adenocarcinoma of cecum 51 0005-051- 0005-051- 83 M Well differentiated 001 002 adenocarcinoma from vilous adenoma 53 0005-053- 0005-053- 81 F Invasive mucinous 001 002 adenocarcinoma of ascending colon 54 0005-054- 0005-054- 76 F Invasive mucinous 001 002 adenocarcinoma of ascending colon 55 0005-055- 0005-055- 73 M Infiltrating mucinous 001 002 adenocarcinoma of

cecum 58 0005-058- 0005-058- 74 M Adenocarcinoma 001 002 59 0005-059- 0005-059- 74 M Adenocarcinoma of 001 002 ascending colon 62 0005-062- 0005-062- 61 M Infiltrating well 001 002 differentiated adenocarcinoma 64 0005-064- 0005-064- 82 M Invasive colonic 001 002 adenocarcinoma of cecum 66 0005-066- 0005-066- 78 M Adenocarcinoma 001 002 71 0005-071- 0005-071- 84 M Adenocarcinoma of 001 002 ascending colon 77 0005-077- 0005-077- 70 M Invasive 001 002 adenocarcinoma of descending colon 78 0005-078- 0005-078- 80 M Invasive 001 002 adenocarcinoma of ascending colon 83 0005-083- 0005-083- 54 F Invasive 001 002 adenocarcinoma of cecum 85 0005-085- 0005-085- 74 F Well differentiated 001 002 mucinous colonic adenocarcinoma 89 0005-089- 0005-089- 71 M Invasive 001 002 adenocarcinoma of transverse colon 91 0005-091- 0005-091- NA NA NA 001 002 93 0005-093- 0005-093- NA NA NA 001 002 96 0005-096- 0005-096- NA NA NA 001 002 99 0005-099- 0005-099- NA NA NA 001 002 100 0005-100- 0005-100- NA NA NA 001 002

101 0005-101- 0005-101- NA NA NA 001 002 *NA- not available

Table 2.4 Cohort Three Metadata

Patient No. Tumour Normal Patient Patient Disease State of TTR TTR Age Gender Patient Tumour ID(Project- ID(Project- Biopsies case-specimen) case-specimen)

1 Test-001- Test-001- 66 M Invasive 002 001 adenocarcinoma of Test-001- Test-001- rectum and anus 003 004 Test-001- 005 Test-001- 006 2 Test-002- Test-002- NA F Adenocarcinoma 005 001 mid-rectum Test-002- Test-002- 006 002 Test-002- 003 Test-002- 004 3 Test-003- Test-003- 61 M Invasive 001 003 adenocarcinoma of Test-003- Test-003- ascending colon 002 004 Test-003- 005 Test-003- 006 4 Test-004- Test-004- 70 M Cecal 005 001 adenocarcinoma Test-004- Test-004- 006 002 Test-004- 003 Test-004- 004

5 Test-005- Test-005- NA M Adenocarcinoma 002 001 sigmoid colon Test-005- Test-005- 003 004 Test-005- 005 Test-005- 006 6 Test-006- Test-006- NA M Mucinous 003 001 adenocarcinoma Test-006- Test-006- upper rectum 004 002 Test-006- 005 Test-006- 006 7 Test-007- Test-007- 90 M Invasive 002 001 adenocarcinoma of Test-007- Test-007- ascending colon 003 004 Test-007- 005 Test-007- 006 8 Test-008- Test-008- 70 M Invasive 002 001 adenocarcinoma of Test-008- Test-008- sigmoid colon 004 003 Test-008- 005 Test-008- 006 *NA- not available

2.1.1 B. Isolation of bacterial strains from CRC biopsy samples

Each biopsy (previously suspended in pre-reduced freezing media at the surgical center) was individually thawed and vigorously vortexed. Bacterial species were isolated from resulting suspension by serial dilution and plating on prepared, degassed agar (see Table 5 for complete list of media types used). In particular, for isolation of Fn, Fusobacterium-selective agar plates containing josamycin, norfloxacin and vancomycin (JVN plates) were used (Brazier et al., 1991)

(Table 5). Inoculated plates were then incubated at 37ºC under anaerobic conditions (10% CO2:

10% H2: 80% N2) in an anaerobic chamber (Ruskinn Bug Box) for 5 days. Isolated colonies were further purified by re-streaking on Fastidious Anaerobic Agar (FAA) supplemented with 5% defibrinated sheep’s blood plates (refer to table below). Following purification, crude DNA was extracted and used as a template in PCR for 16S rRNA gene-based speciation as described in

Section 2.1.2 A.

A strain naming system was established to allow easy identification of all strains isolated from a single patient, as well as the biopsy samples in cases where more than 1 biopsy was taken from the same patient. The strains from all three cohorts were designated “CC” (colorectal cancer), to differentiate them from biopsy samples taken from other diseases and used for separate studies in the laboratory. Cohort One isolates were subsequently numbered according to a patient number, and an alphabetical system was used to indicate different isolates obtained from the same biopsy sample. Thus as an example, CC57F is bacterial isolate “F” from CRC patient 57. Naming for bacterial isolates from cohort two, where more than one biopsy was taken

(one directly from the tumour and one from adjacent colon tissue), involved using additional numbers to indicate the biopsy source: in this case the patient number was followed by either 001

(tumour tissue) or 002 (normal tissue). For example, CC44 001G, refers to the bacterial isolate

(“G”) isolated from a tumour biopsy (“001”) from CRC patient 44. Finally, for the third cohort of bacterial isolates, where multiple biopsies were taken from each patient, a slightly different system was used: here, the designation of the media type used for original isolation was also introduced into the strain name. For example, CC 2/3 TSA3 is the third bacterial isolate that was obtained on tryptic soy agar, from biopsy 3 provided by patient 2. Occasionally, following purification it was found that a mixed culture was still present, in which case after further streak- purification steps, the isolates were designated with lower-case letters. For example, CC 3/3

JVN2a and CC 3/3JVN2b would be isolates obtained after such a secondary purification treatment.

Table 2.5 Media types used for bacterial isolations from colon cancer biopsies.

Code Media Type Source FAA- Acumedia; Lansing, MI FAA Fastidious Anaerobic Agar supplemented with 5% defibrinated sheep’s blood Blood- Hemostat Labs; Dixon, CA

JVN FAA supplemented with josamycin, norfloxacin Antibiotics- Fisher Scientific; and vancomycin Burlington, ON

Fmu FAA supplemented with Porcine Mucin Mucin- Sigma Aldrich; Oakville, ON

TSA Tryptic Soy Agar EMD Millipore; Billerica, MA

CNA Columbia Naladixic Agar BD; Mississauga, ON NA Nutrient Agar

NB Nutrient Broth Agar

BHI Brain Heart Infusion Agar

MRS MRS Agar

D5 FAA supplemented with chemostat effluent (McDonald 2013) from a chemostat community at steady state originally inoculated from Donor 5

D6 FAA supplemented with chemostat effluent from a chemostat community at steady state originally inoculated from Donor 6

MET FAA supplemented with chemostat effluent (Schroeter, 2014) from a chemostat community at steady state originally inoculated from a defined microbial community, MET-1*

*MET-1: Microbial ecosystem therapeutic-1, a set of bacterial strains isolated from healthy patients, designed in the Allen-Vercoe lab

2.1.1 C. Fusobacterium spp. strains and growth conditions

A total of 17 Fn isolates were used for investigations throughout this thesis (Table 1).

The panel includes two Fn type strains ATCC 25586 and ATCC 49256 (Dzink et al., 1990) along with 7 Fn strains [7-1 (EAVG_002), 4-8 (EAVG_001), 11-3-2 (EAVG_003), 13-3C

(EAVG_005), 4-1-13 (EAVG_012), 3-1-33B (EAVG_016) and 3-1-36A2(EAVG_018)] isolated from intestinal biopsy specimens from both healthy and inflammatory bowel disease (IBD) patients (Strauss 2011). The remaining strains were obtained from patients with CRC, either directly from tumour tissue or from healthy surrounding tissue as described in section 2.1.2

A&B.

All Fusobacterium strains were stored at -80ºC, in pre-reduced freezing media (Appendix

I). Resuscitation and propagation of all strains was carried out on FAA (Table 5), or in tryptic soy broth supplemented with hemin (5µg/mL) and menadione (1µg/mL) (TSBsupp) (Sigma

Aldrich, Oakville, ON) (Appendix I). All strains were grown under anaerobic conditions at 37ºC in an anaerobic chamber as above (Section 2.1.1 B)

2.1.1 D. Campylobacter spp. strains and growth conditions

Two Campylobacter isolates were successfully isolated directly, 1 from each of two CRC patients (Table 1). A Campylobacter showae strain was isolated from patient 57 and a

Campylobacter concisus strain was isolated from patient 7. Isolation and characterization of both isolates was done using methods described in detail above (Section 2.1.1 B).

Both Campylobacter spp. isolates were stored at -80ºC in pre-reduced freezing media

(Appendix I) and subsequently grown on FAA agar plates and further cultured in TSBsupp with

38 sodium formate (0.2%w/v) and sodium fumarate (0.3%w/v) (TSBsupp+) (Etoh et al., 1993), under strict anaerobic conditions for 2 days at 37°C.

2.1.1 E. Ethical considerations

For all cohorts, informed consent was obtained by the BCCA_TTR (Watson 2010) to enable the collection of biopsy samples. BCCA_TTR operates as a dedicated biobank with approval from the University of British Columbia–British Columbia Cancer Agency Research

Ethics Board (BCCA REB). The BCCA-TTR platform is governed by Standard Operating

Procedures (SOPs) that meet or exceed the recommendations of international best practice guidelines for repositories (NCI Office of Biorepositories and Biospecmen Research, NCI Best

Practices for Biospecimen Resources).

2.1.2 Identification of CRC-derived strains and the characterization of Fn strains

2.1.2 A. Crude DNA extraction from patient biopsies

DNA was extracted from all distinct colonies from each biopsy (a grand total of over

1100 bacterial isolates). DNA extraction was performed by inoculating and re-suspending one

10µL loopful of each distinct bacterial colony in a microcentrifuge tube with 300 µL TE buffer, pH 8.0 (Appendix I). Resuspended cells were then washed by centrifugation at 14000 x g for 2.5 min, the supernatant was decanted and the pellet was resuspended in 300 µL TE buffer. The samples were then vortexed vigorously and then boiled at 100ºC for 20 min to lyse the cells and release the DNA. This crude DNA preparation was then used immediately as template for PCR.

2.1.2 B. Identification of bacterial isolates using PCR and Illumina sequencing

For all bacterial isolates, including potential Fn candidates, the V3 region of the 16S rRNA gene was amplified using the primers V3kl and V6r as described by Gloor et al., (2010).

These primers include a T3 and T7 universal primer sequence at the 5’ ends of the forward and reverse primers, respectively, to enable direct sequencing from amplicons (Table 6). The PCR master mix included a 1X Thermopol reaction buffer (2mM MgSO4) (NEB; Whitby, ON), dNTP’s (Invitrogen; Burlington, ON) at a concentration of 20 µM, and 0.5 µL Taq DNA polymerase (NEB; Whitby, ON), with 1-2 µL of crude DNA extract (Section 2.1.2A) template.

The reaction conditions were: 94°C for 2 min, (94°C for 30s, 60°C 30s, 70°C 30s) x 30 and 72°C for 5 min. The quality of the various amplicons was determined by agarose gel electrophoresis

(Section 2.1.2 C) and each amplicon was separately quantified using a NanoDrop® ND-8000

(Thermo Scientific; Burlington, ON). PCR amplicons were then individually arrayed into 96 well

40 plates (50 µL of each reaction) and sent to the BCCA-TTR for PCR purification and sequencing using the Illumina GAIIx platform. Annotation and analysis was performed at the BCCA-TTR.

All strains isolated on Fn selective JVN plates were verified using a PCR specific for Fn;

Fn-R6 and All-F6 (Han et al., 2005) (Table 6). These primers amplify a 450bp product from Fn

(as well as mortiferum, ulcerans, and varium) strains, and were used at a 20 µM concentration in a PCR master mix (see above) using the following conditions: 94°C for 2 min, (94°C for 1 min,

68°C 30s, 72°C 1 min) x 30 and 72°C for 10 min and product size was determined by gel electrophoresis (Section 2.1.2C)

2.1.2 C. Agarose gel electrophoresis

PCR products were analyzed by electrophoresis using agarose gel electrophoresis.

Agarose was added to TAE buffer (Appendix I) to a final concentration of 1.0% (w/v) for running 16S rRNA gene PCR or fadA gene products and 2.0% (w/v) for running Fn-specific primers. Ethidium bromide (0.5µg/mL) was added directly to the gel before solidification. The

PCR products were mixed with 5X loading dye. A 1 Kb or 100 bp DNA molecular size ladder

(Thermos Scientific; Burlington, ON) was loaded in reference lane(s). Electrophoresis was carried out by applying at 80 volts across the gel for 30 min in 1X TAE buffer. Bands were visualized using a SynGene G-Box gel documentation system running GeneSnap software

(version 6.08.04; PerkinElmer, Montreal, QC).

2.1.2 D. Genomic DNA extraction for full genome sequencing

DNA was extracted from all verified Fusobacterium spp. and Campylobacter spp. isolates using the Maxwell®16 Cell DNA Purification Kit (Promega; Madison, Wisconsin) according to manufacturer’s instructions. Briefly, isolates were individually cultured in 5 mL

TSBsupp for 24-48h, until OD600 reached between 0.8-1.0 and pelleted by centrifugation at

4000rpm for 5 minutes. After decanting the supernatant, the pellet was resuspended in 1 mL of

TE buffer, pH 8.0. The culture was centrifuged and the pellet resuspended in TE twice more to wash the cells. After the final resuspension in TE buffer, RNaseA (Sigma Aldrich; Oakville, ON) was added to a final concentration of 5µl/mL and 400µL of this bacterial suspension was transferred to a Maxwell®16 DNA Purification cartridge. The remainder of the protocol was carried out according to the Maxwell® kit instructions. The DNA was eluted in elution buffer

(Promega; Madison, Wisconsin) and either used immediately for PCR of the full length 16S rRNA gene or stored at -20ºC. The genomic DNA was then sent to the BCCA-TTR for PCR purification and sequencing using the Illumina GAIIx platform.

2.1.2 E. Identification of Fn subspecies and Campylobacter species by Sanger sequencing

For taxonomy verification of Fusobacterium and Campylobacter species, PCR amplification of an 1502bp region of the 16S rRNA gene sequence was carried out using the universal primers U8F and U1510R as described by Muyzer et al., (1993) (Table 5). The primers were used at a concentration of 20 µM. The PCR master mix used was as above (Section 2.1.2

A) with 1-2 µL of extracted genomic DNA as the template (Section 2.1.2 D). The reaction conditions were: 95ºC 3 min, (98ºC 20 s, 50ºC 15 s, 72ºC 1 min) x 35 and 72ºC 5 min. The quality of the amplicons was determined by agarose gel electrophoresis (Section 2.1.2 C), and amplicons were then purified using the EZ-10 Spin Column PCR Products Purification Kit (Bio

Basic; Markham, ON), quantified on the NanoDrop® ND-8000 (Thermofisher; Burlington, ON) and sent for Sanger sequence analysis following a BigDye® Terminator v.3.1 cycle sequencing

PCR (Thermofisher; Burlington, ON) amplification. Sanger sequencing was carried out at the

Advanced Analysis Center at the University of Guelph. Obtained DNA sequences were compared to the GenBank database (NCBI) using BLASTn

(https://blast.ncbi.nlm.nih.gov/Blast.cgi?PAGE_TYPE=BlastSearch), and the resulting hits allowed approximate speciation of the strains according to their closest match in the database.

Table 2.6 Primers used in this study.

Target Primer Sequence (5’-3’) Product Reference name Size

V3 V3kl 5’ ATTAACCCTCACTAAAGTACGGRAGGCAGCAG 3’ ~750bp Gloor et al., region of 2010 16S rRNA V6r 5’ AATACGACTCACTATAGGGACRACACGAGCTGACGA C 3’

16S U8F 5’ AGAGTTTGATYMTGGCTCAG 3’ RNA ~1502bp Muyzer et (almost al., 1993 full U1510R 5’ GGTTACCTTGTTACGACTT 3’ length)

Fn AllF6 5’ CGGGAGGCAGCAGTGGGGAAT 3’ ~450bp Mi-Kwang specific et al., 2004

FnR6 5’ TTGCTTGGCGCTGAGGTTC 3’

fadA FadAF 5’ TTAGGTGTTTCTGCTTCAGC 3’ ~160bp Han et al., 2005

FadAR 5’ TTACCAGCTCTTAAAGCTTG 3’

2.1.2 F. Detection of fadA gene sequences in Fn strains

PCR was performed on all CRC-derived Fn strains to detect the presence or absence of the fadA gene sequence. Fn genomic DNA extracted using the Maxwell® Cell DNA kit (Section

2.1.2 D) was used as template in a PCR reaction with the fadA forward and reverse primers; these amplify a 160bp region of the fadA gene (Han et al. 2003). These primers were used at a 20 mM concentration with a PCR master mix as above (Section 2.1.2 A). The reaction conditions were: 94ºC 10 min, (94ºC 30 s, 55ºC 20 s, 72ºC 45 s) x 45 and 72ºC 10 min. The presence of a fadA gene amplicon of 160bp was observed by agarose gel electrophoresis (Section 2.1.2 C)

2.1.2 G. Fn phenotypic and phylogenetic analyses

Morphological analysis

Fusobacterium spp. strains were scored for colony morphology on FAA agar plates that had previously been incubated at 37ºC under anaerobic conditions (10% CO2: 10% H2: 80% N2) in an anaerobic chamber for 5 days. Colony morphology was scored based on the following key

(Strauss et al., 2008): 1-mucoid; 2-breadcrumb; 3-umbonate; 4-opaque; 5-pitting; 6-flat; 7- irregular; 8-round.

Isolated Fusobacterium spp. strains were grown separately in 5 mL TSBsupp at 37°C in anaerobic conditions to stationary phase, harvested by centrifugation at 4000rpm for 5 min, and washed once with TE buffer, pH 8.0. For visualization of cell morphology, a small sample of each strain was either smeared onto a microscope slide and examined under phase contrast using a Leica DM2000 LED inverted microscope, or dropped onto a copper grid with carbon-coated formvar film, stained with 1% uranyl acetate (Sigma Aldrich; Oakville, ON) and viewed with a

Phillips CM10 transmission electron microscope at 120kV. Mean cell length was determined by

45 calculating the mean measurements of 20 representative cell lengths using ProgRes® Mac

Capture Pro version 2.7.6 (Jenoptik Optical System; Jena, Germany). Cell morphology was checked and scored based on the following key (Strauss et al., 2008): A-filamentous; B-needle- like; C-short-medium tapered bacilli; D-medium-long rods; E-irregular.

Biochemical analysis

Biochemical profiles for all Fusobacterium spp. strains were assessed using the instructions provided in the API Rapid32A strips, according to manufacturer’s instructions (BioMerieux;

Saint-Laurent, Quebec). Sucrose, fructose, galactose and lactose fermentation by the Fn strains was also confirmed using appropriate tab-tests (Key Scientific; Stamford, TX).

Bile resistance was also measured by observing fusobacterial growth (anaerobically for 5 days at 37ºC) on freshly prepared Bacteroides Bile Esculin (BBE) agar (with no added gentamicin) (Appendix I). All test strains were inoculated on the agar and incubated anaerobically for 5 days at 37°C. Bacterial growth on this media indicates resistance to bile at

20% (v/v). Esculin hydrolysis by bile sensitive strains was also assessed using trypticase soy esculin (TSE) agar (Appendix I). TSE agar was separately inoculated with each Fn strain and incubated at 37°C anaerobically for 5 days. Blackening of the media around zones of growth was indicative of a positive reaction for esculin hydrolysis. Bacteroides fragilis (CC2/3 FAA1) was used as a positive control for both bile tests.

All Fusobacterium spp. strains were assessed for hydrogen sulfide (H2S) production using lead acetate soaked filter paper strips as described by Lee and Simard (1984). Blackening of the filter paper strips was indicative of H2S production.

Antibiotic susceptibility testing was performed for each Fn strain using the Kirby-Bauer disc diffusion method on FAA agar (ThermoScientific (Oxoid); Burlington; BD; Mississauga, ON).

Briefly, individual Fn strains were grown in 5mL TSBsupp broth anaerobically for 24 hours and subsequently spread on FAA plates. Discs were then applied to the plates (one disc per plate) for each Fn strain. Zones of clearance were then measured across the full diameter of the clearance zones and discs. The antibiotic concentrations per disk were as follows: gentamicin (400µg), norfloxacin (20 µg), vancomycin (30 µg), josamycin (20 µg), erythromycin (200 µg), clindamycin (2 µg), imipenem (10 µg), amoxicillin (5 µg), ampicillin (20µg), piperacillin (100

µg), ceftriaxone (30 µg) and ciprofloxacin (5 µg).

Phylogenetic analysis

For phylogenetic analysis of Fn subspecies, the DNA sequences generated by Sanger sequencing of the 1502bp region of the 16S rRNA gene from Section 2.1.2 E. were compared using the ClustalW algorithm (Thompson 1994) on Geneious version 8.1 software (Biomatters;

Auckland, NZ). Phylogenetic trees, corresponding to the alignments, were then constructed using the neighbor-joining method (Saitou and Nei, 1987), and bootstrap analysis (Felsenstein, 1985) of 1,000 permutations was used to verify branch points on all of the trees (Geneious).

2.1.2 H. Campylobacter spp. phenotypic and phylogenetic analyses

Morphological analysis

Both Campylobacter showae (CC57C) and Campylobacter concisus (CC7/5 D55) were assessed for colony morphology on FAA agar plates that had been incubated at 37ºC under anaerobic conditions (10% CO2; 10% H2: 80% N2) in an anaerobic chamber for 5 days. Both strains were grown separately in 5 mL TSBsupp+ at 37°C in anaerobic conditions to stationary

47 phase, harvested by centrifugation at 4000rpm for 5 min, and washed once with TE buffer, pH

8.0. For visualization of cell morphology and flagella structures, a 10µL sample of each strain was dropped onto a copper grid with carbon-coated formvar film, stained with 1% uranyl acetate

(Sigma Aldrich; Oakville, ON) and viewed with a Phillips CM10 transmission electron microscope at 120kV. Additionally, to test for the presence or absence of an extracellular capsule, a 5µL sample of each Campylobacter strain was re-suspended in a drop of 1% (w/v)

Congo red, dried, subsequently stained with 1% (w/v) cobalt blue (Hughes & Smith, 2013), washed with de-ionized H2O and then examined under phase using a Leica DM2000 LED inverted microscope.

Biochemical analysis

Antibiotic susceptibility testing was performed for both Campylobacter strains using the

Kirby-Bauer disc diffusion method on FAA agar (ThermoScientific (Oxoid); Burlington; BD;

Mississauga, ON). Briefly, both CC57C and CC7/5 D55 were grown in 5mL TSBsupp+ anaerobically for 24 hours and subsequently plated on FAA plates. Discs were then applied to the plates (one disc per plate) for each Campylobacter strain. Five clinically relevant antibiotics were used at the following concentrations: gentamycin (400µg), vancomycin (30 µg), erythromycin (200 µg), ampicillin (20µg), and ciprofloxacin (5 µg).

Phylogenetic analysis

For phylogenetic analysis of CC57C and CC7/5 D55, the DNA sequences generated by

Sanger sequencing of the 1502bp region of the 16S rRNA gene from Section 2.1.2 E. were compared using the ClustalW algorithm (Thompson 1994) on Geneious version 8.1 software

(Biomatters; Auckland, NZ). Alignments and tree construction was carried out as above, section

2.1.2 H.

2.1.2 I. Co-aggregation assays by spectrophotometric analysis and phase contrast microscopy

A total of 8 bacterial strains isolated from CRC biopsies, chosen because of their relevance in oral and gut disease etiology, were examined for their ability to co-aggregate with

CRC-derived Fn strains, CC53 CC7/4 Fmu3, and the IBD-derived Fn strain, 7-1 (Table 1).

Bacterial co-aggregation is often measured using spectrophotometric techniques by using the optical density of bacteria in a cuvette. A decrease in OD of two strains incubated together compared to the OD of cultures of the individual strains alone is indicative of co-aggregation because of the increased settling of heavier aggregates to the bottom of the cuvette (Gibbons and

Nygaard, 1970; Handley et al.,, 1987). In addition to using OD, co-aggregation was also examined using phase microscopy. Briefly, all strains were grown in 5 mL TSBsupp under anaerobosis for 1-3 days until the cultures were turbid. Cells were then harvested by centrifugation at 10,000 rpm for 10 min at 4°C, washed in co-aggregation buffer (Appendix I)

(Cisar et al, 1979) three times and finally resuspended and adjusted to a McFarland standard of turbidity between 2.0-3.0 (~108 CFU/mL) (Washington et al., 1972). Equal volumes (0.5 mL) of a Fn strain and the CRC-derived partner strain were added to a 1.5 mL eppendorf tube and vortexed vigorously for 10 s. The tubes were then placed in a shaking incubator at 110 x g for 30 min at 37°C. Individual strains were incubated alone and these served as controls. Following incubation, aggregate cultures were either A) transferred to sterile cuvettes for OD measurements using a Beckman Coulter DU®730 UV/Vis spectrophotometer (Beckman Coulter; Mississauga,

ON); or B) immobilized on agarose-coated microscopy slides and observed by phase microscopy on a Leica DM2000 LED inverted microscope using a Jenoptik ProgRes®CT3 camera. In

49 addition, all suspensions were assigned a score from 0-4 where the scales was as follows: 0, no visible aggregates formed; 1, very small uniform aggregates formed in suspension; 2, definite co- aggregation was seen, but suspension remained turbid; 3, large co-aggregates formed with some settling; and 4, large co-aggregates formed which settled quickly, leaving a clear supernatant

(Cisar et al., 1979; Handley et al, 1985). Each aggregation between two strains was repeated in triplicate to confirm scores.

Co-aggregation statistics

The percentage of co-aggregation was calculated using the following equation (Handley et al., 1987):

(퐹푛 푠푡푟푎𝑖푛 + 퐶푅퐶 푠푡푟푎𝑖푛) 푂퐷600 − 푂퐷600(푀𝑖푥푒푑 푠푡푟푎𝑖푛푠) 2 푥 100 600 (퐹푛 푠푡푟푎𝑖푛 + 퐶푅퐶 푠푡푟푎𝑖푛) 푂퐷 2

The relationship of the percentage of co-aggregation between strains and the negative control was analysed using a Student’s T-test with a confidence level of 95%. For microscopy, cells observed to have direct cell-to-cell contact were defined as co-aggregating.

2.2 Fn interactions with human colonic epithelial cells

2.2.1 Assessing the ability of CRC-derived Fn isolates to invade human epithelial cells in vitro 2.2.1 A. Caco-2 cell culture

Caco-2 (American Type Culture Collection (ATCC), Manassas, VA [ATCC] line HTB-

37TM) is the human colon adenocarcinoma cell line that was used for all of the in vitro studies in this thesis. Caco-2 cells were cultured in Dulbecco’s Modified Eagle Media (DMEM) (Gibco

BRL Life Technologies, Burlington, ON) supplemented with 10% fetal bovine serum (Gibco

BRL Life Technologies, Burlington, ON), 10mM sodium pyruvate (BioWhittaker; Walkersville,

MD), and 5 µg/mL plasmocin prophylactic (InvivoGen; San Diego, CA). Medium was filter- sterilized using a 0.22 µm filter (Corning; Corning, NY). Caco-2 cells were split every 5 days using a 15 min trypsin (Sigma-Aldrich; Oakville, ON) treatment following a brief washing with sterile, pre-warmed PBS supplemented with 0.1 g/L calcium chloride (BDH Chemicals; Poole,

UK). Cells from passages 4-12 were used for all experiments. All tissue culture growth was carried out at 37°C in 5% CO2 atmosphere.

2.2.1 B. Western Immunoblot analysis of polyclonal antibody sera cross-linking

A collection of rat polyclonal antibodies (EAV_AS2 to EAV_AS4) were used which had previously been raised against purified outer membrane protein preparations from a wide variety of gut-derived Fn strains (Strauss, 2011). The specificity of these polyclonal antibodies against

CRC-derived Fn strains isolated as part of this work was determined using Western immunoblotting. Briefly, a suspension of freshly-cultured whole cells from each of the CRC- derived Fn strains in PBS were mixed in equal parts with SDS lysis buffer (Appendix I), boiled at 100°C for 15 minutes, and separately loaded onto 12% SDS-polyacrylamide gels (one gel was

51 run per primary antibody to be tested). After electrophoresis, the separated proteins were transferred to PVDF membranes (Millipore; Bedford, MA) and each membrane was separately incubated with the appropriate primary antibody at a 1/200 dilution in 5 mL blocking buffer

(Appendix I) overnight at 4ºC. After washing each membrane three times in TBST (Appendix I) for 5 min at RT, an alkaline phosphatase conjugated anti-rat secondary antibody (Sigma-Aldrich;

Oakville, ON) was used for detection; after a 1hr incubation step with the secondary antibody, the results were visualized using BCIP/NBT (5-bromo-4-chloro-3’-indolyphosphate/nitro blue tetrazolium) (Roche Diagnostics; Mississauga, ON).

2.2.1 C. Invasion assays using differential immunofluorescence microscopy

Bacterial strains and growth

Fusobacterial strains were grown in TSBsupp to late log phase and normalized for cell number using McFarland standards (Washington et al., 1972).

Preparation of tissue culture cells on slides

To provide a matrix for adherence of cultured cells for use in immunofluorescence microscopy of invasion studies, glass cover-slips were coated with collagen. Briefly, 300 µL of

Type I calf skin collagen (MP Biomedicals, Solon, OH) diluted to 10% (v/v) in sterile PBS was applied to sterile cover-slips (Thermo Scientific; Burlington, ON) in the wells of 24-well plates

(Falcon, Corning Inc.; Durham, NC) and the plates were incubated under laminar flow and UV light exposure in a laminar flow hood for 1hr. Excess collagen solution was then aspirated from each well before tissue culture cells were added.

Caco-2 cells, at a passage of between 4 and 12, were washed to remove antibiotic, seeded and grown to 80-85% confluence on the collagen-coated glass coverslips in the 24-well tissue culture plates and cultured bacteria were added directly to the wells at a multiplicity of infection

(MOI) of 100:1 (bacterial cells: intestinal cells). Infected Caco-2 cells were incubated at 37°C in

5% CO2 for 4hr, the optimal time for invasion (Strauss, 2011), after which the cells were gently washed with sterile, pre-warmed PBS supplemented with 0.1 g/L CaCl2 to remove any non- adherent bacteria. Cells were then fixed using 2.5% (w/v) paraformaldehyde (Sigma-Aldrich;

Oakvile, ON) and blocked overnight at 4°C in non-permeabilizing (non-perm) block (10% normalized goat serum (NGS) in PBS supplemented with 0.1% NaN3) (Thermo Scientific;

Burlington, ON).

Differential staining of Fn and Caco-2 cells

The differential staining technique used was adapted from the protocol described by

Goosney et al., (1999). Briefly, polyclonal antibodies were diluted 1/200 (EAV_AS2) or 1/500

(EAV_AS3, EAV_AS4) in non-perm block solution, applied to appropriate corresponding coverslips and incubated for 30 min at 37°C. After gentle washing with PBS, the coverslips were then incubated with donkey anti-rat Alexa 350 secondary antibody (Molecular Probes®; Thermo

Scientific; Burlington, ON), diluted 1/100 in non-perm block at RT for 30 min. Following further wash steps in PBS, cells were then permeabilized by the addition of 0.1% Triton X-100 (Fisher

Scientific; Burlington, ON) in PBS for 20 min. Once again, coverslips were incubated with the corresponding primary polyclonal antibody, this time diluted in permeabilizing block (10% NGS in PBS + 0.1%Triton X-100) at 37°C for 30 min, followed by washing and then further incubation at RT for 30 min with a mixture of donkey anti-rat Cy3 and Alexa 488 Phalloidin

(Molecular Probes®; Thermo Scientific; Burlington, ON), both diluted to 1/100 in perm block.

Following a final wash step, coverslips were mounted onto glass slides with Mowial (Sigma-

Aldrich; Oakville, ON).

Immunofluorescence microscopy

Each slide was examined at 40X magnification using a Zeiss Axiovert 200 microscope and a Hamamatsu ORCA-R2 camera. Images were taken and analysis was carried out using

VolocityTM software (version 6.3; PerkinElmer; Waltham, MA). Each invasion was carried out on 3 separate occasions using fresh passages of Caco-2 cells and freshly-grown bacteria inocula.

2.2.1 D. Quantification of bacterial invasion: pixel pipeline analysis

Invasion was assessed by examining pixels of different colours representative of internal and external bacteria. This method compared the number of adherent Fn cells to the number of invaded Fn cells, relative to the area of Caco-2 cells, and was carried out by analyzing each field of view image using Photoshop 7 software (Adobe Systems; San Jose CA). A schematic of the process used is depicted in Figure 2.1 A-G using example screen-shots, and a description of the process is given in detail below. Ten fields of view were analyzed for each of the biological replicates using this technique.

Once an invasion image (jpeg file) was opened in Photoshop, a pixel-to-µm reference was established using the scale bar on the jpeg image. To do this, the zoom function was first used to capture the image’s scale bar (Figure 2.1 A), and a white rectangle was created with the same width and height as this bar using the Rectangle Tool. Using the Marquee Tool, the edge of the created rectangle was then traced. In the Histogram sub-menu (Figure 2.1 B) the statistics of the selected image area were displayed and the area of the created rectangle could be equated to pixel number.

The next step involved removing the colour from any areas in the field of view lacking in

Caco-2 cells and this was done using the Polygonal Lasso Tool. The space in the image void of

Caco-2 cells was traced and filled with black to 100% (Figure 2.1 C). A pixel count for the remaining area (Caco-2 cells) was determined using the Select Inverse tool (Figure 2.1 D). Once again, the Histogram tool was used as above to provide the total number of pixels representative of the Caco-2 cell (Figure 2.1 E).

The Marquee Tool was then used to highlight the entire image, and the Colour Range tool was used (with the radio button set to ‘Selection’ and the ‘Selection Preview’ dropdown

55 dialog set to ‘Black Matte’ (Figures 2.1 F & G). Each label colour (i.e. red for invaded bacterial cells, or purple for adherent bacterial cells) was then assessed for image coverage by selecting for these colours individually, and adjusting the sensitivity using the Fuzziness tool so that all of the coloured area became visible. The Histogram tool was then used once again to determine the number of pixels associated with each colour (Figures 2.1 H-J).

In order to determine the area of both internalized (red) and external, adherent (purple) bacterial cells as well as the total area of the Caco-2 cells, the following equation was used:

푇표푡푎푙 푐푒푙푙 푝𝑖푥푒푙푠 Area of Cells = 푥 (푟푒푓푒푟푒푛푐푒 푎푟푒푎)µ푚2 (푟푒푓푒푟푒푛푐푒 푝푥)

564298 푝푥 E.g. Area of cells= 푥 22µ푚 2 9800 푝푥

Then to determine percent invasion, the area of invaded (interior) bacterial cells was divided by the number of non-invaded (exterior) bacterial cells:

퐴푟푒푎 퐼푛푣푎푑푒푑 퐶푒푙푙푠 Percent Invasion = 푥 100 퐴푟푒푎 푁표푛−퐼푛푣푎푑푒푑 퐶푒푙푙푠

Finally, to normalize the percentage of invasion with respect to the total area of Caco-2 cells in each image:

푃푒푟푐푒푛푡 퐼푛푣푎푠𝑖표푛 Normalized value= 퐴푟푒푎 표푓 퐶푎푐표−2 푐푒푙푙푠

Figure 2.1 Screen captures to show how bacterial invasion was quantified using Photoshop 7 and the pixel quantification pipeline developed for work in this thesis

A) Scale bar was zoomed in on for easy viewing

B) The Rectangle Tool was then used to make a square with the same width and length as the predetermined scale bar. The Histogram tool was then used to give a pixels-to-area reference value for all downstream calculations.

C) To remove any colour from voided space where there were no Caco-2 cells, the Polygonal Lasso Tool was used to trace the outline of the voided area. The area was then filled in black.

D) To determine the number of pixels in the total area of Caco-2 cells, the Select Inverse tab was chosen from the popup menu to highlight the Caco-2 cells.

E) Then the Histogram tool was selected to see the total number of pixels selected.

F) The Marquee Tool was used to highlight the entire image. The Color Range tab was then selected from the drop down menu.

G) The radio button was then set to ‘Selection’ and the ‘Selection Preview’ dropdown dialogue was set to ‘Black Matte’.

H) The initial selection was of all the red interior bacterial cells.

I) To add additional tones, other regions of the interior bacterial cells were selected. In addition, the selection was enhanced by sliding the ‘Fuzziness’ toolbar to the right.

J) The Histogram dialogue box was again selected and the total number of pixels within the highlighted interior bacterial cells was recorded.

2.2.1 E. Quantification and comparisons of bacterial adhesion

In addition to invasion analysis, the described Photoshop pipeline (Section 2.1.3 D) also allowed for an assessment of bacterial adhesion to Caco-2 cells by calculating the total area of adherent/invasive Fn bacterial cells (note that all invasive bacterial cells first must have had to become adherent to host cells prior to invasion) and normalizing the adhesion area relative to the area of Caco-2 cells. Therefore, adhesion was also assayed to determine the differences in adhesion capabilities between the CRC-derived Fn strains, relative to the highly adhesive/invasive Fn 7-1 strain. To determine what percent of the Caco-2 area was associated with adherent Fn cells (Adhesion Area), the total area of bacterial cells (interior and exterior) was divided by the area of Caco-2 cells for 10 fields of view:

푇표푡푎푙 푏푎푐푡푒푟𝑖푎푙 푐푒푙푙 푎푟푒푎 Percent of Adhesion Area: x 100 퐴푟푒푎 표푓 퐶푎푐표2 푐푒푙푙푠

Then, to compare the differences in adhesion capabilities between the highly adhesive Fn

7-1 and the CRC-derived Fn strains, the percent adhesion (determined above) for Fn 7-1 was divided by the percent adhesion of each CRC-derived Fn strain

퐴푑ℎ푒푠𝑖표푛 푎푟푒푎 표푓 퐹푛 7−1 표푛 퐶푎푐표−2 푐푒푙푙푠 Adhesion area ratio: 퐴푑ℎ푒푠𝑖표푛 푎푟푒푎 표푓퐶푅퐶 퐹푛 표푛 퐶푎푐표−2 푐푒푙푙푠

2.2.1 F. Statistics

For statistical analysis, the normalized values (Percent Invasion/Caco-2 area) of each

CRC-derived Fn strain after a 4hr invasion were compared strain-to-strain using a 1-way

61 analysis of variance (ANOVA). Tukey’s test was used for post-hoc pair-wise comparisons. All analyses of invasion data were carried out using Microsoft Excel (Microsoft; Redmond, WA). P- values <0.05 were considered to be significant.

To determine the significance for the adhesion area ratio comparisons between Fn 7-1 and the CRC-derived Fn strains, two-tailed P-values were determined using a Student’s T-test at a 95% confidence level between 10 fields of view.

2.2.2 Host Immune Response to Fn and CRC-derived Strains

2.2.2 A. Assessment of cytokine release from Caco-2 cells in response to infection with Fn and other CRC-derived bacterial strains.

Fn and CRC-derived strains were examined for their abilities, both singly and in various combinations, to induce secretion of three human cytokines, IL-6, IL-8 (CXCL8) and TGF-β, from Caco-2 cells in vitro. Cytokine secretions were measured as a result of infection with: A) invasive vs. non-invasive Fn strains; B) IBD-derived Fn strains vs. CRC-derived Fn strains; C)

IBD-derived and CRC-derived Fn strains vs. other gut-associated bacterial species isolated from

CRC biopsies; and D) Fn strains alone vs. Fn strains as part of a mixed community of gut- associated bacterial species isolated from CRC biopsies. The list of bacterial strains used can be found in Table 9.

In 24-well plates, 100% confluent monolayers of fully differentiated Caco-2 cells were infected with various combinations of CRC-derived bacterial isolates and Fn strains. Bacterial strains Fn 7-1, CC 7/4 Fmu3, CC 7/3 MET1, CC92I, and CC57F were grown anaerobically for

24 hrs (to stationary phase) in TSBsupp; for growth of CC57C this medium was further supplemented with 0.2% (w/v) sodium formate and 0.3% (w/v) sodium fumarate. Following growth, bacterial suspensions were centrifuged at 10,000 rpm for 5 min at 4°C, resuspended in sterile TSBsupp at a McFarland standard between 2.0-3.0 (Washington et al., 1972) and used to infect Caco-2 cells at a MOI of 100:1 (bacterial cells: intestinal cells) for a total of 24hr. Media controls for both sterile TSBsupp and DMEM were separately included in the assays. Deoxycholic acid (DCA) (Sigma Aldrich; Oakville, ON) dissolved in sterile PBS to a stock concentration of 2 mg/mL, was used as a stimulant/positive control for secretion of all three assayed cytokines.

After 24h incubations with the bacterial suspensions, Caco-2 cell supernatants were collected and centrifuged at 10,000 rpm for 5 min at 4°C to remove cellular debris, and the cell- free supernatants were frozen at -20°C for later use in enzyme-linked immunosorbent assays

(ELISAs) to measure the respective secreted cytokine concentrations. Infections were each carried out for a total of three biological replicates, and each of these with three technical replicates.

Table 2.7 Bacterial strains used in cytokine secretion assays.

Bacterial Species Strain Designation Disease state of biopsy donor Fusobacterium nucleatum 7-1 (invasive) IBD

4-1-13 (minimally invasive)

7/4 Fmu3 (invasive)

CRC Campylobacter showae CC 57C

Parvimonas micra CC 7/3 MET1

Eikenella corrodens CC 92I

Gemella haemolysins CC 57F

2.2.2 B. Quantification of Cytokine Secretion

Secreted cytokines were measured in triplicate using commercially available sandwich

ELISA kits (eBioscience, Ready-SET-go; San Diego, CA) according to manufacturer’s instructions. All reagents were provided in these kits except for the wash buffer, which consisted of PBS supplemented with 0.05% (v/v) Tween 20 (Bio-Rad; Hercules, CA) and 1M phosphoric acid (Thermo Scientific; Burlington, ON). Briefly, 96-well, high-binding, polystyrene plates

(Corning Costar; Lowell, MA) were coated with the respective capture antibody dissolved in coating buffer to the prescribed concentration, sealed and incubated overnight at 4°C, then washed and blocked for two hours at room temperature using provided reagents. Following aspiration of the wells and further wash steps, standards and thawed aliquots of cell-free supernatant samples were each vortexed and added to triplicate wells and incubated for either 2 hours (TGF-β ELISA) or overnight at 4°C (IL-6 and IL-8 ELISA). Following the prescribed incubation times, detection antibody, pre-titrated avidin-horseradish peroxidise (HRP) enzyme, tetramethylbenzidine (TMB) substrate, and stop solution (2N H2SO4) were added sequentially with incubation and washing carried out according to manufacturer’s instructions.

Absorbance at 450 nm was measured immediately following the addition of stop solution using a Victor 3V 1420 Multilabel Counter (PerkinElmer; Waltham, MA). The mean absorbance for each set of triplicate standards/supernatant samples was calculated and the mean absorbance of the zero standard was subtracted from each. Absorbance values were then normalized using the line of best fit generated by the standard curve. The equation for this line was used to calculate all cytokine concentrations of the supernatant samples.

2.2.2 C. Statistics

All cytokine results were expressed as the mean of three biological replicates. Strain-to- strain comparisons were performed using ANOVA to assess the significance of the differences seen between datasets. Turkey’s test was also used for post-hoc pair-wise comparisons. P-values were calculated between pairwise comparisons using the Student`s T-test with a 95% confidence level. Analyses were carried out using Microsoft Excel (Microsoft; Redmond, WA). P-values

<0.05 were considered to be significant.

2.2.3 Assessing the effects of human cytokines on Fn invasion in vitro

2.2.3 A. IL-27 antibiotic exclusion assays with Fn 7-1

Caco-2 cells were treated with recombinant IL-27 (R&D Systems; Minneapolis, MN) at the biologically relevant concentration of 100 ng/mL (Guzzo et al., 2010; Chen et al., 2013) for

16hr prior to infection with Fn 7-1 for 4hr. Extent of invasion was measured using an antibiotic exclusion assay. Briefly, Caco-2 cells were seeded in a 24-well tissue culture plate (Falcon,

Corning Inc; Durham, NC) and grown to 80-85% confluence. Fn 7-1 was inoculated into 5 mL

TSBsupp and grown until an OD600 of 0.8-1.0 was obtained. The bacterial cells were then centrifuged at 4000 rpm for 5 min and resuspended in TSBsupp to a McFarland standard of 3.0-

4.0 (Washington et al., 1972). This bacterial suspension was then added to the Caco-2 cells at a

MOI of 100:1 (bacterial cells: Caco-2 cells) for 4hr at 37°C in 5% CO2. Following this incubation, as a bacterial viability control, an aliquot of the cell media was diluted in TSBsupp and plated on FAA, and then incubated at 37°C under anaerobiosis. The infected tissue culture cells were gently washed twice with PBS (+ 0.1 g/L CaCl2) and fresh DMEM containing gentamicin

(300 µg/mL) (Sigma-Aldrich; Oakville, ON) was then added to the wells and incubated for 1hr at 37°C in 5% CO2 to kill any bacterial cells remaining external to the cells. Following this incubation, as a bacterial kill control, an aliquot of the cell media was plated onto FAA agar and incubated under anaerobiosis as before, but this time to ensure that the antibiotic treatment had killed all extracellular bacteria.

Intracellular bacteria were released from infected Caco-2 cells by osmotically lysing them with sterile water. Serial dilutions of the lysed cells were performed in sterile, pre-reduced

TSBsupp, spread on FAA plates and incubated for 2 days at 37ºC in an anaerobic chamber (10%

CO2:10% H2: 80% N2) to allow colonies to develop. Bacterial colonies were then enumerated and the CFU/mL was recorded for bacteria. This experiment was performed for a total of three biological replicates with a total of three technical replicates for each.

2.2.3 B. Differential immunofluorescence of Caco-2 cells infected with Fn 7-1 in the presence of human cytokines

In addition to IL-27 (Section 2.1.4 A above), four other recombinant cytokines were added to intestinal Caco-2 cells to assess their effects on Fn invasion: IL-12, IL-10, IL-6, and IL-

8 (CXCL8) (eBioscience; San Diego, CA). In these assays, Caco-2 cells were treated with each cytokine individually for 16hr prior to infection with Fn 7-1 for 4hr. Infected cells were examined via immunofluorescence microscopy as described in Sections 2.1.3 C and 2.1.3 D.

Three biological replicates, each with three technical replicates, were carried out for each invasion assay and 10 fields of view were captured using the Zeiss Axiovert 200 microscope and a Hamamatsu ORCA-R2 camera, for each replicate. Analysis was carried out using the pixel pipeline and Photoshop, as described in Section 2.1.3 D.

2.2.3 C. Statistics

For statistical analyses, comparisons between the normalized values (Percent

Invasion/Caco-2 area) of Fn 7-1 with and without cytokine pre-treatments was performed using a

1-way analysis of variance (ANOVA). A total of 30 fields of view from three biological replicates of the Fn 7-1 invasion were used to determine all normalized values. Tukey-Kramer’s test was used for post-hoc pair-wise comparisons. All analyses of invasion data were carried out using Microsoft Excel (Microsoft; Redmond, WA). P-values <0.05 were considered to be significant.

For antibiotic exclusion assays, two-tailed P-values were determined using the Student`s

T-test with a 95% confidence level. P-values <0.05 were considered to be significant.

2.3 Determining the role of MORN2 domain-associated proteins in Fusobacterium adhesion/invasion 2.3.1 Characterization of MORN2 domains in adhesion and invasion of Fn strains

2.3.1 A. MORN2 domain containing peptide generation

MORN2 peptides were unable to be cloned directly from the Fn 7-1 genome due to an expression/toxicity issue with sub-cloning and transformation. Instead, MORN2 domain containing peptides, with maltose binding proteins attached for stability and solubility, were cloned by collaborators at Lehigh University from a conserved hypothetical protein from

Pseudomonas aeruginosa (http://www.ncbi.nlm.nih.gov/protein/489205027). Once MORN2 domain containing recombinant proteins were purified, they were sent to the University of

Guelph where a bicinchroninic acid (BCA) assay was performed to assess the concentration of the MORN2 domain containing recombinant protein (Appendix II).

2.3.1 B. Alignment of the generated Pseudomonas MORN2 domain containing peptide with MORN2 domains found in Fn

To ensure that the MORN2 domain containing recombinant proteins, cloned from

Pseudomonas aeruginosa, were a reasonable surrogate for the MORN2 domains of Fn 7-1, an alignment was performed using the ClustalW algorithm in Geneious v. 8.1 (Biomatters Limited;

Auckland, NZ) Alignments of the MORN2 domain gene sequences from Pseudomonas aeruginosa and from Fn 7-1 were carried out.

2.3.1 C. Aggregation of Fn strains with MORN2 domain peptide

An aggregation assay was performed to determine if the MORN2 domain containing recombinant proteins increased and/or decreased the ability of Fn 7-1 to self-aggregate. An adapted version of the co-aggregation assay from Section 2.1.2I was used. Briefly, fusobacterial

71 cells were grown to stationary phase in TSBsupp. Cells were collected by centrifugation at 14, 000 rpm for 10 min at 4°C. Cells were then washed three times with co-aggregation buffer (Appendix

I) and re-suspended in co-aggregation buffer to a McFarland standard of 3.0. In Eppendorf tubes

(sealed with parafilm to maintain anaerobiosis), equal volumes of bacterial culture were mixed with 0.2 mg/mL MORN2 domain recombinant protein, vortexed vigorously for 10 sec and then incubated at 37°C while shaking at 110 rpm. The addition of 0.2 mg/mL bovine serum albumin

(BSA) and 0.2 mg/mL maltose binding protein (MBP) were used as the controls.

Fn 7-1 aggregations were assessed by OD measurements using a Coulter DU®730

UV/Vis spectrophotometer (Beckman Coulter; Mississauga, ON), at either a 1hr or 24h time point. The equation by Handley et al., (1987) was simplified to determine the self-aggregation of the Fn 7-1 strain with and without the addition of the MORN2 and BSA/MBP proteins.

푂퐷600(푓푢푠표푏푎푐푡푒푟𝑖푎푙 푐표푛푡푟표푙) − 푂퐷600(푓푢푠표푏푎푐푡푒푟𝑖푎푙 푐푒푙푙푠 푤𝑖푡ℎ 푝푒푝푡𝑖푑푒푠) 푥100 푂퐷600( 푓푢푠표푏푎푐푡푒푟𝑖푎푙 푐표푛푡푟표푙)

Statistical significance was determined using the Student’s T-Test on three biological replicates. Self-aggregation of the fusobacterial cells was also assessed by phase microscopy after immobilizing the co-aggregated cultures on agarose-coated microscopy slides. Slides were observed by phase microscopy on a Leica DM2000 LED inverted microscope (imaged with a

Jenoptik ProgRes®CT3 camera).

2.3.1 D. Competition Caco-2 invasion assay using Fn 7-1 and the MORN2 domain containing peptide

In order to determine if MORN2 domains play a role in facilitating the adhesion of Fn 7-

1 to the Caco-2 intestinal cells in vitro, the purified MORN2 domain containing recombinant

72 protein was used as a competitive binding inhibitor in the conventional invasion assay described in Section 2.2.1 C. Briefly, MORN2 domain containing recombinant proteins (0.2 mg/mL) were added to 85% confluent Caco-2 cells. The Caco-2 cells and MORN2 recombinant proteins were incubated for 30 min at 37°C in 5%CO2, before the Caco-2 cells were washed with PBS (to remove non-adherent MORN2 proteins). Fn 7-1 was then added at a MOI of 100:1 (as described in Section 2.2.1 C). The infected cells were then incubated for 4hrs at 37°C in 5%CO2, before being fixed, stained and imaged as described in Section 2.2.1 C. Invasion images were assessed using the pixel quantification pipeline using the Photoshop 7 software (Adobe Systems; San Jose

CA) as described in Section 2.2.1 D. Statistical analysis was performed on images using 5 fields of view from three different biological replicates. The purpose of this study was to determine if

MORN2 domains are involved directly with adhesion to Caco-2 cell surface proteins, the addition of the MORN2 domain containing recombinant proteins before Fn 7-1 infection would result in the blockage of the epithelial cell binding sites and would thereby decrease the ability of active invasion by Fn 7-1.

2.4 Determining the presence and role of bacteriophage in invasive Fn strains

2.4.1 Induction and characterization of the invasive Fn subspecies animalis 7-1 bacteriophage ɸFunu1 and ɸFunu2 2.4.1 A. Mitomycin C induction.

Fn strain 7-1 was examined for the presence of prophage. Briefly, to induce prophage, an early log-phase TSBsupp broth culture of Fn 7-1 was incubated at 37ºC until its absorbance at 600 nm was between 0.2-0.3, at which time mitomycin C was added at a final concentration of 2.5

µg/mL. Absorbance at 600 nm was measured for 4 hrs, at which time lysis was observed. The lysed culture was then treated with DNase and RNase (1 µg/mL), centrifuged at low speed and the supernatant was sterilized by filtration using a 0.45µm polyethersulfone membrane

(Sambrook et al., 1989; Miller, 1972). Plaque assays used to assess bacteriophage load were also performed on cell free-lysates (Sambrook et al., 1989).

2.4.1 B. Electron microscopy of bacteriophage particles.

The phage suspension from lysate was dropped onto copper grids with carbon-coated formvar film and incubated for 30 s. Excess solution was drained away on filter paper and washed five times with de-ionized water to remove filtrate debris. The grids were then incubated with 1% uranyl acetate for 10 s and the negatively stained phage particles were then viewed with a Phillips CM10 transmission electron microscope at 120kV using a SIS/Olympus Morada 11 megapixel CCD camera.

2.4.1 C. Bacteriophage purification, DNA extraction and whole-genome sequencing.

Bacteriophages in the cell-free filtrate were precipitated using polyethylene glycol (10% w/v) and sodium chloride (1M). Subsequently, the bacteriophages were purified using a cesium

74 chloride (CsCl) gradient (1.2 g/L- 1.6 g/L) and ultracentrifugation (Sambrook and Russel, 2001;

Humphrey et al., 1997). Phage particles forming a band at the 1.3 g/L density were collected and dialysis was performed using SM buffer (Sambrook et al., 1989). Genomic DNA was extracted by incubating the purified bacteriophage particles in SM buffer with Proteinase K (50 µg/mL) and SDS (0.5% w/v) at 56ºC for 1hr followed by a phenol:chloroform (1:1) extraction and DNA precipitation by ethanol (Sambrook and Russel 2001). Genomic DNA was then sent to the Broad

Institute for full genome sequencing, using the Illumina Eco platform (Illumina, San Diego, CA), and de novo sequencing assembly. Annotations of the obtained scaffolds were done using the

Broad Institute’s prokaryotic annotation pipeline.

2.4.1 D. Alignment of ɸFunu1 and ɸFunu2 genomes with reference genome (Fn 7-1)

Both bacteriophage genomes were uploaded into the Integrative Genomics Viewer software (Broad Institute; Boston MA) (Robinson et al., 2011; Thorvaldsottir et al., 2013). This software aligns prophage genomes with the bacterial host genome (Figure 3). In this view, the presence of fadA, MORN2 domain-containing protein genes and hemagluttinin genes within the genome were also shown.

2.4.2 Induction and Characterization of the invasive Fn subspecies nucleatum 13-3C bacteriophage

While assessing the validity of the PHAST software and the phage clustering algorithm software (Cochrane et al., 2015), Fn 13-3C, a highly invasive strain, generated a negative computational result for the presence of bacteriophage. However, during initial phage induction with mitomycin C, bacteriophage particles from Fn 13-3C were visualized by electron microscopy. In order to prove the existence of Fn 13-3C bacteriophage, and to potentially improve the phage prediction software, Fn 13-3C was further considered for the presence of bacteriophage.

2.4.2 A. Mitomycin C induction, purification and subsequent DNA extraction of Fn 13-3C bacteriophage.

Similar to Section 2.3.1 A., an early log-phase TSBsupp broth culture of Fn 13-3C was incubated at 37ºC until its absorbance at 600 nm was between 0.2-0.3, at which time mitomycin

C was added to a final concentration of 5 µg/mL. Absorbance at 600 nm was measured for 4 hrs, at which time lysis was observed. The lysed culture was then treated with DNase and RNase (1

µg/mL), centrifuged at low speed and the supernatant was sterilized by filtration using a 0.45µm polyethersulfone membrane (Sambrook et al., 1989; Miller 1972). Purification of the bacteriophage particles and DNA extraction were performed as described in Section 2.3.1 C.

2.4.2 B. Electron microscopy of Fn 13-3C bacteriophage particles.

The protocol used to prepare copper grids for electron microscopy was the same as described in Section 2.3.1 B, with the exception of a 0.5% uranyl acetate solution being used to

76 negatively stain the phage particles. A Phillips CM10 transmission electron microscope at 120kV was used along with a SIS/Olympus Morada 11 megapixel CCD camera.

2.4.2 C. Fn 13-3C Bacteriophage DNA Sequencing

Next generation sequencing was accomplished using the MiSeq desktop scanner

(Illumina; San Diego, CA) at the AAC Genomics Facility at the University of Guelph. Phage

DNA was simultaneously fragmented and tagged with sequencing adapters in a single enzymatic reaction using the Nextera XT library preparation kit (Illumina; San Diego, CA). Paired end sequencing was performed using the Illumina MiSeq platform. This de novo sequencing generated 2,692,198 reads with a 19,792 X coverage. Assembly of all the reads was done using

Geneious version 8.1 using the de novo assembly function (http://www.geneious.com; Kearse et al., 2012). Alignment of the resulting bacteriophage sequences to the reference bacterial strain genome was also performed using the Geneious software.

2.4.2 D. Fn 13-3C Bacteriophage gene annotation.

Bacterial gene annotation of the single scaffold was performed using the Geneious v.8.1 software to predict all open reading frames (ORFs). All ORFs were then assigned gene product names based on top BLAST hits against SwissProt/EMBL protein databases (≥70% identity and

≥70% query coverage). This was also done using the PHAST software where BLAST results were catered to identifying phage elements. The tRNAs were identified by tRNAscan-SE (Lowe and Eddy 1997). The rRNA genes were predicted using RNAmmer (Lagesen et al., 2007). An additional annotation analyses was performed using the NCBI conserved domain search database

(Marchler-Bauer et al., 2014).

2.4.2 E. Verification of the isolated bacteriophage presence within bacterial genomes

To verify the bacterial host Fn 13-3C, and to determine the host-range specificity of the isolated bacteriophage (in the absence of plaque assay assessment), primers were designed to amplify four of the bacteriophage genes for PCR amplification (Table 2.8). Briefly, as described in Section 2.1.2 A&B, each primer was used at 20µM with a Taq DNA polymerase and 1-2 µL of the targeted crude bacterial DNA extract as templates (including Fn 13-3C). The reaction conditions for all primer sets were: 94°C for 5 min, (94°C for 30s, 60°C 30s, 70°C 30s) x 30 and

72°C for 5 min. The presence of the four amplicons was determined by agarose gel electrophoresis (Section 2.1.2 C).

Table 2.8 Bacteriophage specific primers

Target Gene Primer Sequence Product name Size

Hypothetical PhProF 3' TGCAAGACGACGGCCTTATT 5' 339bp bacteriophage protein gene sequence PhProR 5' GCACACTTTCATCGCTTGCA 3'

Tail fiber gene TfProF 3' CCACGTTTTGTTTTGCGCAC 5' sequence 332bp

TfProR 5' ATACAAGCCTTTGGGGAGCC 3'

Tail shaft gene TsProF 3' GCTGTGCCGCCTTGATATTG 5' 575bp sequence

TsProR 5' TTGCTGTTTGTTAAGCGGGC 3'

Integrase gene IProF 3' CGCCCTTAATAAAGCCGTGC 5' sequence 371bp

IProR 5' GCCACTCGCTTTTGCTATCG 3'

Chapter 3. Isolation and characterization of bacterial species from CRC biopsies

Metagenomic analysis, whereby the presence of a microbe in a sample is inferred from the presence of its sequence signature, has become a sensitive method for identifying novel tumor-associated microbes in a culture-independent manner (Moore et al., 2011). In 2011, two independent studies used metagenomics, high-throughput sequencing technologies and enhanced bioinformatic algorithms to determine that a marked and highly statistically significant over- representation of Fn gene sequences were found in tumour biopsies (Castellarin et al., 2011;

Kostic et al., 2011). These studies proposed that the contribution of Fn to tumorigenesis may stem from this bacterial species’ well documented ability to cause inflammation (Dharmani et al., 2011), invade host cells (Strauss et al., 2011), and allow non-invasive bacterial strains to co- aggregate and co-invade (Edwards et al., 2006). However, the metagenomic data from these studies are not necessarily indicative of a causal relationship between Fn and colon carcinomas but may instead be representative of an opportunistic infection at an immunocompromised host site. In addition, this initial metatranscriptomic analysis of CRC (n = 11) had limited power to detect rarer microbes represented differentially in tumor and matched normal control tissue, or to place the observed differential representation in the context of the larger diversity of the mucosal microbiome.

In 2013, a follow-up study by Warren et al., (in collaboration with our lab), extended these studies with deeper sequencing of a much larger number (n = 130) of colorectal carcinoma and matched normal control tissues. Once again, a significant tumor over-representation of

Fusobacterium spp. sequences was observed. In addition an over-representation in tumors of sequences from additional, less abundant bacteria, including members of the genera

Campylobacter, Leptotrichia and Selenomonas were also detected. The analysis of these data was carried out using co-occurrence networks in order to identify microbe-microbe and host- microbe associations specific to tumors. In particular Fusobacterium, Campylobacter and

Leptotrichia showed patterns of co-occurrence in CRC (Figure 3.1).

Figure 3.1 The co-occurrence of microbes was inferred by pairwise correlation of sequence counts for all genera. Significance was determined by 1,000 iterations of random re-assignment of sequence read pairs to subjects, with re-calculation of Pearson correlation coefficients. Interactions of significantly differentially abundant genera are illustrated here in a network diagram, constructed using Cytoscape (Cline et al., 2007). Pearson R values ranged from a low of 0.51 (Holdemania-Haemophilus) to a high of 0.97 (Fusobacterium-Selenomonas) (Warren et al., 2013)

Each prefix within a node of the network indicates a bacterial genus: Sp: Sphingobium; Sh: Sphingopyxis; Si: Sphingomonas; La: Lactobacillus; Bi: Bifidobacterium; No: Novosphingobium; Se: Selenomonas; Fu: Fusobacterium; Ro: Roseburia; Eg: Eggerthella; Bl: Blautia; Cl: Clostridium; Ve: Veillonella; Ha: Haemophilus; Co: Coprococcus; Ps: Pseudoflavonifractor; Do: Dorea; Ru: Ruminococcus; Bu: Butyrivibrio; Cs: Clostridiales; Ca: Campylobacter; Le: Leptotrichia; Pa: Parabacteroides; Ba: Bacteroides; Rm: Ruminococcaceae; Er: Erysipelotrichaceae; En: Enterococcus; Br: Burkholderia; Ph: Phascolarctobacterium; Lw: Lawsonia; Al: Alistipes; Su: Subdoligranulum; Od: Odoribacter; Po: Porphyromonas; Bk: Burkholderiales; Go: Gordonibacter; Gr: Granulicatella; Di: Dialister; Fa: Faecalibacterium; Eu: Eubacterium; De: Desulfovibrio; Pr: Parvimonas; Pe: Peptostreptococcus; An: Anaerotruncus; Ci: Collinsella; Or: Oribacterium; Ho: Holdemania; Es: Escherichia; Pv: Prevotella; Me: Methylobacterium; Bd: Bradyrhizobium; Ra: Ralstonia.

In order to further define the role of potentially pathogenic bacterial species in carcinogenesis, there is a need to better understand the biology of CRC-associated Fn, as well as bacterial members of the genera that co-occur with Fn. This cannot easily be done using molecular techniques alone, and instead it was necessary to isolate and biologically characterize strains from CRC sources. In this chapter, the isolation and characterization of Fn strains (and to a lesser extent Campylobacter strains) from CRC biopsies is described. The Fn specific findings are then put into context with previous studies of Fn strains isolated from healthy as well as diseased (IBD) gut sites.

Results

3.1 Isolation of bacterial species from CRC biopsies

Tables 2, 3 and 4 (Section 2.1.1A) list the patients in each of the three cohorts studied.

Patient biopsies were isolated from individuals who had been previously diagnosed with colorectal cancer (CRC). A total of 92 biopsies from 35 different patients were immediately frozen for storage, and shipped on dry ice to the University of Guelph where they were maintained at -80ºC. Culturing of bacterial isolates from biopsies was done under anaerobic conditions at 37ºC using various selective and non-selective media types in an attempt to recover as much diversity as possible. Even so, it was not possible to recover viable bacteria from 22 biopsies (from a total of 16 different patients).

3.2 Identification of the CRC bacterial strains

During this study, the goal was to isolate Fusobacterium spp., and particularly Fn, from the CRC patient biopsies. However, the value of isolation of other bacterial strains from these samples was also recognized and over 1100 other bacterial isolates belonging to a total of 124 different bacterial species were also isolated using a variety of different selective and non- selective media types (Table 9). All bacterial isolates were subjected to DNA isolation, PCR amplification and sequencing of the V3 region of 16S rRNA at the BCCA_TRR. Obtained sequences were BLASTed against the database curated by the Ribosomal Database Project

(http://rdp.cme.msu.edu/seqmatch/seqmatch_intro.jsp) in order to infer closest species identity.

Table 3.1 List of bacterial species isolated from each patient in the CRC study.

Cohort Patient Biopsy Isolated Bacteria Oral- Number Number Number associated (Tumour Bacterium or Matched Phylum Genus/Species Normal) 1 014 001 (Tumour) Firmicutes Clostridium hiranonis Erysipelothrix rhusiopathiae Granulicatella adiacens Yes Holdemania filformis Bacteroidetes Bacteroides sp. Prevotella melaninogenica Yes Actinobacteria Atopobium vagina 092 001 (Tumour) Firmicutes Granulicatella adiacens Yes Clostridium sensu stricto Streptococcus parasanguinis Yes Gemella haemolysans Yes Proteobacteria Eikenella corrodens Yes 094 001 (Tumour) Firmicutes Streptococcus oralis Yes Streptococcus sp. Yes Granulicatella adiacens Yes Escherichia coli 070 001 (Tumour) Firmicutes Butyrivibrio crossotus Listeria grayi Clostridium sensu stricto 087 001 (Tumour) Firmicutes Streptococcus parasanguinis Streptococcus oralis Bacteroidetes Bacteroides uniformis Bacteroides sp. Yes Parabacteroides merdae Yes Actinobacteria Rothia dentocariosa Yes 031 001 (Tumour) Firmicutes Anaerococcus sp. Parvimonas sp. Yes Streptococcus pyogenes Yes Bacteroidetes Bacteroides sp. Bacteroides stercoris Parabacteroides sp. Actinobacteria Eggerthella sp. Yes 053 001 (Tumour) Bacteroidetes Bacteroides sp. Fusobacteria Fusobacterium nucleatum Yes 057 001 (Tumour) Firmicutes Clostridium sensu stricto Gemella haemolysans Yes Parvimonas sp. Yes Solobacterium sp. Yes Staphylococcus epidermidis Yes Bacteroidetes Bacteroides sp. Proteobacteria Campylobacter showae Yes Actinobacteria Eggerthella sp. Yes 088 001 (Tumour) Firmicutes Clostridium bartletti

Granulicatella adiacens Yes Streptococcus sp. Yes Bacteroidetes Bacteroides sp. Proteobacteria Propionibacterium sp. Yes Actinobacteria Eggerthella sp. Yes 098 001 (Tumour) Firmicutes Parvimonas sp. Yes Bacteroidetes Prevotella sp. Yes 2 016 001 (Tumour) Firmicutes Bacillus sp. Bacteroidetes Parabacteroides sp. 019 002 (Normal) Firmicutes Clostridium sensu stricto Proteobacteria Shigella sp. 021 001 (Tumour) Firmicutes Enterococcus sp. Solobacterium sp. Yes Bacteroidetes Bacteroides sp. Capnocytophaga sp. 002 (Normal) Firmicutes Enterococcus sp. Granulicatella adiacens Yes Peptinophilius sp. Solobacterium sp. Yes Proteobacteria Escherichia coli Shigella sp. 033 001(Tumour) Firmicutes Clostridium symbiosum 002 (Normal) Firmicutes Bacillus sp. Clostridium XIVa Clostridium symbiosum Paenibacillus Proteobacteria Propionibacterium sp. Yes 040 001 (Tumour) Firmicutes Clostridium XIVa Clostridium cadaveris Streptococcus sp. Yes Proteobacteria Escherichia coli 002 (Normal) Actinobacteria Corynebacterium sp. 043 001 (Tumour) Firmicutes Cellulosilyticum sp. Clostridium XIVa Clostridium clostridioforme Clostridium symbiosum Eubacterium sp. Roseburia sp. 002 (Normal) Firmicutes Clostridium XVIa Clostridium symbiosum Roseburia sp. 044 001 (Tumour) Firmicutes Clostridium cadaveris Bacteroidetes Alistipes sp. Barnesiella sp. Parabacteroides sp. Porphyromonas sp. 051 001 (Tumour) Firmicutes Parvimonas sp. Yes Solobacterium sp. Yes Streptococcus sp. Yes Proteobacteria Escherichia sp. Bacteroidetes Actinomyces sp. Bacteroides sp. Porphyromonas sp. Yes Verrucomicrobia Akkermansia sp. 054 001 (Tumour) Firmicutes Clostridium symbiosum 002 (Normal) Firmicutes Clostridium symbiosum 055 001 (Tumour) Firmicutes Clostridium symbiosum

Hungatella hathewayi Lachnospiraceae-incertae-sedis

Ruminococcus gnavus 002 (Normal) Firmicutes Clostridium symbiosum Ruminococcus gnavus Lachnospiraceae -incertae-sedis 058 001 (Tumour) Firmicutes Granulicatella sp. Yes Streptococcus sp. Yes Proteobacteria Eikenella sp. Yes Escherichia sp. 059 002 (Normal) Firmicutes Ruminococcus sp.

Bacteroidetes Alistipes sp. 064 001 (Tumour) Firmicutes Cellulosilyticum sp. Clostridium sensu stricto 002 (Normal) Firmicutes Staphylococcus sp. Yes Streptococcus sp. Yes Proteobacteria Propionibacterium sp. Yes Neisseria sp. Yes Actinobacteria Actinomyces sp. 066 001 (Tumour) Firmicutes Propionibacterium sp. Yes 002 (Normal) Firmicutes Staphylococcus sp. Yes Streptococcus sp. Yes 071 002 (Normal) Firmicutes Clostridium sensu stricto Proteobacteria Escherichia/Shigella 077 001 (Tumour) Firmicutes Clostridium sensu stricto Clostridium XIVa Coprococcus sp. Turicibacter sp. Proteobacteria Propionibacterium sp. Yes 002 (Normal) Firmicutes Blautia sp. Coprococcus sp. Bacteroidetes Anaerotruncus sp. 078 001 (Tumour) Firmicutes Clostridium sensu stricto Streptococcus sp. Yes 002 (Normal) Firmicutes Clostridium sensu stricto 3 001 001 (Normal) Firmicutes Acidaminococcus intestini Blautia faecis Dorea sp. Veillonella sp. Yes Proteobacteria Escherichia coli Parasutterella excrementihominis Actinobacteria Bifidobacterium longum Bifidobacterium pseudocatenulatum Bacteroidetes Alistipes onderdonkii Bacteroides ovatus Bacteroides stercoris Bacteroides thetaiotaomicron Porphyromonas asaccharolytica Yes 002 (Tumour) Firmicutes Dorea sp. Streptococcus sanguinis Yes Veillonella sp. Yes Bacteroidetes Alistipes shahii Bacteroides pyogenes Bacteroides ovatus Bacteroides thetaiotaomicron Actinobacteria Bifidobacterium longum Bifidobacterium pseudocatenulatum Proteobacteria Escherichia coli

Parasutterella excrementihominis Sutterella sp. Yes 003 (Tumour) Firmicutes Coprobacillus sp. Flavonifractor sp. Streptococcus sp. Yes Actinobacteria Bifidobacterium sp. Collinsella sp. Olsenella sp. Yes Slackia sp. Yes Bacteroidetes Bacteroides sp. Proteobacteria Escherichia/ Shigella 004 (Normal) Actinobacteria Bifidobacterium sp. Proteobacteria Escherichia/Shigella 005 (Normal) Actinobacteria Bifidobacterium sp. Proteobacteria Escherichia/Shigella Eggerthella sp. Yes 006 (Normal) Firmicutes Acidaminococcus intestini Dielma fastidiosa Yes Streptococcus sanguinis Yes Streptococcus anginosus Yes Veillonella parvula Yes Bacteroidetes Alistipes finegoldii Alistipes indistinctus Bacteroides fragilis Bacteroides ovatus Bacteroides stercoris Bacteroides thetaiotaomicron Bacteroides vulgatus Porphyromonas asaccharolytica Yes Actinobacteria Bifidobacterium adolescentis Bifidobacterium longum Bifidobacterium pseudocatenulatum Collinsella aerofaciens Proteobacteria Parasutterella excrementihominis Sutterella sp. Yes Escherichia sp. 002 001 (Normal) Firmicutes Streptococcus anginosus Yes Bacteroidetes Bacteroides fragilis Bacteroides thetaiotaomicron Barnesiella intestinihominis Actinobacteria Eggerthella hongkongensis Yes Proteobacteria Bilophila wadsworthia Escherichia sp. Fusobacteria Fusobacterium nucleatum Yes 002 (Normal) Actinobacteria Paraeggerthella sp. Yes 003 (Normal) Firmicutes Clostridium innocuum Dorea formicigenerans Streptococcus anginosus Yes Bacteroidetes Bacteroides fragilis Escherichia coli Actinobacteria Eggerthella hongkongensis Yes Fusobacteria Fusobacterium nucleatum Yes 004 (Normal) Firmicutes Eubacterium limosum Yes Actinobacteria Eggerthella hongkongensis Yes Bifidobacterium sp. Proteobacteria Escherichia/Shigella 005 (Tumour) Actinobacteria Paraeggerthella sp. 006 (Tumour) Firmicutes Erysipelatoclostridium ramosum

Bacteroidetes Bacteroides fragilis Bacteroides thetaiotaomicron Proteobacteria Escherichia coli Fusobacteria Fusobacterium nucleatum Yes 003 001 (Tumour) Firmicutes Streptococcus pseudopneumoniae Yes Streptococcus sanguinis Yes Streptococcus anginosus Yes Gemella morbillorum Yes Parvimonas micra Yes Bacteroidetes Alistipes putredinis Alistipes shahii

Bacteroides eggerthii Bacteroides fragilis Bacteroides uniformis Bacteroides vulgatus Bacteroides xylanisolvens Barnesiella intestinihominis Parabacteroides distasonis Prevotella denticola Yes Proteobacteria Escherichia sp. 002 (Tumour) Firmicutes Clostridium sensu stricto Proteobacteria Escherichia coli 003 (Normal) Firmicutes Clostridium perfringens Faecalicoccus pleomorphus Streptococcus pseudopneumoniae Yes Streptococcus sanguinis Yes Streptococcus oralis Yes Bacteroidetes Alistipes finegoldii Alistipes putredinis Bacteroides caccae Bacteroides dorei Bacteroides uniformis Bacteroides vulgatus Parabacteroides distasonis Parabacteroides merdae Lentisphaerae Victivallis vadensis 004 (Normal) Firmicutes Gemella morbillorum Yes Granulicatella paraadiacens Yes Parvimonas micra Yes Peptostreptococcus stomatis Yes Bacteroidetes Prevotella denticola Yes Collinsella aerofaciens Slackia isoflavoniconcertens Yes Proteobacteria Escherichia coli 005 (Normal) Firmicutes Clostridium sensu stricto Proteobacteria Escherichia/Shigella 006 (Normal) Firmicutes Granulicatella sp. Yes Proteobacteria Escherichia/Shigella Klebsiella pneumoniae Yes 004 001 (Normal) Firmicutes Streptococcus sp. Yes Bacteroidetes Bacteroides sp. Proteobacteria Escherichia coli Escherichia/Shigella Klebsiella oxytoca Yes 003 (Normal) Firmicutes Streptococcus anginosus Proteobacteria Escherichia coli Actinobacteria Actinomyces odontolyticus Yes 006 (Tumour) Firmicutes Granulicatella adiacens Yes

Stretococcus anginosus Yes

Proteobacteria Escherichia coli Shewanella algae Actinobacteria Eggerthella lenta Yes 005 001 (Normal) Firmicutes Granulicatella adiacens Yes Actinobacteria Bifidobacterium adolescentis Proteobacteria Escherichia coli Klebsiella pneumoniae Yes 003 (Tumour) Firmicutes Granulicatella adiacens Yes Proteobacteria Morganella morganii Yes Klebsiella pneumoniae Yes 004 (Normal) Firmicutes Granulicatella adiacens Yes Streptococcus parasangunis Yes Proteobacteria Klebsiella pneumoniae Yes Morganella morganii Yes 006 002 (Normal) Firmicutes Catenibacterium mitsuokai Yes Bacillus licheniformis Eubacterium limosum Yes Lactobacillus acidophilus Streptococcus salivarius Yes Actinobacteria Eggerthella lenta Yes Proteobacteria Propionibacterium acnes Yes 004 (Tumour) Firmicutes Granulicatella adiacens Yes Peptoniphilus sp. Yes Actinobacteria Bifidobacterium longum Eggerthella lenta Yes Proteobacteria Escherichia coli 006 (Normal) Firmicutes Gemella morbillorum Yes 007 001 (Normal) Actinobacteria Bifidobacterium longum Eggerthella lenta Yes Proteobacteria Klebsiella oxytoca Yes Escherichia coli 002 (Tumour) Firmicutes Clostridium ramosum Granulicatella adiacens Yes Parvimonas micra Yes Proteobacteria Escherichia coli 003 (Tumour) Firmicutes Clostridium ramosum Gemella haemolysans Yes Oribacterium sinus Yes Parvimonas micra Yes Phascolarctobacterium faecium Veillonella dispar Yes Veillonella parvula Yes Streptococcus mitis Bacteroidetes Bacteroides dorei Bacteroides massiliensis Bacteroides thetaiotaomicron Proteobacteria Escherichia coli Fusobacteria Fusobacterium nucleatum Yes 004 (Normal) Firmicutes Clostridium ramosum Gemella haemolysans Yes Granulicatella adiacens Yes Parvimonas micra Yes Solobacterium moorei Yes Streptococcus intermedius Yes Streptococcus mutans Yes Veillonella atypica Yes Veillonella dispar Yes

Veillonella parvula Yes

Proteobacteria Escherichia coli Parasutterella excrementihominis

Bacteroidetes Parabacteroides distasonis Fusobacteria Fusobacterium nucleatum Yes 005 (Normal) Firmicutes Dialister invisus Yes Gemella haemolysans Yes Granulicatella adiacens Yes Erysipelatoclostridium ramosum Eubacterium biforme Parvimonas micra Yes Solobacterium moorei Yes Veillonella parvula Yes Bacteroidetes Alistipes shahii Bacteroides dorei

Bacteroides caccae Bacteroides thetaiotaomicron Bacteroides xylanisolvens Parabacteroides merdae Actinobacteria Eggerthella lenta Yes Proteobacteria Campylobacter concisus Yes Escherichia coli Shigella sonnei Fusobacteria Fusobacterium nucleatum Yes 006 (Normal) Firmicutes Granulicatella adiacens Yes Parvimonas micra Yes Streptococcus anginosus Yes Proteobacteria Escherichia coli Actinobacteria Eggerthella lenta Yes 008 002 (Tumour) Firmicutes Clostridium ramosum Coprobacillus cateniformis Dialister invisus Yes Granulicatella adiacens Yes Veillonella parvula Yes Actinobacteria Bifidobacterium adolescentis Bifidobacterium longum Eggerthella lenta Yes Bacteroidetes Bacteroides dorei Bacteroides ovatus Bacteroides uniformis Parabacteroides distasonis Proteobacteria Parasutterella excrementihominis Verrucomicrobia Akkermansia muciniphila 003 (Normal) Firmicutes Dialister invisus Yes Staphylococcus warneri Yes Actinobacteria Bifidobacterium adolescentis Bifidobacterium pseudocatenulatum Proteobacteria Propionibacterium acnes Yes Parasutterella excrementihominis Verrucomicrobia Akkermansia muciniphila 006 (Normal) Firmicutes Blautia wexlerae Clostridium innocuum Dialister invisus Yes Dorea sp. Enterococcus avium Erysipelatoclostridium ramosum Streptococcus salivaris Yes Bacteroidetes Alistipes shahii

Bacteroides fragilis Bacteroides dorei Bacteroides uniformis Parabacteroides distasonis Actinobacteria Bifidobacterium adolescentis Bifidobacterium longum Yes Bifidobacterium pseudocatenulatum Eggerthella lenta Yes Proteobacteria Parasutterella excrementihominis Verrucomicrobia Akkermansia muciniphila Red: Fusobacterium nucleatum isolates; Blue: Pathogenic isolates

3.3 Characterization of the isolated Fn strains

3.3 A. Comparative analysis of Fn from CRC biopsies

Fn was isolated using JVN (fusobacteria-selective) media and verified by using Fn specific primers (Table 6). As with all CRC-derived isolates, all potential Fn candidates went through DNA isolation, PCR amplification and sequencing of the V3 region of 16S rRNA at the

BCCR_TRR. To infer species identify the obtained sequences were BLASTed against the

Ribosomal Database Project database.

A total of 18 Fn bacterial isolates were obtained from the provided CRC biopsies. Table

3.2 lists details of all Fn isolates, including the patient (and cohort) they were isolated from. This table also indicates whether isolates were from a biopsy extracted directly from tumour tissue or from matched normal tissue adjacent to the tumour (in the latter case, the distance of the sampling site from the tumour and whether this site was distally or proximally located, was recorded). 18 Fn isolates were obtained from only three patients; one from cohort 1 and the other

2 from cohort 3. Of these 18 isolates, 8 were selected for further characterization based on differences in cell and colony morphology, which were likely indicative of different strains, as well as sequencing of the V3 region of the 16S rRNA gene. For subspecies verification of the Fn isolates, as well as for phylogenetic analysis, sequencing of the full-length 16S rRNA gene for each isolate was performed at the University of Guelph Genomics facility. Table 3.3 lists the details of the 8 characterized CRC-derived Fn strains, including their subspecies designation.

This designation was determined using the obtained 16S rRNA gene sequences and BLASTing them against the Ribosomal Database Project database to infer closest Fn relative (for phylogenetic confirmation see Section 3.3 D).

Table 3.2 Details of Fn strain isolates obtained from CRC patients.

Fn Strain Patient Cohort Tumour vs Matched Normal Tissue Designation CC53 F*

CC53 G 053 1 Tumour (colon)

CC53 H

2/1 JVN3*

2/1 FAA3 Normal (rectum; distal: 5cm)

2/1 FAA4

2/3 Fmu1* 002 3 Normal (rectum; proximal: 6cm)

2/3 FmuA*

2/6 JVN3* Tumour (rectum) 2/6 D51

7/3 JVN 3C1* Tumour (ascending colon) 7/3 D5 3A1

7/4 Fmu3* Normal (ascending colon; distal: 15cm)

7/5 JVN 1A4* 007 7/5 JVN 1A2

7/5 JVN 3A1 Normal (ascending colon; proximal:

2.5cm)

7/5 JVN 1B2

7/5 JVN 1B3

* Fn isolates in bold were selected for further characterization. Identical/clonal strains are listed below their selected representative strain in a lighter greyscale.

Table 3.3 16S rRNA gene-based identification of the CRC-derived Fn strains.

Strain Subspecies Closest Fn Relative *

CC53 Fn subsp. vincentii Fn 3-1-36 A2 Identity: 99%

CC2/1 JVN3 Fn subsp. nucleatum Fn 25586 Identity: 98%

CC2/3 Fmu1 Fn subsp. nucleatum Fn 25586 Identity: 99%

CC2/3 FmuA Fn subsp. nucleatum Fn 25586 Identity: 99%

CC2/6 JVN3 Fn subsp. animalis Fn 7-1 Identity: 99%

CC7/3 JVN 3C1 Fn subsp. animalis Fn 4-8 Identity: 99%

CC7/4 Fmu3 Fn subsp. animalis Fn 4-8 Identity: 99%

CC7/5 JVN 1A4 Fn subsp. animalis Fn 4-8 Identity: 99%

*Identities were inferred using full length (1502bp) gene 16S rRNA gene sequences in BLAST alignments against the Ribosomal Database Project.

3.3 B. Detection of fadA gene sequences in Fn strains

DNA from the 8 selected CRC-derived Fn isolates was also used to detect the presence of the fadA gene, which encodes FadA, an important adhesin involved in Fn virulence (Rubenstein, et al., 2013) and previously shown to be ubiquitous amongst IBD-derived Fn (Strauss, 2011).

Detection of the fadA gene sequence was done using a PCR with primers designed to amplify a

150 base pair region within the fadA gene (Han et al., 2005). Not unexpectedly, the targeted fadA sequence was detected in all of the isolates (Figure 3.2).

Figure 3.2 fadA gene amplicons from CRC-derived Fn strains separated by agarose gel electrophoresis Lane L: GeneRuler 100bp DNA ladder (Thermo Fisher; Burlington, ON); Lane 1: 7/3JVN 3C1; Lane 2: 7/4Fmu3; Lane 3: 7/5JVN 1A4; Lane 4: 2/3Fmu1; Lane 5: 2/3 FmuA; Lane 6: 2/6 JVN3; Lane 7: 2/1 JVN3; Lane 8: CC53; Lane 9: 10 CNA (Clostridium difficile) (negative control); Lane 10: Fn 7-1 (positive control); and Lane 11: no DNA template negative control.

L 1 2 3 4 5 6 7 8 9 10 11

500bp 200bp

3.3 C. Morphological and biochemical analysis of Fn isolates

To assess phenotypic traits, all selected CRC-derived Fn isolates were grown on FAA for

5 days under anaerobic conditions and were subsequently scored for colony and cell morphology according to the key described in section 2.1.2 G (Figure 3.3). In addition, biochemical profiles generated using API Rapid32A tests and sugar fermentation tests were obtained. Isolates were also scored for bile resistance, esculin hydrolysis, hydrogen sulfide production and antibiotic susceptibility to 12 antibiotics.

Table 3.4 summarizes phenotypic data for the 8 CRC-derived Fn strains. All isolates were able to ferment sucrose, fructose and galactose. All isolates were able to weakly ferment lactose, with the exception of 2/1 JVN3, which was negative for this test. Bile resistance was found only among animalis and vincentii subgroups, whereas hydrogen sulfide was produced by all Fn strains. All isolates were positive for indole production. All isolates, except 2/1 JVN3 and

2/6 JVN3, were positive for arginine dihydrolase. All isolates were also positive for glutamic acid decarboxylase, with the exception of 2/1 JVN3, 2/3 Fmu1 and 2/3 FmuA. All isolates were negative for raffinose fermentation, mannose fermentation, urease hydrolysis, esculin hydrolysis,

α-arabinosidase hydrolysis, N-acetyl-β-glucosamindase hydrolysis and nitrate reduction.

Table 3.5 summarizes the antibiotic resistance profiles of the 8 CRC-derived Fn strains.

All isolates were resistant to vancomycin. Based on their zones of clearance, some of the isolates appeared to be mildly susceptible to either one or both of josamycin and/or norfloxacin, two antibiotics that Fusobacterium spp. are generally resistant to. However, both of these strains are resistant to these two antibiotics at a lower concentration, as indicated by their respective growth on JVN plates during the initial isolation of the 8 CRC-derived Fn strains. All Fn subspecies nucleatum isolates (2/1 JVN3, 2/3 Fmu1 and 2/3 FmuA) were resistant to ampicillin, piperacillin

98 and ceftriaxone. The mechanism of action for these three antibiotics involve preventing bacterial cell wall synthesis by irreversibly binding to transpeptidases, thus resistance to all three antibiotics in the nucleatum subspecies is not unexpected as these strains appear to have evolved to prevent binding of these antibiotics to their transpeptidase enzymes. The only other antibiotic resistance observed in the CRC-derived cohort were strains 2/1 JVN3 and 2/3FmuA to imipenem, and 7/3 JVN 3C1, 7/4 Fmu3 and 7/5 JVN 1A4 to ciprofloxacin.

Figure 3.3 Phase contrast micrographs showing the diverse cell morphology of the CRC-derived Fn strains. For cell morphology descriptions for each Fn strain see Table 3.4. Scale bars shown in each image represent 30 µm.

100

Table 3.4 Biochemical profiles, and cell & colony morphologies analysis of Fn strains isolated from CRC-derived biopsies. All strains were negative for raffinose and mannose fermentation, urease and esculin hydrolysis, α-arabinosidase hydrolysis, N-acetyl-β- glucosamindase hydrolysis and nitrate reduction. All strains were positive for sucrose and galactose fermentation and for hydrogen sulfide production.

Phylogenetic

Grouping Strain Colony Cell Mean cell

1 2 Morphology Morphology length

3 ARG TRY (µM)(SD ) - -

Bile FRUC LACT L GLU L

Fn. subsp. CC2/1 2, 5, 7 A, E 5.33 - + - +W +W - nucleatum JVN3 (±4.3) CC2/3 2, 5, 7 A, C 4.65 - + +W + +W + Fmu1 (±2.7) CC2/3 1, 4, 6, 8 A, B, C 3.6 - + + + +W + FmuA (±1.6) Fn. subsp. CC2/6 1, 5, 6, 8 B, D 7.9 + + +W +W + + animalis JVN3 (±5.9) CC7/3 2, 5, 7 A, D 20.2 + +W +W + + + JVN (±11.2) 3C1 CC7/4 2, 3, 5, 8 A, D 16.7 + + +W + + + Fmu3 (±10.7) CC7/5 2, 5, 7 A, B 11.4 + + +W + + + JVN (±7.4) 1A4 Fn. subsp. CC53 1, 5, 8 A, D 20.4 + + +W + +W + vincentii (±13.4) 1 Colony Morphology: 1-mucoid; 2-breadcrumb; 3-umbonate; 4-opaque; 5-pitting; 6-flat; 7-irregular; 8-round. 2 Cell Morphology: A-filamentous; B-needle-like; C-short-medium tapered bacilli; D-medium-long rods; E-irregular. 3 SD: Standard deviation of 10 cell length measurement

101

Table 3.5 Antibiotic susceptibility profiles of the CRC-derived Fn isolates.

lactams

Macrolides

Glycopeptide

Cephalosporin

Aminoglycosides

Fluoroquinolones Phylogenetic

Grouping Strain

cillin

Imipenem

Josamycin Ampicillin

Clindamycin

Amoxi Piperacillin

Gentamicin

Norfloxacin Ceftriaxone

Vancomycin

Ciprofloxacin

Erythromycin Fn. subsp. nucleatum CC2/1 S S S S R S R R S S R R JVN3 (10) (12) (30) (8) (20) (10) (17) CC2/3 S S S R S S R R S S R R Fmu1 (25) (12) (40) (47) (32) (10) (18) CC2/3 S S S R R S R R S S R R FmuA (22) (9) (35) (28) (11) (18) Fn. subsp. animalis CC2/6 S S S S S S S S S S R S JVN3 (12) (24) (42) (11) (32) (20) (33) (30) (9) (17) (25) CC7/3 S S S S S S S S S R R S JVN 3C1 (20) (10) (32) (7) (15) (11) (21) (11) (10) (14) CC7/4 2 S S S S S S S R R R S Fmu3 (23) (14) (37) (11) (22) (12) (21) (10) (14) CC7/5 S S S S S S S S S R R S JVN 1A4 (22) (12) (30) (8) (21) (12) (18) (9) (8) (20) Fn. subsp. vincentii CC53 S S S S S S S S S S R S (25) (19) (35) (10) (18) (30) (19) (10) (11) (20) (16)

S: susceptible (zone of clearance mm- diameter including 6 mm disc); and R: resistant.

102

3.3 D. Phylogenetic analysis of CRC-derived and IBD-derived Fn strains

To define phylogenetic relatedness among the isolates, a 1502 bp region of the 16S rRNA gene from each strain was amplified using full-length universal primers, and sequenced using

Sanger chain termination sequencing (Advanced Analysis Centre, University of Guelph).

ClustalW (as part of Geneious v.8.1) was then used to align the obtained sequences, and phylogenetic trees were constructed using the neighbor-joining method and the Tamura-Nei assumption. To compare previously isolated gut Fn isolates with those directly isolated from

CRC biopsies, the 16S rRNA genes from 11 IBD-derived Fn isolates (7-1, 4-8, 3-1-27, 3-1-

36A2, 4-1-13, 3-1-33, 2-1-50, 21-1A, 3-2-4, 11-3-2 and 13-3C ) were included in this phylogenetic analysis. For reference, the 16S rRNA genes from four Fn type strains (Fn subsp. nucleatum (ATCC 25586), Fn subsp. animalis (NCTC 12276), Fn subsp. polymorphum (ATCC

10953) and Fn subsp. vincentii (ATCC 49256), representing each of the four subspecies, were also included in this analysis (Figure 3.4).

The resulting phylogenetic tree showed a clear resolution of CRC-derived and IBD- derived isolates of Fn with the exception of CC53, which grouped in the same clades as Fn subsp. vincentii isolates. All Fn subsp. vincentii strains clustered together, including the type strains (ATCC 49256), with a slight separation of 4-1-13, 3-1-36A2 and the CRC-derived CC53 strain. All CRC-derived Fn subsp. nucleatum strains grouped together with a high confidence.

CRC-derived strains associated with the nucleatum subsp. were found to be relatively distant from the ATCC 25586 type strain for this subspecies. As expected, no CRC-derived Fn strains clustered with the polymorphum clade. The most interesting separation on the tree included the isolates belonging to the animalis subspecies. A clear and distinct separation between Fn subsp. animalis isolates from CRC and IBD can be seen. In addition NCTC 12276 (subsp. animalis) is

103 located in between these two clades of animalis subspecies isolates. All bootstrap values were generally high, and indicated a clear separation of subspecies.

104

Figure 3.4 Phylogenetic tree based on 16S rRNA gene regions for CRC-derived (blue) and IBD-derived (black) Fn isolates, as well as selected type strains. The percentage of replicate trees in which the associated taxa clustered together using the bootstrap test (1000 replicates) are shown next to the branches. Clusters are marked and represent the different Fn subspecies. The scale bar represents a 0.005% difference in nucleotide sequence. Accession numbers for type strains (red) and IBD-derived Fn strains were downloaded from Genbank and are given in brackets.

CRC-derived Fn strains; IBD-derived Fn strains; Fn subspecies type-strains

105

3.4 Isolation and characterization of Campylobacter spp.

Upon publication of the metagenomics follow-up study by Warren et al., (2013), which showed the over-abundance of other bacterial genera that co-occur with Fn in tumour biopsies, it became an important step to carry out a phenotypic and genotypic analysis of the Campylobacter spp. isolated from CRC biopsies as part of this thesis work.

3.4 A. Comparative analysis of Campylobacter spp. from CRC biopsies

As described in Section 2.1.1 B and 2.1.2 B, Campylobacter spp. were isolated from the

CRC biopsies using anaerobic culturing techniques on FAA, NB or D5 media agar plates.

Speciation of the strains was inferred by comparing the amplified V3 region of the 16S rRNA gene (sequenced at the BCCA) to the Ribosomal Database Project to find the closest relative for each isolate.

A total of 5 Campylobacter isolates were obtained from the CRC biopsies. Phenotypic assessment and the alignment of the V3 region of the 16S rRNA gene sequences were used to infer clonality among the isolated strains. Based on these results, 2 isolates (CC57C and 7/5

D55) were chosen for further characterization (Table 3.6).

106

Table 3.6 Full identification of Campylobacter spp. obtained from the CRC biopsies

Campylobacter strain Species Patient Cohort Tumour vs. Matched designation (Identity1) Normal Tissue

CC57 C* Campylobacter 057 1 Tumour (colon) showae (99%)

7/5 D55*

Campylobacter 007 3 Normal (ascending colon; 7/5 D53 concisus proximal: 2.5cm) (99%)

7/5 NB7

7/5 NB6

1 Identities were inferred using the V3 region of the16S rRNA gene sequences and by BLAST analysis against the Ribosomal Database Project. * Campylobacter isolates in bold were selected for further characterization. Identical/clonal strains (as determined by phenotypic and 16S sequencing) were listed below their selected representative strain.

107

3.4 B Phenotypic assessment of CC57C and 7/5 D55

To assess phenotypic traits, both CC57C and 7/5 D55 were grown anaerobically in a

Ruskinn® Bug box for 2 days at 37ºC on FAA agar plates and scored for colony morphology.

Subsequently, a single colony from each strain was separately inoculated in TSBsupp+ for 24 hours (under the same anaerobic conditions as above) and assessed by electron microscopy as described in Section 2.1.2 H. On FAA agar plates, CC57C colonies appear irregular, small, extremely flat and opaque/translucent; 7/5 D55 colonies appear irregular, pinpoint, flat and grey.

By electron microscopy, it was shown that the cell morphologies for both CC57C and 7/5

D55 appear as short, slender rods with one or more helical bend in each cell (Table 3.7). It was also shown by EM that CC57C and 7/5 D55 cells both possess a single, polar flagellum. This is in contrast to the description of the type strain for C. showae, ATCC 51146T, which has 2 to 4 unipolar flagella per cell, but similar to the phenotype for C. concisus, which is reported to have only 1 polar flagella per cell (Etoh et al., 1993). Also included in Table 3.7 are the resulting phase microscopy images for both CC57C and 7/5 D55 using the capsule staining protocol described in Section 2.1.2 H. Neither strain contains a capsule as indicated by the lack of surrounding halo around representative cells.

Table 3.7 summarizes the antibiotic resistance profiles of both CC57C and 7/5 D55 to 5 clinically relevant antibiotics. Both strains shared the same profile as they were resistant to vancomycin, ampicillin, erythromycin and ciprofloxacin and susceptible only to gentamicin. In addition, as determined using the GasPakTM EZ Campy satchets, both strains are able to survive in the presence of small concentrations of oxygen and are therefore both microaerophillic.

108

Table 3.7 Representative transmission electron microscopy images and capsule staining phase microscopy images of both CC57C and 7/5 D55. The arrows highlighted in the EM images point to the unicellular flagellum located in cells from CC57C and 7/5 D55. As shown by the lack of halo surrounding cells in the phase images, cells from both CC57C and 7/5 D55 lack the presence of capsules. Protocols for both procedures can be found in Section 2.1.2 H.

CC57C 7/5 D55 (C.showae) (C. concisus) Representative electron microscopy image

Representative capsule staining phase image

109

Table 3.8 Antibiotic susceptibility profiles for CC57C and 7/5 D55

lactams

Macrolides

Glycopeptide

Aminoglycosides Strain Fluoroquinolones

Ampicillin

Gentamicin

Vancomycin

Ciprofloxacin

Erythromycin

CC57C S R R R R (C. showae) (18)

7/5 D55 S R R R R (C. concisus) (24)

S: susceptible (zone of clearance mm- diameter including 6 mm disc); and R: resistant.

110

3.4 C Genotypic assessment of CC57C and 7/5 D55

To determine the phylogenetic relatedness between CC57C, 7/5 D55 and other

Campylobacter spp. type strains, a 1502bp region of the 16S rRNA gene from both CC57C and

7/5 D55 was amplified using full-length universal primers (as described in Section 2.1.2 H) and sequenced using Sanger chain termination sequencing (Advanced Analysis Centre, University of

Guelph). The alignment of the resulting sequences was performed using ClustalW (as part of

Geneious v8.1). Phylogenetic trees were then constructed on the Geneious (v8.1) software using the neighbor-joining method, the Tamura-Nei assumption and a 1000 replication bootstrap analysis. For reference the 16S rRNA gene was downloaded Genbank for the following

Campylobacter spp. type strains: C. concisus ATCC 33237, C. rectus JCM6302, C. showae

RM3277, C. coli ATCC 33559, C. jejuni ATCC 33560 and the representative outgroup species

Helicobacter hepaticus ATCC 51449.

The resulting phylogenetic tree showed that CRC-derived Campylobacter strains, CC57C and 7/5 D55, appear to cluser with a high resolution within the clade belonging to their respecitive species, as expected.

111

Figure 3.5 Phylogenetic tree based on full length 16S rRNA gene regions of CC57C and 7/5 D55, as well as selected type strains. The percentage of replicate trees in which the associated taxa clustered together using the bootstrap test (1000 replicates) are shown next to the branches. The scale bar represents a 0.02% difference in nucleotide sequence. Accession numbers for Campylobacter spp. type strains were downloaded from Genbank and are given in brackets.

112

The genomic DNA from both CC57C and 7/5 D55 was extracted (as described in Section

2.2.2 D) using the Maxwell®16 Cell DNA Purification Kit (Promega; Madison, Wisconsin) and was then sent to the BCCA-TTR for PCR purification and sequencing using the Illumina MiSeq platform. For the published manuscript (Warren et al., 2013) that showed the over-representation of Campylobacter spp. in addition to Fn in tumour biopsies (as well as the co-occurrence of these two bacterial species in tumours) it was deemed important to characterize a Campylobacter strain isolated directly from a CRC biopsy. At the time, only CC57C has been isolated. The

CC57C genome was therefore annotated using BLASTx (protein database Genbank_nr; Benson et al., 2008) by collaborators at the BCCA and analyzed in conjunction with our lab in time for publication.

The resulting CC57C draft genome assembly showed highest DNA sequence homology with that of the HMP reference C. showae RM3277. However, whole-genome alignments show that these two genomes (C. showae CC57C and C. showae RM3277) are actually fairly substantially diverged, with only 92.5% average nucleotide sequence identity. Certain CC57C sequence contigs harbor predicted genes with associated virulence potential (Table 3.9).

113

Table 3.9 Key predicted genes from the draft CC57C (C. showae) genome* with virulence associations (adapted from Warren et al., 2013)

Contig Contig Genbank Species Predicted gene Sequence iD Size Accession product(s) identity (%) (bp) Number VirB4 98.32 VirB8 99.54 VirB9 95.98 VirB10 97.24 166 21359 AOTD01000166 VirB11/conjugative 99.05 transfer regulon protein

VirD4 94.42

Relaxase/Mobilization 94.5 Campylobacter nuclease domain rectus protein 208 8558 AOTD1000208 RM3267 Nucleotidyltransferase 97.19 family protein 3 37314 AOTD01000003 OmpA/MotB 94.92 157 23429 AOTD01000157 Type IV pilus assembly 90 protein PilZ 33 26377 AOTD01000033 Toxin secretion ATP- 96.05 binding protein 55 17942 AOTD01000055 Acriflavine resistance 92.78 protein 135 11533 AOTD01000012 Macrolide export ATP- 95.1 binding permease protein MacB 195 7327 AOTD01000195 Campylobacter Invasion antigen B 50.33 jejuni subsp. jejuni IA3902 78 33762 AOTD01000219 Campylobacter Beta-lactamase HcpA 98.22 showae RM3277 81 24337 AOTD01000103 Wolinella Multidrug resistance 57.35 succinogenes protein MEXB DSM1740 *inferred from blastx contig alignments vs. Genbank-NR Blue: Possible pathogenicity genes Green: Antibiotic resistance genes

114

3.5 Co-aggregation of Fn gut isolates with other bacterial isolates from human CRC biopsies

A key virulence factor for Fn cells is their ability to co-aggregate and interact with a number of bacterial species in the formation of biofilms. In order to assess the co-aggregative ability of two CRC-derived Fn strains (CC53 and 7/4 Fmu3) and an IBD-derived Fn strain (Fn

7-1) with enteric bacterial strains isolated from CRC biopsies, co-aggregation assays were performed. The enteric bacterial strains tested for co-aggregation with Fn were isolated in the process of culturing CRC biopsies (Section 2.1.2 A), are known pathogenic species and are orally-associated. The enteric isolates tested belonged to the following species: Campylobacter showae (CC57C), Campylobacter concisus (7/5 D55), Parvimonas micra (7/3 MET1, CC57A),

Gemella haemolysins (7/3 CNA4, CC57F), and Eikenella corrodens (CC92I). Previous work by

Strauss (2011) showed that Fn 7-1 was unable to co-aggregate to Clostridium perfringens. A C. perfringens strain (3/4 CNA1) was also isolated from a CRC biopsy, and was therefore used as the negative control for the co-aggregation assays.

Briefly, as described in Section 2.1.2 I, all strains were cultured in TSBsupp broth, washed and re-suspended in co-aggregation buffer. Strains were then mixed in equal volumes and incubated for 30 min at 37ºC at 110 rpm. Optical densities for co-aggregations between two strains, and the individual strains as controls, were measured using a spectrophotometer at λ600.

Co-aggregation percentages were then determined using the equation described by Handley et al., (1987), and significance was assessed using the Student`s T-test relative to the co- aggregation negative control, 3/4 CNA1 (C. perfringens).

Results for the co-aggregation percentages and p-values can be found in Table 3.10.

Based on the percentages of co-aggregation, it is apparent that all Fn strains seem to be capable

115 of a basal level of co-aggregation to all enteric strains tested, as indicated by the co-aggregation to the negative control C. perfringens. This is not unexpected as the ability of Fn to co-aggregate to a wide variety of bacterial species is well documented (Kolenbrander et al., 1989; Bradshaw et al., 1998; Zilm et al., 2007; Ledder et al., 2008; Fardini et al., 2011; Strauss, 2011; Okuda et al.,

2012). However, many other enteric strains tested were significantly more co-aggregative to Fn as indicated by the p-values calculated relative to the negative control (3/4 CNA1). To assess the compatibility of co-aggregation between Fn 7/4 Fmu3 (isolated from patient 7) and enteric bacterial species, strains belonging to the same bacterial species were assessed from both patient

7, as well as from different CRC patients. Results show that strains of the same bacterial species co-aggregate with Fn 7/4 Fmu3 to the same extent regardless of patient isolation.

116

Table 3.10 Enteric bacterial isolates from CRC biopsies examined for their ability to co-aggregate with Fn strains 7-1, CC53 and 7/4 Fmu3. Co-aggregative abilities between strains were determined by spectrophotometric readings and the equation from Section 2.1.2 I (Handley et al., 1987). Significance between Fn strains and enteric isolates were determined using the Student’s T-test relative to the control aggregation between Fn strains and C. perfringens (3/4 CNA1) (P<0.05).

Percent Co-aggregation Percent Co-aggregation Percent Co-aggregation Strain Strain Identity CRC- (%) with Fn 7-1 (%) with CC53 (%) with 7/4 Fmu3 (Species) patient number % ± SD1 P-value % ± SD1 P-value % ± SD1 P-value

3/4 CNA1 Clostridium Patient 3 6.8 ± 4.7 control 16.3 ± 4.8 control 3.6 ± 4.1 control (negative perfringens control) CC 57C Campylobacter Patient 57 25.5 ± 6.0 0.013* 57.5 ± 12.6 0.006* 51.1 ± 3.1 0.0001* showae 7/5 D55 Campylobacter Patient 7 29.5 ± 6.8 0.009* 70.1 ± 5.0 0.0001* 49.2 ± 2.7 0.0001* concisus 7/3 MET1 Parvimonas Patient 7 25.3 ± 3.4 0.005* 69.5 ± 3.4 0.0001* 10.0 ± 2.2 0.087 micra CC 57A Patient 57 18.3 ± 1.3 0.015* 56.5 ± 1.1 0.0001* 11.0 ± 2.6 0.058

7/3 CNA4 Gemella Patient 7 13.0 ± 2.7 0.117 13.6 ± 7.2 0.609 53.3 ± 6.4 0.0003* haemolysins CC 57F Patient 57 16.6 ± 2.9 0.036* 11.8 ± 0.7 0.180 38.8 ± 1.6 0.0001*

CC 92I Eikenella Patient 92 39.3 ± 3.4 0.0006* 67.8 ± 2.1 0.0001* 58.1 ± 3.0 0.0001* corrodens 1 SD: standard deviation from three biological replicates; * values are significant (P<0.05)

117

Figure 3.6 Representative phase microscopy image of co-aggregation between CC53 (Fn) and CC57C (C. showae). Phase-contrast microscopy images of (A) F. nucleatum CC53 alone, (B) C. showae CC57C alone and (C) a mixture of CC53 and CC57C following incubation in co- aggregation buffer. The long, thin cells of CC53 readily self-aggregate when incubated in aggregation buffer alone (panel A), but also show aggregative ability with the much smaller CC57C coccobacilli (panel C). Images were taken using a Leica DM750 microscope fitted with a 100× oil objective, using the Leica Application Suite LAS EZ Version 1.7.0 software (Castellarin et al., 2012).

118

Discussion

In order to isolate CRC-derived Fn strains we obtained biopsy tissues from the BCCA through a collaboration established with Dr. Robert Holt at the BCCA. The BCCA holds a biorepository of stored biopsy tissues from CRC cases. This was the source of biopsy material from cohorts 1 and 2. Unfortunately at the time of storage (up to 15 years from the date of analysis) the samples were neither handled nor stored for optimal microbial viability. This issue was complicated by the fact that the microbes of interest in this study are strictly anaerobic.

Because of this it is not surprising that only a single isolate of Fn (CC53) was recovered from these specimens, from cohort 1. However, in addition to CC53, isolates identifying with several other bacterial species were obtained (including the Campylobacter showae strain CC57C), and the library of isolates obtained, in general, provided confidence to continue the study.

When proceeding with cohort 3, the following suggestions were given to the BCCA to promote optimal microbial viability of the biopsy samples: 1) after aseptic removal from patients, biopsies should be immediately placed in vials with pre-reduced freezing media containing DMSO and glycerol; 2) vials should be immediately snap-frozen in liquid nitrogen and stored at -80ºC; and 3) once biopsies from two to three patients were collected, they should be shipped to the University of Guelph for immediate processing. Adherence to these procedures by BCCA staff ensured minimal bacterial contamination and O2 exposure for any microbes in the biopsy samples, and reduced the amount of time the samples were stored. Many anaerobic bacteria can be sensitive to long term storage where O2 is present, even if they are frozen and not metabolically active (Madigan et al., 2012).

119

Although richness was increased for cohort three, in the future it would be beneficial to pre-screen the biopsy samples by sequencing the 16S rRNA genes to determine all bacterial species present within each biopsy. Then, isolation techniques to culture viable bacteria could be catered to the representative bacterial species pre-determined by community sequencing. This would allow for the isolation of less abundant species that might require unique environmental conditions and nutrients to grow.

Culturing of the biopsies provided by the BCCA allowed isolation of 124 different bacterial species. Many isolated species are normally mutualistic and are a ubiquitous part of the human gastrointestinal flora. These included Bacteroides spp., Bifidobacterium spp., Escherichia spp., Shigella spp., Blautia spp., Morganella spp., Holdemania spp., Akkermansia spp.,

Lactobacillus spp. and Roseburia spp. Interestingly, many of the species isolated from the gut biopsies in this study are known to be common inhabitants of the oral cavity, including species of the genera Veillonella spp., Solobacterium spp., Dialister spp., Prevotella spp., Oribacterium spp., Staphylococcus spp., and Streptococcus spp., as well as the potentially pathogenic

Eggerthella lenta/hongkongenesis, Granulicatella adiacens, Gemella haemolysans, Eikenella corrodens, Parvimonas micra and Fusobacterium nucleatum. Additional gut-associated bacterial species isolated from the CRC biopsies were also known to be potentially pathogenic and associated with various diseases manifestations in human hosts. Clostridium perfringens is known to have an arsenal of toxins capable of causing histotoxic and intestinal infections in humans (Smedley et al., 2004). Klebsiella pneumoniae and K. oxytoca have both been implicated in urinary tract infections and pneumonia. Klebsiella spp. pathogenesis is generally driven by the production of a polysaccharide-rich cell surface that provides protection from the host inflammatory response (Lawlor et al., 2005). The connection between pathogenic oral and

120 enteric bacterial species and the role they play in gastrointestinal diseases warranted further study, leading to our characterization of some of these specific bacterial isolates. We specifically assessed the role of Fn and Campylobacter spp. strains in terms of phenotypic and genotypic characteristics, and the ability of these bacterial strains to co-aggregate with other enteric isolates from the CRC biopsies.

The published manuscript by Warren et al., (2013) confirmed that there is an over- representation of Fusobacterium nucleatum gene sequences in colorectal carcinomas compared to matched normal controls. The study also showed through the use of metagenomics that there was an over-representation in tumours of sequences from additional, less abundant bacteria, including members of the genera Campylobacter, Leptotrichia and Selenomonas. This drives home an important message when considering the discovery of these pathogens by genomic analysis: abundance and association of a bacterial species with a disease phenotype does not indicate causation. However, the abundance profiles may be useful as biomarkers for disease. In this case, three bacterial genera of limited abundance have been shown to be important in identifying the diseased state of CRC patients. This is evident in that there is not only significance in the co-occurrence of Campylobacter, Leptotrichia and Fusobacterium but also that they, as a co-occurring group, define a metagenomic signature of CRC which may be useful in early detection of CRC. In conjunction with the metagenomics assessment of the co- occurrence of bacterial species within tumour biopsies (Warren et al., 2013), co-recovered cultured bacterial species from the same biopsies (Table 3.1) were assessed to give clues about potential bacterial partners that may act synergistically in CRC. The focus was on bacterial species isolated from biopsies with Fn. Unfortunately, no definitive associations between

121 bacterial species that were isolated from the same patient could be made from the cultured bacterial data.

Although it is not surprising that predominantly anaerobic bacteria were isolated from the tumour biopsies – tumours tend to be anaerobic in nature – the strong associations between what are typically considered oral anaerobic bacteria and colorectal carcinoma is intriguing (Xie et al.,

2014). In previous studies of biopsies from the human gut (or feces) of healthy individuals, the number of orally-associated bacteria have been far fewer than those isolated in this particular work where microbes were cultured only from CRC tumour or matched normal biopsies

(Strauss, 2011; Kurokawa et al., 2007). The suggestion that bacteria from the oral cavity may be contributing to the development of disease elsewhere in the human body is not unprecedented

(Koren et al., 2011; Qin et al., 2014; Bajaj et al., 2015). It is not unexpected that oral microbes may be gaining access to the gut of some CRC patients because of a problem with oral health, the extensive formation of plaque biofilms and/or the lack of a gastric acid barrier (Martinsen et al., 2005; Dongari-Bagtzoglou et al., 2008). Future studies should look at prospectively correlating metrics of oral health with CRC status. To further confirm the suspicion that the CRC patient biopsies harbour more orally-associated bacteria than healthy patients, additional culturing of healthy patient biopsies needs to be completed. To determine the origins of tumour isolates comparative analysis by metagenomics of CRC tumours and healthy controls will also be needed. This could be accomplished by taking oral (swab) and gut (biopsy) samples from the same patient, profiling the microbes and comparing clonality between the oral and gut-associated microbial isolates. These studies would give the additional evidence needed to confirm that an orally diseased state has implications in the gut. If deteriorating oral health is indeed indicative of a risk factor in the development of CRC, oral swabbing can be used as a risk assessment strategy

122 by detection of an oral polymicrobial gene signature that is linked to CRC. Possible polymicrobial signatures have been suggested by Warren et al., (2013) with Fn, Campylobacter and Leptotrichia as potential candidates.

A main focus of this study was to characterize Fn strains isolated from CRC biopsies.

Although Fusobacterium spp. have been associated with the gut niche, and gut-associated diseases such as IBD (i.e. F. gonidiformans, F. periodonticum, F. naviforme, F. necrophorum and F. varium), only Fn was isolated from the CRC biopsies. Although Fn strains were only recovered from relatively few CRC patients, when they were recovered often more than one strain was isolated from the respective patient’s biopsies. This suggests that different Fn strains can co-exist with each other in the gut and may behave synergistically. The synergistic relationship between Fn strains has been documented within the human mouth (Thurnheer et al.,,

1999).

Characterization of the CRC-derived Fn isolates showed that they shared a few standard traits common to all Fusobacterium strains: resistance to vancomycin, positive for indole and

H2S production, and heterogeneous in their morphology. As expected, antibiotic resistant profiles varied across all of the strains. Similar to the findings by Strauss (2011), the sub-speciation of the strains showed that there were no isolates that identified with the polymorphum clade and that the majority of the isolates identified with the animalis clade. The animalis isolates were also resistant to bile. Bile resistance indicates a potential adaptation to life within the GI tract, and may differentiate this subspecies accordingly, as it is not often isolated from oral specimens.

Interestingly, phylogenetic analysis in this chapter showed that the Fn subsp. animalis studied could be divided into two distinct clades (separated by a bootstrap value of 96%), representing both the IBD-associated and CRC-associated isolates. This gives evidence to suggest that

123 although these Fn strains share a common ancestor, they have evolved separately from one another.

In addition to the characterization of Fn strains, it was also deemed of importance to characterize the CRC-derived CC57C (C. showae) and 7/5 D55 (C. concisus) strains. Although these particular species have never been associated directly with CRC, they have both been implicated in other inflammatory infections. Similar to many of the bacterial species isolated from the CRC biopsies, C. showae and C. concisus were initially isolated from the human oral cavity (Etoh et al., 1993; Tanner et al., 1981) but have since been associated with the gut (Tay et al., 2013; Zhang et al., 2009). Both strains shared the same antibiotic resistant profiles, were microaerophillic, contained a single polar flagellum and were absent for capsule presence. Based on phylogenetic assessment of the 16S rRNA gene, both CC57C and 7/5 D55 cluster within their own clade (with a bootstrap value of 100), despite being most closely related to different

Campylobacter species.

Since CC57C (C. showae) was sequenced, assembled and annotated for the published manuscript by Warren et al., (2013), it was discovered that this strain carried genes implicated previously in other Campylobacter spp. pathogenicity. One such contig captures the VirB4/D10 operon homologs, components of a Type IV Secretion System (T4SS) specific to anaerobes, which was first described in the plant pathogen Agrobacterium tumefaciens (Iyer et al., 1982).

T4SS are used by pathogenic anaerobic bacteria to translocate DNA and protein substrates across their membranes and into recipient cells (Alvarez-Martinez & Christie, 2009). Other contigs revealed the presence of relaxase and nucleotidyl transferase genes within CC57C operons. The function of the relaxase enzyme is to nick bacterial DNA (at the origin of replication) and to initiate DNA mobilization and transfer (Nash et al., 2011). The nucleotidyl tranferase enzyme is

124 capable of DNA repair (Martin & Keller, 2007). Together the presence of genes coding for the

T4SS, relaxase and nucleotidyl transferase suggests that CC57C may be capable of bacterial nucleic acid transport into other bacterial species and therefore, be capable of bacterial conjugation and recombination. In order to determine if these genes are expressed in CC57C and to further define the role the resulting proteins play in virulence transcriptomic studies should be performed.

Numerous antibiotic resistance genes were also observed within the CC57C genome, including those coding for the beta-lactamase HcpA, the macrolide permease MacB and the multidrug resistance protein MEXB. Based on further evidence from the antibiotic resistance profiles shown in Table 3.8, antibiotic treatment for this particular strain of C. showae could be potentially problematic in the context of disease. Of note, the BLASTx-based annotation of the

C. showae CC57C genome is only a draft and perhaps more potential functionality would be revealed by finishing the genome.

Since the genome of 7/5 D55 (C. concisus) has not yet been sequenced, the literature was able to give clues concerning its virulence potential. In studies aimed at investigating a possible association between C. concisus to IBD, a significantly higher prevalence of C. concisus DNA

(P<0.0001), as well as specific IgG antibodies (P<0.001) were in patients with newly diagnosed

Crohn’s Disease (Zhang et al., 2009). Attempts at elucidating the pathogenic mechanisms produced by C. concisus indicate that owing to its single polar flagellum, C. concisus has the ability to swim through protective host barriers (i.e. saliva and mucin) allowing the bacterium to achieve close proximity with the host epithelial layer (Man et al., 2010). C. concisus isolates have been shown to produce several toxins (Istivan et al., 2008) and can exhibit comparable epithelial adherence, invasion and translocation abilities to that of the pathogenic and invasive C.

125 jejuni (Kalischuka & Inglis, 2011). It would be of importance to also complete the sequencing and full annotation of the 7/5 D55 (C. concisus) strain.

It was also important to assess the ability of the CRC-derived Fn strains to co-aggregate with enteric isolates, as Fn has been shown to be a pivotal secondary colonizer in plaque biofilm formation (Kolenbrander et al., 2001). As expected, Fn was shown to co-aggregate to all enteric bacterial isolates tested, although in some to a lesser extent than others. This result verified the assumption that Fn strains isolated from the gut are as co-aggregative to enteric isolates as oral isolates. Although using OD measurements has been well documented as a means of quantifying co-aggregation between bacterial species (Cisar et al., 1970; Handley et al., 1987) the ability of

Fn to auto-aggregate to such an extent (Merrit et al., 2009) may cause a misrepresentation of the actual Fn bacterial concentration when using optical density. However, estimations of co- aggregation between the represented Fn and enteric bacterial isolates were of importance for the next chapter where the addition of individual bacterial strains along with a microbial consortia

(containing the same bacterial species tested here for co-aggregation) were added separately to a tissue culture cell assay in order to assess the host response to a polymicrobial infection.

It is clear that microbes do not act alone within the gut and that synergy is at play. One future direction for this work should include isolating additional Fn and Campylobacter isolates from CRC biopsies in order to more clearly understand the phenotypic and genotypic traits most commonly associated with them and whether they are relevant to disease. With sequencing advances, whole-genome sequencing and subsequent alignments of these important bacterial species should be completed as they are likely to be helpful in unraveling the relationships between isolates and disease states. In addition, the link between Campylobacter, Leptotrichia, or Fusobacterium spp. associated with CRC could provide a point of intervention in early CRC

126 detection. These species together may provide a set of parsimonious predictors with potential utility in CRC detection and risk assessment via a non-invasive assay that could detect a polymicrobial signature specific to CRC. If oral health is also at play in CRC disease progression, risk assessment could be as simple as swabbing a patient’s mouth. In the next chapter, experiments were conducted to begin assessing potential synergies between CRC- derived Fn, Campylobacter spp. and other potentially pathogenic orally-associated strains.

127

Chapter 4. Determining the invasive potential of CRC- derived Fn isolates and the effects on host cytokine secretion in response to Fn invasion

Bacteria with the ability to adhere to and invade host cells are usually pathogenic and the invasion process provokes inflammatory reactions and immune responses in and local to the invaded host cells (Pizarro-Cerda et al., 2012; Khalifa et al., 2011; Cossart & Sansonetti, 2004;

Schilling et al., 2001; Naumann et al., 1999; Kwaik et al., 1998). Previous work has shown Fn to be an invasive species; the bacteria use adhesins and invasins to actively invade both endothelial and epithelial host cells (Han et al., 2005; Fardini et al., 2011; Rubinstein et al., 2013). Not unsurprisingly, given the known heterogeneity observed within Fusobacterium spp., there is extensive variability between the invasive potentials of different Fn (Strauss et al., 2011;

McGuire et al., 2014).

The invasive, inflammatory potential of Fn isolated from CRC is as yet, not clear. With the necessary resources (a collection of CRC-derived Fn isolates, polyclonal antibodies and an established, in vitro Caco-2 invasion model), the work presented in this chapter was aimed at determining the invasive potential of the CRC-derived Fn isolates, described in Section 3.2B.

Simultaneously, the effect of human cytokines on the invasive ability of Fn isolates was also assessed during the bacterial infection process in vitro.

The work in this chapter addressed the validity of the hypotheses that: A) CRC-derived

Fn isolates are more invasive than isolates derived from healthy individuals; B) Fn invasion evokes an inflammatory response in the invaded host cells, as revealed by the secretion of host

128 pro- and/or anti-inflammatory cytokines; and C) host cytokine molecules can inhibit the ability of Fn to invade.

129

Results

4.1 Assessing the ability of CRC-derived Fn isolates to invade human intestinal epithelial cells in vitro

4.1 A. Invasion by Fn of Caco-2 cells

Assays were performed using the adenocarcinoma Caco-2 colonic cell line model of infection, as described in section 2.2.1 C. Polyclonal antibody sera were previously separately raised in rats against outer membrane surface antigens prepared from 3 pools of Fn strains

(Figure 4.1) (Strauss, 2011). The 3 polyclonal antibody sera pools were tested for activity against the CRC-derived Fn strains using Western Immunoblotting (Section 2.2.1 B). The distantly related bacterium, Bacteroides fragilis, was used as a negative control, and representative IBD- derived Fn strains were used as positive controls (see Figure 4.1). Cross reactivity was found to be low for two of the three polyclonal sera pools, as expected, and likely due to the high level of strain heterogeneity amongst Fn strains (Strauss et al., 2008). However, one polyclonal sera pool

(EAV_AS2) cross-reacted with all of the CRC-derived strains tested (Figure 4.1). We had also previously reported that CRC-derived Fn strain, CC53, showed cross reactivity to the EAV_AS2 sera (Castellarin et al., 2011).

The availability of a polyclonal antibody with reactivity against our novel CRC-isolates allowed us to assess invasion of host cells in tissue culture using immunofluorescence and an antibody-based differential staining method. Briefly, following invasion assays standardized to infection time and bacterial dose, cells were fixed and stained in such a way as to reveal microbial cells which had entered the host cells and to contrast these with microbial cells which had adhered to host cells but had not invaded. Cells were then imaged using an immunofluorescence microscope. This technique was attempted for all CRC-derived Fn isolates,

130 although only images obtained from experiments with Fn subsp. animalis strains 7/3 JVN3C1,

7/4Fmu3 and 7/5 JVN1A4, and the Fn subsp. vincentii CC53 strain cross-reacted with the sera with enough specificity to allow pixel-based quantification of invasion (described in section

2.2.1 D) (See Figure 4.2 for representative images). The invasion capabilities of the remaining stains could be grossly assessed from the images obtained, despite the issues of Fn cell resolution, and thus representative images are also included in this chapter for reference (Figure

4.3).

Pixel assessment (as described in Section 2.2.1D) was used to quantify invasion of host cells by Fn strains. Comparisons of invasive ability were made using the known, highly invasive,

Crohn’s disease-associated Fn 7-1 strain (Strauss et al., 2011) as a positive control. A graphical representation of the quantitative data obtained is also shown in Figure 4.2 and demonstrates that

Fn subsp. animalis strains (7/3 JVN3C1, 7/4Fmu3 and 7/5 JVN1A4) were similarly invasive to the Fn 7-1 positive control. In contrast, strain CC53 (subsp. vincentii) was significantly less invasive than both the control Fn 7-1 strain, as well as the CRC-derived Fn subsp. animalis strains (P<0.001).

131

Figure 4.1 Polyclonal anti-sera, their respective reactive Fn isolates and the corresponding Western blots depicting the reactivity of the four polyclonal antibody sera against the 8 CRC-derived Fn strains. Lane L: PageRuler Prestained Protein Ladder (ThermoFisher; Burlington, ON); Lane 1: 7/3JVN 3C1; Lane 2: 7/4Fmu3; Lane 3: 7/5JVN 1A4; Lane 4: 2/3Fmu1; Lane 5: 2/3 FmuA; Lane 6: 2/1 JVN3; Lane 7: 2/6 JVN3; Lane 8: Positive Controls (EAV_AS2= Fn 7-1; EAV_AS3= Fn 4-8; and EAV_AS4= Fn 4-1-13); Lane 9: Negative Control (Bacillus fragilis)

Western Blots Polyclonal Anti-Sera L 1 2 3 4 5 6 7 8 9

EAV_AS2

EAV_AS3

EAV_AS4

132

Figure 4.2 A: Representative immunofluorescence micrographs of Caco-2 cells (green) infected with Fn strains for 4 hours, fixed, and differentially stained to show bacterial cells either exterior to (purple) or interior within (orange/red) Caco-2 cells. Fn 7-1 (subsp. animalis) strain serves as the positive control. Scale bars = 22µm. B: Bar chart indicating the normalized percentage of invasion after infection with different Fn strains, as measured using pixel quantification, as described in section 2.2.1 D. ANOVA was used for pairwise comparisons between more invasive Fn strains (Fn 7-1, 7/3JVN3C1, 7/4Fmu3 and 7/5JVN1A4) and the less invasive Fn strain (CC53) *P-values <0.001. A

Fn 7-1 CC53 7/3 JVN3C1 7/4 Fmu3 7/5 JVN1A4

133

Figure 4.3 Representative immunofluorescence micrographs of Caco-2 cells infected with Fn strains. For these strains, pixel-based quantification of invasion was not possible because staining of the bacterial cells could not be resolved to a high enough level. Fn 7-1 strain serves as the positive control. Scale bars = 22µm.

134

4.1 B. Adherence analysis between Fn strains

The pixel quantification pipeline developed for this study was also useful to determine the adherence levels of each Fn strain to Caco-2 cells. To do this, the normalized area of adhesive bacterial cells for each of the Fn strains was determined and compared, using Fn 7-1 as a control. (As well as being highly invasive, 7-1 is also highly adherent to Caco-2 cells.)

Although the CRC-derived Fn strains belonging to subsp. animalis were similar in their invasion capacities to the control Fn 7-1 strain, there were differences in their adhesive abilities:

7/3 JVN3C1, 7/4Fmu3 and 7/5 JVN1A4 were significantly less adhesive to the Caco-2 cells than the Fn 7-1 strain (Table 14), despite normalization of MOI during the assay.

The CC53 strain, despite being significantly less invasive than both Fn 7-1 and the CRC- derived Fn subsp. animalis strains, was equally as adherent as the control Fn 7-1 strain (Table

4.1 and Figure 4.2A).

135

Table 4.1 A: The adhesion area (percent of Caco-2 cell area with associated, adherent Fn cells) for each Fn strain (average of 30 representative fields of view) B: Comparison of the area of adhesion between 4 CRC-derived Fn strains (CC53, 7/3 JVN3C1, 7/4Fmu3 and 7/5 JVN1A4) and the highly adhesive Fn 7-1 strain. Comparative analysis was done using the Student’s t-test at the 95% confidence level. A Fn strain Adhesion Area ± SD Average percent of Caco-2 (standard deviation) area with Fn cells adhered

7-1 (control) 0.1208 ± 0.0432 12.08%

CC53 0.1491 ± 0.0398 14.91%

CC 7/3 JVN3C1 0.0311 ± 0.0236 3.11%

CC 7/4 Fmu3 0.0295 ± 0.0179 2.95%

CC 7/5 JVN 1A4 0.0407 ± 0.0223 4.07%

Strain Comparisons Adhesion Ratio P-value

Fn 7-1: CC53 0.810 0.144689

Fn 7-1: 7/3JVN3C1 3.885 0.000018*

Fn 7-1: 7/4Fmu3 4.090 0.000008*

Fn 7-1: 7/5JVN1A4 2.970 0.000059*

*values are significant (P< 0.0001)

136

4.2 Host Cell Immune Response to Fn and CRC-derived strains

4.2 A. IBD- and CRC-derived Fn induced secretion of IL-6, CXCL-8 and TGF-β in vitro

To determine whether CRC- and/or IBD-derived strains could induce the secretion of cytokines in Caco-2 intestinal epithelial cells, invasion assays (as described in section 2.2.1 C), were paired with sandwich enzyme-linked immunosorbent assays (ELISAs) to quantify cytokine concentration as a measure of the host cell inflammatory response to infection. For ELISA analysis, TSBsupp and deoxycholic acid (DCA) were used as the negative and positive controls, respectively. Caco2 cells were monitored through all assays for viability.

Caco-2 cells were infected for 24h with bacterial strains; this longer incubation period, which was greater than the 4 hr period used for standard invasion assays, allowed for maximal expression of cytokines (Strauss, 2011). Three different cytokines were chosen for analysis of the epithelial response: CXCL-8, a hallmark cytokine in IBD (Strauss, 2011; McCormack et al.,

2011); IL-6, of relevance because of its increased secretion from tumour cells in various types of cancers (Suzuki et al., 2011); and TGF-β, which plays both anti- and pro- tumorigenesis roles depending on the cancer (Drabsch and Dijke, 2012). These cytokines have also been directly associated with Fn adhesion/invasion and the host immune/inflammatory response (Strauss et al., 2011; Gemmell & Seymour, 1993).

To compare cytokine secretion in response to Fn infection, measured cytokine responses for known highly invasive (Fn 7-1) and known minimally invasive (Fn 4-1-13) strains were evaluated. Analyses were carried out by determining the fold increase between the secreted cytokine concentration after infection by each Fn strain relative to the negative and positive controls. There was a significant increase in CXCL-8 secretion by Caco-2 cells infected with Fn

4-1-13 versus Fn 7-1 (P<0.005). This result verified the validity of the assay, as previous work

137 by Strauss (2011) showed the same trend. The secretion of IL-6 was also greater when Caco-2 cells were infected with Fn 4-1-13 compared to Fn 7-1 (P< 0.05). There were no significant differences in the secretion of TGF-β when comparing either of these two Fn strains (Figure 4.4).

Next, differences in Caco-2 cell cytokine secretion were compared for cells that were infected with IBD-derived Fn strains (Fn 7-1 or Fn 4-1-13) and the CRC-derived Fn strain (Fn

7/4 Fmu3). The secretion of IL-6 in response to infection with Fn 7/4 Fmu3 was similar to that of Fn 4-1-13, and therefore greater than that stimulated by infection with Fn 7-1 (P< 0.001).

There were no significant differences in the secretion of CXCL-8 or TGF-β by the Caco-2 cells in response to infection with CRC- or IBD-derived Fn strains (Figure 4.4).

138

Figure 4.4 Cytokine secretion by Caco-2 cells in response to challenge with live Fn strains. Data presented are the mean values of fold increase and are the results of 3 biological replicates, each with three technical replicates. Pairwise comparisons between each of the Fn strains fold increases were performed by a 1-way ANOVA and Tukey’s post-hoc analysis; * P< 0.01. Red: Fn 7-1; Blue: Fn 4-1-13; Purple: Fn 7/4 Fmu3 A: CXCL-8 secretion; B: IL-6 secretion; C: TGF-β secretion

20 * 15

10 A

8 FoldIncrease - 0 Fn 7-1: sTSB Fn 7-1: DCA Fn 4-1-13: sTSB Fn 4-1-13: DCA Fn 7/4: sTSB Fn 7/4:DCA

CXCL Comparisons (Strain:Control)

1.5

0.5

C FoldIncrease β β 0

TGF Fn 7-1: sTSB Fn 7-1: DCA Fn 4-1-13: sTSB Fn 4-1-13: DCA Fn 7/4: sTSB Fn 7/4: DCA Comparisons (Strain: Control)

139

4.2 B. Cytokine secretion by Caco-2 cells in response to challenge by Fn in combination with

CRC-derived isolates of other species.

In order to understand whether the Fn driven cytokine secretion profiles, shown in section 4.2 A above, were representative of responses that might occur in the microbially-diverse environment of the human gut, further experiments were carried out to study cytokine secretion in the presence of defined microbial consortia, with or without Fn. To do this, ELISA was used to assess CXCL-8, IL-6 and TFG-β secretion as above in response to incubation with a subset of microbial strains, representative of isolates from CRC tissues, either alone, together, or in the presence of Fn. Caco2 cells were monitored throughout all assays for viability.

For this work, both Fn 7-1(IBD-derived) and Fn 7/4Fmu3 (CRC-derived) were used as representative Fn strains in Caco-2 cell 24 hour infections. A CRC-derived bacterial consortium

(CRCmix) was used: CC57C (Campylobacter showae), CC57F (Gemella haemolysans), CC92I

(Eikenella corrodens) and CC 7/3MET1 (Parvimonas micra). Comparisons were performed using absolute values of cytokine secretion (concentration) as opposed to fold increases which were used in Section 4.2A.

First, each bacterial strain from the CRCmix was added to the Caco-2 cells and incubated for 24 hours at 37ºC in a 5% CO2 atmosphere. The supernatants were then obtained for assessment in cytokine ELISAs as previously described (Section 4.2 A). Each of the Fn and

CRCmix strains were individually cultured in TSBsupp before being separately applied to Caco-2 cells at a MOI of 100:1 (bacterial cells; intestinal epithelial cell). Sterile TSBsupp was added to control wells. ELISA results are presented in Figure 4.5. Both Fn strains induced the secretion of

CXCL-8 (P<0.001), as did CC92I and CC57F (P<0.05). Only Fn 7/4Fmu3 CC92I induced a

140 significant increase in the secretion of IL-6 (P<0.05). Finally, only CC57F induced a significant secretion of TGF-β (P<0.05) (Figure 4.5).

Next, four bacterial strain consortia (Table 4.2) were used to infect Caco-2 cells as above.

The bacterial species selected to represent the different consortia were chosen due to their known statuses as opportunistic oral pathogens, as well as their abilities to co-aggregate to Fn, as addressed in Section 3.5. Each bacterial strain was individually cultured in 5 mL TSBsupp and applied separately to the Caco-2 cells at a MOI of 100:1 as above. Infected Caco-2 cells were incubated at 37ºC at 5% CO2, as above, with control wells incorporated into the assay to which only growth medium (TSBsupp) was added. Results are presented in Figure 4.6. Incubation with each of the microbial consortia resulted in a significant increase in the secretion of all three tested cytokines (P<0.005-P<0.05).

Finally, invasion assays results for all CRC-derived strains and bacterial strain consortia

(Table 4.2) were compared to observe any differences in cytokine secretion between infection of

Caco-2 cells with individual bacterial strains and/or bacterial consortia. ELISA results are presented in Figure 4.7. CXCL-8 secretion by Caco-2 cells was significantly increased when

Caco-2 cells were challenged individually with Fn strains (P<0.001). This trend was not observed when comparing IL-6 or TGF-β secretion by Caco-2 cells in response to infection with individual bacterial strains versus the bacterial consortia (data not shown).

141

Table 4.2 List of the bacterial strains represented in each consortium

Bacterial Consortia Bacterial strain combinations Designations (Strain Name; Species Name) CC57C; Campylobacter showae Consortium A CC7/3 MET1; Parvimonas micra (CRCmix) CC92I; Eikenella corrodens

CC57F; Gemella haemolysans Fn 7-1; Fusobacterium nucleatum Consortium B (IBD-derived) CC57C; Campylobacter showae CC7/3 MET1; Parvimonas micra CC92I; Eikenella corrodens CC57F; Gemella haemolysans Fn 7/4 Fmu3; Fusobacterium nucleatum Consortium C (CRC-derived) CC57C; Campylobacter showae CC7/3 MET1; Parvimonas micra CC92I; Eikenella corrodens CC57F; Gemella haemolysans Fn 7-1; Fusobacterium nucleatum Consortium D (IBD-derived) Fn 7/4 Fmu3; Fusobacterium nucleatum (CRC-derived) CC57C; Campylobacter showae CC7/3 MET1; Parvimonas micra CC92I; Eikenella corrodens CC57F; Gemella haemolysans

142

Figure 4.5 Cytokine secretion by Caco-2 cells in response to challenge with live CRC- or IBD-derived strains. The results shown depict the mean cytokine concentration for 3 biological replicates, each with 3 technical replicates. ANOVA and Tukey’s post hoc analysis were used for pairwise comparisons between each individual strain and the control (no bacteria added), **P<0.001, *P<0.05. A: CXCL-8 secretion; B: IL-6 secretion; C: TGF-β secretion

143

Figure 4.6 Cytokine secretion by Caco-2 cells in response to challenge from four bacterial consortia. Data are expressed as the mean cytokine concentration released from Caco-2 cells from 3 biological replicates, each with 3 technical replicates. Pairwise comparisons between each of the bacterial consortia and the negative control (no bacteria added) were performed using a 1 way ANOVA with Tukey’s post-test, **P<0.005, *P<0.05. See Table 4.2 for strains used in each consortium. The control consisted of TSBsupp (no bacteria added). A: CXCL-8 secretion; B: IL-6 secretion; C: TGF-β secretion

A B

C: C

144

Figure 4.7 CXCL-8 cytokine secretion by Caco-2 cells challenged by with individual bacterial strains and/or bacterial consortia. Pairwise comparisons between (A) Fn strains, (B) CRC-derived strains (CC57F, CC57C, CC92I and CC7/3MET1) and (C) each bacterial consortia were performed by 1-way ANOVA with Tukey’s post-test; *P<0.001. See Table 4.2 for strains used in each consortium. The control consisted of TSBsupp (no bacteria added).

145

4.3 Assessing the effects of human cytokines on Fn invasion in vitro

4.3 A. Determining the effects of IL-27, IL-6, IL-10, IL-12 and CXCL-8 additions to Caco-2 cells on Fn invasion

Host immune factors, including cytokines, play a key role in regulating the ability of bacterial cells to invade host cells, as well as in mediating the inflammation associated with infection and disease. An antibiotic exclusion assay, that was used as a pre-screen, showed that

IL-27 presence caused a reduction in the ability of Fn 7-1 to invade Caco-2 cells in vitro

(Appendix II). Based on these results, it was of interest to determine whether this effect extended to other cytokines of importance in bacterial infection. Incubation of Caco-2 cells with a panel of either pro- and/or anti-inflammatory cytokines (Table 4.3) was performed prior to bacterial infection to determine if these cytokines could impede or enhance Fn invasion in epithelial cells in vitro. Cytokines for this study were selected based on their well described effects on host cell physiology (Table 4.3). For these experiments, Fn 7-1 was selected for study. The differential staining and pixel quantification methods (as described in Section 2.1.2 C and D) were used for assessment of invasion because, although more time consuming than antibiotic exclusion assays, this staining method is a more direct, and thus a more accurate measure of invasion.

The addition of cytokine IL-27 to Caco-2 cells significantly inhibited Fn 7-1 invasion compared to invasion of untreated, control cells (P<0.05) (Figure 4.8). The actual P-value obtained (P=6.8E-16) was, in fact, much more significant than that which had been previously determined by antibiotic exclusion assays (P=0.05) (Appendix II).

The addition of pro-inflammatory cytokines, CXCL-8 and IL-12, also impeded the ability of Fn 7-1 to invade Caco-2 cells in vitro compared to untreated, control cells (P=1.5E-6 and

9.3E-9 respectively). In contrast, the addition of IL- 6, a pleiotropic cytokine, or IL-10, an anti-

146 inflammatory cytokine to host cells did not significantly inhibit nor enhance invasion by Fn 7-1

(Figures 4.8).

147

Table 4.3 Cytokines assayed for their effects on Fn invasion

Cytokine Function Biological Reference Concentration IL-27 A both pro- and anti-inflammatory cytokine 100 ng/mL (Guzzo et al., shown to play a role in mediating 2010) inflammation during many chronic (Chen et al., inflammatory diseases including IDB. Also 2013) plays an important role in regulating the (Wynick et al., activity of B- and T-lymphocytes 2014) IL-12 A pro-inflammatory cytokine that facilitates 10 ng/mL (Vignali and Th1 differentiation and induces the production Kuchroo 2012) of IFNγ and TNFα from natural killer cells. IL-10 An anti-inflammatory cytokine that can 10 ng/mL (Latorre et al., repress pro-inflammatory responses and limit 2014) unnecessary tissue disruptions caused by (Ouyang et al., inflammation. 2011) IL-6 A pleiotropic cytokine with both pro- and anti- 10 ng/mL (Schafer and inflammatory effects. It is important for the Brugge 2007) host response to infection by exerting antigen- (Suzuki et al., specific immune responses. The chronic 2011) effects of IL-6 have also been implicated in cancer. CXCL-8 A pro-inflammatory cytokine that stimulates 10 ng/mL (Brew et al., the migration of neutrophils, monocytes, 1999) lymphocytes and fibroblasts. (Li et al., 2001) Can also regulate tumour growth and metastasis in colonic epithelial cell lines.

148

Figure 4.8 A: Micrographs showing representative images of Caco-2 cells infected by Fn 7-1 with or without cytokine pre-treatments. Each image shows actin (green), interior bacterial cells (red/orange) and exterior bacterial cells (purple). Scale bars = 22µm B: Invasion of Caco-2 cells with Fn 7-1 in the presence (blue) or absence (red) of cytokines pre-treatments. ANOVA and Tukey`s post- hoc analysis was used to infer significance between the presence or absence of each cytokine ***P<0.00001, **P<0.0001, *P<0.001. A

No Cytokine IL-27 (100ng/mL) IL-6 (10ng/mL) IL-10 (10ng/mL) IL-12 (10ng/mL) CXCL8 (10ng/mL) B

149

Discussion

The work performed in this study was designed to determine whether Fn strains derived from CRC tumour tissue (and phenotypically characterized in chapter three), displayed traits associated with virulence in a human intestinal epithelial cell line model. Specifically, the ability of the bacterial strains to adhere to, and to invade host tissues, and to elicit a pro-inflammatory cytokine secretion response, was measured. Also measured was the ability of Fn to invade epithelial cells in response to cytokine presence.

Invasiveness is the ability of a pathogen to enter host cell tissues and encompasses (1) the mechanisms of colonization (adhesion), (2) the production of extracellular factors that promote the active invasion of tissues (i.e. invasins) and (3) the ability to overcome host defense mechanisms. Adherence of bacterial cells to host cells is a prerequisite for invasion and the ability of a bacterial species to adhere and invade host cells contributes to the overall pathogenicity of the bacterial species (Muller et al., 1999). Fn is a highly co-aggregative species, acting as a secondary colonizer in biofilms and containing an arsenal of adhesins that allow this species to adhere to other bacteria as well as both human endothelial and epithelial cells (Han et al., 2000; Kolenbrander, 1989). Invasion is a key virulence trait in Fn (Han et al., 2000;

Bachrach et al., 2004).

Previously in our lab, invasion by Fn strains isolated from the human gut was measured using a Caco-2 cell model and antibody-based differential immunofluorescence (Strauss, 2011;

Goosney et al., 1999). Caco-2 cells are a transformed cell line and although primary cells may have been a better representative for invasion, they are difficult to source and the Caco-2 cells were used in their place. The findings of Strauss (2011) showed that many gut-associated Fn

150 strains belonged to the animalis subsp. and that some isolates were more invasive than others, but invasion efficiency correlated with the disease status (IBD or healthy) of the patient from which the strain originated. Invasion potential of these gut-associated Fn strains was measured by counting the number of Caco-2 cells containing at least one internalized bacterium (Strauss,

2011). Although this method has been published, it represents a conservative approach with limited sensitivity as the number of internalized Fn cells in each of the Caco-2 cells were not accounted for. The study reported in this chapter used a novel method to measure differential staining of bacterial cells by using computer software to discern numbers of coloured pixels.

Differential staining of the bacterial cells resulted in the ability to distinguish between internal and external Fn cells based on colour; external Fn cells were purple and internal Fn cells were red. Therefore, the number of pixels representing each colour was measured and the representative area was determined separately for both invaded and non-invaded Fn cells. In addition to the quantification of bacterial invasion, pixel quantification allowed for the assessment of adhesion of the Fn cells to Caco-2 cells. The area of all Fn cells on and/or in the

Caco-2 cells was measured in pixels and converted to area.

Using this novel, more sensitive quantification method, it was shown that the CRC- derived Fn subsp. animalis strains (7/3 JVN3C1, 7/5 JVN1A4 and 7/4Fmu3) were all similar in their ability to invade Caco-2 cells. The three animalis strains were equally as invasive as the

IBD-derived Fn 7-1 isolate (used as a control for invasion). CC53 (subsp. vincentii) however, was significantly less invasive in the Caco-2 cell model than Fn 7-1, 7/3 JVN3C1, 7/4 Fmu3 and

7/5 JVN1A4 (P<0.001). There was no significant difference between tissue adhesion in CC53 and the control Fn 7-1. Interestingly however, all of the invasive CRC-derived Fn subsp.

151 animalis strains (7/3 JVN3C1, 7/5 JVN1A4 and 7/4Fmu3) were significantly less adhesive than the control Fn 7-1 (P<0.0001).

Similar to findings by Strauss (2011), Fn subsp. animalis strains were isolated from gut biopsies, however the correlation of invasion and the diseased state of the CRC patient could not be addressed as all biopsies isolated for this work were from patients diagnosed with CRC. In this regard, results showed that the invasive potential of CRC-derived Fn isolates varied between strains isolated from tumour tissue and healthy tissue, even within the same patient. For example,

7/3 JVN3C1, 7/4 Fmu3, and 7/5 JVN1A4 which were all isolated from patient 7, were shown to be highly invasive, however only 7/3 JVN3C1 was directly isolated from tumour tissue.

Alternatively, CC53, isolated from patient 53, was isolated from tumour tissue but was minimally invasive. It is possible that, since culture work does not indicate the quantity of a particular strain in a given sample, the gut colonization of CRC-derived Fn isolates had already occurred at the time of diagnosis/biopsy extraction. The ubiquitous presence of Fn throughout the gut of a sick patient may explain why Fn was extracted from both tumour and healthy tissue.

To get a better understanding of the invasive potential of CRC-derived Fn strains, additional isolates from both CRC and healthy patients are needed.

Invasion analysis was only able to be performed on a few of the CRC-derived Fn strains because of the low resolution of bacterial cells seen in the immunofluorescence micrographs of

2/3 FmuA, 2/3 Fmu1, 2/1 JVN3 and 2/6 JVN6. This was probably due to the antibody cross reactivity not being absolute in these strains, resulting in low affinity of binding between bacterial cells and the antibodies. Despite the low resolution, a few trends could still be assessed:

A) 2/3 FmuA appeared to be highly invasive and adhesive; and B) 2/6 JVN3, 2/3 Fmu1 and 2/1

152

JVN3 appeared to invade some Caco-2 cells, as well as display some adherence to the tissue cells.

The cross reactivity of the polyclonal antibodies to the CRC-derived Fn isolates differed greatly from previous work with Fn isolates from IBD patients; for the IBD-derived Fn isolates, there was very little, if any, cross-reactivity observed between Fn strains (Strauss, 2011). It was therefore unexpected to find that all of the CRC-derived Fn isolates cross-reacted with the polyclonal sera, EAV_AS2. The EAV_AS2 antibody sera was raised to a pool of outer membrane proteins from the gut-associated Fn strains: Fn subsp. animalis 7-1 (isolated from the inflamed biopsy of a Crohn’s disease patient; highly invasive), Fn subsp. animalis 4-8 (isolated from a healthy control; minimally invasive) and Fn subsp. nucleatum 11-3-2 (isolated from a healthy patient; minimally invasive). Perhaps a specific outer membrane protein is expressed as part of a gut-dwelling lifestyle in CRC- and IBD-derived Fn isolates. To test this hypothesis, outer membrane preparations of all CRC-derived Fn isolates could be extracted and assessed by analytical chemistry techniques and mass spectrometry to identify all proteins in the preparations. Then outer membrane proteins could be compared between Fn isolates that did, and did not, cross-react with EAV_AS2. Having a better understanding of the common outer membrane proteins used by Fn strains in the gut could be useful for the development of preventative therapies, such as vaccines, aimed these pro-inflammatory, invasive, opportunistic pathogens.

Adherence and invasion of bacteria to host cells often stimulates an inflammatory response that can be assessed through stimulation of cytokine secretion and measurement with commercially available cytokine quantification ELISA kits. In this work, three different cytokines were used to assess the inflammatory response by the host in response to infection

153 with our CRC-derived isolates: CXCL-8, IL-6 and TGF-β. CXCL-8 was chosen as it plays a role in inflammatory diseases, such as IBD and CRC, by recruiting neutrophils to the site of infection, inflammation and trauma (Baggioloni & Clark-Lewis, 1992; McCormack et al., 2001; Knupfer

& Preiss, 2010). IL-6 was chosen because it has been found in elevated levels in CRC patient serum and has been shown to modulate tumour growth by activating signalling pathways in tumour cells (Knupfer & Preiss, 2010: Herbeuval et al., 2004). TGF-β was chosen as it is a negative regulator of inflammation through the promotion of regulatory T-cell populations and can inhibit the growth of tumour cells through the activation of apoptotic cell death or autophagy by various signalling pathways. TGF-β can also play a role in promoting cancer cell progression by regulating epithelial mesenchymal transition, which is the process by which epithelial cells lose their cell polarity and adhesion, and gain migratory and invasive properties (Ikushima &

Miyazono, 2010). In addition to being associated with inflammatory diseases, CXCL-8, IL-6 and

TGF-β were chosen for ELISA analysis as secretion of these cytokines have all been associated with the presence of Fn in host cells (Strauss et al., 2011; Gemmell & Seymour, 1993; Gallimidi et al., 2015).

The ELISA results showed that Fn 7-1 induced significantly less secretion of CXCL-8 from the Caco-2 cells compared to Fn 4-1-13 (P<0.01). This result was expected, and has been shown previously in our laboratory by Strauss (2011). There was no difference in CXCL-8 secretion when Caco-2 cells were infected with either the IBD-derived Fn 7-1 or the CRC- derived 7/4 Fmu3. The concentration of IL-6 secreted from Caco-2 cells when infected with either Fn 4-1-13 or 7/4 Fmu3 was significantly higher than infection with Fn 7-1(P<0.01). There were no significant differences between the three tested Fn strains in their ability to induce the secretion of TGF-β from Caco-2 cells.

154

A variety of situations may have resulted in the decrease of CXCL-8 and IL-6 secretion when Caco-2 cells were infected with Fn 7-1: A) The binding of Fn 7-1 adhesins to Caco-2 surface receptors, and the subsequent invasion into the host cells, may have resulted in cell signalling cascades that could have in turn prevented the expression of the mRNA encoding the

CXCL-8 and IL-6 proteins; B) Fn 7-1 may have directly prevented the secretion of the CXCL-8 and IL-6 proteins; and/or C) Fn 7-1 could have secreted proteases capable of degrading the secreted CXCL-8 and IL-6 proteins. In relation to CRC disease etiology, CXCL-8 and IL-6 concentrations have been shown to be elevated in serum levels of patients diagnosed with CRC.

Therefore, the resulting increase in CXCL-8 and IL-6 concentration by Caco-2 cells in response to the CRC-derived 7/4Fmu3 is expected. In all cases, a basal concentration of TGF-β was secreted by Caco-2 cells; however, the lack of difference in TGF-β concentration induced by infection with Fn is not unsurprising as TGF-β genes have been shown to be constitutively expressed in intestinal epithelial tissues (Li et al., 2012). As a future direction, additional CRC- and IBD-derived Fn strains should be tested for cytokine secretion to determine if these findings are common across Fn strains associated with either disease.

In addition to Fn isolates from CRC tumour tissues, a considerable library of non-Fn isolates (see section 3.2A), also derived from the same biopsy samples, were available. As previously mentioned in chapter three, the strong correlation between oral-associated bacterial strains and colorectal carcinoma biopsy samples was intriguing. Since the CRC-derived isolate library was at our disposal, a preliminary investigation into the potential role that these correlative oral bacterial species play in gastrointestinal cell inflammation could be addressed in our in vitro model. Since Fn has been shown to play an important role in oral polymicrobial infections (Edwards et al., 2006; Zhang et al., 2008; Saito et al., 2008; Polak et al., 2009; Wu et

155 al., 2015), and the likelihood that Fn forms similar associations with oral isolates in the gut, we sought to answer the question: are the Fn-driven cytokine secretion profiles representative of responses that might occur in the polymicrobial environment of the human gut? Ideally, we would have performed co-invasion assays in the Caco-2 tissue cell model with and without all isolated CRC-derived Fn and oral strains, but specific antibodies were not available for the oral strains of interest making this unfeasible given the experimental budget. Instead, we decided to test the host inflammatory response by measuring cytokine secretion by Caco-2 cells in the presence of small, defined microbial consortia with or without Fn.

The microbial consortia consisted of an invasive Fn strain from both IBD (Fn 7-1) and

CRC (7/4 Fmu3) and an additional 4 oral CRC-derived isolates. CC57C (Campylobacter showae) was specifically selected because this species is in an over-abundance in CRC tumour biopsies (Warren et al., 2013). CC57F (Gemella haemolysans), CC92I (Eikenella corrodens) and

7/3 MET1 (Parviomonas micra) were chosen because of their association with various inflammatory infections within the human body including, but not limited to, endocarditis, bacteraemia, necrotizing fasciitis and brain abscesses (Miller et al., 2007; Moazzam et al., 2015;

Kaufhold et al., 1989; Persson & Imfeld, 2008). All oral CRC-derived isolates were also chosen because of their ability to readily co-aggregate with Fn (see Section 3.3A).

ELISA results showed that individually, Fn 7-1, 7/4 Fmu3, CC92I and CC57F infection induced a significant increase in CXCL-8 secretion by Caco-2 cells, compared to the bacteria- free control (P<0.05-0.001). Caco-2 cells secreted a significantly increased concentration of IL-6 when infected by 7/4 Fmu3 and CC92I. Finally, TGF-β secretion was increased significantly by infection with CC57F. However, when the CRC-derived gut isolates were added to the Caco-2 cells together, in various combinations, CXCL-8, IL-6 and TGF-β secretion were significantly

156 increased compared to the bacteria-free control (P<0.05-0.005). When added as a consortium, the

CRC-derived isolates acted in synergy; these CRC-derived strains worked together to promote pro- and anti- inflammatory host responses.

Interestingly, when Fn 7-1and 7/4 Fmu3 were included in the bacterial consortia, the concentration of CXCL-8 secreted by the Caco-2 cells significantly decreased compared to infection with either of the Fn strains individually (P<0.001). Therefore, the ability of Fn to produce a pro-inflammatory response in the host cell is reduced in the presence of these additional CRC-derived isolates. It is important to reinforce that these results were obtained when comparing all bacterial strains of the same concentration in the different consortia. The result may therefore be a modulatory effect of the other CRC-derived bacterial strains on Fn potentially caused by the extensive coaggregation observed amongst the CRC-derived oral and

Fn strains (Kolenbrander et al., 2002). Both tested Fn strains (Fn 7-1 and 7/4Fmu3) were shown to coaggregate to a high degree to all of the CRC-derived oral strains (Section 3.3A). Co- aggregation between Fn and the oral-associated isolates may have resulted in, A) the reduction of CXCL-8 secretion due to competition for adhesion sites (either on Fn cells themselves, or on the Caco-2 cells); B) the release of proteases and/or enzymes by the CRC-derived isolates

(capable of CXCL-8 degradation); or C) the interference of the host cell signalling cascades by the CRC-derived isolates to prevent CXCL-8 expression and/or the release of bacteriocins, aimed at killing the Fn isolates. All of these scenarios would result in a decrease in the pro- inflammatory response by the host cells, and therefore the prolonged survival of the bacterial members of the synergistic polymicrobial unit.

Without budgetary and time constraints, ELISA analysis of the cytokine secretion by

Caco-2 cells on infection with Fn and bacterial consortia should be repeated using only strains

157 isolated from a single patient. Since the results of this study were obtained using strains isolated from different patients, the strains may not have been compatible/complementary to each other, and may have potentially been antagonistic to one another. Based on the co-aggregation assays from Section 3.5, it seems that the same species isolated from different patients co-aggregate similarly to species isolated from the same patients. However, invasion similarities still need to be assessed. In this regard, a single patient would provide a set of bacterial strains that have co- evolved within the same host. As an example, patient 7 (cohort 3) strains would provide an ideal consortium because of the successful isolation of the pathogenic and oral-associated isolates: Fn,

Granulicatella adiacens, Gemella haemolysans, Campylobacter concisus, Eggerthella lenta, and

Klebsiella oxytoca.

In order to determine how the presence of cytokines might interfere with Fn invasion, as a final set of experiments, biologically relevant concentrations of each cytokine were added prior to Caco-2 cell infection by Fn in an invasion assay (as described in section 2.2.1 C). Since we were working with the cytokines themselves, and for reasons of cost, it was only feasible to work with one strain of Fn for this work. The IBD-derived Fn 7-1 was chosen for this preliminary study because, even though it is not a CRC-derived isolate, it is very well characterized and is both highly adhesive and highly invasive, as compared to any individual CRC-derived Fn isolate.

The main cytokine of interest was IL-27 because of its role in the mediation of inflammation during chronic inflammatory and autoimmune diseases (Wynick et al., 2014). IL-

27 mediates inflammation through various cell signalling pathways aimed at regulating T-helper cells (Meka et al., 2015). IL-6 was also chosen because of its prevalence in various types of cancers (Suzuki et al., 2011) and its ability to act in a manner antagonistic to IL-27 (Meka et al.,

158

2015). IL-10 was selected for its ability to suppress the inflammatory response by down- regulating the release of pro-inflammatory cytokines by host cells (Latorre et al., 2014;

Engelhardt & Grimnacher, 2014). IL-12 was selected because it is pivotal in priming the immune system to recognize pathogens by inducing IFNγ (Gee et al., 2008). Finally, CXCL-8 was chosen as it has been suggested that over-expression of CXCL-8 by epithelial cells promotes adhesion and subsequent invasion of some species of bacteria (Kuai et al., 2012).

After treatments with each aforementioned cytokine prior to Fn 7-1 infection, results showed that the addition of IL-27, IL-12 and CXCL-8 to Caco-2 cells significantly reduced the ability of Fn 7-1 to invade Caco-2 cells (P<0.00001, P<0.0001 and P<0.001 respectively). The addition of CXCL-8 and its ability to reduce Fn 7-1 invasion in Caco-2 cells was unexpected.

CXCL-8 has been previously shown to cause a decrease in the expression of E-cadherin mRNA, which resulted in the decreased presence of these adhesive, trans-membrane proteins on the surface of the Caco-2 cells (Kitadai et al., 2000; Tahara, 2000). Fn have been shown to use E- cadherin proteins to adhere to, and subsequently invade, host epithelial cells (Nelson et al., 2005;

Fardini et al., 2011; Rubinstein et al., 2013). Thus, a reduction in E-cadherin proteins on the surface of epithelial cells, caused by the addition of CXCL-8 prior to bacterial infection, may provide a possible explanation for why CXCL-8 decreased the ability of Fn 7-1 to invade Caco-2 cells. As a future direction it would be beneficial to also test the effects these cytokines have on the ability of CRC-derived Fn strains to invade Caco-2 cells. Based on the reduction in invasive capabilities of the IBD-derived Fn 7-1 strain in this study, we expect that similar results will occur in CRC-derived Fn strains that are also highly invasive.

In summary, while the work presented here does begin to demonstrate the invasive potential of CRC-derived Fn strains, the host inflammatory response to those specific Fn strains

159

(and other oral-associated CRC strains), and the effects that host cytokines have on Fn invasion, it is important to emphasize that this was, overall, a preliminary study with a limited number of isolates. It has been shown that the coaggregation of Streptococcus cristatus with Fn can facilitate invasion of the former non-invasive species into cultured host cells, and to alter the host response to Fn invasion (Edwards et al., 2006; Zhang et al., 2011). Thus, in some circumstances, the pathogenicity of Fusobacterium spp. is thought to be related to their ability to act as a vector for facilitation of host tissue infection by co-adherent bacteria (Fardini et al., 2011). Therefore, additional work should be done in order to further define the role of synergy between CRC- derived Fn and oral-associated isolates on host cells in the human gut. In particular, the synergistic relationship between the highly co-aggregative Fn and Campylobacter strains

(isolated from the CRC biopsies) should be assessed in relation to their combined invasive and cytotoxic potential as a shared unit using the described invasion assay/pixel quantification pipeline and cell viability assays.

160

Chapter 5. Determining the role of MORN2 domain- associated proteins in Fusobacterium adhesion/invasion

Genetic variations shared by invasive or non-invasive Fn have not yet been identified.

The lack of genetic tools for manipulating these organisms has resulted in a limited understanding of the genes responsible specifically for adherence to, and invasion of, host cells.

Genetic differences between these two phenotypes were thought to either be extremely subtle or the genes/gene pathways responsible for invasion are not shared by either group of Fn. As such, collaborators at the Broad Institute of MIT applied whole-genome sequencing and comparative analysis to study the evolution of active and passive invasion strategies by 26 Fusobacterium species. They also sought to infer factors associated with active forms of host cell invasion. The results from this study were published in collaboration with our lab in mBio (McGuire et al.,

2014), and are summarized below.

Only some Fusobacterium species can invade host cells independently (active invaders), while others require either helper organisms, co-infection with viruses, or compromised mucosal integrity (passive invaders) (Strauss et al., 2011; Huggan & Murdoch, 2008; Han et al., 2000).

To identify genetic factors associated with active invasion, our collaborators stratified the

Fusobacterium species into active and passive invader clades using a Bayesian approach. Then, information from the literature regarding each species’ invasion potential (Strauss et al., 2011;

Ohkusa et al., 2009; Huggan & Murdoch 2008; Han et al., 2000) was correlated to the different clades. Five distinct clades were identified among Fusobacterium species (Figure 5.1). Three of

161 the clades—clades A (Fn), B (F. periodonticum) and C (F. ulcerans/F. varium) —were composed entirely of species known to actively invade host cells.

Comparisons between active and passive invaders revealed the following information: active invader genomes were, on average, 560 kb larger and contained 257 more genes; active invaders had 1.6-fold more genes with predicted signal peptides (365 versus 233 per genome) and double the membrane-related protein coding capacity; and 44 orthogroups were found to be exclusive to the active invader genomes. Of the 44 orthogroups, 32 were annotated as hypothetical proteins. To gain clues as to the function of these hypothetical proteins, collaborators at the Broad Institute used the GO, Pfam and KEGG functional annotations to identify additional differences between the two phenotypes by comparative analysis of the 26 high quality draft Fusobacterium species genomes (Figure 5.2 A). Of the expanded Pfam domains, MORN2 (membrane occupation and recognition nexus) (Pfam identifier PF07661) was the most intriguing because it represented the most frequent domain in the active invader genomes (32.4 versus 4.8 genes) (Figure 5.2 B). The MORN2 domain itself is 22 to 23 amino acids long, often found in multiple copies per gene and is also highly variable. Members of the same subspecies can differ in their MORN2 gene content by as much as 25% (Figure 5.3). The small number of MORN2 containing genes found in the passive invaders was also present in nearly all active invaders, suggesting that MORN2 genes evolved and expanded in active invader species from an ancestral set of MORN2 genes held by the last common ancestor at the time of adaptive radiation. None of the 697 genes encoding MORN2 domains within our dataset had known function, although 87% were predicted by SignalP (Petersen et al., 2011) to have signal peptides targeting them for export into the extracellular environment or membrane insertion.

This highlights a potential role for these domains at the host-pathogen interface.

162

Another interesting finding by our Broad Institute collaborators was that the actively invading Fusobacterium genomes exhibited an exceptional level of genomic rearrangement, whereas the genomes of passive invaders, with fewer MORN2 containing genes, have fewer rearrangements (p=9e-6), as measured by syntenic fraction. In a whole–genome alignment of the seven fully finished Fn genomes, MORN2 containing genes were also shown to associate with synteny breaks.

Since nothing was known about the function of MORN2 domain containing proteins, and information gathered by comparative analysis hinted at a potential adhesive role at the host- pathogen interface, co-aggregation and invasion competition assays were performed as part of this thesis work. This study used a MORN2 domain containing peptide, in order to better understand the biological role(s) of MORN2 domains in Fn virulence.

163

Figure 5.1 Phylogenetic tree based on nucleotide sequences of 498 core orthogroups, or orthogroups containing exactly one copy from each of the 26 Fusobacterium strains plus the outgroup, Leptotrichia buccalis. Bootstrap values are indicated for each node. The node indicated with an arrow illustrates a 3-way trifurcation (based on bootstrap values and individual orthogroup trees [see Text S1 in the supplemental material]), representing an adaptive radiation. The five clades are outlined with boxes. The clades containing species believed to actively or passively invade host cells are indicated with dark and light gray shading, respectively. The mechanism by which F. mortiferum (in white) can invade cells is unknown. (McGuire et al., 2014)

164

Figure 5.2 Gene categories expanded in active and passive invaders (Q< 0.0005). (A) The top 12 GO terms expanded in the active invaders were largely membrane related. (B) Pfam domains expanded in the active invaders include the known virulence-related adhesins FadA and RadD, as well as a massive expansion of MORN2 domains of unknown function. (C) A different set of adhesins, including the trimeric autotransporter adhesion protein YadA, is expanded in the passive invaders. (D) Clade-specific summary of expanded Pfam domains. The number of dots represents the average number of domains present in a genome from each of the five clades. For MORN2, there can be multiple domains per gene. (McGuire et al., 2014)

165

Figure 5.3 Species distribution of MORN2 orthogroups. Each column represents an orthogroup. The boxes are colored according to the average number of MORN2 domains per protein, and for each orthogroup containing two or more paralogs, the numbers indicate the number of paralogs. There is tremendous variation in the number and structure of proteins containing MORN2 domains across genomes. Active invaders (species names shaded in dark gray) contain longer, more complex MORN2 containing proteins than passive invaders (species names shaded in light gray). Passive invaders contain only a small complement of short, “ancestral” orthogroups, which are present across all species. (McGuire et al., 2014)

166

Results

5.1 Characterization of MORN2 domains in adhesion and invasion of Fn strains

5.1 A. Verification of the MORN2-domain containing protein surrogate

Since MORN2 peptides were unable to be cloned directly from the Fn 7-1 genome, a surrogate MORN2 domain containing protein was cloned and purified from Pseudomonas aeruginosa by the Dr. Berger laboratory at Lehigh University. The cloned P. aeruginosa

MORN2 recombinant protein contained 3 different MORN2 domains, as predicted by the Pfam database. The alignment of the MORN2 domain-containing region of the P. aeruginosa recombinant protein and the Fn MORN2 consensus sequence was assessed by ClustalW on

Geneious v 8.1 (Boimatters Limited; Auckland, NZ) (Figure 5.6). Although the MORN2 domain sequences between P. aeruginosa and Fn were variable, the P. aeruginosa surrogate was deemed acceptable as extensive variability is also shown within different fusobacterial species as well.

167

Figure 5.4 Alignment of one MORN2 domain sequence from P. aeruginosa (pfam0493), one MORN2 domain sequence from Fn 7-1 and the Fn MORN2 domain consensus sequence.

168

5.1 B Competitive aggregation assay

To test if the prior addition of MORN2 recombinant protein could increase the aggregative ability of fusobacterial species, and to indicate if MORN2 domains were capable of influencing self-aggregation, a competitive aggregation assay was performed as described in

Section 2.3.1 C. Briefly, Fn 7-1 was grown in TSBsupp, washed and re-suspended in co- aggregation buffer to a McFarland standard of 3.0 (Washington et al., 1979). MORN2 recombinant proteins were then separately added at a concentration of 0.2 mg/mL to the fusobacterial cells and incubated at 37ºC at 110 rpm for 1 hr and 24 hrs. The addition of 0.2 mg/mL bovine serum albumin (BSA) and maltose binding protein (MBP) were used as the negative controls. MBP represents a better negative control because it was fused to the P. aeruginosa MORN2 recombinant proteins to facilitate purification and to increase the solubility by allowing the protein to fold properly. OD measurements were then taken and assessed using the simplified equation described by Handley et al., (1987). Statistical significance was then determined for Fn 7-1 aggregation between MORN2 recombinant proteins and the control proteins (BSA and MBP) using the Student`s T-test at a 95% confidence level (Table 5.1). There were no significant differences in the aggregation capacity of Fn 7-1 at the 1 hr time point. At the 24 hr time point a significant difference in aggregation was observed between Fn 7-1 and the

BSA control. However there was no difference in aggregation between additions of the MORN2 peptides and MBP. As such, it would appear that MORN2 is unable to increase the aggregative capacity of Fn 7-1.

169

Table 5.1 Fn 7-1 examined for its ability to self-aggregate with the addition of MORN2 peptides, MBP or BSA. Aggregative abilities between Fn 7-1 and the peptides were determined by spectrophotometric readings and the simplified equation from Section 2.3.1 C (Handley et al., 1987). Significance between MORN2 peptide additions and Fn 7-1 aggregations were determined using the Student’s T-test relative to the two control peptides, BSA or MBP (P<0.05).

Peptide Concentration Time Percent Aggregation (%) with Fn 7-1 Addition point % ± SD1 P-value (control to MORN2)

Bovine serum 1 hr 44.00 ± 2.30 0.4697 albumin protein (control) 24 hr 41.45 ± 1.60 0.0005*

Maltose binding 1 hr 46.53 ± 2.89 0.5009 protein 0.2 mg/mL (control) 24hr 62.95 ± 1.82 0.1605

P. aeruginosa 1hr 45.19 ± 1.18 See above MORN2 domain containing peptide 24hr 59.83 ± 2.55 See above

1 SD: standard deviation from three biological replicates; * values are significant (P<0.05)

170

5.1 C. MORN2 competitive infection model

As a preliminary study, an invasion assay (described in Section 2.2.1 C) was adapted to include a 30 min incubation with MORN2 recombinant proteins prior to infection with Fn 7-1.

This preliminary assessment sought to determine if the MORN2 domain containing recombinant proteins could physically bind to the Caco-2 cells at the same binding sites as Fn. The blockage of the binding sites would result in the inability of Fn to adhere and subsequently invade the tissue cells in vitro. Briefly, following the addition of 0.2 mg/mL MORN2 protein, Caco-2 cells were washed three times and then infected with Fn 7-1 (as described in Section 2.3.1 D).

Following a 4 hr infection with the Fn 7-1 strain, the cells were washed, fixed, differentially stained (Section 2.2.1 C), imaged on the Zeiss Axiovert 200 microscope and processed through the pixel quantification pipeline (Section 2.2.1 D). Results, shown in Figure 5.7, show that the addition of MORN2 recombinant proteins did result in a statistically significant decrease in the ability of Fn 7-1 to successfully invade the epithelial tissue cells in vitro.

171

Figure 5.5 A: Representative immunofluorescence micrographs of Caco-2 cells (green) previously incubated with 0.2 mg/mL MORN2 peptides and subsequently infected with Fn 7-1 for 4 hrs and fixed. Cells were then differentially stained to show bacterial cells either exterior to (purple) or interior within (orange/red) Caco-2 cells. Scale bars = 22µm. B: Bar chart indicating the normalized percentage of invasion after incubation with or without MORN2 peptides and subsequent infection with Fn 7-1, as measured using the pixel quantification as described in Section 2.2.1 D. ANOVA was used for pairwise comparisons between Fn 7-1 invasion with and without the addition of MORN2 peptides. I II A

I- no treatment; II- 0.2 mg/mL MORN2 peptide

B 5.00E-05 2 2 - 4.50E-05 4.00E-05 3.50E-05 3.00E-05 2.50E-05

area 2.00E-05

1.50E-05 * 1.00E-05 5.00E-06

0.00E+00 Percent Invasion/ Caco Invasion/ Percent No treatment MORN2 (0.2 mg/mL) Fn 7-1 Invasion 172

Discussion

Our collaborative study with colleagues from the Broad Institute of MIT (McGuire et al.,

2014) represents, by far, the largest comparative genome analysis of Fusobacterium strains.

Once again a high degree of heterogeneity was exhibited by this group of 26 strains representing

7 species of Fusobacterium. Analysis highlighted features unique to individual species and subspecies, providing clues as to how this genetically intractable genus of opportunistic pathogens diverged. Suggestions were also made regarding the selective forces that likely drove divergence. The overall observation that a large number of species-specific genes encoded for membrane proteins and adhesins is consistent with Fusobacterium having an unusual ability to adhere to various ligands (Kolenbrander, 2000), including to host cells (Edwards et al., 2006), and highlighting the host-microbe interface as an important driver of diversity within this genus

(Bolstad et al., 1996).

Focus on differentiating active and passive invader genomes provided a wealth of newly discovered information regarding the Fusobacterium genus. Features enriched in active invader genomes included a massive expansion of genes encoding membrane-associated proteins, including the known virulence adhesion, FadA, and a set of short, repeated membrane-associated protein domains designated MORN2 (Figure 5.2 D). MORN2 domain-containing proteins were encoded within sets of genes with no known function and are incredibly diverse in the organization and grouping of domains.

There appears to have been an ancient event leading to a striking expansion and diversification of MORN2 domains in the active invaders. Further diversity-yielding events, potentially mediated by the repetitive nature of the tandem domain repeats themselves, have

173 additionally diversified and expanded MORN2 domains within each of the clades (Figure 5.1 &

5.2 D). Tandem domain repeats themselves are often facilitators of rapid evolution, usually imparting useful phenotypic consequences, including rapid variation at the microbial cell surface

(Gemayel et al., 2012). Recombination between repeat sequences has been proposed to explain expansion in the number of repetitive elements within genes (Gemayel et al., 2012) and may explain why MORN2 genes from active invaders possess more copies of the MORN2 domain than passive invaders. Recombination between non-randomly distributed, repetitive chromosomal sequences is a widely conserved mechanism to promote genome diversity in prokaryotes, including Helicobacter pylori and Mycobacterium tuberculosis (Aras et al., 2003).

Therefore, it is possible that the repetitive MORN2 regions could be driving this elevated rate of chromosomal rearrangement in Fusobacterium, which could, in turn, allow the bacteria to rapidly adapt to their varied ecological niches and diverse environmental stresses within the host.

The domain organization of MORN2 genes points to an additional role for their protein products in adhesion. Examples of diverse expanded families of adhesins have been observed in other organisms, where variable numbers of repeats in cell wall proteins allow for rapid modulation of adhesive properties, adaptation to the environments, or evasion of the host immune system (Gemayel et al., 2012; Oh et al., 2005; Verstrepen et al., 2005). Often these domains represent subunits, which oligomerize to form a large, variable structure. In H. pylori, genes encoding Sel1-like repeats (SLRs) are involved in adaptation of H. pylori to specific hosts, due to strain-specific variations in the numbers of SLRs (Mittl & Schneider-Brachert, 2007).

SLR genes are similar to MORN2 genes, in that they are of similar length, are poorly conserved, and have repetitive domains, with a similar pattern of conserved residues. The repetitive,

174 modular structure of MORN2 genes points to their involvement in adhesion and further promotes their role in rapid adaptation to diverse environmental conditions.

In attempts to determine if MORN2 containing proteins play a role in Fusobacterium virulence, the aggregation and competitive invasion assays were performed using the surrogate

P. aeruginosa MORN2 containing protein. Results suggest that while MORN2 domains do not appear to aggregate to one another, they do appear to be capable of interfering with the invasive capability of Fn. Whether this interference is directly caused by the MORN2 domains binding to the same epithelial adhesion sites used by Fn, or whether the surrogate proteins were simply obstructing the Fn-specific binding site deserves further scrutiny. The P. aeruginosa MORN2 surrogate protein did vary in amino acid sequence from those seen specifically in Fn. The surrogate also contained a maltose binding protein for added stability and solubility. These factors may have contributed to the prevention of Fn adhesion to the Caco-2 cells (followed by invasion) in vitro.

While insights from this study represent an important step forward in unraveling the mechanisms of Fusobacterium pathogenesis, further research may reveal MORN2 domain function and may indicate whether proteins with MORN2 domains are good anti-vaccine targets.

This would be of high interest considering the association between Fn and human diseases, such as CRC (Castellarin et al., 2012; Kostic et al., 2012), and therefore represent a diagnostic and therapeutic strategy for the detection and treatment of fusobacterial disease.

In addition to the chromosomal rearrangements observed near the MORN2 genes, there was also a spatial correlation between MORN2 genes and mobile elements such as IS elements and transposases (McGuire et al., 2014). This information indicated that mobile genetic elements

175 may be involved in MORN2 diversification. While there was no evidence to suggest that

MORN2 genes were recently acquired, the close proximity of MORN2 and phage related genes lead us to investigate Fn bacteriophage in attempts to further characterize and define the virulence associated factors from Fn. This work can be found in the next chapter.

176

Chapter 6. Determining the presence and role of bacteriophage in invasive Fn strains

The potential importance of Fn in gastrointestinal (GI) diseases such as inflammatory bowel disease and colorectal cancer has become evident (Castellarin et al., 2012; Kostic et al.,

2012; Allen-Vercoe et al., 2011). Fn is unusual in its heterogeneity, with a wide range of phenotypic and genotypic variations evident within the species. For example, a high degree of serovar and ribotype heterogeneity, as well as differences in 16S rRNA gene-based DGGE profiling, have been observed among Fn strains (Thurnheer et al., 1999; Strauss et al., 2008). It is also evident that there may be strain-dependent differences in virulence (e.g. invasive ability)

(Strauss et al., 2011). As such, a greater understanding of the virulence potential of Fn is warranted, given the emergence of this species as an opportunistic pathogen.

Bacteriophages are viruses that can infect only bacteria (Fong, 1941). They are often used as powerful tools for the study of bacterial genetics, and, given their host specificity, are useful in the identification and characterization of their host bacterial species (Machuca et al., 2010).

The contribution of bacteriophages to the pathogenicity of their bacterial hosts has been well documented (Preus et al., 1987; Wagner & Waldor, 2002). Exotoxins are the most widely recognized virulence factor linked to bacteriophage infection, with the most common example being the cholera toxin gene located in the genome of CTXɸ, a bacteriophage from Vibrio cholerae (McLeod et al., 2005). Bacteriophages have also been shown to alter other host bacterial properties including bacterial adhesion, colonization, invasion, the spread through human tissues, resistance to immune defences, resistance to antibiotics and transmissibility among humans (Wagner & Waldor 2002).

177

Therefore, the identification and characterization of Fn bacteriophage may help to define their roles in pathogenesis. A greater understanding of bacteriophage and their associated genomes may help to elucidate the evolution of the species, as well as to delineate a method for strain typing, which for the heterogeneous Fn species is very complex (Thurnheer et al., 1999).

Previous to this work, only 1 Fn bacteriophage, Fnpɸ2, has been described in detail (Machuca et al., 2010). Due to the need for a greater understanding of the virulence potential of Fn and the well documented contribution of bacteriophage to the pathogenicty of their bacterial hosts

(Wagner and Waldorf, 2002; Preus et al., 1987), identification and characterization of two bacteriophage, ɸFunu1 and ɸFunu2, from Fn 7-1 was completed and published (Cochrane et al.,

2015). In addition, this manuscript includes the use of the PHAST bacteriophage prediction software (Zhou et al., 2011) to predict other bacteriophages located in the genomes of 28 other fully sequenced Fusobacterium strains (the same strains tested for MORN2-domain containing proteins from Chapter 5) isolated from patients with IBD or undergoing colorectal cancer screening (Figure 6.1). This bacteriophage prediction work was completed by our collaborators at the Broad Institute at MIT. A bacteriophage sequence clustering algorithm, designed at the

Broad Institute, was also applied to classify the predicted bacteriophage. Computational bacteriophage prediction was performed for two reasons. The first was so that with the sequencing of ɸFunu1 and ɸFunu2 we could determine the validity of the prediction software.

The second reason was to see if the heterogeneity observed in the bacteriophage isolates from the

Fn 7-1strain was also observed in other Fusobacterium strains.

PHAST was able to correctly predict the presence of the two prophage (ɸFunu1 and

ɸFunu2) in Fn 7-1. Based on the bacteriophage predictions in the other strains, it is believed that many strains of Fn harbour temperate bacteriophage within their genomes, and that there is a

178 high level of bacteriophage diversity within the Fusobacterium genus. Across the 29 genomes, a total of 87 predicted bacteriophage were observed (7 “intact”, 14 “questionable”, and 66

“incomplete”) (Figure 6.1). Based on the clustering analysis, Fn strains 11-3-2, 4-1-13 and

F0401 all contain similar bacteriophage sequences that group with ɸFunu1. These were all predicted by PHAST as “incomplete” bacteriophage predictions (PHAST score <70). Sequences similar to ɸFunu2 could not be detected in the sequenced genomes of all other Fn strains sequenced to date.

179

Figure 6.1 Schematic showing the phylogeny of Fusobacterium species and their relevant predicted bacteriophage.

180

Results

6.1 Induction and characterization of the invasive Fn 7-1 bacteriophage ɸFunu1 and ɸFunu2

6.1 A. Mitomycin C induction.

Mitomycin C successfully induced lysis of prophage from Fn 7-1. Two bacteriophages were obtained from this strain and were designated ɸFunu1 and ɸFunu2. Bacteriophage titers in induced lysates were undetermined as both ɸFunu1 and ɸFunu2 were unable to produce plaques using a standard plaque assay and a variety of different Fn strains including the originating strain, Fn 7-1, and isolates representative of the animalis, vincentii and polymorphum subspecies

(data not shown).

6.1 B. Sequence and analysis of Fn 7-1 bacteriophage ɸFunu1.

The genome of ɸFunu1 (Genbank accession no. KR131710) consisted of linear double- stranded DNA (dsDNA), with one scaffold of length of 39,921 bp and GC content of 27%.

ɸFunu1 mapped to a co-linear stretch of the Fn 7-1 genome from positions 810,600 to 854,000.

Annotation of ɸFunu1 revealed 66 coding DNA sequences. Further genome analysis revealed that 71.2% of the ɸFunu1 genes (47 of 66) encoded unique proteins with no reliable identity to database entries. Annotation also revealed that ɸFunu1 had no tRNA genes, one integrase gene, one capsid gene and seven genes associated with DNA replication, recombination and repair. No significant homology to known bacterial virulence genes were detected through comparison of sequences to the Virulence Factors of Pathogenic Bacteria (VFPB) database

(http://www.mgc.ac.cn/VFs/). According to PHAST (http://phast.wishartlab.com/) ɸFunu1 was most similar to BcepMu, a Mu-like Myoviridiae bacteriophage from Burkholderia cenocepacia

(Summer et al., 2004). Further verification from Virfam

181

(http://biodev.cea.fr/virfam/Default.aspx) confirmed ɸFunu1 is most closely related to viruses from the family Myoviridiae.

6.1 C. Sequence and analysis of Fn 7-1 bacteriophage ɸFunu2.

The genome of the ɸFunu2 (Genbank accession no. KR131712) consisted of linear dsDNA comprised of two scaffolds (lengths 38,801 and 1,043 bp), for a total length of 39,844 bp, with GC content of 27.7%. The ɸFunu2 genome mapped to a co-linear stretch of the 7-1 genome from position 2,205,500 to 2,244,400. Annotation by Prodigal revealed 71 coding sequences. Of these 71 genes, 56 (78.9%) encoded hypothetical proteins with no reliable identity to database entries. Genome analysis also revealed ɸFunu2 had no tRNA genes, one integrase gene, one envelope (coat) gene and nine genes associated with DNA replication, recombination and repair. Interestingly, the ɸFunu2 genome includes a toxin secretion/bacteriophage lysis holin gene, yet does not contain a known endolysin gene, and this bacteriophage was unable to produce plaques in a plaque assay with purified bacteriophage on soft agar. Both PHAST and

VIRFAM showed that ɸFunu2 is most similar to SboM-AG3, a bacteriophage that infects

Shigella boydii that also belongs to the Myoviridiae family (Anany et al., 2011). Using the

NUCmer program to align the DNA sequences of ɸFunu1 and ɸFunu2, a short region of sequence similarity was observed between them.

6.1 D. Electron microscopy of the bacteriophage particles.

Two distinct bacteriophage morphologies were observed, representing both ɸFunu1 and

ɸFunu2 (see Figure 6.2 A and B). Both virion morphologies have a rough hexagonal outline indicating an icosahedral nature, as is the case with most virions from the myoviridae family.

The virions’ head size ranges from 30-40 nm in diameter for ɸFunu1 and 50-60 nm for ɸFunu2.

As suggested by the annotation of the ɸFunu1 genome indicating tail proteins, ɸFunu1 is

182 assumed to be the bacteriophage with a tail visible in the electron micrographs. The tail region appears to be 60-120 nm in length and includes a distinct neck region.

6.1 E. ɸFunu1 and ɸFunu2 alignment with their reference parent genome Fn 7-1 .

As previously stated above, ɸFunu1 maps to the co-linear stretch of the Fn 7-1 genome from position 810,600- 854,000 and ɸFunu2 maps from position 2,205,500- 2,244,400 (Figure

6.3).

183

Figure 6.2 Electron Micrographs depicting the bacteriophage isolated from Fn 7-1: A: ɸFunu1; B: ɸFunu2.

184

Figure 6.3 Aligned bacteriophage genomes (ɸFunu1 and ɸFunu2) to their Fn 7-1 bacterial host genome. A: The full Fn 7-1 genome (lengthwise); B: ɸFunu1 prophage alignment (one scaffold); and C: ɸFunu2 prophage alignment (two scaffolds).

185

6.2 Induction and characterization of the bacteriophage from Fn 13-3C

While isolating bacteriophage from Fn 7-1, additional Fn strains were also induced in attempts to identify the presence of additional prophage, seen in the PHAST predictions

(Cochrane et al., 2015). Included in the Fn strains tested was Fn 13-3C, a highly invasive strain isolated from a patient with IBD (Strauss et al., 2008). Upon phage induction, purification and visualization of phage particles by electron microscopy, it was quite clear that the Fn 13-3C genome contained at least one inducible and intact prophage. This was not surprising because most of the Fn strains tested resulted in the visualization of bacteriophage. However, the PHAST bacteriophage prediction and the bacteriophage clustering algorithm (Cochrane et al., 2015) predicted that Fn 13-3C did not contain any bacteriophage. To both improve the bacteriophage prediction software and to determine the annotated predicted protein content in another bacteriophage from a highly invasive Fn strain, the Fn 13-3C bacteriophage DNA was isolated, sequenced, assembled and annotated.

6.2 A. Mitomycin C induction and electron microscopy of the bacteriophage particles.

Mitomycin C successfully induced lysis for Fn 13-3C as indicated by the presence of bacteriophage particles via electron microscopy (Figure 6.4). The virion morphology of this induced prophage, designated ɸNuPh1 (new phage 1), has a rough pentagonal outline indicating an icosahedral nature. The virions’ head size ranges from 50-70 nm in diameter. The tail region appears to be 75-110 nm in length and includes a distinct neck region.

6.2 B. Sequence and analysis of ɸNuPh1.

After de novo assembly of the raw reads using Geneious v.8.1 (Biomatters Limited;

Auckland, NZ), the ɸNuPh1consensus sequence indicated a 40,943bp circular genome with

43.8% GC content. Interestingly however, when using the PHAST software

186

(http://phast.wishartlab.com/), it was determined that the intact prophage genome is only represented by nucleotides 2803-40,768 bp in the circular de novo assembled ɸNuPh1consensus sequence (Figure 6.5). The remaining sequence encoded various proteins not directly associated with bacteriophage, and therefore, not recognized by the PHAST database. The ɸNuPh1did not map to the bacterial Fn 13-3C genome. Annotation of ɸNuPh1 revealed 65 coding DNA sequences. Further genome analysis revealed that 75.3% of the ɸNuPh1 genes (49 of 65) encoded hypothetical proteins with no reliable identity to database entries. Genomic analysis, identifying any BLAST bacteriophage elements, also revealed ɸNuPh1 had no tRNA genes, 1 terminase gene, 1 integrase gene, 2 tail genes, 1 coat protein, 1 plate protein, 9 other bacteriophage related proteins, no rRNA genes and no protease proteins (Figure 6.6). No significant homology to known bacterial virulence genes were detected through comparison of sequences to the Virulence Factors of Pathogenic Bacteria (VFPB) database

(http://www.mgc.ac.cn/VFs/). Both PHAST (http://phast.wishartlab.com/) and VIRFAM

(http://biodev.cea.fr/virfam/Default.aspx) showed that ɸNuPh1 is most similar to Aaphi23, a

Myoviridiae family bacteriophage that infects Actinobacillus actinomycetemcomitans (Resch et al., 2004).

Annotation of the nucleotide sequence outside of the ɸNuPh1 intact prophage genome was also assessed (1-2802 bp). Genome analysis of this region, as predicted by BLASTing the open reading frames (as determined on Geneious r8) shows the presence of both a putative bacteriophage repressor gene and a bacteriophage anti-repressor gene (with homology to

Campylobacter showae/coli), an ATP-dependant protease hsl gene (with homology to

Campylobacter showae), and other various hypothetical protein genes (with homology to

187

Campylobacter showae/rectus/curvus; Haemophilis parasuis; Fusobacterium nucleatum;

Salmonella enterica).

6.2 C The presence of ɸNuPh1 gene amplicons in different bacterial species as determined by PCR analysis

Since ɸNuPh1 did not map directly to the Fn 13-3C genome, to verify the primary host, and to test if ɸNuPh1 gene amplicons could be found in other bacterial species, primers were designed to amplify the presence of four ɸNuPh1 bacteriophage-specific genes. Bacterial strains tested for ɸNuPh1 gene amplicons included Fn 13-3C, a Campylobacter showae strain (found to co-exist with Fn 13-3C), CC57C (C. showae), 7/5 D55 (C. concisus) and Clostridium difficile strain CNA10. The resulting agarose gel electrophoresis image can be found in Figure 6.7. As indicated by the presence of the amplified gene targets, ɸNuPh1 appears to be a bacteriophage belonging to the genus Campylobacter and not Fusobacterium.

In an attempt to separate any co-existing bacteria, the Fn 13-3C isolate was extensively passaged over 50 generations to deliberately cause any residing strains to become lab-adapted and therefore to no longer rely on the Fn 13-3C strain for survival. As expected, over time the margin of the Fn 13-3C colonies became distinct and a separate colony type was able to be purified from the Fn 13-3C strain. Crude DNA was extracted from the newly emerged translucent, irregular colonies and the V3 region of the 16S rRNA gene was amplified and

Sanger sequencing was performed at the Advanced Genomics Facility at the University of

Guelph (as described in Section 2.1.2 A&B). Obtained sequences were BLASTed against the database curated by the Ribosomal Database Project

(http://rdp.cme.msu.edu/seqmatch/seqmatch_intro.jsp) in order to infer closest species identity.

The co-existing bacterial species showed closest homology to Campylobacter showae.

188

Figure 6.4 Electron Micrographs of the purified bacteriophage, ɸNuPh1

head tail

189

Figure 6.5 Linear genome view of the ɸNuPh1 de novo assembly consensus sequence with the highlighted ɸNuPh1 genome highlighted (pink) as a fully intact prophage.

190

Figure 6.6 ɸNuPh1 full genome view showing all the coding DNA sequences, their orientation and their gene types.

191

Figure 6.7 Bacteriophage specific gene amplicons by PCR from ɸNuPh1 separated by agarose gel electrophoresis Lane 1/15/17/31: GeneRuler 100bp DNA ladder (Thermo Fisher; Burlington, ON); Lane 2/9/18/25: C. showae; Lane 3/10/19/26: CC57C; Lane 4/11/20/27: 7/5 D55; Lane 5/12/21/28: Fn 13-3C; Lane 6/13/22/29: Clostridium difficile CNA10 (negative control); Lane 7/14/23/30: no DNA template (negative control); Lane C8/16/24/32: Blank

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

Tail fiber gene Hypothetical phage protein gene

17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32

Tail sheath gene Integrase gene

192

Discussion

Bacteriophage are often analyzed to help give clues to the pathogenicity of bacteria since the discovery of the cholera toxin within ɸCTX (McLeod et al., 2005). Due to the genotypic, phenotypic, phylogenetic and biochemical heterogeneity observed within the Fn genus

(Thurnheer et al., 1999) and the uncertainty about Fusobacterium spp. influence on various inflammatory diseases (Allen-Vercoe et al., 2011), it was hoped that clues about this opportunistic pathogen’s virulence potential could be found within the genome of its harboured bacteriophage. Although both ɸFunu1 and ɸFunu2 genome sequences did not definitively identify with virulence associated genes, many of their genes coded for hypothetical proteins of unknown function. It is possible that these predicted genes are involved in as-yet undetermined virulence mechanisms. Regardless, valuable information has still been obtained concerning the predicted gene content of the two Fn bacteriophages. In addition, the validation of the use of the bacteriophage prediction algorithm PHAST in Fusobacterium has been proven which may be used, together with the bacteriophage clustering algorithm described in the Cochrane et al.

(2015) manuscript for further evaluation of bacteriophage genomes within the Fusobacterium species.

Both ɸFunu1 and ɸFunu2 possess some interesting characteristics. This includes the inability of the bacteriophage lysates to induce plaque formation in a range of different Fn isolates using a simple plaque assay. Although ɸFunu2 does not appear to have any predicted proteins related to a bacteriophage tail, which may provide answers as to why it cannot bind to

Fn and induce active lysis, ɸFunu1 has multiple proteins believed to be involved in tail generation. These tails were observed in the electron micrographs (Figure 6.2 A). However, there may be a defect in these tail proteins in terms of binding capabilities. Perhaps during the

193 isolation and purification of the bacteriophage, pivotal bacteriophage tail proteins, responsible for bacterial attachment, may have been sheared off, rendering the bacteriophage unable to propagate by integration into another fusobacterial cell. Interestingly, ɸFunu2 does appear to have a holin gene, suggesting its ability to form pores in the bacterial cell membranes exposing the peptidoglycan. There is no suggestion of an endolysin gene which would allow the bacteriophage to fully degrade the bacterial cell membrane’s peptidoglycan (Wang et al., 2000).

This gene loss may be responsible for the inability of ɸFunu2 to induce plaque formation. It also may explain the improper formation of the full bacteriophage particle, as seen in some of the electron micrographs (Figure 6.2 B). Since the release of the bacteriophage from Fn 7-1 does not appear to be associated with bacterial cell rupture, it is more difficult to determine the infectivity of these two bacteriophage. It is possible that Fn bacteriophages are able to encase themselves in the Fn bacterial cell outer membrane and bleb out of the bacteria without causing lysis. The phenomenon of a budding bacteriophage has been previously reported for certain mycoplasma viruses (Maniloff et al., 1981).

Although not unexpected, there was a great deal of heterogeneity observed in the two sequenced and annotated bacteriophages, ɸFunu1 and ɸFunu2, as well the heterogeneity observed when looking at the predicted Fusobacterium bacteriophages using the cluster analysis and bacteriophage prediction software. In addition, some of the Fn bacteriophage clusters did not correlate with phylogeny, indicating the possibility of horizontal transfer of mobile bacteriophage elements across different Fn subspecies. These data further suggest that Fn heterogeneity may be related in part to bacteriophage acquisition. The roles of ɸFunu1 and

ɸFunu2 in Fn 7-1 virulence are as-yet undefined, but since Fn strains are known to differ both in

194 virulence (e.g. invasive ability) as well as bacteriophage complement, further investigation of the roles of genes encoded by these bacteriophage genomes is warranted.

Due to the potential implication resident bacteriophage have in perpetuating their bacterial host pathogenicity, it was hoped that although ɸFunu1 and ɸFunu1did not display any virulence associated genes that ɸNuPh1, which was originally thought to have been induced from the highly invasive Fn 13-3C, would give a few clues about Fn virulence potential.

However, after sequencing and assembly of the circular ɸNuPh1 genome, it was shown that there was no homology to the host bacterial Fn 13-3C genome. While annotating the ɸNuPh1 genome, using BLASTx (http://blast.ncbi.nlm.nih.gov/blast/Blast.cgi?PROGRAM=blastx), many of the bacteriophage-associated and hypothetical proteins shared homology to Campylobacter showae strain sequences. As shown in the manuscript we published (Warren et al., 2013), we knew that

Fn and Campylobacter strains can co-occur with one another within a human host. The co- occurrence of these two species and the Campylobacter spp. homology to ɸNuPh1 protein annotations caused us to question the purity of the Fn 13-3C isolate and to ask the following questions: was our Fn 13-3C isolate harbouring a Campylobacter strain (or a different bacterial species); or had the ɸNuPh1 bacteriophage, or at least the genes encoded within the bacteriophage genome, been recently acquired by the Fn 13-3C strain from a Campylobacter strain?

After over 50 generations of the Fn 13-3C isolate had been passaged, it became evident that this particular Fn strain was harbouring an additional bacterial species. After sequencing and

Ribosomal Database alignment of the 16S rRNA gene from this newly acquired bacterial strain, the closest relative was shown to be Campylobacter showae. This particular strain was then given the designation C. showae kc-1.

195

Precedence for broad-host range bacteriophages has been previously published (Jensen et al., 1998; Sano et al., 2004).To officially determine whether ɸNuPh1 belonged to the newly acquired and purified C. showae kc-1 strain, or if ɸNuPh1 participated in bacteriophage jumping to the Fn 13-3C strain, amplification of four different ɸNuPh1-specific bacteriophage-associated genes was carried out using the genomic DNA from both Fn 13-3C and the isolated C. showae kc-l strain. In addition, genomic DNA from both CC57C (C. showae) and 7/5 D55 (C. concisus) were tested to determine if ɸNuPh1 gene amplicons were present in other Campylobacter species. As shown by the pattern on the agarose gel (Figure 6.7), these four ɸNuPh1 gene amplicons were found in various Campylobacter spp., including both C. showae and C. concisus.

The presence of ɸNuPh1 within the Fn 13-3C genome was not found. These findings re-iterate the common theme found throughout this thesis: the importance of synergy between bacterial species in human infections. The Fn 13-3C strain has been previously described as a highly invasive Fn strain. Further research will now need to be performed to test whether the invasive capacity of the Fn 13-3C strain was in fact due to the co-existence/virulence synergy of the C. showae kc-1strain.

While the isolation of bacteriophage from both Fn and Campylobacter further prove the role of synergy between bacterial species in disease, particularly within the human gastrointestinal tract (an exceptionally densely populated ecosystem), no evidence was found concerning the role Fn bacteriophage play in virulence. The next chapter in this thesis lays the framework for the isolation of RNA and transcriptome research for Fn in order to determine gene function of Fn during active adhesion and invasion to host epithelial cells in vitro.

196

Chapter 7. Summary and Significance

Cancer is among the leading causes of human death worldwide. Cancer burden is attributed to the interaction between genetics, the immune system and more recently, infectious microbial agents including bacteria. It is uncertain whether bacteria directly cause cancer, facilitate cancer development or indicate the presence of a cancer (or the risk of developing one).

It is unknown which particular bacterial species are implicated in cancer presence, initiation and/or progression. Colorectal cancer is one of the most prevalent cancers. Given the immense diversity of microbes present within the human gut it is not surprising that the resident bacterial species play a role in the progression of diseases such as CRC. With this knowledge, this study sought to culture bacterial species directly from CRC biopsies. Isolation of over 124 bacterial species was accomplished; including the isolation of Fn. The role of Fn in CRC was of particular interest due to the highly elevated presence of Fn gene sequences in a subset of CRC tumours

(Castellarin et al., 2012; Kostic et al., 2012). As such, CRC-derived Fn strains were further characterized in this study.

Although Fn has been implicated in a wide variety of inflammatory diseases, Fn is also a well recognized benign resident of mucosal surfaces in the absence of pathogenicity. The reason why Fn may in some cases be pathogenic and at other times an apparently benign, commensal organism is not completely understood. This study sought to determine the virulence determinants responsible for converting this commensal bacterium into an opportunistic pathogen. Phenotypic and genotypic characterization of the CRC-derived Fn isolates confirmed the considerable heterogeneity that others have found amongst Fn strains. Wide variations in cell and colony morphology, biochemical profiles, antibiotic susceptibility and invasion potential

197 were observed. A more detailed analysis of these CRC-derived Fn strains at the genomic level would be a benefit in further defining Fn and understanding their virulence properties.

A leading hypothesis to describe the potential importance of gut bacterial species, including Fn, and inflammatory diseases such as CRC is dysbiosis. Dysbiosis is the imbalance between beneficial and harmful bacterial species. Dysbiosis can result from a disruption in the human intestinal epithelial barrier which enables harmful bacterial species to stimulate the mucosal immune system. This can then result in the initiation and perpetuation of intestinal inflammation. Alternatively, dysbiosis could also result when commensal species become functionally altered such that virulent tendencies result in increased adherence, invasion and intracellular persistence within a genetically susceptible host. The role of dysbiosis lends itself to the idea that no single bacterial species results in disease, but rather a group of bacteria that perform the same detrimental function can lead to disease etiology.

It is clear that gut microbes do not act alone and that synergy is at play. Since Fn is able to co-aggregate with a large number of unrelated oral and enteric bacterial species, Fn may not only play a role in the formation of plaque biofilms in the oral niche, but also within the gut.

Studies on biofilm communities have revealed that co-aggregation between bacterial species can lead to altered gene expression profiles and nutritionally beneficial mutualistic relationships.

These relationships enable bacterial species to grow and thrive in niches where they normally cannot. Given the immense diversity of bacterial species in the human gut and the co-aggregative properties of Fn, the potential for the formation of such mutualistic relationships in the gut is high. These mutualistic relationships could contribute to dysbiosis and increased pathogenicity of the commensal species. Therefore, in some circumstances, the pathogenicity of Fn may be

198 related to co-adherent bacteria. While this study demonstrates a preliminary assessment of the potential synergies between CRC-derived bacterial strains and the effects on host cells in vitro, additional work should be done in order to further define the role of synergy between CRC- derived Fn and orally-associated isolates on human gut epithelial cells.

The strong associations between what are typically considered oral anaerobic bacteria and colorectal carcinoma in this study are intriguing. While it is not unreasonable to assume that oral microbes may be gaining access to the gut of some CRC patients, for verification, future studies need to correlate oral health and CRC. If deteriorating oral health is indeed indicative of a risk factor in the development of CRC, preventative measures could be as easy as assessing the microbial health within the oral cavity. The co-occurrence of the oral-associated, but gut-derived

Fn and Campylobacter spp. are of particular interest in regards to oral detection of CRC risk.

Metagenomics and computational assessments of these microbes in CRC tumours have indicated that these two bacterial species co-occur within cancer tumours. Specifically, the link between

Campylobacter, and Fusobacterium spp. associated with CRC could provide a point of intervention in early CRC detection. These species together may provide a set of parsimonious predictors with potential utility in CRC detection and risk assessment via a non-invasive assay that could detect a polymicrobial signature specific to CRC via oral assessment (biomarkers).

Further confirmation of the co-occurrence between Fn and Campylobacter spp. in the gut was shown through the isolation of the Campylobacter isolated ɸPhNu1 bacteriophage. This study was able to isolate a Fn strain that had been harboring a Campylobacter sp. for quite some time within our Fn biorepository.

199

While the correlation of genotype and virulence is rarely clear-cut and as simple as the presence/absence of a gene, this study, in collaboration with The Broad Institute, sought to perform a comparative analysis between active and passive Fn strains. The comparative analysis was able to identify genetic features that distinguish the most invasive forms of Fusobacterium, explain the evolution of active forms of host cell invasion, and most importantly, discover a class of MORN2 genes of unknown function. These genes strongly associate with active forms of host cell invasion and likely represent new strategies for bacterial adherence and invasion of host cells. The insights gained represent an important step toward unraveling the mechanisms of

Fusobacterium pathogenesis and will enable development of diagnostic and therapeutic strategies for the detection and treatment of fusobacterial disease.

Bacteriophage are often analyzed to help give clues to the pathogenicity of bacteria. This study gives the details on the methods and successful characterization of two different bacteriophages that both map to the Fn 7-1 genome. Although neither of the bacteriophage genomes contained virulence associated genes, many of their genes coded for hypothetical proteins of unknown function. It is possible that these predicted genes are involved in as-yet undetermined virulence mechanisms. Perhaps of more interest was the induction of the bacteriophage from C. showae kc-1 which had been ‘hiding’ within our Fn 13-3C stock. This result further proves the influence of bacterial synergy within the human gastrointestinal tract.

Largely due to their genetic intractability, little is known about the mechanisms that have allowed Fn strains to become such pervasive, invasive pathogens. As such, this study has now turned to transcriptomics to determine the gene expression profiles of Fn during active invasion

(Appendix I). While it was no easy endeavor isolating mRNA from Fn cells that had successfully

200 infected Caco-2 cells in vitro, a method was established to allow for the progression of RNA-Seq analysis. Transcriptomics may provide a means to determine if it is the presence of specific genes in Fn strains, or the upregulation of specific gene expression, that can make them pathogenic. The successful identification of gene targets will allow for the generation of virulence determinant based vaccines (Appendix I).

The microbial role in human health is a new frontier in the way we understand our own body as a vast and diverse bacterial ecosystem. The overarching themes derived from this thesis include:

[1] mouth bacteria profiles may be indicators for systemic balances throughout our entire

body, especially within the gastrointestinal tract

[2] bacterial synergy is at play within the gut and its role in disease progression warrants

further study

[3] whether Fn is directly associated with CRC tumorigenesis or is just a resident gut

microbe taking advantage of optimal conditions at a tumour site it represents a prime

candidate for continued CRC study

201

Literature Cited

Allali I, Delgado S, Marron PI, Astudillo A, Yeh JJ, Ghazal H, Amzazi S, Keku T, Azcarate-Peril MA. (2015). Gut microbiome compositional and functional differences between tumor and non-tumor adjacent tissues from cohorts from the US and Spain. Gut Microbes 6:161– 172.

Allen-Vercoe E, Strauss J, Chadee K. (2011). Fusobacterium nucleatum: An emerging gut pathogen? Gut Microbes 2: 294–298.

Anany, H, Lingohr, EJ, Villegas, A, Ackermann, HW, She, YM, Griffiths, MW, Kropinski, AM. (2011). A Shigella boydii bacteriophage which resembles Salmonella phage ViI. Virology Journal 8: 242.

Alvarez-Martinez CE, Christie PJ. (2009). Biological diversity of prokaryotic type IV secretion systems. Microbiol Mol Biol Rev 73:775–808.

Aras RA, Kang J, Tschumi AI, Harasaki Y, Blaser MJ. (2003). Extensive repetitive DNA facilitates prokaryotic genome plasticity. Proc Natl Acad Sci USA 100:13579–13584.

Bachrach G, Rosen G, Bellalou M, Naor R, Sela MN. (2004). Identification of a Fusobacterium nucleatum 65 kDa serine protease. Oral Microbiol Immunol 19: 155-159.

Baggiolini M, Clark-Lewis I. (1992). Interleukin-8, a chemotactic and inflammatory cytokine. FEBS Lett 307:97–101.

Batten M, Kljavin NM, Li J, Walter MJ, de Sauvage FJ, Ghilardi N. (2008). Cutting edge: IL-27 is a potent inducer of IL-10 but not FoxP3 in murine T cells. J Immunol 180:2752–2756.

Ben Haj Khalifa A, Moissenet D, Vu Thien H, Khedher M. (2011). Virulence factors in Pseudomonas aeruginosa: mechanisms and modes of regulation. Ann Biol Clin (Paris) 69:393– 403.

Benson DA, Karsch-Mizrachi I, Lipman DJ, Ostell J, Wheeler DL. (2008). GenBank. Nucleic Acids Res 36: D34-D38.

Bolstad AI, Jensen HB, Bakken V. (1996). Taxonomy, biology, and periodontal aspects of Fusobacterium nucleatum. Clin Microbiol Rev 9: 55-71.

Bradshaw DJ, Marsh PD, Keith Watson G, Allison C. (1998). Role of Fusobacterium nucleatum and coaggregation in anaerobe survival in planktonic and biofilm oral microbial communities during aeration. Infect Immun 66:4729–4732.

Brazier JS, Citron DM, Goldstein EJ. (1991). A selective medium for Fusobacterium spp. J Appl Bacteriol 71: 343-346.

202

Brew R, Southern SA, Flanagan BF, McDicken IW, Christmas SE. (1996). Detection of interleukin-8 mRNA and protein in human colorectal carcinoma cells. Eur J Cancer 32:2142– 2147.

Briani F, Dehò G, Forti F, Ghisotti D. (2001). The plasmid status of satellite bacteriophage P4. Plasmid 45:1–17.

Castellarin M, Warren RL, Freeman JD, Dreolini L, Krzywinski M, Strauss J, Barnes R, Watson P, Allen-Vercoe E, Moore RA, Holt RA. (2012). Fusobacterium nucleatum infection is prevalent in human colorectal carcinoma. Genome Res 22: 299–306.

Chen W, Liu F, Ling Z, Tong X, Xiang C. (2012). Human intestinal lumen and mucosa- associated microbiota in patients with colorectal cancer. PLoS One 7(6): e39743.

Chen Q, Swaminathan S, Yang D, Dai L, Sui H, Yang J, Hornung RL, Wang Y, Huang DW, Hu X, Lempicki R a, Imamichi T. (2013). Interleukin-27 is a potent inhibitor of cis HIV- 1 replication in monocyte-derived dendritic cells via a type I interferon-independent pathway. PLoS One 8:e59194.

Cisar JO, Kolenbrander PE, McIntire FC. (1979). Specificity of coaggregation reactions between human oral streptococci and strains of Actinomyces viscosus or Actinomyces naeslundii. Infect Immun 24:742–752.

Citron DM. (2002). Update on the taxonomy and clinical aspects of the genus Fusobacterium. Clin Infect Dis 35: S22-S27.

Cline MS, Smoot M, Cerami E, Kuchinsky A, Landys N, Workman C, Christmas R, Avila- Campilo I, Creech M, Gross B, Hanspers K, Isserlin R, Kelley R, Killcoyne S, Lotia S, Maere S, Morris J, Ono K, Pavlovic V, Pico AR, Vailaya A, Wang P-L, Adler A, Conklin BR, Hood L, Kuiper M, Sander C, Schmulevich I, Schwikowski B, Warner GJ, Ideker T, Bader GD. (2007). Integration of biological networks and gene expression data using Cytoscape. Nat Protoc 2:2366–2382.

Cochrane K, McGuire AM, Priest ME, Abouelleil A, Cerqueira GC, Lo R, Earl AM, Allen- Veroce E. (2015). Complete genome sequences and analysis of the Fusobacterium nucleatum subspecies animalis 7-1 bacteriophage Funu1 and Funu2. Anaerobe. 38: 125-129.

Conesa A, Götz S, García-Gómez JM, Terol J, Talón, M, Robles M. (2005). Blast2GO: A universal tool for annotation, visualization and analysis in functional genomics research. Bioinformatics, 21(18): 3674–3676. doi:10.1093/bioinformatics/bti610.

Conrads G, Claros MC, Citron DM, Tyrrell KL, Merriam V, Goldstein EJC. (2002). 16S- 23S rDNA internal transcribed spacer sequences for analysis of the phylogenetic relationships among species of the genus Fusobacterium. Int J Syst Evol Microbiol 52:493–499.

203

Corazziari ES. (2009). Intestinal mucus barrier in normal and inflamed colon. J Pediatr Gastroenterol Nutr 48 Suppl 2: S54-S55.

Cossart P, Sansonetti PJ. (2004). Bacterial invasion: the paradigms of enteroinvasive pathogens. Science 304:242–248.

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

Darfeuille-Michaud A, Boudeau J, Bulois P, Neut C, Glasser AL, Barnich N, Bringer M. A, Swidsinski A, Beaugerie L, Colombel JF (2004). High prevalence of adherent-invasive Escherichia coli associated with ileal mucosa in crohn's disease. Gastroenterology 127: 412-421.

Darling ACE, Mau B, Blattner FR, Perna NT. (2004). Mauve: multiple alignment of conserved genomic sequence with rearrangements. Genome Res 14:1394–1403.

DeDantis CE, Lin CC, Mariotto AB, Siegel RL, Stein KD, Kramer JL, Alteri R, Robbins AS, Jemal A. (2014). Cancer treatment and survivorship statistics 64: 252-271.

Dharmani P, Chadee K. (2008). Biologic therapies against inflammatory bowel disease: A dysregulated immune system and the cross talk with gastrointestinal mucosa hold the key. Curr Mol Pharmacol 1: 195-212.

Dharmani P, Strauss J, Ambrose C, Allen-Vercoe E, Chadee K. (2011). Fusobacterium nucleatum infection of colonic cells stimulates MUC2 mucin and tumor necrosis factor alpha. Infect Immun 79: 2597-2607.

Dienz O, Rincon M. (2009). The effects of IL-6 on CD4 T cell responses. Clin Immunol 130:27–33.

Drabsch Y, ten Dijke P. (2012). TGF-β signalling and its role in cancer progression and metastasis. Cancer Metastasis Rev 31:553–568.

Dzink JL, Sheenan MT, Socransky SS. (1990). Proposal of three subspecies of Fusobacterium nucleatum Knorr 1922: Fusobacterium nucleatum subsp. nucleatum subsp. nov., comb. nov.; Fusobacterium nucleatum subsp. polymorphum subsp. nov., nom, rev., comb. nov.; and Fusobacterium nucleatum subsp. vincentii subsp. nov., nom. rev., comb. Nov. Int J Syst Bacteriol 40: 74-78.

Eckmann L, Kagnoff MF, Fierer J. (1993). Epithelial cells secrete the chemokine interleukin-8 in response to bacterial entry. Infect Immun 61:4569–4574.

Edwards AM, Grossman TJ, Rudney JD. (2006). Fusobacterium nucleatum transports noninvasive streptococcus cristatus into human epithelial cells. Infect Immun 74: 654-662.

204

Engelhardt K, Grimbacher B. (2014). IL-10 in humans: lessons from the gut, IL-10/IL-10 receptor deficiencies, and IL-10 polymorphisms. Curr Top Microbiol Immunol 380:1–18.

Enright AJ, Dongen SV, Ouzounis CA. (2002). An efficient algorithm for large-scale detection of protein families. Nucleic Acids Res 30: 1575–1584.

Etoh Y, Dewhirst FE, Paster BJ, Yamamoto A, Goto N. (1993). Campylobacter showae sp. nov., isolated from the human oral cavity. Int J Syst Bacteriol 43: 631–639.

Fardini Y, Wang X, Témoin S, Nithianantham S, Lee D, Shoham M, Han YW. (2011). Fusobacterium nucleatum adhesin FadA binds vascular endothelial cadherin and alters endothelial integrity. Mol Microbiol 82:1468–1480.

Felsenstein J. (1985). Confidence limits on phylogenies: An approach using the bootstrap Evolution 39: 783–791.

Finn RD, Tate J, Mistry J, Coggill PC, Sammut SJ, Hotz HR, Bateman A. (2007). The Pfam protein families database. Nucleic Acids Res 36: D281–D288. doi:10.1093/nar/gkm960.

Flanagan L, Schmid J, Ebert M, Soucek P, Kunicka T, Liska V, Bruha J, Neary P, Dezeeuw N, Tommasino M, Jenab M, Prehn JHM, Hughes DJ. (2014). Fusobacterium nucleatum associates with stages of colorectal neoplasia development, colorectal cancer and disease outcome. Eur J Clin Microbiol Infect Dis 33:1381–1390.

Fujimura KE, Slusher NA, Cabana MD, Lynch SV. (2010). Role of the gut microbiota in defining human health. Expert Rev Anti Infect Ther 8: 435-454.

Fong, J. (1941). Bacteriophage. University of California, Berkeley, California, U.S.A.

Gallimidi AB, Fischman S, Revach B, Bulvik R, Maliutina A, Rubinstein AM, Nussbaum G, Elkin M. (2015). Periodontal pathogens porphyromonas gingivalis and Fusobacterium nucleatum promote tumor progression in an oral-specific chemical carcinogenesis model. Oncotarget 6: 22613-22623.

Gee K, Guzzo C, Che Mat NF, Ma W, Kumar A. (2009). The IL-12 family of cytokines in infection, inflammation and autoimmune disorders. Inflamm Allergy Drug Targets 8:40–52.

Gemayel R, Cho J, Boeynaems S, Verstrepen KJ. (2012). Beyond junk-variable tandem repeats as facilitators of rapid evolution of regulatory and coding sequences. Genes (Basel) 3:461–480.

Gemmell E, Seymour GJ. (1993). Interleukin 1, interleukin 6 and transforming growth factor- beta production by human gingival mononuclear cells following stimulation with Porphyromonas gingivalis and Fusobacterium nucleatum. J Periodontal Res 28:122–129.

205

Gharbia SE, Shah HN. (1992). Fusobacterium nucleatum subsp. fusifome subsp. nov. and Fusobacterium nucleatum subsp. animalis subsp. nov. as Additional Subspecies within Fusobacterium nucleatum. Int J Syst Bacteriol 42:296–298.

Ghosh SK, Gupta S, Jiang B, Weinberg A. (2011). Fusobacterium nucleatum and human beta- defensins modulate the release of antimicrobial chemokine CCL20/macrophage inflammatory protein 3α. Infect Immun 79:4578–4587.

Gibbons RJ, Nygaard M. (1970). Interbacterial aggregation of plaque bacteria. Arch Oral Bio 15: 1397-1400.

Gloor GB, Hummelen R, Macklaim JM, Dickson RJ, Fernandes AD, MacPhee R, Reid G. (2010). Microbiome profiling by illumina sequencing of combinatorial sequence-tagged PCR products. PLoS One 5:e15406.

Goosney DL, De Grado M, Finlay BB. (1999). Putting E. coli on a pedestal: A unique system to study signal transduction and the actin cytoskeleton. Trends Cell Biol 9:11–14.

Greenstein R. J. (2003). Is Crohn's disease caused by a mycobacterium? Comparisons with leprosy, tuberculosis, and Johne's disease. Lancet Infect Dis 3: 507-514.

Gruenheid S, Finlay BB. (2003). Microbial pathogenesis and cytoskeletal function. Nature 422:775–781.

Gumbiner BM. (2005). Regulation of cadherin-mediated adhesion in morphogenesis. Nat Rev Mol Cell Biol 6: 622–634.

Gursoy UK, Könönen E, Uitto VJ. (2008). Intracellular replication of fusobacteria requires new actin filament formation of epithelial cells. Apmis 116:1063–1070.

Guzzo C, Che Mat NF, Gee K. (2010). Interleukin-27 induces a STAT1/3- and NF-kappaB- dependent proinflammatory cytokine profile in human monocytes. J Biol Chem 285:24404– 24411.

Haake S. K., Yoder S., Gerardo S. H. (2006). Efficient gene transfer and targeted mutagenesis in Fusobacterium nucleatum. Plasmid 55: 27-38.

Haft DH, Loftus BJ, Richardson DL, Yang F, Eisen JA, Paulsen IT, White O. (2001). TIGRFAMs: a protein family resource for the functional identification of proteins. Nucleic Acids Res 29: 41–43. doi:10.1093/nar/29.1.41.

Han YW, Redline RW, Li M, Yin L, Hill GB, McCormick TS. (2004). Fusobacterium nucleatum induces premature and term stillbirths in pregnant mice: Implication of oral bacteria in preterm birth. Infect Immun 72: 2272-2279.

206

Han YW, Shi W, Huang GT, Kinder Haake S, Park NH, Kuramitsu H, Genco RJ. (2000). Interactions between periodontal bacteria and human oral epithelial cells: Fusobacterium nucleatum adheres to and invades epithelial cells. Infect Immun 68: 3140-3146.

Han YW, Ikegami A, Rajanna C, Kawsar HI, Zhou Y, Li M, Sojar HT, Genco RJ, Kuramitsu HK, Deng CX. (2005). Identification and characterization of a novel adhesin unique to oral Fusobacteria. J Bacteriol 187: 5330-5340.

Han YW, Ikegami A, Chung P, Zhang L, Deng CX. (2007). Sonoporation is an efficient tool for intracellular fluorescent dextran delivery and one-step double-crossover mutant construction in Fusobacterium nucleatum. Appl Environ Microbiol 73:3677–3683.

Handley PS, Harty DW, Wyatt JE, Brown CR, Doran JP, Gibbs a C. (1987). A comparison of the adhesion, coaggregation and cell-surface hydrophobicity properties of fibrillar and fimbriate strains of Streptococcus salivarius. J Gen Microbiol 133:3207–3217.

Handley PS, Carter PL, Wyatt JE, Hesketh LM. (1985). Surface structures (peritrichous fibrils and tufts of fibrils) found on Streptococcus sanguis strains may be related to their ability to coaggregate with other oral genera. Infect Immun 47:217–227.

Hendrickson BA, Gokhale R, Cho JH. (2002). Clinical aspects and pathophysiology of inflammatory bowel disease. Clin Microbiol Rev 15: 79-94.

Herbeuval J-P, Lelievre E, Lambert C, Dy M, Genin C. (2004). Recruitment of STAT3 for production of IL-10 by colon carcinoma cells induced by macrophage-derived IL-6. J Immunol 172:4630–4636.

Hooper LV, Gordon JI. (2001). Commensal host-bacterial relationships in the gut. Science 292: 1115-1118.

Hope ME, Hold GL, Kain R, El-Omar EM. (2005). Sporadic colorectal cancer--role of the commensal microbiota. FEMS Microbiol Lett 244: 1-7.

Hughes RB, Smith AC. (2013). Capsule Stain Protocols. Protocols. American Society for Microbiology, Washington DC. doi: www.microbelibrary.org. Accessed 10 September 2015.

Humphrey SB, Stanton TB, Jensen NS, Zuerner RL. (1997). Purification and characterization of VSH-1, a generalized transducing bacteriophage of Serpulina hyodysenteriae. J Bacteriol 179:323–329.

Hunter CA, Jones SA. (2015). IL-6 as a keystone cytokine in health and disease. Nat Immunol 16:448–457.

Hyatt D, Chen GL, Locascio PF, Land ML, Larimer FW, Hauser LJ. (2010). Prodigal: prokaryotic gene recognition and translation initiation site identification. BMC Bioinformatics, 11: 119. doi:10.1186/1471-2105-11-119.

207

Ikegami A, Chung P, Han YW. (2009). Complementation of the fadA mutation in Fusobacterium nucleatum demonstrates that the surface-exposed adhesin promotes cellular invasion and placental colonization. Infect Immun 77: 3075-3079.

Ikushima H, Miyazono K. (2010). TGFbeta signalling: a complex web in cancer progression. Nat Rev Cancer 10:415–424.

Istivan TS, Smith SC, Fry BN, Coloe PJ. (2008). Characterization of Campylobacter concisus hemolysins. FEMS Immunol Med Microbiol 54:224–235.

Ito M, Kanno S, Nosho K, Sukawa Y, Mitsuhashi K, Kurihara H, Igarashi H, Takahashi T, Tachibana M, Takahashi H, Yoshii S, Takenouchi T, Hasegawa T, Okita K, Hirata K, Maruyama R, Suzuki H, Imai K, Yamamoto H, Shinomura Y. (2015). Association of Fusobacterium nucleatum with clinical and molecular features in colorectal serrated pathway. Int J Cancer 137 (6): 1258-1268.

Iyer VN, Klee HJ, Nester EW. (1982). Units of genetic expression in the virulence region of a plant tumor-inducing plasmid of Agrobacterium tumefaciens. Mol Gen Genet 188:418–424.

Jensen EC, Schrader HS, Rieland B, Thompson TL, Lee KW, Nickerson KW, Kokjohn TA. (1998). Prevalence of broad-host-range lytic bacteriophages of Sphaerotilus natans, Escherichia coli, and Pseudomonas aeruginosa. Appl Environ Microbiol 64:575–580.

Ji S, Shin JE, Kim YC, Choi Y. (2010). Intracellular degradation of Fusobacterium nucleatum in human gingival epithelial cells. Mol Cells 30: 519-526.

Kado S., Uchida K, Funabashi H, Iwata S, Nagata Y, Ando M, Onoue M, Matsuoka Y, Ohwaki M, Morotomi M. (2001). Intestinal microflora are necessary for development of spontaneous adenocarcinoma of the large intestine in T-cell receptor beta chain and p53 double- knockout mice. Cancer Res 61: 2395-2398.

Kagnoff MF, Eckmann L. (1997). Epithelial cells as sensors for microbial infection. J Clin Invest 100: 6-10.

Kalischuk LD, Inglis GD. (2011). Comparative genotypic and pathogenic examination of Campylobacter concisus isolates from diarrheic and non-diarrheic humans. BMC Microbiol 11:53.

Kamada N, Seo S-U, Chen GY, Núñez G. (2013). Role of the gut microbiota in immunity and inflammatory disease. Nat Rev Immunol 13:321–335.

Kapatral V, Ivanova N, Anderson I, Reznik G, Bhattacharyya A, Gardner WL, Mikhailova N, Lapidus A, Larsen N, D’Souza M, Walunas T, Haselkorn R, Overbeek R, Kyrpides N. (2003). Genome analysis of F. nucleatum sub spp vincentii and its comparison with the genome of F. nucleatum ATCC 25586. Genome Res 13:1180–1189.

208

Kapatral V, Anderson I, Ivanova N, Reznik G, Los T, Lykidis A, Bhattacharyya A, Bartman A, Gardner W. (2002). Genome sequence and analysis of the oral bacterium Fusobacterium nucleatum strain ATCC 25586. J Bacteriol 184: 2005-2018.

Kapalan CW, Lux R, Haake SK, Shi W. (2005). Reporter systems for Fusobacterium nucleatum, an abstract presented at the IADR/AADR/CADR 83rd general session. Baltimore, USA (http://iadr.confex.com/iadr/2005Balt/techprogram/abstract_60920.htm).

Kaplan CW, Lux R, Haake SK, Shi W. (2009). The Fusobacterium nucleatum outer membrane protein RadD is an arginine-inhibitable adhesin required for inter-species adherence and the structured architecture of multispecies biofilm. Mol Microbiol 71: 35-47.

Kaufhold A, Franzen D, Lütticken R. (1989). Endocarditis caused by Gemella haemolysans. Infection 17:385–387.

Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton S, Cooper A, Markowitz S, Duran C, Thierer T, Ashton B, Meintjes P, Drummond A. (2012). Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics 28:1647–1649.

Kelly D, Conway S. (2005). Bacterial modulation of mucosal innate immunity. Mol Immunol 42: 895-901.

Kelly D, Conway S, Aminov R. (2005). Commensal gut bacteria: Mechanisms of immune modulation. Trends Immunol 26: 326-333.

Kearse M, Moir R, Wilson A, Stones-Havas S, Cheung M, Sturrock S, Buxton S, Cooper A, Markowitz S, Duran C, Thierer T, Ashton B, Mentjies P, Drummond, A. (2012). Geneious Basic: an integrated and extendable desktop software platform for the organization and analysis of sequence data. Bioinformatics, 28: 1647-1649.

Kim MK, Park JC, Jeong MJ, Kim DK, Kook JK, Min BM, Lee SE, Choe SJ, Shin YK, Kim HS, Kim BO, Yoo SY, Seong JH. (2004). Multiplex PCR Using Conserved and Species- Specific 16S rDNA Primers for Simultaneous Detection of Fusobacterium nucleatum and Actinobacillus actinomycetemcomitans. J Micro Biotech. 14: 110-115.

Kitadai Y, Haruma K, Mukaida N, Ohmoto Y, Matsutani N, Yasui W, Yamamoto S, Sumii K, Kajiyama G, Fidler IJ, Tahara E. (2000). Regulation of disease-progression genes in human gastric carcinoma cells by interleukin 8. Clin Cancer Res 6:2735–2740.

Klampfer L. (2011). Cytokines, inflammation and colon cancer. Curr Cancer Drug Targets 11:451–464.

Kolenbrander PE, London J. (1993). Adhere today, here tomorrow: Oral bacterial adherence. J Bacteriol 175: 3247-3252.

209

Knüpfer H, Preiss R. (2010). Serum interleukin-6 levels in colorectal cancer patients- a summary of published results. Int J Colorectal Dis 25:135–140.

Kolenbrander PE, Andersen RN. (1989). Inhibition of coaggregation between Fusobacterium nucleatum and Porphyromonas (Bacteroides) gingivalis by lactose and related sugars. Infect Immun 57:3204–3209.

Kolenbrander PE, Andersen RN, Moore LV. (1989). Coaggregation of Fusobacterium nucleatum, Selenomonas flueggei, Selenomonas infelix, Selenomonas noxia, and Selenomonas sputigena with strains from 11 genera of oral bacteria. Infect Immun 57: 3194-3203.

Kolenbrander PE, Andersen RN, Blehert DS, Egland PG, Foster JS, Palmer RJ. (2002). Communication among oral bacteria. Microbiol Mol Biol Rev 66: 486-505.

Koren O, Spor A, Felin J, Fåk F, Stombaugh J, Tremaroli V, Behre CJ, Knight R, Fagerberg B, Ley RE, Bäckhed F. (2011). Human oral, gut, and plaque microbiota in patients with atherosclerosis. Proc Natl Acad Sci USA 108:4592–4598.

Kostic AD, Chun E, Robertson L, Glickman JN, Gallini CA, Michaud M, Clancy TE, Chung DC, Lochhead P, Hold GL, El-Omar EM, Brenner D, Fuchs CS, Meyerson M, Garrett WS. (2013). Fusobacterium nucleatum Potentiates Intestinal Tumorigenesis and Modulates the Tumor-Immune Microenvironment. Cell Host Microbe 14: 207–215.

Kostic AD, Gevers D, Pedamallu CS, Michaud M, Duke F, Earl AM, Ojesina AI, Jung J, Bass AJ, Tabernero J, Baselga J, Liu C, Shivdasani RA, Ogino S, Birren BW, Huttenhower C, Garrett WS, Meyerson M. (2012). Genomic analysis identifies association of Fusobacterium with colorectal carcinoma. Genome Res 22: 292–298.

Krogh A, Larsson B, von Heijne G, Sonnhammer EL. (2001). Predicting transmembrane protein topology with a hidden Markov model: application to complete genomes. J Mol Biol 305:567–580.

Kuai WX, Wang Q, Yang XZ, Zhao Y, Yu R, Tang XJ. (2012). Interleukin-8 associates with adhesion, migration, invasion and chemosensitivity of human gastric cancer cells. World J Gastroenterol 18:979–985.

Kurtz S, Phillippy A, Delcher AL, Smoot M, Shumway M, Antonescu C, Salzberg, SL. (2004). Versatile and open software for comparing large genomes. Gen Bio 5: R12. doi:10.1186/gb-2004-5-2-r12.

Kwaik YA, Gao L-Y, Stone BJ, Harb OS. (1998). Invasion of mammalian and protozoan cells by Legionella pneumophila. Bull Inst Pasteur 96:237–247.

210

Lagesen K, Hallin P, Rødland EA, Stærfeldt HH, Rognes T, Ussery DW. (2007). RNAmmer: Consistent and rapid annotation of ribosomal RNA genes. Nucleic Acids Res 35: 3100–3108. doi:10.1093/nar/gkm160.

Latorre E, Matheus N, Layunta E, Alcalde AI, Mesonero JE. (2014). IL-10 counteracts proinflammatory mediator evoked oxidative stress in Caco-2 cells. Mediators Inflamm 2014: 1-6. doi:10.1155/2014/982639.

Lawlor MS, Hsu J, Rick PD, Miller VL. (2005). Identification of Klebsiella pneumoniae virulence determinants using an intranasal infection model. Mol Microbiol 58:1054–1073.

Ledder RG, Timperley AS, Friswell MK, Macfarlane S, McBain AJ. (2008). Coaggregation between and among human intestinal and oral bacteria. FEMS Microbiol Ecol 66: 630-636.

Lee BH, Simard RE. (1984). Evaluation of methods for detecting the production of H2S, volatile sulfides, and greening by lactobacilli. J Food Sci 49:981–983.

Leonel AJ, Alvarez-Leite JI. (2012). Butyrate: implications for intestinal function. Curr Opin Clin Nutr Metab Care 15: 474–9.

Ley RE, Peterson DA, Gordon JI. (2006). Ecological and evolutionary forces shaping microbial diversity in the human intestine. Cell 124: 837-848.

Li A, Varney ML, Singh RK. (2001). Expression of interleukin 8 and its receptors in human colon carcinoma cells with different metastatic potentials. Clin Cancer Res 7:3298–3304.

Li X, Mai J, Virtue A, Yin Y, Gong R, Sha X, Gutchigian S, Frisch A, Hodge I, Jiang X, Wang H, Yang X-F. (2012). IL-35 is a novel responsive anti-inflammatory cytokine--a new system of categorizing anti-inflammatory cytokines. PLoS One 7:e33628.

Lowe TM, Eddy SR. (1997). tRNAscan-SE: A program for improved detection of transfer RNA genes in genomic sequence. Nucleic Acids Res 25: 955–964. doi:10.1093/nar/25.5.955.

Lui AC, McBride BC, Vovis GF., Smith M. (1979). Site specific endonuclease from Fusobacterium nucleatum. Nucleic Acids Res 6: 1-15.

Maaser C, Egan LJ, Birkenbach MP, Eckmann L, Kagnoff MF. (2004). Expression of Epstein-Barr virus-induced gene 3 and other interleukin-12-related molecules by human intestinal epithelium. Immunology 112:437–445.

Macfarlane S, Dillon JF. (2007). Microbial biofilms in the human gastrointestinal tract. J Appl Microbiol 102: 1187-1196.

Macpherson AJ, Harris NL. (2004). Interactions between commensal intestinal bacteria and the immune system. Nat Rev Immunol 4: 478-485.

211

Machuca P, Daille L, Vinés E, Berrocal L, Bittner, M. (2010). Isolation of a novel bacteriophage specific for the periodontal pathogen Fusobacterium nucleatum. Appl Environ Microbiol 76: 7243–7250. doi:10.1128/AEM.01135-10.

Macuch PJ, Tanner AC. (2000). Campylobacter species in health, gingivitis, and periodontitis. J Dent Res 79:785–792.

Madigan MT, Martinko JM, Stahl DA, Clark DP. (2012). Brock Biology of Microorganisms (13th Edition) Pearson Education Inc., Benjamin Cummings San Francisco CA.

Man SM, Kaakoush NO, Leach ST, Nahidi L, Lu HK, Norman J, Day AS, Zhang L, Mitchell HM. (2010). Host attachment, invasion, and stimulation of proinflammatory cytokines by Campylobacter concisus and other non-Campylobacter jejuni Campylobacter species. J Infect Dis 202:1855–1865.

Mangan PR, Harrington LE, O’Quinn DB, Helms WS, Bullard DC, Elson CO, Hatton RD, Wahl SM, Schoeb TR, Weaver CT. (2006). Transforming growth factor-beta induces development of the T(H)17 lineage. Nature 441:231–234.

Maniloff J, Cadden SP, Putzrath RM. (1981). Maturation of an enveloped budding phage: mycoplasmavirus L2. Prog Clin Bio Res 64: 503-513.

Marchesi JR. (2011). Human distal gut microbiome. Environ Microbiol 13: 3088-3102.

Marchler-Bauer A, Derbyshire MK, Gonzales NR, Lu S, Chitsaz F, Geer LY, Geer RC, He J, Gwadz M, Hurwitz DI, Lanczycki CJ, Lu F, Marchler GH, Song JS, Thanki N, Wang Z, Yamashita RA, Zhang D, Zheng C, Bryant SH. (2014). CDD: NCBI’s conserved domain database. Nucleic Acids Res 43:D222–D226.

Martin G, Keller W. (2007). RNA-specific ribonucleotidyl transferases. RNA 13:1834–1849.

Martinsen TC, Bergh K, Waldum HL. (2005). Gastric juice: A barrier against infectious diseases. Basic Clin Pharmacol Toxicol 96:94–102.

McCormack G, Moriarty D, O’Donoghue DP, McCormick PA, Sheahan K, Baird AW. (2001). Tissue cytokine and chemokine expression in inflammatory bowel disease. Inflamm Res 50:491–495.

McCoy AN, Araújo-Pérez F, Azcárate-Peril A, Yeh JJ, Sandler RS, Keku TO. (2013). Fusobacterium Is Associated with Colorectal Adenomas. PLoS One 8(1): e53653.

McDonald, J. (2013). Development of an In Vitro Fermentation Model to Culture the Human Distal Gut Microbiota. Unpublished doctoral dissertation. University of Guelph, Guelph.

212

McGuire AM, Cochrane K, Griggs AD, Haas BJ, Abeel T, Zeng Q, Nice JB, MacDonald H, Birren BW, Berger BW, Allen-Veroce E, Earl AM. (2014). Evolution of Invasion in a Diverse Set of Fusobacterium Species. Microbiome, 5: 1–11. doi:10.1128/mBio.01864-14.

McGuire A, Griggs A, DesJardins C, Godfrey P, Haas B, Wortmann J, Zucker E, Abeel J, Pearson M, Zheng Q, Strauss J, Allen-Vercoe E, Birren B, Earl A. (2012). Comparative analysis of Fusobacteria. Unpublished.

McLeod SM, Kimsey HH, Davis BM, Waldor MK. (2005). CTX and Vibrio cholerae: Exploring a newly recognized type of phage-host cell relationship. Mol Microbiol, 57: 347–356.

Meddings J. (2008). The significance of the gut barrier in disease. Gut 57: 438-440.

Meka RR, Venkatesha SH, Dudics S, Acharya B, Moudgil KD. (2015). IL-27-induced modulation of autoimmunity and its therapeutic potential. Autoimmun. Rev.14: 1131-1141. doi:org/10.1016/j.autrev.2015.08.001.

Merritt J, Niu G, Okinaga T, Qi F. (2009). Autoaggregation response of Fusobacterium nucleatum. Appl Environ Microbiol 75:7725–7733.

Miller JM. (1972). Experiments in Molecular Genetics. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

Miller AT, Byrn JC, Divino CM, Weber KJ. (2007). Eikenella corrodens causing necrotizing fasciitis after an elective inguinal hernia repair in an adult: a case report and literature review. Am Surg 73:876–879.

Mira A, Pushker R, Legault BA, Moreira D, Rodriguez-Valera F. (2004). Evolutionary relationships of Fusobacterium nucleatum based on phylogenetic analysis and comparative genomics. BMC Evol Biol 4: 50.

Moazzam AA, Rajagopal SM, Sedghizadeh PP, Zada G, Habibian M. (2015). Intracranial bacterial infections of oral origin. J Clin Neurosci. 22: 800-806.

Moore RA, Warren RL, Freeman JD, Gustavsen JA, Chénard C, Friedman JM, Suttle CA., Zhao Y, Holt RA. (2011). The sensitivity of massively parallel sequencing for detecting candidate infectious agents associated with human tissue. PLoS One 6:1–6.

Muller FM, Morschhauser J, Kohler G, Oelschlaeger TA, Ziebuhr W, Hacker J. (1999). Adherence and invasion-two pathogenicity factors in bacterial and fungal pathogens. Mycoses.42 Suppl 1: 39-42.

Murch SH, Braegger CP, Walker-Smith JA, MacDonald TT. (1993). Location of tumour necrosis factor alpha by immunohistochemistry in chronic inflammatory bowel disease. Gut 34: 1705-1709.

213

Murugaiyan G, Mittal A, Lopez-Diego R, Maier LM, Anderson DE, Weiner HL. (2009). IL- 27 is a key regulator of IL-10 and IL-17 production by human CD4+ T cells. J Immunol 183: 2435–2443.

Muyzer G, De Waal E. (1993). Profiling of complex microbial populations by denaturating gradient gel electrophoresis analysis of polymerase chain reaction-amplified genes coding for 16S rRNA. Appl Environ Microbiol 59: 695-700.

Nash RP, Niblock FC, Redinbo MR. (2011). Tyrosine partners coordinate DNA nicking by the Salmonella typhimurium plasmid pCU1 relaxase enzyme. FEBS Lett 585:1216–1222.

Naumann M, Rudel T, Meyer TF. (1999). Host cell interactions and signalling with Neisseria gonorrhoeae. Curr Opin Microbiol 2:62–70.

Nava GM, Stappenbeck TS. (2011). Diversity of the autochthonous colonic microbiota. Gut Microbes 2: 99-104.

Nelson KE, Weinstock GM, Highlander SK, Worley KC, Creasy HH, Wortman JR, Rusch DB, Mitreva M, Sodergren E, Chinwalla AT, Feldgarden M, Gevers D, Haas BJ, Madupu R, Ward D V, Birren BW, Gibbs RA, Methe B, Petrosino JF, Strausberg RL, Sutton GG, White OR, Wilson RK, Durkin S, Giglio MG, Gujja S, Howarth C, Kodira CD, Kyrpides N, Mehta T, Muzny DM, Pearson M, Pepin K, Pati A, Qin X, Yandava C, Zeng Q, Zhang L, Berlin AM, Chen L, Hepburn TA, Johnson J, McCorrison J, Miller J, Minx P, Nusbaum C, Russ C, Sykes SM, Tomlinson CM, Young S, Warren WC, Badger J, Crabtree J, Markowitz VM, Orvis J, Cree A, Ferriera S, Fulton LL, Fulton RS, Gillis M, Hemphill LD, Joshi V, Kovar C, Torralba M, Wetterstrand KA, Abouellleil A, Wollam AM, Buhay CJ, Ding Y, Dugan S, FitzGerald MG, Holder M, Hostetler J, Clifton SW, Allen-Vercoe E, Earl AM, Farmer CN, Liolios K, Surette MG, Xu Q, Pohl C, Wilczek-Boney K, Zhu D. (2010). A catalog of reference genomes from the human microbiome. Science 328:994–999.

Nelson WJ, Drees F, Yamada S. (2005). Interaction of cadherin with the actin cytoskeleton. Novartis Found Symp 269:159–177.

Ogata H, Goto S, Sato K, Fujibuchi W, Bono H, Kanehisa M. (1999). KEGG: Kyoto encyclopedia of genes and genomes. Nucleic Acids Research, 27: 29–34. doi:10.1093/nar/27.1.29.

Oh S-H, Cheng G, Nuessen J a, Jajko R, Yeater KM, Zhao X, Pujol C, Soll DR, Hoyer LL. (2005). Functional specificity of Candida albicans Als3p proteins and clade specificity of ALS3 alleles discriminated by the number of copies of the tandem repeat sequence in the central domain. Microbiology 151:673–81.

O'Hara AM, Shanahan F. (2006). The gut flora as a forgotten organ. EMBO Rep 7:688-693.

214

Ohkusa T, Okayasu I, Ogihara T, Morita K, Ogawa M, Sato N. (2003). Induction of experimental ulcerative colitis by Fusobacterium varium isolated from colonic mucosa of patients with ulcerative colitis. Gut 52: 79-83.

Okuda T, Kokubu E, Kawana T, Saito A, Okuda K, Ishihara K. (2012). Synergy in biofilm formation between Fusobacterium nucleatum and Prevotella species. Anaerobe 18:110–116.

Osborn AM, Böltner D. (2002). When phage, plasmids, and transposons collide: Genomic islands, and conjugative- and mobilizable-transposons as a mosaic continuum. Plasmid 48:202– 212.

O’Toole G, Kaplan HB, Kolter R. (2000). Biofilm formation as microbial development. Annu Rev Microbiol 54:49–79.

Ouyang W, Kolls JK, Zheng Y. (2008). The Biological Functions of T Helper 17 Cell Effector Cytokines in Inflammation. Immunity 28:454–467.

Ouyang W, Rutz S, Crellin NK, Valdez PA, Hymowitz SG. (2011). Regulation and functions of the IL-10 family of cytokines in inflammation and disease. Annu Rev Immunol 29:71–109.

Palmer C, Bik EM, DiGiulio DB, Relman DA, Brown PO. (2007). Development of the human infant intestinal microbiota. PLoS Biol 5: e177.

Parkin DM. (2006). The global health burden of infection-associated cancers in the year 2002. Int J Can 118: 3030-3044.

Persson GR, Imfeld T. (2008). Periodontitis and cardiovascular disease. Ther Umsch. 65: 121- 126.

Petersen TN, Brunak S, von Heijne G, Nielsen H. 2011. SignalP 4.0: discriminating signal peptides from transmembrane regions. Nat Methods 8:785–786.

Pizarro-Cerdá J, Kühbacher A, Cossart P. (2012). Entry of listeria monocytogenes in mammalian epithelial cells: An updated view. Cold Spring Harb Perspect Med 2:1–18.

Podnar J, Deiderick H, Huerta G, Hunicke-Smith S. (2014). Next-generation sequencing RNA-Seq library construction. Curr Prot Mol Bio. 106: 4211-4419.

Polak D, Wilensky A, Shapira L, Halabi A, Goldstein D, Weiss EI, Houri-Haddad Y. (2009). Mouse model of experimental periodontitis induced by Porphyromonas gingivalis Fusobacterium nucleatum infection: Bone loss and host response. J Clin Periodontol 36:406– 410.

Preus HR, Olsen I, Gjermo P. (1987). Bacteriophage infection--a possible mechanism for increased virulence of bacteria associated with rapidly destructive periodontitis. Acta Odontol Scand 45:49–54.

215

Qin N, Yang F, Li A, Prifti E, Chen Y, Shao L, Guo J, Le Chatelier E, Yao J, Wu L, Zhou J, Ni S, Liu L, Pons N, Batto JM, Kennedy SP, Leonard P, Yuan C, Ding W, Chen Y, Hu X, Zheng B, Qian G, Xu W, Ehrlich SD, Zheng S, Li L. (2014). Alterations of the human gut microbiome in liver cirrhosis. Nature 513:59–64.

Resch G, Kulik EM, Dietrich FS, Meyer J. (2004). Complete genomic nucleotide sequence of the temperate bacteriophage Aa Phi 23 of Actinobacillus actinomycetemcomitans. J Bacteriol 186:5523–5528.

Rickard AH, Gilbert P, High NJ, Kolenbrander PE, Handley PS. (2003). Bacterial coaggregation: An integral process in the development of multi-species biofilms. Trends Microbiol 11: 94-100.

Rowland IR. (2009). The role of the gastrointestinal microbiota in colorectal cancer. Curr Pharm Des 15: 1524-1527.

Robinson JT, Thorvaldsdóttir H, Winckler W, Guttman M, Lander ES, Getz G, Mesirov JP. (2011). Integrative genomics viewer. Nat Biotechnol 29:24–26.

Rubinstein MR, Wang X, Liu W, Hao Y, Cai G, Han YW. (2013). Fusobacterium nucleatum Promotes Colorectal Carcinogenesis by Modulating E-Cadherin/β-Catenin Signaling via its FadA Adhesin. Cell Host Microbe 14: 195–206.

Saitou N, Nei M. (1987). The neighbor-joining method: a new method for reconstructing phylogenetic trees. Mol Biol Evol 4: 406–425.

Saito A, Inagaki S, Kimizuka R, Okuda K, Hosaka Y, Nakagawa T, Ishihara K. (2008). Fusobacterium nucleatum enhances invasion of human gingival epithelial and aortic endothelial cells by Porphyromonas gingivalis. FEMS Immunol Med Microbiol 54:349–55.

Sambrook J, Fritsch EF, Maniatis T. (1989). Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, N.Y.

Sambrook J, Russel DW. (2001) Molecular cloning: A laboratory manual. Cold Spring Harbour Laboratory Press, Cold Spring Harbour, N.Y.

Sambuy Y, De Angelis I, Ranaldi G, Scarino ML, Stammati A, Zucco F. (2005). The caco-2 cell line as a model of the intestinal barrier: Influence of cell and culture-related factors on caco- 2 cell functional characteristics. Cell Biol Toxicol 21: 1-26.

Sano E, Carlson S, Wegley L, Rohwer F. (2004). Movement of viruses between biomes. Appl Environ Microbiol 70:5842–5846.

Sansone P, Bromberg J. (2011). Environment, inflammation, and cancer. Curr Opin Genet Dev 21: 80–85.

216

Sasaoka T, Ito M, Yamashita J, Nakajima K, Tanaka I, Narita M, Hara Y, Hada K, Takahashi M, Ohno Y, Matsuo T, Kaneshiro Y, Tanaka H, Kaneko K. (2011). Treatment with IL-27 attenuates experimental colitis through the suppression of the development of IL-17- producing T helper cells. Am J Physiol Gastrointest Liver Physiol 300:G568–G576.

Sautès-Fridman C, Cherfils-Vicini J, Damotte D, Fisson S, Fridman WH, Cremer I, Dieu- Nosjean MC. (2011). Tumor microenvironment is multifaceted. Cancer Metastasis Rev 30:13– 25.

Schafer ZT, Brugge JS. (2007). IL-6 involvement in epithelial cancers. J Clin Invest 117:3660– 3663.

Schilling JD, Mulvey MA, Vincent CD, Lorenz RG, Hultgren SJ. (2001). Bacterial invasion augments epithelial cytokine responses to Escherichia coli through a lipopolysaccharide- dependent mechanism. J Immunol 166:1148–1155.

Schroeter, K. (2014). Adaptation of a Single-Stage Chemostat System to Model Functional Aspects of the Human Distal Gut Microbiota. Unpublished doctoral dissertation. University of Guelph, Guelph.

Sears CL, Pardoll DM. (2011). Perspective: Alpha-bugs, their microbial partners, and the link to colon cancer. J Infect Dis 203: 306-311.

Sekelja M, Berget I, Naes T, Rudi K. (2011). Unveiling an abundant core microbiota in the human adult colon by a phylogroup-independent searching approach. ISME J 5: 519-531.

Sellon RK, Tonkonogy S, Schultz M, Dieleman LA, Grenther W, Balish E, Rennick DM, Sartor RB. (1998). Resident enteric bacteria are necessary for development of spontaneous colitis and immune system activation in interleukin-10-deficient mice. Infect Immun 66: 5224- 5231.

Shanahan F. (2002). The host-microbe interface within the gut. Best Pract Res Clin Gastroenterol 16: 915-931.

Siguier P, Perochon J, Lestrade L, Mahillon J, Chandler M. (2006). ISfinder: the reference centre for bacterial insertion sequences. Nucleic Acids Res 34:D32–D36.

Silva VL, Diniz CG, Cara DC, Santos SG, Nicoli JR, Carvalho MA, Farias LM. (2005). Enhanced pathogenicity of Fusobacterium nucleatum adapted to oxidative stress. Microb Pathog 39: 131-138.

Skerman, VBD, McGowan V, Sneath PHA. (1980). Approved lists of bacterial names. Int J Syst Bacteriol 30: 225-420.

Smedley JG, Fisher DJ, Sayeed S, Chakrabarti G, McClane BA. (2004). The enteric toxins of Clostridium perfringens. Rev Physiol Biochem Pharmacol 152:183–204.

217

Stathopoulou PG, Benakanakere MR, Galicia JC, Kinane DF. (2010). Epithelial cell pro- inflammatory cytokine response differs across dental plaque bacterial species. J Clin Periodontol 37:24–29.

Strauss J, White A, Ambrose C, McDonald J, Allen-Vercoe E. (2008). Phenotypic and genotypic analyses of clinical Fusobacterium nucleatum and Fusobacterium periodonticum isolates from the human gut. Anaerobe 14: 301-309.

Strauss J, Kaplan GG, Beck PL, Rioux K, Panaccione R, Devinney R, Lynch T, Allen- Vercoe E. (2011). Invasive potential of gut mucosa-derived Fusobacterium nucleatum positively correlates with IBD status of the host. Inflamm Bowel Dis 17: 1971-1978.

Strauss J. (2011). Investigating a role for Fusobacterium nucleatum in Inflammatory Bowel Disease. Unpublished doctoral dissertation. University of Guelph, Guelph.

Summer EJ, Gonzalez CF, Carlisle T, Mebane LM, Cass AM, Savva CG, Young R. (2004). Burkholderia cenocepacia phage BcepMu and a family of Mu-like phages encoding potential pathogenesis factors. J Mol Bio 340: 49–65.

Suzuki T, Yoshinaga N, Tanabe S. (2011). Interleukin-6 (IL-6) regulates claudin-2 expression and tight junction permeability in intestinal epithelium. J Biol Chem 286:31263–31271.

Swidsinski A, Khilkin M, Kerjaschki D, Schreiber S, Ortner M, Weber J, Lochs H. (1998). Association between intraepithelial Escherichia coli and colorectal cancer. Gastroenterology 115: 281-286.

Swidsinski A, Ladhoff A, Pernthaler A, Swidsinski S, Loening-Baucke V, Ortner M, Weber J, Hoffmann U, Schreiber S, Dietel M, Lochs H. (2002). Mucosal flora in inflammatory bowel disease. Gastroenterology 122: 44-54.

Tahara E. (2000). Molecular aspects of invasion and metastasis of stomach cancer. Verh Dtsch Ges Pathol 84:43–49.

Tahara T, Yamamoto E, Suzuki H, Maruyama R, Chung W, Garriga J, Jelinek J, Yamano HO, Sugai T, An B, Shureiqi I, Toyota M, Kondo Y, Estecio MRH, Issa JPJ. (2014). Fusobacterium in colonic flora and molecular features of colorectal carcinoma. Cancer Res 74: 1311–1318.

Tanner AC, Badger S, Lai C-H, Listgarten MA, Visconti RA, Socransky SS. (1981). Wolinella gen. nov., Wolinella succinogenes (Vibrio succinogenes) comb. nov., and description of Bacteroides gracilis sp. nov., Wolinella recta sp. nov., Campylobacter concisus sp. nov., and Eikenella corrodens from humans with periodontal disease. Int J Syst Bacteriol 31:432–445.

218

Tannock G, Munro K. (2000). Analysis of the fecal microflora of human subjects consuming a probiotic product containing Lactobacillus rhamnosus DR20. Appl Environ Microbiol 66:2578– 2588.

Tatusov RL, Koonin EV, Lipman DJ. (1997). A genomic perspective on protein families. Science 278: 631–637. doi:10.1126/science.278.5338.631.

Tay AP, Kaakoush NO, Deshpande NP, Chen Z, Mitchell H, Wilkins MR. (2013). Genome Sequence of Campylobacter showae UNSWCD, Isolated from a Patient with Crohn’s Disease. Genome Announcements 1: e00193–12.

Tian W, Arakaki AK, Skolnick J. (2004). EFICAz: A comprehensive approach for accurate genome-scale enzyme function inference. Nucleic Acids Res 32: 6226–6239. doi:10.1093/nar/gkh956.

Timme T, Brennan P. (1984). Induction of bacteriophage from members of the Mycobacterium avium, Mycobacterium intracellulare, Mycobacterium scrofulaceum Serocomplex. J Gen Microbiol 130: 2059–2066.

Thompson JD, Higgins GD, & Gibson TJ. (1994). Clustal w: improving the sensitivity of progressive multiple sequence alignment through sequence weighting, position-specific gap penalties and weight matrix choice. Nucleic Acids Res 22: 4673–4680.

Thompson-Chagoyan OC, Maldonado J, Gil A. (2005). Aetiology of inflammatory bowel disease (IBD): Role of intestinal microbiota and gut-associated lymphoid tissue immune response. Clin Nutr 24: 339-352.

Thorvaldsdóttir H, Robinson JT, Mesirov JP. (2013). Integrative Genomics Viewer (IGV): High-performance genomics data visualization and exploration. Brief Bioinform 14:178–192.

Thurnheer T, Guggenheim B, Gruica B, Gmür R. (1999). Infinite Serovar and Ribotype Heterogeneity Among Oral Fusobacterium nucleatum Strains? Anaerobe 5:79–92.

Verstrepen KJ, Jansen A, Lewitter F, Fink GR. (2005). Intragenic tandem repeats generate functional variability. Nat Genet 37:986–990.

Vignali DAA, Kuchroo VK. 2012. IL-12 family cytokines: immunological playmakers. Nat Immunol 13:722–728.

Vogelstein B, Fearon ER, Hamilton SR, Kern SE, Preisinger AC, Leppert M, Nakamura Y, White R, Smits AM, Bos JL. (1988). Genetic alterations during colorectal-tumor development. N Engl J Med 319: 525-532.

Wagner PL, Waldor MK. (2002). Bacteriophage control of bacterial virulence. Infect Immun 70(8): 3985-3993.

219

Wang L, Christophersen CT, Sorich MJ, Gerber JP, Angley MT, Conlon MA. (2013). Increased abundance of Sutterella spp. and Ruminococcus torques in feces of children with autism spectrum disorder. Mol Autism 4:42.

Wang Z, Gerstein M, Snyder M. (2009). RNA-Seq: a revolutionary tool for transcriptomics. Nat Rev Genet 10:57–63.

Wang IN, Smith DL, Young, R. (2000). Holins: the protein clocks of bacteriophage infections. Ann Rev Microbiol 54: 799–825.

Wapinski I, Pfeffer A, Friedman N, Regev A. (2007). Automatic genome-wide reconstruction of phylogenetic gene trees. Bioinformatics 23:1–9.

Warren RL, Freeman DJ, Pleasance S, Watson P, Moore R a, Cochrane K, Allen-Vercoe E, Holt RA. (2013). Co-occurrence of anaerobic bacteria in colorectal carcinomas. Microbiome 1:16.

Washington JA, Warren E, Karlson AG. (1972). Stability of Barium Sulfate Turbidity Standards. Appl Microbiol 24:1013.

Watson PH. (2010). Canadian Tumour Repository Network. Biopreserv Biobank 8: 181–185.

White JR, Nagarajan N, Pop M. (2009). Statistical methods for detecting differentially abundant features in clinical metagenomic samples. PLoS Comput Biol 5:e1000352.

Wu T, Cen L, Kaplan C, Zhou X, Lux R, Shi W, He X. (2015). Cellular components mediating coadherance of Candida albicans and Fusobacterium nucleatum. J Dent Res. 94: 1432-1438.

Wynick C, Petes C, Gee K. (2014). Interleukin-27 mediates inflammation during chronic disease. J Interferon Cytokine Res 34:741–749.

Xie J, Wu H, Dai C, Pan Q, Ding Z, Hu D, Ji B, Luo Y, Hun X. (2014). Beyond Warburg effect- dual metabolic nature of cancer cells. Sci Rep. 4:4927. Doi:10.1038/srep04927

Xu M, Yamada M, Li M, Liu H, Chen SG, Han YW. (2007). FadA from Fusobacterium nucleatum utilizes both secreted and nonsecreted forms for functional oligomerization for attachment and invasion of host cells. J Biol Chem 282: 25000-25009.

Yang L, Pei Z. (2006). Bacteria, inflammation, and colon cancer. World J Gastroenterol 12: 6741-6746.

Yumoto H, Nakae H, Fujinaka K, Ebisu S, Matsuo T. (1999). Interleukin-6 (IL-6) and IL-8 Are Induced in Human Oral Epithelial Cells in Response to Exposure to Periodontopathic Eikenella corrodens. Infect Immun 67:384–394.

220

Zerbino DR, Birney E. (2008). Velvet: Algorithms for de novo short read assembly using de Bruijn graphs. Genome Res 18: 821–829.

Zhang G, Chen R, Rudney JD. (2008). Streptococcus cristatus attenuates Fusobacterium nucleatum-induced interleukin-8 expression in oral epithelial cells. J Perio Res 43:408–416.

Zhang G, Chen R, Rudney JD. (2011). Streptococcus cristatus modulates the Fusobacterium nucleatum-induced epithelial interleukin-8 response through the nuclear factor-kappa B pathway. J Perio Res 46:558–567.

Zhang JM, An J. (2007). Cytokines, inflammation, and pain. Int Anesthesiol Clin 45:27–37.

Zhang L, Si MM, Day AS, Leach ST, Lemberg DA, Dutt S, Stormon M, Otley A, O’Loughlin EV, Magoffin A, Ng PHY, Mitchell H. (2009). Detection and isolation of Campylobacter species other than C. jejuni from children with Crohn’s disease. J Clin Microbiol 47:453–455.

Zhou Y, Liang Y, Lynch KH, Dennis JJ, Wishart DS. (2011). PHAST: A Fast Phage Search Tool. Nucleic Acids Res 39:347–352.

Zilm PS, Rogers AH. (2007). Co-adhesion and biofilm formation by Fusobacterium nucleatum in response to growth pH. Anaerobe 13:146–52.

221

Appendix I- Method Development for Fn transcriptome studies (RNA-Seq)

A substantial portion of the cancer burden worldwide is attributable to infectious agents such as viruses or bacteria (Parkin, 2006). These microbial agents can directly cause cancers, others can facilitate cancer development, and others may have no causative role but their existence can indicate the presence of a cancer or high risk of developing one. CRC is one of the most prevalent and deadly of all cancers, due in part to the tendency toward diagnosis at a late stage (DeSantis et al., 2014). Fn has been recently found to be highly elevated in a subset of colorectal cancers (Castellarin et al., 2012; Kostic et al., 2012). In addition to the elevation of Fn gene sequences in CRC tumours, Fn is a reasonable candidate for having a facilitative role in tumorigenesis as indicated by the following evidence: Fn can act as a pro-inflammatory agent

(Allen-Vercoe et al., 2011); Fn can be an invasive bacterium that can cause acute oral and gastrointestinal infections (Strauss et al., 2011); the ability of the Fn FadA adhesin to adhere, invade and induce oncogenesis in host cells has been documented (Rubinstein et al., 2013); and

Fn has been shown to modulate the tumour-immune microenvironment (Kostic et al., 2013).

However, Fn is also a well recognized benign resident of mucosal surfaces, particularly the oral cavity, in the absence of pathogenicity (Allen-Vercoe et al., 2011). The reason why Fn may in some cases be pathogenic and at other times an apparently benign, commensal organism is not yet completely understood.

Comparative analysis of our sequenced genomes of Fn strains have revealed diverse and highly variable genomes comprising genes for many hypothetical proteins with no known homologs (McGuire et al., 2014). FadA expression has been detected in patient tumour tissue

222

(Rubinstein et al., 2013) and thus it is plausible that, in response to host cues, Fn upregulates expression of proteins such as FadA to enhance its ability to persist in the gut niche. Further characterization of Fn biology, particularly the aspects involved in virulence, have been slowed by the largely genetically intractable nature of Fn genomes. As such, it is difficult to engineer mutations and genetically characterize important phenotypes via knockout mutations. However, assessing transcriptome profiles for Fn during stages of infection within a host would provide information on novel genes involved in the process. The transcriptome includes the complete set of all RNA molecules in a cell and their quantity for a specific developmental state or physiological condition at a given moment (Wang et al., 2009). Understanding the Fn transcriptome during infection would allow for the interpretation of fundamental elements of the

Fn genome and coincidently would allow for the deeper understanding of its role in inflammatory diseases.

The increasing availability of low-cost sequencing has meant that high throughput sequencing of cDNA libraries generated from mRNA transcripts (RNA-Seq) is emerging as a powerful approach for mapping transcriptomes and profiling gene expression in different bacterial species. We decided to use RNA-Seq together with our established Fn/Caco-2 cell infection model to assess the transcriptional profiles of Fn 7-1 during active invasion. The Fn 7-1 strain, although not isolated from a CRC patient biopsy, was chosen because it has undergone whole-genome sequencing (to facilitate transcript mapping), is our most adhesive and invasive strain (to allow for sufficient yields of internalized bacterial RNA), has been extensively characterized in our laboratory and has recently been shown to induce tumorigenesis in germ- free mice (Kostic et al., 2013).

223

So far we have only a rudimentary idea of Fn gene expression during infection. Therefore the overall goal of this study was to identify genes associated with Fn virulence, and to determine how expression levels of these genes were modulated during infection. We hypothesized that Fn virulence is directly linked to its invasive capacity, and that invasive Fn isolates require select sets of genes that endow these strains with this (and perhaps other) pathogenic properties, setting them apart from commensal isolates. Our alternative hypothesis is that there is no difference in the virulence gene repertoires of commensal and invasive strains, but genes associated with pathogenicity are upregulated under certain circumstances to promote invasion and tumorigenesis.

RNA extraction from Fn is not straightforward. Fn is a Gram negative bacterial species with an outer membrane and as such, lends itself to ineffective lysis during the initial step of

RNA extraction. Fn cells can also release RNA inhibitors, such as RNAses, that are capable of degrading the RNA before it can be properly preserved. In addition, extraction of RNA from Fn cells inside of the Caco-2 cells increases the difficulty and time needed to isolate the bacterial

RNA. This also increases contamination opportunities leading to RNA degradation.

In order to proceed with RNA-Seq analysis, various methods for the isolation of high quality mRNA from bacterial Fn, as well as from the host epithelial cells were studied in order to find the most efficient method for this critical step of the RNA-Seq process. The finalized method developed for mRNA extraction is discussed below.

224

Materials and Methods

A. Isolating total RNA from invaded Fn 7-1 bacterial cells

Adherence and Invasion Caco-2-cell model

Caco-2 cells were grown to 85% confluence in 150cm2 tissue culture flasks (as described in Section 2.1.3 A). Fn 7-1 was grown to stationary phase in TSBsupp and added at a MOI of

100:1 to the Caco-2 cells (Figure 7 A). The flasks were then incubated at 37°C in 5%CO2 for either 4 hrs, 8hrs, 12hrs or 24 hrs. After the incubation, cells were washed with sterile PBS (+0.1 g/L CaCl2) and kill media was added (SFM + 0.5mg/mL gentamicin) to kill any bacterial cells that had not invaded the Caco-2 cells (Figure 7 B). The cells were incubated with the kill media for 30 min and subsequently washed twice with PBS (Figure 7 C). Trypsin (Sigma-Aldrich;

Oakville, ON) was then added to the flasks and incubated for 15 min at 37°C to re-suspend

Caco-2 cells adherent to the flask wall. The cells were then quenched with Caco-2 cell media

(Appendix I) and transferred to 50 mL conicals where they were centrifuged at 4000 rpm for 2 min at 4°C and washed with PBS (Figure 7 D). The Caco-2 cells were then lysed with 0.1%

Nonidet P-40 in PBS (Sigma-Aldrich; Oakville, ON), to release interior Fn 7-1 cells, vortexed and placed on a LabQuakeTM Tube Rotator (ThermoFisher; Burlington, ON) for 5 minutes at 4°C

(Figure 7 E). Cells were then pelleted by centrifugation (Figure 7 F) and washed with 1X

Versene (Appendix I) (Figure 7 G). Finally, to isolate total RNA from the Caco-2 and Fn 7-1 bacterial cells, TRIzol® Reagent (Ambion; ThermoFisher Scientific; Burlington, ON) was added at 1mL per 50-100mg of pelleted sample, homogenized for 5 min at RT and then frozen at -80°C until RNA extraction was performed on the samples (Figure 7 H).

225

The following controls were also homogenized with TRIzol® reagent to isolate total

RNA during the invasion assays (Figure 7; Figure 8): C1) Fn 7-1 in TSBsupp; C2) Caco-2 cells, without the addition of bacteria; C3) non-adhered Fn 7-1; C4) adherent Fn7-1 cells at the 4hr time point, collected after trypsinization; C5) lysed Caco-2 cells; and C6) Fn 7-1 in Caco-2 cell media after 8 hr incubation. For the 12 hr and 24 hr incubation steps, washing the Caco-2 cells before trypsinization was not possible as the tissue culture cells were already lifted off the flasks and had lost adherence capacity. Instead, all Caco-2 cells and Fn 7-1 cells were collected and subjected to centrifugation, washing with 1X Versene and mixed with TRIzol® reagent. The protocol for RNA isolation of all invasion samples and the respective controls were repeated with three biological replicates.

RNA extraction, DNase treatment and analysis of RNA quality

RNA was separated from DNA and cellular debris by adding 0.2 mL of chloroform per 1 mL of TRIzol® reagent, used in Section 2.5.1A for homogenization of the samples. Each sample was then vortexed vigorously for 15 s and then incubated for 3 min at RT before being centrifuged at 12,000 xg for 15 min at 4°C. RNA remained exclusively in the aqueous phase of

TRIzol® preps (the colourless upper phase). This portion was removed carefully, avoiding drawing any of the interphase or organic layer, into fresh RNase-free Eppendorf tubes

(ThermoFisher; Burlington, ON). The RNA was then precipitated by adding 100% isopropanol, incubated at RT for 15 min and pelleted by centrifugation at 12,000 xg for 10 min at 4°C. The supernatant was carefully removed, leaving behind the RNA pellet, which was washed three times with 75% EtOH, vortexed gently and re-pelleted by centrifugation at 7500 xg for 5 min at

4°C. Following the wash steps, RNA was air dried for 5 min, ensuring the pellet did not dry completely, and then resuspended in 50 µL nuclease-free, de-ionized H20. Each sample was then

226 subjected to a DNase treatment using the Turbo DNA-freeTM kit (ThermoFisher Scientific;

Burlington, ON) to remove any residual DNA, following the manufacturer’s instructions. All samples (including controls) were then sent to be analysed for RNA quality at the AAC

Genomics Facility at the University of Guelph using the Agilent 2100 Bioanalyzer, before continuing onto to the final mRNA purification step (Figure 2.2 I&J).

B. Ribodepletion of total RNA samples mRNA purification

To remove ribosomal RNA from the total RNA samples, prior to RNAseq analysis, the purification of bacterial mRNA was accomplished using the MICROBexpressTM kit

(ThermoFisher Scientific; Burlington, ON). This kit uses capture oligonucleotides to bind bacterial 16S and 23S rRNA subunits and subsequently remove the rRNA hybrids using derivitized magnetic beads. All reagents were provided with the kit and the manufacturer’s instructions were used for this procedure. Briefly, up to 10µg of total RNA was precipitated and resuspended in binding buffer containing the capture oligonucleotides. The RNA mixture was then heated to 70°C for 10 min to denature any secondary structures in the RNA, including the

16S and 23S rRNA subunits. The mixture was then transferred to an annealing temperature of

37°C for 15 min, which allowed the oligonucleotides to hybridize to homologous regions of the rRNA. Next, magnetic beads were added to the RNA mixture and incubated at 37°C for another

15 min. At this step the magnetic beads anneal to the capture oligonucleotides. After the incubation the magnetic beads and bound rRNA subunits were removed using a magnetic stand.

The supernatant, which contained the enriched mRNA, was then precipitated and resuspended in nuclease-free, de-ionized H2O before being frozen and stored at -80°C. A small aliquot was removed prior to freezing to be sent for RNA quality analysis on the Agilent 2100 bioanalyzer.

227

Figure AIa. Schematic of Fn 7-1 RNA isolation from infected Caco-2 tissue cells

228

229

230

Results

A. Isolation and purification of high quality mRNA from Fn 7-1 and Caco-2 cells from an adhesion/invasion model

Isolation of total RNA from bacterial and eukaryotic cells from a Fn 7-1 invasion model in Caco-

2 tissue cells

Using the adherence and invasion Caco-2 epithelial cell model, total RNA (tRNA) was extracted using TRIzol®. For each of the invasion and control parameters, described in detail in

Section 2.5.1 A, RNA was extracted from three separate biological replicates and were assessed independently for quality. Quality was assessed using the Agilent 2100 bioanalyzer. Samples deemed to be of a high RNA quality, based on the Agilent tRNA assay (i.e. a high amount of prokaryotic and eukaryotic ribosomal RNA (rRNA) and undegraded RNA samples), underwent ribodepletion. Unfortunately, RNA integrity numbers (RIN) could not be used to accurately assess the quality of the RNA in this eukaryotic/prokaryotic cell interaction study, as the RIN algorithm is unable to differentiate eukaryotic/prokaryotic rRNA. This creates serious quality index underestimations.

Ribodepletion of total bacterial RNA to generate high quality bacterial mRNA

All samples were purified using the bacterial messenger RNA (mRNA) kit,

MICROBexpressTM (ThermoFisher Scientific; Burlington, ON). This kit successfully removed the bacterial rRNA from the samples using derivitized magnetic beads. The removal of bacterial rRNA and the final quality of the mRNA was assessed once again using the Agilent 2100 bioanalyzer, this time using the mRNA assay. Again, RIN numbers could not be used to assess quality and so, the absence of the 16S and 23S rRNA bands on the resulting chromotagrams were used, along with the gel electrophoresis analysis which assessed the degradation of the

231 remaining mRNA product (Figure 7.1). During RNA-Seq analysis, an algorithm will be used to separate the eukaryotic and prokaryotic RNA reads, as well as to remove the eukaryotic rRNA reads. An algorithm does not yet exist for the systematic downstream genomic removal of bacterial rRNA reads, so it was done manually, ahead of time, using the MICROBexpressTM kit.

232

Figure AIb. Representative chromotographs, data and gel electrophoresis images for A) a degraded invaded Fn total RNA sample; B) a high quality invaded Fn total RNA sample; and C) a high quality invaded Fn messenger RNA sample (post rRNA removal).

233

Discussion and Future Directions

In order to determine the gene expression of the highly invasive Fn 7-1 during the infection process by RNA-Seq analysis, high quality mRNA was isolated and purified from both

Fn 7-1 bacterial and Caco-2 epithelial cells. Due to the unpredictable and labile nature of RNA, and the fact that high quality RNA was needed from both the Fn prokaryotic and Caco-2 eukaryotic cells, RNA extraction protocol development was an elaborate endeavour and involved many troubleshooting and optimization steps. Various RNA extraction and purification protocols and kits (Trizol®, Maxwell®, Qiagen RNAeasy®, Ambion Purelink®) were tested to determine the most efficient way to isolate high quality total RNA from both the adhesive and internalized

Fn cells and from the host cells. It was determined that using Trizol® (phenol), a chloroform/isopropanol extraction and ethanol washes produced the highest yield of good quality, high integrity, non-degraded RNA from both the prokaryotic and eukaryotic cells. In addition due to the lack of a downstream bioinformatic algorithm available for in silico removal of bacterial rRNA, the bacterial ribodepletion was accomplished using MICROBexpressTM. This meant additional opportunities for RNA degradation and contamination from environmental

RNAses. Despite these hurdles and technical conditions, RNA was successfully isolated from all adhesion/invasion time points and it was of a high enough integrity to proceed with RNA-Seq analysis.

There is one caveat to this method. Since bacteria are capable of rapid transcriptional responses to the environment, the feasibility of this approach depends upon the ability to recover biologically relevant bacterial RNA. Therefore, serious consideration was given to prevent gene expression changes associated with preparative procedures. These considerations included

234 working quickly, avoiding contamination, keeping all samples and reagents ice cold and including RNAlater® (ThermoFisher Scientific; Burlington, ON) in all of my wash and lysis buffers.

This study is currently still underway. The RNA samples are currently undergoing cDNA library generation. The next steps in the RNA-Seq pipeline involve the construction of strand- specific (ss)RNA-Seq libraries and subsequent molecular barcoding of the libraries for multiplexed sequencing on the Illumina platform. ssRNA-Seq provides state-of-the-art methodology to discriminate between genes that overlap or are encoded on opposite strands

(Podnar et al., 2014).There is well-established methodology for RNA-Seq at the BCCA Genome

Sciences Center and our collaborators in the Dr. Holt laboratory have sequenced over 9,000

RNA-Seq libraries to date, including the RNA-Seq studies that led to the initial identification of the association of Fn and colorectal cancer (Castellarin et al., 2012; Warren et al., 2013). For the present study, barcoded cDNA libraries will also be generated from each isolate grown in

TSBsupp broth culture to mid-log phase as control (no infection) libraries. Differential expression will be determined by comparing normalized read counts obtained from alignment of the RNA-

Seq output to the Fn 7-1 genome, and also by de novo transcript assembly, which will increase sensitivity for determining the structure and relative abundance of novel transcripts that may be present in Fn 7-1.

Although our Caco-2 model is well developed, standardized and suitable for the needs of this study, it can be argued that it is not representative of in vivo infections and that an animal model should be used instead. Therefore, future directions of a successful project will include development of a suitable animal model and its use to assess the potential to develop neutralizing antibodies against identified virulence determinants in vaccination strategies. Targeted

235 vaccination against specific virulence genes, as opposed to a general anti-Fn vaccine, will avoid the potential problem of vaccine inefficiency caused by tolerization to commensal Fn.

236

Appendix II - Media and Buffers

Fastidious Anaerobic Agar (FAA) 46 g FAA (Acumedia)

1 L ddH2O Autoclave at 121ºC x 15 min Cool to 47ºC Add 50mL defibrinated sheep’s blood (5% v/v) Pour into petri dishes

JVN agar 46 g FAA (Acumedia)

1 L ddH2O Autoclave at 121ºC x 15 min Cool to 50ºC Add 50mL defibrinated sheep’s blood (5% v/v) 3 mg Josamycin (3µg/mL final concentration) 4 mg Vancomycin (4µg/mL final concentration) 1 mg Norfloxacin (1µg/mL final concentration) Pour into petri dishes

Supplemented TSB (TSBsupp) 30 g Trypic soy broth

1 L ddH2O Autoclave at 121ºC x 15 min Cool to 50ºC Add: Hemin to a final concentration of 5µg/mL Menadione to a final concentration of 1µg/mL

237

Supplemented TSB for Campylobacter spp. (TSBsupp+) 30 g Trypic soy broth

1 L ddH2O Autoclave at 121ºC x 15 min Cool to 50ºC Add: Hemin to a final concentration of 5µg/mL Menadione to a final concentration of 1µg/mL Sodium Formate at 0.2% (w/v) Sodium Fumarate at 0.3 (w/v)

Freezing Media for bacterial strains 60 g Skim milk powder 5 mL DMSO (dimethyl sulfoxide) 5 mL Glycerol

Add ddH2O to 500 mL Autoclave at 121ºC x 15 min Aliquot 1.5 mL into cryovials Add bacteria to cryovial Store at -80 ºC

Bacteroides Bile Esculin (BBE) agar 40 g Trypticase soy agar 20 g Oxgall 1 g Esculin 0.5 g Ferric ammonium citrate 2 mL Hemin (5mg/mL)

Add ddH2O to 1L Adjust to pH 7.0 Autoclave at 121ºC x 15 min Cool to 50ºC and pour plates

238

Trypticase Soy Esculin (TSE) agar 4g Trypticase soy agar 1 g Esculin 0.5 g Ferric ammonium citrate 2 mL Hemin (5 mg/mL)

Add ddH2O to 1L Adjust to pH 7.0 Autoclave at 121ºC x 15 min Cool to 50ºC and pour plates

Agarose gels for electrophoresis 4 g agarose (for a 2% gel) or 2 g (for a 1% gel) 200 mL TAE buffer Microwave to dissolve agarose Pour agarose into gel casting tray Add 2 µL ethidium bromide

TAE (Tris-acetate-EDTA) buffer Make 50X TAE buffer: 242 g Tris base 750 mL Glacial acetic acid 100 mL of 0.5 M EDTA, pH 8.0

Add ddH2O to 1L

Dissolve Tris base in approximately 750 mL of ddH2O. Carefully add glacial acetic acid and

EDTA and make up to 1 L with remaining ddH2O. Dilute the 50X stock to 1X with ddH2O for agarose gel electrophoresis.

TE (Tris/EDTA) Buffer 10 mM Tris-HCL pH 8.0 1 mM ETDA pH 8.0

239

SDS-PAGE lysis buffer

0.5 mL ddH2O 2.5 mL 0.5 M Tris-HCL pH 6.8 2 mL 20% (v/v) SDS 2 mL Glycerol 0.5 mg Bromophenol Blue 2 mL Beta-mercaptoethanol

SDS (20% w/v) 200 g electrophoresis-grade SDS

900 mL ddH2O Heat to 68ºC and stir Adjust to pH 7.2

Adjust volume to 1 L ddH2O and store at room temperature

Non-permeabilizing (Non-perm) Block 10% (v/v) Normal goat serum PBS

Permeabilizing Block 10% (v/v) Normal goal serum 0.1% (w/v) Triton X100 PBS

TBST buffer 50 mM Tris buffer 150 mM Nacl 0.05% (v/v) Tween 20

Add ddH2O to 1L Adjust to pH 7.6

240

Co-Aggregation Buffer 0.01 M Tris-HCL adjusted to pH 8.0

0.001 M CaCl2

0.001 M MgCl2 0.15 M NaCl

2% (v/v) NaN3 Make up to 1L in sterile water

Caco-2 cell media 500 mL DMEM/High glucose (HyClone Laboratoried) 50 mL Fetal Bovine Serum Filter sterilize and keep at 4ºC

PBS+ (for tissue culture)

Add 0.1 g CaCl2 to 1L PBS (phosphate buffered saline; ThermoFisher Scientific) Filter sterilize and keep at 4ºC

1X Versene 2.0 g EDTA 8.0 g NaCl

11.5 g Na2HPO4 2.0 g KCl

2.0 g KH2PO4 2.0 g Dextrose

Make up to 1 L ddH2O and filter sterilize

241

Appendix III- Supplementary Data

Antibiotic exclusion assay pre-screen to determine the effects of IL-27 addition to Cac02 cells on Fn invasion

In order to test the hypothesis that IL-27 would negatively affect the ability of Fn to invade intestinal epithelial cells, the invasion of Caco-2 cells by the highly invasive, IBD-derived strain Fn 7-1 was examined in the presence of IL-27 (100 ng/µL). For this work, only Fn 7-1 was assessed as the antibiotic exclusion assay was used as a pre-screen for the more sensitive invasion assay using differential staining and immunofluorescence, and the Fn 7-1 strain is sensitive to gentamicin. During the assay, following invasion, gentamicin was added to the Caco-

2 cell culture media where it killed bacterial cells that had not invaded.

The addition of the cytokine IL-27 to the Caco-2 cells, prior to infection with Fn 7-1, significantly inhibited bacterial invasion when compared to invasion of untreated, control cells

(P<0.005) (Figure A1).

242

Figure AIIIa. Invasion of Caco-2 cells with Fn 7-1 in the presence (blue) or absence (red) of IL- 27 (100 ng/µL). Cells were incubated with bacteria for 4 hours before gentamicin treatment. A two-tailed Student’s t-test was used to infer significance between absence or presence of IL-27 *P<0.05.

243

Bicinchroninic acid (BCA) assay for P. aeruginosa MORN2 peptide quantification

The BCA protein assay is used for quantification of total protein in a sample. The principle of this method is that proteins can reduce Cu+2 to Cu+1 in an alkaline solution (the biuret reaction) and result in a purple color formation by bicinchoninic acid. The PierceTM BCA protein assay kit (ThermoFisher Scientific; Burlington, ON) and manufacturer’s directions were followed. Briefly, a dilution series of the target protein of unknown concentration and bovine serum albumin protein (BSA) standards were made up in BCA Reagent A and 25µL aliquots were added to a 96-well plate. Then 200µL of the provided working reagent was added to each well. The 96-well plate was then incubated at 60ºC for 30 min before absorbance was read at

562nm on a Victor 3V 1420 Multilabel Counter (PerkinElmer; Waltham, MA). Construct a standard curve using the values obtained from the BSA standard using Excel. Determine the equation of the best fit line. Use the equation to calculate the µg/mL of the unknown protein.

244