The Effect of Treponema Denticola on Porphyromonas Gingivalis Phenotypes and Transcriptome

The Effect of Treponema denticola on Porphyromonas gingivalis Phenotypes and Transcriptome

Lin Xin Kin ORCID identifier 0000-0003-2378-1711

Submitted in total fulfilment of the requirements of the degree of Doctor of Philosophy

October 2018

Oral Health CRC Melbourne Dental School Faculty of Medicine, Dentistry and Health Sciences The University of Melbourne

Abstract

Chronic periodontitis is an inflammatory, bacterial biofilm-associated disease resulting in destruction of the tooth’s supporting tissues. The imminent progression of chronic periodontitis can be predicted by the levels of Porphyromonas gingivalis and Treponema denticola in subgingival plaque. Living in a complex oral polymicrobial community, these two bacterial species display close association via physical interaction and metabolic cooperativity in the biosynthesis and cross-feeding of growth substrates. These interspecies interactions result in the coordination of their physiological activities, some of which exhibit complementary and combinatory effects in enhancing their growth and virulence factors. A previous study demonstrated that coculture of T. denticola and P. gingivalis in a continuous system led to upregulation of T. denticola glycine utilisation systems. Likewise, P. gingivalis increased the production of free glycine by proteolytic hydrolysis of peptide-bound glycine during growth in T. denticola conditioned medium (TdCM), suggesting cross-feeding of glycine from P. gingivalis to T. denticola. Free glycine is an important nutrient source for T. denticola, contributing to a dramatic increase in T. denticola growth rate and biomass. Given that P. gingivalis glycine release was stimulated in TdCM, this study aimed to determine and characterise the T. denticola stimulatory factors that had been released into TdCM. TdCM was fractionated by size filtration and reversed-phase high-performance liquid chromatography (RP-HPLC) to determine the most active stimulatory fractions. The free glycine produced by P. gingivalis in different TdCM fractions was then quantified using a glycine enzyme-linked immunosorbent assay (ELISA) kit and liquid chromatography-triple quadrupole mass spectrometry (LC-QQQ-MS). As several RP-HPLC fractions of TdCM appeared to stimulate the release of free glycine by P. gingivalis, this indicated that there was no specific signalling molecule that stimulated P. gingivalis to produce glycine. Instead, the release of glycine by P. gingivalis was likely due to P. gingivalis proteinases that further digested the peptides that had been partially processed by T. denticola proteases during growth in the medium. Candidate peptidases of P. gingivalis that could be involved in the hydrolysis of glycine-containing peptides into free glycine were selected by bioinformatic predictions of the localisation and specificities of P. gingivalis putative peptidases. Interestingly, inactivation of PG0753 and PG1788 that encode putative PrtQ collagenase and C1 family cysteine peptidase respectively, showed a reduction in the rate of free glycine production relative to P. gingivalis

calculated cell number in OB:CM, relative to wild type. These results indicated that these two peptidases might play a role in the release of free glycine by P. gingivalis in the presence of T. denticola stimulatory factors. Other potential multimodal interactions of T. denticola and P. gingivalis were examined by investigating the differential gene expression profiles of P. gingivalis during growth in TdCM, compared to oral bacterial growth medium (OBGM). There were a total of 132 genes that showed differential expression, which included transcripts related to metabolic pathways, signal transduction systems, transcription regulatory systems, nucleic acid interacting protein encoding genes, transporters and hypothetical protein encoding genes. This work generated more testable hypotheses of the molecular effects of T. denticola on P. gingivalis gene expression and as well as potential mechanisms that might contribute to P. gingivalis and T. denticola physiological interactions.

Declaration

This is to certify that:

(i) the thesis comprises only my original work towards the PhD except where indicated in the Preface,

(ii) due acknowledgement has been made in the text to all other material used,

(iii) the thesis is fewer than 100,000 words in length, exclusive of tables, maps, bibliographies and appendices.

Lin Xin Kin

III

Preface

This thesis is submitted for the degree of Doctor of Philosophy at the University of Melbourne. The research described herein was conducted as a continuation of Tan et al. (2014) work: Porphyromonas gingivalis and Treponema denticola Exhibit Metabolic Symbioses. This project was performed under the supervision of Dr. Nada Slakeski, Prof. Stuart Dashper and Dr. Catherine A. Butler in the Melbourne Dental School, from July 2014 until July 2018.

Part of the contents of Chapter 1 have been published in a review article with Hong Min Ng as first co-author: *Ng, H. M., *Kin, L. X., Dashper, S. G., Slakeski, N., Butler, C. A. and Reynolds, E. C. (2016) 'Bacterial interactions in pathogenic subgingival plaque', Microbial Pathogenesis, 94, 60-69.

For Chapter 3, Dr. Kheng Tan guided me in the preparation of TdCMs and processing of samples for the metabolomic experiments. Ms. Komal Kanojia and Dr. Dedreia Tull from Metabolomics Australia provided service and assistance in the generation and analysis of data on metabolomics.

Chapter 4, genomic library preparation and Ion Torrent sequencing were performed by Brigitte Hoffmann. The resulting sequencing reads were analysed by Dr. Catherine A. Butler.

For Chapter 5, Micromon Monash and Monash Bioinformatics were paid to perform the RNA sequencing and analyse the results. RNA sequencing library preparation and sequencing were performed by Mr. Scott Coutts from Micromon Monash. The raw files were analysed by Mr. Kirill Tsyganov and Assoc. Prof. David Powell from the Monash Bioinformatics platform and output was shared on the Degust web tool.

Acknowledgements

To Yong Kai, my soulmate, companion, cheerleader, supporter, teacher and househusband for your support, encouragement and enthusiasm that have made our journeys in Australia fruitful and exciting. I would like to thank my mom, dad and family members for their love, understanding and moral support to me in the pursuit of this study and constantly reminding me to finish what I started. Thank you my supervisors, Dr. Nada Slakeski for accepting me as her PhD student, Prof. Stuart Dashper for entrusting me with this project and Dr. Catherine Butler for her warm hugs, patience and support in the laboratory. I am very appreciative for their dedication in proofreading and reviewing my thesis drafts. My sincere gratitude to Assoc. Prof Paul Veith, who has provided helpful suggestions and advice on the research direction. His guidance has been a valuable input in this work. I am grateful to all staff at Metabolomics Australia, especially Komal Kanojia and Dr. Dedreia Tull for their time and patience in helping me with the metabolomic data analysis. Mr. Scott Coutts and Mr. Mark Cauchi from Micromon Monash for performing the library preparation and RNA sequencing adeptly. Mr. Kirill Tsyganov and Assoc. Prof. David Powell from the Monash Bioinformatic platform for their quick response in answering questions I have on the RNA sequencing data. I would like to extend my gratitude to all past and present lab members. Dr. Kheng Tan for his suggestions and guidance in reproducing his work and continuing his research project. Mr. David Stanton for his guidance on the RP-HPLC. Ms. Brigitte Hoffmann for her contribution on the Ion Torrent sequencing. Dr. Alexis Gonzalez for tirelessly optimizing the protocol for bacterial cell enumeration using the flow cytometry. A great thank you to Ms. Deanne Catmull, Dr. Tanya D’Cruze, Dr. Christine Seers, Ms. Caroline Moore, Mr. Steve Cleal, Ms. Sze Wei Liu, Mr. William Singleton, Dr. Brent Ward and Dr. Jacqueline Heath, who have provided technical assistance and support in my studies. I am also truly thankful to other PhD students and friends, who provide me inspirations, emotional support and sweet memories. It is my pleasure to have both Hong Min and May Ju as my comrades in completing this PhD journey with me, and to have all the desserts they both made and shared with me. Thanks to Jia Min, Lilian, Mathew, Ros, Nidhi, Sayeed, Grace and Jason for all the lunch breaks, chats and laughter together.

Finally, I am grateful to the University of Melbourne for the Fee Remission Scholarship (MIFRS) and Melbourne International Research Scholarship (MIRS). I appreciated Melbourne Dental School, The University of Melbourne for the provision of studentship stipend support for the extension of my PhD candidature. Laureate Prof. Eric C. Reynolds AO and Oral Health Cooperative Research Centres (CRC) for the top-up scholarship and funding support in this research project.

Table of Contents

Abstract ...... I Declaration ...... III Preface...... IV Acknowledgements ...... V Table of Contents ...... VII List of Tables ...... XII List of Figures ...... XII Abbreviations ...... XV

Chapter 1 Introduction ...... 1 1.1 Chronic periodontitis ...... 1 1.2 Oral polymicrobial communities ...... 2 1.3 Intercellular communication ...... 6 1.4 Porphyromonas gingivalis ...... 10 1.4.1 Gingipains ...... 10 1.4.2 Outer membrane vesicles ...... 11 1.4.3 Mediators of physical interactions ...... 12 1.4.4 Roles of proteases in P. gingivalis...... 14 1.4.5 Peptide and amino acid metabolism ...... 18 1.5 Treponema denticola ...... 20 1.5.1 Chemotaxis and motility ...... 20 1.5.2 Mediators of physical interactions ...... 23 1.5.3 Roles of proteases in T. denticola ...... 26 1.5.4 T. denticola peptide and amino acid metabolism ...... 29 1.6 P. gingivalis and T. denticola metabolic interactions ...... 31 1.7 Aims of this study ...... 35

Chapter 2 Materials and methods ...... 36 2.1 Bacterial strains and growth conditions ...... 36 2.2 Bacterial cell number enumeration ...... 40 VII

2.3 Molecular biology techniques ...... 41 2.3.1 Extraction of bacterial genomic and plasmid DNA ...... 41 2.3.2 Polymerase Chain Reaction ...... 41 2.3.3 Molecular cloning ...... 46 2.3.4 Agarose gel electrophoresis ...... 46 2.3.5 Nucleotide sequencing ...... 46 2.4 Bacterial transformation ...... 47 2.4.1 Escherichia coli heat-shock transformation ...... 47 2.4.2 P. gingivalis W50 transformation ...... 47 2.5 Whole genome sequencing using the Ion Torrent Personal Genome Machine ...... 48 2.6 Preparation of T. denticola conditioned medium ...... 48 2.6.1 Characterisation of T. denticola stimulatory factors ...... 49 2.6.2 Size filtration of TdCM...... 49 2.6.3 Preparation of a simplified form of TdCM ...... 50 2.6.4 Reversed-phase high-performance liquid chromatography ...... 51 2.7 Determination of free glycine concentration ...... 52 2.7.1 Glycine enzyme-linked immunosorbent assay ...... 52 2.7.2 Liquid chromatography-triple quadrupole mass spectrometry analyses ...... 53 2.8 RNA sequencing ...... 54 2.9 Bioinformatics ...... 55 2.10 Statistical analysis ...... 56

Chapter 3 Characterisation of T. denticola factors that stimulate free glycine production by P. gingivalis ...... 57 3.1 Introduction ...... 57 3.2 Results ...... 61 3.2.1 P. gingivalis growth and glycine production ...... 61 3.2.1.1 Nature of T. denticola stimulatory factors ...... 63 3.2.1.2 Simplified forms of TdCM ...... 66 3.2.1.3 TdCM_5F(OB_5F) profile in RP-HPLC fractionation ...... 71 3.2.2 Metabolomic analyses of targeted amine groups ...... 78 3.2.2.1 Free glycine concentration quantification ...... 78 3.2.2.2 Quantification of amine groups in the media ...... 80 3.3 Discussion ...... 84 VIII

3.4 Conclusion ...... 90

Chapter 4 Characterisation of free glycine release by P. gingivalis peptidase mutants ...... 91 4.1 Introduction ...... 91 4.2 Results ...... 93 4.2.1 Selection of P. gingivalis peptidase targets ...... 93 4.2.1.1 Localisation predictions of P. gingivalis proteases ...... 93 4.2.1.2 Determination of P. gingivalis peptidases orthologous to glycine-cleaving peptidases...... 96 4.2.2 Generation of P. gingivalis peptidase mutants ...... 98 4.2.2.1 PCR confirmation of P. gingivalis peptidase mutants ...... 100 4.2.2.2 Whole genome sequencing of P. gingivalis peptidase mutants ...... 102 4.2.3 Characterisation of the glycine-releasing ability of P. gingivalis peptidase mutants 103 4.2.3.1 Growth curves of P. gingivalis peptidase mutants in OBGM and ...... 103 4.2.3.2 Enumeration of P. gingivalis wild type and peptidase mutants in OBGM and OB:CM ...... 105 4.2.3.3 Effects of P. gingivalis peptidase mutants on glycine release ...... 106 4.3 Discussion ...... 107 4.4 Conclusion ...... 109

Chapter 5 Effect of T. denticola conditioned medium on P. gingivalis gene expression...... 110 5.1 Introduction ...... 110 5.2 Results and Discussion ...... 112 5.2.1 RNA extraction and sequencing of P. gingivalis ...... 112 5.2.2 Classification of differentially expressed genes ...... 124 5.2.2.1 Metabolic pathways ...... 124 5.2.2.1.1 Succinate pathways ...... 124 5.2.2.1.2 Glycine related pathways ...... 128 5.2.2.2 P. gingivalis proteases ...... 134 5.2.2.3 Transcription regulators and non-regulatory DNA-binding proteins ...... 137 5.2.2.4 Transporter and efflux systems ...... 141 5.2.2.5 Hypothetical protein encoding genes ...... 145 IX

5.3 Conclusion ...... 148

Chapter 6 Conclusion ...... 149

Bibliography ...... 152 Appendices ...... 194 Appendix I ...... 194 Appendix II ...... 196 Appendix III ...... 199 Appendix A Bioinformatic analyses on PG1607 and PG1788 ...... 210 A.1 Porphyromonas gingivalis C1 cysteine peptidases ...... 210 A.2 Bioinformatic analysis on peptidases PG1605 and PG1788 ...... 212 A.2.1 Sequence comparisons of PG1605 with its homologues ...... 212 A.2.2 Sequence comparisons of PG1788 with its homologues ...... 219 A.2.3 PG1605 modelled structure ...... 223 A.3.2 PG1788 modelled structure ...... 226 Bibiolography (cited in Appendice A) ...... 234

List of Tables Table 2.1 Bacterial strains and plasmids used in this study………………………………….38 Table 2.2: Primers for PCR amplification……………………………………………………43 Table 2.3: Parameters for the flow gradient of solvent B during the separation of molecules in TdCM5F(OB_5F) using semipreparative C18 RP-HPLC…………………………………...52 Table 3.1 pH of RP-HPLC fractions of TdCM5F(OB_5F)………………………………….74 Table 4.1: Summary table of the cleavage specificity, localisation and other important descriptions of the characterised P. gingivalis peptidases…………………………………...92 Table 4.2: Localisation prediction and N-terminal signal peptide prediction of P. gingivalis proteases……………………………………………………………………………………...94 Table 4.3: Rate of free glycine production by different P. gingivalis strains in OBGM and OB:CM……………………………………………………………………………………...107 Table 5.1: Total RNA quality and quantity of P. gingivalis in OBGM and OB:CM……….115 Table 5.2: Summary of the sequencing library with number of assigned reads and percentage of duplications per sample…………………………………………………………………..116 Table 5.3: P. gingivalis genes that were differentially expressed during growth in OB:CM, compared to OBGM………………………………………………………………………...118 Table 5.4: P. gingivalis protease-encoding genes that demonstrated the highest average level of expression in OBGM and OB:CM……………………………………………………….136 Table 5.5: Summary table of the hypothetical protein encoding genes that showed DGE during P. gingivalis growth in OB:CM, relative to OBGM………………………………...147 Table I.1: Relative peak abundance of amino acids determined at initial time point and after P. gingivalis 48 h growth in OB:CM and OBGM………………………………………….195 Table I.2: Relative peak abundance of other amine groups determined at initial time point and after P. gingivalis 48 h growth in OB:CM and OBGM…………………………………….195 Table II.1: Common nucleotide polymorphisms of P. gingivalis peptidase mutant strains (PG0753-, PG1605- and PG1788-) determined by whole genome sequencing analysis…...196 Table III.1: Details P. gingivalis differential gene expression profile during growth in OB:CM, relative to OBGM…………………………………………………………………199 Table III.2: Summary table of genes, metabolic pathways, which include the involvement of metabolic substrates and products that showed DGE during P. gingivalis growth in OB:CM relative to OBGM…………………………………………………………………………...205 Table A.1: Bacterial species and genera containing the highest number of C1 family peptidases based on bacterial species with complete genome sequences…………………..211

List of Figures Figure 1.1: Schematic representation of proposed bacterial accretion in polymicrobial biofIlm community development……………………………………………………………...5 Figure 1.2: P. gingivalis amino acid metabolism…………………………………………….20 Figure 1.3: Schematic diagram of T. denticola chemotactic components and signalling pathways……………………………………………………………………………………...23 Figure 1.4: Glutathione stepwise degradation and catabolism by T. denticola……………....29 Figure 1.5: Proposed T. denticola glycine catabolic pathways……………………………...31 Figure 1.6: Schematic diagram of metabolic interactions between T. denticola and P. gingivalis……………………………………………………………………………………..34 Figure 2.1: Flow chart representing the preparation steps for the minimised forms of TdCM and their designations………………………………………………………………………...51 Figure 3.1: Diagrammatic descriptions of possible scenarios where T. denticola stimulates P. gingivalis to increase free glycine production in TdCM……………………………………..60 Figure 3.2: P. gingivalis growth curve in different media…………………………………...61 Figure 3.3: Average free glycine measurement using a glycine ELISA kit………………….62 Figure 3.4: Ratio of glycine concentration after P. gingivalis 40 h growth, relative to basal glycine concentration in OBGM, OB:CM and OB:PBS……………………………………..63 Figure 3.5: Schematic diagram of the steps performed for different treatments of T. denticola conditioned medium (TdCM)………………………………………………………………...64 Figure 3.6: Growth of P. gingivalis in OB:CM mixture, where TdCMs were subjected to different experimental treatments…………………………………………………………….65 Figure 3.7: Ratio of glycine concentration after P. gingivalis 40 h growth in differently processed TdCM to their respective basal glycine concentration……………………………66 Figure 3.8: Preparation of simplified forms of TdCMs and their annotations……………….67 Figure 3.9: P. gingivalis growth curve in TdCMs obtained by growth of T. denticola in modified OBGM……………………………………………………………………………..68 Figure 3.10: Average ratio of free glycine concentration after 40 h of P. gingivalis growth in different TdCMs mixtures……………………………………………………………………70 Figure 3.11: RP-HPLC fractionation of TdCM5F(OB_5F).. ………………………………..71 Figure 3.12: Chromatogram of OB_5F and TdCM5F(OB_5F)……………………………...73 Figure 3.13: Growth curve of P. gingivalis in different RP-HPLC fractions of TdCM5F(OB_5F)…………………………………………………………………………….75 XII

Figure 3.14: Average ratio of glycine concentration after P. gingivalis 50 h growth to the basal glycine concentration in different RP-HPLC fractions of TdCM5F(OB_5F)…………77 Figure 3.15: Calibration graph of the matrix standards for the determination of absolute glycine concentration in each sample ………………………………………………………..78 Figure 3.16: Glycine concentrations measured at the initial time point and after P. gingivalis 48 h growth in different size filtrates of TdCM……………………………………………...79 Figure 3.17: Average ratio of glycine concentration (mM) after 48 h growth of P. gingivalis in different TdCM size fractions mixed with OBGM……………………………………… ..79 Figure 3.18: Ratio in the peak abundance of amino acids after P. gingivalis 48 h growth in different size-filtrates of TdCM……………………………………………………………...81 Figure 3.19: Ratio in the peak abundance of each amine group after P. gingivalis 48 h growth in different size-filtrates of TdCM…………………………………………………………...83 Figure 3.20: T. denticola and P. gingivalis metabolic enzymes involved in the amino acid catabolism pathways…………………………………………………………………………89 Figure 4.1: Cleavage site sequence logo of peptidases………………………………………97 Figure 4.2: Schematic representation of the relative position of the peptidase targeted and the neighbouring open reading frames…………………………………………………………...99 Figure 4.3: PCR amplicons of P. gingivalis peptidase mutant strains, PG0445-, PG0753-, PG1605- and PG1788-, resolved by agarose gel electrophoresis…………………………..101 Figure 4.4: Growth of P. gingivalis wild type and PG0445-, PG0753-, PG1605- and PG1788- strains in OBGM and OB:CM………………………………………………………………104 Figure 4.5: Comparison of P. gingivalis cell count methods. ………………………………106 Figure 5.1: Quality and integrity of total RNA isolated from P. gingivalis during growth in OBGM or OB:CM…………………………………………………………………………..114 Figure 5.2: Examination of RNA for possible source of gDNA contamination……………115 Figure 5.3: Quality control reports of the RNA sequencing library………………………...116 Figure 5.4: Volcano plot of P. gingivalis transcripts that were detected in the RNAseq analysis……………………………………………………………………………………...117 Figure 5.5: Butyrate metabolism of P. gingivalis W83……………………………………..127 Figure 5.6: P. gingivalis serine, threonine and glycine biosynthesis superpathways to one carbon pathways…………………………………………………………………………….130 Figure 5.7: P. gingivalis ubiquinone and other terpenoid-quinone biosynthesis pathways...133 Figure 5.8: DNA binding protein-encoding genes distribution around the sequenced genome of P. gingivalis W83………………………………………………………………………...140 Figure 5.9: Schematic representation of efflux transporters in Gram negative bacteria……144 XIII

Figure I.1: Linear regression of all data points collected for different experimental set-up for OBGM, OB:CM and OB:PBS……………………………………………………………...195 Figure II.1: Linear regression of change in free glycine concentration as a function of P. gingivalis cell number in (a) OBGM and (b) OB:CM from the time interval of 30 to 40 h of growth……………………………………………………………………………………….199 Figure A.1: Assignment of specific hits and superfamilies of PG1605 in the BLASTp search………………………………………………………………………………………..212 Figure A.2: Phylogenetic tree generated by neighbor-joining using BLOSUM6 by ClustalWS alignment of the amino acid sequences of PG1605 and homologues………………………213 Figure A.3: Multiple sequence alignment of PG1605 against other C1B peptidase family homologues was performed using ClustalO in default parameters…………………………216 Figure A.4: Consensus sequence representing the phylum of Firmicutes and Bacteroidetes, in comparison to yeast and human BHs amino acid sequences. ………………………………218 Figure A.5: Phylogenetic tree of PG1788 against other C1 family member by neighbour- joining of BLOSUM62……………………………………………………………………...220 Figure A.6: Multiple sequence alignment of PG1788 with other C1 peptidases homologues using ClustalO with default parameters…………………………………………………….222 Figure A.7: PG1605 modelled structure by I-TASSER and structural alignment with 1CB5A………………………………………………………………………………………225 Figure A.8: PG1788 modelled structure by I-TASSER and structure alignment with 3PW3A……………………………………………………………………………………...229 Figure A.9: PG1788 modelled structure alignment with 4K7C and 1A6R…………………231

XIV

Abbreviations

°C Degrees Celsius

A260 Absorbance at wavelength 260 nm

A280 Absorbance at wavelength 280 nm ABC ATP-binding cassette ACN Acetonitrile AHL N-acylhomoserine lactone AI-2 Autoinducer-2 ANOVA One way analysis of variance Ap Ampicillin ATP Adenosine 5’-triphosphate ATCC American Type Culture Collection

BA Biogenic amine BSA Bovine serum albumin BDPs BSA-digested peptides BH Bleomycin hydrolase BHI Brain heart infusion broth bp Base pair BLASTp Protein BLAST

CDD Conserved domain database CDM Chemically defined medium Cfp Cytoplasmic filaments

5,10-CH2-THF Methylenetetrahydrofolate COG Clusters of orthologous groups of proteins CSP Competence-stimulating peptides CTD C-terminal domain CTLP Chymotrypsin-like protease complex

DIW De-ionised water DGE Differential Gene Expression DNA Deoxyribonucleic acid dNTP Deoxynucleotidyltriphosphate DPP Dipeptidylpeptidase DTT 1,4-Dithiothreitol DUF Domain of unknown function

ECF Extracytoplasmic function EDTA Ethylenediaminetetraacetic acid ELISA Enzyme-linked immunosorbent assay Em Erythromycin

FD Freeze-dried FDR False discovery rate Fe-S Iron–sulphur g Relative centrifugal force GCS Glycine cleavage system GC-MS Gas chromatography mass spectrometry gDNA Genomic DNA GSH Glutathione h Hour HBA Horse blood agar HOMD Human Oral Microbiome Database HTH Helix-turn-hellix i.d Internal diameter IPTG Isopropyl β-D-1-thiogalactopyranoside IL Interleukin

XVI

KEGG Kyoto Encyclopedia of Genes and Genomes Kgp P. gingivais gingipain K

LB Lysogeny broth LBA LB agar LC-MS/MS Liquid chromatography tandem mass spectrometry LC-QQQ-MS Liquid chromatography triple quadrupole mass spectrometry LPS Lipopolysaccharide

MATE Multidrug and toxic compound extrusion MCPs Methyl-accepting chemotaxis proteins MDR Multidrug resistance MFS Major facilitator superfamily min Minutes MK Menaquinone MOPS (3-(N-morpholino)propanesulfonic acid) mRNA Messenger ribonucleic acid MSA Multiple sequence alignment Msp Major sheath protein MWCO Molecular weight cut-off nt Nucleotide NAD+ Nicotinamide adenine dinucleotide NAM N-acetyl muramic acid NCBI National Centre for Biotechnology Information NEB New England Biolabs

OBGM Oral bacterial growth medium

OD650 Optical density measurement at wavelength 650 nm OMVs Outer membrane vesicles O/N Overnight ORF Open reading frame XVII

p Plasmid PBS Phosphate buffered saline PCR Polymerase chain reaction PDB Protein Data Bank PF Periplasmic flagella ProtK Proteinase K PrtP Dentilisin Ptp Prolyl tripeptidylpeptidase qRT-PCR Quantitative reverse transcription PCR QS Quorum sensing

Rgp Gingipains R RMS Root mean square RNA Ribonucleic acid RNAseq RNA sequencing RND Resistance nodulation division RP-HPLC Reversed-phase high-performance liquid chromatography rpm Revolution per minute RT Retention time s Seconds SDS-PAGE Sodium dodecyl sulfate polyacrylamide gel electrophoresis SEM Scanning electron microscopy SOE Splicing by overlap extension SMR Small multidrug resistance SNP Single nucleotide polymorphism spp. Species SUF Sulfur mobilization t Time XVIII

TAE Tris-acetate buffer TCSTS Two component signal transduction system TdCM T. denticola conditioned medium TdSF T. denticola stimulatory factors TF Transcription factor TFA Trifluoroacetic acid THF Tetrahydrofolate TFF Tangential flow filtration

TPCK L-1-Tosylamide-2-phenylethyl chloromethyl ketone TPP Thiamine pyrophosphate T9SS Type IX secretion system

U Units

V Volts Vol Volume

XIX

1 Chapter 1 Introduction

2 1.1 Chronic periodontitis 3 Chronic periodontitis is caused by microbial-associated inflammation of the tooth’s 4 structural supports, including gingiva, periodontal ligaments, and eventually leads to bone and 5 tooth loss if left untreated. This gum disease was the sixth-most prevalent condition afflicting 6 10.8% people worldwide in 2010 (Eke et al. 2012), with higher prevalence among the elderly, 7 lower socio-economic groups and individuals in developing countries. Based on the assessment 8 criteria of having at least two interdental sites with ≥4 mm clinical attachment loss and ≥ 5 mm 9 probing depth, the Australian National Survey of Adult Oral Health reported that 22.9% of 10 Australian adults had severe periodontitis between the years of 2004 to 2006 (Adelaide 2009, 11 Mdala et al. 2014). Specifically, chronic periodontitis afflicted 24.5%, 43.6% and 60.8% of 12 Australian adults in the age groups of 35-54, 55-74 and ≥75 years respectively (Adelaide 2009). 13 The risk factors of periodontitis can be multifactorial, including genetics, diet, oral 14 hygiene, age, tobacco and alcohol use, as well as other environmental factors (Van Dyke and 15 Dave 2005, Genco and Borgnakke 2013, Yousef 2014). More recently, chronic periodontitis 16 has been reported as an important indicator and risk factor associated with other systemic 17 diseases, such as diabetes, rheumatoid arthritis, cardiovascular disease, atherosclerosis, and 18 preterm birth (Gully et al. 2014, Hujoel et al. 2000, Jeffcoat et al. 2001, Lalla et al. 2003, 19 Meurman et al. 2004, Mealey and Oates 2006, Seymour et al. 2007). Shifting of the subgingival 20 plaque microbiota population from a high abundance of commensal Gram positive 21 streptococcal spp. to a group of pathogenic Gram negative bacteria, is one of the major 22 contributors to chronic periodontal disease progression (Nath and Raveendran 2013, Jiao et al. 23 2014). Scaling, root planning and administration of antibiotics are the nonsurgical treatments 24 commonly employed for the removal of dental plaque to control chronic periodontitis. 25 However, given the polymicrobial nature of disease development, a better understanding of the 26 pathogenic microbial aetiology would enable more targeted approaches for modulation of the 27 bacterial composition, abundance and their interactomes in the subgingival dental plaque so as 28 to restore the balance of the oral microbiota. 29

1 1.2 Oral polymicrobial communities 2 Oral polymicrobial communities cover a wide array of genetically distinct groups of 3 bacteria interacting synergistically and antagonistically in a confined space with limited 4 resources in order to achieve selective advantages (Kolenbrander et al. 2010). The complexity 5 of the oral polymicrobial biofilm is established through sequential spatial and temporal 6 organization of at least 700 bacterial species to ensure functional bacterial community 7 development and maintenance (Chen et al. 2010). Unlike indigenous oral bacteria, the 8 introduction of allochthonous bacteria to the oral cavity by food ingestion is transient because 9 they lack binding specificity towards the oral host and bacterial ligands, thus can be easily 10 removed by salivary flow and oral hygiene practices (Kolenbrander et al. 2010). Adherence of 11 early colonisers, such as Streptococcus, Actinomyces and Veillonella species to host surface 12 receptors on epithelial cells or the salivary pellicles that coat the tooth surfaces, are the initial 13 step for the establishment of oral dental plaque (Cheaib et al. 2016, Nobbs et al. 2011, 14 Kolenbrander et al. 2002). This in turn allows binding and accretion of the secondary colonisers 15 into a mature heterotypic biofilm (Kolenbrander et al. 2002, Kolenbrander and London 1993). 16 Close associations between oral bacteria can be achieved by coadhesion or coaggregation with 17 compatible partners, and these partners generally exhibit physical, metabolic and chemical 18 interactions within the multispecies bacterial communities (Guo et al. 2014). Colonisation by 19 the pioneering species plays an important role in modifying the pH, oxygen levels and nutrient 20 availability in the niche and providing new binding surfaces for the succeeding bacteria. Being 21 the most abundant bacteria in the oral cavity, streptococci metabolize carbohydrates found in 22 food or endogenous salivary glycoproteins into lactate, acetate and formate, which significantly 23 reduce the pH of the microenvironment (Beighton et al. 2013). However, Veillonella spp. lacks 24 the ability to utilise carbohydrates and preferentially utilises lactate and pyruvate as the carbon 25 and energy sources (Chalmers et al. 2008). This metabolic preference of Veillonella spp. can 26 be fulfilled by association with streptococci (Bradshaw and Marsh 1998), while their metabolic 27 cross-feeding interactions can optimize the nutrient utilisation efficiencies in the niche and 28 reduce the acidity of the surroundings. Furthermore, commensalism can be observed between 29 Actinomyces spp. and S. gordonii. For example, S. gordonii is auxotrophic for arginine under 30 low arginine concentration in aerobic conditions; coaggregation of Actinomyces naeslundii or 31 Actinomyces oris with Streptococcus gordonii enabled stabilization of the S. gordonii arginine 32 biosynthesis genes, thus protecting S. gordonii from protein carbamylation and sustaining its 33 aerobic growth when exogenous arginine became depleted (Jakubovics et al. 2008, Jakubovics 2

1 et al. 2008b, Jakubovics et al. 2015). These contact-dependent and protease-mediated 2 interactions of Actinomyces spp. with S. gordonii affect its arginine regulatory responses, which 3 in turn allows S. gordonii to elicit a cellular response that corresponds to arginine restrictive 4 conditions for the stabilisation of its arginine biosynthesis pathways, as well as regulation of 5 other gene expression, cellular growth and biofilm formation (Jakubovics et al. 2015, 6 Mohammed at al. 2018). Furthermore, A. naeslundii produces catalase during coaggregation 7 with S. gordonii, which protects S. gordonii from oxidative damage induced by self-produced

8 H2O2 (Jakubovics et al. 2008b). These interactions enhance S. gordonii growth and also typify 9 the physiological compatibility required during interspecies bacterial coaggregation. 10 The pioneering species create new surface binding receptors and modify the 11 environmental conditions with their metabolites, aiding the recruitment and colonisation of 12 succeeding oral bacterial species. Fusobacterium nucleatum is known as a “bridge species” as 13 it is able to mediate coaggregation and coadhesion with the early and late colonisers. Being a 14 middle coloniser, F. nucleatum is aerotolerant, able to thrive under a broad pH range (pH 5-7) 15 and plays a role in transforming the microenvironment into one favouring the growth of late 16 colonisers (Fig. 1.1). It is able to reduce oxygen tension and has acid neutralising properties 17 due to the production of ammonia from amino acid metabolism (Takahashi et al. 1997, Diaz et 18 al. 2002). Successive colonisers in the subgingival plaque can comprise of Eubacterium 19 species, Aggregatibacter actinomycetemcomitans, Prevotella intermedia, Selenomonas 20 species, Tannerella forsythia, Porphyromonas gingivalis and Treponema denticola as 21 illustrated in Figure 1.1. Similar to the aforementioned examples, multimodal interactions 22 between bacterial species are required to develop heterotypic polymicrobial communities. For 23 instance, S. gordonii increased lactate utilisation by A. actinomycetemcomitans, which is its 24 preferred substrate during coculture (Ramsey et al. 2011), as well as providing A. 25 actinomycetemcomitans with electron acceptors for respiratory metabolism, which in turn 26 enhanced its biomass, persistence and virulence (Stacy et al. 2016). However, A. 27 actinomycetemcomitans is equipped with both catalase and dispersin B enzyme as its defensive

28 and protective mechanisms to alleviate the oxidative stress effects of H2O2 produced by oral 29 streptococci during their coaggregation (Stacy et al. 2014). These physiological interactions 30 promote and sustain the heterogeneity observed in polymicrobial subgingival biofilm. 31 The red complex species are composed of P. gingivalis, T. denticola and T. forsythia 32 that are found to be associated with chronic periodontitis (Socransky and Haffajee 1992, 33 Socransky et al. 1998, Byrne et al. 2009). These bacteria have also been recognized as

1 pathobionts as they demonstrated enhanced virulence under dysbiotic conditions, which 2 occurred by a shift in the balance of the oral microbiome diversity and proportion, leading to 3 disease manifestation (Lamont and Hajishengallis 2015). In addition, many studies reported 4 the low number of T. denticola and P. gingivalis in healthy sites, but their proportions and 5 number increased with severity of the disease (Holt and Ebersole 2005, Byrne et al. 2009). 6 Mouse model experiments have also verified their synergistic virulence by comparing the 7 lesion sizes and the bone loss effects resulting from coinfection and monoinfection of the 8 periodontopathogens (Kesavalu et al. 1998, Orth et al. 2011). To promote their colonisation 9 and establishment in the subgingival region, T. denticola and P. gingivalis express a wide 10 variety of ligands on their cell surfaces that bind to many host and bacterial surface receptors, 11 which are described in detail in later sections. P. gingivalis also interacts with other bacterial 12 species, such as Streptococcus species, F. nucleatum, P. intermedia, T. forsythia, Actinomyces 13 species and A. actinomycetemcomitans (Kuboniwa et al. 2017, Nagayama et al. 2001, Diaz et 14 al. 2002, Kuboniwa et al. 2006, Simionato et al. 2006, Maeda et al. 2008, Takahashi et al. 15 1997, Takahashi and Sato 2002, Szafrański et al. 2015), whereby P. gingivalis specific 16 partnerships with different bacterial species can dictate its persistence and prevalence in the 17 subgingival plaque. In addition, these bacteria can potentially be accessory pathogens that 18 enhance P. gingivalis virulence in the dysbiotic oral polymicrobial communities. Whilst the 19 red complex species have been characterised for their virulence and pathogenic involvement 20 in chronic periodontitis, metagenomic analyses also revealed that other bacterial colonisers, 21 which include Filifactor alocis, Peptostreptococcus stomatis, Parvimonas micra, Prevotella, 22 Eubacterium, Selenomonas, Megasphaera and other Treponema species, could be associated 23 with chronic periodontitis (Kumar et al. 2006, Colombo et al. 2009, Griffen et al. 2012, 24 Abusleme et al. 2013). However, the physiological activities and virulence of these bacterial 25 species have not been well characterised. 26 Given the complexity of oral polymicrobial communities, on-going studies have been 27 conducted to elucidate the molecular mechanisms adopted by these pathobionts to tip the 28 balance of the oral microbial communities towards favouring their propagation. Chemical 29 interactions are an important aspect for the development of a polymicrobial biofilm community 30 as these diffusible molecules can regulate bacterial phenotypic expression and virulence 31 factors. In the following section, the virulence factors of P. gingivalis and T. denticola that 32 contribute to their colonisation and persistence at disease sites are addressed. Some of the 33 pathogenic mechanisms can also lead to the destruction of epithelial tissues, host

1 immunomodulatory factors and the accumulation of cytotoxic metabolites that further 2 aggravate the periodontal disease conditions. Understanding the polymicrobial network 3 interactions and development could shed light on potential effective therapeutics options and 4 identify suitable biological markers for the diagnosis of periodontal disease progression.

7 8 Figure 1.1: Schematic representation of proposed bacterial accretion in polymicrobial biofilm 9 community development. Image reproduced from Kolenbrander et al. (2010).

19 5

1 1.3 Intercellular communication 2 Quorum sensing (QS) is achieved by the release of small diffusible signalling molecules 3 that regulate bacterial gene expression and coordinate physiological activities in a density- 4 dependent manner. As dental plaque is comprised of a complex bacterial biofilm community, 5 it is one of the best platforms for the study of bacterial signalling molecules. The mechanism 6 of intercellular communication serves as an important asset for the facilitation of the 7 spatiotemporal development, roles and interactions of each constituent in an oral multispecies 8 biofilm (Kolenbrander et al. 2010). Intercellular cross-talking in the environment regulates 9 bacterial virulence mechanisms, including bacterial cell competency for genetic 10 transformation, conjugation, secondary metabolite production, antibiotic secretion, motility 11 and biofilm formation, as well as promoting symbiosis and providing bacteria with an 12 additional adaptation ability via the global gene regulatory networks (Miller and Bassler 2001, 13 Burgess et al. 2002). From an evolutionary perspective, the long term benefits of signalling 14 communication in regulating behaviours between bacterial species outweighs the costs of 15 producing signalling molecules (Keller and Surette 2006). The term “communication” used in 16 this context only applies to the extent of the production of a molecule (signal) from the emitter, 17 which in turn triggers a specific response from the receiver. 18 The commonly known species-specific signalling molecules are N-acylhomoserine 19 lactone (AHL) released by Gram negative bacteria and modified peptides by Gram positive 20 bacteria. AHL synthase LuxI and AHL receptor protein LuxR are the main components for 21 AHL production and detection respectively, which trigger a series of intraspecific regulatory 22 activities. To date, there are no Gram negative oral bacteria that possesses the ability of utilising 23 AHL for intercellular communication (Whittaker et al. 1996, Guo et al. 2014). 24 Oral Streptococcus spp. are found to be active producers of competence stimulating 25 peptides (CSP). The fundamental components of CSP-mediated QS are ComC, which produces 26 CSP from the modification of peptide precursors, ComAB oligopeptide transporter for the 27 export of CSP, and ComD/E, the two component signal transduction system for the detection 28 of and response to CSP respectively (Suntharalingam and Cvitkovitch 2005). CSP is involved 29 in bacterial horizontal gene transfer by inducing DNA uptake in naturally competent 30 streptococci in a cell-density dependent mode. CSP also confers S. mutans with an acid 31 tolerance response (Li et al. 2001) and is also required for the production of bacteriocin that 32 acts as a competitive mechanism in highly populated bacterial biofilm (van der Ploeg 2005). 33 Although CSP was first characterised as a species-specific signalling molecule, later studies 6

1 discovered that S. mutans CSP is indirectly involved in heterotypic biofilm development with 2 Candida albicans. Evidence for this includes the fact that a S. mutans comCDE deficient strain 3 failed to initiate monospecies biofilm formation, which is proposed to be due to the loss of 4 genetic competency (Li et al. 2002), but exhibited increased biofilm with Candida albicans, 5 which might be due to the accumulation of extracellular DNA that helped stabilize the biofilm 6 structure (Jack et al. 2015). CSP negatively affects C. albicans morphological transition from 7 yeast-to-hypha (Jarosz et al. 2009), as well as inhibiting its monospecies biofilm formation 8 (Jack et al. 2015). In a mixed-species biofilm with C. albicans, S. gordonii CSP signalling 9 pathways became upregulated along with its competency genes, bacteriocin and mutacin 10 production (Sztajer et al. 2014). 11 Autoinducer-2 (AI-2) is a universal signalling molecule employed by both Gram 12 positive and Gram negative bacteria that serves as a modulator in controlling bacterial gene 13 expression. AI-2 is normally found to regulate bacterial physiological activities and pathogenic 14 mechanisms, such as stress responses, adhesion properties and biofilm formation under varying 15 environmental conditions (Hardie and Heurlier 2008). LuxS, which is conserved across a 16 diverse range of bacteria, plays a dual-role as a methionine metabolic pathway protein and 17 interspecies signalling molecule producer. AI-2 arises from the spontaneous cyclization of 4,5- 18 dihydroxy-2,3-pentanedione (DPD), which is derived from the catabolic product S- 19 ribosylhomocysteine by LuxS catalytic activity (Schauder et al. 2001). This interspecies 20 signalling molecule is structurally interconvertible, existing in various forms in equilibrium 21 that provides flexibility for the binding of different receptors on a variety of bacteria. The 22 natural occurrence of AI-2 provides the technical convenience of using the standard 23 bioluminescence reporter strain Vibrio harveyi BB170 for the screening of AI-2 producing 24 bacteria (Frias et al. 2001). This biosensor assay resulted in the determination of S. mutans, S. 25 gordonii, F. nucleatum, A. actinomycetemcomitans and S. oralis as AI-2 producing bacterial 26 species (Frias et al. 2001). In depth studies into the role of AI-2 in multispecies communities 27 revealed that AI-2 concentrations can be utilised to monitor the overall bacterial density and 28 regulate oral bacterial establishment in a mixed-species population. For example, early 29 colonisers S. oralis and A. naeslundii can only achieve mutualistic growth when the bioactive 30 threshold of AI-2 is reached (Rickard et al. 2006). AI-2 affects S. gordonii monospecies biofilm 31 architecture and biovolume; likewise it controls the proportional abundance of S. gordonii and 32 S. oralis in a dual-species biofilm (Cuadra-Saenz et al. 2012). Although A. 33 actinomycetemcomitans are found to physically attach to C. albicans, the production of AI-2

1 inhibits C. albicans hypha and biofilm formation in a dose-dependent fashion (Bachtiar et al. 2 2014). During close association of S. gordonii and Veillonella atypica, interspecies signalling 3 molecules produced by V. atypica result in an increased biofilm (Izumi and Futoshi 2015), as 4 well as an enhanced gene expression of S. gordonii α-amylase that could increase its 5 carbohydrate utilisation (Egland et al. 2004). As a result, lactate produced via S. gordonii 6 carbohydrate fermentation could be cross-fed to V. atypica. Their metabolic relationship 7 demands close engagement as it is mediated by short-range chemical interactions, by which 8 these signalling molecules become rapidly diluted and fall below the threshold required for a 9 response. 10 Interestingly, P. gingivalis possesses a functional LuxS orthologue for the generation 11 of AI-2, which is detectable using the V. harveyi biosensor and luminescence assay (Chung et 12 al. 2001, Burgess et al. 2002). P. gingivalis AI-2 production is regulated by environmental 13 osmolarity and cell growth phase, as maximum AI-2 production was detected during mid 14 exponential phase, but the molecules became degraded at stationary phase or when NaCl 15 concentration increased in the environment (Chung et al. 2001). P. gingivalis LuxS protein has 16 a regulatory role in stress-related genes that improve its survival under higher temperature, pH

17 and H2O2 levels (Yuan et al. 2005). The mutation of luxS in two independent studies showed 18 contradictory outcomes in the transcriptional and phenotypic profiles of the resulting mutants. 19 Chung et al. (2001) showed an upregulation of P. gingivalis gingipain encoding gene rgpA and 20 a gene that showed homology with Pseudomonas fluorescens hemin acquisition protein, while 21 other study observed attenuation of gingipains and hemagglutinin activities in the luxS mutant 22 (Burgess et al. 2002). However, both studies agreed that P. gingivalis luxS has a role in 23 modulating hemin acquisition and protease activities. Undoubtedly, AI-2 generated by other 24 oral bacteria also has an impact on P. gingivalis growth and biofilm formation. The addition of 25 F. nucleatum AI-2 enhanced monospecies biofilm development by P. gingivalis and T. 26 forsythia in a dose-dependent manner, but had no effect on T. denticola biofilm growth (Jang 27 et al. 2013). The upregulation of red complex species adhesins by F. nucleatum AI-2 enhanced 28 their coaggregation and mixed-species biofilm formation with F. nucleatum and as expected, 29 the biofilm was inhibited by QS inhibitors (Jang et al. 2013). Both S. gordonii and P. gingivalis 30 luxS deficient strains became impaired in dual-species biofilm formation; however, the biofilm 31 formation ability is restored by complementation of either of these mutant strains to confer the 32 AI-2 producing properties (McNab et al. 2003). Interestingly, Wang et al. (2011) demonstrated 33 that P. gingivalis and T. denticola inactivate S. mutans CSP, which in turn disrupts S. mutans

1 CSP stimulating properties and increases its susceptibility to environmental stress factors 2 (Wang et al. 2011). This finding suggested that periodontopathogens exploit the QS 3 mechanisms employed by S. mutans to antagonize its colonisation. 4 Although P. gingivalis lacks the orthologous genes for AHL-mediated quorum sensing, 5 it does possess a LuxR homologue that generally pairs with LuxI AHL synthase as part of the 6 regulatory circuit. Investigation of P. gingivalis composite regulatory pathways reveals that 7 this LuxR homologue, annotated as CdhR, suppresses the transcriptional expression of minor 8 fimbrillin (mfa) and luxS by direct binding to their respective promoter regions (Chawla et al. 9 2010). This interaction resulted in restricted heterotypic biofilm development between P. 10 gingivalis and S. gordonii (Chawla et al. 2010). Although the complete regulatory mechanisms 11 have not yet been decoded, tyrosine phosphatase (Ltp1) is recognized as an essential 12 component that is required for the initiation of the entire regulatory process (Chawla et al. 13 2010). AHL analogues have been found to inhibit P. gingivalis biofilm formation (Asahi et al. 14 2010, Asahi et al. 2012). These findings suggest that P. gingivalis might have additional 15 mechanisms for the regulation of gene expression or altered AHL-like signalling pathways. 16 Intercellular communications are critical for the regulation of bacterial phenotypic 17 characteristics, which include adhesin production, biofilm formation, bacterial colonisation, 18 metabolic interactions and virulence factor expression (Miller and Bassler 2001, Kolenbrander 19 et al. 2002, Henke and Bassler 2004, Guo et al. 2014). However, there are many factors to be 20 considered during investigation of signalling molecules and their influence on bacterial 21 community development as their varying properties and efficiencies rely on their molecular 22 specificity, relative distance, local and global concentrations to initiate specific responses. 23 While some have argued that AI-2 signalling might not be a true representative of bacterial 24 signalling communication due to the involvement of AI-2 and LuxS in metabolic pathways 25 (Diggle et al. 2007), other studies open up more possibilities of AI-2 molecule being a mediator 26 in interspecies communication that can affect bacterial synergistic and antagonistic interactions 27 (Kuramitsu et al. 2007, Guo et al. 2014). 28 In conclusion, signalling communication is crucial for the establishment of multispecies 29 bacterial communities and manipulation of signalling molecules could affect the oral microbial 30 ecological equilibrium. 31

1 1.4 Porphyromonas gingivalis 2 P. gingivalis is a nonmotile, black pigmented coccobacillus that is readily cultivated on 3 blood agar and protein-rich medium under anaerobic conditions. Being recognized as an 4 aetiologic agent of chronic periodontitis, P. gingivalis possesses a number of virulence factors, 5 such as proteases, outer membrane vesicles, fimbriae, hemagglutinins and cytotoxic metabolic 6 end products that are found to be detrimental to host cells (Holt et al. 1999, Nobuhiro 2000, 7 Kah Yan et al. 2016). It has recently been labelled as a “keystone pathogen” owing to its ability 8 to modulate host immune mechanisms leading to an imbalance of the immunological barrier 9 when in low abundance (Hajishengallis et al. 2012). The ability of P. gingivalis to subvert host 10 immune defence systems can in turn create an environment that selects for the proliferation of 11 pathobionts and restricts the growth of commensals during the initiation and progression of 12 chronic periodontitis (Hajishengallis and Lamont 2014).

13 1.4.1 Gingipains 14 The arginine specific cysteine proteinases, RgpA and RgpB, and the lysine-specific 15 cysteine proteinase, Kgp, collectively known as the gingipains, are the major virulence factors 16 of P. gingivalis. The translated gingipains are composed of a signal peptide, propeptide, 17 catalytic domain, immunoglobulin (IgG)-like domain, hemagglutinin domains (absent in 18 RgpB) with adhesin-binding motifs found between the domains and C-terminal domain (CTD). 19 Proteolytic processing of the RgpA, RgpB and Kgp polypeptides by sequential cleavage of 20 propeptides, catalytic and hemagglutinin domains is important for the maturation of the 21 gingipains. The processed proteinase and hemagglutinin domains of RgpA and Kgp remain 22 non-covalently bound as a large proteinase adhesin complex on the cell surface (Guo et al. 23 2010, Potempa et al. 2000). The matured RgpB is found largely as a membrane-associated 24 protein, which migrates through SDS-PAGE gel as a series of different size bands. This size 25 difference is due to differences in post-translational modification by the addition of glucan 26 moieties onto the structures for membrane attachment, whilst a discrete C-terminally truncated 27 soluble RgpB may also be found in the culture supernatant (Guo et al. 2010, Potempa et al. 28 2000, Seers et al. 2006). 29 These enzymes account for at least 85% of P. gingivalis proteolytic activities (Potempa 30 et al. 1997). The cell- and vesicle-associated gingipains function as endopeptidases, which are 31 responsible for the degradation of host proteins. Metabolically, these gingipains cleave a range 32 of host structural proteins into smaller peptides, as well as the hydrolysis of transferrin and 33 hemoglobin for release of the essential micronutrients hemin and iron (Lamont and Jenkinson 10

1 1998, Dashper et al. 2004). In addition, gingipains also contribute to invasion of host epithelial 2 tissues, as well as inactivation of host protease inhibitors leading to the exacerbation of host 3 and bacterial proteolytic activities (Nakayama et al. 1996, Potempa et al. 2000, Sugawara et 4 al. 2000b, Potempa et al. 2003, Wilensky et al. 2015). These proteases also participate in the 5 post-translational proteolytic processing of cellular components i.e. maturation of fimbriae, and 6 activations of other proteases and collagenases (Potempa et al. 2003, Guo et al. 2010). Several 7 lines of evidence also showed that gingipains are important for the evasion and dysregulation 8 of host immune defence responses by dysregulation of the cytokine network (Imamura 2003), 9 degradation of monocyte CD14 receptor (Sugawara et al. 2000a) and disturbance of neutrophil 10 migration to P. gingivalis at the infection site, thus protecting the bacterium against 11 phagocytosis (Jagels et al. 1996). The proteolytic cleavage of host proteinase inhibitors also 12 lead to the dysregulation of host proteases, which contribute to an uncontrollable activation of 13 inflammatory responses (Potempa et al. 2003, Guo et al. 2010). 14 In situ metatranscriptomic analysis confirmed the over-expression of these proteinase 15 genes at periodontal disease sites, in comparison to the healthy sites, which inferred their 16 pathogenic roles in aggravating disease conditions (Szafrański et al. 2015). As gingipains are 17 regarded as the key virulence factors of P. gingivalis (Imamura 2003, O'Brien-Simpson et al. 18 2003, Potempa et al. 2003, Guo et al. 2010), they have been presented as suitable diagnostic 19 biomarkers of disease severity, as well as attractive drug and vaccination targets for the 20 treatment and prevention of P. gingivalis associated chronic periodontitis (Yasaki-Inagaki et 21 al. 2006, Jong and van der Reijden 2010, O’Brien-Simpson et al. 2016).

22 1.4.2 Outer membrane vesicles 23 P. gingivalis outer membrane vesicles (OMVs) are spherical single bilayer membranes 24 containing bioactive components that are formed by the blebbing of the outer membrane (Gui 25 et al. 2016, Kulp and Kuehn. 2010). P. gingivalis virulence factors, which include antigenic 26 lipopolysaccharide (LPS), lipoproteins, proteases, fimbriae and hemagglutinins that are from 27 the periplasm or outer membrane are found to be enriched in the OMVs (Veith et al. 2014). A 28 recent study also showed that DNA was found in P. gingivalis OMVs, which can be carried 29 into host cells (Bitto et al. 2017). OMVs from P. gingivalis have been shown to modulate 30 immune responses, transmit virulence factors and lead to impairment and destruction of host 31 cells (Nakao et al. 2011, Xie 2015, Cecil et al. 2016, Waller et al. 2016). Moreover, P. 32 gingivalis vesicles can also mediate bacterial interspecies interactions. OMVs of P. gingivalis 33 have been found to aggregate with the antecedent colonisers, such as streptococci, A. 11

1 naeslundii, Actinomyces viscosus and F. nucleatum (Hiratsuka et al. 2008), as well as facilitate 2 the adherence of P. gingivalis to T. forsythia and aid T. forsythia in the attachment and invasion 3 of epithelial cells (Inagaki et al. 2006). This observation also invited speculation that OMVs 4 not only mediate adherence of interspecies bacteria, but also allow the acquisition of additional 5 phenotypes. The ability of OMVs released by P. gingivalis to bind to antibacterial agents i.e. 6 chlorhexidine, suggested that OMVs could confer additional resistance and protection 7 mechanisms to other subgingival bacteria (Grenier et al. 1995). The production and 8 composition of P. gingivalis OMVs can be modulated under different conditions, such as 9 during growth in heme-limiting conditions and when the cells enter stationary phase (Veith et 10 al. 2018). The release of bacterial OMVs that are enriched with nutrient-scavenging factors 11 and proteolytic enzymes could help increase their nutrient acquisition ability especially at 12 distant, small and immunologically restricted sites (Kulp and Kuehn 2010, Gui et al. 2016), 13 which would otherwise be inaccessible to bacterial cells.

14 1.4.3 Mediators of physical interactions 15 P. gingivalis long fimbriae FimA, gingipains and hemagglutinin domains mediate 16 attachment of P. gingivalis to a variety of host substrates i.e. epithelial cells, salivary 17 components, fibrinogen, fibronectin, glycoproteins and proline-rich proteins (Lamont and 18 Jenkinson 1998, Kataoka et al. 1997, Chen et al. 2001). Studies demonstrate physical 19 interactions of P. gingivalis with antecedent oral bacterial colonisers, such as Streptococcus, 20 Actinomyces, F. nucleatum and Prevotella intermedia that ensure its incorporation into 21 polymicrobial communities (Kamaguchi et al. 2001, Nagayama et al. 2001); whilst achieving 22 mutualistic growth and heterotypic biofilm formation with specific early colonising partners. 23 P. gingivalis long fimbriae have been reported as an essential adhesin that mediates 24 coaggregation to Streptococcus spp. and Actinomyces spp., as well as other pathobionts, 25 including T. denticola (Hashimoto et al. 2003, Enersen et al. 2013). For example, P. gingivalis 26 FimA is a 43 kDa long fimbrial major subunit protein that exhibits binding affinity for the 27 glyceraldehyde-3-phosphate dehydrogenase (GAPDH) of Streptococcus oralis with an

28 association constant (Ka) that is ten-fold higher than for the host components tested, which 29 included haemoglobin and extracellular matrix proteins i.e. laminin, fibronectin, elastin, 30 fibrinogen and vitronectin (Nakamura et al. 1999, Amano 2003, Maeda et al. 2004). Due to the 31 conservation of the GAPDH sequence, many streptococci with high GAPDH activities, such 32 as S. gordonii, Streptococcus sanguinis and Streptococcus parasanguinis are found to 33 coaggregate with P. gingivalis FimA relatively strongly (Maeda et al. 2004b). In addition, P. 12

1 gingivalis and streptococci physical association is assisted by additional pair-wise ligands and 2 receptors to secure their intimate interactions. The 67 kDa minor fimbriae (Mfa) expressed on 3 P. gingivalis showed strong binding affinity and specificity towards a 26 amino acid epitope, 4 known as the BAR motif, on SspA/B adhesins of Streptococcus species (Demuth et al. 2001, 5 Lamont et al. 2002). Mfa fimbriae are also important for autoaggregation and microcolony 6 development, ensuring the attachment of P. gingivalis to substratum via major fimbriae (Lin et 7 al. 2006). Molecular inhibitors targeting mfa1 and fimA expression successfully reduced the 8 accumulation of P. gingivalis onto a S. gordonii substratum (Wright et al. 2014). These studies 9 indicated that P. gingivalis is closely associated with S. gordonii, which is consistent with the 10 detection of P. gingivalis in the presence of S. gordonii in a heterotypic biofilm (Kuboniwa et 11 al. 2006). Interestingly, P. gingivalis fimA transcription is repressed upon contact with 12 Streptococcus cristatus arginine deiminase (ArcA), which inhibits P. gingivalis surface 13 attachment and heterotypic biofilm formation (Wu and Xie 2010, Cugini et al. 2013). This 14 phenomenon has also been suggested as an intergeneric bacterial contact-dependent 15 communication that negatively regulates P. gingivalis biofilm formation. The negative 16 correlation of their distribution in the subgingival region also suggests that S. cristatus is an 17 antagonistic partner for P. gingivalis, while the ArcA of S. cristatus could be a potential 18 candidate inhibitor of P. gingivalis biofilm formation (Wang et al. 2009). 19 In addition, P. gingivalis hemagglutinin (HagA) is a surface adhesin protein that is 20 involved in hemagglutination and the binding of hemoglobin. Repeated adhesin domains found 21 C-terminal to the catalytic domains of the gingipains Kgp and RgpA, exhibit sequence 22 similarity with HagA (Veith et al. 2002). These domains have been referred to as cleaved 23 adhesin domains and DUF 2436 (Nan et al. 2010, Dashper et al. 2017). A P. gingivalis triple 24 proteinase mutant rgpA-rgpB- kgp-, the use of amino acids arginine and lysine or the combined 25 action of Rgp-specific and Kgp-specific inhibitors completely abolished P. gingivalis 26 coaggregation activities with other bacteria, suggesting that gingipains were involved in the 27 adhesion (Abe et al. 2004). However, the attenuation of P. gingivalis coaggregation and biofilm 28 formation via mutation of the gingipains was largely due to the requirement of the gingipains 29 for proteolytic processing and presentation of its adhesion factor (Abe et al. 2004, Ito et al. 30 2010). Other studies specifically showed that RgpB, but not RgpA and Kgp, was responsible 31 for the binding of P. gingivalis to S. gordonii and A. naeslundii (Tokuda et al. 1996), as well 32 as biofilm formation with T. denticola (Yamada et al. 2005). However, the coaggregation 33 activities between P. gingivalis rgpB- with T. denticola were unaffected (Yamada et al. 2005).

1 Detailed investigation revealed that the 44 kDa Hgp44 adhesin domains found in RgpA, Kgp 2 and hemagglutinin A (HagA) are important factors for the coaggregation of P. gingivalis with 3 T. denticola (Ito et al. 2010). 4 P. gingivalis HmuR serves as a dual-function protein that acts as a major hemin uptake 5 protein and adhesin for three-species community development between P. gingivalis, S. 6 gordonii and F. nucleatum, but was unimportant for mono- or dual-species biofilm formation 7 (Kuboniwa et al. 2009). Hbp35 is another multifunctional hemin binding protein that 8 contributes to hemagglutination, autoaggregation and coaggregation with A. viscosus, as well 9 as allowing P. gingivalis attachment to host gingival epithelial cells (Hiratsuka et al. 1992, 10 Shibata et al. 2003, Hiratsuka et al. 2010). The mutation of hbp35 influenced other binding 11 factors i.e. impaired OMVs aggregation activity, reduced rgpA, rgpB, kgp expression, which 12 in turn negatively affected fimbriae maturation (Hiratsuka et al. 2008, Hiratsuka et al. 2010). 13 Taken together, the above-mentioned findings show that P. gingivalis possesses multivalent 14 ligands of long fimbriae, minor fimbriae, the gingipains, HmuR and Hbp35 that perform 15 multifunctional and combinatory activities in securing P. gingivalis interactions and integration 16 within the multispecies biofilm.

17 1.4.4 Roles of proteases in P. gingivalis 18 P. gingivalis is asaccharolytic and derives energy from the fermentation of peptides and 19 amino acids. Having highly active proteolytic enzymes is selectively advantageous for P. 20 gingivalis given the nature of its habitat, which consists of abundant proteinaceous substrates 21 (Takahashi et al. 2000). During disease progression, dysregulation of the host immune response 22 causes host proteinases to work in concert with the pathobiont proteinases to break down 23 proteins into oligopeptides, which are in turn further hydrolysed into di- or tripeptides and 24 amino acids for bacterial assimilation and growth. P. gingivalis has been demonstrated to 25 degrade a wide array of host substances, ranging from extracellular matrix, serum albumin, 26 collagen, keratin, fibronectin, fibrinogen and immunoglobulin (Lamont and Jenkinson 1998, 27 Holt and Ebersole 2005, Guo et al. 2010). Although 42 genes predicted to encode proteases 28 were identified in P. gingivalis W83 via whole genome sequencing (Nelson et al. 2003), many 29 of them remain uncharacterised. Several P. gingivalis cysteine proteinases (periodontain, Tpr, 30 PrtH, PrtT, PrtC, RgpA, RgpB, Kgp) and serine proteinases mainly composed of the 31 dipeptidylpeptidases (DPPs), which include DPPIII, DPPIV, DPP5, DPP7, DPP11 and prolyl 32 tripeptidylpeptidase A (PtpA), have been characterised.

1 In addition to the gingipains, other P. gingivalis cysteine proteinases are also likely to 2 contribute to its pathogenesis. P. gingivalis thiol protease (Tpr) was demonstrated to hydrolyse 3 azocoll, casein, fibrinogen, gelatin, bovine serum albumin, heat-treated collagen, complement 4 proteins and LL-37 antimicrobial peptide, which suggested its role in P. gingivalis nutrient 5 acquisition and disruption of the host immune system (Bourgeau et al. 1992, Potempa et al. 6 2000, Staniec et al. 2015). The upregulation of tpr expression during growth in a peptide- 7 limited medium was observed, likewise, the supplementation of peptides, but not free amino 8 acids, suppressed the transcriptional expression of tpr (Lu and McBride 1998). 9 Purified recombinant PrtT exhibited optimal hydrolytic activities on casein and gelatin 10 in the presence of reducing agents and Ca2+ ions at pH 8.5 (Otogoto and Kuramitsu 1993). The 11 minor proteinases, PrtT and Tpr, were found to be weakly expressed during normal P. 12 gingivalis growth, however the transcriptional upregulation of prtT in an rgpA mutant strain 13 demonstrated the potential to compensate for the loss of RgpA activities. In addition, the 14 expression of prtT and rgpA were reciprocally regulated, with prtT expression down-regulated 15 under iron and nutrient depleted conditions (Tokuda et al. 1998). 16 Periodontain, which shares significant amino acid sequence identity with S. pyogenes 17 streptopain and PrtT, plays a role in inactivating the host proteinase inhibitor serpin; thus, 18 preventing the regulation of human neutrophils elastase activity and aggravates tissue damage 19 (Nelson et al. 1999). However, the periodontain substrate specificity has not been identified. 20 The PrtC collagenase has been shown to cleave phenylazobenzyl-oxycarbonyl (Pz)- 21 Pro-Leu-Gly-Pro peptide and soluble Type 1 collagen. Early studies predicted that the prtC 22 gene was cotranscribed with prtT based on Northern blot analysis (Takahashi et al. 1991, 23 Otogoto and Kuramitsu 1993), which led to the postulation that PrtC could work cooperatively 24 with PrtT to degrade collagen. However, P. gingivalis operon and regulon predictions 25 suggested that prtC is cotranscribed with genes encoding metabolic biosynthesis enzymes 26 involved in the purine, cofactors and coenzyme metabolic processes (Alm et al. 2004, Arkin 27 2011). 28 A series of DPPs and tripeptidylpeptidases are proposed to work in combination with 29 endopeptidases and proteinases to ensure the efficiency of di- and tripeptide assimilation for 30 bacterial growth. To date, five DPPs: DPPIII, DPPIV, DPP5, DPP7 and DPP11 have been 31 characterised from P. gingivalis, yet bioinformatics and considerable experimental evidence 32 indicate that there are other peptidases in operation.

1 P. gingivalis DPPIII is localized in the cytoplasm and belongs to the M49 2 metallopeptidase family with a conserved zinc-binding motif (Ohara-Nemoto et al. 2014). P. 3 gingivalis recombinant DPPIII cleaves and releases dipeptides with preference for Arg in the 4 P1 position from the N-termini of oligopeptides (Ohara-Nemoto et al. 2014), as well as 5 hydrolysing peptide fragments from bioactive proteins (Hromić-Jahjefendić et al. 2017). 6 Distinct from other DPPIII orthologues is the presence of an Armadillo (ARM) type fold at the 7 C-terminus of P. gingivalis DPPIII (Hromić-Jahjefendić et al. 2017). This ARM repeat domain 8 is also presents in the AlkD family of alkylpurine bacterial DNA glycosylases that are involved 9 in DNA repair. However, the roles of the DPPIII ARM repeat domain remain unknown as it 10 could not complement the alkylation repair function of a DNA repair-deficient E. coli, yet the 11 absence of ARM repeat domain in recombinant P. gingivalis DPPIII resulted in a decreased 12 catalytic efficiency (Hromić-Jahjefendić et al. 2017). 13 DPPIV is a surface associated serine proteinase, previously known as the glycylprolyl 14 peptidase that specifically cleaves at a proline or alanine residue in the second position from 15 the amino terminus. DPPIV has been shown to partially hydrolyse collagen type I for the 16 release of Gly-Pro dipeptides, degrade biologically active peptides β-casomorphin and the N- 17 terminal region of IL-1β and IL-2 (Banbula et al. 2000). Although DPPIV does not exhibit 18 hydrolytic activities towards connective tissues, it enhances the activities of host-derived 19 gelatinase and collagenases to degrade type I collagen, which later becomes susceptible to 20 further degradation (Kumagai et al. 2005). Also, DPPIV plays a role in mediating the binding 21 of P. gingivalis to fibronectin for colonisation of disease sites, while blocking the adhesion of 22 human gingival fibroblasts to fibronectin (Kumagai et al. 2005). This could perturb the 23 recovery process at the inflammatory site. 24 DPP7 is a novel type of DPP found in P. gingivalis that preferentially cleaves both 25 aliphatic (AVILM except GP) and aromatic (FYW) residues at the penultimate position from 26 the amino-terminus (Banbula et al. 2001). DPP7 also releases dipeptides with a non- 27 hydrophobic P1 residue if the N-terminal (P2) residue is hydrophobic (Rouf et al. 2013). More 28 recently, DPP5 that was previously annotated as a prolyl oligopeptidase, yet did not show an 29 ability to hydrolyse proline-containing peptides, demonstrated hydrolytic activities towards 30 hydrophobic amino acids and Ala at the P1 position. Its activities could complement the 31 proteolytic activities of DPP7 (Ohara-Nemoto et al. 2014). However, there is a slight variation 32 in their specificity towards different peptides, where DPP5 has no preference at P2 but DPP7 33 exhibits hydrophobic preferences at the P1 and P2 positions. Interestingly, DPP5 is located in

1 the periplasm, which has been suggested to maximize the efficiency of dipeptides binding to 2 the transporters localized in the inner membrane (Ohara-Nemoto et al. 2014). DPP11 possesses

3 the specificity NH2-Yaa-D/E -↓-(Xaa)n, where Xaa represents any amino acid and Yaa is any 4 amino acid, except Pro. DPP11 is speculated to be important for the release of 5 glutamylglutamate and aspartylaspartate for P. gingivalis utilisation (Ohara-Nemoto et al. 6 2011). 7 The surface located P. gingivalis prolyl tripeptidylpeptidase A (PtpA) is responsible for 8 the liberation of tripeptides with proline in the third position from the amino-terminus, but not 9 when Pro is also at the fourth position. For example, synthetic oligopeptides, interleukin 6 and

10 cystatin C with the NH2-Xaa-Xaa-Pro-Yaa sequence can be degraded by PtpA (Banbula et al. 11 1999). Other carboxypeptidases and aminopeptidases that are mostly localized in the cell 12 cytoplasm are important for the cleavage of internalized di- and tripeptides into free amino 13 acids i.e. arginine/lysine metallocarboxypeptidase is responsible for the release of arginine and 14 lysine for bacterial growth (Chen et al. 2002, Masuda et al. 2002). 15 P. gingivalis has been shown to increase the rate of peptide hydrolysis and the release 16 of free glycine in cell free T. denticola conditioned medium (Tan et al. 2014). Based on this 17 observation, it is possible that the accumulation of glycine is a result of the upregulation of a 18 series of peptidases that release free glycine. This result indirectly suggests that P. gingivalis 19 proteinase activities might be nutritionally beneficial to other oral bacteria and play an 20 important role in supporting the propagation of other bacteria by degradation of complex 21 proteins. 22 Cooperative actions of P. gingivalis proteases assist in complex protein degradation to 23 maximize the efficiency of peptide utilisation. A DPPIV, DPP7, PtpA triple mutant and a 24 DPPIV-5-7-11 quadruple mutant showed retardation in growth, suggesting that DPPs and PtpA 25 play roles in nutrient acquisition (Oda et al. 2009, Ohara-Nemoto et al. 2014). Although P. 26 gingivalis has not been reported to utilise proline or proline-containing peptides (Takahashi et 27 al. 2000), DPPIV and PtpA roles in cleaving N-terminal proline/hydroxylproline-containing 28 peptides cannot be discounted as they help increase the accessibility of peptides that would 29 otherwise be resistant to other proteases. Interestingly, the hydrolysis of Met-Leu-4- 30 Methylcoumaryl-7-amide by the DPPIV-5-7-11 quadruple mutant was higher than the wild 31 type strain, which suggests the presence of unknown peptidase(s) to compensate the loss of the 32 four DPPs (Ohara-Nemoto et al. 2014). Although many other P. gingivalis proteinases remain 33 uncharacterised biochemically and functionally, their activities are conceivably

1 complementary and maximize the efficiency of polypeptide degradation for nutrient 2 acquisition. The proteases may exhibit complementary activities under different environmental 3 conditions, such as variation in salt concentrations, nutrient availability and pH. There is also 4 the possibility that the minor proteases are only induced under certain environmental conditions 5 making identification of their physiological roles difficult. 6 The wide range and activity of the peptidases coupled with the action of high 7 specificity transporters for the uptake of di- and tripeptides, which are subsequently broken 8 down into free amino acids intracellularly for metabolism, is an important indication of the 9 demands of P. gingivalis for peptides and amino acids for survival and growth. In addition, P. 10 gingivalis proteases confer upon the bacterium growth-promoting traits under inflammatory 11 conditions, which in turn allow the bacteria to thrive as a predominant member in the dysbiotic 12 subgingival plaque. This strategy provides a competitive advantage and improves fitness of P. 13 gingivalis in the polymicrobial oral environment, especially when external peptides become 14 scarce. Although the proteolytic system of P. gingivalis has long been regarded as one of its 15 major virulence factors, intense research efforts are still required to decode their properties and 16 the regulation of these proteinases in bacterial pathogenesis.

17 1.4.5 Peptide and amino acid metabolism 18 Information regarding the energy production pathways of P. gingivalis is mostly 19 available in the form of genome-scale metabolic networks informed by genomic, proteomic 20 and transcriptomic analyses (Nelson et al. 2003, Naito et al. 2008, Hendrickson et al. 2009, 21 Mazumdar et al. 2009, Høvik et al. 2012). However, few studies have performed the necessary 22 enzymatic and physiological testing required to confirm their metabolic properties. P. 23 gingivalis preferentially utilises peptides over amino acids for nutrition, cell materials and 24 energy production. Elevated levels of peptides in diseased periodontal sites supports the notion 25 that extracellular proteases could be responsible for the accumulation of peptides in the gingival 26 crevice (Barnes et al. 2011). The aforementioned extensive protease activities will release a 27 wide range of di- and tripeptide metabolic substrates. The peptide substrates tested in the forms 28 of glutamylglutamate and aspartylaspartate favour the growth of P. gingivalis (Takahashi et al. 29 2000). Central metabolic pathways for glutamate and aspartate metabolism based on metabolic 30 enzyme activities, stoichiometric calculations and proteomic analyses have been proposed 31 (Takahashi et al. 2000, Dashper et al. 2009). Takahashi et al. (2000) also showed that P. 32 gingivalis generates butyrate from aspartate and succinate from glutamate, while ammonia, 33 acetate and propionate are common end products for both substrates (Fig. 1.2). Following the 18

1 above findings, the authors continued to define the peptides preferentially utilised by P. 2 gingivalis. Dipeptides were the preferred substrates, although tetrapeptide utilisation also 3 generated a substantial amount of metabolic end products, which could be due to the hydrolysis 4 of tetrapeptides to dipeptides (Takahashi and Sato 2001). It should be noted that identifying 5 the potential proteases responsible for the release of these dipeptides would give a better 6 understanding of the P. gingivalis nutrient acquisition strategy. Valylvaline and leucylleucine 7 can be metabolized into isobutyrate and isovalerate respectively and other common end 8 products, provided that glutamylglutamate and aspartylaspartate were present (Takahashi and 9 Sato 2001). This suggests metabolic linkage within the peptide utilisation process. 10 An earlier study conducted for the design of P. gingivalis chemically defined media 11 showed that the addition of free amino acids alone failed to support the growth of P. gingivalis, 12 and the lack of evidence of amino acid transport systems led to the controversial statement that 13 P. gingivalis does not utilise free amino acids (Milner et al. 1996). Instead, peptides are 14 required for its growth. Given the fastidious nutrient requirements for P. gingivalis growth and 15 the complexity of media rich in proteins, peptides and amino acids, the utilisation of amino 16 acids and peptides could be challenging to elucidate. In particular, free serine, threonine, 17 arginine, aspartate/asparagine, glutamate/glutamine and leucine have been demonstrated to be 18 consumed by P. gingivalis, although it is conceivable that peptides are required as the major 19 energy sources (Dashper et al. 2001). Many studies agreed that P. gingivalis does not utilise 20 sulphur-containing amino acids and proline. However, cysteine is considered essential for 21 many cysteine proteinases and sodium ion-motive force driven serine/threonine transport 22 (Chen et al. 1992, Dashper et al. 2001). Upon utilisation of peptides and amino acids, P. 23 gingivalis produces cytotoxic end products, especially ammonia, butyrate and propionate in 24 millimolar concentrations that adversely affect host cell activities and immune defence 25 mechanisms (Takahashi et al. 2000).

1 2 Figure 1.2: P. gingivalis amino acid metabolism. Image reproduced from Meuric et al. 2010.

3 1.5 Treponema denticola 4 T. denticola is a helically-shaped anaerobic bacterium that is the most well 5 characterised of the oral spirochetes due to being more readily cultivated and amenable to 6 genetic manipulation, than other Treponema spp. commonly isolated from the subgingival 7 plaque of periodontitis patients, such as T. socranskii, T. vincentii and T. pectinovorum (Moter 8 et al. 1998, Dewhirst et al. 2000, Asai et al. 2002). T. denticola possesses outer membrane 9 structures distinct from Gram negative bacteria as it lacks LPS, instead containing lipoteichoic 10 acids that are found in Gram positive bacteria (Ishihara and Okuda 1999, Holt and Ebersole 11 2005). Thus its outer cellular region is known as an outer sheath. T. denticola increases from 12 <1% of the total bacterial population of the subgingival community in a healthy individual to 13 become a highly prominent member of the pathogenic subgingival microbial community (Chan 14 and McLaughlin 2000, Visser and Ellen 2011). This suggests that growth of T. denticola is 15 favourable under a dysbiotic environment and/or guided migration to favorable sites supports 16 T. denticola proliferation. As it is known to be associated with chronic periodontitis, the 17 virulence factors of T. denticola have been the subject of a number of reviews (Fenno and 18 McBride 1998, Holt and Ebersole 2005, Ishihara 2010, Dashper et al. 2011, Visser and Ellen 19 2011).

20 1.5.1 Chemotaxis and motility 21 Based on the Human Oral Microbiome Database (HOMD), approximately 11% (43 out 22 of 405 oral bacterial taxa with annotated genomes) of oral bacterial genomes were identified 20

1 to encode the chemotaxis core components, CheA, CheW and transmembrane receptors (Erbse 2 and Falke 2009, Chen et al. 2010), suggesting that chemotaxis is not a rare physiological 3 property exhibited in oral polymicrobial biofilms. Community-wide transcriptomic and 4 genomic profiling showed an over-representation of chemotaxis and motility genes in the 5 enrichment functional analysis from periodontitis samples compared to healthy samples (Wang 6 et al. 2013, Duran-Pinedo et al. 2014). These studies highlight the importance of bacterial 7 chemotaxis and motility as potential virulence factors in the pathogenic subgingival microbial 8 community. 9 Two percent of the T. denticola genome encodes putative chemotaxis and flagellar 10 related proteins (Seshadri et al. 2004). The central chemotaxis pathway of T. denticola consists 11 of histidine kinase (CheA), accessory protein (CheW), phosphatase (CheX) and motor response 12 regulator (CheY). The genes encoding these proteins were shown to be cotranscribed in an 13 operon (Greene and Stamm 1999, Sim et al. 2005). In addition, methyltransferase (CheR) and 14 phosphatase (CheX) are also found as the regulators of T. denticola chemotaxis pathways, as 15 shown in Fig. 1.3. There are also at least 20 methyl accepting chemotaxis protein (MCP) 16 complexes that are coupled with specific periplasmic binding proteins or bind directly to 17 ligands for the sensing of environmental stimuli. As T. denticola is attracted towards glucose 18 and serum, especially its albumin component, this may allow translocation of T. denticola to 19 nutrient-rich sites according to the chemotactic substrate gradient concentration (Umemoto et 20 al. 2001, Ruby et al. 2008). Experimental analyses of a T. denticola cheA mutant showed a 21 reduction in its swarming motility and impairment in its ability to swim towards nutrient 22 sources (Lux et al. 2002). Mutation of the identified MCP-encoding genes, dmcA and dmcB, 23 resulted in a decrease in T. denticola chemoattraction towards nutrient sources, especially 24 serum. In addition, inactivation of dmcA and dmcB was also shown to affect the methylation 25 profile of proteins, in comparison to wild type (Kataoka et al. 1997, Li et al. 1999). 26 T. denticola periplasmic flagella (PF) are located within the periplasmic space, and are 27 composed of a basal body-motor complex, flagellar hook (FlgE) and flagellar filaments that 28 are composed of a core of three flagellin proteins, FlaB1, FlaB2 and FlaB3 covered with a 29 sheath of FlaA proteins (Li et al. 2000, Charon and Goldstein 2002, Limberger 2004, Charon 30 et al. 2012). Unlike externally flagellated motile bacteria, the PF of T. denticola are important 31 in shaping the cellular corkscrew morphology of this organism and its localisation within the 32 cell body confers protection from host immunological detection (Cockayne et al. 1989). They 33 also enable the bacterium to swim in highly viscous environments (Ruby et al. 1997). The

1 response regulator of the chemotaxis pathways, CheY, is coupled to the T. denticola motility 2 system by interacting with the FliM switch protein, which regulates the rotational direction of 3 the PF (Fig. 1.3; Lux et al. 2000, Sim et al. 2005). In vitro studies illustrated that T. denticola 4 chemotaxis is important for the guidance of swarming motility, penetration of host epithelial 5 cells and invasion of mammalian host tissues (Lux et al. 2001, Lux et al. 2002, Charon et al. 6 2012). A T. denticola mutant defective in the flagellar hook protein FlgE showed attenuation 7 in both its ability to coaggregate with P. gingivalis and form mixed-species biofilm (Vesey and 8 Kuramitsu 2004, Yamada et al. 2005). The significant attenuation of the T. denticola flgE 9 mutant in biofilm formation might be due to the loss of its ability to swarm or migrate through 10 a viscous environment (O'Toole et al. 2000). In addition, investigations into T. denticola 11 flagellin proteins showed that their expression levels, post-translational modifications and their 12 length can affect the motility rate of T. denticola strains, which in turn demonstrated that lower 13 motility strains achieved higher bacterial cell densities, protease activities and more 14 associations with gingival epithelial cells (Nagano et al. 2017) 15 Given that T. denticola is a highly fastidious bacterium with regard to nutrient 16 requirements and growth conditions, chemotaxis and motility could be essential characteristics 17 for T. denticola to persist in a highly populated and competitive environment. 18 Metatranscriptomic analyses, along with phenotypic characterisation of T. denticola 19 chemotaxis and motility mutant strains, support the conjecture that T. denticola chemotaxis 20 and motility phenotypes support T. denticola virulence (Wang et al. 2013, Duran-Pinedo et al. 21 2014, Lux et al. 2002, Charon et al. 2012). In addition, the ability of T. denticola to sense and 22 migrate to favourable environments suggests ecological and competitive advantages for T. 23 denticola. It is also hypothesized that T. denticola chemotaxis and motility phenotypes mediate 24 T. denticola movement towards synergistic partners and away from antagonistic partners, thus 25 enabling interspecies interactions of T. denticola within the oral polymicrobial community. 26 Likewise, these phenotypic properties might also be essential for the expansion of 27 polymicrobial biofilm development by creating transient pore openings to increase nutrient 28 flow into heterotypic biofilm microcolonies.

1 2 Figure 1.3: Schematic diagram of the chemotactic components and signalling pathways of T. 3 denticola. MCP sensory receptors are responsible for the perception of external stimuli, the 4 signal is then transduced to a histidine kinase (CheA), response regulator (CheY) and 5 methylation regulator (CheR) that allow the cell to adapt to the existing stimuli state. CheW is 6 an accessory protein of the chemotaxis complex that is associated with the dimers of CheA. 7 CheX functions as a phosphatase to restore the phosphorylated level of CheY to prestimulus 8 level. Extracted from Sim et al. 2005.

9 1.5.2 Mediators of physical interactions 10 The outer sheath components of T. denticola are essential for its integration within 11 multispecies biofilm communities, by coaggregating with F. nucleatum (Kolenbrander et al. 12 1995, Rosen et al. 2008), Prevotella intermedia, Parvimonas micra (Cogoni et al. 2012) and 13 T. forsythia (Sano et al. 2014). In addition, the outer sheath components are also involved in 14 physical interactions with other host cells and proteins, such as gingival fibroblasts and 15 extracellular proteins that exist in the form of fibronectin, laminin, gelatin and type I and II 16 collagen (Fenno 2012). 17 Major sheath protein (Msp) is one of the most abundant outer sheath proteins and is a 18 major antigenic protein that provides T. denticola with adhesive, proteolytic and cytotoxic 19 properties. It exists as a 53 kDa or 64 kDa monomer, depending on the T. denticola strain, that 20 forms a detergent-stable 150-200 kDa oligomeric complex and visible as a hexagonal array in 21 the outer surface by electron microscopy (Egli et al. 1993, Nakamura et al. 1993, Fenno et al. 22 1996). Msp is involved in coaggregation with P. gingivalis and F. nucleatum (Rosen et al. 23 2008), as well as exhibiting porin formation ability on host epithelial cells (Mathers et al.

1 1996). Msp also contributes to hemagglutination, hemooxidization and hemolysis of 2 erythrocytes (Fenno et al. 1998a). 3 Chymotryptic-like protease (CTLP) complex comprises of a 72 kDa prolyl- 4 phenylalanine-specific-protease (also known as CTLP, dentilisin and PrtP), two auxiliary 5 polypeptides of 40 kDa PrcA1 and 30 kDa PrcA2 (proteolytic processed forms of PrcA by 6 PrtP), and a 22 kDa PrcB, which is a lipoprotein that interacts with PrtP pre-proteinase and is 7 required for its enzymatic activity, and Msp (Ishihara et al. 1996, Lee et al. 2002, Godovikova 8 et al. 2010). CTLP complex is commonly known for its proteolytic and adhesive properties; 9 proteolytic activities will be described in Section 1.5.3. Bamford et al. (2007) showed that the 10 inhibition of T. denticola chymotrypsin-like activity did not abolish CTLP complex adherence 11 to fibrinogen, instead it enhanced overall binding, suggesting a different active site for CTLP 12 binding and proteolysis. Although the adherence and proteolytic domains of CTLP complex 13 have not yet been distinguished, intact CTLP complex is required for both adherence and 14 hydrolytic activities that may however function independently. CTLP complex has also been 15 implicated in T. denticola adherence to host epithelial cells (Ellen et al. 2000). In particular, T. 16 denticola CTLP complex showed binding specificity towards fibrinogen Aα and Bβ domains, 17 while P. gingivalis preferentially bound the ϒ domain, thus circumventing adherence 18 competition and allowing coexistence on the fibrinogen for metabolic synergy in fibrinogen 19 degradation, hence increasing nutrient accessibility (Bamford et al. 2007). Consequently, T. 20 denticola and P. gingivalis inhibit blood clot formation and interfere with the wound healing 21 process at periodontal sites. 22 Whilst the CTLP complex allows T. denticola binding and degradation of fibrinogen, 23 Msp also demonstrated a minor contributory role to these properties as T. denticola deficient 24 in Msp production showed a reduction in fibrinogen binding and proteolysis, suggesting that 25 Msp and CTLP complex could be acting synergistically (Edwards et al. 2005, Bamford et al. 26 2007). Although the molecular mechanisms of Msp and CTLP complex formation have not 27 been completely elucidated, the protein expression of CTLP and Msp is interdependent in order 28 to maintain the stability and oligomerisation of the outer membrane complexes (Lee et al. 29 2002). A msp mutant strain that produced truncated Msp and lower Msp expression level 30 resulted in the loss of CTLP complex (Fenno et al. 1998b); likewise, mutation of prcA or prtP 31 caused a reduction in msp transcription and affected the localisation and structural organization 32 of Msp (Lee et al. 2002, Bian et al. 2005). Further characterisation of the CTLP complex 33 revealed that the abundance of PrcB is at a consistent level with PrtP, while the interaction of

1 PrcA2 with Msp may be responsible for the generation of a high-molecular weight outer 2 membrane complex consisting of Msp and the CTLP complex (Godovikova et al. 2011). Also, 3 it is important to note that the antigenicity of Msp and CTLP complex vary remarkably across 4 different T. denticola strains, hence giving rise to distinct specificity and efficiency in host 5 protein binding and heterotypic biofilm formation. Therefore, the pore-forming Msp and CTLP 6 complex of T. denticola that exhibit adhesive and proteolytic activities are recognized as major 7 virulence factors (Ishihara 2010, Dashper et al. 2011). 8 Other T. denticola surface proteins that exhibit adhesive properties include BspA, 9 TDE2508 and OppA. Ikegami et al. (2004) reported that the T. denticola surface associated 10 leucine-rich repeat protein, now annotated as BspA (TDE2258), allowed coaggregation with 11 T. forsythia, attachment to human epithelial cells and tissue penetration (Ikegami et al. 2004). 12 Following mutation of bspA, T. denticola became deficient in swarming ability, which might 13 reduce biofilm formation but this has not been verified experimentally. Recently, TDE2508 14 was identified as a major outer membrane protein that might be able to regulate T. denticola 15 adhesive properties as a mutant strain demonstrated an enhancement in both biofilm formation 16 and adherence to gingival epithelial cells (Abiko et al. 2014). Likewise, the T. denticola ATP- 17 binding cassette-type (ABC) peptide transporter OppA (TDE1071) is an antigenic protein 18 expressed on the T. denticola cell surface that binds soluble fibronectin and plasminogen, but 19 not immobilized substrates, and may be important for the uptake of peptide nutrients (Fenno et 20 al. 2000). 21 The inner membrane anchored cytoplasmic filaments are composed of the CfpA protein 22 and run parallel to the tightly organized flagellar filaments (Izard et al. 2004, Izard et al. 2008). 23 CfpA has a major role in cell division and was also shown to have an effect on T. denticola 24 swarming (Izard et al. 2001). In addition, a T. denticola cfpA deficient strain showed reduced 25 fibronectin binding, P. gingivalis coaggregation and mixed-species biofilm formation (Vesey 26 and Kuramitsu 2004, Yamada et al. 2005, O'Toole et al. 2000). 27 Whilst T. denticola does not form biofilm on inert surfaces in a static biofilm system, 28 the presence of P. gingivalis and fibronectin as the substratum mediates T. denticola biofilm 29 formation (Vesey and Kuramitsu 2004). P. gingivalis and T. denticola form mixed-species 30 biofilms with increased thickness and biomass, as well as exhibiting tighter adherence to the 31 substratum than monospecies biofilms (Kuramitsu et al. 2005, Zhu et al. 2013), wherein, P. 32 gingivalis is found in the inner layer of the biofilm while T. denticola is localized to the exterior 33 region of the dual-species static biofilm (Kuramitsu et al. 2005, Yamada et al. 2005). P.

1 gingivalis fimbriae and gingipains bind to T. denticola dentilisin and Msp respectively, which 2 is consistent with the observation that these two periodontopathogens are consistently found 3 colocalizing and coaggregating in the same environment (Hashimoto et al. 2003, Rosen et al. 4 2008).

5 1.5.3 Roles of proteases in T. denticola 6 T. denticola possesses a number of proteases known to be involved in periodontal tissue 7 destruction. Many studies have demonstrated that T. denticola proteinases (PrtP, PrtA, PrtB) 8 and peptidases (prolyl oligopeptidase, prolyl aminopeptidase, oligopeptidase B, ϒ- 9 glutamyltransferase, leucyl aminopeptidase and cystalysin) are involved in various 10 pathophysiological processes, including penetration and invasion of epithelial tissues and 11 gingival tissues, cytopathic effects on epithelial cells and fibroblasts, dysregulation of tissue 12 homeostasis, and degradation of host complex proteins. Concomitantly, these enzymes also 13 play important roles in providing metabolic substrates in the form of peptides and free amino 14 acids for T. denticola growth and energy generation. 15 The outer sheath of T. denticola is decorated with a wide array of versatile and effective 16 proteinases capable of degrading complex host proteins (Rosen et al. 1995, Veith et al. 2009). 17 Triton X-114 extraction of the outer sheath performed by Rosen et al. (1994) showed that cell 18 surface associated proteinases with apparent molecular masses ranging from 91 to 228 kDa 19 exhibited fibrinolytic activities (Rosen et al. 1994). T. denticola outer sheath-associated 20 proteinases that have been characterised include CTLP, prolyl oligopeptidase (POPase), 21 proline iminopeptidase (PIPase) and endo-acting oligopeptidase (OPase). 22 CTLP shares common biochemical properties and the same catalytic triad as 23 chymotrypsin enzymes; however, it exhibits a relatively wider range of substrate specificities 24 as compared to chymotrypsin enzymes, which specifically cleave peptide amide bonds with 25 aromatic amino acids (Phe, Tyr, Trp) at their N-terminus of the peptide (Fenno and McBride 26 1998). The CTLP complex has been implicated as a pathogenic factor contributing to the 27 penetration of epithelial cell monolayers, cytotoxic effects on epithelial cells and fibroblasts, 28 as well as hydrolyzing a wide array of host proteins, including bioactive peptides i.e. bradykinin 29 and substance P (Mäkinen et al. 1995, Dashper et al. 2011), serum proteins i.e. serum albumin, 30 transferrin, IgA and IgG, and various cytokines i.e interleukin-1β (IL-1β), IL-6, IL-8 and tumor 31 necrosis alpha factor (Miyamoto et al. 2006). The CTLP complex displays hydrolytic activities 32 towards host structural proteins in the extracellular matrix, such as fibrinogen, fibronectin, 33 transferrin, laminin and collagen type IV (Bamford et al. 2007). CTLP complex is also involved 26

1 in the activation of host pro-matrix metalloproteinase-2 (pro-MMP-2) that is constitutively 2 expressed in periodontal ligament cells, which in turn degrades intact fibronectin into smaller 3 fragments (Miao et al. 2011). The fragmented fibronectin induces the expression of periodontal 4 ligament cell collagenase, stromelysin, urolinase plasminogen activator, induces apoptosis and 5 suppresses osteoblast differentiation of periodontal ligament cells, all of which significantly 6 disrupt tissue homeostasis and enhances periodontal tissue destruction (Miao et al. 2011). The 7 CTLP complex has also been demonstrated to modulate the host proteinases plasminogen and 8 fibrinolytic enzymes, which leads to further degradation of host proteinase inhibitors and 9 extracellular matrix tissues (Rosen et al. 1994, Grenier 1996). Other chymotrypsin-like 10 proteinases in T. denticola that have been purified for characterisation include PrtA and PrtB. 11 PrtA is a 67 kDa protease that hydrolyzes type IV collagen, laminin and fibronectin, whilst a 12 30 kDa PrtB hydrolyzes serum albumin and casein (Que and Kuramitsu 1990, Arakawa and 13 Kuramitsu 1994). 14 T. denticola possesses a prolyl oligopeptidase (POPase) that has been shown to cleave 15 oligopeptide substrates at a proline residue, with a minimum hydrolysable peptide size being a 16 tetrapeptide to a maximum substrate size of approximately 3 kDa. POPase was shown to 17 hydrolyse bonds at the carboxyl side of proline in several human bioactive peptides, including 18 substance P, neurotensin, angiotensins, oxytoxin, vasopressin, bradykinin and human 19 endothelin fragment (Mäkinen et al. 1994). Proline iminopeptidase (PIPase), also known as 20 prolyl aminopeptidase, specifically cleaves at the Pro-Y bond of dipeptides, where Y is 21 preferably Arg, Lys, Gln, Asn or Ala (Mäkinen et al. 1996). PIPase is probably involved in the 22 downstream activities of the peptidolytic cascade to release proline and other amino acids for 23 metabolism. T. denticola exhibits hydrolytic activities against furylacryloyl-Leu-Gly-Pro-Ala 24 (FALGPA) and phenylazobenzyl-oxycarbonyl-L-leucylglycyl-L-prolyl-D-arginine (PZ- 25 PLGPA), which are synthetic substrates used for screening bacteria with collagenase-like 26 activity (Que and Kuramitsu 1990, Mäkinen et al. 1990, Söderling et al. 1994). Likewise, T. 27 denticola cell extracts also demonstrated the ability to hydrolyse gelatin, soluble type I 28 collagen, collagen-derived polypeptides (Mäkinen et al. 1990, Söderling et al. 1994); however, 29 the FALGPA-peptidase that had been purified to homogeneity was unable to degrade collagens 30 or other synthetic collagenase substrates, but can cleave bradykinin and other bioactive 31 components (Mäkinen et al. 1992, Mäkinen et al. 1995). 32 T. denticola also demonstrated proteolytic activities against N-α-benzoyl-L-arginine-2- 33 naphthylamine (BANA) and N-α-benzoyl-L-arginine-p-nitroaniline (BApNA) substrates

1 (Ohta et al. 1986, Mäkinen et al. 1995). This catalytic enzyme was identified as the surface- 2 associated oligopeptidase, known as OpdB or OPase (Mäkinen et al. 1995, Fenno et al. 2001). 3 T. denticola OpdB is unable to hydrolyse proteins, but can degrade peptides of varying sizes 4 and show specificities in the cleavage of peptides with arginine or lysine at the P1 position, 5 with a relatively slower hydrolysis rate for the latter (Mäkinen et al. 1995). The knockout strain 6 of opdB showed complete abolishment of BANA hydrolysis, as well as demonstrating an 7 extended lag phase with slightly slower growth rate than the wild type (Fenno et al. 2001). 8 T. denticola possesses catalytic enzymes for the hydrolysis of glutathione (GSH). 9 Physiologically, GSH is an important thiol-source that regulates cellular oxidative stress and is 10 present in the concentration range of 0.5-2.5 mM in gingival crevicular fluid (Bains and Bains 11 2015, Chapple et al. 2002). T. denticola γ-glutamyl-transpeptidase (GGT) and 12 cysteinylglycinase (CGase) are essential for the sequential hydrolysis of GSH into the free 13 amino acids glutamate, glycine and cysteine, while cysteine is further hydrolysed into hydrogen

14 sulfide (H2S), ammonia and pyruvate by cystalysin (Fig. 1.4). The 27 kDa GGT was first 15 identified and purified by Mäkinen and Mäkinen using the chromogenic substrate N-γ- 16 glutamyl-4-nitroaniline (GNA) (Mäkinen and Mäkinen 1997). The recombinant GGT was later 17 found to be responsible for the hydrolysis of γ-glutamyl linkages in GSH to produce glutamate 18 and cysteinylglycine (Chu et al. 2003). Interestingly, the GNA substrate exhibited competitive

19 inhibition for GGT activity towards GSH, leading to decreased H2S production (Chu et al. 20 2003). The 52 kDa CGase is specifically involved in the hydrolysis of cysteinylglycine into 21 cysteine and glycine (Fig. 1.4; Chu et al. 2008). Cystalysin is a 46 kDa cytoplasm-localized L- 22 cysteine desulfhydrase, which participates in the hemolysis of human erythrocytes, 23 hemooxidation of hemoglobin and removes sulfhydryl and amino groups from S-containing

24 compounds, producing pyruvate, H2S and ammonia (Chu et al. 1997, Chu et al. 1999).

25 Hemolysis of erythrocytes is due to cystalysin activities that elevate the concentration of H2S, 26 which diffuse directly into the erythrocyte membrane, damage the membrane and increase its 27 permeability (Krupka et al. 2000, Chu et al. 1999). As a result, the erythrocyte cytoplasmic 28 content, which contain rich source of peptides and haemoglobin become accessible to T. 29 denticola. The versatility of cystalysin in the degradation of S-containing compounds, 30 combined with its hemolytic properties, suggest that it could be a potential antimicrobial target 31 due to its important role in periodontal disease progression (Cellini et al. 2003, Spyrakis et al. 32 2014).

1 It is conceivable that T. denticola proteolytic enzymes are not only important for the 2 survival and growth of T. denticola in the environment, but also contribute to significant 3 periodontal tissue destruction. Many of the proteinases and peptidases discussed above seem 4 to display complementarity and redundancy in function, resulting in the hydrolysis of complex 5 host proteins into oligopeptides and amino acids. The complexity of proteolytic enzymes in T. 6 denticola is demonstrated by the diverse enzymatic activities detected using screening assays 7 (Mäkinen and Mäkinen 1996, Seshadri et al. 2004), however, many of their physiological 8 properties and roles in pathogenesis either remain uncharacterised or only superficially 9 characterised. Therefore, more research efforts are required for a better understanding of the 10 roles of T. denticola proteolytic enzymes in the initiation and progression of periodontal 11 disease. 12

13 14 Figure 1.4: Glutathione stepwise degradation and catabolism by T. denticola. The enzyme 15 names are italicised. Image reproduced from Chu et al. 2002.

17 1.5.4 T. denticola peptide and amino acid metabolism 18 The extensive proteolytic activities of T. denticola confers this bacterium with 19 competitive and adaptive advantages for proliferation in proteinaceous environments; the 20 degradation of host proteins into a variety of peptides and free amino acids supports its nutrient 21 requirements. T. denticola primarily utilises free amino acids such as serine, alanine, valine, 22 cysteine and glycine as carbon and energy sources (Hespell and Canalepa 1971). The only 23 monosaccharide reported to be catabolized by T. denticola via the Embden-Meyerhoff 24 Parnass/glycolysis pathway is glucose (Hespell and Canalepa 1971). Early studies showed that 25 T. denticola possesses similar enzymatic activities to P. gingivalis for the assimilation of 26 arginine to generate citrulline, ATP and ornithine (Blakemore and Canaleparola 1976). 27 Ornithine can then undergo dissimilation into putrescine via a decarboxylation reaction or into 28 proline and ammonia via a deamination reaction; these pathways exist as a protective 29 mechanism to prevent the accumulation of inhibitory concentrations of ornithine (Leschine and

1 Canaleparola 1980). Recently, the importance of glycine for T. denticola growth was revisited. 2 Tan et al. (2014) demonstrated that the supplementation of exogenous glycine significantly 3 enhanced the growth of T. denticola, while releasing lactate and acetate as the metabolic end 4 products via the reversible glycine reductase system. Glycine could also be utilised by the 5 glycine cleavage system, which is important for the maintenance of the NAD+ and NADH ratio 6 in equilibrium, as well as providing the cofactor folate in different redox states (Fig. 1.5). 7 GSH metabolism is a unique capability found in all strains of T. denticola, but not in 8 other Treponema spp. or other oral bacteria (Chu et al. 2002). As mentioned above, T. denticola 9 is equipped with a set of enzymes specifically involved in the stepwise degradation and 10 metabolism of GSH, which results in the release of free amino acids glutamate, glycine and 11 cysteine; while cystalysin is responsible for the catabolism of cysteine (Chu et al. 2002). 12 Regulated by the amount of iron and thiol-containing compounds, cystalysin hemooxidative 13 and hemolysis activities could also help fulfil T. denticola iron requirement and provide a 14 competitive advantage for its growth (Chu et al. 1999). In addition, other degraded products of 15 GSH, which include glutamate, glycine, pyruvate and ammonia can also be utilised as nutrient 16 substrates by T. denticola, some of which have been shown to enhance T. denticola growth 17 (Chu et al. 2002, Veith et al. 2009, Tan et al. 2014). Other detected fermentation products of 18 T. denticola include acetate, lactate, succinate, formate, pyruvate, ethanol and carbon dioxide 19 (Hespell and Canalepa 1971). The degradation and catabolism of human serum proteins,

20 cysteine and glutathione give rise to an increase of H2S (greater than 2 mM in the periodontal 21 disease pockets) and ammonia level, which are cytotoxic to mammalian cells, destructive to 22 lymphocytes, macrophages and neutrophils, as well as contributing to tissue homeostasis 23 dysregulation (Chu 1999). In addition, T. denticola is an active producer of methyl mercaptan, 24 which is a highly toxic and volatile S-containing compound, via the α,γ-elimination of L- 25 methionine by L-methione-α-deamino-γ-mercaptomethane lyase (METase) (Fukamachi et al. 26 2005). Methyl mercaptan can cause oral malodour and also play an important role in 27 periodontitis by contributing to the degradation of epithelial cell components (Yaegaki 2008). 28

1 2 Figure 1.5: Proposed T. denticola glycine catabolic pathways. Arrow (1,2) glycine cleavage 3 system, (3) deamination of serine to produce pyruvate, (4,5,6,7) pyruvate interconversion 4 pathways, (8) glycine reductase system. Image reproduced from Tan et al. 2014.

5 6 Bacterial metabolism is an important determinant of bacterial persistency in the 7 environment, yet relatively little is known about the molecular mechanisms of T. denticola 8 metabolic pathways. Understanding how T. denticola metabolises peptides and amino acids 9 could help provide insights into how T. denticola proliferate and increase in abundance as 10 periodontal disease progresses. Both the activities of T. denticola proteinases and its unique 11 metabolic pathways are important virulence factors that contribute to the accumulation of 12 cytotoxic end products and destruction of host tissues.

13 1.6 P. gingivalis and T. denticola metabolic interactions 14 Bacteria growing in a planktonic state or homotypic biofilm often show limited 15 metabolic capability, whereas bacteria living in a consortium can achieve metabolic 16 coordination and expansion of their metabolic capacity. Close association of T. denticola and 17 P. gingivalis in the same microenvironment opens up the possibilities for synergistic metabolic 18 interactions. Metabolic cooperativities between P. gingivalis and T. denticola have been 19 demonstrated to include the complementary metabolic biosynthesis pathways and proteinase 20 functions, to increase the production and release of metabolic substrates, as well as the cross- 21 feeding of metabolic substrates.

1 Metabolic synergistic interactions have been observed between P. gingivalis and T. 2 denticola in the biosynthesis and cross-feeding of short chain fatty acids. T. denticola produces 3 succinic acid and P. gingivalis produces isobutyric acid, which can support the growth of P. 4 gingivalis and T. denticola respectively (Grenier 1992). Glutamate is another potential 5 substrate that might play a role in the synergistic interactions between these two 6 periodontopathogens as P. gingivalis glutamate catabolic enzyme levels were increased in a 7 mixed-species biofilm (Zainal-Abidin et al. 2012). Bioinformatic analysis of the bacterial 8 metabolic pathways revealed that the butyric acid biosynthesis genes of the red complex 9 species can be complementary to one another, leading to increased production of butyric acid 10 (Endo et al. 2015). Likewise, P. gingivalis showed downregulation of fatty acid biosynthesis 11 genes fabG, fabF and acpP during coculture with T. denticola (Tan et al. 2014), suggesting 12 that complementary activities of fatty acid biosynthesis pathways during interspecies 13 association could be one of the factors that contribute to higher cellular biomass. 14 T. denticola showed upregulation of glycine cleavage and reductase pathways during 15 coculture with P. gingivalis, while P. gingivalis upregulated thiamine biosynthesis pathways 16 under similar conditions (Fig. 1.6; Tan et al. 2014). These results suggested metabolic cross- 17 feeding reactions of glycine and thiamine from P. gingivalis to T. denticola. Although the 18 molecular mechanisms involved in these interactions have not been fully elucidated, it was 19 hypothesized that T. denticola produced stimulatory factor(s) that increased P. gingivalis 20 proteinase expression to help release free glycine from its peptide-bound forms, which in turn 21 enhanced the growth of T. denticola. However, transcription of the genes encoding the T. 22 denticola glycine reductase complex selenoproteins (TDE2119, TDE2120) is repressed upon 23 contact with P. gingivalis, which might serve as an immune evasion strategy (Sarkar et al. 24 2014). Hence, it was postulated that glycine might serve as an important metabolic substrate 25 and chemoattractant for T. denticola towards the glycine-producing P. gingivalis during the 26 transitory stages of dual-species biofilm formation prior to their physical contact. In a mixed 27 species biofilm, T. denticola showed an increase in the abundance of glycine cleavage enzymes 28 for the utilisation of glycine (Zainal-Abidin et al. 2012). Lactate and acetate, as the glycine 29 catabolic end products could potentially be utilised by both T. denticola and P. gingivalis. 30 A significant upregulation of T. denticola cystalysin production during coculture with 31 P. gingivalis was also observed (Tan et al. 2014). This could possibly lead to an increase in

32 cysteine catabolic activities, hence contributing to a higher level of H2S, ammonia and pyruvate 33 production during coculture. Whilst it has been shown that GSH was inhibitory to P. gingivalis

1 (Kato et al. 2008), the degradation of GSH by recombinant T. denticola GGT into 2 cysteinylglycine enabled P. gingivalis to utilise it as a nutritional substrate (Chu et al. 2003). 3 As a consequence, P. gingivalis exhibited enhanced cytotoxic products release as well as 4 hemooxidative and hemolytic activities (Chu et al. 2003). Although the mechanism by which

5 P. gingivalis catabolizes Cys-Gly into H2S is still unknown, it is conceivable that P. gingivalis 6 possesses its own peptidase rather than the T. denticola CGase catalyzing the enzymatic events 7 (Chu et al. 2008). In addition, the T. denticola CTLP complex and P. gingivalis gingipains 8 have been found to break down a range of host proteinase inhibitors and activate plasminogen, 9 indirectly modulating host proteolytic activities and triggering inflammatory responses 10 (Grenier 1996). Based on all of the evidence presented, T. denticola and P. gingivalis proteases 11 and peptidases are likely to function complementarily for proteolytic degradation of host cells 12 and proteins, whilst concomitantly disrupting host regulatory activities, leading to the 13 uncontrolled activities of host proteases, thereby increasing the release of metabolic substrates. 14 Amino acids and peptides are the carbon and energy sources of T. denticola and P. 15 gingivalis, likewise, utilisation of these substrates results in increased production and release 16 of cytotoxic end products into the environment. The accumulation of cytotoxic substances,

17 such as ammonia, H2S and butyrate dysregulates host tissue homeostasis and immune 18 responses that aggravate periodontal disease conditions. High levels of these substances are not 19 only toxic to the host cells, but also play a role in competitive exclusion of oral commensals,

20 while favoring the persistence of periodontopathogens that are H2S- and alkaline-tolerant 21 (Takahashi 2005). These processes may in part act as a selective force for the propagation of 22 T. denticola and P. gingivalis while reducing the diversity of the oral microbiota as the disease 23 progresses. 24 The colocalisation of these bacterial species is a factor that helps shape their nutritional 25 acquisition strategies, whereby close association of T. denticola and P. gingivalis allows 26 synergistic effects on host protein degradation, as well as maximizing the efficiency of 27 metabolic substrate uptake and utilisation by reducing the diffusion of the generated substrates. 28 P. gingivalis and T. denticola proteinases and metabolic pathways are responsible for fulfilling 29 their nutrient requirements and increasing production of cytotoxic end products, which may 30 directly or indirectly contribute to the modification of the oral environment, modulation of host 31 inflammatory responses and damage of the host periodontal tissues. Collectively, T. denticola 32 and P. gingivalis synergistic interactions contribute to increased biomass and enhanced 33 virulence during coculture and polymicrobial biofilm formation.

2 3 Figure 1.6: Schematic diagram illustrating metabolic interactions between T. denticola (green 4 spiral) and P. gingivalis (purple coccobacillus) in a polymicrobial biofilm. Extracted from Ng 5 et al. (2016)

1 1.7 Aims of this study 2 A previous study in this laboratory demonstrated that P. gingivalis increased free glycine 3 production by proteolytic degradation of glycine-containing peptides in T. denticola 4 conditioned medium (TdCM). It was hypothesized that the increased free glycine production 5 by P. gingivalis was due to the presence of T. denticola signalling molecules in TdCM, which 6 engaged P. gingivalis in a glycine cross-feeding reaction with T. denticola. This work was 7 performed to examine the molecular factors in TdCM that contribute to the free glycine- 8 releasing activity of P. gingivalis, as well as to determine the P. gingivalis peptidases that are 9 involved in the process of releasing free glycine. The gene expression of P. gingivalis in TdCM 10 was studied to gain molecular insights into P. gingivalis and T. denticola multimodal 11 interactions. 12 13 Thesis Aim: To elucidate synergistic interactions between T. denticola and P. gingivalis 14 15 Objectives of this study: 16 1. To characterise T. denticola stimulatory factors that result in an increased production 17 of free glycine by P. gingivalis. 18 2. To identify P. gingivalis peptidases involved in the release of free glycine from glycine- 19 containing peptides. 20 3. To determine the differential gene expression of P. gingivalis in response to growth in 21 cell free TdCM.

22 23 24 25 26 27 28 29

1 Chapter 2 Materials and methods

2 2.1 Bacterial strains and growth conditions 3 Bacterial strains and plasmids used in this study are listed in Table 2.1. All chemicals 4 used for media preparation were purchased from Sigma Aldrich (Australia) unless stated 5 otherwise. P. gingivalis was grown in Brain Heart Infusion (BHI, Bacto, BD, DifcoTM, Becton, 6 Dickinson and Co, USA) broth supplemented with 5 µg/mL hemin and 5 mg/mL L-cysteine, 7 and on horse blood agar (HBA) plates containing 40 g/L Blood Agar base No. 2 (Oxoid Ltd., 8 England) with 10% (v/v) defibrinated horse blood (BioLab, Vic Australia). 9 T. denticola and P. gingivalis were grown in Oral Bacterial Growth Medium (OBGM) 10 (Veith et al. 2009) formulated with the following components per litre of medium: 12.5 g BHI, 11 10 g Tryptone Soya Broth (Oxoid), 7.5 g Yeast Extract (Oxoid), 2 g sodium chloride (Chem- 12 Supply Pty. Ltd., Australia), 2 g D-glucose, 2 g ascorbic acid, 1 g sodium pyruvate, 0.25 g L- 13 asparagine and 0.5 g sodium thioglycolate were mixed. The mixture was pH adjusted to 7.4 14 using 5 M KOH before the volume was made up to 800 mL with de-ionised water (DIW; Milli- 15 Q®, Merck Millipore, Australia) for autoclave sterilisation at 121°C, 100 kPa for 25 min as the 16 medium base. If required, 0.8% (w/v) agarose (Life Technologies, Thermo Fisher Scientific, 17 Australia) was added prior to autoclaving for OBGM agar. Then, the base media (or agar 18 medium) was supplemented with OBGM supplemental medium consisting of 6 mg thiamine 19 pyrophosphate, 2 g of sodium bicarbonate, 0.5% (v/v) volatile fatty acids mixture stock (0.5% 20 (v/v) isobutyric acid, DL-2-methylbutyric acid, isovaleric acid and valeric acid in 0.1 M KOH), 21 2 g ammonium sulfate, 5 mg hemin, 1 mg of menadione, 2% (v/v) filtered rabbit serum that 22 had been heat-inactivated at 55°C for 30 min, 1 g of cysteine hydrochloride and pH adjusted 23 to 7.4. The final volume was made up to 200 mL with DIW before filter-sterilisation via a 0.22 24 µm-size Steritop filter (Merck Millipore, Thermo Fisher Scientific, Australia). All P. gingivalis 25 and T. denticola bacterial strains were grown in an anaerobic chamber (MG500 anaerobic 26 workstation; Don Whitley Scientific Pty. Ltd., NSW, Australia) with gas composition of 5%

27 H2, 10% CO2 in N2 at 37°C. Erythromycin (Em) and ampicillin (Ap) were supplemented at a 28 final concentration of 10 µg/mL and 5 µg/mL respectively in the growth media of different P. 29 gingivalis mutant strains when necessary. 30 For the preparation of glycerol stocks, T. denticola and P. gingivalis log phase cultures

31 with an OD650 ~0.15 and ~0.6 respectively, as measured by a Cary 50 UV-Vis 32 Spectrophotometer (Varian Inc. Scientific Instrument, USA) were concentrated 12.5-fold and

1 4-fold respectively by centrifugation at 4,000 × g for 15 min at 4°C and resuspended in a lower 2 volume. Then, 1 mL and 500 µL of the concentrated cultures of T. denticola and P. gingivalis 3 respectively were added to 500 µL glycerol, snap-frozen with liquid nitrogen and stored at - 4 80°C. The glycerol stocks of T. denticola were cultured in 40 mL of pre-warmed OBGM, while 5 P. gingivalis stocks were streaked on HBA plates to initiate cell growth. Routine passaging of 6 T. denticola cells was performed to maintain the viability of cells, where 5% (v/v) from the 7 previous culture was used to inoculate fresh pre-warmed OBGM, up to a maximum of 5 8 passages. For experimental testing, 1.4 × 108 P. gingivalis log phase cells from the secondary 9 culture were inoculated into a 40 mL OBGM. In a 2 ml medium, 7 × 106 P. gingivalis cells 10 were inoculated and grown in a CELLSTAR® cell culture 12-well polystyrene plate (Greiner 11 BioOne GmbH, Germany). 12 Escherichia coli strains were grown in Lysogeny Broth [LB: 1% (w/v) Tryptone 13 (Oxoid), 0.5% (w/v) Yeast Extract (Oxoid) and 0.5% (w/v) sodium chloride (Chem-Supply)] 14 or on LB agar plates made with LB broth containing 1.5% (w/v) Bacto agar (BD, DifcoTM). E. 15 coli cultures were incubated aerobically at 37°C in a shaking incubator (NT INFORS Minitron, 16 INFORS AG-CH 4103 Bottmingen, Infors AG, Switzerland), shaking at 180-220 rpm. Media 17 were supplemented with Ap at a final concentration of 100 µg/mL for the selection of E. coli 18 transformed cells. 19

Table 2.1: Bacterial strains and plasmids used in this study

Bacterial Background Plasmid Description Antibiotic Reference or strain strain resistance source* selectiona,b Plasmids NA NA pGEM-T easy T-overhang cloning plasmid; ApR Ap Promega, USA NA NA pGEM-T easy::cepA pGEM-T easy vector with Bacteroides fragilis cephalosporinase Ap This laboratory encoding gene (cepA) gene Dr Christine Seers NA NA pVA2198 Mob+; 9.2 kbp E. coli and Bacteroides shuttle vector with source of Em (Fletcher et al. 1995) ermF-ermAM cassette; SpR, EmR NA NA pLK0445 PG0445 flanking regions ligated with cepA was TA cloned into Ap This study pGEM-T easy vector NA NA pLK0753 PG0753 flanking regions ligated with cepA was TA cloned into Ap This study pGEM-T easy vector NA NA pLK1605 PG1605 flanking regions ligated with ermF was TA cloned into Ap This study pGEM-T easy vector NA NA pLK1788 PG1788 flanking regions ligated with cepA was TA cloned into Ap This study pGEM-T easy vector E. coli α-select NA NA F- deoR endA1 recA1 relA1 gyrA96 hsdR17(rk-, mk+) supE44 thi- NA Bioline, England 1 phoA Δ(lacZYA-argF)U169 Φ80lacZΔM15 λ- ECR839 α-select pLK0445 E. coli α-select transformed with pLK0445 Ap This study ECR840 α-select pLK0753 E. coli α-select transformed with pLK0753 Ap This study ECR841 α-select pLK1605 E. coli α-select transformed with pLK1605 Ap This study ECR842 α-select pLK1605 E. coli α-select transformed with pLK1788 Ap This study P. gingivalis W50 NA NA Wild type NA This laboratory ECR843 W50 NA P. gingivalis PG0445- deletion mutant. The PG0445 predicted ORF Ap This study (PG0445-)c was replaced with cepA; ∆PG0445::cepA 38

Bacterial Background Plasmid Description Antibiotic Reference or strain strain resistance source* selectiona,b ECR844 W50 NA P. gingivalis PG0753- deletion mutant. The PG0753 predicted ORF Ap This study (PG0753-)c was replaced with cepA with promoter; ∆PG0753::cepA ECR845 W50 NA P. gingivalis PG1605- deletion mutant. The PG1605 predicted ORF Em This study (PG1605-)c was replaced with an ermF with promoter; ∆PG1605::ermF ECR846 W50 NA P. gingivalis PG1788- deletion mutant. The PG1788 predicted ORF Ap This study (PG1788-)c was replaced with cepA with promoter; ∆PG1788::cepA T. denticola ATCC NA NA Wild type NA This laboratory 35405 NA: Not applicable * Strains obtained in this laboratory and constructed in this study were added to the culture collection of the Oral Health Cooperative Research Center, The University of Melbourne, Australia. a. Antibiotic resistance selection in E. coli strains: 100 µg/mL ampicillin (Ap) and 50 µg/mL erythromycin (Em) b. Antibiotic resistance selection in P. gingivalis strains: 5 µg/mL Ap and 10 µg/mL Em c. Superscript (-) denotes the deletion mutants

2.2 Bacterial cell number enumeration Cell numbers of P. gingivalis and T. denticola in OBGM were determined by measurements of the optical density of culture at 650 nm wavelength (OD650) using a Cary 50

UV-Vis spectrophotometer and calculated based on the OD650 derived equations, P. gingivalis 9 8 cell number/mL = (2×10 ×OD650)+2×10 , within an OD650 range of 0.1-1.0 and T. denticola 9 7 cell number/mL = (3×10 ×OD650)+9×10 , within an OD650 range of 0.1-0.25 (Orth et al. 2010). A plate count method was also employed to validate P. gingivalis cell numbers in different growth media, by which P. gingivalis culture was serially diluted using the corresponding medium until an estimated countable number of colony forming units (cfu) i.e. 30-300 cfu/plate was attained. Then, 50-200 µL of the diluted culture was spread onto HBA plate and incubated for a maximum of 10 days under anaerobic conditions before counting of colony forming units. P. gingivalis cell counts using the live/dead fluorescence stain coupled with flow cytometry was performed according to methods by Orth et al. (2010) with slight modifications. Basically, P. gingivalis cultures were serially diluted to an expected range of 2 x 105 to 2 x 106 cells/mL with 0.85% (w/v) NaCl, then 200 µL of P. gingivalis diluted culture was transferred to a CELLSTAR® cell culture 96-well round bottom plate (Greiner bio-one GmbH, Germany). The cells were mixed with green-fluorescent DNA intercalating dye SYTO9 (Life Technologies) and red-fluorescent DNA intercalating dye propidium iodide (Life Technologies) in a final dye to culture volume ratio of 1:1000. Then, the samples were incubated in the dark for 5 min before measurement using a Flow Cell Lab Quanta SC MPL Flow Cytometer (Beckman CoulterTM Inc., NSW, Australia) with an excitation wavelength of 488 nm. The green fluorescence of live cells was measured through a 525 nm filter, while the red fluorescently stained DNA of membrane-compromised or dead cells was measured through a 575 nm filter. The number of P. gingivalis live cells/mL was determined by the Cell Lab Quanta Collection and Analysis softwares.

2.3 Molecular biology techniques 2.3.1 Extraction of bacterial genomic and plasmid DNA Chromosomal or plasmid DNAs were purified using DNeasy Blood and Tissue kit (QIAGEN, Germany) or QIAprep Spin Miniprep kit (QIAGEN) respectively as per the manufacturer’s instructions. HiPure Plasmid Midiprep kit (Invitrogen, Life Technologies Pty. Ltd., Australia) was used for the isolation of high yield plasmids from bacterial clones according to the manufacturer’s instruction. DNA was eluted with low Buffer TE (2 mM Tris- Cl, 0.1 mM ethylenediaminetetraacetic acid (EDTA), pH 9.0) and DNA concentration was determined by measurement of 2 µL DNA aliquots at the absorbance reading of 260 nm TM wavelength (A260) using a NanoDrop Spectrophotometer ND-1000 (Thermo Fisher Scientific Inc.). 2.3.2 Polymerase Chain Reaction The design of polymerase chain reaction (PCR) primers to amplify from genomic DNA was based on the reference sequence of P. gingivalis W83 available from the National Center for Biotechnology Information (NCBI, https://www.ncbi.nlm.nih.gov; Reference Sequence: NC_002950.2). The nucleotide sequences of ermF in the shuttle vector pHS17 and cepA resistance cassettes were sourced from GenBank using accession numbers AF219231 and L13472 respectively (Benson et al. 2017). All primers were designed to have melting temperatures of approximately ~55-60°C based on the salt adjusted melting temperature calculation using OligoCalc (Kibbe 2007). Lyophilised oligonucleotide primers (GeneWorks Pty. Ltd., Australia) were resuspended to a concentration of 100 µM in DIW as primer stocks listed in Table 2.2 P. gingivalis chromosomal DNA was used as the DNA template for PCR amplification. An erythromycin resistance cassette (ermFAM cassette) driven by its promoter was amplified from pVA2198 (Fletcher et al. 1995), in which ErmF is expressed for the selection of P. gingivalis transformants. The cepA gene that encodes ampicillin resistance was obtained by PCR amplification of the Bacteroides fragilis cephalosporinase encoding gene (cepA) that had been cloned into the pGEM-T vector by Dr. Christine Seers. A G-Storm GS1 thermal cycler (GeneWorks Pty. Ltd., Australia) was used for all PCR reactions. The lid was continuously heated at 111°C during the PCR cycles and the PCR cycles did not commence until the thermal block was preheated to 94°C. PCRs were performed using approximately 50-100 ng of DNA template, 2.5 µL each of 10 µM forward and reverse primers, 1 µL of 10 mM dNTPs, 1 × Phusion High Fidelity (HF) buffer, 1 U of Phusion HF DNA 41

polymerase (NEB, New England Biolabs Inc., USA) and DIW in a final volume of 50 µL per reaction. The cycling program consisted of an initial denaturation of 2 min at 98°C, followed by 30 cycles of 30 s at 98°C, 30 s at 60°C and 30 s/kbp at 72°C, then a final extension for 8 min at 72°C, and storage of samples in the PCR block at 4°C. Splicing by overlap extension (SOEing) PCR was performed using two equimolar DNA templates, 1 × Herculase II reaction buffer, 0.5 µL of 100 mM dNTP, 1.25 µL of 10 µM forward and reverse primers respectively, 1 µL of Herculase II Fusion DNA polymerase (Agilent Technologies, USA) and DIW in a final volume of 50 µL according to the manufacturer’s instructions. The SOEing cycling program was set with denaturation at 95°C for 2 min, followed by 95°C for 20 s, 60°C for 20 s, 72°C for 30 s/kbp extension. The first five cycles were without the addition of primers, and the program was resumed for another 30 cycles after the addition of primers; the amplification was completed with a final extension for another 8 min at 72°C. The final SOE PCR was performed using High Fidelity Platinum Taq DNA Polymerase (Invitrogen) to obtain an A-tail on the PCR product for ligation into pGEM-T easy vector (Promega Corporation, Wisconsin, USA). The PCR mixture consisted of 1 × HF PCR buffer,

2 µL of 50 mM MgSO4, 1 µL of 10 mM dNTP mix, 1 µL each of 10 µM forward and reverse primers, two PCR fragments to join in equimolar ratio, 1 U of Platinum Taq DNA HF polymerase and DIW in a final volume of 50 µL. PCR conditions were as previously described for SOE PCR, except the denaturation and elongation settings, which were at 94°C for 30 s and 68°C for 1 min/kbp respectively. PCR products of the expected size were purified from the PCR solutions or gel purified using column based clean-up kits such as the QIAquick gel extraction kit (QIAGEN) as per the manufacturer’s instructions.

Table 2.2: Primers for PCR amplification

Primer Sequence (5' - 3') Nucleotide Product of Amplification position * PG0445F1_FP AGACCTATGGCTTTACGGGC 484882a 510 bp upstream of PG0445 predicted start codon cepAPG0445F1_RP TGTATAAGTCTTTTTTGCATATTCTTTGTTTCGTTGTTATGCTTC 485391a PG0445cepA_FP ATAACAACGAAACAAAGAATATGCAAAAAAGACTTATACATTTATCCAT 436c 903 bp based on Bf cepA PG0445F2cepA_RP GATCGTCATTCGTGCGATGATTAATCTATCTGTTGCGTTACGTATT 1338c cepAPG0445F2_FP TAACGCAACAGATAGATTAATCATCGCACGAATGACGATCC 486610a 527 bp downstream of PG0445 predicted transcript PG0445F2_RP TAGCGCGTCTCAGCATTGTC 487136a PG0445confo_FP AAGCGGACGAGTACCATCG 484395a 3152 bp based on W83 sequence PG0445confo_RP CTTACAGGCAGAATTGCCTTG 487546a PG0753F1_FP CGACTGCGCTTTCGTCGTAAT 797891a 550 bp upstream of PG0753 predicted start codon cepAPG0753F1_RP CCTTAACTCTTTTTGACGTCAGGCCAAATCGCCTGCGTC 798440a PG0753cepA_FP TGACGCAGGCGATTTGGCCTGACGTCAAAAAGAGTTAAGGAAAG 288c 1049 bp including promoter region of Bf cepA PG0753F2cepA_RP TTACTCTGTTTCCCCTCGTATTAATCTATCTGTTGCGTTACGTAT 1338c cepAPG0753F2_FP TAACGCAACAGATAGATTAATACGAGGGGAAACAGAGTAAG 800349a 500 bp downstream of PG0753 predicted transcript PG0753F2_RP GTCGAACTGACGTTGTAAGCAG 800848a PG0753confo_FP TGGGCATGCATCCCTACATG 797502a 3777 bp based on W83 sequence PG0753confo_RP GTTCCTTGACTACCGGACGT 801278a PG1605 F1_FP ATCTCCTTTCGGAAGATTGCAG 1684762a 498 bp upstream of PG1605 predicted transcript EmF PG1605F1_RP CGGAAGCTATCGGGGGTACCCGATCCTGCTACGGAAGCAG 1685258a

Primer Sequence (5' - 3') Nucleotide Product of Amplification position * PG1605 EmF_FP CTGCTTCCGTAGCAGGATCGGGTACCCCCGATAGCTTCC 293b 801 bp based on pHS17 PG1605F2 EmF_RP GGAGGAGTAGGGCGAGAGTTTTACGAAGGATGAAATTTTTCAGG 1093b EmF PG1605F2_FP AAAAATTTCATCCTTCGTAAAACTCTCGCCCTACTCCTCC 1686663a 505 bp downstream of PG1605 predicted start codon PG1605F2_RP CGTTACGATCACAACTGTCGG 1687167a PG1605confo_FP TAGCCAATCTGGCCTCCTTC 1684321a 3316 bp based on W83 sequence PG1605confo_RP GCTCCTTGAATTTAGCGGTTC 1687636a PG1788F1_FP TGTGCCCAGAGACGTTACATC 1875576a 609 bp upstream of PG1788 predicted start codon# cepAPG1788F1_RP CCTTAACTCTTTTTGACGTCTCGAGTCCCAAATATCGTCCAT 1876115a PG1788cepA_FP GGACGATATTTGGGACTCGAGACGTCAAAAAGAGTTAAGGAAAG 436c 903 bp based on Bf cepA PG1788F2cepA_RP TGAAATACAGGGGATATGATTTAATCTATCTGTTGCGTTACGTATT 1338c cepAPG1788F2_FP TAACGCAACAGATAGATTAAATCATATCCCCTGTATTTCAGAC 1877388a 705 bp downstream of PG1788 predicted transcript PG1788F2_RP CTTGTCCGGCACCTCGGTAG 1878092a PG1788confo_FP GTGCTTTGGTAGTGGGTTCG 1875120a 3306 bp based on W83 sequence PG1788confo_RP CAACCGACTGTACGGACTTC 1878425a UnivFPd GTAAAACGACGGCCAGTG 2977d NA UnivRPd GGAAACAGCTATGACCATGA 173d NA

Underlined and bold nucleotide sequences represent overlapping sequences utilised for SOEing PCR a. P. gingivalis W83 complete sequence (NCBI Reference sequence: NC_002950.2) b. Shuttle vector pHS17 containing ermF and ermAM genes with accession number AF219231 44

c. Bacteroides fragilis (Bf) cephalosporinase (cepA) gene with accession number L13472 d. pGEM-T Easy vector target sequence * Based on the 5’ position of the primer relative to the 5’ targeted sequence position # The first flanking segment of PG1788 did not include 70 bp upstream from the annotated start site in NCBI, as it was based on the Uniprot PG1788 (Q7MTY9) transcriptional start site

2.3.3 Molecular cloning The molecular cloning techniques performed in this work include ligation and restriction digestion, which are described as follows. Final SOE products were ligated into 50 ng pGEM-T easy vector (Promega) in a 3:1 insert to vector molar ratio with 1 × rapid ligase buffer, 1 µL of T4 DNA ligase (Promega) and DIW to a final volume of 10 µL at 4°C overnight (O/N). Plasmids were linearized in a reaction that consisted of 1 × Cut Smart buffer (NEB), 1 × BSA (NEB), 10 U SacI-HF (NEB) per 1 µg plasmid DNA and DIW for 2 h incubation at 37°C. The resulting linearized plasmids were purified using ethanol precipitation, which involved the addition of 0.1 volumes (vol) of 3 M sodium acetate (pH 5.2) to the restriction digestion mixture, followed by 3 vol of 100% ice-cold ethanol and incubation on ice for 30 min. The plasmids were pelleted by centrifugation at 16,000 × g for 30 min at 4°C and washed thrice with 1 mL of 70% ice-cold ethanol. Then, the pellet was air-dried and resuspended in 50 µL of DIW for use. Plasmid DNA was resuspended in low TE buffer and the concentration was determined at a wavelength of 260 nm using a NanoDrop Spectrophotometer (Thermo Fisher Scientific Inc.). 2.3.4 Agarose gel electrophoresis DNA samples (genomic DNA, plasmid DNA, PCR amplicons) were mixed with 0.2 vol of 5 × DNA loading dye (Bioline) prior to loading onto agarose gels. DNA was resolved in 0.8-1.0% (w/v) agarose (Life Technologies) containing 0.5 × SYBR Safe DNA gel stain (Invitrogen) by gel electrophoresis in 1 × Tris-acetate buffer (TAE; 40 mM Tris-acetate, 1 mM EDTA, pH 8.0) and electrophoresed at 100-120 V for 30-40 min. Then, DNA was visualized using a LAS-3000 Imager (Fujifilm, Berthold Australia Pty.Ltd., Australia). HyperLadderTM 1 kb (Bioline) was used as the standard for DNA size approximation. 2.3.5 Nucleotide sequencing Plasmid DNA from insert positive clones were sequenced using UnivFP, UnivRP and relevant internal primers (Table 2.2) to ensure complete coverage of the insertional construct at the Melbourne Translational Genomics Platform sequencing facility (University of Melbourne) or at the Australian Genome Research Facility (AGRF; University of Melbourne). The chromatogram files provided following the nucleotide sequencing were visualized using Chromas Lite (Technelysium Pty. Ltd., QLD, Australia).

2.4 Bacterial transformation 2.4.1 Escherichia coli heat-shock transformation Chemically competent α-Select Gold E. coli Competent Cells (Bioline) were mixed with 5 µL of mutational construct ligated into the pGEM-T easy vector (Promega) and incubated on ice for 20 min. The cells were heat-shocked in a water bath at 42°C for 45 s, then immediately transferred to ice. The transformed cells were allowed to recover in SOC medium [2% (w/v) Bacto®-tryptone, 0.5% (w/v) Bacto®-yeast extract, 1% (v/v) 1 M NaCl, 0.25% (v/v)

1 M KCl, 1% (v/v) 2 M MgCl2, 1% (v/v) 2 M glucose] at 37°C with shaking at 180 rpm (NT INFORS Minitron) for 2 h. Then, the recovered transformant cells were selected with 100 µg/mL Ap and screened with blue/white colony selection using 50 µL of 0.25 mM 5-bromo- 4-chloro-3-indolyl-β-D-galactopyranoside (X-gal) solution and 100 µL of 100 mM isopropyl β-D-1-thiogalactopyranoside (IPTG) spread onto 20 mL LB agar plates. White colonies were selected to further verify the presence of insert in the pGEM-T Easy vector by colony PCR using MyTaq High Sensitivity DNA Polymerase (Invitrogen, Life Technologies Corporation, Australia) with UnivFP and UnivRP (Table 2.2) following the manufacturer’s instructions. Colony PCR was carried out with longer (≥2 min) initial denaturation to ensure cell lysis, while the PCR thermal cycling reaction was carried out as the PCR protocols described above. 2.4.2 P. gingivalis W50 transformation

Two hundred milliliters of early log phase P. gingivalis W50 cells (OD650 ~0.3-0.65) were washed twice with an equivalent volume of ice-cold electroporation (EP) buffer [10%

(v/v) glycerol, 1 mM MgCl2] and the resulting washed pellet was resuspended in 400 µL of ice-cold EP buffer. Aliquots of 80 µL of P. gingivalis washed cell suspensions were incubated with 500 ng of linear suicide vector containing the target constructs on ice for 5 min. The mixture was transferred to a 0.1 cm gap cuvette (Bio-Rad Inc., USA) and electroporated at 1.8 kV, 200 Ω resistance and 25 µF capacitance. P. gingivalis cells were allowed to recover O/N in BHI supplemented with 0.5 mg/mL cysteine and 5 mg/mL hemin in the anaerobic chamber before plating on HBA plates containing 5 µg/mL Ap or 10 µg/mL Em. Colonies appeared after 10 days of anaerobic incubation, and the cells were passaged onto fresh HBA plates with corresponding antibiotics for further manipulation. The genomic DNA (gDNA) of P. gingivalis transformants was isolated using DNeasy Blood and Tissue kit (QIAGEN) and subjected to PCR to verify the required homologous recombination event in the chromosome.

2.5 Whole genome sequencing using the Ion Torrent Personal Genome Machine Chromosomal DNA of P. gingivalis PG0753-, PG1605- and PG1788- strains was isolated using a DNeasy Blood and Tissue kit (QIAGEN) as per the manufacturer’s instruction. The quality and integrity of gDNA were evaluated using a NanoDropTM spectrophotometer (Thermo Fisher Scientific Inc.), QubitTM dsDNA HS assay kit (Invitrogen) and 0.8% (w/v) agarose gel electrophoresis. Library preparation began with 1 µg of DNA (in 130 µL) sheared at 50 W for 90 s in a Covaris M220 Focused-ultrasonicatorTM (TrendBio, VIC, Australia) to generate 400 bp DNA fragments. The size and quantity of fragmented DNA was assessed using the HT DNA High Sensitivity LabChip reagent kit (Perkin Elmer Inc., USA) with the LabChip GXII Touch (Perkin Elmer) to ensure that a peak fragment size of 400 bp was generated. Then, the DNA fragments were end-repaired using Ion XpressTM Plus Fragment Library kit and purified with AgencourtTM AMPure XP beads (Beckman Coulter, California, USA) according to the manufacturer’s instruction. After assigning each sample with an individual Ion Express Barcode adaptor, the libraries were size selected from 407 to 543 bp using the Pippin PrepTM Instrument with 2% Marker B Agarose Gel Cassettes (Sage Science Inc., USA). The size- selected and barcoded libraries were purified with Agencourt AMPure XP beads and quantitated by qPCR using the Ion Library TaqMan® Quantitation kit. Each of the libraries were diluted to 26 pM and pooled for preparation of template-positive Ion Sphere™ Particles (ISPs) containing 400 bp average insert libraries, using the Ion PGM™ Hi-Q™ View OT2 Kit and the Ion OneTouch™ 2 Instrument. These template-positive ISPs were then enriched with the Ion OneTouch™ ES, before loading onto an Ion 318™ Chip v2 and subjected to sequencing using the Ion PGM™ System with the Ion PGM™ Hi-Q™ View Sequencing Kit. The resulting sequencing reads were downloaded from the Torrent Server and the sequences were analyzed using Geneious 8.1.9 (Biomatters Ltd., New Zealand) with reference to P. gingivalis W83 genome reference sequence (NCBI Reference Sequence: CP025932). 2.6 Preparation of T. denticola conditioned medium The protocol for the preparation of T. denticola conditioned medium (TdCM) was as described in Tan et al. (2014). Briefly, T. denticola ATCC 35405 was grown in OBGM under anaerobic conditions for 7 days until reaching stationary phase with an approximate OD650 of 0.25. The culture supernatant collected after centrifugation at 4,000 × g for 20 min at 4oC, was subjected to vacuum filtration through a 0.22 µm Steritop filter (Millipore), followed by a 0.1 µm filter (Sarstedt AG & Co., Germany) to obtain TdCM. OBGM and OB:PBS (OBGM mixed

with PBS in a 1:1 volume ratio) were used as the negative controls, while OBGM mixed with TdCM in a 1:1 volume ratio, labelled as OB:CM, was the positive control. 2.6.1 Characterisation of T. denticola stimulatory factors Several treatments were performed on TdCM to determine the physical stability of T. denticola stimulatory factors (1) TdCM was heat-treated in a 65°C incubator for 20 h to examine its heat stability. (2) TdCM was stored at –30°C for three months to examine the long term storage stability of TdCM in frozen solution. (3) TdCM was frozen at –80°C, then lyophilised O/N to determine the volatility of the TdCM stimulatory effect. (4) TdCM was treated with 50 µg/mL Proteinase K (ProtK; Sigma) at 42oC for 20 h. ProtK was removed by 5 kDa tangential flow filtration (TFF) and the filtrate (TdCMProtK) was collected in order to determine the susceptibility of T. denticola stimulatory factors to ProtK degradation. These differently treated TdCM were mixed with OBGM in a 1:1 volume ratio for P. gingivalis growth. 2.6.2 Size filtration of TdCM For quantification of amine groups in the metabolomics profiling of P. gingivalis, TdCM was subjected to 5,000 × g centrifugal filtration through a 10 kDa molecular weight cut- off (MWCO) Amicon Ultra-15 centrifugal filter unit (Millipore) for 30 min at room temperature. The filtrate was then subjected to centrifugation through a 3 kDa MWCO Amicon Ultra-15 centrifugal filter unit (Millipore) at room temperature for 45 min using a JA-12 rotor (Beckman CoulterTM Inc.). The filtrate from the 10 kDa MWCO centrifugal filter would consist of molecules of <10 kDa in size, whereas the retentate and filtrate of the 3 kDa MWCO centrifugal filter would contain molecules of <10 kDa and <3 kDa, respectively. The retentate from the 10 kDa and 3 kDa centrifugal filters were annotated as TdCM10R and TdCM3-10F respectively, whilst the 3 kDa filtrate of TdCM was denoted as TdCM3F. The volume of these different size-fractionations of TdCM varied depending on the centrifugal filter unit used; the OBGM was mixed with the TdCM10R, TdCM3-10F and TdCM3F fractions in volume ratios of 31:5, 11:7 and 13:5 respectively to obtain OB:CM10R, OB:CM3-10F and OB:CM3F for the metabolomics analysis.

2.6.3 Preparation of a simplified form of TdCM OBGM was filtered through a 10 kDa MWCO Amicon Ultra-15 centrifugal filter (Millipore) by centrifugation at 5,000 × g at room temperature for 30 min using a JA-12 rotor (Beckman CoulterTM Inc.) to collect OB_10F, which was then used to grow T. denticola to generate TdCM(OB_10F). In addition, a MinimateTM TFF (Pall Corporation) with MinimateTM TFF capsule of size 5 kDa cut-off was used to filter OBGM in order to reduce the complexity of the medium used for growth of T. denticola. The 5 kDa TFF capsule was setup with a pump (Masterflex® Console Drive, Cole-Parmer, USA) with pressure below 2 bar and the system was first flushed with DIW, followed by preconditioning the system with OBGM before starting to collect the filtrate, OB_5F. OB_10F and OB_5F were treated with 50 µg/mL recombinant ProtK (Sigma) at 42°C for 20 h, then ProtK was removed by another round of 10 kDa centrifugal filtration or 5 kDa TFF to obtain OB_10F_ProtK and OB_5F_ProtK. The conditioned medium resulting from T. denticola growth for 7 days in OB_10F, OB_10F_ProtK, OB_5F or OB_5F_ProtK was designated as TdCM(OB_10F), TdCM(OB_10F_ProtK), TdCM(OB_5F) and TdCM(OB_5F_ProtK) respectively. Furthermore, TdCM(OB_5F) and TdCM(OB_5F_ProtK) were filtered through a 5 kDa TFF to obtain their respective 5 kDa filtrates, which were called TdCM5F(OB_5F) and TdCM5F(OB_5F_ProtK). Finally, the MinimateTM TFF capsule was rinsed with 500 mL DIW and conditioned with 0.1 M NaCl for 20 min before storage at 4oC. A flow chart representing the steps taken for the preparation of different TdCM is shown in Figure 2.1.

Figure 2.1: Flow chart representing the preparation steps for the minimised forms of TdCM and their designations. Filtrate of OBGM that had been filtered through a 10 kDa or 5 kDa MWCO filters, named OB_10F and OB_5F respectively, were used for T. denticola growth. In addition, OB_10F and OB_5F were also treated with ProtK to obtain OB_10F_ProtK and OB_5F_ProtK repectively, for T. denticola growth. TdCMs collected after T. denticola 7 days growth in OB_5F and OB_5F_ProtK were further filtered through a 5 kDa MWCO TFF to generate TdCM5F(OB_5F) and TdCM5F(OB_5F_ProtK). These TdCMs were mixed with OBGM in 1:1 volume ratio for experimental testing of their stimulatory effects on P. gingivalis. 2.6.4 Reversed-phase high-performance liquid chromatography Molecules in OB_5F and TdCM5F(OB_5F) were separated using reversed-phase high- performance liquid chromatography (RP-HPLC), which was accomplished on an Agilent 1100 ® series instrument equipped with a Vydac silica C18 5 µm particle size analytical column with 4.6 mm internal diameter (i.d) × 250 mm for initial screening or semipreparative column with 10 mm i.d × 250 mm (Grace Davison Discovery Sciences, W.R. Grace & Co., USA) for fractionation of TdCM5F(OB_5F). The signal of eluent was monitored by a UV-Vis detector at wavelength of 220 nm, coupled with Agilent ChemStation software. Samples of 1.9 mL were injected with a flow rate of 3.5 mL/min and varying gradients of Buffer B [80% (v/v) acetonitrile in 0.1% (v/v) trifluoroacetic acid] were used for sample elution. The gradient for elution is displayed in Table 2.3. Fraction 1 with retention time (RT), RT 3-13 min, Fraction 2 with RT 13-23 min, Fraction 3 with RT 23-33 min, Fraction 4 with RT 33-43 min and Fraction 5 with RT 43-50 min were manually collected. Two specific fractions around RT 18 min and 38 min that contained distinct peaks in TdCM5F(OB_5F), but not in OB_5F, were also

collected separately. Fractions collected were lyophilised and resuspended in 1.9 mL of 1 × PBS, the pH of each fraction was measured and adjusted to ~7.4 using 1 M KOH. The fractions were filtered with 0.22 µm filters (Millipore) for further testing. Table 2.3: Parameters for the flow gradient of solvent B during the separation of molecules in TdCM5F(OB_5F) using semipreparative C18 RP-HPLC. The maximum pressure was maintained at 220 bar.

Step Time (min) % Solvent B Flow Rate (mL/min) 1 0 5 0.1 2 1.0 5 3.5 3 4.5 5 3.5 4 39.5 40 3.5 5 40.0 100 3.5 6 47.0 100 3.5 7 49.0 5 3.5 8 59.5 5 3.5 9 60.0 5 0.1

2.7 Determination of free glycine concentration

2.7.1 Glycine enzyme-linked immunosorbent assay Two 1 mL aliquots of P. gingivalis cultures (40 mL) were simultaneously collected at various time points, t = 0, 20, 30, 40, 50 h. Each sample was snap-frozen with liquid nitrogen and stored at -80°C for glycine quantification; while the other was used for the measurement of OD650. Similarly, 100 µL and 60 µL sample aliquots were collected for glycine quantification and OD650 measurements using a 50 mm window quartz cuvette (Variance, Cary) respectively, from P. gingivalis grown in 2 mL cultures. Quantification of free glycine using a glycine enzyme-linked immunosorbent assay (ELISA) K0713 (ImmunoDiagnostik, Germany) or glycine ELISA BAE-2100 (Labor Diagnostika Nord GmbH & Co.KG, Germany) kit supplied with all required reagents were performed as per the manufacturer’s instructions. Briefly, all samples, standards and controls were diluted with DIW. Aliquots of diluted samples were derivatized with a cross-linking agent in dimethylsulfoxide at room temperature by shaking at approximately 600 rpm (Micromixer MX4, Bacto Laboratories Pty Ltd, NSW, Australia) for 2 h. All downstream shaking steps were consistently performed at approximately 600 rpm at room temperature. The derivatized

samples were reduced with reducing solution provided prior to transfer to the antigen precoated solid phase on the 96-well strips provided, then glycine anti-serum (rabbit anti-glycine antibody) was added into each well for O/N incubation at 4oC. The ELISA plates were washed thrice with 300 µL 1 × wash buffer (consisting of non-ionic detergent at physiological pH) using a Model 405TM LS Microplate Washer (BioTek, Millennium Science New Zealand Pty. Ltd.), by which 70 mL buffer was first primed through the entire system before usage. To each well was then added 100 µL of goat anti-rabbit immunoglobulins conjugated with peroxidase and incubated at room temperature for 30 min with shaking. The ELISA plates were then washed to remove free antigen and free antigen-antiserum complexes. Then, anti-rabbit IgG- peroxidase conjugate with 3,3’,5,5’-tetramethylbenzidine (TMB) chromogenic substrate was added for 20 min incubation with shaking. The reaction was quenched by the addition of 100 µL of 0.25 M sulfuric acid, which caused a color change from light blue to yellow. The reaction plate was measured at A450 using a microplate reader (Wallac Victor3 model 1420, Perkin Elmer, USA) with Victor Workstation software Version 3.2. The 4-parameter logarithm analysis was performed using MyAssay (https://www.myassays.com/). 2.7.2 Liquid chromatography-triple quadrupole mass spectrometry analyses Amine group quantification before and after P. gingivalis growth in OBGM, OB:CM, OB:CM10R, OB:CM3F, OB:CM3R and OB:PBS was performed using liquid chromatography triple quadrupole mass spectrometry (LC-QQQ-MS) using protocols described in Boughton et al. (2011) and Tan et al. (2014). Firstly, 20 µL of bacterial culture was mixed with 80 µL of 13 cold methanol:chloroform (3:1) containing 83.88 µM internal standards of C3-alanine. The mixture was vortexed and centrifuged at 18,000 × g (Beckman CoulterTM Microfuge® 22R Centrifuge) at -10°C for 10 min. The supernatant was collected and added to 40 µL of DIW, resulting in 1:3:3 chloroform:methanol:water biphasic partition, this mixture was centrifuged at 18,000 × g at -10°C for 10 min. Then, 90 µL of the upper aqueous phase was diluted 10-fold in 50% (v/v) methanol. A 10 µL aliquot of diluted sample was derivatized with borate buffer containing 35.7 µM internal standard 13C, 15N-valine and added with 20 µL of 10 mM 6- aminoquinolyl-N-hydrosysuccinimidyl carbamate derivatisation reagent. The samples were incubated at 55oC for 10 min at 750 rpm (Eppendorf Thermomix® C) and centrifuged at 0oC for 5 min at 18,000 × g. The supernatants were collected for analysis by injecting 4 µL of samples for LC-QQQ-MS analysis. Analyses were performed using Mass Hunter Version 7 software and graphical presentations of data were performed using Excel.

2.8 RNA sequencing

Log phase cells of P. gingivalis W50 (OD650 ~0.5) in OBGM and OB:CM were added to 0.2 vol of 5% (v/v) phenol in absolute ethanol at room temperature and centrifuged at 8,000 × g for 15 min at room temperature. Following removal of supernatant, each cell pellet was snap-frozen with liquid nitrogen before storage at -80°C. Total RNA extraction was performed using TRIzolTM Reagent (Ambion, Invitrogen). Briefly, 1.4 mL TRIzol was added to each frozen pellet consisting ~1010 P. gingivalis cells and then transferred to 2 mL tubes with prechilled glass beads (MP BioMatrix B). The samples were homogenized in a Precellys24 homogenizer (Bertin Corp., USA) at 6500 rpm for 23 s twice with a 5 s pause between cycles and then quickly chilled on ice for 2 min. Samples were then mixed by inversion at room temperature for 5 min to allow complete dissociation of nucleoprotein complexes. The samples were centrifuged at 12,000 × g for 5 min and the resulting supernatant was mixed with 20% (v/v) chloroform per 1 mL of TRIzolTM Reagent. All following centrifugations were performed at 12,000 × g and 4°C unless stated otherwise. The tubes were shaken vigorously by hand for 15 s, incubated at room temperature for 3 min and then centrifuged for 10 min. The mixture separated into three phases and the upper colorless aqueous phase was collected and mixed with 50% (v/v) isopropanol per 1 mL of TRIzolTM Reagent used. The mixture was incubated for 10 min at room temperature and centrifuged. The isopropanol was removed and the nucleic acid pellet was washed with 75% ethanol by centrifugation at 7,600 × g for 5 min, then the pellet was air-dried before resuspending in 60 µL nuclease-free water. RNA extracted with TRIzolTM was subjected to rigorous DNase treatment according to the Turbo DNA-freeTM protocol (Ambion, Life Technologies), then the sample was subjected to column clean-up using the NucleoSpin RNA II kit (Macherey-Nagel GmbH & Co. KG, Germany) as per the manufacturer’s instructions. To prevent degradation of RNA, Ribosafe RNase inhibitor (Bioline) was added to each RNA sample following manufacturer’s instructions. RNA sample quality control was performed to check for gDNA contamination by subjecting the RNA samples to 35 cycles of PCR using PG0445 and PG1605confo forward and reverse primer pairs (Table 2.2), the PCR products were then examined on a 0.8% (w/v) agarose gel following electrophoresis at 100 V for 30 min, The RNA A260/280 and A260/230 ratios were examined using a NanoDrop Spectrophotometer. RNA integrity, quality score and quantification were verified using the DNA 5K/RNA Charge Variant Assay Labchip (PerkinElmer) on the LabChip GXII Touch 24 instrument (Perkin Elmer). The preparation of reaction and samples was based on the low throughput, HT RNA High Sens assay according

to the manufacturer’s instruction. The quality and quantity of RNA was reviewed using the LabChip GX Reviewer. RNA samples with an RNA integrity score above 8.0 were submitted for analysis. RNA samples were diluted to 450 ng/µL based on the spectrophotometric calculations. Then, 35 µL RNA aliquots were added to RNAstable (Biomatrica, Inc.) and vacuum dried using a rotational vacuum concentrator (RVC 2-25 CDplus, Martin Christ Gefriertrocknungsanlagen GmbH, Germany) without heating for 2.5 h. The dried RNA was stored at room temperature and protected from moisture with desiccant packs in a moisture barrier heat sealed bag for the shipping of RNA. RNA library preparations using Epicentre ScriptSeq Complete Bacteria and NextSeq Illumina sequencing in Mid-Output mode and single-read 75 bp format were performed at Micromon (Monash University), whereas the data analysis was performed by the Monash Bioinformatics Platform (Monash University). 2.9 Bioinformatics Resources in the National Center for Biotechnology Information (NCBI; https://www.ncbi.nlm.nih.gov/) were utilised, especially the BLASTp suite and Conserved Domain Database (CDD) for determination of protein homologues and their conserved domains (Marchler-Bauer et al. 2015, Marchler-Bauer et al. 2017). The Microbial Transcriptome Database (MTD: http://bioinformatics.forsyth.org/mtd/) and MicrobesOnline (http://www.microbesonline.org/) were referred to for examination of P. gingivalis W83 transcriptional profiles and gene information (Alm et al. 2004, Arkin 2011, Høvik et al. 2012). The MEROPS Peptidase Database Release 12.0 was used as the primary resource for predictions of bacterial proteases and peptidases (Rawlings et al. 2014). The database for Annotation, Visualization, and Integrated Discovery (DAVID) online software was used for functional annotation of genes, gene clustering and pathways analysis (Dennis et al. 2003). PSORTb, SignalP and Phobius were employed for the predictions of protein cellular localisation, signal peptide and transmembrane topology (Käll et al. 2007, Yu et al. 2010, Petersen et al. 2011). The BIOCYC database collection and Kyoto Encyclopedia of Genes and Genome (KEGG) were employed for metabolic pathways comparative and enrichment analysis (Caspi et al. 2016, Kanehisa et al. 2016). Predicted Prokaryotic Transcription Factors (P2TF) database was referred to for the prediction of P. gingivalis W83 s transcription regulatory factors (Ortet et al. 2012). Multiple sequence alignments (MSA), followed by phylogenetic tree construction of PG1605 and PG1788 with their candidate homologues were performed using ClustalWS (Sievers and Higgins 2014) and neighbour-joining parameters respectively in

Jalview (Barton et al. 2009). Protein structural modelling was performed using iterative threading assembly refinement, ITASSER (Yang et al. 2015), while PyMol (The PyMOL Molecular Graphics System, Version 1.7.4 Schrödinger, LLC.) was used for protein structural visualizations. 2.10 Statistical analysis Data visualization and statistical analysis were performed using R (Version 3.4.2) or with Excel. The normality and homogeneity of residuals were assessed by the Shapiro-Wilk test and Bartlett’s test respectively (Bartlett 1937, Shapiro and Wilk 1965). If residuals were normally distributed and with equal variance, then one-way Analysis of Variance (ANOVA) was performed, followed by the post-hoc Tukey test (Tukey 1949) for determination of pairwise differences between samples. Residuals that showed non-standard distribution and unequal variance were subjected to a non-parametric counterpart of the one-way ANOVA – the Kruskal-Wallis test (Kruskal and Wallis 1952). If the p-value was <0.05, then post-hoc analysis of Conover-Iman test was performed (Conover and Iman 1981). The null hypotheses are rejected, suggesting a difference between variables, at a statistical significance level of 0.05. Data transformation by log (glycine concentration ratio) was performed to reduce the variance of samples and the transformed data subjected to similar statistical analysis for determination of the sample significance level.

Chapter 3 Characterisation of T. denticola factors that stimulate free glycine production by P. gingivalis

3.1 Introduction Bacterial cells often live in polymicrobial biofilm settings that constitute highly diverse microbial communities. Oral bacterial species in human subgingival plaque demonstrate interdependencies and multimodal interactions with their mutualistic, synergistic and/or antagonistic partners, as well as host interactions, to achieve homeostasis in the environment (Kolenbrander et al. 2006, Kolenbrander et al. 2010). There are different modes of social interactions that lead to an overall increase of growth, biomass and fitness advantageous to a polymicrobial community. These include (1) physical association for adherence, colonisation and establishment, (2) metabolic cooperation, such as cross-feeding, syntrophy and sharing of public goods and (3) communal dependency by releasing antimicrobial agents and matrix by- products into the environment that confer additional protection and food sources to the members of polymicrobial community (Kuramitsu et al. 2007, Yung-Hua et al. 2017). The above-described mechanisms support mutualistic relationships between different bacterial species. However, there are also competitive or exploitative interactions that exist among bacterial species. For example, the production of antimicrobial substances that act as a defence mechanism against antagonistic partners or the secretion of toxic metabolic by-products that result in growth suppression or inhibition of bacterial species. Often, bacterial social interactions are not straightforward; mounting evidence shows the involvement of both cooperative and competitive interactions among the same interacting species. For instance, Streptococcus gordonii showed cross-feeding of L-lactate to Aggregatibacter actinomycetemcomitans, which is its preferred carbon source (Brown and Whiteley 2007). However, S. gordonii also produced high levels of the antimicrobial product hydrogen peroxide, H2O2, which induced oxidative stress to A. actinomycetemcomitans (Stacy et al.

2014). In response to the H2O2 produced by S. gordonii, A. actinomycetemcomitans produced catalase and dispersin B enzymes as its defensive and protective mechanisms to alleviate the oxidative stress effects, as well as to mediate spatial separation from its synergistic partner (Stacy et al. 2014). These regulated multimodal interactions improved the fitness advantage of A. actinomycetemcomitans as S. gordonii enhanced the availability of oxygen to act as electron

acceptors for a shift from low-energy yielding fermentation to respiratory metabolism in the pathogen (Stacy et al. 2016). P. gingivalis and T. denticola exist as components of a pathogenic polymicrobial subgingival biofilm that contributes substantially to the progression of chronic periodontitis. In vivo studies demonstrated that P. gingivalis and T. denticola dual-species inocula enhanced virulence in a mouse periodontitis model compared with monospecies inoculation (Orth et al. 2011). In vitro experiments with P. gingivalis, T. denticola and Tannerella forsythia showed an increase in the biomass, thickness and coverage area of the polymicrobial biofilm compared with their monospecies counterparts in a flow cell setting (Zhu et al. 2013). Molecular and mechanistic investigations revealed that P. gingivalis and T. denticola cooperate metabolically and possess multivalent physical binding sites for interspecies cell-cell adherence. These species exhibit metabolic interactions by fatty acids cross-feeding (Grenier 1992), as well as thiamine pyrophosphate cofactor and glycine exchange from P. gingivalis to T. denticola (Tan et al. 2014). Synergistic interactions between T. denticola and P. gingivalis demonstrated in numerous studies are reflective of their ecological roles in dysbiotic subgingival biofilms where the relative abundance of T. denticola and P. gingivalis increase during chronic periodontitis progression (Byrne et al. 2009). Therefore, it is of interest to gain a better understanding of the underlying cooperative mechanisms that support their synergistic interactions. The hypothesis that T. denticola releases signalling molecules or stimulatory factors to engage P. gingivalis to produce free glycine, a preferred amino acid source for T. denticola, was proposed based on the work of Tan et al. (2014). In summary, Tan et al. (2014) showed upregulation of the T. denticola glycine reductase and cleavage pathways during T. denticola and P. gingivalis coculture in a continuous system, in comparison with monoculture. This result suggested an increased utilisation of glycine by T. denticola under these conditions. Furthermore, the concentration of free glycine increased during growth of P. gingivalis in a 1:1 volume ratio mixture of OBGM with cell-free T. denticola conditioned medium (TdCM), termed OB:CM. It was also shown that P. gingivalis exhibited more complete or faster hydrolysis of glycine-containing peptides leading to an increase of free glycine during growth in OB:CM (Tan et al. 2014). Proposed mechanisms for P. gingivalis to increase free glycine content in OB:CM are described in detail in Fig. 3.1. In this study, TdCM was simplified and fractionated using reversed-phase high- performance liquid chromatography (RP-HPLC) to examine the stimulatory effects of each fraction that contribute to P. gingivalis free glycine production. Furthermore, the absolute free glycine concentration in different molecular size fractions of TdCM were determined using 58

liquid chromatography coupled with triple-quadrupole mass spectrometry (LC-QQQ-MS). LC- QQQ-MS analysis also provided a high-throughput screening of the changes of amine group profiles in the different media after P. gingivalis growth. These approaches indicated that TdSF consist of non-sequence specific peptides, which could be hydrolysed by P. gingivalis peptidases to release free glycine. These data also provided useful insights into the fate of amino acids and other biogenic amine (BA) groups in the media that are affected by both T. denticola and P. gingivalis.

Figure 3.1: Diagrammatic descriptions of possible scenarios where T. denticola stimulates P. gingivalis to increase free glycine production. (1) P. gingivalis (purple coccobacilli) produces free glycine by upregulation of its glycine biosynthetic pathways in response to T. denticola (green spirochaete) signalling molecule(s). (2) P. gingivalis increases expression or activities of proteases upon sensing T. denticola signalling molecules to increase the hydrolysis of environmental peptides, thereby releasing free glycine. (3) Glycine-containing oligopeptides partially hydrolysed by T. denticola proteases become further hydrolysed by an extensive array of P. gingivalis extracellular peptidases, thereby releasing free glycine. Red spheres represent glycine residues, green spheres represent unspecified amino acids in T. denticola produced peptides, cyan spheres represent the T. denticola produced signalling molecules and blue spheres represent unspecified amino acids in existing peptides in the environment. The green and purple pacman in symbols indicate different extracellular proteases of T. denticola and P. gingivalis that hydrolyse the peptides in the growth medium. Purple wide arrows represent the signal molecule receptors on P. gingivalis. The orange arrows indicate the potential target sites of molecules, while the black and double headed black arrows indicate the directionality of peptides hydrolysed by bacterial proteases.

1 3.2 Results

2 3.2.1 P. gingivalis growth and glycine production 3 Growth curves of P. gingivalis in different growth media were constructed by

4 monitoring the OD650 of P. gingivalis at t = 0, 20, 30, 40, 50, 70 h (Fig. 3.2) in OBGM, OB:PBS 5 (OBGM mixed with PBS in a 1:1 volume ratio) and OB:CM (OBGM mixed with TdCM in a 6 1:1 volume ratio). In OBGM, P. gingivalis showed a lag phase over the first 20 h, achieved a 7 log growth phase between 20 to 30 h, followed by stationary phase (Fig. 3.2). During the 40 to 8 70 h period, the optical density of the P. gingivalis culture declined, which was likely due to 9 cell lysis. Growth of P. gingivalis in OB:CM, in comparison to OB:PBS, showed that TdCM 10 initially had a negative effect on the growth of P. gingivalis. As shown in Fig. 3.2, P. gingivalis

11 growth plateaued at a cell density equivalent to an OD650 of approximately 0.6 after 30 h of 12 growth in OB:PBS that has half the nutrient content of OBGM, whereas P. gingivalis grown in 13 OB:CM showed a lower cell density at the end of lag phase from 0 to 20 h and achieved log 14 growth phase between 20 to 40 h after inoculation. Growth of P. gingivalis in OB:CM

15 eventually plateaued at a cell density equivalent to an OD650 of approximately 1.0 (Fig 3.2). 16

2.0 1.8

1.6 )

650 1.4 1.2 1.0 0.8

Cell density (OD densityCell 0.6 0.4 0.2 0.0 0 10 20 30 40 50 60 70 17 Time (h) 18 Figure 3.2: P. gingivalis growth curve in different media (a) Growth curve of P. gingivalis in 19 OBGM (black line), OB:PBS (grey line) and OB:CM (red line) generated in this current study. 20 Each data point was represented by n = 10 for OBGM, n = 3 for OB:PBS and n = 14 for 21 OB:CM, and the error bars represent the standard deviation of OD650. 22 23 To examine the reliability and accuracy of the ELISA kit for glycine measurement in 24 this system, exogeneous free glycine was supplemented into TdCM, which was depleted of free

1 glycine. Then, the concentration of free glycine in TdCM determined by the ELISA kit was 2 plotted against the expected free glycine concentration in TdCM. The measured free glycine 3 concentration was presented in a linear regression line with a slope of 0.9892 to the expected 4 free glycine concentration in TdCM (Fig. 3.3), implying that the free glycine measurement 5 using the ELISA kit was accurate as the data correlated to the expected free glycine 6 concentration in the background medium. The low-level of cross-reactivity of antibody 7 provided by the ELISA kit against other substances i.e. D-serine, L-cysteine, β-Alanine, γ- 8 aminobutyric acid (GABA) (as stated by the manufacturer) could be the factor that contributes 9 to the y-intercept being higher than zero. Thus, the equation derived from this standard curve 10 was applied to all glycine data collected to account for this deviation in free glycine 11 quantification.

12 13 Figure 3.3: Average free glycine measurement using a glycine ELISA kit. TdCM, which was 14 depleted of free glycine was supplemented with exogeneous free glycine at 0.5, 1, 1.5, 2, 3, 4 15 and 5 mM concentrations. The concentration of free glycine was determined from the glycine 16 standards provided by the kit. Linear regression of the measured glycine concentration was 17 plotted against the expected free glycine concentration. The dotted line represents the linear 18 regression trendline and the slope of the regression line was 0.9892 as indicated in the linear 19 equation. The error bars represent the standard deviation of glycine concentration measured. 20 Three independent experiments were performed to collect the data points. 21 22 Free glycine concentration at the t= 0, 20, 30, 40 and 50 h in various media 23 combinations were measured directly using glycine ELISA kits (Fig. I.1). P. gingivalis showed 24 the highest average increase of glycine from 30 to 40 h time points in OB:CM (Fig. I.1), which 25 resulted in an approximately 2.8-fold higher free glycine ratio based on the free glycine 26 concentration determined at 40 h relative to its basal concentration (Fig. 3.4). For this reason,

1 t = 40 h was selected as the time point for glycine measurement in subsequent experiments. P. 2 gingivalis produced an average of 1.5- and 1.2-fold higher glycine ratio after 40 h growth in 3 OBGM and OB:PBS, respectively (Fig. 3.4). Based on the statistical analysis of the free glycine 4 ratio between samples, OB:CM was shown to be significantly (adjusted p-value <0.001) higher 5 than OBGM and OB:PBS (Fig. 3.4). Further, the rate of free glycine production relative to P. 6 gingivalis cell number in OB:CM was 1.8- and 2.7-fold higher than in OBGM and OB:PBS 7 respectively (Fig. I.2). However, P. gingivalis exhibited a different cell growth phase in 8 OB:PBS than other media reaching stationary phase after 30 h growth in OB:PBS. Thus, 9 OB:PBS was not included as the negative control medium in the following experiments. 10 OB:CM was used as the positive control, while OBGM represented the negative control 11 medium.

12 13 Figure 3.4: Ratio of glycine concentration after P. gingivalis 40 h growth, relative to basal 14 glycine concentration in OBGM, OB:CM and OB:PBS. These data were pooled from all 15 independent experimental set-ups to illustrate the reproducibility and consistency of the 16 positive and negative controls used, where n = 10 for OBGM, n = 14 for OB:CM, and n = 3 17 for OB:PBS, error bars represent the standard deviation. One-way ANOVA and post-hoc 18 analysis of Conover-Iman test were performed for statistical significance analysis of the data. 19 The “*” symbol represents an adjusted p-value <0.001 between pair-wise comparison of 20 samples.

21 3.2.1.1 Nature of T. denticola stimulatory factors

22 As a first step towards identifying the TdSF in TdCM that were responsible for the 23 enhancement of free glycine production by P. gingivalis, the physical stability and 24 characteristics of the TdSF were examined, as shown in the schematic representation in Fig. 25 3.5. TdCMs were treated differently, as follows (1) Proteinase K (ProtK) hydrolysis, followed 63

1 by removal of ProtK using a 5 kDa tangential flow filtration (TFF) to produce a 5 kDa filtrate 2 of TdCMProtK, (2) heat treatment at 65°C, (3) lyophilisation and (4) freezing at -30°C for three 3 months. Frozen and thawed TdCM resulted in precipitation of the medium and the sample was 4 unable to be used for downstream experiment. 5

6 7 Figure 3.5: Schematic diagram of the steps performed for different treatments of T. denticola 8 conditioned medium (TdCM). TdCM was obtained by growth of T. denticola in OBGM for 7 9 days. (a) TdCM was treated with ProtK for 20 h by incubation at 42°C, ProtK was later removed 10 by 5 kDa centrifugal filtration. (b) TdCM was heat treated at 65oC for 20 h. (c) TdCM was 11 frozen at -80°C and freeze-dried for storage. Lyophilised TdCM was reconstituted by 12 resuspension in equivolume of 1 × PBS. (d) TdCM was frozen at -30°C for three months. 13 These TdCMs were mixed with OBGM in 1:1 volume ratios for experimental testings.

14 The treated TdCMs were mixed with OBGM in 1:1 volume ratios. These media did 15 not affect P. gingivalis growth, as compared with OB:CM (Fig. 3.6). Lyophilisation and heat 16 treatment of TdCM did not have a significant effect on TdSF (Fig. 3.7), suggesting that TdSF 17 are heat stable and can be stored in the lyophilised form. TdCMs that had been treated with 18 ProtK were unable to stimulate P. gingivalis to produce free glycine after 40 h growth (Fig. 19 3.7). However, it should be noted that the initial average glycine content in OB:CMProtK_5F 20 (1.40 ± 0.06 mM) was higher than any other medium combination, including OB:CM (0.94 ± 21 0.29 mM). This suggested that the free glycine is released from the peptide-bound form as a 22 result of ProtK hydrolysis. Taken together, these data suggested that TdSF likely consist of 23 glycine-containing peptides in TdCM.

Figure 3.6: Growth of P. gingivalis in OB:CM mixture, where TdCMs were subjected to different experimental treatments, including freeze-dried (CMFD, dark brown line), heat treated TdCM at 65°C for 20 h (CMheat, yellow line), or treatment of TdCM with ProtK at 42°C for 20 h and the ProtK was then removed by 5 kDa TFF (CMProtK_5F, purple line) before use. Each of the treated TdCMs were mixed in a 1:1 volume ratio with OBGM. There were four biological replicates for OB:CMFD, while the other samples were made of biological triplicates. The OB:CM (red line) from Fig. 3.2 were included for comparison.

Figure 3.7: Ratio of glycine concentration after P. gingivalis 40 h growth in differently processed TdCM to their respective basal glycine concentration. OBGM was mixed in a 1:1 volume ratio with TdCM that had been treated in the following manner: frozen at -80°C and freeze-dried (OB:CMFD), heated at 65°C (OB:CMheat) and treated with ProtK at 42°C, followed by 5 kDa TFF for the removal of ProtK (OB:CMProtK_5F). There were four biological replicates for OB:CMFD, while the other samples were made of biological triplicates. The error bar of the graph represents the standard deviation. OB:CM from Fig. 3.4 were included for comparison. The “*” symbol represents an adjusted p-value <0.01 between pair-wise comparison of sample in the statistical analysis of samples using post-hoc analysis of Conover-Iman test.

3.2.1.2 Simplified forms of TdCM To enable identification of T. denticola stimulatory peptides, the complexity of the media used to grow T. denticola was simplified. Firstly, OBGM was filtered through a 10 kDa centrifugal device or 5 kDa tangential flow filtration (TFF) to obtain OB_10F and OB_5F respectively (Fig. 3.8). This step resulted in the removal of serum albumin (66.5 kDa), which is a major protein source in OBGM, and other protein sources in the media that are larger than the molecular weight cut-off (MWCO) of the filter used, while retaining oligopeptides as the carbon and energy sources for T. denticola growth. Then, T. denticola was grown in each of the size-filtrates of OBGM to obtain TdCM(OB_10F) and TdCM(OB_5F). In addition, T. denticola was grown in aliquots of OB_10F and OB_5F that were pretreated with ProtK to degrade protein and peptide substrates in the filtered media, producing TdCM(OB_10F_ProtK) and TdCM(OB_5F_ProtK) (Fig. 3.8). The resultant TdCMs obtained by growth of T. denticola in OB_5F and OB_5F_ProtK were further filtered via a 5 kDa TFF to remove proteins produced by T. denticola that might have a contributory role to glycine production. These TdCMs were called TdCM5F(OB_5F) and TdCM5F(OB_5F_ProtK) respectively.

Figure 3.8: Preparation of simplified forms of TdCMs and their annotations. Filtrate of OBGM that had been filtered through a 10 kDa or 5 kDa filters, named OB_10F and OB_5F respectively, were treated with ProtK to obtain OB_10F_ProtK and OB_5F_ProtK, which were used for T. denticola growth. TdCMs collected after T. denticola 7 days growth in OB_5F and OB_5F_ProtK were further filtered through a 5 kDa TFF to generate TdCM5F(OB_5F) and TdCM5F(OB_5F_ProtK). These TdCMs were mixed with OBGM in 1:1 volume ratio for experimental testing of their stimulatory effects on P. gingivalis. This flow chart was replicated from Fig. 2.1.

As previously shown, OB:CM contains growth suppressing molecules that have a negative effect on P. gingivalis initial growth (Fig. 3.2). Growth of P. gingivalis in all TdCM derivatives mixed with OBGM also had a similar cell density at the lag phase (Fig. 3.9). This observation reinforced that growth of T. denticola in different OBGM derivatives consistently resulted in the production of a growth suppressing molecule(s) against P. gingivalis. Interestingly, P. gingivalis achieved the highest growth rate in OB:CM(OB_10F_ProtK) and reached the highest OD650 of 1.0 after 40 h growth, as compared with other TdCMs (Fig. 3.9). This could be due to an increased bioavailability of peptide substrates for P. gingivalis utilisation in the higher molecular weight fraction of OBGM after ProtK treatment.

1.4

1.2

1.0

) 0.8 650

0.6

0.4

Cell density (OD densityCell 0.2

0.0 0 10 20 30 40 50 -0.2

-0.4 Time (h)

Figure 3.9: P. gingivalis growth curve in TdCMs obtained by growth of T. denticola in modified OBGM. Different TdCMs were mixed with OBGM in a 1:1 volume ratio, consisting of OB:CM(OB_10F) (dark red line), OB:CM(OB_5F) (dark blue line), OB:CM(OB_10F_ProtK) (orange line), OB:CM(OB_5F_ProtK) (cyan line), OB:CM5F(OB_5F) (green lines) and OB:CM5F(OB_5F_ProtK) (yellow line) respectively. The OB:CM (red line) from Fig. 3.2 were included for comparison. The data points were represented by an average of biological triplicates.

All TdCMs tested stimulated P. gingivalis to produce a >1.5-fold glycine ratio at 40 h compared with 0 h, except for P. gingivalis growth in OB:CM(OB_5F_ProtK), where only an approximate 1.4-fold glycine ratio was achieved (Fig. 3.10). Statistical pair-wise comparison of these data indicated that P. gingivalis growth in OB:CM(OB_5F_ProtK) and OB:CM5F(OB_5F_ProtK) resulted in a significantly lower ratio of free glycine, compared with the positive control medium, OB:CM. This result also suggested that other simplified forms of TdCM derivatives could be used for downstream fractionation analysis. Based on this study, TdCM5F(OB_5F) contributed to an increase in P. gingivalis free glycine production and was simpler to produce for large scale experiments, thus this simplified TdCM was selected for further fractionation and analysis.

Figure 3.10: Average ratio of free glycine concentration after 40 h of P. gingivalis growth in different TdCMs mixtures. Data representing OB:CM were replicated from Fig. 3.4, while each of the other samples were represented in biological triplicates. The error bars represent the standard deviation. The “*” symbol represents an adjusted p-value <0.01 between pair-wise comparison of sample using post-hoc analysis of Conover- Iman test.

3.2.1.3 TdCM_5F(OB_5F) profile in RP-HPLC fractionation To substantiate the nature and source of TdSF, a simplified form of TdCM was used for further fractionation. As TdCM5F(OB_5F) exhibited a comparable stimulatory effect to OB:CM, TdCM5F(OB_5F) was fractionated via a C18 reversed-phase high-performance liquid chromatography (RP-HPLC). The derived fractions were then used to determine their effect on P. gingivalis free glycine production (Fig. 3.11).

Figure 3.11: Overview of RP-HPLC fractionation of TdCM5F(OB_5F) for the free glycine analysis. TdCM5F(OB_5F) was fractionated using a preparative C18 RP-HPLC column and molecules with different retention time were all collected for testing. All fractions were freeze- dried and resuspended with PBS. P. gingivalis growth in different media was measured at OD650 using a spectrophotometer and the free glycine was quantified using a glycine ELISA kit.

Signal profiles between TdCM5F(OB_5F) and OB_5F medium control were compared by examining the differences in the absorbance at wavelength 220 nm (A220). The overlaid chromatograms of OB_5F and TdCM5F(OB_5F) showed that both media were complex and difficult to resolve into individual baseline peaks (Fig. 3.12a). The high A220 signal in the void volume indicated that many molecules, which can be composed of polar and charged macromolecules did not interact with the C18 stationary phase, resulting in direct elution of molecules from the column in the buffering system (Fig. 3.12). Although A220 is normally used for the detection of peptide bonds, other metabolites containing carboxyl and amino functional groups in the growth medium could also contribute to the far-UV range detection, such as biogenic amines (BAs) and fatty acids (Burr and Miller 1941, Schmid 2005). Therefore, it should be noted that the peak area displayed in the chromatogram may not directly correspond to the abundance and diversity of peptides found in the respective fractions as there may be other metabolites present. TdCM5F(OB_5F) showed an overall reduction in the absorbance readout as compared with OB_5F (Fig. 3.12a), suggesting that these molecules were diminished after T. denticola growth in OB_5F. Interestingly, three peaks in TdCM5F(OB_5F), which eluted at retention time (RT) 18–20 min (RT18) and 38 min (RT38) respectively, were distinct from the OB_5F medium control run (Fig. 3.12a). This observation suggested that T. denticola produced novel molecules during growth or alternatively there were peak shifts as a result of chemical modification or processing of the pre-existing molecules in OB_5F.

(a)

(b)

Figure 3.12: Chromatogram of OB_5F and TdCM5F(OB_5F). (a) Overlay chromatogram of OB_5F (blue) and TdCM5F(OB_5F) (red) that were generated by fractionation of molecules using an analytical column (b) C18 semi-preparative RP-HPLC A220 signal profile of TdCM5F(OB_5F). Fractions, F1, F2, F3, F4 and F5 were collected in 10 min time intervals, starting from t=3 min of separation as shown in the chromatogram. RT18 and RT38 were distinctive peaks absent from OB_5F and were collected specifically. Right axis %B represents the elution gradient of solvent B (80% acetonitrile and 0.1% trifluoroacetic acid).

Fraction F1, F2, F3, F4 and F5 were collected at retention time ranges of 3-13 min, 13- 23 min, 23-33 min, 33-43 min and 43-50 min respectively (Fig. 3.12b). Fractions F1 to F4 showed a high diversity and abundance of molecules that absorbed at wavelength 220 nm, whereas F5 was seemingly composed of two unresolved peaks with no obvious background absorbance. All fractions that had been lyophilised were resuspended in 1 × PBS and pH adjusted in volumes equal to the medium volume used per run. RT18 and RT38, which fell within fractions F2 and F4 respectively, were programmed to be collected automatically (Fig. 3.12b). Then, RT18 and RT38 were concentrated 3-fold relative to their initial injecting volume of TdCM5F(OB_5F) to test the efficacy of these peaks in stimulating P. gingivalis to produce glycine. Fraction 1 (F1) that showed a low initial pH (Table 4.1), which is likely due to residual trifluoroacetic acid (TFA), was not used for the following experiment. Other fractions were then used to determine its contribution to P. gingivalis free glycine production.

Table 3.1: pH of RP-HPLC fractions of TdCM5F(OB_5F). The lyophilised form of F1-F5 fractions were resuspended in 1 × PBS in equivolume as sample injection (2.9 mL/run), while RT18 and RT38 were three-fold concentrated in 1 × PBS. All of the resuspended fractions were pH adjusted to pH 7.4 using KOH before mixing with OBGM for P. gingivalis growth.

RP-HPLC Fraction Average pH F1 2.3 F2 6.2 F3 6.4 F4 7.0 F5 7.3 RT18 7.0 RT38 7.2

This experimental setup, which was downscaled to a 2 mL culture showed that P. gingivalis in OBGM reached an OD650 that was double the growth in OB:CM5F(OB_5F) (data not shown), with is consistent with the results obtained for the 40 mL setup. This experiment also demonstrated that growth of P. gingivalis in OB:CM was suppressed. The RP-HPLC fractions of TdCM_5F(OB_5F) mixed with OBGM did not exhibit growth suppressing effect towards P. gingivalis (Fig. 3.13). P. gingivalis growth in OB:CM and all RP-HPLC fractions of TdCM5F(OB_5F) mixed with OBGM resulted in a cell density equivalent to OD650 0.7–0.8 at 50 h (Fig. 3.13), hence this time point was used for comparison of the change in glycine by P. gingivalis in different fractions of TdCM5F(OB_5F).

Figure 3.13: Growth of P. gingivalis W50 in different RP-HPLC fractions of TdCM5F(OB_5F). All RP-HPLC fractions, F2 (cyan line), F3 (dark orange line), F4 (yellow line), F5 (green line), RT18 (pink dotted line), RT38 (green dotted line) were mixed with OBGM and P. gingivalis was inoculated in a total volume of 2 mL media, in comparison with the positive control OB:CM5F (red line) in a similar setting. Final cell density of P. gingivalis in OBGM was double that in OB:CM5F(OB_5F), but is not shown in this diagram.

There was a 1.94-fold increase in average glycine ratio after P. gingivalis 50 h growth in the positive control medium OB:CM5F(OB_5F) (Fig. 3.14), which exhibited an identical trend to the 40 mL culture. Growth of P. gingivalis in the negative control medium (OBGM) for 50 h resulted an average of 1.4-fold glycine concentration ratio (Fig. 3.14). The glycine ratio in OB:CM5F(OB_5F) was significantly (adjusted p-value <0.05) higher than OBGM. Both OB:RT38 and OB:F4 media combination showed significantly higher glycine ratio than OBGM (Fig. 3.14), suggesting that these fractions are stimulatory to P. gingivalis free glycine production. Similarly, the ratio of free glycine after P. gingivalis 50 h growth in OB:RT18 was also significantly higher than the negative control medium (Fig. 3.14), suggesting that the distinctive peaks in TdCM(OB_5F) can stimulate P. gingivalis to generate a higher concentration of free glycine. From the chromatogram (Fig. 3.12a), OB:F3 seemed to have no observable T. denticola produced molecules and yet P. gingivalis grown in OB:F3 produced a significantly higher ratio of glycine concentration than OBGM. Conversely, P. gingivalis growth in OB:F5 did not contribute to a significant increase of glycine ratio, which suggested that the molecular peaks detected around RT of 45 min were irrelevant for P. gingivalis glycine production. Upon compilation of all observations, it appears that the TdSF were comprised of non-sequence specific peptides. The results did not support the initial hypothesis proposed by Tan et al. (2014) that signalling molecules produced by T. denticola are involved in regulating P. gingivalis free glycine production. Rather, the results showed that non-specific peptides are present in TdCM, which may be produced via the action of T. denticola proteolytic enzymes on proteins and peptides derived from the growth medium.

Figure 3.14: Average ratio of glycine concentration after P. gingivalis 50 h growth to the basal glycine concentration in different RP-HPLC fractions of TdCM5F(OB_5F). Fractions F1, F2, F3, F4 and F5 were mixed with OBGM in a 1:1 volume ratio to obtain OB:F2, OB:F3, OB:F4 and OB:F5 respectively. RT18 and RT38 peaks were concentrated thrice before mixing with OBGM. OBGM as the negative control and OB:CM5F(OB_5F) as the positive control of the experiment. This data were pooled from all independent experimental set-up, where n = 9 for OBGM, n = 12 for OB:CM5F(OB_5F) and n = 5 for OB:F5, while the rest of the samples have n = 6. The error bars represent the standard deviation. One-way ANOVA and post-hoc analysis of Conover-Iman test were performed for statistical significance analysis of data. The medium variable with “*” symbol represents an adjusted p-value <0.01 compared to OBGM.

1 3.2.2 Metabolomic analyses of targeted amine groups

2 3.2.2.1 Free glycine concentration quantification 3 Metabolomic analyses were performed to determine P. gingivalis amine group profiles 4 and quantify the absolute concentration of free glycine in different molecular size fractions of 5 TdCM. TdCM was filtered through a 10 kDa centrifugal filter to obtain the 10 kDa retentate 6 and filtrate fractions. The 10 kDa retentate fraction was named TdCM10R, whilst the 10 kDa 7 filtrate was further filtered through a 3 kDa centrifugal filter to obtain the 3 kDa retentate and 8 filtrate fractions, which were named as TdCM3-10F and TdCM3F respectively. 9 The absolute concentration of free glycine in each medium was determined from the 10 linear calibration graph of matrix standards (Fig. 3.15), which were obtained by 11 supplementation of exogenous free glycine to a concentration of 10 mM into TdCM that had 12 been depleted of glycine, followed by serial dilution of preceding standards. 13 14

15 16 Figure 3.15: Calibration graph of the matrix standards for the determination of absolute glycine 17 concentration in each sample, by which T. denticola conditioned medium was supplemented 18 with 1.25, 2.5, 5 or 10 mM of glycine. This linear regression was generated by Mass Hunter 19 program and was used for determination of glycine content in different P. gingivalis growth 20 media.

21 The absolute glycine concentration at t = 0 and 48 h after P. gingivalis growth in 22 OB:CM, OB:CM10R, OB:CM3-10F, OB:CM3F and OBGM are depicted in Fig. 3.16. Results 23 showed that the glycine concentration remained unchanged in OBGM over 48 h, whereas free 24 glycine was significantly increased after P. gingivalis 48 h growth in OB:CM. The data are 25 also presented in alternate fashion in Fig. 3.17, where the change in glycine ratio for each

1 medium is compared. Statistical analysis on the data indicated that these molecular size 2 fractions of TdCM had contributed to a significantly higher ratio of glycine after P. gingivalis 3 48 h growth, compared with OBGM (Fig. 3.17). Given that all of the TdCM size fractions 4 showed an increase in glycine concentration after P. gingivalis growth, one would conclude 5 that the TdSF are less than 3 kDa in size.

6 7 Figure 3.16: Glycine concentrations measured at the initial time point and after P. gingivalis 8 48 h growth in different size filtrates of TdCM, which was quantified using a LC-QQQ-MS. 9 These data represent biological quadruplicates. The error bars represent the standard deviation.

10 11 Figure 3.17: Average ratio of glycine concentration (mM) after 48 h growth of P. gingivalis in 12 different TdCM size fractions mixed with OBGM. OB:CM (CM, a), OB:CM10R (CM10R, b), 13 OB:CM3-10F(CM3-10F, c), OB:CM3F (CM3F, d), OB (OBGM, e). These data were 14 represented in biological quadruplicates with error bars representing the standard deviation. 15 The letters indicate significant difference (adjusted p-value <0.05) compared with other media, 16 while “*” symbol represents an adjusted p-value <0.001 using post-hoc analysis of Tukey test.

1 3.2.2.2 Quantification of amine groups in the media 2 Quantitative analyses of targeted amine groups, which included amino acids, peptides 3 and polyamines, were performed using LC-QQQ-MS by Metabolomics Australia (MA) 4 according to Boughton et al. (2011). Changes in relative abundance of amine groups by P. 5 gingivalis after 48 h growth in OB:CM and OBGM were examined. Only metabolites that 6 passed the following described criteria were included for comparative analysis of the relative 7 peak abundances between samples: (1) a signal response above 1000 a.u. (2) pooled biological 8 controls displayed less than 25% coefficient variance (3) metabolites that were provided in the 9 solvent standards as prepared by MA (4) metabolites that were relevant to the biological 10 systems. A total of 32 amine groups were further analysed in all samples based on these 11 selection criteria, whilst eight metabolites were co-eluted and hence could not be discriminated. 12 Relative peak abundances of amine groups in OBGM and OB:CM at the initial time 13 point were assessed (Table I.1, Table I.2). The data indicated that the relative peak abundances 14 of aspartate, asparagine and histidine in OB:CM were approximately half of that in OBGM 15 (Table I.1), which suggested that these amino acids were completely used by T. denticola after 16 7 days growth in OBGM. Conversely, the initial peak abundances of proline was 3-fold higher 17 in OB:CM than OBGM, suggesting the release of proline via T. denticola proteolytic activities. 18 Other 19 P. gingivalis growth over 48 h did not affect the ratio of relative peak abundances of 20 most amino acids in OB:CM than OBGM, except for glutamine and proline. Ratios of 21 glutamine relative peak abundance achieved an approximate 2-fold increase after P. gingivalis 22 growth in both media (Fig. 3.18). There was a >4-fold increase of proline ratio in OBGM, 23 which was significantly higher (p-value < 0.01) than proline ratio in OB:CM that showed a 2- 24 fold increase. This suggested that P. gingivalis did not utilise proline and the ratio of proline 25 could in turn correlate to P. gingivalis proteolytic activities in releasing amino acids from 26 peptide-bound forms. Interestingly, the ratio of homoserine/threonine relative peak abundance 27 was significantly lower in OBGM (0.48 ± 0.07) after P. gingivalis growth than OB:CM (0.75 28 ± 0.06), suggesting increased utilisation of these amino acids in the negative control medium. 29 30

Figure 3.18: Ratio of relative peak abundance of amino acids after P. gingivalis 48 h growth in different size-filtrates of TdCM. OB:CM (red bar), and OBGM (black bar). These data represent biological quadruplicates with error bars representing the standard deviation. The “*” symbol indicates significant difference (p-value <0.01) between OB:CM and OBGM using a two-tailed paired t-test.

The initial relative abundances of glutathione (GSH) and ornithine in OB:CM were half of that in OBGM, suggesting that TdCM is completely depleted of these metabolites. Metabolites that were expected to be produced by T. denticola, which include citrulline, putrescine and cadaverine were 7-, 139- and 10-fold higher in OB:CM than in OBGM (Table I.2). Other metabolites, such as tyramine, 2-aminobutyric acid/GABA, agmatine, taurine, cystathionine and dihydroxyphenylalanine remained unchanged by T. denticola, with the initial peak abundance in OB:CM similar to OBGM (Table I.2). The ratio of relative peak abundance of other amine groups, which included taurine, agmatine, GSH, GABA, ornithine, dihydroxyphenylalanine, cystathionine, putrescine, cadaverine, spermidine, PABA/tyramine and citrulline, were also examined and compared after P. gingivalis 48 h growth in OBGM and OB:CM (Fig. 3.19). While most amine groups remained unchanged by P. gingivalis, the ratio of relative peak abundance of ornithine and citrulline were increased by P. gingivalis in both media. Relative peak abundances of ornithine ratios were found to increase to approximately 4.3-fold after P. gingivalis growth in OBGM and OB:CM. There was an average of 17-fold increase in ratio of citrulline relative peak abundance in OBGM, but to a lesser extent with ratio of 2.4 in OB:CM (Fig. 3.19). The ratio differences between these two media were significantly different (p-value <0.01), which suggested that citrulline was produced at a higher rate by P. gingivalis in OBGM. Interestingly, the peak abundance ratio of GSH, a tripeptide consisting of glutamate, cysteine and glycine, was significantly decreased by P. gingivalis over 48 h in OB:CM (0.16 ± 0.02) than OBGM (0.74 ± 0.11) (Fig. 3.19).

Figure 3.19: Ratio in the peak abundance of each amine group after P. gingivalis 48 h growth in different size-filtrates of TdCM. OB:CM (red bar) and OBGM (black bar). These data represent biological quadruplicates with error bars representing the standard deviation. The “*” symbol indicates significant difference (p-value <0.01) between OB:CM and OBGM using two-tailed paired t-test.

1 3.3 Discussion 2 Glycine is a preferred amino acid substrate for T. denticola, and exogeneous 3 supplementation of free glycine has been demonstrated to enhance its growth (Tan et al. 2014). 4 Likewise, P. gingivalis is equipped with a glycine cleavage system and it is also likely to utilise 5 glycine. P. gingivalis free glycine utilisation might occur at later time points, which was 6 indicated by a decrease in glycine during stationary phase in this study (Fig. I.1). Otherwise, 7 P. gingivalis might be internalising and hydrolysing glycine-containing peptides intracellularly 8 and catabolising glycine into metabolic intermediates or precursors. Physiologically, free 9 glycine is scarce, but is enriched in matrices such as collagen with the repetitive Gly-Pro-Xaa 10 and Gly-Xaa-Hyp motifs (Xaa and Hyp represent any amino acids and hydroxyproline 11 respectively) or found in other peptide forms. Therefore, the availability of free glycine for T. 12 denticola utilisation is likely dependent on host and bacterial proteases to degrade the 13 exogeneous glycine-containing peptides. In this context, the combinatory proteolytic activities 14 of T. denticola and P. gingivalis might be pertinent for the release of free glycine. An earlier 15 study using an oral polymicrobial inoculum grown in mucin has shown that the diversity and 16 complementarity of enzymatic activities found in different bacterial species led to increasing 17 accessibility of nutrients and proliferation of the microbial consortium (Bradshaw et al. 1994). 18 An extensive array of proteolytic enzymes present in the culture will likely increase the 19 efficiency of glycine release from glycine-bound macromolecules. As an expansion of this 20 study, mutational analysis of P. gingivalis peptidases that are predicted to be involved in 21 glycine production, as well as RNA sequencing of P. gingivalis grown in OB:CM relative to 22 OBGM were investigated and described in Chapter 4 and 5 respectively. 23 Reductionism and elimination approaches are normally taken to investigate interspecies 24 relationships by using a chemically defined minimal medium for growth of bacteria. Yet, the 25 formulation for the T. denticola chemically defined medium is complex, constituting at least 26 55 components (Wyss 1992), as T. denticola has fastidious nutritional requirements. This also 27 implies that T. denticola is a polymicrobial- and host-associated bacterial species. The 28 micronutrients, i.e. vitamins, other cofactors, nucleobases and nucleosides, and metal ions that 29 are required by T. denticola are not ubiquitous, but can be sourced from the host or other 30 subgingival plaque community members by cross-feeding or complementation of biosynthetic 31 pathways. 32 Attempts to reduce the complexity of OBGM and TdCM were performed using size- 33 filtration. OBGM used for T. denticola growth was size-filtered to separate the molecules less

1 than 5 kDa from other macromolecules, including the primary protein component in the 2 medium, rabbit serum albumin. A simplified form of TdCM, TdCM5F(OB_5F) was further 3 fractionated using RP-HPLC fractions to determine the most active TdCM fractions that 4 contribute to P. gingivalis glycine production. However, both OB_5F and TdCM5F (OB_5F) 5 were still deemed complex and showed high absorbance readings (Fig. 3.12). Usage of OBGM, 6 an undefined and complex medium, contributed to batch-to-batch variation in its composition 7 and hence increased variability between samples. This resulted in large variances between 8 independent experimental replicates and posed a series of complications for determination of 9 the peptides of interest, as the diversity and abundance of peptides in the media varied. 10 Nevertheless, the results of this study revealed that the events of P. gingivalis increased free 11 glycine production were not explicitly triggered by a molecule produced by T. denticola. It is 12 most likely that the increase in free glycine by P. gingivalis is due to the activity of its proteases 13 that act upon the highly diverse and non-sequence specific glycine-containing peptides in 14 OB:CM. 15 The consistent observation of lower cell density at the end of P. gingivalis lag phase in 16 OB:CM led to the speculation that T. denticola is secreting growth suppressing molecules into 17 the culture supernatant. The possibility that nutrient depletion by T. denticola in TdCM 18 contributed to the P. gingivalis reduced cell density during the lag phase was eliminated as 19 OB:PBS that has less nutrients than OB:CM had no effect on P. gingivalis lag phase.. As 20 competition is one of the mechanisms involved in shaping polymicrobial communities, many 21 microbial consortia have been demonstrated to produce antagonistic molecules to survive and 22 grow in complex and diverse polymicrobial communities. Various competitive mechanisms 23 have been demonstrated in bacterial interspecies interactions (Stubbendieck and Straight 24 2016). Given that T. denticola possesses a number of bacteriocin biosynthesis genes and toxic 25 metabolic end products as detailed in KEGG (Kanehisa and Goto 2000), it is possible that these 26 growth inhibiting molecules were released into the T. denticola spent medium and initially 27 suppressed P. gingivalis growth. Bacteria living in a polymicrobial community often have 28 strategies to mitigate the inhibitory effects of toxin molecules released by neighbouring cells.. 29 However, as the production of toxin metabolites is energetically costly and involves a highly 30 regulated process, bacteria are unlikely to synthesize and release toxins in a non-specific or 31 undirected manner. Therefore, it is far more likely that the suppressive effect of TdCM towards 32 P. gingivalis is due to the accumulation of T. denticola metabolic end products. 33 Bacterial metabolic cooperation by increasing the overlapping, complementary and 34 diversification of enzymatic activities can lead to nutrient niche expansion and increase the 85

1 nutrient utilisation efficiency. As P. gingivalis and T. denticola are both utilising amino acids 2 and peptides as their primary carbon and energy sources, a diverse array of proteolytic enzymes 3 are essentially their nutrient acquisition strategy by hydrolysing complex protein structure into 4 small peptides and amino acids. Releasing free glycine from protein complex might not be 5 useful for P. gingivalis, but this process can increase the fitness advantage of T. denticola, 6 especially in a highly competitve environment. 7 A high-throughput screening of the change in relative abundance of amine groups in 8 different TdCMs could provide insights into the dynamics of amino acids and BA groups 9 during P. gingivalis growth and elucidate potential metabolic interactions between P. gingivalis 10 and T. denticola. There was a dramatic decrease of GSH during P. gingivalis growth in OB:CM 11 but not in OBGM, suggesting the involvement of T. denticola sequential GSH degradation 12 enzymes (Chu et al. 2002, Chu et al. 2003, Chu et al. 2008) in breaking down GSH for P. 13 gingivalis. The presence of T. denticola outer membrane vesicles (OMVs) may contribute to 14 this outcome and this theory is supported by the detection of γ-glutamyl-transpeptidase (GGT) 15 and cysteinylglycinase (CGase) in T. denticola OMVs (Veith et al. 2009). Close association of 16 T. denticola and P. gingivalis in dysbiotic polymicrobial communities might contribute to 17 synergistic metabolic interactions by increasing the degradation and utilisation of GSH and 18 free glycine, which in turn affects the host oxidative regulatory efficiency. Depletion of GSH 19 can result in the decrease of antioxidant level and lead to further accumulation of reactive 2- 20 oxygen species, i.e. free radicals O and peroxides H2O2 that have been found to increase in 21 diseased gingival crevicular fluid and saliva (Bains and Bains 2015). Consequently, 22 mechanisms for glycine scavenging and sequestration between the host and bacteria can be an 23 important determinant for the maintenance of redox balance in the cell, which in turn can affect 24 host-bacterial and bacterial-bacterial homeostasis. 25 Glutamate and aspartate are metabolised by P. gingivalis (Takahashi et al. 2000) and 26 glutamate levels were found to be greatly reduced in the GCF of periodontitis patients, which 27 supports the notion that P. gingivalis and other periodontally associated bacteria prefer the 28 utilisation of glutamate as a nutrient source (Narda et al. 2008). However, no net change of 29 these two metabolites was observed after P. gingivalis growth in different TdCMs 30 combinations. This might be due to an equivalent rate of proteolysis and internalisation of 31 amino acids for utilisation. Otherwise, it might be owing to the intrinsic nature of P. gingivalis 32 that preferably transports peptides, instead of free amino acids (Takahashi et al. 2000, 33 Takahashi and Sato 2001, Takahashi and Sato 2002). As such, P. gingivalis is equipped with a 34 wide range of exopeptidases and oligopeptide transporters to ensure efficient internalisation of 86

1 oligopeptides (Nelson et al. 2003, Nemoto and Ohara-Nemoto 2016). Intracellular peptidolytic 2 enzymes of P. gingivalis would further degrade the di-, tri- and oligopeptides into amino acids 3 for catabolism. This competitive mechanism employed by P. gingivalis in polymicrobial 4 communities is essential in securing its source of nutrients, energy and biomass precursors. The 5 increase of proline after T. denticola and P. gingivalis growth suggested that these two bacterial 6 species do not preferably utilise proline in free and peptide-bound forms. While P. gingivalis 7 is predicted to lack of proline-biosynthesis pathways, P. gingivalis possesses 8 dipeptidylpeptidase IV that shows high specifically and efficiency in the hydrolysis of proline 9 or hydroxyproline-containing peptides (Rea et al. 2004, Banbula et al. 2000). The free proline 10 could be released in the extracellular milieu as a side product of their proteolytic activities. 11 This data also indirectly indicated that P. gingivalis extracellular proteases could contribute to 12 an overall increase in nutrient acquisition strategy as these proteases could potentially be 13 exploited by other bacteria in the communities as public goods to increase the accessibility of 14 nutrients. Further studies into the fate of peptides in each medium could help illustrate P. 15 gingivalis nutrient utilisation preference, as well as their proteases capability in nutrient 16 acquisition. 17 A marked increase in polyamines, such as ornithine and citrulline by P. gingivalis in 18 all growth media, as well as putrescine by T. denticola suggested extensive utilisation of 19 arginine by both bacterial species. This theory is proposed as the biosynthesis of ornithine, 20 putrescine, spermidine and agmatine are interconnected with arginine, which acts as the amino 21 acid precursor in a complex metabolic network (Fig. 3.20). Free arginine was excluded from 22 this analysis as the signals were at background level (<1000 a.u.) of detection. However, 23 arginine levels are higher in the gingival crevicular fluid of adult periodontitis patients, which 24 has been correlated to the proteolytic activities of gingipains from P. gingivalis (Narda et al. 25 2008). An increase in peak abundance ratios of these polyamines suggested that arginine exists 26 in peptide forms, which only become available for utilisation during P. gingivalis growth. 27 However, P. gingivalis enzymes responsible for the deamination of arginine to citrulline were 28 not found in the KEGG database (Fig. 3.21). In the polyamine biosynthesis metabolic network, 29 P. gingivalis also lacks ornithine decarboxylase for the conversion of ornithine to putrescine, 30 however, T. denticola is able to synthesise putrescine (Fig. 3.20). Although the roles of these 31 polyamines in bacterial physiology are rarely studied, a high level of putrescine can in turn lead 32 to T. denticola increased production of succinate and butyrate, which are cytotoxic to host cells. 33 Further characterisation of these enzymes and metabolic pathways are warranted to validate 34 these hypothesis. 87

1 Biosynthesis of BA groups from amino acid precursors have been listed as diagnostic 2 biomarkers to monitor periodontal disease development, progression and severity (Taba et al. 3 2005, China et al. 2012). Decarboxylation of amino acids for the biosynthesis of BA groups 4 are primarily modulated by host-associated bacteria and have been shown to affect host 5 homeostasis and its immunomodulatory activities. It is important to elucidate the mechanisms 6 of interactions between species in polymicrobial communities in their nutrient acquisition 7 strategies, as well as their metabolic interactions that allow them to thrive under dysbiotic 8 conditions. Numerous studies have shown that polyamines (putrescine, cadaverine, 9 spermidine, spermine) are involved in host cell invasion and survival, bacteriocin production 10 and biofilm formation (Shah and Swiatlo 2008, Di Martino et al. 2013, Espinel et al. 2016). 11 Knowledge of the metabolic capabilities of dysbiotic polymicrobial communities and the effect 12 of BAs on bacterial physiology are still lacking. These data have provided insights into the 13 dynamics of amine groups, as well as the potential metabolic interactions between P. gingivalis 14 and T. denticola. Further investigations into their metabolic interactions, as well as the 15 mechanisms involved are required to gain a better understanding of how these bacterial species 16 interact to become more competitive and cause enhanced virulence under dysbiotic conditions. 17 18

Figure 3.20: T. denticola and P. gingivalis metabolic enzymes involved in the amino acid catabolism pathways. This information is based on the genome of T. denticola ATCC35405 and P. gingivalis W83 from the KEGG and BIOCYC databases. The protein-encoding genes of T. denticola and P. gingivalis are represented in common gene identifier initiate with letter “PG” and “TDE” respectively.

1 3.4 Conclusion 2 Investigations into the TdSFs in TdCM that resulted in an increase of free glycine 3 production by P. gingivalis indicated that the stimulatory factors are non-sequence specific 4 peptides. The net increase of free glycine after P. gingivalis growth is likely contributed to by 5 a diverse array of glycine-containing peptides in all fractions of TdCM. This study also 6 repeatedly demonstrated that TdCM exhibits an initial suppressive effect on P. gingivalis 7 growth, which led to the hypothesis that T. denticola and P. gingivalis might also demonstrate 8 competitive interactions, possibly mediated by metabolic end products, including biogenic 9 amines. Regulation of the levels of amine-containing compounds, such as GSH, spermidine, 10 ornithine and putrescine by P. gingivalis in a medium-dependent manner could be part of the 11 metabolic interactions between P. gingivalis and T. denticola. 12 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28

1 Chapter 4 Characterisation of free glycine

2 release by P. gingivalis peptidase mutants

3 4.1 Introduction 4 P. gingivalis is an asaccharolytic bacterium that relies on amino acids as carbon and 5 energy sources for growth. The extensive array of proteinases in P. gingivalis increase the 6 accessibility of nutrients by degrading host proteins, followed by the sequential digestion of 7 oligopeptides into tri-, dipeptides and amino acids for acquisition. 8 Gingipains (RgpA, RgpB, Kgp) are the primary proteolytic system of P. gingivalis that 9 initiate the process of degradation of extracellular matrix, serum-derived proteins and 10 hemoglobin, which leads to the production of oligopeptides for downstream digestion 11 (Ruggiero et al. 2013, O’Brien-Simpson et al. 2003, Dashper et al. 2004). Following the initial 12 proteolytic degradation of complex proteins by gingipains or other proteases into oligopeptides, 13 numerous oligopeptidases, carboxypeptidases and aminopeptidases work in concert to produce 14 tri- and dipeptides for P. gingivalis internalization and utilisation (Nemoto and Ohara-Nemoto 15 2016). Peptidases of P. gingivalis that have been characterised include; dipeptidyl peptidase 16 (DPP) IV, DPP5, DPP7, DPP11, prolyl tripeptidyl peptidase A (PtpA), as well as acylpeptidyl 17 oligopeptidase (AOP) and carboxypeptidase (CPG70), are described in detail in Chapter 1 and 18 summarized in Table 4.1 below. Less well known are the exopeptidases (carboxy- and 19 aminopeptidases) of P. gingivalis that are required for the generation of free amino acids from 20 peptide substrates. Besides acting as important macronutrient extracting factors for P. 21 gingivalis, extracellularly localized peptidases are postulated to have a role in the release of 22 peptides and free amino acids for other oral bacterial species. The efficiency and specificity of 23 amino acid release and utilisation by the oral subgingival polymicrobial community can 24 indirectly govern their amino acid fermentation pathways and therefore generate cytotoxic end 25 products that exacerbate disease conditions. 26 Tan et al. (2014) demonstrated that P. gingivalis increased the rate of hydrolysis of 27 glycine-containing peptides in OB:CM, compared with OB:PBS, therefore it is postulated that 28 P. gingivalis peptidases play important roles in the release of free glycine from glycine-bound 29 peptides. This study was conducted in parallel with Chapter 3 to corroborate one of the 30 hypotheses stated in Fig. 3.1, which was to examine the roles of P. gingivalis peptidases that 31 are potentially involved in the release of free glycine in OB:CM. Predictive analysis of P. 32 gingivalis peptidase and protease localisation and specificity were performed to guide the

1 selection of P. gingivalis peptidases that are likely to have contributory roles in the cleavage 2 of glycine-containing peptides leading to an increased release of free glycine. Four protease- 3 encoding genes of P. gingivalis; PG0445, PG0753, PG1605 and PG1788, were selected for 4 inactivation. Then, the mutant strains were grown in OBGM and OB:CM to examine and 5 compare their free glycine release to that of P. gingivalis wild type. 6 7 Table 4.1: Summary table of the cleavage specificity and other important descriptions of the 8 characterised P. gingivalis peptidases. Peptidase Substrate specificity (N to C) Description References

DPPIII NH2-Xaa-Xaa- -(Xaa)n Preferably cleaves Arg at P1 position Hromić-Jahjefendić et al. 2017 DPPIV NH2-Xaa-Pro/Ala- -Yaa-(Xaa)n Ala at the penultimate position will Rea et al. 2004 decrease the peptidase hydrolysis Banbula et al. 2000 efficiency, as compared to Pro and hydroxyproline DPP5 NH2-Xaa-Yaa- -(Xaa)n Preferably cleaves Ala and hydrophobic Ohara-Nemoto et al. residues at P1 position 2014 DPP7 NH2-Xaa-Zaa- -Yaa-(Xaa)n Does not cleave substrates with a Banbula et al. 2001 blocked N-terminus and long oligopeptides; prefers hydrophobic amino acid at P2 position DPP11 NH2-Yaa-Asp/Glu- -(Xaa)n Cleaves aspartate or glutamate at P1 Ohara-Nemoto et al. position and preferably cleaves 2011 hydrophobic amino acids at P2 position

AOP (Na)Xaa-Yaa- -Xaa Preferably liberates di- and tripeptides Nemoto et al. 2016 from N-terminally modified oligopeptides with hydrophobic amino acids at the P1 position PtpA NH2-Xaa-Xaa-Pro- -Yaa-(Xaa)n Liberates N-terminal tripeptide; does not Banbula et al. 1999 cleave a blocked N-terminus or when Pro at P2 position CPG70 (Xaa)n-Yaa- -Lys/Arg-COOH Predominantly cleaves Lys at P1’ Chen et al. 2002 position and does not cleave peptides Masuda et al. 2002 with Pro at P2 position 9 Xaa – all amino acids 10 Yaa – any amino acids except Pro 11 Zaa – All charged amino acids (Asp, Glu, Asn, Gln, Lys, Arg) and hydrophobic side chain 12 amino acids (Ala, Val, Ile, Leu, Met, Phe, Tyr, Trp), except Pro and Gly. 13 Na – N-terminally modified oligopeptides, including acylated, benzyloxycarbonyl substrates 14 P2-P1- -P1’-P2’, whereby “ ” represents cleavage site 15 16 17 18 19

1 4.2 Results

2 4.2.1 Selection of P. gingivalis peptidase targets

3 4.2.1.1 Localisation predictions of P. gingivalis proteases 4 It is proposed that P. gingivalis proteases and peptidases that are responsible for the 5 hydrolysis of oligopeptides into free glycine in TdCM are localized extracellularly. Therefore, 6 prediction of the cellular location of P. gingivalis peptidases was taken as a preliminary step 7 into understanding the potential contributing roles of these peptidases in the modification of 8 the peptide-based environment, as well as their effects on P. gingivalis growth. 9 The initial step involved the identification and compilation of potential proteases in P. 10 gingivalis. The P. gingivalis W83 complete genome sequence in the NCBI database (Reference 11 sequence: NC_002950.2) showed 52 protein encoding genes with “peptidase”, “protease”, 12 “collagenase”, as well as “trypsin” and “gingipain”-containing keywords in the gene product 13 annotations. A search of the MEROPS Peptidase Database Release 12.0 resulted in a list of 70 14 putative peptidase and non-peptidase homologues in P. gingivalis W83 strain (Rawlings et al. 15 2014). Eight non-peptidase homologues that were distinguished by the unacceptable 16 replacement or absence of characteristic active site residues from their peptidase counterparts 17 (Rawlings and Morton 2008), as well as two other homologues that lack the keywords, which 18 include GMP synthase (PG0589) and probable imidazolonepropionase (PG0328), were 19 excluded from the analysis. As most P. gingivalis peptidases are currently uncharacterised and 20 their physiological properties undetermined, information from both resources were combined 21 in this study. Moreover, it was noted that PrtT and Kgp that are annotated as “T9SS C-terminal 22 target domain-containing protein” and “DUF2436 domain-containing protein” respectively in 23 the NCBI database and were not found in the MEROPS search, thus were added to the P. 24 gingivalis peptidase list. Based on this approach, a total of 68 unique potential proteases and 25 peptidases were identified and used for prediction of P. gingivalis peptidase localization (Table 26 4.2). 27 Amino acid sequences of the selected peptidases were analyzed via PSORTb version 28 3.0, a bacterial protein subcellular localisation prediction tool (Yu et al. 2010). The confidence 29 of prediction was indicated in an output score ranging from scale 0–10, whereby a score >7.5 30 was considered confident enough for the determination of peptidase cellular location, whilst a 31 score <7.5 was considered as unknown. In summary, 24 peptidases of P. gingivalis were 32 predicted to be localized in the cytoplasm, 17 in the cytoplasmic membrane, three peptidases

1 in the periplasmic region, two in the outer membrane, one located in the extracellular region 2 and 21 proteases were of unknown location (Table 4.2). 3 Table 4.2: Localisation prediction and N-terminal signal peptide prediction of P. gingivalis 4 proteases (sourced from NCBI database and MEROPS) by PSORTb and signalP. Gene Product name used PSORTb SignalP PG0010 ATP-dependent Clp protease ClpC Cytoplasmic No PG0088 subfamily M16B unassigned peptidases Cytoplasmic No PG0137 Pep581 peptidase Cytoplasmic No PG0159 Endopeptidase (PepO) Cytoplasmic No PG0227 DNA repair protein RadA (Escherichia coli) Cytoplasmic No PG0317 family M49 unassigned peptidases Cytoplasmic No PG0418 peptidase Clp (type 1) Cytoplasmic No PG0445 peptidase T Cytoplasmic No PG0528 family C44 unassigned peptidases Cytoplasmic No PG0537 Pep581 peptidase Cytoplasmic No PG0561 subfamily M20F unassigned peptidases Cytoplasmic No PG0620 family S16 unassigned peptidases Cytoplasmic No PG0713 family C26 unassigned peptidases Cytoplasmic No PG0753 collagenase (Salmonella-type) Cytoplasmic No PG0758 peptidyl-dipeptidase Dcp Cytoplasmic No PG0889 peptidase M24 family protein Cytoplasmic No PG1055 Tpr peptidase Cytoplasmic No PG1210 subfamily M24B unassigned peptidases Cytoplasmic No PG1337 UmuD protein Cytoplasmic No PG1542 collagenase Cytoplasmic No PG1654 vanX D-Ala-D-Ala dipeptidase Cytoplasmic Yes PG1701 family C26 unassigned peptidases Cytoplasmic No PG1789 peptidyl-dipeptidase Dcp Cytoplasmic No PG1917 subfamily M24A unassigned peptidases Cytoplasmic No PG0011 family S12 unassigned peptidases Cytoplasmic Membrane Yes PG0026 PG0026 peptidase (PorU) Cytoplasmic Membrane Yes PG0047 family M41 unassigned peptidases Cytoplasmic Membrane No PG0235 C-terminal processing peptidase-3 Cytoplasmic Membrane Yes PG0383 RseP peptidase Cytoplasmic Membrane No PG0518 At3g26085 (Arabidopsis thaliana) Cytoplasmic Membrane No PG0639 subfamily S49B unassigned peptidases Cytoplasmic Membrane No PG1060 subfamily S41A unassigned peptidases Cytoplasmic Membrane Yes PG1403 family S54 unassigned peptidases Cytoplasmic Membrane No PG1404 At5g25752 rhomboid peptidase Cytoplasmic Membrane No PG1598 signal peptidase II Cytoplasmic Membrane No PG1855 subfamily S41A unassigned peptidases Cytoplasmic Membrane No PG2000 signal peptidase I Cytoplasmic Membrane No PG2001 subfamily S26A unassigned peptidases Cytoplasmic Membrane No

PG2024 gingipain RgpA Cytoplasmic Membrane Yes PG2185 family M56 unassigned peptidases Cytoplasmic Membrane No PG2197 Oma1 peptidase Cytoplasmic Membrane No PG0593 subfamily S1C unassigned peptidases Periplasmic Yes PG0724 dipeptidyl-peptidase 5 Periplasmic Yes PG1422 family S13 unassigned peptidases Periplasmic No PG1427 periodontain Outer Membrane Yes PG2192 subfamily M23B unassigned peptidases Outer Membrane Yes PG0553 pepK (Porphyromonas gingivalis) Extracellular Yes PG0196 PqqL protein (Escherichi coli) Unknown No PG0232 CPG70 carboxypeptidase Unknown Yes PG0410 family C25 unassigned peptidases Unknown No PG0491 dipeptidyl-peptidase 7 Unknown Yes PG0503 S9 family peptidase Unknown Yes PG0506 gingipain R2 (RgpB) Unknown Yes PG0721 family C40 unassigned peptidases Unknown No PG0740 family C40 unassigned peptidases Unknown No PG0840 subfamily S1C unassigned peptidases Unknown No PG0956 peptidase M24 Unknown No PG1004 subfamily S9C unassigned peptidases Unknown Yes PG1283 dipeptidyl-peptidase 11 Unknown Yes PG1313 family C69 unassigned peptidases Unknown Yes PG1361 prolyl tripeptidyl peptidase (PtpA) Unknown Yes PG1548 PrtT Unknown No PG1605 subfamily C1B unassigned peptidases Unknown Yes PG1620 peptidase S41 Unknown Yes PG1754 subfamily S9C unassigned peptidases Unknown Yes PG1788 subfamily C1A unassigned peptidases Unknown Yes PG1840 neutral zinc metallopeptidase Unknown No PG1844 Kgp Unknown Yes 1 2 Furthermore, all proteases were examined for the presence of an N-terminal signal 3 peptide using a signal peptide predictor program - SignalP 4.0 (Petersen et al. 2011). 4 Consistently, all putative proteases that were predicted to be localised in the outer membrane 5 and extracellular regions by PSORTb were predicted to possess an N-terminal signal peptide. 6 Further, manual curation of proteases that are categorised in the “cytoplasmic membrane” and 7 “unknown” predictive localisation group using signalP showed that an approximate 29% and 8 62% proteases respectively had a signal peptide (Table 4.2) starting at a nearby in-frame 9 initiation codon. The presence of an N-terminal signal peptide on a targeted protein typically 10 indicates that the protein will be exported from the cytoplasm via a Sec translocon, whereupon

1 the protein can end up anchored on the cytoplasmic or outer membrane, retained in the 2 periplasmic space or released extracellularly into the environment (Ivankov et al. 2013).

3 4.2.1.2 Determination of P. gingivalis peptidases orthologous to glycine- 4 cleaving peptidases

5 P. gingivalis peptidases that were predicted to be localized in the extracellular or outer 6 membrane regions are of interest for examination of their potential role in the release of free 7 glycine from glycine-containing peptides. As summarized above, P. gingivalis was predicted 8 to have 44 proteases located outside of the cytoplasm. Twelve of these proteinases and 9 peptidases; RgpA, PorU, DPP5, periodontain, CPG70, DPP7, RgpA, RgpB, DPP11, PtpA, PrtT 10 and Kgp (Table 4.2), have been characterised previously and shown to be irrelevant for the 11 hydrolysis of glycine-containing peptides, while the remaining 32 of the predicted proteases 12 have not been characterised. 13 In search of peptidase candidates that can cleave glycine-containing peptides, PG1605 14 that showed 35% overall sequence identity to Lactococcus lactis aminopeptidase C (PepC), 15 which has been demonstrated to hydrolyse polyglycine and glycine-containing peptides 16 (Mistou and Gripon 1998), was selected as a target peptidase. Although lactococcal PepC 17 demonstrated cleavage activity on dipeptides, it exhibited higher specific activity as an 18 aminopeptidase on polyglycine peptide, which suggested that it is composed of extended 19 substrate binding sites (Mistou and Gripon 1998). Both PG1605 and L. lactis PepC are 20 members of the C1 family of clan CA cysteine peptidases. Other members of clan CA, such as 21 ubiquitin-specific peptidases, bacteriocin-processing peptidases and their homologues, are also 22 capable of cleaving glycine-containing substrates with high specificity (Fig. 4.1). In addition 23 to PG1605, PG1788 was also a predicted C1 family cysteine peptidase that is encoded in the 24 P. gingivalis W83 genome and was thus selected as another potential target for investigation. 25 Furthermore, bioinformatic analyses on peptidase PG1605 and PG1788 primary sequences and 26 modelled structures were performed to gain insights into their potentially distinctive biological 27 properties. 28 Based on information sourced from the MEROPS peptidase database, only homologues 29 of M20.003 assigned peptidase of P. gingivalis W83 showed cleavage specificity for Gly at P1 30 and P1’ positions of the tested substrates (Fig. 4.1, Mori et al. 2005, Miller and Broder 2004). 31 Therefore, PepT (PG0445), a predicted M20.003 peptidase of P. gingivalis, was also included 32 as a potential candidate for further study. Microarray data from coculture of P. gingivalis with 33 T. denticola in a continuous culture system showed increased expression of PG0753 and

1 PG0383 protease encoding genes, in comparison to the monoculture (Tan et al. 2014). 2 Therefore, a putative PrtQ collagenase (PG0753) belonging to the U32 peptidase family was 3 selected as one of the candidates for characterisation of its glycine-releasing effect in OB:CM. 4 PG0383, annotated as a regulated intramembrane proteolytic (RIP) metalloprotease RseP, was 5 considered irrelevant for glycine release and was therefore not selected. Although PG0445 and 6 PG0753 were predicted to be located in the cytoplasm by PSORTb, they were both selected for 7 further investigation as the possibility of glycine-containing peptides being hydrolysed 8 intracellularly and then exported into the environment as free glycine, cannot be ruled out. 9 Furthermore, leaderless bacterial proteins have been shown to be secreted into the extracellular 10 milieu by the non-conventional secretion mechanism, such as pore formation, especially during 11 cellular stress (Bendtsen et al. 2005, Rabouille 2017). Based on the above interpretation, 12 PG0445, PG0753, PG1605 and PG1788 were selected for inactivation and the resulting mutant 13 strains were subjected to experimental analyses of their glycine-releasing ability in OBGM and 14 OB:CM. 15 16 (a)

17 18 (b)

19 20 Figure 4.1: Cleavage site sequence logo of peptidases (a) C19.022 ubiquitin-specific peptidase 21 demonstrated cleavage of substrate peptides with double glycine occupying at the P2 and P1 22 position (b) M20.003 peptidase demonstrated cleavage activities on tripeptide substrates that 23 were composed of Ala, Gly, Leu and Met residues in the sequence. Label 1 to 6 on the x-axis 24 indicates P4, P3, P2, P1, P1’and P2’ positions respectively of the amino acids in the peptide 25 substrates (sourced from MEROPS).

26 27 28 97

1 2 4.2.2 Generation of P. gingivalis peptidase mutants 3 P. gingivalis peptidase mutants were created by allelic exchange of the chromosomal 4 gene encoding the peptidase of interest with an antibiotic resistance cassette. The 5 recombination cassettes for inactivation of P. gingivalis peptidases was constructed by SOE 6 PCR of three PCR fragments, comprised of approximately 500 bp sequences from the upstream 7 and downstream flanking regions of the targeted gene, as well as a gene encoding an antibiotic 8 resistance cassette. These PCR products were assembled using SOE PCR with the final product 9 A-tailed with polymerase and ligated into pGEM-T Easy. The constructs with the correct 10 sequence were linearized and used for P. gingivalis W50 transformation. The replacement of 11 peptidase target with an antibiotic resistance cassette was achieved by homologous 12 recombination of the flanking regions in the linearized vector with the genomic sequence of 13 the wild type. 14 A schematic overview of the genomic organizations of PG0445-, PG0753-, PG1605- 15 and PG1788- mutant strains around the homologous recombination sites is presented in Fig. 16 4.2. The gene encoding PG0753 that is located at the end of an operon, as well as PG1605 and 17 PG1788 that exist as open reading frames (ORFs) on their own, were completely replaced by 18 a gene encoding antibiotic resistance driven by cepA/emF own promoter. This is to ensure 19 strong expression of the antibiotic resistance phenotype during the exertion of the selective 20 pressure. Using the P. gingivalis W83 reference genome, MicrobesOnline and BIOCYC 21 databases predicted that PG0445 is the last ORF of an operon. However, a high and continuous 22 level of transcription from the ORF of PG0443 to PG0451 was shown in the transcriptomics 23 profile of P. gingivalis W83 during growth in different media (Høvik et al. 2012). Furthermore, 24 PG0444 was predicted to be co-expressed with PG0445, which encodes an oligopeptide 25 transporter (Alm et al. 2004, Arkin 2011, Caspi et al. 2016). Due to the likelihood that PG0445 26 may be located within an operon, PG0445 was replaced with the cepA ORF without its own 27 promoter to prevent polar effects. In this case, cepA expression relied on the native promoter 28 of PG0445 and expression of the downstream genes should not be affected. 29 30 31 32 33

1 2 (a)

3 4 (b)

5 6 (c)

7 8 (d)

9 10 11 Figure 4.2: Schematic representation of the relative position of P. gingivalis peptidase gene 12 targeted for replacement and the neighbouring open reading frames in the mutated genome. (a) 13 PG0445 replacement with cepA ORF (B) PG0753 replacement with cepA cassette (c) PG1605 14 replacement with emF cassette (d) PG1788 replacement with cepA cassette. Open arrows in 15 white, blue, pink and yellow represent the neighbouring genes of targeted operon, genes of the 16 same operon as the target, erythromycin (emF) or ampicillin (cepA) resistance cassette 17 respectively. P’ represents the promoter of the antibiotic resistance cassette. Not drawn to scale.

18 19 20 21 22 23 24 25 26 27 28

1 4.2.2.1 PCR confirmation of P. gingivalis peptidase mutants 2 Eight to ten transformants of P. gingivalis PG0445-, PG0753-, PG1605- and PG1788- 3 mutant strains were selected and plated onto HBA supplemented with Em or Ap, followed by 4 inoculation into BHI supplemented with Ap or Em for growth. Genomic DNA (gDNA) of P. 5 gingivalis mutant strains was isolated, quantified and amplified by PCR using appropriate 6 primers targeting a region within the antibiotic resistance gene paired with their respective 7 external construct region. The primers used for PCR confirmation of the mutant clones are 8 listed in Table 2.2. All transformants produced PCR products of the expected size, 9 demonstrating that homologous recombination at the flanking regions of the targeted gene was 10 successful (Fig. 4.3). There was no spontaneous antibiotic resistance mutant detected amongst 11 the transformants. Wild type gDNA was used as the negative control, which showed no PCR 12 amplification product as expected. Amplification of wild type gDNA using both primers 13 targeting the external construct regions generated amplicons of the correct size (Fig. 4.3). 14 Transformants PG0445- clone 1, PG0753- clone 1, PG1605- clone 1 and PG1788- clone 15 8 that had been confirmed by PCR were selected for further phenotypic characterisation (Fig. 16 4.3). Whole genome sequencing was performed using gDNA of PG0753-, PG1605- and 17 PG1788- to determine the presence of secondary mutations that could give rise to phenotypic 18 variations. 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33

100

1 (a) I II III 2 M 1 2 3 4 5 6 7 8 1 2 3 4 5 6 7 8 1 2 3 3 kb 4.0 3.0 2.5 4 2.0 1.5 5 1.0 6 (b) III IV 7 M 4 5 6 7 8 9 10 1 2 3 4 5 6 7 8 9 10 WT1 WT2 8 kb 4.0 3.0 2.5 9 2.0 1.5 10 1.0

11 (c) V VI

12 M WT3 WT 1 2 3 4 5 6 7 8 WT 1 2 3 4 5 6 7 8 kb 4 13 3 2.5 14 2 1.5 15 1 16 (d) VII VIII 17 kb M 1 2 3 4 5 6 7 8 WT 1 2 3 4 5 6 7 8 WT WT4 18 4 3 2.5 19 2 1.5 20 1 21

22 Figure 4.3: PCR amplicons of P. gingivalis peptidase mutant strains; PG0045-, PG0753-, 23 PG1605- and PG1788-, resolved by agarose gel electrophoresis. (a) Eight transformants 24 (labelled 1-8) of PG0445- were selected and amplified with (I) cepAPG0445F1_FP paired with 25 PG0445confo_RP and (II) cepAPG0445F2_RP paired with PG0445confo_FP. (a)(b) Ten 26 transformants (labelled 1-10) of PG1788- were selected and amplified with (III) 27 cepAPG0753_FP paired with PG0753confo_RP and (IV) cepAPG0753F2_RP paired with 28 PG0753confo_FP. WT1 and WT2 represent the amplicons of wild type gDNA using 29 PG0445confo_F/RP and PG0753confo_F/RP, respectively. (c) Eight transformants (labelled 30 1-8) of PG1605- were selected and wild type negative control (WT) were amplified with (V) 31 emFPG1605_RP paired with PG1605confo_FP and (VI) emFPG1605F1_FP paired with 32 PG1605confo_RP, respectively. WT3 represents the amplicon of wild type gDNA using 33 PG1605confo_F/RP. (d) Eight transformants (labelled 1-8) of PG1788- were selected and wild 34 type negative control (WT) were amplified with (VII) cepAFPG1788F1_FP paired with 35 PG1788confo_RP and (VIII) cepAPG1605_RP paired with PG1788confo_FP, respectively. 36 WT4 represents the amplicon of wild type gDNA using PG1788confo F/RP. “M” denotes DNA 37 molecular weight markers in kbp.

101

1 4.2.2.2 Whole genome sequencing of P. gingivalis peptidase mutants 2 Whole genome sequencing of PG0753-, PG1605- and PG1788- strains in reference to 3 the latest genome release of P. gingivalis W83 (NCBI reference sequence: CP0256932) further 4 confirmed the identity of each peptidase mutant, which showed deletion of the corresponding 5 gene. This analysis also revealed common sequence variations across all strains as shown in 6 Table II.1, which suggested that nucleotide polymorphisms occurred between P. gingivalis 7 W50 laboratory strain and P. gingivalis W83 strain and/or nucleotide variations might have 8 existed in the background strain used for transformation. Common nucleotide variations that 9 can result in frameshift mutations were detected at locus tag CF003_0196, which encodes a 10 zinc-dependent peptidase, CF003_0479 and CF003_2168, which encode hypothetical proteins. 11 An insertion of three cytosine nucleotides were detected in the locus tag CF003_1844, which 12 was typical for gingipain K (Kgp) encoding genes that have varying NPNP motif length in the 13 hemagglutinin domain (Table II.1)(Shibata et al. 1999). In addition, all determined single 14 nucleotide polymorphisms (SNP) in protein encoding genes were expected to cause amino acid 15 substitution mutations. The mutants that were sequenced were used for downstream phenotypic 16 characterisation experiments, as these mutations are unlikely to cause detrimental effects to P. 17 gingivalis growth. 18

102

1 4.2.3 Characterisation of the glycine-releasing ability of P. gingivalis

2 peptidase mutants

3 4.2.3.1 Growth curves of P. gingivalis peptidase mutants in OBGM and

4 OB:CM 5 P. gingivalis PG0753- and PG1605- peptidase mutants showed no distinct difference in

6 growth in OBGM relative to the wild type, based on OD650 measurements at t = 0, 20, 30, 40,

7 50 and 70 h. All three strains reached a maximum OD650 of 1.8 by 40 h (Fig. 4.4). On the - 8 contrary, PG1788 growth plateaued around OD650 ~1.5 after 40 h, whereas the cell density of - 9 PG0445 reached OD650 ~1.6 and then declined at a relatively rapid rate, in comparison to the 10 other strains. 11 P. gingivalis wild type demonstrated the highest growth relative to other peptidase

12 mutants in OB:CM and growth plateaued around OD650 ~1.0 after 50 h of growth (Fig. 4.4). 13 The growth trend of PG1605- strain was comparatively similar to the wild type; however, 14 PG1788- strain was severely compromised, showing a 30 h lag phase and slow growth rate, in 15 OB:CM. Although PG0753- grew better than the PG1788- strain, it also showed a slower 16 growth rate in comparison to the other strains (Fig. 4.4). In addition, there was a consistently

17 lower OD650 measurement at the end of lag phase for all P. gingivalis mutant strains grown in 18 OB:CM, similar to wild type. These observations led to the speculation that T. denticola had 19 produced growth suppressing factors against P. gingivalis, which especially impacted the 20 growth of PG0753- and PG1788- strains. It is also possible that T. denticola had utilised 21 essential metabolites that could not be produced by both mutants, thus compromising P. 22 gingivalis growth in OB:CM. The most reduced cell density of PG1788- strain lag phase in 23 OB:CM compared to other strains suggested that PG1788 might be required as part of a 24 mechanism to alleviate the growth suppressing effect against P. gingivalis.

103

2.0

1.8

1.6

1.4 )

650 1.2

1.0

0.8 Cell density (OD density Cell 0.6

0.4

0.2

0.0 0 10 20 30 40 50 60 70 Time (h) Figure 4.4. Growth of P. gingivalis wild type (red line) and PG0445- (brown line), PG0753- (green line), PG1605- (black line) and PG1788- (blue line) strains in OBGM (squares) and OB:CM (triangles). Bacterial strains were grown under anaerobic conditions at 37°C in a static culture and the cell growth was determined by measurement of optical density at a wavelength of 650 nm (OD650) using a spectrophotometer at time points of 0, 20, 30, 40, 50, 70 h. All samples with n = 8, except for PG0445- that has n = 4.

104

1 4.2.3.2 Enumeration of P. gingivalis wild type and peptidase mutants in

2 OBGM and OB:CM

3 Accurate enumeration of P. gingivalis cell number is an important aspect of this study. 4 Previously, the efficiency of free glycine release by P. gingivalis was expressed as the relative

5 change of glycine normalized against cell number derived from OD650 measurements in 6 different media (Tan et al. 2014). In order to examine the robustness of P. gingivalis cell

7 number calculations based on the following OD650 derived equation: P. gingivalis cell number/ 9 8 8 mL = (2 × 10 cells/mL × OD650) + 2 × 10 (Orth et al. 2010), these calculations were compared 9 to other cell enumeration methods. The cell density of P. gingivalis wild type and PG1788- 10 strain in OBGM and OB:CM obtained from viable plate counts and live/dead stain cell

11 counting using flow cytometry were employed and compared to the aforementioned OD650 12 measurement method for cell enumeration. 13 All cell count methods showed agreement when log phase (30 h) P. gingivalis wild type 14 and PG1788- in OBGM were compared (Fig. 4.5). Interestingly, the plate count of cell viability

15 corresponded well with the cell number calculated from the OD650 derived equation for P. 16 gingivalis W50 and PG1788- mutant after 50 h growth in OBGM (Fig. 4.5). However, 17 stationary phase cells under the same conditions appeared to be over-represented by flow 18 cytometry. This is probably due to the insensitivity of the flow cytometer used for 19 differentiation of large cell debris and membrane compromised cells from healthy living cells, 20 which suggested that this flow cytometry method employed is unsuitable for the cell count of 21 unhealthy bacterial cells. 22 All three methods showed similar trends in predicting cell numbers of P. gingivalis 23 wild type in OB:CM as in OBGM. P. gingivalis wild type and PG1788- strains were estimated 24 to constitute of 3.7×108 and 6×107 cells/mL respectively after 30 h growth in OB:CM, as 25 determined by the plate count method (Fig. 4.5). The calculated cell density of P. gingivalis - 26 wild type and PG1788 mutant after 30 h growth in OB:CM using the OD650 derived equation 27 were comparable. This experiment demonstrated that the cell number of different P. gingivalis 28 strains under these two growth conditions tested can be estimated based on calculation from

29 the OD650 derived equation.

105

1 2 Figure 4.5: Comparison of P. gingivalis cell count methods. Methods included P. gingivalis 3 plate count of viability (blue bar), cell number calculation based on OD650 (orange bar) and 4 live/dead stain counting by flow cytometer (grey bar). All three methods were derived from 5 serially diluted cells to ensure the cell range fell within the dynamic range for OD650 6 measurement and enumeration using flow cytometry method and countable cell range for 7 colony counting method. Experimental data were represented in biological triplicates.

8 4.2.3.3 Effects of P. gingivalis peptidase mutants on glycine release 9 In this study, changes in free glycine concentration during growth of P. gingivalis 10 peptidase mutant and wild type strains in OB:CM were examined at t = 30 and 40 h. Glycine 11 concentrations were not measured beyond the 50 h time point as there was concern that the 12 glycine concentration might be affected by cellular materials generated by cell lysis. 13 The rate of change of free glycine against P. gingivalis cell number was used to examine 14 the rate of free glycine production by different P. gingivalis strains during log phase growth 15 (Table 4.3, Fig. II.1). P. gingivalis wild type demonstrated an approximately 0.7 mM increase 16 in the rate of free glycine production per 109 cells in OB:CM (Table 4.3). Mutation of peptidase 17 encoding genes resulted in change of free glycine production rate during the 30 to 40 h growth 18 of these mutant strains in OB:CM. PG0445- produced the highest amount of free glycine in 19 OB:CM, which showed a production rate of 0.831 mM/109 cells that was higher than the wild 20 type. PG1605- strain that had a similar growth as wild type in OB:CM showed a slight reduction 21 in the rate of free glycine production at approximately 0.546 mM/109 cells (Table 4.3). Growth 22 of PG0753- and PG1788- in OB:CM from 30 to 40 h depicted an overall negative rate of glycine 23 production, suggesting that these two strains did not contribute to the production of free glycine

106

1 in this medium. PG0753- strain showed a 27-fold reduction in the rate of free glycine 2 production in OB:CM compared to wild type, while PG1788- showed a 4.5-fold reduction in 3 the rate of free glycine production in OB:CM. These results suggested that PG0753 and 4 PG1788 peptidases were likely to be involved in the release of free glycine during P. gingivalis 5 growth in OB:CM. 6 Table 4.3: Rate of free glycine production by different P. gingivalis strains in OB:CM. The rate 7 of free glycine production for each P. gingivalis strain was determined from the slope of a 8 linear regression line that indicates the change of free glycine concentration from the time 9 9 interval of 30 to 40 h against 10 P. gingivalis cells calculated from the OD650 derived equation. Rate of glycine production Strain Linear equation (mM/109 cells) W50 y=0.656x-0.0053 0.656 PG0445- y=0.8312x-0.3651 0.831 PG0753- y=-0.0238x+0.7195 -0.024 PG1605- y=0.5458x-0.0819 0.546 PG1788- y=-0.146x+0.4955 -0.146 10

11 4.3 Discussion 12 As glycine is an important growth substrate for T. denticola, which is found to be 13 closely associated with P. gingivalis, determination of P. gingivalis peptidases that are 14 potentially involved in the generation of glycine is important for a better understanding of their 15 synergistic interactions. Tan et al. (2014) demonstrated an increase in P. gingivalis hydrolytic 16 activities against glycine-containing peptides during growth in OB:CM. It is hypothesised that 17 cooperative proteolytic activities of P. gingivalis and T. denticola are one of the mechanisms 18 required for the increased production of free glycine, which in turn enhanced T. denticola 19 growth. This study was performed with the intention of determining P. gingivalis 20 uncharacterised peptidases that might play a role in the release of free glycine during P. 21 gingivalis growth in OB:CM. 22 P. gingivalis was predicted to possess a total of 68 putative proteases, which were used 23 for localisation predictive analysis. Four putative peptidases were selected for study, PG0445, 24 PG0753, PG1605 and PG1788. PG0445 and PG0753 were predicted with high confidence to 25 be localized in the cytoplasm and the absence of N-terminal signal peptide sequences 26 corroborated their localisation. A previous study also revealed that PG0445 was localized 27 cytoplasmically and was associated with a tetratricopeptide repeat protein, PG1385 (Kondo et 28 al. 2010). However, PG0753, which is annotated as a collagenase-like peptidase PrtQ, was not

107

1 expected to be localized in the cytoplasm, as collagenases typically have a role in the processing 2 of collagen in the extracellular milieu. PG1605 and PG1788 were predicted to contain N- 3 terminal signal peptides that signalled the exportation of proteins via the Sec translocon and 4 were predicted to be localized non-cytoplasmically using Phobius, other transmembrane 5 topology and signal peptide predictive software (Käll et al. 2007). In addition, PG1788 was 6 found in the lumen of P. gingivalis OMVs, which indicated that it is contained in the cellular 7 periplasm (Veith et al. 2014). Peptidases that can be packaged in the OMVs could also be 8 regarded as extracellular peptidases as OMVs have been found to be distributed to distant sites 9 from bacterial colonisation regions (Bonnington and Kuehn 2014, Xie 2015). Thus, it is 10 postulated that secreted proteases could act as public goods for neighbouring bacterial cells 11 within the polymicrobial biofilm community, where bacterial cells will have to acquire 12 nutrients in a highly competitive environment. 13 Deletion of PG0753 that encodes a potential PrtQ collagenase-like peptidase, resulted 14 in a mutant strain that demonstrated a dramatic decrease of free glycine production rate during 15 the 30 to 40 h growth in OB:CM, compared with the wild type. This suggested that PrtQ might 16 have a role in increasing the accessibility of glycine in OB:CM during early log growth phase. 17 This observation also implied that PG0753 has a role in hydrolysing glycine-containing 18 peptides and that the absence of PG0753 has altered the process of peptide degradation and 19 glycine generation. Whilst PrtC that has been shown to function as a true collagenase (Kato et 20 al. 1992), the gene expression of putative PrtQ collagenase that showed upregulation during P. 21 gingivalis coculture with T. denticola might be important for the processing of T. denticola 22 generated glycine-containing peptides in this conditioned medium (Tan et al. 2013). 23 Interestingly, PG1788- strain also exhibited a decline of free glycine production rate and a 24 lower cell density at the lag phase, as well as slower growth rate in OB:CM. This suggested 25 that peptidase PG1788 and PG0753 are likely to play a role in the release of free glycine from 26 peptides in TdCM. PG1605- strain showed a similar extent of free glycine production rate as 27 wild type in OB:CM. By contrast, PG0445- showed a higher free glycine production rate during 28 the 30 to 40 h growth in OB:CM. The increased release of glycine in PG0445- remains 29 unknown, however, deletion of PG0445 may in turn decrease steric hindrance to glycine 30 residues adjacent to the specific PG0445 cleavage site and alternatively increase the activity of 31 other peptidases. Indeed, it has been demonstrated that an additional peptidase mutation 32 (dpp11) in P. gingivalis triple mutant strain (dppIV-5-7-) enhanced the Met-Leu-MCA substrate

108

1 hydrolysis (Nemoto et al. 2016). These data suggested that PG0445 and PG1605 peptidases 2 were not essential for releasing glycine from peptides in TdCM.

3 4.4 Conclusion 4 P. gingivalis proteases have different roles in affecting the dynamics of global amine 5 group profile, including modulating the overall free glycine production in different growth 6 media. Inactivation of P. gingivalis peptidases for the investigation of their putative roles in 7 modifying the environmental glycine content can be confounding as P. gingivalis encodes 8 numerous proteinases and peptidases with complementary and overlapping cleavage activities. 9 Both PG0753- and PG1788- strains showed a reduction in glycine production rate during log 10 phase growth in OB:CM. In addition, deletion of PG1788 resulted in growth retardation in 11 OB:CM, suggesting that it has a role in the processing of the proposed growth suppressing 12 molecules produced by T. denticola or is essential for P. gingivalis nutrient acquisition in 13 OB:CM. Further experimental studies on PG0753 and PG1788 are warranted to determine the 14 biochemical and functional properties these two peptidases. 15 16 17 18 19 20 21 22 23 24 25 26 27 28

109

1 Chapter 5 Effect of T. denticola conditioned

2 medium on P. gingivalis gene expression

3 5.1 Introduction 4 Studies on transcriptional control of P gingivalis gene expression have played a 5 significant role in furthering our understanding of its virulence mechanisms. Messenger RNA 6 (mRNA) expression of a few select genes is typically undertaken by quantitative reverse 7 transcription polymerase chain reaction (qRT-PCR), however global transcriptomic 8 technologies enable the study of an organism’s transcriptome, which includes both coding and 9 non-coding RNA under a given set of conditions. Such technologies include (i) microarray 10 analysis, which investigates the change in an organism’s mRNA profile, and (ii) RNA 11 sequencing analysis, which determines the change in all RNA, including small and non-coding 12 RNA that cannot be identified in microarrays. The majority of global transcriptomic studies on 13 P. gingivalis have used microarrays, which use specific probes to target each mRNA sequence 14 of the genome simultaneously. Whilst a powerful technology, microarrays have a limited 15 dynamic range of detection due to background and saturation signals, and are reliant on specific 16 probes that are limited to genome annotation. Thus non-coding and small RNAs are not 17 detected. Microarray results provide a preliminary screening mechanism, which must then be 18 validated by qRT-PCR. More recently however, with a reduction in sequencing costs, an 19 alternative method, RNA sequencing (RNAseq) has been used to provide a more 20 comprehensive snapshot of an organism’s global transcriptional profile, including quantitative 21 analysis of the levels of transcription and non-coding RNAs. RNAseq also has a high dynamic 22 range of expression, demonstrating high sensitivity and specificity for detection of low (or 23 high) abundance transcripts, unknown transcripts and variants. 24 Transcriptional gene expression profiles of P. gingivalis have provided insights into 25 gene regulation, metabolism, micronutrient acquisition and virulence mechanisms using 26 microarrays (Dashper et al. 2009, Lo et al. 2009, Yamamoto et al. 2011, Lewis et al. 2009, 27 Romero-Lastra et al. 2017) and RNAseq (Høvik et al. 2012, Hirano et al. 2012). RNAseq 28 (Hendrickson et al. 2017, Kuboniwa et al. 2017), microarray (Simionato et al. 2006) and qRT- 29 PCR (Maeda et al. 2015) analyses of P. gingivalis interspecies interactions with Streptococcus 30 spp. indicated that P. gingivalis regulates its gene expression for cell-cell communication,

110

1 metabolism, colonisation and virulence factors during the heterotypic community 2 development. 3 P. gingivalis and T. denticola are the major etiological agents in the subgingival 4 polymicrobial biofilm of chronic periodontitis. Many studies have shown that P. gingivalis and 5 T. denticola exhibit close association and metabolic interactions. For example, T. denticola 6 produces succinic acid that is utilised by P. gingivalis, while P. gingivalis produces isobutyric 7 acid that is utilised by T. denticola (Grenier 1992). Although P. gingivalis is intimately 8 associated with T. denticola during chronic periodontitis progression, transcriptional profiles 9 of P. gingivalis for multimodal interactions with T. denticola are less well-studied, with a single 10 microarray analysis available (Tan et al. 2014). This microarray analysis of P. gingivalis and 11 T. denticola grown in monoculture and coculture in a continuous system revealed that P. 12 gingivalis upregulates thiamine biosynthesis pathways in the presence of T. denticola, while T. 13 denticola upregulates glycine cleavage pathways during coculture with P. gingivalis (Tan et 14 al. 2014). Given that P. gingivalis showed enhancement in the hydrolysis of glycine- 15 containing peptides in T. denticola conditioned medium (TdCM) (Tan et al. 2014), it is 16 hypothesised that an increased expression of P. gingivalis proteinases and peptidases could 17 contribute to the glycine cross-feeding reaction to its synergistic partner T. denticola. 18 This study will expand upon the results of Tan et al. (2014), which assessed mRNA 19 expression changes in P. gingivalis grown in planktonic co culture with T. denticola. Here, the 20 more comprehensive and sensitive RNAseq methodology was used to determine differential 21 gene expression (DGE) profiles of P. gingivalis during growth in OBGM and OB:CM, 22 providing insights into the multimodal interactions between P. gingivalis and T. denticola, but 23 without direct cell contact. In particular, this analysis aims to investigate: (1) potential 24 cooperative metabolic pathways of P. gingivalis with T. denticola, (2) P. gingivalis peptidases 25 that could be involved in the hydrolysis of glycine-bound peptides into free glycine for T. 26 denticola utilisation, (3) P. gingivalis putative immunity proteins and mechanisms that confer 27 P. gingivalis with an ability to survive and proliferate in the presence of the growth suppressing 28 factors in TdCM (4) P. gingivalis regulatory systems and hypothetical genes that become 29 activated or repressed under growth in T. denticola conditioned media. 30

111

1 5.2 Results and Discussion

2 5.2.1 RNA extraction and sequencing of P. gingivalis 3 RNA of P. gingivalis was extracted from three biological replicates during log phase

4 growth (OD650 ~0.5) in OBGM and OB:CM. The integrity of P. gingivalis RNA was first 5 examined on a 0.8 % (w/v) agarose gel stained with SYBR Safe (Fig. 5.1a), followed by 6 evaluation of the RNA quality and quantity using a RNA High Sensitivity assay on LabChip 7 (Fig 5.1b, Table 5.1). All RNA samples showed distinctive peaks and baseline separations of 8 the 5S (~0.12 kbp), 16S (~1.5 kbp) and 23S (~2.9 kbp) rRNAs, which are the indicators of the 9 overall RNA quality. The area ratio of 23S:16S that is >1.8 is considered to be RNA of high 10 quality as the size of 23S rRNA is approximately twice of the 16S rRNA. Similarly, all samples 11 showed a RNA quality score >8 (Table 5.1). The isolated RNA samples were also examined 12 for possible gDNA contamination by subjecting the RNA to 35 cycles of PCR amplification 13 without the production of any amplicon, indicating that the RNA was DNA-free (Fig. 5.2). The 14 resulting high quality RNA samples were sent to Micromon at Monash University for library 15 preparation and RNA sequencing. 16 Raw data processing and analysis of the DGE of P. gingivalis in OB:CM, relative to 17 OBGM, were performed by the Monash Bioinformatics platform. Reference sequences of P. 18 gingivalis W83 (NCBI reference sequence: NC_002950.2) was used for this analysis as the 19 complete sequence annotations of P. gingivalis W50 is not yet available. In the quality control 20 (QC) reports, all samples showed approximately 20 million reads with >90% mapped to the 21 gene features, except for the third biological sample of OBGM (OB_3) that had about half the 22 reads (~9.04 million counts) of the other samples, as well as the highest duplication and lowest 23 assignment rates (Table 5.2, Fig. 5.3). Nevertheless, OB_3 was included in the differential gene 24 expression analysis as the counts per million (CPM) standardisation of OB_3, which was used 25 for the fold-change analysis, was similar to the other samples (Table III.1) and was deemed 26 suitable for inclusion in the bioinformatics analysis. A high percentage of duplication for each 27 sample was normal for high reads with small genome size of P. gingivalis with 2.343 Mbp 28 (Table 5.2), indicating deep sequencing for each library. The unassigned reads were composed 29 of approximately 5% that were ambiguous and 4% with no feature indicating that the region of 30 the genome has not been annotated, while less than 1% of the reads were unmapped to the 31 reference sequence (Fig. 5.3).

112

1 There were a total of 1866 transcripts (out of 1915 protein coding genes) detected in 2 this analysis based on the new locus tag assigned unique identifiers, of which 132 protein- 3 encoding genes were differentially expressed in OB:CM, relative to OBGM. Based on the false

4 discovery rate (FDR) cut-off of <0.05 and fold-change log2 >0.585 (>1.5-fold), 45 transcripts 5 were shown to be significantly upregulated, while 87 transcripts were significantly 6 downregulated in OB:CM when compared to OBGM (Table 5.2, Table III.1). Due to the high 7 depth of coverage obtained from this RNAseq analysis, these data represent a comprehensive 8 transcriptomics profile of P. gingivalis during growth in medium enriched with T. denticola 9 produced effector molecules. TdCM used in this experiment would more closely resemble a 10 condition where T. denticola was present in high abundance and pre-modified the environment 11 with its diffusible molecules, which in turn would induce P. gingivalis regulations of gene 12 expression. 13 P. gingivalis genes that showed differential expression were inspected and compared 14 in multiple relevant databases for further classification and characterisation of their functional 15 roles. These differentially expressed genes were categorized into various specialized functional 16 groups, which included metabolic pathways (succinate pathways, one-carbon metabolism, 17 menaquionone biosynthesis), signal transduction and regulatory systems, transport systems, as 18 well as hypothetical protein encoding genes. Each functional category was analysed and 19 speculations upon their putative roles in relation to P. gingivalis potential multimodal 20 interactions with T. denticola were made. 21 22 23 24 25 26 27 28 29 30 31 32

113

1 (a)

2 3 (b)

4 5 Figure 5.1: Quality and integrity of total RNA isolated from P. gingivalis during growth in 6 OBGM or OB:CM (a) Total RNA on 0.8% (w/v) agarose gel stained with 0.5 × SYBR Safe. 7 Label “M” denotes DNA molecular weight markers in kbp. (b) A representative 8 electropherogram of the total RNA extracted from the first biological replicate of OBGM 9 (OB_1) using the RNA High Sensitivity assay on a Labchip GX Reviewer, indicated were the 10 peaks corresponded to 5S, 16S and 23S rRNA, as well as the lower marker (LM).

11 12 13 14 15 16 17 18 19 20 21 114

1 2 Figure 5.2: Examination of RNA for possible source of gDNA contamination. RNA samples 3 were subjected to 35 cycles of PCR that targeted PG0445 and PG1605 using primer pairs 4 PG0445confo_F/RP and PG1605confo_F/RP, respectively. Positive control of P. gingivalis 5 gDNA material and representative samples of RNA extracted from OBGM_1 (OB_1) and 6 OB:CM_1 (CM_1) were subjected to 0.8% (w/v) agarose gel electrophesis. Label M denotes 7 DNA molecular weight markers in kbp.

8 9 Table 5.1: Total RNA quality and quantity of P. gingivalis in OBGM (OB) and OB:CM in 10 three biological replicates each. The tabulated data was generated by the LabChip GX Touch 11 data analysis software.

Total rRNA Area rRNA Height RNA Peak Conc. RNA Ratio Ratio rRNA Fast Quality Sample Count (ng/ul) Area [28S/18S] [28S/18S] Area Ratio Score 5S Area 5S %Total 18S Area 18S %Total 28S Area 28S %Total OB_1 9 90.05 96.29 1.93 1.52 0.11 8.2 17.86 18.60% 16.72 17.40% 32.29 33.50% OB_2 7 75.71 82.05 2.05 1.64 0.07 8.3 16.2 19.70% 13.1 16.00% 26.88 32.80% OB_3 6 70.79 76.86 1.94 1.56 0.09 8.1 3.49 4.50% 12.23 15.90% 23.72 30.90% OB: CM_1 7 78.75 86.01 1.82 1.45 0.12 8.4 9.56 11.10% 14.66 17.00% 26.75 31.10% OB: CM_2 9 86.31 94.91 1.96 1.65 0.1 8.3 2.61 2.80% 15.31 16.10% 30.03 31.60% OB: CM_3 6 81.51 85.43 2.07 1.66 0.12 8.2 5.28 6.20% 13.73 16.10% 28.43 33.30% RNA STD12 480 500.5 13 14 15 16 17 18 19

115

1 Table 5.2: Summary of the sequencing library with number of assigned reads and percentage 2 of duplications per sample.

3 4

5 6 Figure 5.3: Quality control reports of the RNA sequencing library. Percentage representations 7 of the feature counts assignments of the library. This information was obtained from the 8 multiQC report prepared by the Monash Bioinformatics Platform. OB and CM represented 9 OBGM and OB:CM samples respectively in three biological replicates.

116

1 2 Figure 5.4: Volcano plot of P. gingivalis transcripts that were detected in the RNAseq analysis. 3 The dotted lines represent the cut-off of false discovery rate (FDR) < 0.05 and fold change log2 4 >0.585. Genes that showed differential expression during P. gingivalis growth in OB:CM, 5 relative to OBGM, were indicated in red dots. This image was sourced from the Degust web 6 tool generated by Monash Bioinformatics Platform.

7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 117

Table 5.3: P. gingivalis genes that were differentially expressed (fold change ≥1.5, FDR <0.05) during growth in OB:CM, compared to OBGM.

New locus Gene ID Product Fold FDR AveExpr Avg CM Avg OB COG tag change CPM CPM categories PG_RS03660 hypothetical protein -1.54 0.03 6.99 105.49 154.68 Unassigned PG_RS05505 hypothetical protein -1.64 0.02 9.89 767.38 1176.93 Unassigned PG_RS00430 hypothetical protein -1.64 0.03 5.31 32.39 49.60 Unassigned PG_RS10565 hypothetical protein -1.67 0.02 6.23 60.02 95.03 Unassigned PG_RS06450 hypothetical protein -1.98 0.01 7.83 167.44 315.89 Unassigned PG_RS04805 hypothetical protein -2.50 0.01 3.47 7.31 17.06 Unassigned PG_RS10535 hypothetical protein -2.71 0.01 3.65 8.32 20.02 Unassigned PG_RS10570 hypothetical protein -3.26 0.02 2.82 4.31 12.81 Unassigned PG_RS10595 hypothetical protein -4.38 0.01 -1.65 0.14 0.62 Unassigned PG_RS00055 PG0010b ATP-dependent Clp protease ClpC -1.73 0.03 10.82 1446.97 2320.64 O PG_RS00115 PG0025 2-hydroxyhepta-2,4-diene-1,7-dioate isomerase -2.19 0.01 4.33 14.41 29.05 Q PG_RS00210 PG0045 molecular chaperone HtpG 2.15 0.00 8.83 684.77 303.76 O PG_RS00220 PG0047b cell division protein FtsH -2.05 0.01 11.19 1711.78 3254.30 O PG_RS00270 PG0059 DUF2007 domain-containing protein -1.63 0.02 4.49 18.12 27.85 Unassigned PG_RS00285 PG0062b helicase -1.83 0.01 5.11 26.52 45.58 Unassigned PG_RS00290 PG0063d TolC family protein 2.49 0.00 3.58 19.60 7.31 MU PG_RS00295 PG0064d CusA/CzcA family heavy metal efflux RND 2.47 0.00 4.15 28.60 11.24 P transporter PG_RS00300 PG0065d efflux RND transporter periplasmic adaptor subunit 2.85 0.00 3.11 14.71 5.10 M PG_RS00360 PG0081e hypothetical protein -1.94 0.01 9.80 667.67 1212.45 Unassigned PG_RS00365 PG0082 glycoside hydrolase -2.16 0.01 7.34 116.06 235.07 Unassigned PG_RS00370 PG0083 PorT family protein -2.03 0.01 5.57 34.71 65.92 Unassigned PG_RS00375 PG0084a L-serine ammonia-lyase -1.79 0.01 6.21 57.98 96.25 E

118

PG_RS00380 PG0085 alpha-galactosidase -1.53 0.03 5.81 47.14 67.89 Unassigned PG_RS00385 PG0086 DEAD/DEAH box helicase -1.52 0.01 9.19 488.56 698.46 LK PG_RS00595 PG0128 LPS kinase -1.56 0.04 3.75 11.27 16.32 Unassigned PG_RS00665 PG0143 apolipoprotein acyltransferase -2.13 0.00 5.50 32.27 64.81 R PG_RS00670 PG0144 agmatine deiminase family protein -2.08 0.02 4.76 20.21 38.45 E PG_RS00950 PG0209a formate transporter -3.21 0.00 4.08 9.97 29.50 P PG_RS01045 PG0228e hypothetical protein 1.50 0.01 6.53 117.12 73.79 E PG_RS01160 PG0257d Fe-S cluster assembly protein SufB -1.82 0.04 11.65 2493.50 4252.82 O PG_RS01170 PG0259d Fe-S cluster assembly protein SufD -1.71 0.03 8.27 247.66 393.20 O PG_RS01285 PG0287 membrane protein -1.68 0.01 5.86 46.38 73.34 Unassigned PG_RS01290 PG0288 hypothetical protein -1.61 0.01 8.42 278.26 424.58 Unassigned PG_RS01560 PG0350 hypothetical protein 1.53 0.01 9.00 652.49 403.55 Unassigned PG_RS01590 PG0358 aspartate carbamoyltransferase regulatory subunit -1.67 0.01 8.95 396.70 619.35 F PG_RS01595 PG0359 flavin reductase -1.73 0.01 5.89 46.47 76.46 R PG_RS01865 PG0419 DUF2807 domain-containing protein -1.56 0.04 7.31 130.56 196.67 Unassigned PG_RS01870 PG0421 DUF2807 domain-containing protein -2.27 0.02 8.83 324.14 681.21 Unassigned PG_RS01920 PG0432 SAM-dependent methyltransferase -2.39 0.02 10.51 1022.30 2187.68 J PG_RS01925 PG0433 S-adenosylmethionine-dependent -2.52 0.02 5.69 35.80 79.81 R methyltransferase PG_RS01930 PG0434e hypothetical protein -2.09 0.02 10.29 925.98 1754.15 Unassigned PG_RS02305 PG0520 molecular chaperone GroEL 1.51 0.01 10.61 1984.13 1241.84 O PG_RS02310 PG0521 molecular chaperone GroES 1.85 0.04 9.48 1007.00 527.57 O PG_RS02410 PG0539 efflux RND transporter periplasmic adaptor subunit -1.55 0.03 4.79 23.14 33.49 M PG_RS02510 PG0565e hypothetical protein -1.90 0.05 9.13 425.46 766.53 Unassigned PG_RS02515 PG0566c histidinol-phosphate aminotransferase 1.64 0.01 5.77 72.21 41.38 Unassigned PG_RS02540 PG0574e hypothetical protein 1.55 0.01 5.50 58.27 35.25 Unassigned PG_RS02595 PG0585 aspartyl-tRNA amidotransferase subunit B 1.70 0.01 8.14 376.65 213.71 S

119

PG_RS02735 PG0620b endopeptidase La -2.02 0.01 10.68 1202.23 2258.49 O PG_RS02740 PG0621 DUF1599 domain-containing protein -1.98 0.01 5.95 45.19 85.46 Unassigned PG_RS02935 PG0668 TonB-dependent receptor 1.60 0.01 8.65 521.28 311.39 P PG_RS02940 PG0669 heme-binding protein 1.58 0.01 10.64 2059.70 1235.52 H PG_RS02985 PG0679d TolC family protein -4.34 0.00 5.69 25.71 105.08 MU PG_RS02990 PG0680d efflux RND transporter periplasmic adaptor subunit -6.10 0.00 6.05 27.65 159.56 M PG_RS03000 PG0682d ABC transporter permease -6.02 0.00 5.80 23.47 133.94 V PG_RS03005 PG0683d ABC transporter permease -4.45 0.00 4.16 8.76 36.58 V PG_RS03010 PG0684d ABC transporter permease -4.53 0.00 5.55 22.82 97.68 Unassigned PG_RS03015 PG0685d ABC transporter ATP-binding protein -2.93 0.00 3.48 6.73 18.66 V PG_RS03020 PG0686c hypothetical protein -2.12 0.03 11.96 2957.40 5661.20 S PG_RS03030 PG0689a NAD-dependent alcohol dehydrogenase 1.66 0.02 9.01 683.91 396.89 C PG_RS03040 PG0691a NifU family protein 1.73 0.01 4.71 35.54 19.39 O PG_RS03045 PG0692a 4-hydroxybutyryl-CoA dehydratase 2.09 0.01 12.35 7791.32 3545.50 Q PG_RS03145 PG0717e hypothetical protein 4.62 0.00 8.64 884.28 184.21 Unassigned PG_RS03150 PG0718e hypothetical protein 5.40 0.00 4.07 40.05 7.03 Unassigned PG_RS03600 PG0822 nuclear transport factor 2 family protein -1.79 0.04 5.30 30.51 52.31 R PG_RS03650 PG0832 nucleoid-associated protein -2.35 0.01 9.14 381.12 856.86 Unassigned PG_RS03655 PG0833 DUF3732 domain-containing protein -1.97 0.04 11.10 1627.10 3045.43 Unassigned PG_RS03735 PG0849e hypothetical protein 1.62 0.02 5.27 50.79 29.43 Unassigned PG_RS03785 PG0860c transcriptional regulator 1.58 0.03 5.83 74.45 44.39 Unassigned PG_RS03915 PG0886a 2-amino-4-hydroxy-6- -2.21 0.03 10.53 1050.23 2205.90 H hydroxymethyldihydropteridine pyrophosphokinase PG_RS03920 PG0889b peptidase M24 family protein -1.74 0.02 8.84 365.31 587.87 E PG_RS03960 PG0899 cytochrome D ubiquinol oxidase subunit II -1.68 0.01 8.17 231.02 364.58 C

120

PG_RS04260 PG0964 CDP-diacylglycerol-serine-O- 1.53 0.04 4.43 27.84 17.06 Unassigned phosphatidyltransferase PG_RS04520 PG1025d DUF2813 domain-containing protein 1.58 0.04 7.58 247.42 151.30 R PG_RS04565 PG1036 excinuclease ABC subunit A -1.74 0.02 10.57 1200.60 1948.27 L PG_RS04600 PG1043c ferrous iron transporter B -2.10 0.01 7.09 97.36 193.77 P PG_RS04605 PG1044c DNA-binding protein -1.86 0.02 8.76 336.99 577.21 K PG_RS04675 PG1063c transcriptional regulator 1.72 0.01 5.33 54.13 30.28 Unassigned PG_RS04785 PG1085e hypothetical protein -1.97 0.04 9.66 589.01 1171.94 Unassigned PG_RS04795 PG1088 N-acetyltransferase -1.76 0.05 7.73 170.28 275.26 R PG_RS04935 PG1114 aspartate 1-decarboxylase -1.53 0.02 7.07 113.13 162.50 H PG_RS04945 PG1116a bifunctional methylenetetrahydrofolate -1.87 0.03 8.13 216.02 377.56 H dehydrogenase/ methenyltetrahydrofolate cyclohydrolase PG_RS04955 PG1118 chaperone protein ClpB 1.73 0.01 6.13 94.69 51.86 O PG_RS04980 PG1124 cobalamin adenosyltransferase -1.73 0.05 10.33 1027.20 1661.02 S PG_RS04995 PG1127c transcriptional regulator 1.63 0.02 9.28 816.96 477.29 K PG_RS05010 PG1130e hypothetical protein -1.91 0.05 9.12 423.82 764.45 Unassigned PG_RS05035 PG1134 thioredoxin-disulphide reductase -1.67 0.04 10.24 988.15 1525.79 O PG_RS05130 PG1155 ADP-heptose--LPS heptosyltransferase -1.74 0.01 7.65 157.76 258.00 M PG_RS05170 PG1165e hypothetical protein 1.64 0.03 6.35 108.77 62.62 Unassigned PG_RS05210 PG1176d ABC transporter ATP-binding protein -1.69 0.04 3.56 9.58 14.95 V PG_RS05320 PG1208 molecular chaperone DnaK 1.96 0.00 9.43 999.43 480.03 O PG_RS05625 PG1276c histidinol-phosphate aminotransferase 1.76 0.01 5.68 70.52 37.64 Unassigned PG_RS05635 PG1278a phosphoserine aminotransferase 1.52 0.03 10.12 1408.27 884.96 HE PG_RS05815 PG1321a formate--tetrahydrofolate ligase -2.71 0.01 9.36 424.86 1052.19 F PG_RS05835 PG1326 hemagglutinin 1.57 0.04 7.18 188.46 114.26 Unassigned PG_RS06065 PG1379d ABC transporter substrate-binding protein 1.55 0.02 6.36 105.83 64.67 P

121

PG_RS06080 PG1382d porin 1.55 0.01 8.34 416.35 254.15 Unassigned PG_RS06095 PG1385e hypothetical protein 1.56 0.01 10.37 1702.11 1041.25 Unassigned PG_RS06220 PG1414 TonB-dependent receptor 1.52 0.05 8.06 340.29 215.59 P PG_RS06455 PG1469 restriction endonuclease subunit M -1.61 0.05 11.49 2360.45 3569.40 V PG_RS06490 PG1476 conjugative transposon protein TraM -1.66 0.04 3.09 6.89 10.62 Unassigned PG_RS06570 PG1492 GLPGLI family protein -1.52 0.05 12.67 5468.01 7863.12 Unassigned PG_RS06575 PG1493e hypothetical protein -1.54 0.02 10.36 1089.90 1595.18 Unassigned PG_RS06705 PG1521a O-succinylbenzoic acid--CoA ligase -1.74 0.01 3.98 12.31 20.31 IQ PG_RS06800 PG1541a 2-amino-4-hydroxy-6- -2.19 0.02 5.02 23.52 47.21 H hydroxymethyldihydropteridine pyrophosphokinase PG_RS06805 PG1542b collagenase-like protease -2.31 0.02 10.00 716.61 1515.28 O PG_RS06815 PG1544 peroxide stress protein YaaA -1.62 0.04 5.45 35.47 54.37 S PG_RS06855 PG1554e hypothetical protein -1.94 0.04 2.55 4.45 8.04 Unassigned PG_RS06865 PG1556 DUF2149 domain-containing protein -1.93 0.01 3.44 8.00 14.71 S PG_RS07215 PG1640d MATE family multidrug exporter -1.99 0.03 6.25 57.27 106.05 V PG_RS07220 PG1641d low molecular weight phosphotyrosine protein -2.14 0.04 5.70 38.68 75.13 T phosphatase PG_RS07225 PG1642d copper-translocating P-type ATPase -2.71 0.01 10.12 731.52 1784.63 Unassigned PG_RS07540 PG1713 rhodanese-like domain-containing protein 1.56 0.01 5.90 76.62 46.58 P PG_RS07715 PG1755 class I fructose-bisphosphate aldolase 1.76 0.01 11.59 4241.17 2266.00 G PG_RS07730 PG1758 30S ribosomal protein S15 -2.17 0.01 9.48 504.83 1024.02 J PG_RS08065 PG1829 long-chain-fatty-acid--CoA ligase 1.98 0.00 6.42 124.61 59.61 I PG_RS08090 PG1837 hemagglutinin A -1.53 0.02 12.36 4369.44 6364.55 Unassigned PG_RS08100 PG1842 acetyltransferase -2.15 0.02 5.83 42.03 81.25 R PG_RS08250 PG1881 hypothetical protein 1.65 0.01 8.68 540.00 311.54 Unassigned PG_RS08400 PG1908 GLPGLI family protein -1.56 0.02 9.51 607.68 886.40 Unassigned

122

PG_RS08715 PG1974e hypothetical protein -1.74 0.03 6.79 88.15 142.58 Unassigned PG_RS08850 PG2006e hypothetical protein -2.73 0.01 5.31 25.35 64.89 Unassigned PG_RS08860 PG2009 DNA repair protein RecO 1.51 0.01 5.54 58.98 36.80 L PG_RS09040 PG2047c helicase -1.75 0.01 7.43 135.00 222.55 Unassigned PG_RS09230 PG2082d MFS transporter 1.64 0.01 10.04 1386.01 814.64 E PG_RS09270 PG2090d cation transporter 1.53 0.05 4.35 26.30 16.06 P PG_RS09385 PG2119 gfo/Idh/MocA family oxidoreductase 1.53 0.01 7.88 299.93 185.63 R PG_RS09670 PG2177a NADH:ubiquinone reductase (Na(+)-transporting) 1.75 0.01 9.38 905.95 496.04 C subunit F PG_RS09680 PG2179a NADH:ubiquinone reductase (Na(+)-transporting) 1.53 0.03 5.23 48.21 29.81 C subunit D PG_RS09710 PG2187a 1,4-dihydroxy-2-naphthoate octaprenyltransferase -2.39 0.01 3.40 7.08 15.88 H PG_RS09725 PG2190d phosphonate ABC transporter ATP-binding protein 1.63 0.04 9.33 847.80 496.37 D PG_RS09785 PG2205 2-dehydropantoate 2-reductase -2.02 0.04 7.82 171.50 317.96 H Superscript at gene ID indicates the section by which the protein encoding genes were addressed a. Metabolic pathways b. Peptidases and proteinases c. Signal transduction and transcription regulators d. Transporters e. Hypothetical protein

* One-letter abbreviations for the functional COG categories: Information storage and processing (J, translation, ribosomal structure and biogenesis; K, transcription; L, replication, recombination and repair), Cellular processes and signalling (D, cell cycle control, cell division, chromosome partitioning; M, cell wall/membrane/envelope biogenesis; N, cell motility; O; post-translational modification, protein turnover, chaperones; T, signal transduction mechanisms; U, intracellular trafficking, secretion and vesicular transport; V, defence mechanisms), metabolism (C, energy production and conversion; E, amino acid transport and metabolism; F, nucleotide transport and metabolism; G, carbohydrate transport and metabolism; H, coenzyme transport and metabolism; I, lipid transport and metabolism; P, inorganic ion transport and metabolism; Q, secondary metabolites biosynthesis) and poorly or uncharacterised (R, general function prediction only; S, function unknown) and unassigned COG. 123

5.2.2 Classification of differentially expressed genes 5.2.2.1 Metabolic pathways A total of 25 respiratory chain components and metabolic enzyme encoding genes showed >1.5-fold change (FDR <0.05) during P. gingivalis growth in OB:CM, relative to OBGM (Table 5.3, Table III.2), of which 12 showed significant upregulation and 13 were significantly downregulated. These metabolism related genes were enriched in pathways corresponding to cofactors, prosthetic groups, biosynthesis of electron carriers, vitamin biosynthesis, one-carbon compound assimilation, reductant biosynthesis, amino acid biosynthesis and melibiose degradation. A number of metabolic enzyme-encoding genes are likely to be involved in multiple reactions and pathways, as shown in Table III.2. While there were many single intermediary genes in the metabolic pathways that showed differential expression, multiple enzyme-encoding genes delineating the entire pathway provided a higher confidence in deducing potential metabolic change and physiological variations exhibited by P. gingivalis during growth in OB:CM. For example, a number of metabolic enzyme encoding genes that were differentially expressed correspond to P. gingivalis succinate catabolism and one-carbon metabolism, which are both categorized in the carbon metabolism module in KEGG. Of the cofactor and vitamin biosynthesis pathways enriched for transcriptional change, menaquinone-biosynthetic enzyme encoding genes were significantly reduced in their expression levels. These pathways were examined to gain insights into possible metabolic interactions between P. gingivalis and T. denticola. 5.2.2.1.1 Succinate pathways P. gingivalis showed significant upregulation of genes encoding enzymes in succinate utilisation pathways during growth in OB:CM, relative to OBGM (Fig. 5.5). This result supports the findings of an earlier study that inferred cross-feeding of succinate from T. denticola to P. gingivalis by detection of succinic acid in T. denticola spent medium and supplementation of succinic acid to increase P. gingivalis growth (Grenier 1992). P. gingivalis genes that encode acyl-CoA reductase and succinate semialdehyde reductase have been shown to be essential for P. gingivalis growth, as well as being associated with the production of isobutyrate and isovalerate fatty acids (Yoshida et al. 2015, Yoshida et al. 2016). In P. gingivalis butyrate (also known as butanoate) metabolic pathways, genes encoding succinate semialdehyde reductase (PG0689) and 4-hydroxybutyryl-CoA dehydratase (PG0692) that are

124

transcribed in an operon comprised of PG0687-92 were significantly upregulated, while the transcripts of acyl-CoA reductase (PG0687) and 4-hydroxybutyryl-CoA transferase (PG0690) were increased 1.36-fold and 1.41-fold respectively in OB:CM, relative to OBGM (Fig 5.5). Although PG0691 was not predicted to be involved in the butyrate metabolism, PG0691 that encodes a NifU-family protein was also significantly upregulated (Table 4.3). Increased expression of these metabolic genes could lead to accumulation of butyrate (Fig. 5.5). Although butyrate can also be derived from P. gingivalis via metabolism of glutamate, aspartate or their peptide moieties (Takahashi et al. 2000), the metabolic cross-feeding of succinate to P. gingivalis can increase the level and rate of butyrate production (Yoshida et al. 2015, Yoshida et al. 2016). An increase of this highly cytotoxic product has been demonstrated to be one of the factors contributing to the exacerbation of disease (Scragg et al. 1994, Kuniyasu and Tomoko 2009). Hence, it is postulated that metabolic exchange between P. gingivalis and T. denticola can lead to increased cytotoxic end products and contribute to disruptive effects on host cell activities, as well as immune defence mechanisms. Short chain fatty acid cross-feeding interactions in complex polymicrobial communities have been demonstrated to provide a competitive advantage to members of the consortium. For example, microbiota-derived succinate enabled Salmonella enterica serovar Typhimurium and Clostridium difficile to colonise and infect the gut of gnotobiotic mice as succinate can be an important carbon source for their survival and growth in a polymicrobial gut microbiota community (Ferreyra et al. 2014, Spiga et al. 2017). Although succinate is the catabolic product of many fermentative pathways and might not be utilised during growth in nutrient rich environments, bacteria that can utilise succinate during growth in complex polymicrobial biofilms could gain a fitness advantage. Furthermore, other bacterial species that have a close association with P. gingivalis, such as S. gordonii, T. forsythia, A. actinomycetemcomitans and F. nucleatum are also capable of producing succinate, suggesting that succinate could be an important carbon source for P. gingivalis in the oral polymicrobial biofilm. Likewise, it has been postulated that succinate could serve as a partial alternative to rescue cell growth under heme limited conditions (Baughn and Malamy 2002, Dashper et al. 2009). Thus, P. gingivalis succinate utilizing pathway might enable metabolic interaction of P. gingivalis with other oral bacterial species that produce short chain fatty acids, as well as to increase nutrient utilisation efficiency in the polymicrobial community. It might be especially beneficial to P. gingivalis to grow and colonise in regions that had been pre-occupied by early and late colonisers of other

125

bacterial species in the polymicrobial community, as with the competitive advantages confer on Salmonella Typhimurium and Clostridium difficile (Ferreyra et al. 2014, Spiga et al. 2017).

126

Figure 5.5: Butyrate metabolism of P. gingivalis W83. The metabolic genes and pathway reactions of P. gingivalis W83 (purple bordered box) were sourced from KEGG. The red and orange filled boxes correspond to metabolic genes that showed significant upregulation (>1.5-fold change, FDR <0.05) and >1.4-fold increase in expression, respectively during P. gingivalis growth in OB:CM, relative to OBGM. The black arrows correspond to the metabolic enzymes encoding genes that did not show differential expression.

127

5.2.2.1.2 Glycine related pathways Glycine can be synthesized in the glycine, serine and threonine superpathways by one- step reversible conversion of serine or threonine to glycine (Fig 5.6). PG0474 that encodes a threonine aldolase, which catalyses the interconversion of threonine to glycine, showed a 1.3- fold upregulation, whilst PG1278 that encodes a phosphoserine aminotransferase, which is involved in the serine biosynthesis pathway, showed significant upregulation during growth in OB:CM. Likewise PG0084 (sda), a monocistronic gene that encodes L-serine ammonia lyase, which is responsible for the degradation of serine to CO2 and pyruvate, showed significant downregulation (Fig. 5.6). Increased gene expression of a serine biosynthesis enzyme and decreased gene expression of a putative serine degradation pathway were observed in this study, suggesting accumulation of a glycine precursor during P. gingivalis growth in OB:CM. Other pathways that could affect the glycine level, include the glycine reductase system and the glycine cleavage system (GCS). P. gingivalis lacks the glycine reductase system. Nevertheless, glycine can also be generated or utilised via the reversible GCS, depending on the equilibrium state of metabolites. Biochemically, glycine can be produced by oxidation of serine, which is coupled with the reduction of tetrahydrofolate (THF) into 5,10-methyleneTHF

(5,10-CH2-THF) in the GCS (Fig. 5.6). Thus, this reaction is also integrated as part of the one carbon metabolism that involves the transfer of one carbon molecules. In addition, one-carbon metabolism is also intertwined with other pathways, which include homocysteine and methionine conversion pathways, as well as purine nucleotides and thymidylate biosynthesis pathways, each of which are dependent upon different redox states of folate. In the one carbon metabolic pathway, P. gingivalis showed significant reduction of folD and fhs expression that are responsible for the generation of N10-formylTHF (N10-fTHF) during growth in OB:CM. On the contrary, T. denticola N10-fTHF-generating enzyme encoding genes showed significant upregulation during coculture with P. gingivalis (Tan et al. 2014). While most bacterial species are only equipped with one N10-fTHF generating enzyme, P. gingivalis and T. denticola were found to encode both the N10-fTHF generating enzymes that utilise different pathways and substrates. P. gingivalis fhs (PG1321), a monocistronic gene, showed a 2.7-fold reduction in expression during growth in OB:CM, compared with OBGM (Table 5.3, Fig. 5.6). Fhs, also known as a fTHF synthetase, is predominantly present in facultative or strict anaerobes for the incorporation of formate into the one-carbon pathway (Sah et al. 2015); whereby, the source

128

of formate could be generated via the reaction of pyruvate lyase into formate and acetyl-CoA under anaerobic or fermentative pathways. In addition, a gene-encoding a putative formate transporter (PG0209) that is involved in the bidirectional transport of formate was also significantly downregulated (Table 5.3). Interestingly, the expression of fhs and PG0209 have been demonstrated to be elevated during P. gingivalis growth in microaerophilic conditions, in comparison to anaerobic conditions (Lewis et al. 2009). THF ligase (FolD, PG1116) is a mono-, bi- or trifunctional enzyme that carries out the sequential steps for reversible conversion of 5,10-CH2-THF to 5,10-methenyl-THF (5-CH- THF), followed by the formation of N10-fTHF via its dehydrogenase and cyclohydrolase activities. FolD also plays an important role in maintaining the ratio of NADP+ to NADPH, as + the oxidation of 5,10-CH2-THF to 5-CH-THF is accompanied with the reduction of NADP . This one-carbon metabolism related folD was significantly downregulated (1.87-fold) in P. gingivalis grown in OB:CM (Table 5.3, Fig. 5.6). The decreased expression of these pathways is expected to result in a reduction of N10- fTHF, which is required for the formylation of aminoacylated tRNA initiator by formyl-Met- tRNAfMET transferase (Fmt) and de novo biosynthesis of purine nucleotides (Sah et al. 2015). T. denticola showed a higher rate of free glycine utilisation based on its glycine uptake rate (Tan et al. 2014), thus it is speculated that T. denticola is more efficient in the production of N10-fTHF via its GCS. Compiling data from this study and that of Tan et al. (2014), it is hypothesized that the glycine cross-feeding reaction from P. gingivalis to T. denticola might be related to the one carbon metabolism cycle. In addition, this RNAseq analysis also demonstrated significant downregulation (>2- fold) of PG0886 and PG1541 (folK) expression that are involved in the biosynthesis pathway of 6-hydroxymethyl-dihydropterin diphosphate (Table 5.3, Table III.2), which is required as an intermediate for pterin biosynthesis. Only PG0886 that is predicted to be expressed in the operon of PG0882-6 showed differential expression, while the remaining genes in the operon displayed a <1.3-fold change. Consistently, the transcription levels of PG1541-PG1542 were significantly decreased in OB:CM. However, the expression levels of the first four genes of the predicted PG1536-43 operon were no different in either medium, while PG1540 and PG1543 showed a 1.6-fold and 1.4-fold decrease in expression in OB:CM, compared with OBGM. The significant decrease in expression of enzyme-encoding genes for pterin biosynthesis could

129

translate into a reduction of pterin, which is required as a folate precursor and hence would affect the downstream one-carbon metabolism.

Figure 5.6: P. gingivalis serine, threonine and glycine biosynthesis superpathways to one carbon pathways. The red and cyan proteins (with coloured double arrows) are encoded by metabolic genes that showed significant (>1.5-fold, FDR <0.05) upregulation and downregulation, respectively. The orange protein-encoding genes (with red double arrow) had a ~1.3-fold increase in expression, while the black protein-encoding genes (with black double arrow) showed no change in expression during P. gingivalis growth in OB:CM, relative to OBGM.

130

5.2.2.1.3 Menaquinone biosynthesis pathways

Menaquinone (MK; Vitamin K2) plays an essential role in several anaerobic electron transport systems. It functions as an essential cofactor for the electron transport chain that is commonly found within the inner cytoplasmic membrane for bacterial anaerobic respiration. P. gingivalis W50 possesses the ability to synthesize vitamin K based on the putative metabolic pathways analysis in the genome of P. gingivalis W83 and its ability to grow in the absence of an exogenous supply of vitamin K (Tan 2012). However, other strains of P. gingivalis that lack these pathways required the supplementation of vitamin K for growth (Wyss 1992). This could provide P. gingivalis W50 and W83 strains with competitive advantages to survive and compete in the polymicrobial community by incorporating MK for anaerobic respiration. MK or its intermediates are also important for the growth of oral Bifidobacterium, Porphyromonas and Prevotella spp. (Marcotte and Lavoie 1998, Isawa et al. 2002, Hojo et al. 2009). Genes encoding enzymes in the MK biosynthesis pathways, PG1521 and PG1524 that are expressed as part of the PG1521-26 operon showed decreased expression during P. gingivalis growth in OB:CM, relative to OBGM (Fig. 5.7). PG1526, which encodes a putative 1,4-dihydroxy-2- naphthoyl-CoA synthase (MenB in the MK pathways), is predicted to be the first gene of the PG1521-26 operon and showed an approximately 1.2-fold increase in expression. This small change in PG1526 expression could be due to a potential small RNA transcript that was detected in the complementary strand of PG1526 (Høvik et al. 2012), which might contribute to the sequence reads across the gene and thus suggest a higher level of expression in OB:CM. Likewise, PG2187 that encodes a 4-dihydroxy-2-naphthoate octaprenyltransferase (MenA in the MK pathways) is an ORF from the PG2187-90 operon on the antisense strand. This gene showed a 2.39-fold decrease in expression (FDR <0.01) in OB:CM, compared with OBGM (Table 5.3, Fig. 5.7). The other downstream genes of PG2187 were not differentially expressed under these conditions, except for PG2190 that demonstrated a 1.6-fold increased in expression. Examination of transcriptional profiles available online show that the transcription level of the first third of PG2187 attained a similarly high level of abundance as the downstream genes, while the following two-thirds of the gene had an approximately two to three-fold reduction in transcription levels (Høvik et al. 2012), which might be associated with the secondary structure of PG2187. It has been shown that secondary structures can introduce bias to transcriptional levels (Price et al. 2017).

131

Interestingly, PG2177 and PG2179 that encode NADH:ubiquinone reductase [Na(+)- transporting] subunit F and subunit D respectively showed significant upregulation (Table 5.3). These were the only transcripts in PG2177-82 operon that showed differential expression, while PG2178 showed a 1.47-fold increase in expression and PG2180-82 had a fold change of <1.25. These subunits are part of the Nqr enzyme complex, which is involved in the extrusion of Na+ ions coupled with the reduction of MK and oxidation of NADH in the bacterial anaerobic respiratory chain (Meuric et al. 2010). Based on these results, one would speculate that MK might be cross-fed from T. denticola to P. gingivalis, especially during cell lysis of T. denticola, which would result in a release of MK or its precursors into the medium. The metabolic gene clusters for MK biosynthesis were absent in T. denticola ATCC 35405 based on the pathway searches of the KEGG database, while T. denticola requires the supplementation of vitamin K for growth (Wyss 1992, Tan 2012). Thus, it is speculated that vitamin K was taken-up by T. denticola during growth in OBGM, while T. denticola released the modified form of MK into the medium, presumably by cell lysis, for P. gingivalis utilisation. However, P. gingivalis MK biosynthesis genes were also significantly downregulated during polyphosphate treatment (Ji-Hoi et al. 2014). As polyphosphate exerts an antibacterial effect via inhibition of bacterial hemin acquisition, the decreased transcription of MK biosynthesis genes in this context might be due to a lower hemin concentration in OB:CM as a result of T. denticola hemin utilisation during growth. Otherwise, T. denticola-produced growth suppressing molecules could also be a contributory factor for P. gingivalis decreased MK biosynthesis. Likewise, another study has also speculated that antimicrobials produced by salivary bifidobacteria could be inhibitory to P. gingivalis MK production ability, as well as compete for MK with P. gingivalis, which in turn inhibits P. gingivalis growth (Hojo et al. 2007). Therefore, the decreased expression of P. gingivalis MK biosynthesis gene clusters in OB:CM might not be simply due to a cross-feeding reaction from T. denticola, but could be governed by the hemin level in OB:CM or T. denticola produced inhibitory molecules. However, these hypothesis will require further experimental testing.

132

Figure 5.7: P. gingivalis W83 ubiquinone and other terpenoid-quinone biosynthesis pathways. All gene identifiers (purple bordered box) of P. gingivalis W83. The gene identifiers (in green filled box) indicate significant downregulation (>1.5-fold change, FRD <0.05), while gene identifier (light blue filled box) indicate >1.4-fold decrease in gene expression, but did not fulfil the 1.5-fold cut-off set, respectively during P. gingivalis growth in OB:CM, when compared with OBGM.

133

5.2.2.2 P. gingivalis proteases In this RNAseq analysis, the expression of all sixty-seven genes encoding putative proteases listed in Chapter 4 were detected during P. gingivalis growth in OBGM and OB:CM, with the exception of PG1840, a neutral zinc metallopeptidase. P. gingivalis proteinases and peptidases that displayed the highest average expression (AveExpr = log2 Average CPM) in OBGM and OB:CM are shown in Table 5.4. As expected, gingipains (RgpA, RgpB and Kgp) showed the highest expression in OBGM and OB:CM (Table 5.4), which was at least 2-fold higher than the average expression (AveExpr = 6.56) of all other transcripts detected, exemplifying the importance of these proteases for P. gingivalis survival and growth. In addition, P. gingivalis putative ATP-dependent proteases, which include the FtsH protease (PG0047), Lon protease (PG0620) and Clp proteases subunits (PG0418 and PG0010) also showed relatively high expression levels (Table 5.4). Although the functional properties of these P. gingivalis ATP-dependent proteases have not been characterised, homologues of these annotated proteases are known to be essential for the regulation of cellular activities i.e. degradation of misfolded proteins, especially during stress-induced conditions, recycling of amino acids, fine-tuning of regulatory proteins levels and cell cycling activities (Tsilibaris et al. 2006). For example, the Clp protease complex is part of an adaptive mechanism to modulate cellular protein expression under different conditions i.e. host epithelial cells invasion and environmental stress conditions (Capestany et al. 2008, Tsilibaris et al. 2006, Olinares et al. 2011). Other highly expressed proteases under these culturing conditions include PrtT and PG1788 (Table 5.4). Numerous P. gingivalis proteases showed significant downregulation in OB:CM, relative to OBGM (Table 5.3). PrtC (PG1542) that has been shown to hydrolyse Type I collagenase (Takahashi et al. 1991, Kato et al. 1992) was downregulated 2.3-fold in OB:CM. The significant reduction of P. gingivalis ATPase protease (PG0047, PG0620 and PG0010) expression might be inherently regulated by P. gingivalis in response to environmental factors. PG0889, annotated as an aminopeptidase P family protein in the NCBI also showed significant reduction in expression during growth in OB:CM. Other P. gingivalis proteases that showed >1.3-fold decrease in expression in OB:CM, relative to OBGM, included PG1654, PorU, PG1701, PG0445 and Kgp (data not shown). However, none of the P. gingivalis protease genes showed significant upregulation during growth in OB:CM, compared to OBGM. The genes encoding PrtT and PG1788 showed

134

the highest increase in expression levels, which only achieved an approximately 1.4- and 1.3- fold increase in gene expression respectively (Table 5.4), which is deemed to be below the significance threshold. An increase in PG1788 and prtT transcripts that were already high, suggests a prominent role for these peptidases during P. gingivalis growth in OB:CM. This observation is in accord with the experimental data determined in this study (Chapter 4), where the PG1788 null strain showed retarded growth in OB:CM, as well as a decrease in free glycine production in OB:CM. The gene expression levels and differential expression of other targeted peptidases (PG0445, PG0753 and PG1605) of P. gingivalis used in this study (Chapter 4) were also examined. This RNAseq analysis showed that the average expression of PG0753 was relatively high, with an average expression level of 10.5 (data not shown). Although the transcription of PG0753 was upregulated during T. denticola and P. gingivalis coculture in a continuous system (Tan et al. 2014), P. gingivalis did not show increased expression of PG0753 in TdCM. The average expression levels of PG1605 and PG0445 in both media were relatively low (AveExpr ~5 – 5.5) and did not show DGE in OB:CM, compared to OBGM. Taken together, P. gingivalis gene expression profiles of putative proteases and peptidases were not associated with an increase in P. gingivalis free glycine production ability, which disapproves the hypothesis that TdSF in TdCM positively regulate a subset of P. gingivalis peptidases that contribute to the hydrolysis of glycine-containing peptides into free glycine. Proteases non-relevant to glycine release showed significant downregulation during P. gingivalis growth in OB:CM. The significant decrease in gene expression of P. gingivalis PrtC and ATPase-dependent proteases suggested complementary and cooperative activities of T. denticola proteases in the digestion of peptides in the medium, as well as external factors in TdCM that might regulate their expression.

135

Table 5.4: P. gingivalis protease-encoding genes that demonstrated the highest average gene expression levels in OBGM and OB:CM. Locus tag Gene Product Fold FDR AveExpr Avg CM Avg OB change CPM CPM PG_RS08940 PG2024 RgpA -1.039 0.823 14.982 32698.424 32078.123 PG_RS08105 PG1844 Kgp -1.307 0.099 14.318 18378.510 22815.452 PG_RS02240 PG0506 RgpB 1.096 0.571 13.518 12668.817 10926.819 PG_RS07885 PG1788 aminopeptidase 1.291 0.126 11.758 4069.142 2967.081 PG_RS02400 PG0537 aminoacyl-histidine dipeptidase 1.132 0.409 11.404 2962.279 2509.416 PG_RS00220 PG0047 cell division protein FtsH -2.050 0.014 11.188 1711.781 3254.297 PG_RS01860 PG0418 ATP-dependent Clp protease -1.008 0.968 11.010 2132.997 2028.864 proteolytic subunit PG_RS06835 PG1548 PrtT 1.353 0.069 10.826 2188.135 1513.170 PG_RS00055 PG0010 ATP-dependent Clp protease -1.734 0.035 10.823 1446.972 2320.642 ClpC PG_RS02735 PG0620 endopeptidase La -2.020 0.009 10.677 1202.229 2258.486

136

5.2.2.3 Transcription regulators and non-regulatory DNA-binding proteins DNA binding proteins contain DNA binding domains that interact with DNA, either specifically via binding to targeted DNA sequences or non-specifically with a general affinity (Siggers and Gordân 2014). DNA binding proteins include transcription factors (TFs), polymerases, nucleases and recombinases. Of these, proteins that bind non-specifically to DNA may be grouped into the non-regulatory DNA binding proteins. As chromosomal DNA is a dynamic structure that is generally compact, non-regulatory DNA binding proteins enable diverse DNA architectural processes, such as DNA replication, bending and compaction. TFs are known as DNA regulatory proteins that may work alone or in complex with other proteins to regulate the transcription of the adjacent genes by binding to specific DNA promoter sequences, resulting in the activation or repression of regulon expressions (Ortet et al. 2012). Since these specific DNA sequence motif may be found throughout the genomes, changes in the gene expression of TFs can in turn result in a simultaneous change in the expression of numerous genes. Furthermore, TFs can act in a coordinated manner to control a wide array of biological processes, such as cell death, cell division and cell stress response (Loewen et al. 2004, Perez and Groisman 2009, Yesudhas et al. 2017). In the Predicted Prokaryotic Transcription Factors database (P2TF), transcription factors are further categorised into one-component system (OCS), DNA-binding response regulator (RR), sigma factors (SF) and transcriptional regulators (TR; Fig. 5.8, Ortet et al. 2012). In P. gingivalis W83, there are 29 predicted helix-turn-helix (HTH) domain-containing transcription regulators, which are sub-divided into different collections of transcriptional regulator families i.e. ArgR, AsnC, Fur, GntR, LuxR, MerR, Mga, NrdR, PadR, Rrf2, HxlR, MarR, unclassified members, AraC, TetR and Xre (Fig. 5.8, Ortet et al. 2012). In addition, P. gingivalis is predicted to possess 14 other non-regulatory DNA-binding proteins (ODP) in the P2TF database, which consisted of Bhl, DnaA, Fis and unclassified members (Fig. 5.8, Ortet et al. 2012). In this RNAseq analysis, three predicted transcriptional regulators and three non- regulatory DNA-binding protein-encoding genes showed differential expression during P. gingivalis growth in OB:CM, compared with OBGM (Table 5.3, Fig. 5.8). The transcripts of predicted TFs and non-regulatory DNA-binding protein in P2TF that remain undetected in this study include genes encoding a putative DNA-binding protein PG0546, histidine kinase PG1432 and transcriptional regulator PG2125.

137

Amongst the predicted non-regulatory DNA-binding protein-encoding genes that showed DGE were PG2047, PG0566 and PG1276 (Fig. 5.8). PG2047 encodes a putative helicase that is made up of a combinations of domains, such as P-loop_NTPase, TraI and RecD superfamilies, and a HTH domain at its C-terminus, was downregulated (Marchler-Bauer et al. 2015, Marchler-Bauer et al. 2017). Decreased gene expression of a helicase might suggest lower DNA replication rates as helicase is responsible for DNA unwinding, which could be a factor contributing to the longer generation time of P. gingivalis in OB:CM . PG0566 and PG1276 that encode histone-like family DNA-binding proteins (or annotated as histidinol- phosphate aminotransferase in the NCBI) were significantly upregulated during P. gingivalis growth in OB:CM. These histone-like DNA-binding proteins (HU) have not been characterised; nevertheless, they are predicted to have an integration host factor and be a member of the non-specific domain HU_IHF that might have a role in affecting DNA conformation for the regulation of various cellular processes (Marchler-Bauer et al. 2017). However, both of these DNA-binding proteins that consisted of 168 amino acids exhibited distinct domain architecture from the HU that functions as a major non-sequence-specific DNA-binding domain protein that is associated with nucleoid (Priyadarshini et al. 2013). Three transcription regulators determined in P2TF showed DGE, which belong to the Xre and AsnC family. The transcriptional expression of two members of the xenobiotic resistance transcription regulators (Xre family), PG0860 and PG1063, showed significant upregulation during P. gingivalis growth in OB:CM, relative to OBGM (Table 5.3, Fig. 5.8). As both P. gingivalis Xre transcriptional regulators have not been characterised, it is difficult to elucidate their roles in this context. Another transcriptional regulator that showed significant upregulation during P. gingivalis growth in OB:CM was PG1127. PG1127 is the only member of the putative AsnC transcriptional regulator in P. gingivalis genome. While PG1127 has not been addressed specifically in the literature, its regulatory properties could presumably be inferred from the other characterised members of AsnC that modulate cellular metabolism, especially amino acid metabolism (Yokoyama et al. 2006, Knoten et al. 2011, Ji et al. 2015). Out of six extracytoplasmic function (ECF) sigma factors identified in the P. gingivalis W83 genome (Nelson et al. 2003), PG0985 that is annotated as a sigma-70 family RNA polymerase sigma factor encoding gene showed a 1.45-fold increase in expression (Fig. 5.8). ECF sigma factors are important in sensing and responding to signals in the extracellular environment, which in turn mediate P. gingivalis gene expression. PG0985 is predicted to be

138

cotranscribed with two hypothetical protein encoding genes, PG0986 and PG0987. PG0987 possesses DUF4252 and MutS domains, whilst the PG0986 showed no conserved domain. Although other ECF sigma factors (PG1660 and PG0162) of P. gingivalis demonstrated recognition to their own promoter for autoregulation of their own gene expression (Dou et al. 2016, Dou et al. 2018), PG0985 activities might be controlled by its cognate anti-sigma factor from downstream protein encoding genes as with most cases of ECF sigma factor regulation (De Las Peñas et al. 1997, Asai 2018, Kwak et al. 2018). Inactivation of PG0985 homologue in P. gingivalis ATCC 33277 strain (PGN_0970) showed no difference in the phenotype assays screened, which included biofilm formation, autoaggregation, hemagglutination, gingipain activities and OMV production, when compared with wild type (Onozawa et al. 2015, Kazutaka et al.). Likewise, P. gingivalis cells subjected to environmental stress factors i.e. hydrogen peroxide and nitric oxide did not show upregulation of this ECF sigma factor (Boutrin et al. 2012). However, PG0985 was found to be downregulated under microaerophilic conditions, and showed increased expression in a LuxS-deficient strain and PG1660PG0162- double mutant (Yuan et al. 2005, Dou et al. 2018). This evidence suggested that PG0985 might act as a redundant regulatory system that becomes upregulated when the canonical signalling pathways become disrupted or other ECF sigma factors become inactivated. However, PG0985 could also be upregulated in response to T. denticola produced molecules in TdCM, which in turn lead to specialized promoter target recognition for the mRNA synthesis. Although not identified as a TF in P2TF, PG1044 that encodes PgMntR showed a 1.86- fold decrease in expression in OB:CM, compared to OBGM (Table 5.3). PgMntR acts as a transcriptional metalloregulator for the expression of the only manganese transporter (PG1043, FeoB2) encoded in the P. gingivalis genome (Zhang et al. 2016). PgMntR has been demonstrated to bind with manganese and ferrous metal ions; while the PgMntR in complex with Mn2+ resulted in an increase of binding affinity to its own promoter, which caused transcriptional repression; Fe2+ binding to the regulator lead to depression of the transcription (Zhang et al. 2016). Downregulation of pgmntR suggested a high content of Mn2+ or other metal divalent cations intracellularly (Dashper et al. 2005, He et al. 2006). In this regard, the DGE of pgmntR was also associated with a significant downregulation of manganese transporter transcript PG1043, which is cotranscribed with pgmntR. This change in expression can in turn regulate the protein levels of manganese transporter and prevent further internalization of metal ions.

139

Differential expression of these transcriptional factors during P. gingivalis growth in OB:CM indicate an extensive change of the P. gingivalis regulatory network and gene expression landscape. These regulatory networks can be comprehensive and complex to coordinate expression of specialized genes that can be essential for P. gingivalis survival and adaptation in an environment that is enriched with T. denticola produced effectors. Some of these regulatory components might be relevant to the modulation of phenotypes for P. gingivalis heterotypic community development with T. denticola, which warrant further experimental characterisation.

Figure 5.8: DNA binding protein-encoding genes distribution around the sequenced genome of P. gingivalis W83. These are subcategorised as transcription regulator (TR), other DNA binding protein (ODP), one component system (OCS), sigma factor (SF), response regulator (RR). Red open arrows indicate significantly upregulated genes and the blue open arrow represents a significantly downregulated gene during P. gingivalis growth in OB:CM, compared with OBGM. Extracted and modified from P2TF database (Ortet et al. 2012).

140

5.2.2.4 Transporter and efflux systems Transporters are membrane protein components that allow the import and export of substrates to and from the cytoplasm of cells. While importers are commonly recognized for the internalization of metabolic substrates and precursors, efflux systems can be the defence mechanisms employed by bacteria to ensure their survival in the presence of toxic metabolites or antimicrobial substances. These efflux pumps are also involved in the maintenance of cellular homeostasis, by regulating the process of intracellular metabolite detoxification and preventing the accumulation of a broad range of toxic compounds from the environment, as well as allowing the export of bacterial virulence factors and intercellular signal trafficking (Kuroda and Tsuchiya 2009, Martinez et al. 2009, Beis 2015, Locher 2016). Currently, the structures of five families of bacterial multidrug resistance (MDR) pumps have been solved, comprising the ATP-binding cassette (ABC), major facilitator superfamily (MFS), small multidrug resistance (SMR), multidrug and toxic compound extrusion (MATE) and resistance nodulation division (RND) families (Fig. 5.9; Blanco et al. 2016, Du et al. 2015). While MATE, ABC, MFS and SMR efflux superfamily members consist of an inner membrane protein complex for the removal of toxic compounds from the cytoplasm, the RND superfamily forms a tripartite or tetrapartite complex for the direct export of toxic metabolites from the periplasm or cytoplasm into the extracellular milieu (Fig. 5.9). ABC transporters utilise ATP as the energy source to drive the conformational change of the transmembrane domains to allow the import of substrates or export of toxic metabolites from the cytoplasm (Locher 2016). ABC transporters have been reviewed extensively and show high structural and mechanistic diversity that can be classified into Type I, Type II and Type III ABC importers, and B-family efflux (Davidson et al. 2008, Beis 2015, Wilkens 2015, Locher 2016). There were eight genes encoding putative ABC superfamily transporter components of P. gingivalis W83 that showed differential expression during growth in OB:CM. Five genes were downregulated and three genes were upregulated (Table 5.3). PG0682, PG0683 and PG0684, which were some of the most highly downregulated genes with 6.02-, 4.45- and 4.53-fold decreases in expression under these conditions, are currently annotated as ABC transporter permeases in the NCBI and predicted to be expressed in an operon (Table 5.3). PG0685 that encodes an ATP transporter ATP-binding protein and is predicted to transcribe as a monocitronic gene in the same orientation as the adjacent PG0682- 4 operon also showed a 2.93-fold decrease in its expression (Table 5.3). PG0682-4 are

141

predicted to function as a multi-subunit ABC transporter that allows lipoprotein translocation across the membrane as they showed sequence similarities to the FtsX-like superfamily protein that utilises ATP to transport lipid targets from the inner to outer membrane (Marchler-Bauer et al. 2017, Li et al. 2012). Interestingly, PG1176 encoding a putative lipid A binding component in the ABC subfamily also showed a significant decrease in expression, suggesting that P. gingivalis might show reduced lipid-based substrate transport. Genes encoding the putative components/subunits of the ABC transporter complex, which are normally associated with the permease of ABC importers, that showed significant upregulation of gene expression in OB:CM, compared with OBGM, are listed below. For example, PG1379 that encodes a ABC transporter substrate-binding protein, which is likely to be transcribed as part of an operon consisting of PG1379-82, was significantly upregulated to a similar extent as PG1382 that encodes a porin (Table 5.3). Based on the NCBI Conserved Domain Database (CDD) search, PG1379 is predicted to resemble the architecture of a type II periplasmic binding fold superfamily, which is commonly found in solute binding domain- containing proteins that are coupled to an ABC importer. Moreover, this type of solute binding protein could serve as a receptor in the transport, signal transduction and channel gating (Marchler-Bauer et al. 2015, Marchler-Bauer et al. 2017). Other potential genes encoding ABC transporter components that showed DGE, include PG2190 and PG1025 that encode a putative phosphonate ABC transporter ATP-binding protein and a DUF2813 domain-containing binding proteins of the ABC superfamily, respectively (Elbourne et al. 2017). As Type I and II ABC importers generally function by coupling to a binding protein (Wilkens 2015, Locher 2016), the upregulation of gene expression of these ABC transporter components might correspond to an increase in substrate binding and internalization of the preferred molecules, rather than the export of molecules. Likwise, metatranscriptomic profile of P. gingivalis also demonstrated a significant upregulation of other ABC transporters in vivo, comparative to laboratory culturing conditions, suggesting nutrient uptake competition between members of polymicrobial (Deng et al. 2018). In addition to the DGE of ABC transporter components, PG2082 and PG2090 that encode proteins classified as a MFS transporter and cation diffusion facilitator family transporter respectively were significantly upregulated (Table 5.3). In addition, PG1640 that encodes a protein with predicted MATE_DinF_like subfamily domain (Marchler-Bauer et al. 2017) was downregulated by 2-fold during P. gingivalis growth in OB:CM. Likewise, PG1641

142

and PG1642 that encode a low molecular weight phosphotyrosine phosphatase and copper- transporting P-type ATPase respectively, as well as predicted to express in the same operon as PG1640 were downregulated greater than 2-fold. MATE transporter is a generic secondary transport system that employs Na+/H+ ion gradient for the efflux of antimicrobial agents (Kuroda and Tsuchiya 2009), instead of using ATP hydrolysis. However, the predicted cotranscription of PG1640 with genes encoding ATPase and phosphatase proteins that are required for the catalysis of ATP hydrolysis and protein dephosphorylation activities respectively, suggested that this transporter might have alternative functional properties that are yet to be characterised. On the other hand, the other five transcripts encoding putative members of MATE multidrug transporters, which include PG0006, PG0636, PG0827, PG1117 and PG1446, showed no change in expression under the two different conditions. The RND family is also known as a tripartite efflux complex as it is generally composed of an outer membrane channel (TolC), efflux transporter periplasmic adaptor belonging to the membrane fusion protein family and RND pump that span the inner membrane (Fig. 5.10). These components are normally expressed in an operon under the control of the same operational transcription unit to ensure stoichiometric balance of these products for their assembly into a typical tripartite complex (Xu et al. 2011, Anes et al. 2015). PG0063-5 that encode CusA/CzcA family heavy metal efflux (HME)-RND showed a greater than 2-fold increase in expression. PG0063 and PG0065 corresponded to the TolC and periplasmic adaptor proteins respectively; whereas PG0064 is annotated as a Cus/CzcA RND pump, by which the CusA system is more specialized in the binding and removal of monovalent ions, while the CzcA system is responsible for the efflux of divalent and heavy metal ions, such as cobalt, zinc, cadmium (Long et al. 2012, Moraleda-Munoz et al. 2010). Thus, one would speculate that the upregulation of transcription of this system is predicted to prevent the accumulation of monovalent or divalent metals intracellularly as high levels of metal can lead to cell toxicity. Furthermore, PG0539 encoding a periplasmic adaptor component of a putative hydrophobic and amphiphilic (HAE)-RND transporter subfamily of P. gingivalis was also differentially regulated. Although this gene is predicted to be transcribed with PG0538 and PG0540 that encode TolC and an efflux pump respectively, only this single transcript demonstrated significant downregulation (Table 5.3). A previous study on PG0538, PG0539, PG0540, which was then named as XepC, XepA and XepB respectively, showed that they are involved in the export of xenobiotics (Ikeda and Yoshimura 2002). Mutation of xepB resulted

143

in an increased susceptibility of P. gingivalis to the antimicrobial agents tested, which included ampicillin, tetracycline, rifampicin, norfloxacin, minocycline, ciprofloxacin, acriflavine and berberine, some of which are used as antibiotics treatment for periodontitis (Ikeda and Yoshimura 2002, Inoue et al. 2015). It should be noted that many efflux systems are regulated by transcriptional regulators and are intertwined with complex cellular regulatory processes (Hwang et al. 2012, Grkovic et al. 2001, Issa et al. 2018). Differential expression of transcriptional regulators i.e. Xre associated with multidrug resistance as described above are postulated to be concomitantly related to the change in expression of these efflux systems and thus might affect the multidrug resistance profile of P. gingivalis growth in OB:CM. Consistent with previous observations that TdCM exhibits growth suppressive effects towards P. gingivalis, upregulation of these putative efflux transporters might be the essential determinants for P. gingivalis survival and proliferation under a condition that contained T. denticola produced effector molecules and metabolites. Therefore, it is important to investigate the roles of these efflux transporters, especially during P. gingivalis association with T. denticola.

Figure 5.9: Schematic representation of efflux transporters in Gram negative bacteria. ABC, MF, MATE and SMR superfamily efflux transporters only spanned the inner membrane (IM), while the RND transporter that assembled into a tripartite complex are composed of the efflux pump in the inner membrane, membrane fusion protein and TolC outer membrane protein for the extrusion of molecules into the extracellular milieu. This diagram is extracted from (Blanco et al. 2016).

144

5.2.2.5 Hypothetical protein encoding genes Hypothetical protein encoding genes constituted approximately 22% of the DGE profile of P. gingivalis during growth in OB:CM, in comparison to OBGM. Of the 20 transcripts determined using the common gene identifier, six showed upregulation and 14 genes were downregulated (Table 5.3, Table 5.5). The hypothetical protein encoding genes with commonly gene identifier were predicted to comprise of seven monocistronic genes and 13 belong to the polycistronic genes (Table 5.5). Furthermore, all nine transcripts that were annotated as hypothetical protein encoding genes from the new locus tag in the NCBI were downregulated (Table 5.3). These hypothetical protein sequences were then searched via the CDD in the NCBI, which found that none of the new locus tag annotated ORF showed conserved domains, except for PG_RS10535 that consists of 48 amino acids and is predicted to encode a DUF1661 superfamily. This domain with unknown function was also annotated in 19 other P. gingivalis hypothetical proteins, which included PG0080, PG0174, PG0524, PG1722, PG1980, PG2064 and PG2212 that were speculated to have DNA-binding properties due to the presence of a zinc finger motif in their sequences (Dou et al. 2014). Other putative DNA-binding proteins of unknown function that showed decreased expression were PG0565 and PG0833. HB1, ASXL, restriction endonuclease (HARE)-HTH superfamily was detected at the N-terminus of PG0565, while PG0833 that is predicted to be in an operon with a nucleoid associated protein, PG0832 consisted of a P-loop NTPase domain superfamily, SMC superfamily and DUF3732 domain- containing superfamily. Certain hypothetical protein encoding genes that were differentially regulated contained domains that were predicted to be components of membrane protein complexes. Examples of putative membrane protein component encoding genes that showed upregulation were PG1493 that encodes a TonB-dependent receptor Plug domain and PG0621 that encodes a protein of 182 amino acids with two DUF1599 superfamily domains in the protein sequence (Table 5.5). PG0621 is predicted to be transcribed in the PG0620-0625 operon, along with other putative membrane-associated proteins i.e. the DoxX family protein and SPOR domain- containing protein (Yahashiri et al. 2017). Similarly, PG0574 that showed no conserved domains, is located as the second ORF of an operon that is predicted to consist of a total of 13 ORFs. This operon primarily encodes Mur enzymes that are essential for bacterial cell wall biosynthesis and Fts cell division proteins. PG0717-0718 that encodes multiple PepSY-like

145

superfamily domains were downregulated; this PepSY-like domain containing proteins showed similarity to the beta-lactamase inhibitor of PepST protein (Marchler-Bauer et al. 2015, Marchler-Bauer et al. 2017). These hypothetical proteins are likely to be associated with the expression of P. gingivalis phenotypes, which include putative surface structures, DNA-binding proteins and regulation of cell division, suggesting that these proteins might play a role in mediating P. gingivalis and T. denticola interactions. Therefore, characterising these hypothetical proteins in future studies could help provide a better understanding of the molecular and mechanistic basis of P. gingivalis in polymicrobial interactions.

146

Table 5.5: Summary table of hypothetical protein encoding genes that showed DGE during P. gingivalis growth in OB:CM, relative to OBGM.

New locus tag Gene Gene/ Start End (bp)a Stranda Length Predicted domainb Fold chg FDR Opa (bp)a (bp)a PG_RS00360 PG0081 3 93945 96122 + 725 Bacterial Ig-like domain, Carboxypeptidase-regulatory like -1.919 0.011 domain PG_RS00670 PG0144 2 166733 167758 - 341 Agmatine deiminase family protein -2.110 0.019 PG_RS01930 PG0434 1 471524 472603 - 359 hypothetical protein -2.124 0.019 PG_RS02510 PG0565 1 621861 622733 - 290 HARE-HTH superfamily -1.849 0.045 PG_RS02540 PG0574 13 629158 629631 + 157 hypothetical protein 1.551 0.012 PG_RS02740 PG0621 6 675532 676080 + 182 DUF1599 superfamily -1.977 0.010 PG_RS03020 PG0686 1 741221 742774 - 517 DUF438 superfamily, DUF1858 superfamily -2.172 0.034 PG_RS03145 PG0717 2 770081 771184 - 367 putative lipoprotein, PepSY-like superfamily 4.622 0.000 PG_RS03150 PG0718 2 771226 771663 - 145 PepSY-like superfamily 5.404 0.000 PG_RS03655 PG0833 2 889798 891663 - 621 DUF3732 domain-containing protein -1.954 0.036 PG_RS03735 PG0849 4 910562 911332 - 256 hypothetical protein 1.625 0.022 PG_RS04785 PG1085 1 1153343 1153591 + 82 hypothetical protein -1.960 0.037 PG_RS05010 PG1130 1 1208193 1210190 + 665 TPR domain-containing protein -1.860 0.045 PG_RS05170 PG1165 2 1247601 1248317 - 238 hypothetical protein 1.645 0.034 PG_RS06095 PG1385 2 1465223 1466425 - 400 TPR domain-containing protein 1.565 0.011 PG_RS06575 PG1493 2 1566923 1569598 + 891 TonB-dependent Receptor Plug domain -1.240 0.015 PG_RS06855 PG1554 4 1635404 1636081 + 225 hypothetical protein -1.918 0.036 PG_RS06865 PG1556 2 1636735 1637061 + 108 DUF2149 domain containing protein -1.902 0.007 PG_RS08715 PG1974 1 2062054 2063145 - 363 hypothetical protein -1.591 0.034 PG_RS08850 PG2006 1 2094727 2095086 + 119 hypothetical protein -2.895 0.009 a: sourced from DOOR Database of Prokaryotic Operons b: determined from NCBI Conserved domain database (CDD)

147

1 5.3 Conclusion 2 RNAseq is a powerful approach in providing insights into P. gingivalis’s 3 physiological changes during growth in TdCM. These data strengthen evidence of P. gingivalis 4 and T. denticola interspecies interactions, which P. gingivalis demonstrated DGE of a number 5 of signal transduction components, as well as putative resistance mechanisms against T. 6 denticola produced effector molecules during growth in OB:CM. Although this study aimed to 7 shed light on the potential molecular and mechanistic synergism between P. gingivalis and T. 8 denticola, this approach has also revealed competitive mechanisms between P. gingivalis and 9 T. denticola. This investigation has provided a hypothetical framework to formulate relevant 10 experiments to answer more questions on their multifaceted interactions, by which T. denticola 11 produced diffusible effector molecules might be affecting P. gingivalis regulatory activities, as 12 well as contributing to its phenotypic change. However, it should be borne in mind that the 13 usage of TdCM mixed with OBGM for P. gingivalis growth would resemble conditions where 14 T. denticola is present in high abundance and has pre-modified the environment with its 15 intermediates, metabolic end products and secreted proteins, rather than demonstrating gene 16 expression of P. gingivalis during association with T. denticola. In addition, RT-PCR and 17 experimental assessment of P. gingivalis phenotypes should be performed to further validate 18 these data. 19 20 21 22 23 24 25 26 27 28 29

148

1 Chapter 6 Conclusion

2 Chronic periodontitis is an inflammatory disease associated with a shift of the 3 polymicrobial biofilm community from a healthy to diseased state. The compositional and 4 abundance change of the oral microbiota also demonstrated an overall alteration in their 5 functional properties during proliferation under dysbiotic conditions, as determined via 6 metatranscriptomics (Duran-Pinedo et al. 2014, Szafrański et al. 2015, Deng et al. 2017). The 7 red complex species, which encompass P. gingivalis, T. denticola and T. forsythia have been 8 found in high association with the progression and severity of chronic periodontitis (Socransky 9 and Haffajee 1992, Byrne et al. 2009). Previous studies have shown that P. gingivalis and T. 10 denticola exhibited physical association and metabolic interactions which include cross- 11 feeding of fatty acids, thiamine and glycine (Grenier 1992, Tan et al. 2014). A mouse model 12 study demonstrated that coinoculation of P. gingivalis and T. denticola enhanced virulence 13 (Orth et al. 2011), T. denticola and P. gingivalis dual-species biofilms showed significantly 14 higher biomass and thickness than their monospecies counterparts (Zhu et al. 2013). These 15 studies displayed synergistic interactions between P. gingivalis and T. denticola under both in 16 vitro and in vivo conditions. 17 A recent study demonstrated that P. gingivalis increased free glycine production in the 18 presence of T. denticola (Tan et al. 2014), thus it was of interest to investigate the molecular 19 basis of this metabolic interaction. Firstly, P. gingivalis was grown in OBGM mixed with T. 20 denticola conditioned medium (TdCM) that had been shown to stimulate the P. gingivalis 21 release of free glycine (Tan et al. 2014). This study showed that TdCM had a negative effect 22 on P. gingivalis growth, which resulted in reduced cell density of P. gingivalis lag phase, 23 suggesting the presence of T. denticola produced growth suppressing factors. In an attempt to 24 identify TdSF, TdCM was simplified and fractionated with RP-HPLC. However, each RP- 25 HPLC fraction (F2, F3, F4) of a simplified TdCM could stimulate P. gingivalis to generate a 26 relatively high amount of free glycine. This observation led to the interpretation that TdSF were 27 composed of non-sequence specific peptides. The non-sequence specific peptides in TdCM 28 fraction could be generated from proteins and oligopeptides in OBGM that had been digested 29 with T. denticola proteinases and peptidases, followed by further digestion of these peptides 30 by P. gingivalis peptidases to increase free glycine release. The possible requirement for T. 31 denticola protease and peptidase digestion to make suitable substrates for P. gingivalis further

149

1 digestion and release of free glycine shows a complementarity in their protease activities. 2 Another potential cooperative interaction between T. denticola and P. gingivalis that is 3 depicted in this study was the enhanced utilisation of glutathione (GSH) by P. gingivalis during 4 growth in TdCM and >10 kDa size fraction of TdCM that had been mixed with OBGM. Whilst 5 GSH was completedly degraded by T. denticola, P. gingivalis that lacks the enzymes is unable 6 to degrade GSH in OBGM. The release of T. denticola GSH utilisation enzymes, γ-glutamyl- 7 transpeptidase and cysteinylglycinase, in OMVs into TdCM (Chu et al. 2003, Chu et al. 2008, 8 Veith et al. 2009), aid the degradation of GSH into glutamate, glycine and cysteine for P. 9 gingivalis. This study reinforced the importance of metabolic cooperativity between bacterial 10 species in maximising nutrient utilisation efficiency in the oral polymicrobial community. 11 In addition to the pursuit of TdSF determination, the P. gingivalis proteases involved in 12 the degradation of glycine-containing peptides into free glycine were also investigated. Given 13 the extensive array of putative proteases encoded in the P. gingivalis genome, bioinformatic 14 analyses were performed. As a result, PG1605 and PG1788 were selected as targets for further 15 study, because both were predicted to have an N-terminal signal peptide, be located outside of 16 the cytoplasm, and demonstrated homology to PepC, a cysteine aminopeptidase from L. lactis 17 capable of polyglycine cleavage (Mistou and Gripon 1998). In addition, PG0753, which 18 showed upregulation during P. gingivalis coculture with T. denticola (Tan et al. 2014), and 19 PG0445, which is homologous to peptidases that showed a preference for glycine cleavage at 20 the P1 position, were also selected for further study. The genes encoding these four putative 21 proteases were each inactivated on the P. gingivalis genome, then these mutants and wild type 22 were grown in OBGM and OB:CM to compare their glycine releasing abilities. All P. 23 gingivalis strains showed consistent lag phase during growth in OB:CM, however, the growth 24 suppressing effect against PG1788- strain was more predominent. This observation suggested 25 that PG1788 might be required as part of an immunity mechanism to alleviate T. denticola 26 growth suppressing factors against P. gingivalis. While PG0445- and PG1605- were shown to 27 be irrelevant for P. gingivalis glycine production in OB:CM, PG1605- and PG0753- strains 28 showed a reduction in free glycine release in OB:CM in comparison to wild type, indicating 29 that they played a role in the release of free glycine. 30 Differential gene expression of P. gingivalis during growth in OBGM and OB:CM was 31 investigated to gain a better understanding of the effect of T. denticola on the P. gingivalis 32 transcriptome. A total of 132 genes were differentially regulated, which were classified into

150

1 different functional properties. Although it was previously hypothesized that P. gingivalis 2 increased glycine production by up-regulating proteolytic digestion of glycine-containing 3 peptides in the medium, this transcriptomic analysis did not show a differential change of gene 4 expression of any peptidases that might correspond to glycine release. Instead, there was an 5 increase in transcription of the genes encoding the threonine to glycine one-step conversion 6 pathway. Whilst genes encoding P. gingivalis glycine cleavage systems displayed significant 7 downregulation, there was also a downregulation of genes encoding serine utilisation pathways 8 other than the oxidation of serine to glycine pathway. The transcriptomic data also supported 9 early studies of T. denticola succinic acid cross-feeding to P. gingivalis by showing 10 upregulation of genes encoding the P. gingivalis succinic acid utilisation pathway. Moreover, 11 a number of transcriptional factors and signal transduction components were differentially 12 expressed, suggesting that P. gingivalis could be sensing and responding to T. denticola 13 effector molecules in TdCM. Change in transcription levels of these regulators might affect the 14 expression levels of the neighbouring operon and other regulons. This study also demonstrated 15 for the first time that growth in TdCM caused P. gingivalis to upregulate genes encoding an 16 efflux transporter system. Given that P. gingivalis consistently demonstrated a reduced cell 17 density during lag phase period of growth in OB:CM, the efflux transporter system could be 18 required for the extrusion of T. denticola produced toxic metabolites, and represent a P. 19 gingivalis survival mechanism. However, experimental validation is required to verify 20 phenotypic changes of P. gingivalis in the presence of T. denticola. In conclusion, this study 21 demonstrated P. gingivalis global gene expression in response to T. denticola modified 22 conditions, which in turn provided insights into their potential molecular and mechanistic 23 interactions. These findings also refined current understanding on P. gingivalis and T. denticola 24 interactions and possibly help develop new strategies in disrupting their multimodal 25 interactions.

151

1 Bibliography 2 3 4 Abe, N., Baba, A., Takii, R., Nakayama, K., Kamaguchi, A., Shibata, Y., Abiko, Y., 5 Okamoto, K., Kadowaki, T. and Yamamoto, K. (2004) 'Roles of Arg- and Lys- 6 gingipains in coaggregation of Porphyromonas gingivalis: identification of its 7 responsible molecules in translation products of rgpA, kgp, and hagA genes', 8 Biological Chemistry, 385(11), 1041-1047. 9 10 Abiko, Y., Nagano, K., Yoshida, Y. and Yoshimura, F. (2014) 'Major Membrane Protein 11 TDE2508 Regulates Adhesive Potency in Treponema denticola', Plos One, 9(2). 12 13 Abusleme, L., Diaz, P. I., Silva, N., Dupuy, A. K., Strausbaugh, L. D., Dutzan, N., Gamonal, 14 J. and Burleson, J. A. (2013) 'The subgingival microbiome in health and periodontitis 15 and its relationship with community biomass and inflammation', Isme Journal, 7(5), 16 1016-1025. 17 18 Adelaide, Australian Research Centre for Population Oral Health. The University of 19 Adelaide. (2009) 'Periodontal diseases in the Australian adult population', Australian 20 Dental Association, 54(4), 390-393.

21 Alhasawi, A., Castonguay, Z., Appanna, N. D., Auger, C. and Appanna, V. D. (2015) 22 'Glycine metabolism and anti-oxidative defence mechanisms in Pseudomonas 23 fluorescens', Microbiological Research, 171, 26-31.

24 Alm E. J., Huang, K. H., Price, M. N., Koche, R.P., Keller, K., Dubchak, I. L. and Arkin, 25 A.P. (2004) 'The MicrobesOnline Web site for comparative genomics', Genome 26 Research, 15, 1015-1022. 27 28 Alvarez-Carreno, C., Becerra, A., Lazcano, A. and Alva, V. (2018) 'Structure, function and 29 evolution of the hemerythrin-like domain superfamily', Protein Science, 27(4), 848- 30 860. 31 32 Amano, A. (2003) 'Molecular interaction of Porphyromonas gingivalis with host cells: 33 Implication for the microbial pathogenesis of periodontal disease', Journal of 34 Periodontology, 74(1), 90-96. 35 36 Anes, J., McCusker, M. P., Fanning, S. and Martins, M. (2015) 'The Ins and Outs of RND 37 Efflux Pumps in Escherichia coli', Frontiers In Microbiology, 6(587). 38

152

1 Arakawa, S. and Kuramitsu, H. K. (1994) 'Cloning and Sequence-Analysis of a 2 Chymotrypsin-like Protease from Treponema denticola', Infection and Immunity, 3 62(8), 3424-3433. 4 5 Arkin, A. (2011) 'MicrobesOnline: an integrated portal for comparative and functional 6 genomics'. 7 8 Arnold, W. K., Savage, C. R., Antonicello, A. D. and Stevenson, B. 'Apparent Role for 9 Borrelia burgdorferi LuxS during Mammalian Infection', Infection and Immunity, 10 83(4), 1347-1353. 11 12 Asahi, Y., Noiri, Y., Igarashi, J., Asai, H., Suga, H. and Ebisu, S. (2010) 'Effects of N-acyl 13 homoserine lactone analogues on Porphyromonas gingivalis biofilm formation', 14 Journal of Periodontal Research, 45(2), 255-261. 15 16 Asahi, Y., Noiri, Y., Igarashi, J., Suga, H., Azakami, H. and Ebisu, S. (2012) 'Synergistic 17 effects of antibiotics and an N-acyl homoserine lactone analog on Porphyromonas 18 gingivalis biofilms', Journal of Applied Microbiology, 112(2), 404-411. 19 20 Asai, K. (2018) 'Anti-sigma factor-mediated cell surface stress responses in Bacillus subtilis', 21 Genes & Genetic Systems, 92(5), 223-234. 22 23 Asai, Y., Jinno, T., Igarashi, H., Ohyama, Y. and Ogawa, T. (2002) 'Detection and 24 quantification of oral treponemes in subgingival plaque by real-time PCR', Journal of 25 Clinical Microbiology, 40(9), 3334-3340. 26 27 Avraham, B. e., Zhang, H. P., Florin, E. L., Shelley, M. P., Eshel, B.-J. and Harry, L. S. 28 (2009) 'Deadly Competition between Sibling Bacterial Colonies', Proceedings of the 29 National Academy of Sciences of the United States of America, (2), 428. 30 31 Bachtiar, E. W., Bachtiar, B. M., Jarosz, L. M., Amir, L. R., Sunarto, H., Ganin, H., Meijler, 32 M. M. and Krom, B. P. (2014) 'Al-2 of Aggregatibacter actinomycetemcomitans 33 inhibits Candida albicans biofilm formation', Frontiers in Cellular and Infection 34 Microbiology, 4. 35 36 Bains, V. K. and Bains, R. (2015) 'The antioxidant master glutathione and periodontal health', 37 Dental Research Journal, 12(5), 389-405. 38 39 Bamford, C. V., Fenno, J. C., Jenkinson, H. F. and Dymock, D. (2007) 'The chymotrypsin- 40 like protease complex of Treponema denticola ATCC 35405 mediates fibrinogen 41 adherence and degradation', Infection and Immunity, 75(9), 4364-4372.

153

1 2 Banbula, A., Bugno, M., Goldstein, J., Yen, J., Nelson, D., Travis, J. and Potempa, J. (2000) 3 'Emerging family of proline-specific peptidases of Porphyromonas gingivalis: 4 Purification and characterisation of serine dipeptidyl peptidase, a structural and 5 functional homologue of mammalian prolyl dipeptidyl peptidase IV', Infection and 6 Immunity, 68(3), 1176-1182. 7 8 Banbula, A., Mak, P., Bugno, M., Silberring, J., Dubin, A., Nelson, D., Travis, J. and 9 Potempa, J. (1999) 'Prolyl tripeptidyl peptidase from Porphyromonas gingivalis - A 10 novel enzyme with possible pathological implications for the development of 11 periodontitis', Journal of Biological Chemistry, 274(14), 9246-9252. 12 13 Banbula, A., Yen, J., Oleksy, A., Mak, P., Bugno, M., Travis, J. and Potempa, J. (2001) 14 'Porphyromonas gingivalis DPP-7 represents a novel type of dipeptidylpeptidase', 15 Journal of Biological Chemistry, 276(9), 6299-6305. 16 17 Barnes, V. M., Ciancio, S. G., Shibly, O., Xu, T., Devizio, W., Trivedi, H. M., Guo, L. and 18 Jonsson, T. J. (2011) 'Metabolomics Reveals Elevated Macromolecular Degradation 19 in Periodontal Disease', Journal of Dental Research, 90(11), 1293-1297. 20 21 Bartlett, M. S. (1937) 'Properties of Sufficiency and Statistical Tests', Proceedings of the 22 Royal Society of London. Series A, Mathematical and Physical Sciences, (901), 268. 23 24 Barton, G., Cole, C., Procter, J., Drozdetskiy, A., Cole, C., Procter, J. and Barton, G. J. 25 (2015) JPred4: a protein secondary structure prediction server, Nucleid Acids 26 Research: Oxford Univ Press. 27 28 Barton, G., Martin, D., Procter, J., Martin, D., Barton, G., Waterhouse, A. M., Procter, J. B., 29 Martin, D. M. A., Barton, G. J. and Clamp, M. (2009) Jalview Version 2-a multiple 30 sequence alignment editor and analysis workbench. 31 32 Baughn, A. D. and Malamy, M. H. (2002) 'A mitochondrial-like aconitase in the bacterium 33 Bacteroides fragilis: implications for the evolution of the mitochondrial Krebs cycle', 34 Proceedings of the National Academy of Sciences of the United States, (7), 4662. 35 36 Beighton, D., Mughal, S. R. and Do, T. (2013) Bacterial Catabolism of Salivary Substrates, 37 Caster Academic Press, Norfolk, UK. 38 39 Beis, K. (2015) 'Structural basis for the mechanism of ABC transporters', Biochemical 40 Society Transactions, 43, 889-893. 41

154

1 Bendtsen, J. D., Kiemer, L., Fausbøll, A. and Brunak, S. (2005) 'Non-classical protein 2 secretion in bacteria', Bmc Microbiology, 5, 58-58. 3 4 Benson, D. A., Cavanaugh, M., Clark, K., Karsch-Mizrachi, I., Lipman, D. J., Ostell, J. and 5 Sayers, E. W. (2017) 'GenBank', Nucleic acids research, 45(D1), D37-D42. 6 7 Bian, X. L., Wang, H. T., Ning, Y., Lee, S. Y. and Fenno, J. C. (2005) 'Mutagenesis of a 8 novel gene in the prcA-prtP protease locus affects expression of Treponema denticola 9 membrane complexes', Infection and Immunity, 73(2), 1252-1255. 10 11 Bitto, N. J., Chapman, R., Choi, J., Ferrero, R. L., Pidot, S., Costin, A., Lo, C., D'Cruze, T., 12 Reynolds, E. C., Dashper, S. G., Turnbull, L., Whitchurch, C. B. and Stacey, K. J. 13 (2017) 'Bacterial membrane vesicles transport their DNA cargo into host cells', 14 Scientific Reports, 7(1). 15 16 Blakemore, R. P. and Canaleparola, E. (1976) 'ARGININE CATABOLISM BY 17 TREPONEMA-DENTICOLA', Journal of Bacteriology, 128(2), 616-622. 18 19 Blanco, P., Hernando-Amado, S., Reales-Calderon, J. A., Corona, F., Lira, F., Alcalde-Rico, 20 M., Bernardini, A., Sanchez, M. B. and Martinez, J. L. (2016) 'Bacterial Multidrug 21 Efflux Pumps: Much More Than Antibiotic Resistance Determinants', 22 Microorganisms, 4(1). 23 24 Bonnington, K. E. and Kuehn, M. J. (2014) 'Review: Protein selection and export via outer 25 membrane vesicles', BBA - Molecular Cell Research, 1843, 1612-1619. 26 27 Boughton, B. A., Callahan, D. L., Silva, C., Bowne, J., Nahid, A., Rupasinghe, T., Tull, D. 28 L., McConville, M. J., Bacic, A. and Roessner, U. (2011) 'Comprehensive profiling 29 and quantitation of amine group containing metabolites', Analytical Chemistry, 30 83(19), 7523-7530. 31 32 Bourgeau, G., Lapointe, H., Peloquin, P. and Mayrand, D. (1992) 'CLONING, 33 EXPRESSION, AND SEQUENCING OF A PROTEASE GENE (TPR) FROM 34 PORPHYROMONAS-GINGIVALIS W83 IN ESCHERICHIA-COLI', Infection and 35 Immunity, 60(8), 3186-3192. 36 37 Boutrin, M.-C., Wang, C., Aruni, W., Li, X. and Fletcher, H. M. (2012) 'Nitric Oxide Stress 38 Resistance in Porphyromonas gingivalis Is Mediated by a Putative Hydroxylamine 39 Reductase', Journal of Bacteriology, 194(6), 1582-1592. 40

155

1 Bradshaw, D. J., Homer, K. A., Marsh, P. D. and Beighton, D. (1994) 'Metabolic cooperation 2 in oral microbial communities during growth on mucin', Microbiology, (12), 3407. 3 4 Bradshaw, D. J. and Marsh, P. D. (1998) 'Analysis of pH-Driven Disruption of Oral 5 Microbial Communities in vitro', Caries Research, 32(6), 456-462. 6 7 Brown, S. A. and Whiteley, M. (2007) 'A Novel Exclusion Mechanism for Carbon Resource 8 Partitioning in Aggregatibacter actinomycetemcomitans', Journal of Bacteriology, 9 189(17), 6407-6414. 10 11 Burgess, N. A., Kirke, D. F., Williams, P., Winzer, K., Hardie, K. R., Meyers, N. L., Aduse- 12 Opoku, J., Curtis, M. A. and Camara, M. (2002) 'LuxS-dependent quorum sensing in 13 Porphyromonas gingivalis modulates protease and haemagglutinin activities but is not 14 essential for virulence', Microbiology-Sgm, 148, 763-772. 15 16 Burr, G. O. and Miller, E. S. (1941) 'Ultraviolet Absorption Spectra of Fatty Acids and their 17 Application to Chemical Problems', Chemical Reviews, 29(2), 419-438. 18 19 Byrne, S. J., Dashper, S. G., Darby, I. B., Adams, G. G., Hoffmann, B. and Reynolds, E. C. 20 (2009) 'Progression of chronic periodontitis can be predicted by the levels of 21 Porphyromonas gingivalis and Treponema denticola in subgingival plaque', Oral 22 Microbiology and Immunology, 24(6), 469-477. 23 24 C. J. R. Loewen, a., M. L. Gaspar, a., S. A. Jesch, a., C. Delon, a., N. T. Ktistakis, a., S. A. 25 Henry, a. and T. P. Levine, a. (2004) 'Phospholipid Metabolism Regulated by a 26 Transcription Factor Sensing Phosphatidic Acid', Science, (5677), 1644. 27 28 29 Capestany, C. A., Tribble, G. D., Maeda, K., Demuth, D. R. and Lamont, R. J. (2008) 'Role 30 of the Clp system in stress tolerance, biofilm formation, and intracellular invasion in 31 Porphyromonas gingivalis', Journal of Bacteriology, 190(4), 1436-1446. 32 33 Caspi, R., Billington, R., Ferrer, L., Foerster, H., Fulcher, C. A., Keseler, I. M., Kothari, A., 34 Krummenacker, M., Latendresse, M., Mueller, L. A., Ong, Q., Paley, S., Subhraveti, 35 P., Weaver, D. S. and Karp, P. D. (2016) 'The MetaCyc database of metabolic 36 pathways and enzymes and the BioCyc collection of pathway/genome databases', 37 Nucleic acids research, 44(D1), D471-D480. 38 39 Cecil, J. D., O’Brien-Simpson, N. M., Lenzo, J. C., Holden, J. A., Chen, Y.-Y., Singleton, 40 W., Gause, K. T., Yan, Y., Caruso, F. and Reynolds, E. C. (2016) 'Differential 41 Responses of Pattern Recognition Receptors to Outer Membrane Vesicles of Three 42 Periodontal Pathogens', Plos One, 11(4), 1-20.

156

1 2 Cellini, B., Bertoldi, M. and Voltattorni, C. B. (2003) 'Treponema denticola cystalysin 3 catalyzes beta-desulfination of L-cysteine sulfinic acid and beta-decarboxylation of L- 4 aspartate and oxalacetate', Febs Letters, 554(3), 306-310. 5 6 Chapple, I. L. C., Brock, G., Eftimiadi, C. and Matthews, J. B. (2002) 'Glutathione in 7 gingival crevicular fluid and its relation to local antioxidant capacity in periodontal 8 health and disease', Molecular Pathology, 55(6), 367-373. 9 10 Charon, N. W., Cockburn, A., Li, C., Liu, J., Miller, K. A., Miller, M. R., Motaleb, M. A. and 11 Wolgemuth, C. W. (2012) 'The Unique Paradigm of Spirochete Motility and 12 Chemotaxis', Annual Review of Microbiology, Vol 66, 66, 349-370. 13 14 Charon, N. W. and Goldstein, S. F. (2002) 'Genetics of motility and chemotaxis of a 15 fascinating group of bacteria: the spirochetes', Annual Review of Genetics, 36, 47-73. 16 17 Chawla, A., Hirano, T., Bainbridge, B. W., Demuth, D. R., Xie, H. and Lamont, R. J. (2010) 18 'Community signalling between Streptococcus gordonii and Porphyromonas 19 gingivalis is controlled by the transcriptional regulator CdhR', Molecular 20 Microbiology, 78(6), 1510-1522.

21 Cheaib, Z., Rakmathulina, E., Lussi, A. and Eick, S. (2016) 'Impact of Acquired Pellicle 22 Modification on Adhesion of Early Colonizers', Caries Research, 49(6), 626-632.

23 Chen, T., Nakayama, K., Belliveau, L. and Duncan, M. J. (2001) 'Porphyromonas gingivalis 24 gingipains and adhesion to epithelial cells', Infection and Immunity, 69(5), 3048-3056. 25 26 Chen, T., Yu, W.-H., Izard, J., Baranova, O. V., Lakshmanan, A. and Dewhirst, F. E. (2010) 27 'The Human Oral Microbiome Database: a web accessible resource for investigating 28 oral microbe taxonomic and genomic information', Database: The Journal Of 29 Biological Databases And Curation, 2010, baq013-baq013. 30 31 Chen, Y. Y., Cross, K. J., Paolini, R. A., Fielding, J. E., Slakeski, N. and Reynolds, E. C. 32 (2002) 'CPG70 is a novel basic metallocarboxypeptidase with C-terminal polycystic 33 kidney disease domains from Porphyromonas gingivalis', Journal of Biological 34 Chemistry, 277(26), 23433-23440. 35 36 Chen, Z. X., Potempa, J., Polanowski, A., Wikstrom, M. and Travis, J. (1992) 37 'PURIFICATION AND CHARACTERISATION OF A 50-KDA CYSTEINE 38 PROTEINASE (GINGIPAIN) FROM PORPHYROMONAS-GINGIVALIS', Journal 39 of Biological Chemistry, 267(26), 18896-18901. 40

157

1 China, K., Dabra, S. and Kaushik, A. (2012) 'Salivary enzymes as diagnostic markers for 2 detection of gingival/periodontal disease and their correlation with the severity of the 3 disease', Journal of Indian Society of Periodontology, (3). 4 5 Chionh, Y. T., Arulmuruganar, A., Venditti, E., Ng, G. Z., Han, J.-X., Entwisle, C., Ang, C.- 6 S., Colaco, C. A., McNulty, S. and Sutton, P. (2014) 'Heat shock protein complex 7 vaccination induces protection against Helicobacter pylori without exogenous 8 adjuvant', Vaccine, 32, 2350-2358. 9 10 Chu, L., Ebersole, J. L. and Holt, S. C. (1999) 'Hemoxidation and binding of the 46-kDa 11 cystalysin of Treponema denticola leads to a cysteine-dependent hemolysis of human 12 erythrocytes', Oral Microbiology and Immunology, 14(5), 293-303. 13 14 Chu, L., Ebersole, J. L., Kurzban, G. P. and Holt, S. C. (1997) 'Cystalysin, a 46-kilodalton 15 cysteine desulfhydrase from Treponema denticola, with hemolytic and hemoxidative 16 activities', Infection and Immunity, 65(8), 3231-3238. 17 18 Chu, L. R., Dong, Z., Xu, X. P., Cochran, D. L. and Ebersole, J. L. (2002) 'Role of 19 glutathione metabolism of Treponema denticola in bacterial growth and virulence 20 expression', Infection and Immunity, 70(3), 1113-1120. 21 22 Chu, L. R., Ebersole, J. L., Kurzban, G. P. and Holt, S. C. (1999) 'Cystalysin, a 46-kDa L- 23 cysteine desulfhydrase from Treponema denticola: Biochemical and biophysical 24 characterisation', Clinical Infectious Diseases, 28(3), 442-450. 25 26 Chu, L. R., Lai, Y. L., Xu, X. P., Eddy, S., Yang, S., Song, L. and Kolodrubetz, D. (2008) 'A 27 52-kDa leucyl aminopeptidase from Treponema denticola is a cysteinylglycinase that 28 mediates the second step of glutathione metabolism', Journal of Biological Chemistry, 29 283(28), 19351-19358. 30 31 Chu, L. R., Xu, X. P., Dong, Z., Cappelli, D. and Ebersole, J. L. (2003) 'Role for recombinant 32 gamma-glutamyltransferase from Treponema denticola in glutathione metabolism', 33 Infection and Immunity, 71(1), 335-342. 34 35 Chung, W. S. O., Park, Y., Lamont, R. J., McNab, R., Barbieri, B. and Demuth, D. R. (2001) 36 'Signalling system in Porphyromonas gingivalis based on a LuxS protein', Journal of 37 Bacteriology, 183(13), 3903-3909. 38 39 40 Cockayne, A., Sanger, R., Ivic, A., Strugnell, R. A., Macdougall, J. H., Russell, R. R. B. and 41 Penn, C. W. (1989) 'Antigenic and Structural-Analysis of Treponema denticola', 42 Microbiology, 135, 3209-3218.

158

1 2 Cogoni, V., Morgan-Smith, A., Fenno, J. C., Jenkinson, H. F. and Dymock, D. (2012) 3 'Treponema denticola chymotrypsin-like proteinase (CTLP) integrates spirochaetes 4 within oral microbial communities', Microbiology-Sgm, 158, 759-770. 5 6 Colombo, A. P. V., Boches, S. K., Cotton, S. L., Goodson, J. M., Kent, R., Haffajee, A. D., 7 Socransky, S. S., Hasturk, H., Van Dyke, T. E. and Dewhirst, F. (2009) 'Comparisons 8 of Subgingival Microbial Profiles of Refractory Periodontitis, Severe Periodontitis, 9 and Periodontal Health Using the Human Oral Microbe Identification Microarray', 10 Journal of Periodontology, 80(9), 1421-1432. 11 12 Conover, W. J. and Iman, R. L. (1981) 'Rank Transformations as a Bridge Between 13 Parametric and Nonparametric Statistics', The American Statistician, (3), 124. 14 15 Cuadra-Saenz, G., Rao, D. L., Underwood, A. J., Belapure, S. A., Campagna, S. R., Sun, Z. 16 C., Tammariello, S. and Rickard, A. H. (2012) 'Autoinducer-2 influences interactions 17 amongst pioneer colonizing streptococci in oral biofilms', Microbiology-Sgm, 158, 18 1783-1795. 19 20 Cugini, C., Stephens, D. N., Nguyen, D., Kantarci, A. and Davey, M. E. (2013) 'Arginine 21 deiminase inhibits Porphyromonas gingivalis surface attachment', Microbiology-Sgm, 22 159, 275-285. 23 24 Dashper, S. G., Ang, C. S., Veith, P. D., Mitchell, H. L., Lo, A. W. H., Seers, C. A., Walsh, 25 K. A., Slakeski, N., Chen, D., Lissel, J. P., Butler, C. A., O'Brien-Simpson, N. M., 26 Barr, I. G. and Reynolds, E. C. (2009) 'Response of Porphyromonas gingivalis to 27 Heme Limitation in Continuous Culture', Journal of Bacteriology, 191(3), 1044-1055. 28 29 Dashper, S. G., Brownfield, L., Slakeski, N., Zilm, P. S., Rogers, A. H. and Reynolds, E. C. 30 (2001) 'Sodium ion-driven serine/threonine transport in Porphyromonas gingivalis', 31 Journal of Bacteriology, 183(14), 4142-4148. 32 33 Dashper, S. G., Butler, C. A., Lissel, J. P., Paolini, R. A., Hoffmann, B., Veith, P. D., O 34 Brien-Simpson, N. M., Snelgrove, S. L., Tsiros, J. T. and Reynolds, E. C. (2005) 'A 35 Novel Porphyromonas gingivalis FeoB Plays a Role in Manganese Accumulation', 36 Journal of Biological Chemistry, 280(30), 28095-28102. 37 38 Dashper, S. G., Cross, K. J., Slakeski, N., Lissel, P., Aulakh, P., Moore, C. and Reynolds, E. 39 C. (2004) 'Hemoglobin hydrolysis and heme acquisition by Porphyromonas 40 gingivalis', Oral Microbiology and Immunology, 19(1), 50-56. 41 42 Dashper, S. G., Mitchell, H. L., Seers, C. A., Cross, K. J., Cleal, S. M., Reynolds, E., 43 Gladman, S. L., Seemann, T., Bulach, D. M. and Chandry, P. S. (2017) 159

1 'Porphyromonas gingivalis Uses Specific Domain Rearrangements and Allelic 2 Exchange to Generate Diversity in Surface Virulence Factors', Frontiers In 3 Microbiology, 8. 4 5 Dashper, S. G., Seers, C. A., Tan, K. H. and Reynolds, E. C. (2011) 'Virulence Factors of the 6 Oral Spirochete Treponema denticola', Journal of Dental Research, 90(6), 691-703. 7 8 Davidson, A. L., Dassa, E., Orelle, C. and Chen, J. (2008) 'Structure, function, and evolution 9 of bacterial ATP-binding cassette systems', Microbiology and Molecular Biology 10 Review, 72(2), 317-364. 11 12 de Kievit, T. R. and Iglewski, B. H. (2000) 'Bacterial Quorum Sensing in Pathogenic 13 Relationships', American Society for Microbiology, 68(9), 4839-4849. 14 15 De Las Peñas, A., Connolly, L. and Gross, C. A. (1997) 'The sigmaE-mediated response to 16 extracytoplasmic stress in Escherichia coli is transduced by RseA and RseB, two 17 negative regulators of sigmaE', Molecular Microbiology, 24(2), 373-385. 18 19 Demuth, D. R., Irvine, D. C., Costerton, J. W., Cook, G. S. and Lamont, R. J. (2001) 'Discrete 20 protein determinant directs the species-specific adherence of Porphyromonas 21 gingivalis to oral streptococci', Infection and Immunity, 69(9), 5736-5741. 22 23 Deng, Z.-L., Szafrański, S. P., Jarek, M., Bhuju, S. and Wagner-Döbler, I. (2017) 'Dysbiosis 24 in chronic periodontitis: Key microbial players and interactions with the human host', 25 Scientific Reports, 7(1), 3703-3703. 26 27 Deng, Z.-L., Sztajer, H., Jarek, M., Bhuju, S. and Wagner-Doebler, I. (2018) 'Worlds Apart - 28 Transcriptome Profiles of Key Oral Microbes in the Periodontal Pocket Compared to 29 Single Laboratory Culture Reflect Synergistic Interactions', Frontiers In 30 Microbiology, 9. 31 32 Dennis, G., Jr., Sherman, B. T., Hosack, D. A., Yang, J., Gao, W., Lane, H. C. and Lempicki, 33 R. A. (2003) 'DAVID: Database for Annotation, Visualization, and Integrated 34 Discovery', Genome Biology, 4(5), P3-P3. 35 36 Dewhirst, F. E., Tamer, M. A., Ericson, R. E., Lau, C. N., Levanos, V. A., Boches, S. K., 37 Galvin, J. L. and Paster, B. J. (2000) 'The diversity of periodontal spirochetes by 16S 38 rRNA analysis', Oral Microbiology & Immunology, 15(3), 196. 39

160

1 Di Martino, M. L., Campilongo, R., Casalino, M., Micheli, G., Colonna, B. and Prosseda, G. 2 (2013) 'Mini review: Polyamines: Emerging players in bacteria–host interactions', 3 International Journal of Medical Microbiology, 303, 484-491. 4 5 Diaz, P. I., Zilm, P. S. and Rogers, A. H. (2002) 'Fusobacterium nucleatum supports the 6 growth of Porphyromonas gingivalis in oxygenated and carbon-dioxide-depleted 7 environments', Microbiology-Sgm, 148, 467-472. 8 9 Diggle, S. P., Gardner, A., West, S. A. and Griffin, A. S. (2007) 'Evolutionary theory of 10 bacterial quorum sensing: when is a signal not a signal?', Philosophical Transactions 11 of the Royal Society B-Biological Sciences, 362(1483), 1241-1249. 12 13 Dou, Y., Aruni, W., Luo, T., Roy, F., Wang, C. and Fletcher, H. M. (2014) 'Involvement of 14 PG2212 zinc finger protein in the regulation of oxidative stress resistance in 15 Porphyromonas gingivalis W83', Journal of Bacteriology, 196(23), 4057-4070. 16 17 Dou, Y., Aruni, W., Muthiah, A., Roy, F., Wang, C. and Fletcher, H. M. (2016) 'Studies of 18 the extracytoplasmic function sigma factor PG0162 in Porphyromonas gingivalis', 19 Molecular Oral Microbiology, 31(3), 270-283. 20 21 Dou, Y., Rutanhira, H., Chen, X., Mishra, A., Wang, C. and Fletcher, H. M. (2018) 'Role of 22 extracytoplasmic function sigma factor PG1660 (RpoE) in the oxidative stress 23 resistance regulatory network of Porphyromonas gingivalis', Molecular Oral 24 Microbiology, (1), 89. 25 26 Du, D., van Veen, H. W., Murakami, S., Pos, K. M. and Luisi, B. F. (2015) 'Structure, 27 mechanism and cooperation of bacterial multidrug transporters', Current Opinion in 28 Structural Biology, 33, 76-91. 29 30 Duran-Pinedo, A. E., Chen, T., Teles, R., Starr, J. R., Wang, X., Krishnan, K. and Frias- 31 Lopez, J. (2014) 'Community-wide transcriptome of the oral microbiome in subjects 32 with and without periodontitis', ISME Journal: Multidisciplinary Journal of Microbial 33 Ecology, 8(8), 1659-1672. 34 35 Edwards, A. M., Jenkinson, H. F., Woodward, M. J. and Dymock, D. (2005) 'Binding 36 properties and adhesion-mediating regions of the major sheath protein of Treponema 37 denticola ATCC 35405', Infection and Immunity, 73(5), 2891-2898. 38 39 Egland, P. G., Palmer, R. J. and Kolenbrander, P. E. (2004) 'Interspecies communication in 40 Streptococcus gordonii-Veillonella atypica biofilms: Signalling in flow conditions 41 requires juxtaposition', Proceedings of the National Academy of Sciences of the 42 United States of America, 101(48), 16917-16922.

161

1 2 Egli, C., Leung, W. K., Muller, K. H., Hancock, R. E. W. and McBride, B. C. (1993) 'PORE- 3 FORMING PROPERTIES OF THE MAJOR 53-KILODALTON SURFACE- 4 ANTIGEN FROM THE OUTER SHEATH OF TREPONEMA-DENTICOLA', 5 Infection and Immunity, 61(5), 1694-1699. 6 7 Eke, P. I., Wei, L., Thornton-Evans, G. O. and Genco, R. J. (2012) 'Prevalence of 8 Periodontitis in Adults in the United States: 2009 and 2010', Journal of Dental 9 Research, 91(10), 914-920. 10 11 Elbourne, L. D. H., Tetu, S. G., Hassan, K. A. and Paulsen, I. T. (2017) 'TransportDB 2.0: a 12 database for exploring membrane transporters in sequenced genomes from all 13 domains of life', Nucleic acids research, 45(D1), pD320-pD324. 14 15 Ellen, R. P., Ko, K. S. C., Lo, C. M., Grove, D. A. and Ishihara, K. (2000) 'Insertional 16 inactivation of the prtP gene of Treponema denticola confirms dentilisin's disruption 17 of epithelial junctions', Journal of Molecular Microbiology and Biotechnology, 2(4), 18 581-586.

19 Endo, A., Watanabe, T., Ogata, N., Nozawa, T., Aikawa, C., Arakawa, S., Maruyama, F., 20 Izumi, Y. and Nakagawa, I. (2015) 'Comparative genome analysis and identification 21 of competitive and cooperative interactions in a polymicrobial disease', Isme Journal, 22 9(3), 629-642.

23 Enersen, M., Kazuhiko, N. and Atsuo, A. (2013) 'Porphyromonas gingivalis fimbriae', 24 Journal of Oral Microbiology, 5, 1-10. 25 26 Erbse, A. H. and Falke, J. J. (2009) 'The core signalling proteins of bacterial chemotaxis 27 assemble to form an ultrastable complex', Biochemistry, (29), 6975. 28 29 Espinel, I. C., Guerra, P. R. and Jelsbak, L. (2016) 'Multiple roles of putrescine and 30 spermidine in stress resistance and virulence of Salmonella enterica serovar 31 Typhimurium', Microbial Pathogenesis, 95, 117-123. 32 33 Fenno, J. C. (2012) 'Treponema denticola interactions with host proteins', Journal of Oral 34 Microbiology, 4, 1-13. 35 36 Fenno, J. C., Hannam, P. M., Leung, W. K., Tamura, M., Uitto, V. J. and McBride, B. C. 37 (1998a) 'Cytopathic effects of the major surface protein and the chymotrypsinlike 38 protease of Treponema denticola', Infection and Immunity, 66(5), 1869-1877. 39 40 Fenno, J. C. and McBride, B. C. (1998) 'Virulence factors of oral treponemes', Anaerobe, 41 4(1), 1-17. 42

162

1 Fenno, J. C., Lee, S. Y., Bayer, C. H. and Ning, Y. (2001) 'The opdB locus encodes the 2 trypsin-like peptidase activity of Treponema denticola', Infection and Immunity, 3 69(10), 6193-6200. 4 5 Fenno, J. C., Muller, K. H. and McBride, B. C. (1996) 'Sequence analysis, expression, and 6 binding activity of recombinant major outer sheath protein (Msp) of Treponema 7 denticola', Journal of Bacteriology, 178(9), 2489-2497. 8 9 Fenno, J. C., Tamura, M., Hannam, P. M., Wong, G. W. K., Chan, R. A. and McBride, B. C. 10 (2000) 'Identification of a Treponema denticola OppA homologue that binds host 11 proteins present in the subgingival environment', Infection and Immunity, 68(4), 1884- 12 1892. 13 14 Fenno, J. C., Wong, G. W. K., Hannam, P. M. and McBride, B. C. (1998b) 'Mutagenesis of 15 outer membrane virulence determinants of the oral spirochete Treponema denticola', 16 Fems Microbiology Letters, 163(2), 209-215. 17 18 Ferreyra, J. A., Wu, K. J., Hryckowian, A. J., Bouley, D. M., Weimer, B. C. and Sonnenburg, 19 J. L. (2014) 'Gut microbiota-produced succinate promotes C. difficile infection after 20 antibiotic treatment or motility disturbance', Cell Host & Microbe, 16(6), 770-777. 21 22 Fletcher, H. M., Schenkein, H. A., Morgan, R. M., Bailey, K. A., Berry, C. R. and Macrina, 23 F. L. (1995) 'VIRULENCE OF A PORPHYROMONAS-GINGIVALIS W83 24 MUTANT DEFECTIVE IN THE PRTH GENE', Infection and Immunity, 63(4), 25 1521-1528. 26 27 Frederick, J., Marconi, R. T. and Sarkar, J. The Hpk2-Rrp2 two-component regulatory system 28 of Treponema denticola: a potential regulator of environmental and adaptive 29 responses. 30 31 Frederick, J. R., Sarkar, J., McDowell, J. V. and Marconi, R. T. (2011) 'Molecular Signalling 32 Mechanisms of the Periopathogen, Treponema denticola', Journal of Dental 33 Research, 90(10), 1155-1163. 34 35 Frias, J., Olle, E. and Alsina, M. (2001) 'Periodontal pathogens produce quorum sensing 36 signal molecules', Infection and Immunity, 69(5), 3431-3434. 37 38 Fukamachi, H., Nakano, Y., Okano, S., Shibata, Y., Abiko, Y. and Yamashita, Y. (2005) 39 'High production of methyl mercaptan by L-methionine-alpha-deamino-gamma- 40 mercaptomethane lyase from Treponema denticola', Biochemical and Biophysical 41 Research Communications, 331(1), 127-131. 42

163

1 2 Genco, R. J. and Borgnakke, W. S. (2013) 'Risk factors for periodontal disease', 3 Periodontology 2000, 62(1), 59-94. 4 5 Glew, M. D., Veith, P. D., Peng, B., Chen, Y.-Y., Gorasia, D. G., Yang, Q., Slakeski, N., 6 Chen, D., Moore, C., Crawford, S. and Reynolds, E. C. (2012) 'PG0026 Is the C- 7 terminal Signal Peptidase of a Novel Secretion System of Porphyromonas gingivalis', 8 287(29), 24605-24617. 9 10 Godovikova, V., Goetting-Minesky, M. P. and Fenno, J. C. (2011) 'Composition and 11 Localisation of Treponema denticola Outer Membrane Complexes', Infection and 12 Immunity, 79(12), 4868-4875. 13 14 Godovikova, V., Wang, H. T., Goetting-Minesky, M. P., Ning, Y., Capone, R. F., Slater, C. 15 K. and Fenno, J. C. (2010) 'Treponema denticola PrcB Is Required for Expression and 16 Activity of the PrcA-PrtP (Dentilisin) Complex', Journal of Bacteriology, 192(13), 17 3337-3344. 18 19 González-Pastor, J. E. (2011) 'Cannibalism: a social behavior in sporulating Bacillus subtilis', 20 FEMS Microbiology Reviews, 35(3), 415-424. 21 22 Greene, S. R. and Stamm, L. V. (1999) 'Molecular characterisation of a chemotaxis operon in 23 the oral spirochete, Treponema denticola', Gene, 232, 59-68. 24 25 Grenier, D. (1992) 'Nutritional interactions between 2 suspected periodontopathogens, 26 Treponema denticola and Porphyromonas gingivalis', Infection and Immunity, 60(12), 27 5298-5301. 28 29 Grenier, D. (1996) 'Degradation of host protease inhibitors and activation of plasminogen by 30 proteolytic enzymes from Porphyromonas gingivalis and Treponema denticola', 31 Microbiology-Uk, 142, 955-961. 32 33 Grenier, D., Bertrand, J. and Mayrand, D. (1995) 'Porphyromonas gingivalis Outer- 34 Membrane Vesicles Promote Bacterial-Resistance To Chlorhexidine', Oral 35 Microbiology and Immunology, 10(5), 319-320. 36 37 Griffen, A. L., Beall, C. J., Campbell, J. H., Firestone, N. D., Kumar, P. S., Yang, Z. K., 38 Podar, M. and Leys, E. J. (2012) 'Distinct and complex bacterial profiles in human 39 periodontitis and health revealed by 16S pyrosequencing', Isme Journal, 6(6), 1176- 40 1185. 41

164

1 Grkovic, S., Brown, M. H. and Skurray, R. A. (2001) 'Transcriptional regulation of multidrug 2 efflux pumps in bacteria', Seminars in Cell & Developmental Biology, 12(3), 225-237. 3 4 Gui, M. J., Dashper, S. G., Slakeski, N., Chen, Y. Y. and Reynolds, E. C. (2016) 'Spheres of 5 influence: Porphyromonas gingivalis outer membrane vesicles', Molecular Oral 6 Microbiology, 31(5), 365-378. 7 8 Gully, N., Bright, R., Marino, V., Marchant, C., Cantley, M., Haynes, D., Butler, C., Dashper, 9 S., Reynolds, E. and Bartold, M. (2014) 'Porphyromonas gingivalis Peptidylarginine 10 Deiminase, a Key Contributor in the Pathogenesis of Experimental Periodontal 11 Disease and Experimental Arthritis', Plos One, 9(6). 12 13 Guo, L., He, X. and Shi, W. (2014) 'Intercellular communications in multispecies oral 14 microbial communities', Frontiers In Microbiology, 5, 328-328. 15 16 Guo, Y. H., Nguyen, K. A. and Potempa, J. (2010) 'Dichotomy of gingipains action as 17 virulence factors: from cleaving substrates with the precision of a surgeon's knife to a 18 meat chopper-like brutal degradation of proteins', Periodontology 2000, 54, 15-44. 19 20 Hajishengallis, G., Darveau, R. P. and Curtis, M. A. (2012) 'The keystone-pathogen 21 hypothesis', Nature Reviews Microbiology, 10(10), 717-725. 22 23 Hajishengallis, G. and Lamont, R. J. (2014) 'Breaking bad: Manipulation of the host response 24 by Porphyromonas gingivalis', European Journal of Immunology, 44(2), 328-338. 25 26 Hardie, K. R. and Heurlier, K. (2008) 'Establishing bacterial communities by 'word of mouth': 27 LuxS and autoinducer 2 in biofilm development', Nature Reviews Microbiology, 6(8), 28 635-643. 29 30 Hashimoto, M., Ogawa, S., Asai, Y., Takai, Y. and Ogawa, T. (2003) 'Binding of 31 Porphyromonas gingivalis fimbriae to Treponema denticola dentilisin', Fems 32 Microbiology Letters, 226(2), 267-271. 33 34 He, J., Miyazaki, H., Anaya, C., Yu, F., Yeudall, W. A. and Lewis, J. P. (2006) 'Role of 35 Porphyromonas gingivalis FeoB2 in metal uptake and oxidative stress protection', 36 Infection and Immunity, 74(7), 4214-4223. 37 38 Hendrickson, E. L., Beck, D. A. C., Hackett, M., Miller, D. P., Wang, Q., Lamont, R. J. and 39 Whiteley, M. (2017) 'Insights into Dynamic Polymicrobial Synergy Revealed by 40 Time-Coursed RNA-Seq', Frontiers in Microbiology, 8, 261. 41

165

1 Hendrickson, E. L., Xia, Q. W., Wang, T. S., Lamont, R. J. and Hackett, M. (2009) 'Pathway 2 analysis for intracellular Porphyromonas gingivalis using a strain ATCC 33277 3 specific database', Bmc Microbiology, 9, 10. 4 5 Henke, J. M. and Bassler, B. L. (2004) 'Bacterial social engagements', Trends in Cell Biology, 6 14(11), 648-656. 7 8 Hespell, R. B. and Canalepa.E (1971) 'AMINO ACID AND GLUCOSE FERMENTATION 9 BY TREPONEMA-DENTICOLA', Archiv Fur Mikrobiologie, 78(3), 234-&. 10 11 Hirano, T., Beck, D., Demuth, D., Hackett, M. and Lamont, R. (2012) 'Deep Sequencing of 12 Porphyromonas gingivalis and comparative transcriptome analysis of a LuxS mutant', 13 Frontiers in Cellular and Infection Microbiology, 2(79). 14 Hiratsuka, K., Abiko, Y., Hayakawa, M., Ito, T., Sasahara, H. and Takiguchi, H. (1992) 'Role 15 of Porphyromonas gingivalis 40-Kda Outer Membrane Protein in the aggregation of 16 Porphyromonas gingivalis Vesicles and Actinomyces viscosus', Archives of Oral 17 Biology, 37(9), 717-724. 18 19 Hiratsuka, K., Hayakawa, M., Kiyama-Kishikawa, M., Sasaki, Y., Hirai, T. and Abiko, Y. 20 (2008) 'Role of the hemin-binding protein 35 (HBP35) of Porphyromonas gingivalis 21 in coaggregation', Microbial Pathogenesis, 44(4), 320-328. 22 23 Hiratsuka, K., Kiyama-Kishikawa, M. and Abiko, Y. (2010) 'Hemin-binding protein 35 24 (HBP35) plays an important role in bacteria-mammalian cells interactions in 25 Porphyromonas gingivalis', Microbial Pathogenesis, 48(3-4), 116-123. 26 27 Hojo, K., Nagaoka, S., Murata, S., Taketomo, N., Ohshima, T. and Maeda, N. (2007) 28 'Reduction of vitamin K concentration by salivary Bifidobacterium strains and their 29 possible nutritional competition with Porphyromonas gingivalis', Journal of Applied 30 Microbiology, 103(5), 1969-1974. 31 32 Hojo, K., Nagaoka, S., Ohshima, T. and Maeda, N. (2009) 'Bacterial Interactions in Dental 33 Biofilm Development', Critical Reviews in Oral Biology & Medicine, 88(11), 982- 34 990. 35 36 Holt, S. C. and Ebersole, J. L. (2005) 'Porphyromonas gingivalis, Treponema denticola, and 37 Tannerella forsythia: the 'red complex', a prototype polybacterial pathogenic 38 consortium in periodontitis', Periodontology 2000, 38, 72-122. 39 40 Holt, S. C., Kesavalu, L., Walker, S. and Genco, C. A. (1999) 'Virulence factors of 41 Porphyromonas gingivalis', Periodontology, 20, 168-238.

166

1 2 Hosogi, Y. and Duncan, M. J. (2005) 'Gene Expression in Porphyromonas gingivalis after 3 Contact with Human Epithelial Cells', Infection and Immunity, 73(4), 2327-2335. 4 5 Høvik, H., Yu, W.-H., Olsen, I. and Chen, T. (2012) 'Comprehensive Transcriptome Analysis 6 of the Periodontopathogenic Bacterium Porphyromonas gingivalis W83', Journal of 7 Bacteriology, (1). 8 9 Hromić-Jahjefendić, A., Jajčanin Jozić, N., Kazazić, S., Grabar Branilović, M., Karačić, Z., 10 Schrittwieser, J. H., Das, K. M. P., Tomin, M., Oberer, M., Gruber, K., Abramić, M. 11 and Tomić, S. (2017) 'A novel Porphyromonas gingivalis enzyme: An atypical 12 dipeptidyl peptidase III with an ARM repeat domain', Plos One, 12(11), 1-27. 13 14 Hujoel, P. P., Drangsholt, M., Spiekerman, C. and DeRouen, T. A. (2000) 'Periodontal 15 disease and coronary heart disease risk', Jama-Journal of the American Medical 16 Association, 284(11), 1406-1410. 17 18 Hwang, S., Zhang, Q., Ryu, S. and Jeon, B. (2012) 'Transcriptional regulation of the 19 CmeABC multidrug efflux pump and the KatA catalase by CosR in Campylobacter 20 jejuni', Journal of Bacteriology, 194(24), 6883-6891. 21 22 Ikeda, T. and Yoshimura, F. (2002) 'A resistance-nodulation-cell division family xenobiotic 23 efflux pump in an obligate anaerobe, Porphyromonas gingivalis', Antimicrobial 24 Agents and Chemotherapy, 46(10), 3257-3260. 25 26 Ikegami, A., Honma, K., Sharma, A. and Kuramitsu, H. K. (2004) 'Multiple functions of the 27 leucine-rich repeat protein LrrA of Treponema denticola', Infection and Immunity, 28 72(8), 4619-4627. 29 30 Imamura, T. (2003) 'The role of gingipains in the pathogenesis of periodontal disease', 31 Journal of Periodontology, 74(1), 111-118. 32 33 Inagaki, S., Onishi, S., Kuramitsu, H. K. and Sharma, A. (2006) 'Porphyromonas gingivalis 34 vesicles enhance attachment, and the leucine-rich repeat BspA protein is required for 35 invasion of epithelial cells by "Tannerella forsythia"', Infection and Immunity, 74(9), 36 5023-5028. 37 38 Inoue, T., Nakayama, M., Ono, M., Shimizu, Y., Ohara, N., Taguchi, Y., Kano, K. and 39 Kuroda, T. (2015) 'Characterisation of the tripartite drug efflux pumps of 40 Porphyromonas gingivalis ATCC 33277', New Microbiologica, 38(97), 101-107. 41

167

1 Isabella, V. M. and Clark, V. L. (2011) 'Deep sequencing-based analysis of the anaerobic 2 stimulon in Neisseria gonorrhoeae', Bmc Genomics, 12, 24. 3 4 Ishihara, K. (2010) 'Virulence factors of Treponema denticola', Periodontology 2000, 54(1), 5 117-135. 6 7 Ishihara, K., Miura, T., Kuramitsu, H. K. and Okuda, K. (1996) 'Characterisation of the 8 Treponema denticola prtP gene encoding a prolyl-phenylalanine-specific protease 9 (Dentilisin)', Infection and Immunity, 64(12), 5178-5186. 10 11 Ishihara, K. and Okuda, K. (1999) 'Molecular pathogenesis of the cell surface proteins and 12 lipids from Treponema denticola', Fems Microbiology Letters, 181(2), 199-204. 13 14 Isawa, K., Hojo, K., Yoda, N., Kamiyama, T., Makino, S., Saito, M., Sugano, H., Mizoguchi, 15 C., Kurama, S., Shibasaki, M., Endo, N. and Sato, Y. (2002) 'Isolation and 16 identification of a new bifidogenic growth stimulator produced by Propionibacterium 17 freudenreichii ET-3', Bioscience, Biotechnology, and Biochemistry, 66(3), 679-681. 18 19 Issa, K. H. B., Phan, G. and Broutin, I. (2018) 'Functional Mechanism of the Efflux Pumps 20 Transcription Regulators From Pseudomonas aeruginosa Based on 3D Structures', 21 Frontiers in Molecular Biosciences, 5(57). 22 23 Ito, R., Ishihara, K., Shoji, M., Nakayama, K. and Okuda, K. (2010) 'Hemagglutinin/Adhesin 24 domains of Porphyromonas gingivalis play key roles in coaggregation with 25 Treponema denticola', Fems Immunology and Medical Microbiology, 60(3), 251-260. 26 27 Ivankov, D. N., Payne, S. H., Galperin, M. Y., Bonissone, S., Pevzner, P. A. and Frishman, 28 D. (2013) 'How many signal peptides are there in bacteria?', Environmental 29 Microbiology, 15(4), 983-990. 30 31 Izard, J., Hsieh, C.-E., Limberger, R. J., Mannella, C. A. and Marko, M. (2008) 'Native 32 cellular architecture of Treponema denticola revealed by cryo-electron tomography', 33 Journal of Structural Biology, 163, 10-17. 34 35 Izard, J., McEwen, B. F., Barnard, R. M., Portuese, T., Samsonoff, W. A. and Limberger, R. 36 J. (2004) 'Tomographic reconstruction of treponemal cytoplasmic filaments reveals 37 novel bridging and anchoring components', Molecular Microbiology, 51(3), 609-618. 38

168

1 Izumi, M. and Futoshi, N. (2015) 'Interaction between Streptococcus spp. and Veillonella 2 tobetsuensis in the Early Stages of Oral Biofilm Formation', Journal of Bacteriology, 3 197(13), 2104-2111. 4 5 Jack, A. A., Daniels, D. E., Jepson, M. A., Vickerman, M. M., Lamont, R. J., Jenkinson, H. F. 6 and Nobbs, A. H. (2015) 'Streptococcus gordonii comCDE (competence) operon 7 modulates biofilm formation with Candida albicans', Microbiology (Reading, 8 England), 161(Pt 2), 411-21. 9 10 Jagels, M. A., Travis, J., Potempa, J., Pike, R. and Hugli, T. E. (1996) 'Proteolytic 11 inactivation of the leukocyte C5a receptor by proteinases derived from 12 Porphyromonas gingivalis', Infection and Immunity, 64(6), 1984-1991. 13 14 Jakubovics, N. S., Gill, S. R., Iobst, S. E., Vickerman, M. M. and Kolenbrander, P. E. (2008) 15 'Regulation of gene expression in a mixed-genus community: Stabilized arginine 16 biosynthesis in Streptococcus gordonii by coaggregation with Actinomyces 17 naeslundii', Journal of Bacteriology, 190(10), 3646-3657. 18 19 Jakubovics, N. S., Gill, S. R., Vickerman, M. M. and Kolenbrander, P. E. (2008b) 'Role of 20 hydrogen peroxide in competition and cooperation between Streptococcus gordonii 21 and Actinomyces naeslundii', FEMS Microbiology Ecology, 66(3), 637-644. 22 23 Jakubovics, N. S., Robinson, J. C., Samarian, D. S., Kolderman, E., Yassin, S. A., 24 Bettampadi, D., Bashton, M. and Rickard, A. H. (2015) 'Critical roles of arginine in 25 growth and biofilm development by Streptococcus gordonii', Molecular 26 Microbiology, (2), 281. 27 28 Jang, Y. J., Choi, Y. J., Lee, S. H., Jun, H. K. and Choi, B. K. (2013) 'Autoinducer 2 of 29 Fusobacterium nucleatum as a target molecule to inhibit biofilm formation of 30 periodontopathogens', Archives of Oral Biology, 58(1), 17-27. 31 32 Jarosz, L. M., Deng, D. M., van der Mei, H. C., Crielaard, W. and Krom, B. P. (2009) 33 'Streptococcus mutans Competence-Stimulating Peptide Inhibits Candida albicans 34 Hypha Formation', Eukaryotic Cell, 8(11), 1658-1664. 35 36 Jeffcoat, M. K., Geurs, N. C., Reddy, M. S., Cliver, S. P., Goldenberg, R. L. and Hauth, J. C. 37 (2001) 'Periodontal infection and preterm birth - Results of a prospective study', 38 Journal of the American Dental Association, 132(7), 875-880. 39 40 Ji-Hoi, M., Jae-Hyung, L. and Jin-Yong, L. (2014) 'Microarray analysis of the transcriptional 41 responses of Porphyromonas gingivalis to polyphosphate', Bmc Microbiology, 14(1), 42 1-21.

169

1 2 Ji, A. J., Jaekyung, H. and Jeong-Il, O. (2015) 'Regulation mechanism of the ald gene 3 encoding alanine dehydrogenase in Mycobacterium smegmatis and Mycobacterium 4 tuberculosis by the Lrp/AsnC family regulator AldR', Journal of Bacteriology, (19- 5 20), 3142. 6 7 Jiao, Y., Hasegawa, M. and Inohara, N. (2014) 'The Role of Oral Pathobionts in Dysbiosis 8 during Periodontitis Development', Journal of Dental Research, 93(6), 539-546. 9 10 João, A., McCusker, M. P., Seamus, e. and Martins, M. (2015) 'The Ins and Outs of RND 11 Efflux Pumps in Escherichia coli', Frontiers In Microbiology, 6(587). 12 13 Jong, R. A. and van der Reijden, W. A. (2010) 'Feasibility and therapeutic strategies of 14 vaccines against Porphyromonas gingivalis', Expert Review of Vaccines, 9(2), 193- 15 208. 16 17 Kah Yan, H., Kok Gan, C. and Keang Peng, S. (2016) 'Porphyromonas gingivalis: An 18 Overview of Periodontopathic Pathogen Below the Gum Line', Frontiers in 19 Microbiology, Vol 7 (2016). 20 21 Käll, L., Krogh, A. and Sonnhammer, E. L. L. (2007) 'Advantages of combined 22 transmembrane topology and signal peptide prediction--the Phobius web server', 23 Nucleic acids research, 35(suppl_2), W429-W432. 24 25 Kamaguchi, A., Nakayama, K., Ohyama, T., Watanabe, T., Okamoto, M. and Baba, H. 26 (2001) 'Coaggregation of Porphyromonas gingivalis and Prevotella intermedia', 27 Microbiology and Immunology, 45(9), 649-656. 28 29 Kanehisa, M., Furumichi, M., Tanabe, M., Sato, Y. and Kawashima, M. (2016) 'KEGG as a 30 reference resource for gene and protein annotation', Nucleic Acids Research, 44(D1), 31 D457-62. 32 33 Kanehisa, M. and Goto, S. (2000) 'Kyoto Encyclopedia of Genes and Genomes', Nucleic 34 acids research, 28, 27-30. 35 36 Kataoka, K., Amano, A., Kuboniwa, M., Horie, H., Nagata, H. and Shizukuishi, S. (1997) 37 'Active sites of salivary proline-rich protein for binding to Porphyromonas gingivalis 38 fimbriae', Infection and Immunity, 65(8), 3159-3164. 39

170

1 Kataoka, M., Li, H., Arakawa, S. and Kuramitsu, H. (1997) 'Characterisation of a methyl- 2 accepting chemotaxis protein gene, dmcA, from the oral spirochete Treponema 3 denticola', Infection and Immunity, 65(10), 4011-4016. 4 5 Kato, C., Mikami, M. and Natsuno, T. (2008) 'Participation of glutathione in the elimination 6 of Porphyromonas gingivalis in vivo', Oral Microbiology and Immunology, 23(6), 7 441-448. 8 9 Kato, T., Takahashi, N. and Kuramitsu, H. K. (1992) 'Sequence analysis and characterisation 10 of the Porphyromonas gingivalis prtC gene, which expresses a novel collagenase 11 activity', Journal of Bacteriology, (11-12), 3889. 12 13 Kazutaka, F., Yuichiro, K., Eitoyo, K., Kazuko, O.-S. and Kazuyuki, I. (2017) 'Effect of 14 extracytoplasmic function sigma factors on autoaggregation, hemagglutination, and 15 cell surface properties of Porphyromonas gingivalis', Plos One, 12(9), e0185027. 16 17 Keller, L. and Surette, M. G. (2006) 'Communication in bacteria: an ecological and 18 evolutionary perspective', Nature Reviews Microbiology, 4(4), 249-258. 19 20 Kesavalu, L., Holt, S. C. and Ebersole, J. L. (1998) 'Virulence of a polymicrobic complex, 21 Treponema denticola and Porphyromonas gingivalis, in a murine model', Oral 22 Microbiology and Immunology, 13(6), 373-377. 23 24 Khan, M. A., Ahmad, N., Zafar, A. U., Nasir, I. A. and Qadir, M. A. (2011) 'Isolation and 25 screening of alkaline protease producing bacteria and physio-chemical 26 characterization of the enzyme', African Journal of Biotechnology, 10(6), 6203-6212. 27 28 Kibbe, W. A. (2007) 'OligoCalc: an online oligonucleotide properties calculator', Nucleic 29 acids research, 35(Web Server issue), W43-W46. 30 31 Kigure, T., Saito, A., Seida, K., Yamada, S., Ishihara, K. and Okuda, K. (1995) 32 'DISTRIBUTION OF PORPHYROMONAS-GINGIVALIS AND TREPONEMA- 33 DENTICOLA IN HUMAN SUBGINGIVAL PLAQUE AT DIFFERENT 34 PERIODONTAL POCKET DEPTHS EXAMINED BY 35 IMMUNOHISTOCHEMICAL METHODS', Journal of Periodontal Research, 30(5), 36 332-341. 37 38 Knoten, C. A., Hudson, L. L., Coleman, J. P., Farrow, J. M., III and Pesci, E. C. (2011) 39 'KynR, a Lrp/AsnC-Type Transcriptional Regulator, Directly Controls the 40 Kynurenine Pathway in Pseudomonas aeruginosa', Journal of Bacteriology, 193(23), 41 6567-6575. 42

171

1 Kolenbrander, P. E., Andersen, R. N., Blehert, D. S., Egland, P. G., Foster, J. S. and Palmer, 2 R. J. (2002) 'Communication among oral bacteria', Microbiology and Molecular 3 Biology Reviews, 66(3), 486-+. 4 5 Kolenbrander, P. E. and London, J. (1993) 'ADHERE TODAY, HERE TOMORROW - 6 ORAL BACTERIAL ADHERENCE', Journal of Bacteriology, 175(11), 3247-3252. 7 8 Kolenbrander, P. E., Palmer, R. J., Periasamy, S. and Jakubovics, N. S. (2010) 'Oral 9 multispecies biofilm development and the key role of cell-cell distance', Nature 10 Reviews Microbiology, 8(7), 471-480. 11 12 Kolenbrander, P. E., Palmer, R. J., Rickard, A. H., Jakubovics, N. S., Chalmers, N. I. and 13 Diaz, P. I. (2006) 'Bacterial interactions and successions during plaque development', 14 Periodontology, 42(1), 47-79. 15 16 Kolenbrander, P. E., Parrish, K. D., Andersen, R. N. and Greenberg, E. P. (1995) 17 'INTERGENERIC COAGGREGATION OF ORAL TREPONEMA SPP WITH 18 FUSOBACTERIUM SPP AND INTRAGENERIC COAGGREGATION AMONG 19 FUSOBACTERIUM SPP', Infection and Immunity, 63(12), 4584-4588. 20 21 Kondo, Y., Ohara, N., Sato, K., Yoshimura, M., Yukitake, H., Naito, M., Fujiwara, T. and 22 Nakayama, K. (2010) 'Tetratricopeptide Repeat Protein-Associated Proteins 23 Contribute to the Virulence of Porphyromonas gingivalis', Infection and Immunity, 24 78(6), 2846-2856. 25 26 Kuboniwa, M., Hendrickson, E. L., Xia, Q. W., Wang, T. S., Xie, H., Hackett, M. and 27 Lamont, R. J. (2009) 'Proteomics of Porphyromonas gingivalis within a model oral 28 microbial community', Bmc Microbiology, 9. 29 30 Kuboniwa, M., Houser, J. R., Hendrickson, E. L., Wang, Q., Alghamdi, S. A., Sakanaka, A., 31 Miller, D. P., Hutcherson, J. A., Wang, T., Beck, D. A. C., Whiteley, M., Amano, A., 32 Wang, H., Marcotte, E. M., Hackett, M. and Lamont, R. J. (2017) 'Metabolic crosstalk 33 regulates Porphyromonas gingivalis colonization and virulence during oral 34 polymicrobial infection', Nature Microbiology, 2(11), 1493-1499. 35 36 Kuboniwa, M., Tribble, G. D., James, C. E., Kilic, A. O., Tao, L., Herzberg, M. C., 37 Shizukuishi, S. and Lamont, R. J. (2006) 'Streptococcus gordonii utilises several 38 distinct gene functions to recruit Porphyromonas gingivalis into a mixed community', 39 Molecular Microbiology, 60(1), 121-139. 40 41 Kulp, A. and Kuehn, M. J. (2010) 'Biological Functions and Biogenesis of Secreted Bacterial 42 Outer Membrane Vesicles', Annual Review of Microbiology,Vol 64, 163-184.

172

1 2 Kumagai, Y., Yagishita, H., Yajima, A., Okamoto, T. and Konishi, K. (2005) 'Molecular 3 mechanism for connective tissue destruction by dipeptidyl aminopeptidase IV 4 produced by the periodontal pathogen Porphyromonas gingivalis', Infection and 5 Immunity, 73(5), 2655-2664. 6 7 Kumar, P. S., Leys, E. J., Bryk, J. M., Martinez, F. J., Moeschberger, M. L. and Griffen, A. L. 8 (2006) 'Changes in Periodontal Health Status Are Associated with Bacterial 9 Community Shifts as Assessed by Quantitative 16S Cloning and Sequencing', Journal 10 of clinical microbiology JCM, 44(10). 11 12 Kuniyasu, O. and Tomoko, K.-O. (2009) 'Effects of butyric acid on the periodontal tissue', 13 Japanese Dental Science Review, 45(2),75-82. 14 15 Kuramitsu, H. K., Chen, W. and Ikegami, A. (2005) 'Biofilm formation by the 16 periodontopathic bacteria Treponema denticola and Porphyromonas gingivalis', 17 Journal of Periodontology, 76(11), 2047-2051. 18 19 Kuramitsu, H. K., He, X., Lux, R., Anderson, M. H. and Shi, W. (2007) 'Interspecies 20 Interactions within Oral Microbial Communities', Microbiology and Molecular 21 Biology Reviews, 71(4), 653-670. 22 23 Kuroda, T. and Tsuchiya, T. (2009) 'Review: Multidrug efflux transporters in the MATE 24 family', BBA - Proteins and Proteomics, 1794, 763-768. 25 26 Kwak, M.-K., Ryu, H.-B., Song, S.-H., Lee, J.-W. and Kang, S.-O. (2018) 'Anti-sigma factor 27 YlaD regulates transcriptional activity of sigma factor YlaC and sporulation via 28 manganese-dependent redox-sensing molecular switch in Bacillus subtilis', The 29 Biochemical Journal, 475(13), 2127-2151. 30 31 Lai, Y. and Chu, L. (2008) 'Novel Mechanism for Conditional Aerobic Growth of the 32 Anaerobic Bacterium Treponema denticola', Applied and Environmental 33 Microbiology, 74(1), 73-79. 34 35 Lalla, E., Lamster, I. B., Hofmann, M. A., Bucciarelli, L., Jerud, A. P., Tucker, S., Lu, Y., 36 Papapanou, P. N. and Schmidt, A. M. (2003) 'Oral infection with a periodontal 37 pathogen accelerates early atherosclerosis in apolipoprotein E-null mice', 38 Arteriosclerosis Thrombosis and Vascular Biology, 23(8), 1405-1411. 39 40 Lamont, R. J., El-Sabaeny, A., Park, Y., Cook, G. S., Costerton, J. W. and Demuth, D. R. 41 (2002) 'Role of the Streptococcus gordonii SspB protein in the development of 42 Porphyromonas gingivalis biofilms on streptococcal substrates', Microbiology 43 (Reading, England), 148(Pt 6), 1627-1636. 173

1 2 Lamont, R. J. and Hajishengallis, G. (2015) 'Review: Polymicrobial synergy and dysbiosis in 3 inflammatory disease', Trends in Molecular Medicine, 21, 172-183. 4 5 Lamont, R. J. and Jenkinson, H. F. (1998) 'Life below the gum line: Pathogenic mechanisms 6 of Porphyromonas gingivalis', Microbiology and Molecular Biology Reviews, 62(4), 7 1244-+. 8 9 Lee, S. Y., Bian, X. L., Wong, G. W. K., Hannam, P. M., McBride, B. C. and Fenno, J. C. 10 (2002) 'Cleavage of Treponema denticola PrcA polypeptide to yield protease 11 complex-associated proteins Prca1 and Prca2 is dependent on PrtP', Journal of 12 Bacteriology, 184(14), 3864-3870. 13 14 Leschine, S. B. and Canaleparola, E. (1980) 'ORNITHINE DISSIMILATION BY 15 TREPONEMA-DENTICOLA', Current Microbiology, 3(5), 305-310. 16 17 Lewis, J. P., Iyer, D. and Anaya-Bergman, C. (2009) 'Adaptation of Porphyromonas 18 gingivalis to microaerophilic conditions involves increased consumption of formate 19 and reduced utilisation of lactate', Microbiology (Reading, England), 155(Pt 11), 20 3758-3774. 21 22 Li, C., Moteleb, M. A., Sal, M., Goldstein, F. and Charon, N. (2000) 'Spirochete periplasmic 23 flagella and motility', Journal Molecular Microbiology and Biotechnology, 2(4), 24 3445-354. 25 26 Li, H., Arakawa, S., Deng, Q. D. and Kuramitsu, H. (1999) 'Characterisation of a novel 27 methyl-accepting chemotaxis gene, dmcB, from the oral spirochete Treponema 28 denticola', Infection and Immunity, 67(2), 694-699. 29 30 Li, W., Cong, Q., Pei, J., Kinch, L. N. and Grishin, N. V. (2012) 'The ABC transporters in 31 Candidatus Liberibacter asiaticus', Proteins, 80(11), 2614-2628. 32 Li, Y. H., Hanna, M. N., Svensater, G., Ellen, R. P. and Cvitkovitch, D. G. (2001) 'Cell 33 density modulates acid adaptation in Streptococcus mutans: Implications for survival 34 in biofilms', Journal of Bacteriology, 183(23), 6875-6884. 35 36 Li, Y. H., Tang, N., Aspiras, M. B., Lau, P. C. Y., Lee, J. H., Ellen, R. P. and Cvitkovitch, D. 37 G. (2002) 'A quorum-sensing signalling system essential for genetic competence in 38 Streptococcus mutans is involved in biofilm formation', Journal of Bacteriology, 39 184(10), 2699-2708. 40 41 Limberger, R. J. (2004) 'The Periplasmic Flagellum of Spirochetes', Journal of Molecular 42 Microbiology & Biotechnology, 7(1/2), 30-40.

174

1 2 Lin, X. H., Wu, J. and Xie, H. (2006) 'Porphyromonas gingivalis minor fimbriae are required 3 for cell-cell interactions', Infection and Immunity, 74(10), 6011-6015. 4 5 Lo, A. W., Seers, C. A., Boyce, J. D., Dashper, S. G., Slakeski, N., Lissel, J. P. and Reynolds, 6 E. C. (2009) 'Comparative transcriptomic analysis of Porphyromonas gingivalis 7 biofilm and planktonic cells', Bmc Microbiology, 9, 18-18. 8 9 Locher, K. P. (2016) 'Mechanistic diversity in ATP-binding cassette (ABC) transporters', 10 Nature Structural and Molecular Biology, 23(6), 487-493.

11 12 Long, F., Su, C. C., Lei, H. T., Bolla, J. R., Do, S. V. and Yu, E. W. (2012) 'Structure and 13 mechanism of the tripartite CusCBA heavy-metal efflux complex', Philosophical 14 Transactions of the Royal Society B-Biological Sciences, 367(1592), 1047-1058. 15 16 Low, L. Y., Harrison, P. F., Lin, Y. H., Boyce, J. D., Rood, J. I. and Cheung, J. K. (2016) 17 'RNA-seq analysis of virR and revR mutants of Clostridium perfringens', Bmc 18 Genomics, 17, 14. 19 20 Lu, B. Q. and McBride, B. C. (1998) 'Expression of the tpr protease gene of Porphyromonas 21 gingivalis is regulated by peptide nutrients', Infection and Immunity, 66(11), 5147- 22 5156. 23 24 Lundgren, B. R., Thornton, W., Dornan, M. H., Villegas-Penaranda, L. R., Boddy, C. N. and 25 Nomura, C. T. (2013) 'Gene PA2449 is essential for glycine metabolism and 26 pyocyanin biosynthesis in Pseudomonas aeruginosa PAO1', Journal of Bacteriology, 27 (9-10), 2087. 28 29 Lux, R., Miller, J. N., Park, N. H. and Shi, W. Y. (2001) 'Motility and chemotaxis in tissue 30 penetration of oral epithelial cell layers by Treponema denticola', Infection and 31 Immunity, 69(10), 6276-6283. 32 33 Lux, R., Moter, A. and Shi, W. (2000) 'Chemotaxis in Pathogenic Spirochetes: Directed 34 Movement Toward Targeting Tissues?', Journal of Molecular Microbiology and 35 Biotechnology, 2(4), 355-364. 36 37 Lux, R., Sim, J. H., Tsai, J. P. and Shi, W. (2002) 'Construction and characterisation of a 38 cheA mutant of Treponema denticola', Journal of Bacteriology, 184(11), 3130-3134. 39 40 Maeda, K., Nagata, H., Kuboniwa, M., Kataoka, K., Nishida, N., Tanaka, M. and 41 Shizukuishi, S. (2004) 'Characterisation of binding of Streptococcus oralis 175

1 glyceraldehyde-3-phosphate dehydrogenase to Porphyromonas gingivalis major 2 fimbriae', Infection and Immunity, 72(9), 5475-5477. 3 4 Maeda, K., Nagata, H., Nonaka, A., Kataoka, K., Tanaka, M. and Shizukuishi, S. (2004b) 5 'Oral streptococcal glyceraldehyde-3-phosphate dehydrogenase mediates interaction 6 with Porphyromonas gingivalis fimbriae', Microbes and Infection, 6(13), 1163-1170. 7 8 Maeda, K., Nagata, H., Ojima, M. and Amano, A. (2015) 'Proteomic and Transcriptional 9 Analysis of Interaction between Oral Microbiota Porphyromonas gingivalis and 10 Streptococcus oralis', Journal of Proteome Research, 14 (1), 82-94. 11 12 Maeda, K., Tribble, G. D., Tucker, C. M., Anaya, C., Shizukuishi, S., Lewis, J. P., Demuth, 13 D. R. and Lamont, R. J. (2008) 'A Porphyromonas gingivalis tyrosine phosphatase is 14 a multifunctional regulator of virulence attributes', Molecular Microbiology, 69(5), 15 1153-1164. 16 17 Mäkinen, K. K., Chen, C. Y. and Mäkinen, P. L. (1996) 'Proline iminopeptidase form the 18 outer cell envelope of the human oral spirochete Treponema denticola ATCC 35405', 19 Infection and Immunity, 64(3), 702-708. 20 21 Mäkinen, K. K. and Mäkinen, P. L. (1996) 'The peptidolytic capacity of the spirochete 22 system', Medical Microbiology and Immunology, 185(1), 1-10. 23 24 Mäkinen, K. K., Mäkinen, P. L., Loesche, W. J. and Syed, S. A. (1995) 'Purification and 25 general properties of an oligopeptidase from Treponema denticola ATCC 35405 - A 26 human oral spirochete', Archives of Biochemistry and Biophysics, 316(2), 689-698.

27 Mäkinen, K. K., Mäkinen, P. L. and Syed, S. A. (1992) ' Purification and Substrate 28 Specificity of an Endopeptidase from the Human Oral Spirochete Treponema 29 denticola ATCC 35405, Active on Furylacryloyl-Leu-Gly-Pro-Ala and Bradykinin ', 30 Journal of Biological Chemistry, 267(20), 14285-14293.

31 Mäkinen, K. K., Syed, S. A., Salvador, S. L. and Mäkinen, P. L. (1990) 'Hydrolysis of the 32 Leu-Gly bond of phenylbenzyl-oxycarbonyl-L-Pro-Leu-Gly-L-Pro-D-Arg (a substrate 33 of microbial collagenase) by treponemes isolated from the subgingival plaque of 34 periodontitis patients', Current Microbiology, 20(1), 69-74. 35 36 Mäkinen, P. L. and Mäkinen, K. K. (1997) 'gamma-Glutamyltransferase from the outer cell 37 envelope of Treponema denticola ATCC 35405', Infection and Immunity, 65(2), 685- 38 691. 39

176

1 Mäkinen, P. L., Mäkinen, K. K. and Syed, S. A. (1994) 'An endo-acting proline-specific 2 oligopeptidase from Treponema denticola ATCC 35405 - Evidence of hydrolysis of 3 human bioactive peptides', Infection and Immunity, 62(11), 4938-4947. 4 5 Mäkinen, P. L., Mäkinen, K. K. and Syed, S. A. (1995) ' Role of the chymotrypsin-like 6 membrane-associated proteinase from Treponema denticola ATCC 35405 in 7 inactivation of bioactive peptides ', Infection and Immunity, 63(9), 3567-3575. 8 9 Marchler-Bauer, A., Bo, Y., Han, L., He, J., Lanczycki, C. J., Lu, S., Chitsaz, F., Derbyshire, 10 M. K., Geer, R. C., Gonzales, N. R., Gwadz, M., Hurwitz, D. I., Lu, F., Marchler, G. 11 H., Song, J. S., Thanki, N., Wang, Z., Yamashita, R. A., Zhang, D., Zheng, C., Geer, 12 L. Y. and Bryant, S. H. (2017) 'CDD/SPARCLE: functional classification of proteins 13 via subfamily domain architectures', Nucleic acids research, 45 4. 14 15 Marchler-Bauer, A., Derbyshire, M. K., Gonzales, N. R., Lu, S., Chitsaz, F., Geer, L. Y., 16 Geer, R. C., He, J., Gwadz, M. and Hurwitz, D. I. (2015) 'CDD: NCBI's conserved 17 domain database', Nucleic acids research, 43(D1), D222-D226. 18 19 Martinez, J. L., Sánchez, M. B., Martínez-Solano, L., Hernandez, A., Garmendia, L., Fajardo, 20 A. and Alvarez-Ortega, C. (2009) 'Functional role of bacterial multidrug efflux pumps 21 in microbial natural ecosystems', FEMS Microbiology Reviews, 33(2), 430-449. 22 23 Masuda, K., Yoshioka, M., Hinode, D. and Nakamura, R. (2002) 'Purification and 24 characterisation of arginine carboxypeptidase produced by Porphyromonas 25 gingivalis', Infection and Immunity, 70(4), 1807-1815. 26 27 Mazumdar, V., Snitkin, E. S., Amar, S. and Segre, D. (2009) 'Metabolic Network Model of a 28 Human Oral Pathogen', Journal of Bacteriology, 191(1), 74-90. 29 30 McNab, R., Ford, S. K., El-Sabaeny, A., Barbieri, B., Cook, G. S. and Lamont, R. J. (2003) 31 'LuxS-based signalling in Streptococcus gordonii: Autoinducer 2 controls 32 carbohydrate metabolism and biofilm formation with Porphyromonas gingivalis', 33 Journal of Bacteriology, 185(1), 274-284. 34 35 Mdala, I., Olsen, I., Haffajee, A. D., Socransky, S. S., Thoresen, M. and Blasio, B. F. (2014) 36 'Comparing clinical attachment level and pocket depth for predicting periodontal 37 disease progression in healthy sites of patients with chronic periodontitis using multi- 38 state Markov models', Journal of Clinial Periodontology, 41(9), 837-845. 39 40 Mealey, B. L. and Oates, T. W. (2006) 'Diabetes mellitus anal periodontal diseases', Journal 41 of Periodontology, 77(8), 1289-1303. 42

177

1 2 Meuric, V., Rouillon, A., Chandad, F. and Bonnaure-Mallett, M. (2010) 'Putative respiratory 3 chain of Porphyromonas gingivalis', Future Microbiology, 5(5), 717-734. 4 5 Meurman, J. H., Sanz, M. and Janket, S. J. (2004) 'Oral health, atherosclerosis, and 6 cardiovascular disease', Critical Reviews in Oral Biology & Medicine, 15(6), 403-413. 7 8 Miao, D., Fenno, J. C., Timm, J. C., Joo, N. E. and Kapila, Y. L. (2011) 'The Treponema 9 denticola Chymotrypsin-Like Protease Dentilisin Induces Matrix Metalloproteinase- 10 2-Dependent Fibronectin Fragmentation in Periodontal Ligament Cells', Infection and 11 Immunity, 79(2), 806-811. 12 13 Miller, C. G and Broder D.H. (2004) 'Peptidase T', In Handbook of Proteolytic Enzymes, 2 14 edn, 946-948. Elsevier, London. 15 Miller, D. P., Frederick, J. R., Marconi, R. T. and Sarkar, J. (2014) The Treponema denticola 16 AtcR LytTR domain-containing response regulator interacts with three architecturally 17 distinct promoter elements: implications for understanding the molecular signalling 18 mechanisms that drive the progression of periodontal disease. Molecular Oral 19 Microbiology, 29(5), 219-232. 20 21 Miller, M. B. and Bassler, B. L. (2001) 'Quorum sensing in bacteria', Annual Review of 22 Microbiology, 55, 165-199. 23 24 Milner, P., Batten, J. E. and Curtis, M. A. (1996) 'Development of a simple chemically 25 defined medium for Porphyromonas gingivalis: Requirement for alpha-ketoglutarate', 26 Fems Microbiology Letters, 140(2-3), 125-130. 27 28 Mistou, M. Y. and Gripon, J. C. (1998) 'Catalytic properties of the cysteine aminopeptidase 29 PepC, a bacterial bleomycin hydrolase', Biochimica Et Biophysica Acta, 1383(1), 63- 30 70. 31 32 Miyamoto, M., Ishihara, K. and Okuda, K. (2006) 'The Treponema denticola surface protease 33 dentilisin degrades interleukin-1 beta (IL-1 beta), IL-6, and tumor necrosis factor 34 alpha', Infection and Immunity, 74(4), 2462-2467. 35 36 Moraleda-Munoz, A., Perez, J., Extremera, A. L. and Munoz-Dorado, J. (2010) 'Differential 37 Regulation of Six Heavy Metal Efflux Systems in the Response of Myxococcus 38 xanthus to Copper', Applied and Environmental Microbiology, 76(18), 6069-6076. 39

178

1 Mori, S., Nirasawa, S., Komba, S. and Kasumi, T. (2005) 'Characterization and kinetic 2 analysis of enzyme–substrate recognition by three recombinant lactococcal 3 tripeptidases', Biochimica Et Biophysica Acta, 1748(1), 26-34. 4 5 Moter, A., Hoenig, C., Choi, B. K., Riep, B. and Gobel, U. B. (1998) 'Molecular 6 epidemiology of oral treponemes associated with periodontal disease', Journal of 7 Clinical Microbiology, 36(5), 1399-1403. 8 9 Nagano, K., Hasegawa, Y., Yoshida, Y. and Yoshimura, F. (2017) 'Comparative analysis of 10 motility and other properties of Treponema denticola strains', Microbial 11 Pathogenesis, 102, 82-88. 12 13 Nagayama, M., Sato, M., Yamaguchi, R., Tokuda, C. and Takeuchi, H. (2001) 'Evaluation of 14 co-aggregation among Streptococcus mitis, Fusobacterium nucleatum and 15 Porphyromonas gingivalis', Letters in Applied Microbiology, 33(2), 122-125. 16 17 Naito, M., Hirakawa, H., Yamashita, A., Ohara, N., Shoji, M., Yukitake, H., Nakayama, K., 18 Toh, H., Yoshimura, F., Kuhara, S., Hattori, M., Hayashi, T. and Nakayama, K. 19 (2008) 'Determination of the Genome Sequence of Porphyromonas gingivalis Strain 20 ATCC 33277 and Genomic Comparison with Strain W83 Revealed Extensive 21 Genome Rearrangements in P-gingivalis', DNA Research, 15(4), 215-225. 22 23 Nakamura, T., Amano, A., Nakagawa, I. and Hamada, S. (1999) 'Specific interactions 24 between Porphyromonas gingivalis fimbriae and human extracellular matrix proteins', 25 Fems Microbiology Letters, 175(2), 267-272. 26 27 Nakamura, Y., Umemoto, T., Nakatani, Y., Namikawa, I. and Wadood, A. (1993) 28 'COMMON AND SPECIFIC ANTIGENS OF SEVERAL TREPONEMES 29 DETECTED BY POLYCLONAL ANTISERA AGAINST MAJOR CELLULAR 30 PROTEINS', Oral Microbiology and Immunology, 8(5), 288-294. 31 32 Nakao, R., Hasegawa, H., Ochiai, K., Takashiba, S., Ainai, A., Ohnishi, M., Watanabe, H. 33 and Senpuku, H. (2011) 'Outer Membrane Vesicles of Porphyromonas gingivalis 34 Elicit a Mucosal Immune Response', Plos One, 6(10), 1-7. 35 36 Nakayama, K., Yoshimura, F., Kadowaki, T. and Yamamoto, K. (1996) 'Involvement of 37 Arginine-Specific Cysteine Proteinase (Arg-Gingipain) in Fimbriation of 38 Porphyromonas gingivalis', Journal of Bacteriology, 178(10), 2818-2824. 39 40 Nan, L., Yu, P., Nadkarni, M. A., Ghadikolaee, N. B., Ky-Anh, N., Mihwa, L., Hunter, N. 41 and Collyer, C. A. (2010) 'Structure determination and analysis of a haemolytic

179

1 gingipain adhesin domain from Porphyromonas gingivalis', Molecular Microbiology, 2 76(4), 861-873. 3 4 Narda, T., Natalia, A., Belkis, Q., Elizabeth, S., Luis Enrique, G. and Luis, H. (2008) 5 'Arginine and glutamate levels in the gingival crevicular fluid from patients with 6 chronic periodontitis', Brazilian Dental Journal, (4), 318. 7 8 Nath, S. and Raveendran, R. (2013) 'Microbial dysbiosis in periodontitis', Journal of Indian 9 Society of Periodontology, (4). 10 11 Nelson, D., Potempa, J., Kordula, T. and Travis, J. (1999) 'Purification and characterisation 12 of a novel cysteine proteinase (periodontain) from Porphyromonas gingivalis - 13 Evidence for a role in the inactivation of human alpha(1)-proteinase inhibitor', 14 Journal of Biological Chemistry, 274(18), 12245-12251. 15 16 Nelson, K. E., Fleischmann, R. D., DeBoy, R. T., Paulsen, I. T., Fouts, D. E., Eisen, J. A., 17 Daugherty, S. C., Dodson, R. J., Durkin, A. S., Gwinn, M., Haft, D. H., Kolonay, J. 18 F., Nelson, W. C., Mason, T., Tallon, L., Gray, J., Granger, D., Tettelin, H., Dong, H., 19 Galvin, J. L., Duncan, M. J., Dewhirst, F. E. and Fraser, C. M. (2003) 'Complete 20 genome sequence of the oral pathogenic bacterium Porphyromonas gingivalis strain 21 W83', Journal of Bacteriology, 185(18), 5591-5601. 22 23 Nemoto, T. K. and Ohara-Nemoto, Y. (2016) 'Review Article: Exopeptidases and gingipains 24 in Porphyromonas gingivalis as prerequisites for its amino acid metabolism', 25 Japanese Dental Science Review, 52, 22-29. 26 27 Nemoto, T. K., Ohara-Nemoto, Y., Bezerra, G. A., Shimoyama, Y. and Kimura, S. (2016) 'A 28 Porphyromonas gingivalis Periplasmic Novel Exopeptidase, Acylpeptidyl 29 Oligopeptidase, Releases N-Acylated Di- and Tripeptides from Oligopeptides', The 30 Journal Of Biological Chemistry, 291(11), 5913-5925. 31 32 Ng, H. M., Kin, L. X., Dashper, S. G., Slakeski, N., Butler, C. A. and Reynolds, E. C. (2016) 33 'Bacterial interactions in pathogenic subgingival plaque', Microbial Pathogenesis, 94, 34 60-69 35 36 Nobbs, A. H., Jenkinson, H. F., Jakubovics, N. S., Nobbs, A. H., Jenkinson, H. F. and 37 Jakubovics, N. S. (2011) 'Stick to your gums: mechanisms of oral microbial 38 adherence', Journal of Dental Research, 90(11), 1271-1278. 39

180

1 O'Brien-Simpson, N. M., Veith, P. D., Dashper, S. G. and Reynolds, E. C. (2003) 2 'Porphyromonas gingivalis gingipains: The molecular teeth of a microbial vampire', 3 Current Protein & Peptide Science, 4(6), 409-426. 4 5 O'Toole, G., Kaplan, H. B. and Kolter, R. (2000) 'Biofilm formation as microbial 6 development', Annual Review of Microbiology, 54, 49-79. 7 8 O’Brien-Simpson, N. M., Holden, J. A., Lenzo, J. C., Tan, Y., Brammar, G. C., Walsh, K. A., 9 Singleton, W., Orth, R. K. H., Slakeski, N., Cross, K. J., Darby, I. B., Becher, D., 10 Rowe, T., Morelli, A. B., Hammet, A., Nash, A., Brown, A., Ma, B., Vingadassalom, 11 D., McCluskey, J., Kleanthous, H. and Reynolds, E. C. (2016) 'A therapeutic 12 Porphyromonas gingivalis gingipain vaccine induces neutralising IgG1 antibodies 13 that protect against experimental periodontitis', Npj Vaccines, 1, 16022. 14 15 16 Oda, H., Saiki, K., Tonosaki, M., Yajima, A. and Konishi, K. (2009) 'Participation of the 17 secreted dipeptidyl and tripeptidyl aminopeptidases in asaccharolytic growth of 18 Porphyromonas gingivalis', Journal of Periodontal Research, 44(3), 362-367. 19 20 Ohara-Nemoto, Y., Rouf, S. M. A., Naito, M., Yanase, A., Tetsuo, F., Ono, T., Kobayakawa, 21 T., Shimoyama, Y., Kimura, S., Nakayama, K., Saiki, K., Konishi, K. and Nemoto, T. 22 K. (2014) 'Identification and Characterisation of Prokaryotic Dipeptidyl-peptidase 5 23 from Porphyromonas gingivalis', Journal of Biological Chemistry, 289(9), 5436- 24 5448. 25 26 Ohara-Nemoto, Y., Shimoyama, Y., Kimura, S., Kon, A., Haraga, H., Ono, T. and Nemoto, 27 T. K. (2011) 'Asp- and Glu-specific Novel Dipeptidyl Peptidase 11 of Porphyromonas 28 gingivalis Ensures Utilisation of Proteinaceous Energy Sources', Journal of Biological 29 Chemistry, 286(44), 38115-38127. 30 31 Ohta, K., Mäkinen, K. K. and Loesche, W. J. (1986) 'PURIFICATION AND 32 CHARACTERISATION OF AN ENZYME PRODUCED BY TREPONEMA- 33 DENTICOLA CAPABLE OF HYDROLYZING SYNTHETIC TRYPSIN 34 SUBSTRATES', Infection and Immunity, 53(1), 213-220. 35 36 Olinares, P. D. B., Kim, J. and van Wijk, K. J. (2011) 'Review: The Clp protease system; a 37 central component of the chloroplast protease network', BBA - Bioenergetics, 1807, 38 999-1011. 39 40 Onozawa, S., Kikuchi, Y., Shibayama, K., Kokubu, E., Nakayama, M., Inoue, T., Nakano, 41 K., Shibata, Y., Ohara, N., Nakayama, K., Ishihara, K., Kawakami, T. and Hasegawa, 42 H. (2015) 'Role of extracytoplasmic function sigma factors in biofilm formation of 43 Porphyromonas gingivalis', BMC Oral Health, 15, 4-4.

181

1 2 Ortet, P., De Luca, G., Whitworth, D. E. and Barakat, M. (2012) 'P2TF: a comprehensive 3 resource for analysis of prokaryotic transcription factors', Bmc Genomics, 13, 628- 4 628. 5 6 Orth, R., O'Brien-Simpson, N., Dashper, S., Walsh, K. and Reynolds, E. (2010) 'An efficient 7 method for enumerating oral spirochetes using flow cytometry', Journal of 8 Microbiological Methods, 80(2), 123-128. 9 10 Orth, R. K. H., O'Brien-Simpson, N. M., Dashper, S. G. and Reynolds, E. C. (2011) 11 'Synergistic virulence of Porphyromonas gingivalis and Treponema denticola in a 12 murine periodontitis model', Molecular Oral Microbiology, 26(4), 229-240. 13 14 Otogoto, J. I. and Kuramitsu, H. K. (1993) 'ISOLATION AND CHARACTERISATION OF 15 THE PORPHYROMONAS-GINGIVALIS PRTT GENE, CODING FOR 16 PROTEASE ACTIVITY', Infection and Immunity, 61(1), 117-123. 17 18 Passalacqua, K. D., Varadarajan, A., Ondov, B. D., Okou, D. T., Zwick, M. E. and Bergman, 19 N. H. (2009) 'Structure and Complexity of a Bacterial Transcriptome', Journal of 20 Bacteriology, 191(10), 3203-3211. 21 22 Perez, J. C. and Groisman, E. A. (2009) 'Review: Evolution of Transcriptional Regulatory 23 Circuits in Bacteria', Cell, 138, 233-244. 24 25 Petersen, T. N., Brunak, S., von Heijne, G. and Nielsen, H. (2011) 'SignalP 4.0: 26 discriminating signal peptides from transmembrane regions', Nature Methods, 8(10), 27 785-786. 28 29 Potempa, J., Banbula, A. and Travis, J. (2000) 'Role of bacterial proteinases in matrix 30 destruction and modulation of host responses', Periodontology 2000, 24, 153-192. 31 32 Potempa, J., Pike, R. and Travis, J. (1997) 'Titration and mapping of the active site of 33 cysteine proteinases from Porphyromonas gingivalis (Gingipains) using peptidyl 34 chloromethanes', Biological Chemistry, 378(3-4), 223-230. 35 36 Potempa, J., Sroka, A., Imamura, T. and Travis, J. (2003) 'Gingipains, the Major Cysteine 37 Proteinases and Virulence Factors of Porphyromonas gingivalis: Structure, Function 38 and Assembly of Multidomain Protein Complexes', Current Protein & Peptide 39 Science, 4(6), 397-408. 40 41 Price, A., Garhyan, J. and Gibas, C. (2017) 'The impact of RNA secondary structure on read 42 start locations on the Illumina sequencing platform', Plos One, 12(2), 1-11.

182

1 2 Pritchard, G. G. and Coolbear, T. (1993) 'The physiology and biochemistry of the proteolytic 3 system in lactic acid bacteria', FEMS Microbiology Reviews, 12(1-3), 179-206. 4 5 Priyadarshini, R., Cugini, C., Arndt, A., Chen, T., Tjokro, N. O., Goodman, S. D. and Davey, 6 M. E. (2013) 'The nucleoid-associated protein HUβ affects global gene expression in 7 Porphyromonas gingivalis', Microbiology (Reading, England), 159(Pt 2), 219-229. 8 9 Que, X. C. and Kuramitsu, H. K. (1990) ' Isolation and characterisation of the Treponema 10 denticola prtA gene coding for chymotrypsinlike protease activity and detection of a 11 closely linked gene encoding PZ-PLGPA-hydrolyzing activity', Infection and 12 Immunity, 58(12), 4099-4105.

13 14 Rabouille, C. (2017) 'Pathways of Unconventional Protein Secretion', Trends in Cell Biology, 15 (3), 230. 16 Rawlings, N. D. and Morton, F. R. (2008) 'The MEROPS batch BLAST: A tool to detect 17 peptidases and their non-peptidase homologues in a genome', Biochimie, 90(2), 243- 18 259. 19 20 Rawlings, N. D., Waller, M., Barrett, A. J. and Bateman, A. (2014) 'MEROPS: the database 21 of proteolytic enzymes, their substrates and inhibitors', Nucleic Acids Research, 22 42(D1), D503-D509. 23 24 Rea, D., Lambeir, A.-M., Kumagai, Y., De Meester, I., Scharpe, S., Fül, oml and p, V. (2004) 25 'Expression, purification and preliminary crystallographic analysis of dipeptidyl 26 peptidase IV from Porphyromonas gingivalis', Acta Crystallographica: Section D 27 (Wiley-Blackwell), 60(10), 1871-1873. 28 29 Rickard, A. H., Palmer, R. J., Blehert, D. S., Campagna, S. R., Semmelhack, M. F., Egland, 30 P. G., Bassler, B. L. and Kolenbrander, P. E. (2006) 'Autoinducer 2: a concentration- 31 dependent signal for mutualistic bacterial biofilm growth', Molecular Microbiology, 32 60(6), 1446-1456. 33 34 Romero-Lastra, P., Sanchez, M. C., Ribeiro-Vidal, H., Llama-Palacios, A., Figuero, E., 35 Herrera, D. and Sanz, M. (2017) Comparative gene expression analysis of 36 Porphyromonas gingivalis ATCC 33277 in planktonic and biofilms states. Plos One, 37 12(4). 38 39 Rosen, G., Genzler, T. and Sela, M. N. (2008) 'Coaggregation of Treponema denticola with 40 Porphyromonas gingivalis and Fusobacterium nucleatum is mediated by the major 41 outer sheath protein of Treponema denticola', Fems Microbiology Letters, 289(1), 59- 42 66. 183

1 2 Rosen, G., Naor, R., Kutner, S. and Sela, M. N. (1994) 'CHARACTERISATION OF 3 FIBRINOLYTIC-ACTIVITIES OF TREPONEMA-DENTICOLA', Infection and 4 Immunity, 62(5), 1749-1754. 5 6 Rosen, G., Naor, R., Rahamim, E., Yishai, R. and Sela, M. N. (1995) 'PROTEASES OF 7 TREPONEMA-DENTICOLA OUTER SHEATH AND EXTRACELLULAR 8 VESICLES', Infection and Immunity, 63(10), 3973-3979. 9 10 Rouf, S. M. A., Ohara-Nemoto, Y., Hoshino, T., Fujiwara, T., Ono, T. and Nemoto, T. K. 11 (2013) 'Discrimination based on Gly and Arg/Ser at position 673 between dipeptidyl- 12 peptidase (DPP) 7 and DPP11, widely distributed DPPs in pathogenic and 13 environmental gram-negative bacteria', Biochimie, 95(4), 824-832. 14 15 Ruby, J. D., Li, H., Kuramitsu, H., Norris, S. J., Goldstein, S. F., Buttle, K. F. and Charon, N. 16 W. (1997) 'Relationship of Treponema denticola periplasmic flagella to irregular cell 17 morphology', Journal of Bacteriology, 179(5), 1628-35. 18 19 Ruby, J. D., Lux, R., Shi, W., Charon, N. W. and Dasanayake, A. (2008) 'Effect of glucose 20 on Treponema denticola cell behavior', Molecular Oral Microbiology, 23(3), 234-8. 21 22 Ruggiero, S., Cosgarea, R., Potempa, J., Potempa, B., Eick, S. and Chiquet, M. (2013) 23 'Cleavage of extracellular matrix in periodontitis: Gingipains differentially affect cell 24 adhesion activities of fibronectin and tenascin-C', BBA - Molecular Basis of Disease, 25 1832, 517-526. 26 27 Sah, S., Aluri, S., Rex, K. and Varshney, U. (2015) 'One-Carbon Metabolic Pathway 28 Rewiring in Escherichia coli Reveals an Evolutionary Advantage of 10- 29 Formyltetrahydrofolate Synthetase (Fhs) in Survival under Hypoxia', Journal of 30 Bacteriology, (4). 31 32 Sano, Y., Okamoto-Shibayama, K., Tanaka, K., Ito, R., Shintani, S., Yakushiji, M. and 33 Ishihara, K. (2014) 'Dentilisin involvement in coaggregation between Treponema 34 denticola and Tannerella forsythia', Anaerobe, 30, 45-50. 35 36 Sarkar, J., McHardy, I. H., Simanian, E. J., Shi, W. Y. and Lux, R. (2014) 'Transcriptional 37 Responses of Treponema denticola to Other Oral Bacterial Species', Plos One, 9(2). 38 39 Schauder, S., Shokat, K., Surette, M. G. and Bassler, B. L. (2001) 'The LuxS family of 40 bacterial autoinducers: biosynthesis of a novel quorum-sensing signal molecule', 41 Molecular Microbiology, 41(2), 463-476. 42

184

1 Schmid, F. X. (2005) Biological Macromolecules: UV–visible Spectrophotometry [Topic 2 overview], John Wiley & Sons, Inc. 3 4 Scragg, M. A., Cannon, S. J. and Williams, D. M. (1994) 'Comparative Cytotoxic Effects of 5 Short-chain Fatty Acids Produced by Periodontal Pathogens on Two Cultured 6 Fibroblast Lines', in Great Britain, 1994, JOHN WILEY & SONS, 83. 7 8 Seers, C. A., Slakeski, N., Veith, P. D., Nikolof, T., Chen, Y.-Y., Dashper, S. G. and 9 Reynolds, E. C. (2006) 'The RgpB C-terminal domain has a role in attachment of 10 RgpB to the outer membrane and belongs to a novel C-terminal-domain family found 11 in Porphyromonas gingivalis', Journal of Bacteriology, 188(17), 6376-6386. 12 13 Seshadri, G., Myers, G. S. A., Tettelin, H., Eisen, J. A., Heidelberg, J. F., Dodson, R. J., 14 Davidsen, T. M., DeBoy, R. T., Fouts, D. E., Haft, D. H., Selengut, J., Ren, Q. H., 15 Brinkac, L. M., Madupu, R., Kolonay, J., Durkin, S. A., Daugherty, S. C., Shetty, J., 16 Shvartsbeyn, A., Gebregeorgis, E., Geer, K., Tsegaye, G., Malek, J., Ayodeji, B., 17 Shatsman, S., McLeod, M. P., Smajs, D., Howell, J. K., Pal, S., Amin, A., Vashisth, 18 P., McNeill, T. Z., Xiang, Q., Sodergren, E., Baca, E., Weinstock, G. M., Norris, S. J., 19 Fraser, C. M. and Paulsen, I. T. (2004) 'Comparison of the genome Treponema 20 denticola with of the oral pathogen other spirochete genomes', Proceedings of the 21 National Academy of Sciences of the United States of America, 101(15), 5646-5651. 22 23 Seymour, G. J., Ford, P. J., Cullinan, M. P., Leishman, S. and Yamazaki, K. (2007) 24 'Relationship between periodontal infections and systemic disease', Clinical 25 Microbiology and Infection, 13, 3-10. 26 27 Shah, P. and Swiatlo, E. (2008) 'A multifaceted role for polyamines in bacterial pathogens', 28 Molecular Microbiology, 68(1), 4-16. 29 30 Shapiro, S. S. and Wilk, M. B. (1965) 'An analysis of variance test for normality (complete 31 samples)†', Biometrika, 52(3/4), 591. 32 33 Sharma, K. M., Kumar, R., Panwar, S. and Kumar, A. (2017) 'Microbial alkaline proteases: 34 Optimization of production parameters and their properties', Journal of Genetic 35 Engineering and Biotechnology, (1), 115. 36 37 Shibata, Y., Hayakawa, M., Takiguchi, H., Shiroza, T. and Abiko, Y. (1999) 'Determination 38 and characterisation of the hemagglutinin-associated short motifs found in 39 Porphyromonas gingivalis multiple gene products', Journal od Biological Chemistry, 40 274(8), 5012-5020. 41 42 Shibata, Y., Hiratsuka, K., Hayakawa, M., Shiroza, T., Takiguchi, H., Nagatsuka, Y. and 43 Abiko, Y. (2003) 'A 35-kDa co-aggregation factor is a hemin binding protein in 185

1 Porphyromonas gingivalis', Biochemical and Biophysical Research Communications, 2 300(2), 351-356. 3 4 Sievers, F. and Higgins, D. G. (2014) 'Clustal omega', Current Protocols In Bioinformatics, 5 48, 3.13.1-16. 6 7 Siggers, T. and Gordân, R. (2014) 'Protein–DNA binding: complexities and multi-protein 8 codes', Nucleic acids research, (4), 2099. 9 10 Sim, J. H., Shi, W. and Lux, R. (2005) 'Protein-protein interactions in the chemotaxis 11 signalling pathway of Treponema denticola', Microbiology, 151(Pt 6), 1801-7. 12 13 Simionato, M. R., Tucker, C. M., Kuboniwa, M., Lamont, G., Demuth, D. R., Tribble, G. D. 14 and Lamont, R. J. (2006) 'Porphyromonas gingivalis Genes Involved in Community 15 Development with Streptococcus gordonii', Infection and immunity IAI, 74(11), 6419- 16 6428. 17 18 Socransky, S. S. and Haffajee, A. D. (1992) 'THE BACTERIAL ETIOLOGY OF 19 DESTRUCTIVE PERIODONTAL-DISEASE - CURRENT CONCEPTS', Journal of 20 Periodontology, 63(4), 322-331. 21 22 Socransky, S. S., Haffajee, A. D., Cugini, M. A., Smith, C. and Kent, R. L. (1998) 'Microbial 23 complexes in subgingival plaque', Journal of Clinical Periodontology, 25(2), 134- 24 144. 25 26 Söderling, E., Syed, S., Mäkinen, P. L. and Mäkinen, K. K. (1994) 'Proteolytic Activity of 27 Treponemes from the Subgingival Plaque of Periodontitis Patients', Microbial 28 Ecology in Health and Disease, 7(2), 71. 29 30 Spiga, L., Winter, M. G., Zhu, W., Hughes, E. R., Gillis, C. C., Winter, S. E., de Carvalho, T. 31 F., Santos, R. L., Behrendt, C. L., Hooper, L. V., Kim, J., Chessa, D., Andrews- 32 Polymenis, H. L. and Beiting, D. P. (2017) 'An Oxidative Central Metabolism 33 Enables Salmonella to Utilise Microbiota-Derived Succinate', Cell Host & Microbe, 34 22(3), 291-308. 35 36 Spyrakis, F., Cellini, B., Bruno, S., Benedetti, P., Carosati, E., Cruciani, G., Micheli, F., 37 Felici, A., Cozzini, P., Kellogg, G. E., Voltattorni, C. B. and Mozzarelli, A. (2014) 38 'Targeting Cystalysin, a Virulence Factor of Treponema denticola-Supported 39 Periodontitis', Chemmedchem, 9(7), 1501-1511. 40

186

1 Stacy, A., Everett, J., Jorth, P., Trivedi, U., Rumbaugh, K. P. and Whiteley, M. (2014) 2 'Bacterial fight-and-flight responses enhance virulence in a polymicrobial infection', 3 Proceedings of the National Academy of Sciences of the United States of America, 4 111(21), 7819-7824. 5 6 Stacy, A., Fleming, D., Lamont, R., J.,, Rumbaugh, K., P., and Whiteley, M., Whiteley 7 (2016) 'A Commensal Bacterium Promotes Virulence of an Opportunistic Pathogen 8 via Cross-Respiration', mBio, 7(3). 9 10 Staniec, D., Ksiazek, M., Thøgersen, I. B., Enghild, J. J., Sroka, A., Bryzek, D., Bogyo, M., 11 Abrahamson, M. and Potempa, J. (2015) 'Calcium Regulates the Activity and 12 Structural Stability of Tpr, a Bacterial Calpain-like Peptidase', The Journal Of 13 Biological Chemistry, 290(45), 27248-27260. 14 15 Stevenson, B. and Babb, K. (2002) 'LuxS-Mediated Quorum Sensing in Borrelia burgdorferi, 16 the Lyme Disease Spirochete', Infection and Immunity, 70(8), 4099-4105. 17 18 Stevenson, B., von Lackum, K., Wattier, R. L., McAlister, J. D., Miller, J. C. and Babb, K. 19 (2003) 'Review: Quorum sensing by the Lyme disease spirochete', Microbes and 20 Infection, 5, 991-997. 21 22 Stubbendieck, R. M. and Straight, P. D. (2016) 'Multifaceted interfaces of bacterial 23 competition', Journal of Bacteriology, (15-16), 2145. 24 25 Sugawara, S., Nemoto, E., Tada, H., Miyake, K., Imamura, T. and Takada, H. (2000a) 26 'Proteolysis of human monocyte CD14 by cysteine proteinases (Gingipains) from 27 Porphyromonas gingivalis leading to lipopolysaccharide hyporesponsiveness', 28 Journal of Immunology, 165(1), 411-418. 29 30 Suntharalingam, P. and Cvitkovitch, D. G. (2005) 'Quorum sensing in streptococcal biofilm 31 formation', Trends in Microbiology, 13(1), 3-6. 32 33 Szafrański, S. P., Deng, Z.-L., Tomasch, J., Jarek, M., Bhuju, S., Meisinger, C., Kühnisch, J., 34 Sztajer, H. and Wagner-Döbler, I. (2015) 'Functional biomarkers for chronic 35 periodontitis and insights into the roles of Prevotella nigrescens and Fusobacterium 36 nucleatum; a metatranscriptome analysis', NPJ Biofilms And Microbiomes, 1, 15017- 37 15017. 38 39 Sztajer, H., Szafranski, S. P., Tomasch, J., Reck, M., Nimtz, M., Rohde, M. and Wagner- 40 Dobler, I. (2014) 'Cross-feeding and interkingdom communication in dual-species 41 biofilms of Streptococcus mutans and Candida albicans', Isme Journal, 8(11), 2256- 42 2271.

187

1 2 Taba, M., Kinney, J., Kim, A. S. and Giannobile, W. V. (2005) 'Diagnostic Biomarkers for 3 Oral and Periodontal Diseases', Dental Clinics of North America, 49(3), 551-572. 4 5 Takahashi, N. (2005) 'Microbial ecosystem in the oral cavity: Metabolic diversity in an 6 ecological niche and its relationship with oral diseases', International Congress 7 Series, 1284, 103-112. 8 9 Takahashi, N., Kato, T. and Kuramitsu, H. K. (1991) 'ISOLATION AND PRELIMINARY 10 CHARACTERISATION OF THE PORPHYROMONAS-GINGIVALIS-PRTC 11 GENE EXPRESSING COLLAGENASE ACTIVITY', Fems Microbiology Letters, 12 84(2), 135-138. 13 14 Takahashi, N., Saito, K., Schachtele, C. F. and Yamada, T. (1997) 'Acid tolerance and acid- 15 neutralizing activity of Porphyromonas gingivalis, Prevotella intermedia and 16 Fusobacterium nucleatum', Oral Microbiology and Immunology, 12(6), 323-328. 17 18 Takahashi, N. and Sato, T. (2001) 'Preferential utilisation of dipeptides by Porphyromonas 19 gingivalis', Journal of Dental Research, 80(5), 1425-1429. 20 21 Takahashi, N. and Sato, T. (2002) 'Dipeptide utilisation by the periodontal pathogens 22 Porphyromonas gingivalis, Prevotella intermedia, Prevotella nigrescens and 23 Fusobacterium nucleatum', Oral Microbiology and Immunology, 17(1), 50-54. 24 25 Takahashi, N., Sato, T. and Yamada, T. (2000) 'Metabolic pathways for cytotoxic end 26 product formation from glutamate- and aspartate-containing peptides by 27 Porphyromonas gingivalis', Journal of Bacteriology, 182(17), 4704-4710. 28 29 Tan, K. H. (2012) Interactions between Porphyromonas gingivalis and Treponema denticola, 30 unpublished thesis (Doctor of Philosophy), University of Melbourne. 31 32 Tan, K. H., Seers, C. A., Dashper, S. G., Mitchell, H. L., Pyke, J. S., Meuric, V., Slakeski, N., 33 Cleal, S. M., Chambers, J. L., McConville, M. J. and Reynolds, E. C. (2014) 34 'Porphyromonas gingivalis and Treponema denticola Exhibit Metabolic Symbioses', 35 Plos Pathogens, 10(3), 11. 36 37 Tokuda, M., Chen, W., Karunakaran, T. and Kuramitsu, H. K. (1998) 'Regulation of protease 38 expression in Porphyromonas gingivalis', Infection and Immunity, 66(11), 5232-5237. 39 40 Tokuda, M., Duncan, M., Cho, M. I. and Kuramitsu, H. K. (1996) 'Role of Porphyromonas 41 gingivalis protease activity in colonization of oral surfaces', Infection and Immunity, 42 64(10), 4067-4073.

188

1 2 Tsilibaris, V., Maenhaut-Michel, G. and Van Melderen, L. (2006) 'Mini-review: Biological 3 roles of the Lon ATP-dependent protease', Research in Microbiology, 157, 701-713. 4 5 Tukey, J. W. (1949) 'Comparing Individual Means in the Analysis of Variance', Biometrics, 6 (2), 99. 7 8 Umemoto, T., Jinno, T., Taiji, Y. and Ogawa, T. (2001) 'Chemotaxis of oral treponemes 9 toward sera and albumin of rabbit', Microbiology and Immunology, 45(8), 571-7. 10 11 van der Ploeg, J. R. (2005) 'Regulation of bacteriocin production in Streptococcus mutans by 12 the quorum-sensing system required for development of genetic competence', Journal 13 of Bacteriology, 187(12), 3980-3989. 14 15 Van Dyke, T. E. and Dave, S. (2005) 'Risk Factors for Periodontitis', International Academy 16 of Periodontology, 7(1), 3-7. 17 18 Veith, P. D., Chen, Y.-Y., Gorasia, D. G., Chen, D., Glew, M. D., O'Brien-Simpson, N. M., 19 Cecil, J. D., Holden, J. A. and Reynolds, E. C. (2014) 'Porphyromonas gingivalis 20 outer membrane vesicles exclusively contain outer membrane and periplasmic 21 proteins and carry a cargo enriched with virulence factors', Journal of Proteome 22 Research, 13(5), 2420-2432. 23 24 Veith, P. D., Dashper, S. G., O'Brien-Simpson, N. M., Paolini, R. A., Orth, R., Walsh, K. A. 25 and Reynolds, E. C. (2009) 'Major proteins and antigens of Treponema denticola', 26 Biochimica Et Biophysica Acta-Proteins and Proteomics, 1794(10), 1421-1432. 27 28 Veith, P. D., Luong, C., Tan, K. H., Dashper, S. G. and Reynolds, E. C. (2018) 'Outer 29 Membrane Vesicle Proteome of Porphyromonas gingivalis Is Differentially 30 Modulated Relative to the Outer Membrane in Response to Heme Availability', 31 Journal of Proteome Research. 32 33 Veith, P. D., Talbo, G. H., Slakeski, N., Dashper, S. G., Moore, C., Paolini, R. A. and 34 Reynolds, E. C. (2002) 'Major outer membrane proteins and proteolytic processing of 35 RgpA and Kgp of Porphyromonas gingivalis W50', Biochemical Journal, 363, 105- 36 115. 37 38 Vercellino, I., Rezabkova, L., Olieric, V., Polyhach, Y., Weinert, T., Kammerer, R. A., 39 Jeschke, G. and Korkhov, V. M. (2017) 'Role of the nucleotidyl cyclase helical 40 domain in catalytically active dimer formation', Proceedings of the National Academy 41 of Sciences of the United States, (46), 9821.

189

1 2 Vesey, P. M. and Kuramitsu, H. K. (2004) 'Genetic analysis of Treponema denticola ATCC 3 35405 biofilm formation', Microbiology-Sgm, 150, 2401-2407. 4 5 Visser, M. B. and Ellen, R. P. (2011) 'Review: New insights into the emerging role of oral 6 spirochaetes in periodontal disease', Clinical Microbiology and Infection, 17, 502- 7 512. 8 9 Wai-Leung, N. and Bassler, B. L. (2009) 'Bacterial quorum-sensing network architectures', 10 Annual Review of Genetics, 197. 11 12 Waller, T., Kesper, L., Hirschfeld, J., Koelpin, J., Uebele, J., Hoerauf, A., Bekeredjian-Ding, 13 I., Dommisch, H., Jepsen, S., Oldenburg, J. and Deschner, J. (2016) 'Porphyromonas 14 gingivalis Outer Membrane Vesicles Induce Selective Tumor Necrosis Factor 15 Tolerance in a Toll-Like Receptor 4-and mTOR-Dependent Manner', Infection and 16 Immunity, 84(4), 1194-1204. 17 18 Wang, B. Y., Alvarez, P., Hong, J. and Kuramitsu, H. K. (2011) 'Periodontal pathogens 19 interfere with quorum-sensing-dependent virulence properties in Streptococcus 20 mutans', Journal of Periodontal Research, 46(1), 105-110. 21 22 Wang, B. Y., Wu, J., Lamont, R. J., Lin, X. H. and Xie, H. (2009) 'Negative Correlation of 23 Distributions of Streptococcus cristatus and Porphyromonas gingivalis in Subgingival 24 Plaque', Journal of Clinical Microbiology, 47(12), 3902-3906. 25 26 Wang J., Qi, J., Zhao, H., He, S., Zhang, Y., Wei, S. and Zhao, F. (2013) 'Metagenomic 27 sequencing reveals microbiota and its functional potential associated with periodontal 28 disease', Scientific Reports, 3(1843). 29 30 Whittaker, C. J., Klier, C. M. and Kolenbrander, P. E. (1996) 'Mechanisms of adhesion by 31 oral bacteria', Annual Review of Microbiology, 50, 513-552. 32 33 Wilensky, A., Tzach-Nahman, R., Potempa, J., Shapira, L. and Nussbaum, G. (2015) 34 'Porphyromonas gingivalis Gingipains Selectively Reduce CD14 Expression, Leading 35 to Macrophage Hyporesponsiveness to Bacterial Infection', Journal of Innate 36 Immunity, 7(2), 127-135. 37 38 Wilkens, S. (2015) 'Structure and mechanism of ABC transporters', F1000prime Reports, 7, 39 14-14. 40

190

1 William H. Kruskal, a. and W. Allen Wallis, a. (1952) 'Use of Ranks in One-Criterion 2 Variance Analysis', Journal of the American Statistical Association, (260), 583. 3 4 Wright, C. J., Wu, H., Melander, R. J., Melander, C. and Lamont, J. (2014) 'Disruption of 5 heterotypic community development by Porphyromonas gingivalis with small 6 molecule inhibitors', Molecular Oral Microbiology, 29(5), 185-193. 7 Wu, J. and Xie, H. (2010) 'Role of Arginine Deiminase of Streptococcus cristatus in 8 Porphyromonas gingivalis Colonization', Antimicrobial Agents and Chemotherapy, 9 54(11), 4694-4698. 10 Wyss, C. (1992) 'GROWTH OF PORPHYROMONAS-GINGIVALIS, TREPONEMA- 11 DENTICOLA, T-PECTINOVORUM, T-SOCRANSKII, AND T-VINCENTII IN A 12 CHEMICALLY DEFINED MEDIUM', Journal of Clinical Microbiology, 30(9), 13 2225-2229. 14 15 Xie, H. (2015) 'Biogenesis and function of Porphyromonas gingivalis outer membrane 16 vesicles', Future Microbiology, 10(9), - 1527. 17 18 Xu, Y., Lee, M., Moeller, A., Song, S., Yoon, B.-Y., Kim, H.-M., Jun, S.-Y., Lee, K. and Ha, 19 N.-C. (2011) 'Funnel-like Hexameric Assembly of the Periplasmic Adapter Protein in 20 the Tripartite Multidrug Efflux Pump in Gram-negative Bacteria', Journal of 21 Biological Chemistry, 286(20), 17910-17920. 22 23 Yaegaki, K. (2008) 'Oral malodorous compounds are periodontally pathogenic and 24 carcinogenic', Japanese Dental Science Review, 44, 100-108. 25 26 Yahashiri, A., Weiss, D. S. and Jorgenson, M. A. (2017) 'The SPOR Domain, a Widely 27 Conserved Peptidoglycan Binding Domain That Targets Proteins to the Site of Cell 28 Division', Journal of Bacteriology, 199(14), 10p. 29 30 Yamada, M., Ikegami, A. and Kuramitsu, H. K. (2005) 'Synergistic biofilm formation by 31 Treponema denticola and Porphyromonas gingivalis', Fems Microbiology Letters, 32 250(2), 271-277. 33 34 Yamamoto, R., Noiri, Y., Yamaguchi, M., Asahi, Y., Maezono, H. and Ebisu, S. (2011) 35 'Time course of gene expression during Porphyromonas gingivalis strain ATCC 36 33277 biofilm formation', Applied and Environmental Microbiology, (18), 6733. 37 38 Yanamandra, S. S., Sarrafee, S. S., Anaya-Bergman, C., Jones, K. and Lewis, J. P. (2012) 39 'Role of the Porphyromonas gingivalis extracytoplasmic function sigma factor, SigH', 40 Molecular Oral Microbiology, 27(3), 202-219.

191

1 2 Yang, J., Yan, R., Roy, A., Xu, D., Poisson, J. and Zhang, Y. (2015) 'The I-TASSER Suite: 3 protein structure and function prediction', Nature Methods, 12(1), 7-8. 4 5 Yasaki-Inagaki, Y., Inagaki, S., Yamada, S., Okuda, K. and Ishihara, K. (2006) 'Production 6 of protective antibodies against Porphyromonas gingivalis strains by immunization 7 with recombinant gingipain domains', Fems Immunology and Medical Microbiology, 8 47(2), 287-295. 9 10 Yesudhas, D., Batool, M., Anwar, M. A., Panneerselvam, S. and Choi, S. (2017) 'Proteins 11 Recognizing DNA: Structural Uniqueness and Versatility of DNA-Binding Domains 12 in Stem Cell Transcription Factors', 8. 13 14 Yokoyama, K., Ishijima, S. A., Clowney, L., Koike, H., Aramaki, H., Tanaka, C., Makino, K. 15 and Suzuki, M. (2006) 'Feast/famine regulatory proteins (FFRPs): Escherichia coli 16 Lrp, AsnC and related archaeal transcription factors', FEMS Microbiology Reviews, 17 30(1), 89-108. 18 19 Yoshida, Y., Sato, M., Kezuka, Y., Hasegawa, Y., Nagano, K., Takebe, J. and Yoshimura, F. 20 (2016) 'Acyl-CoA reductase PGN_0723 utilises succinyl-CoA to generate succinate 21 semialdehyde in a butyrate-producing pathway of Porphyromonas gingivalis', 22 Archives of Biochemistry and Biophysics, 596, 138-148. 23 24 Yoshida, Y., Sato, M., Nagano, K., Hasegawa, Y., Okamoto, T. and Yoshimura, F. (2015) 25 'Production of 4-hydroxybutyrate from succinate semialdehyde in butyrate 26 biosynthesis in Porphyromonas gingivalis', BBA - General Subjects, 1850, 2582- 27 2591. 28 29 Yousef, A. A. (2014) 'Risk Factors of Periodontal Disease: Review of the Literature', 30 International Journal of Dentistry, 2014(182513). 31 Yu, N. Y., Wagner, J. R., Laird, M. R., Melli, G., Rey, S., Lo, R., Dao, P., Sahinalp, S. C., 32 Ester, M. and Foster, L. J. (2010) 'PSORTb 3.0: improved protein subcellular 33 localisation prediction with refined localisation subcategories and predictive 34 capabilities for all prokaryotes', Bioinformatics, 26(13), 1608-1615. 35 36 Yuan, L. H., Hillman, J. D. and Progulske-Fox, A. (2005) 'Microarray analysis of quorum- 37 sensing-regulated genes in Porphyromonas gingivalis', Infection and Immunity, 73(7), 38 4146-4154. 39 40 Yung-Hua, L., Xingxing, H. and Xiao-Lin, T. (2017) 'Recent advances in dental biofilm: 41 impacts of microbial interactions on the biofilm ecology and pathogenesis', AIMS 42 Bioengineering, 4(3),335-350.

192

1 2 Zainal-Abidin, Z., Veith, P. D., Dashper, S. G., Zhu, Y., Catmull, D. V., Chen, Y.-Y., 3 Heryanto, D. C., Chen, D., Pyke, J. S., Tan, K., Mitchell, H. L. and Reynolds, E. C. 4 (2012) 'Differential Proteomic Analysis of a Polymicrobial Biofilm', Journal of 5 Proteome Research, 11(9), 4449-4464. 6 7 Zhang, L., Butler, C. A., Khan, H. S. G., Dashper, S. G., Seers, C. A., Veith, P. D., Zhang, J.- 8 G. and Reynolds, E. C. (2016) 'Characterisation of the Porphyromonas gingivalis 9 Manganese Transport Regulator Orthologue', Plos One, 11(3), 1-22. 10 11 Zhang, Y. (2009) 'I-TASSER: Fully automated protein structure prediction in CASP8', 12 Proteins, 77(SUPP/9), 100-113. 13 14 15 Zhu, Y., Dashper, S. G., Chen, Y. Y., Crawford, S., Slakeski, N. and Reynolds, E. C. (2013) 16 'Porphyromonas gingivalis and Treponema denticola Synergistic Polymicrobial 17 Biofilm Development', Plos One, 8(8), 8.

193

Appendices

Appendix I

Figure I.1: Linear regression of all data points collected for different experimental set-up for OBGM, OB:CM and OB:PBS.

2.8 2.6 2.4 2.2 y = 0.2824x + 0.0342 2.0 1.8

1.6 y = 0.5059x + 0.2487 1.4 1.2

1.0 [Glycine] (mM) [Glycine]

Δ 0.8 0.6 0.4

0.2 y = 0.1888x - 0.0242 0.0 -0.2 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 9 Cell number (10 ) Figure I.2: Linear regression of change in free glycine concentration as a function of P. gingivalis cell number in OBGM (black dots), OB:CM (red dots) and OB:PBS (grey dots). A linear regression trendline and equation that indicates the change of free glycine concentration 9 against 10 P. gingivalis cells as calculated from the OD650 derived equation were included.

194

Table I.1: Relative peak abundance of amino acids determined at initial time point and after P. gingivalis 48 h growth in OB:CM and OBGM. Most essential amino acids were included in the analysis, except arginine, alanine, valine, cysteine and methionine. Homeserine and threonine showed coelution using the methodology and hence the relative abundance determined corresponded to both metabolites.

4- Homoserine Phenyl- Labels Histidine Asparagine Aspartate Serine Glutamine Glutamate Proline Lysine2+ Lysine Tyrosine Methionine Isoleucine Hydroxyproline/ Tryptophan / Threonine alanine Leucine

OB:CM

0 6.07E+04 9.65E+06 4.93E+05 1.13E+07 5.49E+05 2.18E+07 3.61E+07 2.49E+07 3.42E+07 1.31E+06 7.48E+06 8.18E+06 5.13E+07 2.10E+08 4.10E+07 3.88E+06

48 5.51E+04 7.90E+06 4.14E+05 8.17E+06 1.01E+06 2.10E+07 2.71E+07 4.42E+07 3.35E+07 1.15E+06 8.04E+06 7.13E+06 5.23E+07 1.97E+08 3.38E+07 3.42E+06

OBGM

0 1.40E+05 1.88E+07 1.02E+06 1.95E+07 4.00E+05 2.28E+07 3.57E+07 7.96E+06 3.10E+07 1.13E+06 6.66E+06 1.02E+07 4.21E+07 1.90E+08 3.66E+07 5.89E+06

48 8.79E+04 1.40E+07 7.51E+05 8.62E+06 8.77E+05 1.85E+07 1.68E+07 3.86E+07 2.55E+07 8.89E+05 6.40E+06 6.90E+06 3.25E+07 1.40E+08 2.44E+07 4.13E+06

Table I.2: Relative peak abundance of other amine groups determined at initial time point and after P. gingivalis 48 h growth in OB:CM and OBGM. 2-aminobutyric acid/GABA and PABA/Tyramine showed co-elution and hence the relative abundance determined corresponded to both metabolites.

2-Aminobutyric acid Dihydroxy- PABA/ Labels Taurine Agmatine Citrulline Glutathione Ornithine2+ Cystathionine2+ Putrescine2+ Cadaverine2+ Spermidine3+ / GABA phenylalanine Tyramine OB:CM 0 2.70E+04 4.52E+05 8.90E+05 4.84E+04 1.64E+06 1.17E+06 1.51E+04 1.08E+05 1.17E+07 3.94E+05 9.87E+03 1.13E+05 48 2.30E+04 5.73E+05 2.14E+06 7.88E+03 1.49E+06 5.04E+06 1.54E+04 1.12E+05 1.03E+07 3.58E+05 8.81E+03 9.60E+04 OBGM 0 2.72E+04 4.90E+05 1.27E+05 9.60E+04 1.73E+06 1.76E+06 1.13E+04 1.22E+05 8.41E+04 3.76E+04 6.35E+03 1.08E+05 48 1.98E+04 4.59E+05 2.15E+06 7.01E+04 1.33E+06 7.40E+06 1.19E+04 1.13E+05 4.58E+04 2.62E+04 5.59E+03 7.25E+04

195

Appendix II Table II.1: Common nucleotide polymorphisms of P. gingivalis peptidase mutant strains (PG0753-, PG1605- and PG1788-) determined by whole genome sequencing analysis.

Variant Amino Nucleotide Variant P-Value Acid CDS Locus_tag Position Change Coverage Polymorphism Type Strand-Bias Frequency (approximate) Change Position Codon Change Protein Effect protein_id CF003_0196 235296 (C)3 -> (C)2 124 Deletion (tandem repeat) 52.90% 82.30% 7.6E-59 2254 Frame Shift AUR45513.1 CF003_0217 252997 A -> C 108 SNP (transversion) 57.00% 99.10% 1.0E-321 Y -> S 368 TAT -> TCT Substitution AUR46393.1 CF003_0224 258048 C -> T 119 SNP (transition) 50.80% 99.20% 1.0E-354 V -> I 295 GTA -> ATA Substitution AUR46707.1 CF003_0279 312732 G -> C 90 SNP (transversion) 58.40% 98.90% 1.3E-276 H -> D 316 CAT -> GAT Substitution AUR45581.1 CF003_0373 405325 T -> A 95 SNP (transversion) 55.30% 98.90% 1E-282 S -> T 481 TCG -> ACG Substitution AUR46669.1 CF003_0373 405342 T -> G 97 SNP (transversion) 56.30% 99.00% 4E-279 D -> E 498 GAT -> GAG Substitution AUR46669.1 CF003_0375 406364 T -> G 119 SNP (transversion) 57.60% 99.20% 1.6E-366 F -> L 291 TTT -> TTG Substitution AUR46955.1 CF003_0429 467147 T -> C 103 SNP (transition) 51.00% 99.00% 2.5E-286 V -> A 908 GTA -> GCA Substitution AUR45677.1 CF003_0479 519971 -G 52 Deletion 53.10% 94.20% 4E-39 125 Frame Shift AUR46547.1 CF003_0488 528619 G -> C 97 SNP (transversion) 50.00% 99.00% 1E-288 S -> T 332 AGT -> ACT Substitution AUR46206.1 CF003_1180 1152559 C -> A 90 SNP (transversion) 53.90% 98.90% 4E-232 H -> N 604 CAC -> AAC Substitution AUR45563.1 CF003_1180 1152853 A -> G 110 SNP (transition) 60.60% 99.10% 1.0E-327 T -> A 898 ACC -> GCC Substitution AUR45563.1 CF003_1117 1224166 A -> G 94 SNP (transition) 61.70% 100.00% 6.3E-264 L -> S 1241 TTG -> TCG Substitution AUR45903.1 CF003_0999 1358392 G -> C 76 SNP (transversion) 52.00% 98.70% 1E-225 T -> R 101 ACA -> AGA Substitution AUR46854.1 CF003_0987 1367462 T -> C 81 SNP (transition) 51.20% 98.80% 1E-248 T -> A 58 ACC -> GCC Substitution AUR46525.1 CF003_0827 1531351 A -> T 96 SNP (transversion) 54.20% 100.00% 6.3E-20 I -> N 5 ATC -> AAC Substitution AUR45890.1 CF003_1548 1626078 C -> G 124 SNP (transversion) 50.40% 99.20% 1.9E-232 R -> G 1099 CGC -> GGC Substitution AUR45552.1 CF003_1844 1944280 +CCC 122 Insertion 52.10% 99.20% 1.1E-57 N -> NP 4461 AAT -> AAC,CCT Insertion AUR45469.1 CF003_2168 2274529 C -> A 119 SNP (transversion) 58.50% 99.20% 2.5E-378 C -> F 278 TGT -> TTT Substitution AUR46755.1 CF003_2168 2274773 +A 97 Insertion 53.60% 100.00% 2.5E-175 35 Frame Shift AUR46755.1 CF003_2174 2282858 T -> C 129 SNP (transition) 53.90% 99.20% 6.3E-372 R -> G 334 AGG -> GGG Substitution AUR45705.1

196

CF003_2185 2293042 C -> T 121 SNP (transition) 51.70% 99.20% 7.6E-215 S -> N 821 AGT -> AAT Substitution AUR45938.1 CF003_2219 2334168 G -> C 73 SNP (transversion) 54.20% 98.60% 9.1E-136 A -> P 559 GCA -> CCA Substitution AUR45816.1

197

Figure II.1: Linear regression of change in free glycine concentration as a function of P. gingivalis cell number in OB:CM from the time interval of 30 to 40 h of growth. P. gingivalis W50 – red cross symbol, PG0445- – brown circle symbol, PG0753- – green triangular symbol, PG1605- – black square symbol and PG1788- – blue plus symbol with each dot point of lower and higher cell density representing a biological replicate. A linear regression trendline that indicates the change of free glycine concentration against 109 P. gingivalis cells as calculated from the OD650 derived equation was included.

198

Appendix III Table III.1: Details of P. gingivalis differential gene expression profile during growth in OB:CM relative to OBGM, which showed protein- encoding genes with an at least 1.5-fold difference in their expression levels and false discovery rate (FDR) <0.05. Each condition was represented by biological triplicates. “CM” and “OB” represent the number of reads for each sample in OB:CM and OBGM respectively. CPM represents counts-per-million for each sample normalised for the library size.

CM rel to CM1 CM2 CM3 OB1 OB2 OB3 Gene.Name OB (log2) FDR AveExpr P value CM1 CM2 CM3 OB1 OB2 OB3 CPM CPM CPM CPM CPM CPM PG_RS02990 -2.6094 2.66E-06 6.052598 1.43E-09 578 647 545 3025 3532 1421 27.472 29.968 25.514 156.154 165.259 157.273 PG_RS03000 -2.59057 7.05E-06 5.800104 7.55E-09 469 589 445 2253 3311 1180 22.291 27.281 20.832 116.302 154.919 130.6 PG_RS02985 -2.11608 3.4E-05 5.689412 7.30E-08 504 650 493 1993 2632 806 23.955 30.107 23.079 102.881 123.149 89.206 PG_RS03010 -2.17843 3.4E-05 5.549921 6.24E-08 416 584 462 1640 2448 848 19.772 27.05 21.628 84.659 114.54 93.855 PG_RS03150 2.433918 6.04E-05 4.070224 1.94E-07 922 779 860 135 148 65 43.822 36.082 40.26 6.969 6.925 7.194 PG_RS03005 -2.15267 6.04E-05 4.155393 1.72E-07 168 210 183 709 913 275 7.985 9.727 8.567 36.599 42.718 30.436 PG_RS03145 2.208554 0.000495 8.6367 1.86E-06 18782 19438 18367 4563 3964 1189 892.69 900.332 859.831 235.547 185.472 131.596 PG_RS03015 -1.55335 0.000731 3.480998 3.13E-06 124 167 140 323 459 161 5.894 7.735 6.554 16.674 21.476 17.819 PG_RS00950 -1.6808 0.001644 4.081274 7.93E-06 230 260 148 545 692 253 10.932 12.043 6.928 28.134 32.378 28.001 PG_RS05320 0.971036 0.002383 9.43364 1.28E-05 21780 22892 19285 8838 10234 4563 1035.182 1060.314 902.806 456.228 478.84 505.022 PG_RS00290 1.314385 0.003099 3.575478 1.86E-05 424 414 416 139 190 53 20.152 19.176 19.475 7.175 8.89 5.866 PG_RS00300 1.513111 0.003099 3.105772 1.99E-05 299 359 284 79 105 57 14.211 16.628 13.295 4.078 4.913 6.309 PG_RS00295 1.303366 0.003279 4.15175 2.28E-05 586 702 543 166 270 113 27.852 32.515 25.42 8.569 12.633 12.507 PG_RS08065 0.985273 0.003688 6.424298 2.77E-05 2228 2902 2852 1058 1316 566 105.895 134.415 133.513 54.615 61.574 62.644 PG_RS00210 1.104411 0.004116 8.825869 3.31E-05 15787 14212 13793 5042 6349 3198 750.341 658.273 645.704 260.274 297.064 353.947 PG_RS00665 -1.08755 0.004653 5.502169 3.99E-05 577 864 627 1081 1477 628 27.424 40.019 29.352 55.802 69.108 69.506 PG_RS04955 0.790522 0.005085 6.128705 4.63E-05 1924 2284 1855 964 1124 481 91.446 105.791 86.84 49.763 52.591 53.236 PG_RS09040 -0.80369 0.006044 7.431236 5.83E-05 3115 3032 2489 3911 5311 1963 148.053 140.437 116.52 201.89 248.497 217.26 PG_RS03045 1.060976 0.006097 12.35124 6.53E-05 141423 167523 189963 67562 85880 28286 6721.698 7759.35 8892.911 3487.628 4018.248 3130.628

199

PG_RS04805 -1.32282 0.006097 3.472919 6.36E-05 184 155 128 369 360 138 8.745 7.179 5.992 19.048 16.844 15.274 PG_RS05130 -0.79669 0.006311 7.651399 7.44E-05 3657 3576 2859 4702 5763 2364 173.814 165.634 133.841 242.723 269.646 261.642 PG_RS00285 -0.86809 0.006311 5.111402 7.44E-05 549 651 498 813 1148 371 26.093 30.153 23.313 41.968 53.714 41.061 PG_RS04600 -1.07381 0.006342 7.093858 7.82E-05 2312 2170 1745 4012 3643 1841 109.887 100.51 81.69 207.104 170.453 203.758 PG_RS06865 -0.95108 0.00671 3.439185 8.63E-05 157 182 173 255 347 133 7.462 8.43 8.099 13.163 16.236 14.72 PG_RS07715 0.818791 0.006781 11.59387 9.08E-05 91571 98958 80909 39721 52339 20769 4352.281 4583.548 3787.667 2050.444 2448.895 2298.664 PG_RS09670 0.804926 0.006813 9.38003 9.49E-05 20197 20027 17736 8461 12812 4083 959.944 927.613 830.291 436.767 599.462 451.897 PG_RS08250 0.718433 0.0073 8.675842 0.00011 10925 11243 12389 5575 7582 2639 519.255 520.755 579.978 287.788 354.755 292.078 PG_RS09385 0.615393 0.0073 7.880003 0.000107 6281 6571 6342 3331 4440 1601 298.53 304.356 296.894 171.95 207.744 177.195 PG_RS07730 -1.11587 0.00734 9.477808 0.000114 8900 12446 11001 22412 18057 9670 423.008 576.475 515 1156.933 844.871 1070.253 PG_RS00370 -1.02201 0.007441 5.56998 0.000123 828 809 583 1396 1376 554 39.354 37.471 27.293 72.063 64.382 61.315 PG_RS09710 -1.25538 0.007441 3.396351 0.000125 128 187 139 362 323 125 6.084 8.661 6.507 18.687 15.113 13.835 PG_RS05625 0.813084 0.007441 5.680041 0.000128 1499 1580 1434 765 919 275 71.246 73.183 67.131 39.49 42.999 30.436 PG_RS06705 -0.79673 0.008266 3.982761 0.000146 252 274 262 356 481 181 11.977 12.691 12.265 18.377 22.506 20.033 PG_RS05815 -1.4404 0.008456 9.357103 0.000154 11946 8770 6421 22654 19537 9695 567.782 406.21 300.592 1169.425 914.119 1073.02 PG_RS07540 0.639591 0.008805 5.897309 0.000167 1561 1790 1554 817 1118 409 74.193 82.909 72.749 42.174 52.31 45.267 PG_RS00115 -1.12782 0.008805 4.334169 0.000177 307 382 234 587 723 208 14.591 17.694 10.954 30.302 33.829 23.021 PG_RS04675 0.782179 0.008805 5.33453 0.000184 1114 1155 1195 486 705 296 52.947 53.497 55.943 25.088 32.986 32.761 PG_RS02515 0.713201 0.008805 5.768913 0.000179 1538 1605 1478 842 967 320 73.1 74.341 69.191 43.465 45.245 35.417 PG_RS08850 -1.44762 0.008805 5.309481 0.000184 642 573 406 1608 1393 420 30.514 26.54 19.006 83.007 65.177 46.485 PG_RS02935 0.676242 0.008895 8.646676 0.000191 10941 11679 10742 5528 8029 2468 520.015 540.949 502.875 285.362 375.67 273.152 PG_RS01590 -0.73588 0.009194 8.949421 0.000211 8785 9159 7441 11707 12536 6028 417.543 424.228 348.342 604.329 586.548 667.165 PG_RS02735 -1.01439 0.009194 10.67653 0.000221 30842 24272 21715 44582 44791 21489 1465.89 1124.233 1016.564 2301.374 2095.731 2378.352 PG_RS03650 -1.23547 0.009194 9.140818 0.000217 7694 7679 9014 20777 13711 7739 365.688 355.677 421.981 1072.533 641.525 856.534 PG_RS01595 -0.78821 0.009194 5.890796 0.000208 935 1159 882 1341 1592 774 44.44 53.683 41.29 69.224 74.488 85.664 PG_RS03960 -0.74723 0.009765 8.172091 0.000251 5534 5190 4051 6719 8689 3075 263.026 240.391 189.643 346.842 406.55 340.334

200

PG_RS01285 -0.74525 0.009765 5.860303 0.000267 1006 1050 912 1545 1690 553 47.814 48.634 42.694 79.755 79.074 61.205 PG_RS03040 0.793605 0.009765 4.709557 0.000262 730 730 814 329 485 167 34.696 33.812 38.107 16.983 22.693 18.483 PG_RS06080 0.636867 0.009765 8.341117 0.000258 8924 8962 8754 4943 6163 1978 424.149 415.103 409.809 255.163 288.361 218.92 PG_RS07225 -1.43655 0.009765 10.12421 0.000241 19929 17177 9650 36061 37553 15679 947.206 795.606 451.754 1861.51 1757.071 1735.315 PG_RS01290 -0.68288 0.009813 8.417288 0.000279 5600 6746 5472 8578 10140 3221 266.163 312.462 256.166 442.806 474.441 356.493 PG_RS01560 0.614421 0.009813 9.001097 0.000273 13018 13853 14891 7679 9327 3414 618.733 641.645 697.106 396.399 436.402 377.853 PG_RS02740 -0.98867 0.010161 5.946142 0.000304 993 952 946 2048 1765 615 47.196 44.095 44.286 105.72 82.583 68.067 PG_RS08860 0.595596 0.010161 5.540363 0.000305 1213 1317 1245 689 873 307 57.653 61.001 58.283 35.567 40.847 33.978 PG_RS02305 0.594883 0.010546 10.61446 0.000336 41483 45601 39915 22974 28456 10916 1971.647 2112.152 1868.577 1185.944 1331.431 1208.157 PG_RS00360 -0.95937 0.010546 9.802824 0.000334 15393 14656 12658 22386 21886 13171 731.614 678.838 592.57 1155.591 1024.026 1457.735 PG_RS06450 -0.98763 0.010546 7.834588 0.000328 3222 3524 3972 7478 5825 2612 153.139 163.225 185.945 386.023 272.547 289.09 PG_RS06095 0.645727 0.011139 10.37402 0.000364 37126 36372 35397 19001 25745 8478 1764.563 1684.683 1657.072 980.853 1204.585 938.325 PG_RS10535 -1.43749 0.011546 3.648825 0.000384 247 174 110 425 453 153 11.74 8.059 5.15 21.939 21.195 16.934 PG_RS02940 0.660609 0.012063 10.6387 0.000424 43541 43853 44398 25293 26540 10473 2069.462 2031.188 2078.444 1305.654 1241.783 1159.127 PG_RS00365 -1.11312 0.012063 7.342524 0.000431 2932 2650 1839 5489 5087 1661 139.355 122.743 86.091 283.348 238.016 183.836 PG_RS02540 0.633075 0.012063 5.497167 0.000431 1140 1352 1239 685 856 274 54.183 62.622 58.002 35.36 40.051 30.326 PG_RS10595 -2.13024 0.012063 -1.64519 0.000433 2 3 4 12 17 4 0.095 0.139 0.187 0.619 0.795 0.443 PG_RS01045 0.588434 0.012146 6.531543 0.000443 2366 2799 2334 1265 1855 626 112.454 129.644 109.264 65.301 86.794 69.284 PG_RS09230 0.715233 0.013216 10.0439 0.000489 28477 28778 31435 14374 21342 6355 1353.484 1332.943 1471.595 742.002 998.573 703.356 PG_RS00385 -0.60238 0.013817 9.188544 0.000518 10705 10304 10245 12510 14880 6807 508.798 477.262 479.608 645.781 696.222 753.383 PG_RS02595 0.7621 0.014471 8.136476 0.00058 8557 7644 7886 3182 5203 2109 406.706 354.056 369.174 164.258 243.444 233.419 PG_RS00220 -1.03596 0.014471 11.18839 0.000582 44271 36008 29123 62986 59621 33628 2104.158 1667.823 1363.361 3251.41 2789.613 3721.868 PG_RS00375 -0.83644 0.014471 6.208737 0.000565 1309 1469 933 2007 2119 777 62.216 68.041 43.677 103.604 99.146 85.997 PG_RS06575 -0.61991 0.015024 10.36251 0.00062 21947 23865 23950 32855 35353 12969 1043.12 1105.382 1121.193 1696.013 1654.135 1435.378 PG_RS03030 0.727124 0.015321 9.014731 0.000649 14029 14373 15363 7867 10069 2832 666.785 665.73 719.202 406.104 471.119 313.439 PG_RS01925 -1.33077 0.016374 5.692975 0.000702 1032 824 431 1587 1803 661 49.05 38.166 20.177 81.923 84.361 73.158

201

PG_RS06805 -1.20757 0.017033 9.995921 0.000758 20161 13718 11881 30206 25737 16104 958.233 635.392 556.196 1559.268 1204.211 1782.353 PG_RS05505 -0.71325 0.017033 9.887207 0.000751 13862 17477 17811 23966 23081 10966 658.847 809.502 833.803 1237.152 1079.939 1213.691 PG_RS08090 -0.6132 0.017804 12.35978 0.000801 88013 101073 90649 115033 156296 52789 4183.173 4681.511 4243.634 5938.135 7312.949 5842.562 PG_RS06065 0.628027 0.01821 6.361905 0.00084 2087 2264 2423 1202 1659 491 99.193 104.864 113.43 62.049 77.623 54.343 PG_RS00670 -1.05501 0.018511 4.76348 0.000893 297 607 393 644 1017 312 14.116 28.115 18.398 33.244 47.585 34.531 PG_RS10565 -0.74221 0.018511 6.227517 0.000888 1204 1449 1190 2149 2116 679 57.225 67.115 55.709 110.934 99.006 75.15 PG_RS01930 -1.06185 0.018623 10.28735 0.000916 24728 21621 12842 32011 39201 16045 1175.298 1001.444 601.184 1652.445 1834.18 1775.823 PG_RS10570 -1.70419 0.018623 2.82305 0.000918 103 117 56 339 289 67 4.895 5.419 2.622 17.5 13.522 7.415 PG_RS04565 -0.79776 0.020033 10.57031 0.001041 30350 23795 22582 39840 38012 18158 1442.506 1102.14 1057.152 2056.586 1778.547 2009.685 PG_RS04605 -0.8936 0.020033 8.763529 0.001041 8933 7531 5075 11179 13182 4859 424.577 348.822 237.581 577.073 616.774 537.783 PG_RS01920 -1.25844 0.022142 10.50562 0.001194 30213 21134 13928 44398 40128 21627 1435.995 978.887 652.024 2291.876 1877.553 2393.625 PG_RS00270 -0.70201 0.022142 4.485864 0.00123 380 413 367 623 580 219 18.061 19.129 17.181 32.16 27.138 24.238 PG_RS08100 -1.1042 0.022142 5.830999 0.001211 1199 937 549 1504 1880 706 56.987 43.4 25.701 77.638 87.964 78.138 PG_RS06800 -1.13394 0.022142 5.020018 0.001246 663 470 369 1066 1134 303 31.512 21.77 17.274 55.028 53.059 33.535 PG_RS03920 -0.7952 0.022142 8.843318 0.001207 9468 7742 6138 11915 11712 5426 450.005 358.595 287.344 615.066 547.994 600.537 PG_RS04935 -0.61553 0.022142 7.074966 0.001216 2113 2974 2162 3286 3364 1450 100.429 137.75 101.212 169.627 157.399 160.483 PG_RS03735 0.700255 0.022183 5.268867 0.00126 931 1064 1257 593 618 260 44.25 49.282 58.845 30.611 28.916 28.776 PG_RS04995 0.70653 0.022554 9.276771 0.001293 18062 16342 17847 7363 10697 4981 858.469 756.931 835.488 380.087 500.503 551.285 PG_RS01870 -1.18551 0.023803 8.831166 0.001403 8897 7567 4252 14715 16252 4731 422.866 350.489 199.053 759.605 760.416 523.616 PG_RS08400 -0.63815 0.024503 9.513759 0.001458 10783 13550 14588 18207 18211 7836 512.506 627.61 682.921 939.866 852.076 867.27 PG_RS09680 0.614243 0.026306 5.233145 0.001593 889 1171 1028 486 774 254 42.253 54.239 48.125 25.088 36.215 28.112 PG_RS05635 0.600719 0.026823 10.1199 0.001667 29859 27583 32641 17784 20463 7042 1419.169 1277.593 1528.053 918.03 957.445 779.392 PG_RS03785 0.656608 0.029586 5.832026 0.001903 1486 1589 1690 906 1132 302 70.628 73.599 79.116 46.769 52.965 33.425 PG_RS03915 -1.14125 0.030962 10.53134 0.002091 25662 22626 18862 34760 34600 28953 1219.69 1047.994 883.004 1794.351 1618.903 3204.45 PG_RS01170 -0.77256 0.031472 8.269975 0.002151 6528 5181 4117 8457 7850 3395 310.27 239.974 192.733 436.56 367.294 375.751 PG_RS07215 -0.99328 0.031759 6.25307 0.002213 1470 1349 843 2574 2018 821 69.868 62.483 39.464 132.873 94.42 90.866

202

PG_RS00380 -0.61735 0.032624 5.812359 0.002308 1039 1133 845 1475 1505 516 49.383 52.478 39.558 76.141 70.418 57.11 PG_RS00430 -0.71655 0.033041 5.312087 0.002373 529 784 763 1081 1065 390 25.143 36.313 35.719 55.802 49.83 43.164 PG_RS02410 -0.63555 0.033949 4.785602 0.002502 481 585 416 670 833 243 22.861 27.096 19.475 34.586 38.975 26.895 PG_RS03020 -1.08582 0.033949 11.96123 0.002496 85637 62861 40380 121466 108345 50995 4070.244 2911.603 1890.346 6270.214 5069.365 5644.007 PG_RS08715 -0.79542 0.033949 6.786778 0.002565 1404 1962 2282 3068 3361 1013 66.731 90.876 106.829 158.374 157.258 112.116 PG_RS03660 -0.62672 0.034405 6.988193 0.002648 2063 2319 2371 3561 3276 1147 98.052 107.412 110.996 183.823 153.281 126.947 PG_RS04945 -0.90263 0.034405 8.126115 0.002662 5396 5233 3187 8573 8597 2601 256.467 242.383 149.196 442.548 402.246 287.873 PG_RS05170 0.717679 0.034405 6.345193 0.002637 2017 2500 2449 1302 1642 396 95.866 115.795 114.647 67.211 76.828 43.828 PG_RS00055 -0.79374 0.034627 10.82269 0.002709 37253 32202 23044 49026 44829 21085 1770.599 1491.536 1078.78 2530.778 2097.509 2333.638 PG_RS07220 -1.0954 0.035635 5.702998 0.002865 1030 940 503 1733 1744 491 48.955 43.539 23.547 89.459 81.6 54.343 PG_RS03655 -0.97683 0.035957 11.09533 0.002929 33319 31889 38891 56748 46748 36318 1583.62 1477.038 1820.64 2929.397 2187.297 4019.591 PG_RS04520 0.664244 0.036314 7.583101 0.003094 4814 5207 5816 2406 3016 1704 228.805 241.178 272.27 124.2 141.116 188.595 PG_RS05210 -0.7577 0.036314 3.561167 0.00303 190 264 160 271 390 114 9.031 12.228 7.49 13.989 18.248 12.617 PG_RS06855 -0.959 0.036314 2.549965 0.003078 79 125 81 128 235 59 3.755 5.79 3.792 6.608 10.995 6.53 PG_RS05035 -0.74112 0.036314 10.24379 0.003053 25829 22108 15227 29999 33660 13136 1227.627 1024.001 712.835 1548.583 1574.921 1453.862 PG_RS04785 -0.98012 0.037159 9.663323 0.003226 12331 12070 13284 32643 21725 7357 586.08 559.06 621.876 1685.069 1016.493 814.255 PG_RS09725 0.70467 0.037159 9.325959 0.003226 20025 17638 16548 7383 10751 5466 951.769 816.959 774.677 381.119 503.03 604.964 PG_RS05835 0.654087 0.039202 7.178087 0.003445 3772 3885 4404 2209 3098 757 179.28 179.946 206.168 114.031 144.953 83.783 PG_RS06815 -0.69883 0.040414 5.445086 0.003595 905 754 608 1087 1038 528 43.014 34.924 28.463 56.112 48.567 58.438 PG_RS03600 -0.83988 0.042057 5.301122 0.003832 628 615 709 1349 967 380 29.848 28.486 33.191 69.637 45.245 42.058 PG_RS00595 -0.64053 0.042981 3.749864 0.004054 209 311 202 320 369 137 9.934 14.405 9.456 16.519 17.265 15.163 PG_RS02310 0.885185 0.042981 9.483438 0.004016 20302 21829 22322 7908 9402 6637 964.934 1011.078 1044.98 408.22 439.911 734.568 PG_RS01865 -0.64273 0.043732 7.313976 0.004163 2962 2695 2693 3515 3568 2183 140.781 124.827 126.07 181.448 166.944 241.609 PG_RS01160 -0.8672 0.043732 11.6463 0.004239 64781 49578 44969 78686 71912 48175 3078.978 2296.36 2105.175 4061.862 3364.698 5331.896 PG_RS06490 -0.73467 0.043732 3.087271 0.004209 166 163 112 223 236 84 7.89 7.55 5.243 11.512 11.042 9.297 PG_RS09785 -1.0124 0.044009 7.820174 0.004379 4983 3562 2407 7580 6931 2153 236.837 164.985 112.681 391.288 324.295 238.289

203

PG_RS04260 0.616254 0.044499 4.427823 0.004483 559 627 596 299 485 118 26.569 29.041 27.901 15.435 22.693 13.06 PG_RS09270 0.613401 0.045221 4.345211 0.004653 603 578 501 288 442 114 28.66 26.772 23.454 14.867 20.681 12.617 PG_RS02510 -0.92467 0.045484 9.125978 0.004753 8003 8535 10695 13338 11885 9532 380.375 395.325 500.675 688.523 556.088 1054.979 PG_RS05010 -0.92981 0.045484 9.119445 0.004748 7515 8877 10747 12934 12103 9572 357.181 411.166 503.109 667.668 566.288 1059.406 PG_RS06220 0.602679 0.045618 8.060493 0.004837 6852 7919 7015 4101 5961 1411 325.669 366.793 328.4 211.698 278.91 156.166 PG_RS04980 -0.78857 0.045618 10.32766 0.004821 28421 20196 16989 37334 34721 12932 1350.823 935.441 795.321 1927.224 1624.564 1431.283 PG_RS06455 -0.68724 0.045618 11.49326 0.004882 44214 50004 56902 71642 62923 36736 2101.449 2316.091 2663.805 3698.242 2944.111 4065.854 PG_RS04795 -0.81244 0.047439 7.731772 0.005412 4841 3356 2677 5702 5097 2647 230.088 155.444 125.321 294.344 238.484 292.964 PG_RS06570 -0.60386 0.04885 12.67401 0.005602 109349 113304 127286 171863 149696 69693 5197.252 5248.028 5958.755 8871.765 7004.141 7713.457

204

Table III.2: Summary table of genes, metabolic pathways, which include the involvement of metabolic substrates and products that showed DGE during P. gingivalis growth in OB:CM relative to OBGM.

Locus tag Log2 Pathways of gene Reactions of gene chg PG_RS00375 -0.84 L-serine degradation L-serine → pyruvate + ammonium (sda) L-serine → 2-aminoprop-2-enoate + H+ + H2O PG_RS00380 -0.62 melibiose degradation an α-D-galactoside + H2O → α-D-galactose + a non galactosylated galactose acceptor p-nitrophenyl-α-D-galactopyranoside + H2O → D-galactopyranose + 4-nitrophenol + H+ melibiose + H2O → D-galactopyranose + D-glucopyranose raffinose + H2O → α-D-galactose + sucrose melibionate + H2O → α-D-galactose + D-gluconate stachyose + H2O → raffinose + α-D-galactose PG_RS01590 -0.74 UMP biosynthesis II L-aspartate + carbamoyl phosphate → N-carbamoyl-L-aspartate + phosphate + H+ PG_RS01595 -0.79 reduced riboflavin + NADP+ = riboflavin + NADPH + 2 H+ FMNH2 + NAD(P)+ ← FMN + NAD(P)H + 2 H+ FMNH2 + NAD+ = FMN + NADH + 2 H+ FMNH2 + NADP+ = FMN + NADPH + 2 H+ PG_RS02515 0.71 L-histidine biosynthesis L-histidinol-phosphate + 2-oxoglutarate ↔ imidazole acetol-phosphate + L-glutamate PG_RS03030 0.73 superpathway of purine acetaldehyde + coenzyme A + NAD+ ↔ acetyl-CoA + NADH + H+ deoxyribonucleosides degradation 2'-deoxy-α-D-ribose 1- 4-hydroxybutanoate + NAD+ ← succinate semialdehyde + NADH + H+ phosphate degradation pyruvate fermentation to methionol + NAD+ = 3-methylthiopropanal + NADH + H+ ethanol III a secondary alcohol + NAD+ ↔ a ketone + NADH + H+ 3,4-dihydroxyphenylglycol + NAD+ = 3,4-dihydroxyphenylglycolaldehyde + NADH + H+

205

2-methylbutanol + NAD+ = 2-methylbutanal + NADH + H+ isobutanol + NAD+ ↔ isobutanal + NADH + H+ 5-hydroxytryptophol + NAD+ = 5-hydroxyindole acetaldehyde + NADH + H+ a primary alcohol + NAD+ = an aldehyde + NADH + H+ ethanol + NAD+ ↔ acetaldehyde + NADH + H+ butan-1-ol + NAD+ ↔ butan-1-al + NADH + H+ 3-methylbutanol + NAD+ ↔ 3-methylbutanal + NADH + H+ indole-3-glycol + NAD+ = indole-3-glycol aldehyde + NADH + H+ phytol + NAD+ = phytenal + NADH + H+ 4-tyrosol + NAD+ = (4-hydroxyphenyl)acetaldehyde + NADH + H+ propan-1-ol + NAD+ = 1-propanal + NADH + H+ 3-methoxy-4-hydroxyphenylglycol + NAD+ ↔ 3-methoxy-4-hydroxyphenylglycolaldehyde + NADH + H+ 2-phenylethanol + NAD+ = phenylacetaldehyde + NADH + H+ PG_RS03045 1.06 4-hydroxybutanoyl-CoA = crotonyl-CoA + H2O vinylacetyl-CoA = crotonyl-CoA PG_RS03915 -1.14 superpathway of 6-(hydroxymethyl)-7,8-dihydropterin + ATP → (7,8-dihydropterin-6-yl)methyl diphosphate + AMP + H+ (PG0886) tetrahydrofolate biosynthesis 6-hydroxymethyl- 2-amino-6-[1-hydroxyethyl]-7-methyl-7,8-dihydropterin + ATP = [1-(2-amino-7-methyl-4-oxo-7,8- dihydropterin diphosphate dihydro-3H-pteridin-6-yl)]ethyl diphosphate + AMP + H+ biosynthesis I PG_RS03960 0.75 succinate to cytochrome 2 an ubiquinol[membrane] + oxygen[in] + 4 H+[in] → 2 a ubiquinone[membrane] + 4 H+[out] + 2 H2O[in] bd oxidase electron transfer 2 an ubiquinol[membrane] + oxygen[in] → 2 a ubiquinone[membrane] + 2 H2O[in] PG_RS04260 0.62 phosphatidylethanolamine a CDP-diacylglycerol + L-serine → CMP + a 3-O-sn-phosphatidyl-L-serine + H+ biosynthesis I

206

PG_RS04935 -0.62 β-alanine biosynthesis III L-aspartate + H+ → β-alanine + CO2 PG_RS04945 -0.90 N10-formyl- a 5,10-methylene-tetrahydrofolate + NADP+ ↔ a 5,10-methenyltetrahydrofolate + NADPH tetrahydrofolate biosynthesis formate assimilation into a 5,10-methenyltetrahydrofolate + H2O ↔ an N10-formyl-tetrahydrofolate + H+ 5,10- methylenetetrahydrofolate PG_RS04980 -0.79 adenosylcobalamin ATP + cobinamide → adenosylcobinamide + PPPi salvage from cobinamide II adenosylcobalamin ATP + cob(I)alamin → adenosylcobalamin + PPPi salvage from cobalamin ATP + cob(I)yrinate a,c-diamide → adenosyl-cobyrinate a,c-diamide + PPPi PG_RS05035 -0.74 thioredoxin pathway a reduced thioredoxin + NADP+ ← an oxidized thioredoxin + NADPH + H+ PG_RS05320 0.97 [2Fe-2S] iron-sulfur a [co-chaperone]-[scaffold protein-(2Fe-2S)] complex + an [Fe-S cluster biosynthesis chaperone]-ATP → cluster biosynthesis a [chaperone-ATP]-[co-chaperone]-[scaffold protein-(2Fe-2S)] complex a [chaperone-ATP]-[co-chaperone]-[scaffold protein-(2Fe-2S)] complex + an apo-iron-sulfur protein → a [chaperone-ADP]-[disordered-form scaffold protein] complex + an [Fe-S cluster biosynthesis co- chaperone] + an [2Fe-2S] cluster protein + phosphate a [chaperone-ADP]-[disordered-form scaffold protein] complex + ATP → an [Fe-S cluster biosynthesis chaperone]-ATP + a [disordered-form [Fe-S] cluster scaffold protein] + ADP PG_RS05625 0.81 L-histidine biosynthesis L-histidinol-phosphate + 2-oxoglutarate ↔ imidazole acetol-phosphate + L-glutamate PG_RS05635 0.60 superpathway of L-serine 3-phospho-L-serine + 2-oxoglutarate ← L-glutamate + 3-phospho-hydroxypyruvate and glycine biosynthesis I L-serine biosynthesis (3R)-3-hydroxy-2-oxo-4 phosphonooxybutanoate + L-glutamate → 4-phospho-hydroxy-L-threonine + 2- oxoglutarate pyridoxal 5'-phosphate biosynthesis I

207

PG_RS05815 -1.44 formate assimilation into a tetrahydrofolate + ATP + formate ↔ an N10-formyl-tetrahydrofolate + ADP + phosphate 5,10- methylenetetrahydrofolate PG_RS06705 -0.80 1,4-dihydroxy-2- 2-succinylbenzoate + ATP + coenzyme A → 4-(2'-carboxyphenyl)-4-oxobutyryl-CoA + AMP + naphthoate biosynthesis diphosphate PG_RS06800 -1.13 superpathway of 6-(hydroxymethyl)-7,8-dihydropterin + ATP → (7,8-dihydropterin-6-yl)methyl diphosphate + AMP + H+ tetrahydrofolate biosynthesis 6-hydroxymethyl- 2-amino-6-[1-hydroxyethyl]-7-methyl-7,8-dihydropterin + ATP = [1-(2-amino-7-methyl-4-oxo-7,8- dihydropterin diphosphate dihydro-3H-pteridin-6-yl)]ethyl diphosphate + AMP + H+ biosynthesis I PG_RS07715 0.82 Calvin-Benson-Bassham β-D-fructofuranose 1-phosphate = glycerone phosphate + D-glyceraldehyde cycle gluconeogenesis I β-D-fructose 1,6-bisphosphate ↔ glycerone phosphate + D-glyceraldehyde 3-phosphate glycolysis I (from glucose 6-phosphate) PG_RS08065 0.99 long-chain fatty acid a long-chain fatty acid + ATP + coenzyme A → a long-chain acyl-CoA + AMP + diphosphate activation fatty acid β-oxidation I a 2,3,4-saturated fatty acid + ATP + coenzyme A + H+ → a 2,3,4-saturated fatty acyl CoA + AMP + diphosphate + H+ stearate biosynthesis II oleate + ATP + coenzyme A + H+ → oleoyl-CoA + AMP + diphosphate + H+ (bacteria and plants) palmitate biosynthesis II a 2,3,4-saturated fatty acid + ATP + coenzyme A → a 2,3,4-saturated fatty acyl CoA + AMP + diphosphate (bacteria and plants) decanoate + ATP + coenzyme A → decanoyl-CoA + AMP + diphosphate a (2R)-2-hydroxy even numbered straight chain 2,3,4-saturated fatty acid + ATP + coenzyme A → a (R)-2- hydroxy even numbered straight chain 2,3,4-saturated fatty acyl CoA + AMP + diphosphate cis-vaccenate + ATP + coenzyme A + H+ → cis-vaccenoyl-CoA + AMP + diphosphate + H+ palmitoleate + ATP + coenzyme A + H+ → palmitoleoyl-CoA + AMP + diphosphate + H+

208

an odd numbered straight chain 2,3,4-saturated fatty acid + ATP + coenzyme A → an odd numbered straight chain 2,3,4-saturated fatty acyl CoA + AMP + diphosphate oleate + ATP + coenzyme A → oleoyl-CoA + AMP + diphosphate palmitate + ATP + coenzyme A → palmitoyl-CoA + AMP + diphosphate 18-hydroxyoleate + ATP + coenzyme A = 18-hydroxyoleoyl-CoA + AMP + diphosphate 16-hydroxypalmitate + ATP + coenzyme A = 16-hydroxypalmitoyl-CoA + AMP + diphosphate stearate + ATP + coenzyme A → stearoyl-CoA + AMP + diphosphate octanoate + ATP + coenzyme A → octanoyl-CoA + AMP + diphosphate 7-hydroxylaurate + ATP + coenzyme A → 7-hydroxylauroyl-CoA + AMP + diphosphate α,ω-9Z-octadecenedioate + ATP + coenzyme A = ω-carboxy-(9Z)-octadec-9-enoyl-CoA + AMP + diphosphate a very-long-chain fatty acid + ATP + coenzyme A → a very long chain fatty acyl-CoA + AMP + diphosphate (15Z)-tetracosenoate + ATP + coenzyme A → (Z)-15-tetracosenoyl-CoA + AMP + diphosphate linoleate + ATP + coenzyme A → linoleoyl-CoA + AMP + diphosphate a 2-methyl branched 2,3,4-saturated fatty acid + ATP + coenzyme A → a 2-methyl branched 2,3,4-saturated fatty acyl-CoA + AMP + diphosphate icosapentaenoate + ATP + coenzyme A → icosapentaenoyl-CoA + AMP + diphosphate a 3-methyl-branched 2,3,4-saturated fatty acid + ATP + coenzyme A → a 3-methyl-branched 2,3,4-saturated fatty acyl-CoA + AMP + diphosphate 18-hydroxystearate + ATP + coenzyme A = 18-hydroxystearoyl-CoA + AMP + diphosphate pristanate + ATP + coenzyme A → pristanoyl-CoA + AMP + diphosphate α-linolenate + ATP + coenzyme A → α-linolenoyl-CoA + AMP + diphosphate phytenate + ATP + coenzyme A → phytenoyl-CoA + AMP + diphosphate laurate + ATP + coenzyme A → lauroyl-CoA + AMP + diphosphate PG_RS09670 0.80 Na+[in] + NADH + a ubiquinone + H+ → Na+[out] + NAD+ + an ubiquinol PG_RS09680 0.61 Na+[in] + NADH + a ubiquinone + H+ → Na+[out] + NAD+ + an ubiquinol PG_RS09785 -1.01 phosphopantothenate (R)-pantoate + NADP+ ← 2-dehydropantoate + NADPH + H+ biosynthesis I

209

Appendix A Bioinformatic analyses on PG1605 and PG1788 A.1 Porphyromonas gingivalis C1 cysteine peptidases P. gingivalis was found to possess a number of physiologically important peptidases in the clan CA of cysteine peptidases in the MEROPS database, which include Tpr, PrtT, periodontain and bacteriocin-processing peptidases (Rawlings et al. 2014). These peptidases are further classified into distinct families and subfamilies that exhibit diverse biological properties and activation mechanisms under various conditions. Most of the P. gingivalis peptidases in the clan CA have been studied, except for PG1605 and PG1788 that are categorised as an unassigned subfamily of C1B and C1A respectively. Members of the C1A subfamily are widespread across all phylogenetic kingdoms and are especially well represented in plants, mammals and protozoa. Although MEROPS has currently denoted 220 C1A holotypes or founding members with unique identification number, there are only three bacterial C1A representatives, which include Cwp13 of Clostridium difficile (C01.167), CysP peptidase of Mycoplasma sp. (C01.156) and CPXaC peptidase of Xanthomonas sp. (C01.164), suggesting that the bacterial C1A peptidases are rather poorly examined in this subfamily. Peptidases from the C1B subfamily have fewer holotypes as compared to the C1A subfamily. The peptidases of C1B subfamily are primarily represented by bleomycin hydrolases (BHs) and peptidases (PepC, PepE, PepG, PepW) from the Firmicutes phylum. Members of the C1B subfamily of bacterial origin are primarily studied from Lactobacillus and Lactococcus spp. as these bacterial peptidases have agricultural and industrial implications in casein hydrolysis for quality production of cheese (Pritchard and Coolbear 1993). The prevalence of C1 family peptidases determined from the complete genome sequences of bacteria suggest that these peptidases might play essential biological roles among the prokaryotes. Furthermore, bacterial species with the highest count of C1 peptidases were found to be pathogenic bacteria or human associated microbiota (Table A.1a), which suggests that some of these peptidases could be virulence determinants. Yet, little is known about their roles due to the lack of structural, biochemical and functional analyses of the bacterial C1 family peptidases. Compilation of the total number of annotated C1 peptidases by bacterial genera (881 genera in total) revealed that there are 553 genera without C1 peptidases, while Porphyromonas has the 8th highest number of C1 peptidases identified among the bacterial genera in the MEROPS database (Table A.1b). In addition, other oral bacterial species that

210

belong to the genus Prevotella, Capnocytophaga and Streptococcus were also identified to encode a significant number of C1 peptidase genes (Table A.1b). However, it should be noted that the actual number of C1 peptidases could vary between species and the total count of C1 peptidases grouped by genus can be biased depending on the availability of bacteria with complete genome sequences. Table A.1: (a) Bacteria; species containing the highest number of C1 family peptidases based on bacterial species with complete genome sequences (b) Top 20 list of bacterial genera containing the most C1 family peptidases, based on bacterial species with complete genome sequences. This Information was sourced from the MEROPS Peptidase Database Release 12.0 database. (a) Organism Counts of C1 peptidase Gardnerella vaginalis 16 Chlamydia trachomatis 15 Parabacteroides distasonis 11 Lactobacillus johnsonii 9 Bacillus cereus 7 Finegoldia magna 7 (b) Genus Counts of C1 Count of Species Lactobacillus 203 79 Prevotella 202 68 Bacteroides 196 74 Bifidobacterium 93 44 Streptococcus 91 82 Clostridium 88 96 Parabacteroides 44 10 Porphyromonas 41 19 Mycoplasma 39 40 Eubacterium 38 26 Enterococcus 36 27 Leptospira 35 25 Corynebacterium 34 33 Capnocytophaga 31 17 Xanthomonas 30 16 Flavobacterium 28 24 Atopobium 22 9 Alistipes 22 14 Pseudomonas 22 47 Propionibacterium 21 18

211

A.2 Bioinformatic analysis on peptidase PG1605 and PG1788 PG1605 and PG1788 are C1 peptidases of P. gingivalis that are annotated as aminopeptidase in the NCBI database and predicted to localize in the non-cytoplasmic regions (Table 4.2), however their biochemical and functional properties remain undocumented. Therefore, it is of interest to gain a better understanding of their physiological activation mechanisms by bioinformatic analyses of the sequences and predicted structures of PG1605 and PG1788. A.2.1 Sequence comparisons of PG1605 with its homologues Amino acid sequence comparisons of PG1605 with its orthologues were performed to gain insights into the functionally important residues of PG1605. A BLASTp search with the full length amino acid sequence of PG1605 (NCBI Accession number: WP_005874606.1) provided a list of proteins that were homologous to PG1605, as well as assigning it to the C1 peptidase superfamily (Fig. A.1). The top 100 BLASTp hits were peptidases that belong to the Bacteroidetes phylum, which showed full length coverage of the PG1605 protein sequence with an at least 59% sequence identity (E-value = 0). Representatives of peptidases from different Bacteroidetes spp. were selected for multiple sequence alignment (MSA) and phylogenetic comparison with PG1605, while the P. gingivalis strains that accounted for 22 hits were not included. To my knowledge, none of the Bacteroidetes peptidases that have highly similar sequences to PG1605 have been experimentally characterised. Therefore, full length amino acid sequences from the characterised C1B peptidase homologues that belong to the Firmicutes, as well as human and yeast BHs, were included for further analysis. BHs are referred to as a primary archetype of C1B peptidases, which are well-characterised due to their ability to inactivate bleomycin, an anti-cancer drug.

Figure A.1: Assignment of specific hits and superfamilies of PG1605 in the BLASTp search. The putative active sites (Gln87, Cys93, His386, Asn408) and trimer interface (Lys67, Gln74, Phe75, Ser76, Glu185, Thr187, Thr250, Ser252, Arg254, Val260, Lys403) of PG1605 were inferred in this search result.

212

Expectedly, PG1605 is most closely related to peptidases of Bacteroidetes origin in accordance to their distance on the phylogenetic tree (Fig. A.2). Likewise, the Firmicutes peptidases were clustered in a clade on the phylogenetic tree (Fig. A.2). Oligopeptidase PepE did not show significant deviation from PepW, PepG and PepC aminopeptidases of the Lactobacillus and Lactococcus spp. Further, PepW of L. rhamnosis that has been structurally solved was branching from an internal node with PepC Lactobacilllus helveticus and Lactobacillus delbrueckii. Human and yeast BHs that showed 38% and 31% sequence identity to PG1605 respectively were the first to diverge from the root, as compared with other bacterial C1B peptidases (Fig. A.2).

Figure A.2: Phylogenetic tree generated by neighbor-joining using BLOSUM6 by ClustalWS alignment of the amino acid sequences of PG1605 and homologues. Abbreviation of labels PepC of Pe, Porphyromonas endodontalis (WP_004332057.1); Pm, Porphyromonas macacae (WP_036850061.1); Pdentalis, Prevotella dentalis (WP_005844626.1 Pd); Pdenticola, Prevotella denticola (WP_029216475.1); Bt, Bacteroides thetaiotaomicron (WP_061473815.1); Pgulae, Porphyromonas gulae (WP_039419063.1); BMHL_Bt, BH of B. thetaiotiomicron (WP_016267729.1); Prevotella sp. (WP_009227266.1); Bf, Bacteroides fragilis (WP_032574186.1); Lh, Lactobacilllus helveticus (Q10744); Ll, Lactococcus lactis (Q04723); Ld, Lactobacillus delbrueckii (Q48543); PepW_Ld, PepW of L. delbrueckii (P94868); PepG_Ld, PepG of L. delbrueckii (P94869); Lm, Listeria monoxytogenes (O69192); St, Streptococcus thermophiles (Q56115); PepE_Lh, PepE of L. helveticus (P94870); BMHL_Yeast, Yeast bleomycin hydrolase (Q01532) and BMHL_human, human BH (Q13867).

213

MSA of PG1605 with the selected peptidases showed highly conserved intermittent stretches of amino acid sequence throughout the protein length, especially around the predicted active sites (Gln87, Cys93, His386, Asn408) (Fig. A.3). Regions and residues that showed high conservation across all peptidase sequences likely correspond to the unique and preserved functional roles of the C1B subfamily peptidases. Consensus sequences of the Firmicutes and Bacteroidetes peptidases were presented, in comparison to human and yeast BHs, to demonstrate potential sequence divergence and conservation in C1B peptidases (Fig. A.4). Peptidases selected from Bacteroidetes for analysis showed the presence of an extended N-terminal sequence, which consisted of 26–29 amino acid residues (Fig. A.4). The N-terminal signal peptide showed a conserved motif of AQXXZGG/A, where X represents any amino acid and Z represents a charged or polar residue (Fig. A.4). This conserved sequence was similar to the leader peptide of bacteriocins or translocator-associated bacteriocin leader peptidases that contain double glycine or Gly at P2 and Ala at P1 of the C-terminal end of the leader peptide sequence, suggesting that PG1605 might be processed similarly to these bacteriocins and leader peptidases (Ennahar et al. 2000, Rawlings and Salvesen 2012). This analysis implies that these cysteine peptidases amongst the Bacteroidetes might be secreted from the cytoplasm and potentially function as an extracellular cysteine peptidases. On the contrary, the peptidases of the Firmicutes that have been found to be localized in the cytoplasmic region lack the extended N-terminal signalling sequence. Other sequence features of C1B peptidases were also scrutinized and compared with PG1605. In general, C1B peptidases are synthesized without propeptides and do not contain disulphide bonds (Rawlings and Salvesen 2012). PG1605 contains two cysteine residues in the sequence but one is predicted to act as a nucleophile for the catalytic activity. Likewise, regions around the PG1605 putative catalytic Cys residue are conserved, which is followed by an aromatic, hydrophobic amino acid as observed in other C1B peptidases (Rawlings and Salvesen 2012). The C-terminal sequences were conserved in the aminopeptidases of Lactobacillus and Lactococcus spp., as well as human and yeast BHs. The last residue (Lys454) at the C-terminus of yeast BH was removed by autocatalytic activation mechanism, while the four residues (Gly450, Ala451, Leu452, Ala453) in the C-terminal arm have been shown to be the determinant of their aminopeptidase properties (Zheng et al. 1998). In particular, Gly450 allows the flexibility of the C-terminal arm and governs the positioning of substrates for

214

cleavage at the catalytic site (Zheng et al. 1998). Deletion of the Gly-Ala-Leu-Ala residues of yeast BH and PepC resulted in the conversion of the aminopeptidase to oligopeptidase activities as the C-terminus of the aminopeptidase has been shown to be important for the positioning of oligopeptide substrates for cleavage of the N-terminus amino acid (Zheng et al. 1998, Mata et al. 1999). Mechanistically, the carboxylate group at the C-terminus of BHs and PepC is responsible for the binding and positioning of the N-terminus of a peptide substrate at the P1 position. PepG, PepE and PepW have a shorter aligned C-terminal region in comparison to the characterised aminopeptidases (Fig. A.3), suggesting that they function as endopeptidases, which is the case for PepE and PepG (Mengjin et al. 2010). PG1605 and other representative peptidases of Bacteroidetes are found to have different C-terminal motifs from Firmicutes and BHs. The replacement of Gly450 in the C-terminus arm of BHs to Phe443 in the Bacteroidetes peptidases (Fig. A.4) suggests that the C-terminus of the Bacteroidetes peptidases is not flexible and could not support the role in assisting peptide substrate mobilization and positioning in the catalytic site as with the case for Gly450 in BHs. Based on the MSA analysis, Bacteroidetes cysteine peptidases are suggested to be non-cytoplasmic and function as aminopeptidases in the C1B subfamily.

215

Figure A.3: Multiple sequence alignment of PG1605 against other C1B peptidase family homologues was performed using ClustalO in default parameters. The colored amino acid residues are based on the Clustal default coloring scheme, which hydrophobic residues (Ala, Ile, Leu, Met, Phe, Trp, Val, Cys) are in blue, aromatic residues (His, Tyr) is in cyan, positively charged residues (Lys, Arg) are in red, negatively charge residues (Glu, Asp) are in magenta, polar residues (Asn, Qln, Ser, Thr) are in green, Gly is in orange, Pro is in yellow and white as unconserved category. Residues highlighted in black indicate the active site (Gln, Cys, His, Asn) of the peptidases.

216

Figure A.3 continued.

217

Figure A.4: Consensus sequence representing the phylum of Firmicutes and Bacteroidetes, in comparison to yeast and human BHs amino acid sequences. Boxed in red are the catalytic residues of peptidases. Boxed in purple and blue are the Lys residues of yeast BH that contribute to DNA- binding property and high electropositive potential in spatial proximity to DNA binding site, respectively. The consensus sequences of peptidases from the Bacteroidetes and Firmicutes were generated from the Bacteroidetes and Firmicutes amino acid sequences in Fig. 4.10 by Jalview. Boxed in green in the consensus sequence of Bacteroidetes peptidases were the conserved sequence motif of the N-terminal signal peptide.

218

A.2.2 Sequence comparisons of PG1788 with its homologues Similar approaches as described above were taken for the MSA and phylogenetic tree construction for PG1788 peptidase. A BLASTp search using the PG1788 amino acid sequence (Accession number: WP_005874142.1) showed PG1788 homologues with full length coverage of the protein with at least 56% sequence identity (E-value = 2e-163). Ten peptidase homologues representing different phylogenetic families were selected for the phylogenetic tree construction (Fig. A.5). Aminopeptidase C of Parabacteroidetes distasonis (BDI_2249) that has been structurally solved and showed 52% overall sequence identity to PG1788 was also included in the analysis. In addition, archetypes of C1A subfamily cysteine proteases i.e. papains and cathepsins, as well as bacterial representatives of the C1A subfamily and certain C1B members that have been characterised were arbitrarily included in the analysis. On the phylogenetic tree, bacterial cysteine peptidases were found to cluster in a clade, whereby PG1788 was more closely related to homologues determined from the BLASTp results (Fig. A.5). PG1605 and its C1B homologues were found to segregate from a more recent node from PG1788 homologues, whereas the bacterial representatives of C1A peptidases, Cwp13 of Peptoclostridium difficile and CysP peptidase of Mycoplasma sp., diverged from the clade earlier (Fig. A.5). Papains and cathepsins appeared to arise in parallel or in linear arrangement on the phylogenetic tree, but the Xanthomonas campestris C1A subfamily peptidase showed early divergence from all the other well-characterised peptidases, suggesting that this cysteine peptidase has different physiological properties from the rest of the peptidases (Fig. A.5). PG1788 showed 21% (E-value = 2e-09) and 19% (E-value = 4e-09) sequence identity to yeast and human BHs respectively with full sequence coverage, but only partial coverage with 31% sequence identity to the papain of Carica papaya (query coverage 39% and E-value = 3e-05) and 33% sequence identity to human cathepsin B (21% query coverage and E-value = 6e-04). Those conserved regions primarily included the catalytic residues (Gln, Cys, His, Asn), which were aligned in all selected C1 family peptidases; otherwise, gaps were introduced to compensate for the unaligned regions (not shown). Therefore, the founding members of the C1A subfamily that showed large gaps in the MSA analysis with PG1788, as well as the PG1788 homologues that have not been characterised were excluded from the MSA visualization presented below (not shown).

219

Figure A.5: Phylogenetic tree of PG1788 against other C1 family member by neighbor-joining of BLOSUM62. Full names and the accession numbers for the abbreviated labels used for phylogenetic tree construction are as follows: AP, Aminopeptidase; Pgordonii, Parabacteroides gordonii (WP_028727934.1); Pp, Paludibacter propionicigenes (WP_013445363.1); Pc, Porphyromonas cangingivalis (WP_036844245.1); Bf, Bacteroides faecichinchillae (SHF28025.1); Tf, Tannerella forsythia (WP_060831328.1); Pe, P. endodontalis (WP_004332304.1); Pa, Porphyromonas asaccharolytica (WP_004330520.1); Pgulae, P. gulae (WP_018964538.1); Bt, B. thetaiotaomicron (WP_072067540.1); Pc, Porphyromoans crevioricanis (WP_036890018.1); PG1788 (Q7MTY9), PG1605 (Q7MUC5); BMHL_Human (Q13867), CSAP_Pd, Cell surface associated protease of Peptoclostridium difficile (C9YQ11); BLH1_Yeast (Q01532); BDI_2249 of P. distasonis (A6LE66); PepW_Lr, PepW of L. rhamnosis (PDB code 4K7C); PAPA1, Papain of Carica papaya (P00784); PAPA4, Glycyl endopeptidase of papain (P05994); CATL2,_HUMAN, Human Cathepsin L2 (O60911); CATL1_Human, Human Cathepsin L1 (P07711); CATS_Human, Human Cathepsin S (P25774); CATK_Human, Human Cathepsin K (P43235); CATH_Human, Human Pro-cathepsin H (P09668); CATB-Human, Human cathepsin B (P07858); PepT1_Dp, Dermatophagoides pteronyssinus Peptidase 1 (P08176); CP_Xc, X. campestris cysteine protease (Q8P7B8); CATC_RAT, Rattus norvegicus dipeptidyl peptidase 1, (P80067); Falcipain-3_Pf, Falcipain-3 of Plasmodium falciparum (Q9NAW4); C1_Ms, Mycoplasma synoviae Peptidase C1 (A0A0E3NC65).

220

Included in the MSA analysis were P. distasonis BDI_2249 that has been solved structurally and was shown to be in the same clade as PG1788 (Fig. A.5), as well as the other four representatives from the C1B subfamily that shared a more recent node with PG1788 homologues (Fig. A.6). As described in section A.1.1, the C-terminal arm of BHs and PepC is their primary feature that allows substrate position specificity, which was relatively conserved across most characterised C1B subfamily representatives. However, the bacterial C1A unassigned subfamily peptidases lack the C-terminus arm or showed shorter C-termini (Fig. A.6), suggesting different activation and catalytic mechanisms for these peptidases. The high overall sequence identity of PG1788 with P. distasonis BDI_2249 implied that BDI_2249 could be employed as a reference for the understanding of PG1788 mechanistic and functional properties. The metal binding site (Cys56, Asp298, His340) of BDI_2249 was compared to PG1788, which showed high sequence conservation between the orthologues (Fig. A.6). This suggested that PG1788 might also require metal ions for the coordination of its activities. Although the crystal structure of BDI_2249 has been published on Protein Data Bank (PDB), literature information on this peptidase is still unavailable at present. Biochemical characterisation and kinetic assays of this peptidase that is highly similar to PG1788 would be extremely helpful in understanding the functional properties of PG1788.

221

Figure A.6: Multiple sequence alignment of PG1788 with other C1 peptidases homologues using ClustalO with default parameters. The colored amino acid residues are based on the Clustal default coloring scheme, which hydrophobic residues (Ala, Ile, Leu, Met, Phe, Trp, Val, Cys) are in blue, aromatic residues (His, Tyr) is in cyan, positively charged residues (Lys, Arg) are in red, negatively charge residues (Glu, Asp) are in magenta, polar residues (Asn, Qln, Ser, Thr) are in green, Gly is in orange, Pro is in yellow and white as unconserved category.

222

A.2.3 PG1605 modelled structure PG1605 tertiary structure modelling was performed using I-TASSER servers, which adopt a threading approach to model protein structure (Yang et al. 2015). I-TASSER modeling identifies structure templates from a meta-server threading approach, then performs structure assembly and simulates clustering of the structural conformations with high pair-wise similarity and eventually selects for the highest structure density as the targets (Zhang 2009). The top structure templates that had been determined with the highest significance in threading alignments determined by different threading programs include 1CB5, mutant form of a human BH C73A/ΔE455 (PDB: 2CB5) and 4K7C, while the top five structural analogs of PG1605 final modelled structure provided by I-TASSER were 1CB5, mutant form of yeast BH C73A (PBD: 1A6R), 4K7C, 3PW3 and 3GCB. PG1605 was predicted to comprise 18 α-helices and 12 β-strands, with short and intermittent β-strands interspersing the α-helices at the protein C- terminus (Fig. A.7a; Zhang 2009). Consistently, the predicted catalytic Cys73 and His366 residues of PG1605 are localized at the start of the fourth helix and eleventh β-strand, which is similar to other clan CA cysteine peptidases (Rawlings and Salvesen 2012). Alignment of the PG1605 modelled structure with 1CB5 showed a root mean square deviation (RMS) of 0.3 Å, which indicated that the overall protein folds were in high agreement with one another (Fig. A.7b). The potential N-terminal signal sequence of PG1605 formed an extended arm conformation that is shown to be exposed in the solution (Fig. A.7a). This enables the localisation and proteolytic cleavage processing of the PG1605 leader peptide sequence. Furthermore, the core secondary structure elements of PG1605 were highly conserved with the 1CB5 structure, while the active site of PG1605 (Gln87, Cys93, His386, Asn408) was modelled to be accessible at the cleft in a monomer as with the aligned chain A of 1CB5 (Fig. A.7b). In addition, the annotated trimer interface of PG1605 (Fig. A.1) showed similar folds at and around the amino acid residues of the 1CB5 oligomerization sites (Fig. A.7c), suggesting that the modelled PG1605 will have a quaternary structure resembling that of human BH. Human BH is active in a homohexameric form with its active sites embedded in the central cavity of the protein barrel structure that restricts the accessibility of large peptides into the active clefts (Rawlings and Salvesen 2012). Thus, since PG1605 is predicted with high confidence to be similar to human BH structure, PG1605 may similarly be involved in the cleavage of short peptides.

223

The cleavage specificity and activity of BH is dependent on its C-terminal arm that extends into the central channel for selective interaction with the amide group of the substrate, whereby the flexible Gly450 that forms a G bulge, directs the positioning of substrate to the catalytic sites (Zheng et al. 1998, O'Farrell et al. 1999). The C-terminus arm of PG1605 is represented by Phe463-Ala464-Pro465-Glu466 residues that are in positional replacement of Gly450-Ala451-Leu452-Ala453 in yeast BH (Fig. A.7b), which is responsible for the mobilization and positioning of substrate at the catalytic site. The presence of bulky amino acid Phe463 at the C-terminal of PG1605 is expected to cause rigidity to the mobilization arm. Furthermore, the bulky residues Pro465 and Glu466 may further restrict accessibility. Thus, PG1605 is predicted to lack the flexible C-terminal arm and function as a strict aminopeptidase. Moreover, autocatalytic processing of Gln467 from the C-terminus may be required for its self- activation mechanism as the Gln residue might hinder the peptide substrates from accessing the active sites (Fig. A.7b), which was similar to its characterised orthologues (Zheng et al. 1998). Due to the high structural conservation between the modelled PG1605 and 1CB5, the potential functional properties of PG1605 could be inferred from human BH.

(a)

224

(b)

(c)

Figure A.7 PG1605 modelled structure by I-TASSER and structural alignment with 1CB5A. (a) PG1605 modelled structural representation of secondary elements, where helices, sheets and loops are represented in red, yellow and green respectively. (b) Alignment of PG1605 structural model (green cartoon) with 1CB5A (magenta cartoon); zoomed in region is the catalytic site (top right box) and C-terminal arm (bottom right box) of the C1B peptidases and 1CB5A represented in sticks, respectively. The labelled amino acid residues are of the PG1605 sequence. The color of atoms representing stick in red, blue, white and yellow were oxygen, nitrogen, hydrogen and sulfur residues respectively. (c) Structure model of PG1605 aligned with chain A of 1CB5 in a homotrimer. Chain B and C of 1CB5 are presented in grey cartoon. Trimer interface of PG1605 indicated in the Uniprot database (red spheres), while other possible intermolecular interactions of chain A to chain B and C in 1CB5 were represented in light purple spheres.

225

A.3.2 PG1788 modelled structure The modelled structure of PG1788 (Accession number: WP_005874142.1) using I- TASSER consisted of 15 α-helices and nine β-strands (Fig. A.8a). The top structural templates identified from ten different threading programs displayed in the I-TASSER output for the modeling of PG1788, included yeast BH with point mutation C73A (PDB: 1A6R), BDI_2249 (PDB: 3PW3), a mutant form of human BH C73S/ΔE455 (PDB: 2CB5) and PepW of L. rhamnosis (PDB: 4K7C). However, the top ten templates provided by different threading programs for the structural modeling of PG1788 did not show a majority agreement as had occurred with the 1CB5 template for PG1605 modeling. Of the five final models generated by the I-TASSER server, PG1788 modelled structure that showed the highest confidence score was retrieved for further analysis. The structural analogs of the PG1788 modelled structure corresponded to the templates mentioned above, as well as the yeast BH mutant C73A/ΔK454 (PDB: 3GCB), procathepsins (PDB: 2PBH, 2O6X, 2C0Y) and C1 peptidases domain- containing protein (PDB: 5JT8). Therefore, the PG1788 selected model (Fig. A.8a) was compared with various peptidase structures, 3PW3, 1A6R and wild type human BH (1CB5), to gain insights into its structural conformations. The potential N-terminal signal sequence of PG1788 is shown to be rather exposed in the solution relative to its overall conformation (Fig. A.8a) and seems to be readily accessible for localisation targeting and possible cleavage processing. This modelled N-terminal structure of PG1788 also correlated with previous predictive localisation analysis that shows the presence of an N-terminal signal peptide and localisation in the periplasm (Veith et al. 2014). Although PG1788 has five cysteine residues in the sequence, none of the cysteine residues are close enough in proximity to allow the formation of disulphide bonds in the modelled structure (not shown), which is distinctive from the C1A subfamily peptidases that contain several disulphide bonds (Cambra et al. 2012, Seo et al. 2013). The MSA analysis of PG1788 suggested that the C-terminus is rather short in comparison with the C-terminal arms of C1B peptidases (Fig. A.6), while PG1788 modelled structure demomstrated an exposed C-terminus (Fig. A.8a). Thus, PG1788 C-terminus is unlikely to be involved in the positioning of peptide substrate as with the C1B peptidases for their cleavage activities. PG1788 modelled structure showed a RMS of 2.768 Å with chain A of 3PW3 (Fig. A.8b). The solved crystal structure 3PW3 was truncated at its N-terminus for 31 amino acid residues and formed a homohexamer as the BHs. Some of the structural deviations between

226

PG1788 and 3PW3 may partly be accounted for by the interactions of the 3PW3 structure with its ligands Zn2+, K+, chloride and ethane-1,2-diol. Metal ion Zn2+ was found to be localized in the active sites of 3PW3, indicating that metal ion is an important functional component of this peptidase (Fig. A.8b). In addition, regions that showed the inclusion of two K+ metal ions, adjacent to the active sites, were found to be enriched for acidic amino acids in the pocket of 3PW3 that hold the metal ions (Fig. A.8c). Therefore, the vacuum electrostatic potential of PG1788 was computed by PyMol to examine the conservation of Zn2+ and K+ metal binding sites in this cysteine peptidase. Indeed, PG1788 pockets that corresponded to the metal ion binding sites, relative to the position of 3PW3, demonstrated strong negative electrostatic potential (Fig. A.8d), which suggested that these regions might be able to bind to the cationic metal ions. Given the structural and physiochemical conservation of PG1788 modelled structure in the metal binding sites, it is likely that PG1788 requires metal ion binding for its activation and activities. Likewise, numerous cysteine peptidases (Tpr and PrtT) of P. gingivalis have been found to be activated or regulated by metal ions (Otogoto and Kuramitsu 1993, Staniec et al. 2015). The prediction of potential metal binding sites in PG1788 suggested that this C1 peptidase of P. gingivalis might be activated under different conditions to PG1605 that is not predicted to contain a metal binding site. In addition, the PG1788 modelled structure was aligned with 4K7C and 1A6R with RMS of 2.077 Å and 1.097 Å respectively (Fig. A.9a,b). Regions around the putative active sites of PG1788 were shown to be structurally conserved with 3PW3 and 1A6R structures (Fig. A.9b, Fig. A.9b). Common deviations of PG1788 with other structural homologues were the regions (residues 39-66, 296-344, 364-373) that are composed of loops, which are likely to be highly flexible and are largely responsible for the dynamics and physiological properties of the protein i.e. engaging in protein-protein interactions. The backbone structure at the putative catalytic site of PG1788 is also in alignment with 1A6R that is without metal ion binding (Fig. A.9b), suggesting that the modelled PG1788 might also exhibit functional properties similar to a yeast BH. PG1788 structure showed shorter helix-turn-helix (HTH) and antiparallel beta strand at the periphery of the protein, as compared to 1A6R and 4K7C (Fig. A.9a,b). Whereas, the loop regions of the PG1788 modelled structure were more similar to 1A6R (Fig. A.9b), than 4K7C (Fig. A.9a) and 3PW3 (Fig. A.9b).

227

(a) C

(b)

228

Figure A.8: PG1788 modelled structure by I-TASSER and structure alignment with 3PW3A (a) PG1788 modelled structure, where helices, sheets and loops are colored in red, yellow and green (b) Alignment of PG1788 modelled structure from I-TASSER (cyan cartoon) with its structural analogues, 3PW3A (red cartoon). Bound to the structure of 3PW3A were zinc ion (yellow sphere) and two potassium ions (orange spheres) with a close-up view at the active regions of PG1788 and 3PW3A. (c) Metal binding regions of 3PW3A were aligned with PG1788, which demonstrate structural conservation around the metal binding regions of PG1788 relative to 3PW3A. (d) Vacuum electrostatic potential of PG1788 modelled structure computed by PyMol. The gradient of electrostatic potential from strong negative to strong positive is indicated in the scale bar below from red to blue.

229

(a)

(b)

230

Figure A.9: PG1788 modelled structure alignment with 4K7C and 1A6R. (a) PG1788 (cyan cartoon) was aligned with PepW of L. rhamnosis (4K7C; orange cartoon). Circled regions showed structural deviations of PG1788 from 4K7C, which were mainly around the (1) loop region, (2) shorter antiparallel beta strand and (3) shorter HTH than the 4K7C. (b) PG1788 is aligned to variant form of yeast BH (1A6R; yellow cartoon) and their different orientations were shown. A close-up view of their active sites are depicted in a box, which also shows structural conservation. Circled regions representing the structural deviation of PG1788 fom 1A6R, which include (1) loop region, (2) shorter antiparallel beta strand (3) shorter HTH of PG1788, in comparison to 1A6R.

231

Examination and characterisation of cysteine peptidases encoded in the P. gingivalis genome are important as these peptidases can be essential virulence factors of P. gingivalis. For instance, gingipains, C25 peptidase and PorU signal peptidase are part of a pathologically and physiologically important clan CD of cysteine peptidases. In addition, periodontain that belongs to clan CA was able to inactivate human proteinase inhibitors that are the regulators for human neutrophils elastase, which in turn cause abnormal connective tissue destruction (Nelson et al. 1999). Tpr, another clan CA calpain-like peptidase that is localized in the outer membrane fraction, was shown to be able to degrade complement proteins and antimicrobial peptides, as well as suggested to be involved in the generation of metabolic peptides for amino acid incorporation (Staniec et al. 2015). However, PG1605 and PG1788 that are classified as members of the clan CA cysteine peptidases and segregated into the C1B and C1A subfamilies respectively remain uncharacterised thus far. Bioinformatic analyses of PG1605 and PG1788 in this study suggested that these two peptidases are likely to have different activation mechanisms and functioning conditions. The modelled structure of PG1605 was likely to be immature for functioning as Gln447 in the C- terminal arm was modelled to be inserted adjacent to the active site, which might hinder the peptide substrates from accessing to the catalytic region. Autocatalytic processing by removing the last Gln447 residue from the C-terminus might be required for its self-activation, which was proposed in reference to the activation mechanism of BHs and lactococcal PepC by first operating as a carboxypeptidase to cleave the last residue from the C-terminus (Zheng et al. 1998). Also, it is conceivable that PG1605 functions as an aminopeptidase based on the comparative sequence analyses that showed a Phe443 replacement in PG1605 instead of the essential Gly residue that gives rise to flexibility of the characteristic C-terminal arm of BHs and peptidases of the Lactococcus and Lactobacillus spp. Some other functional properties of lactococcal peptidases can also be applied to PG1605 as they exhibit a relatively high overall sequence similarity. It was determined that some of the conserved amino acids of PG1605 are localized around the cleft region and were in close proximity to the active sites, which suggested that these residues might have a role in governing the specificity and affinity for substrate binding at extended binding subsites. Insights into PG1788 functional properties were mainly in reference to PepC of P. distasonis, BDI_2249, as they shared highly conserved sequences. Furthermore, PG1788 demonstrated strong negative electrostatic potentials at its putative metal binding sites, relative

232

to the metal ion binding positions in 3PW3, suggesting that the activities of PG1788 are activated or regulated by metal ions. As stimulation of Tpr and PrtT activities require the presence of calcium ions, as well as its autoprocessing activation mechanism (Otogoto and Kuramitsu 1993, Madden et al. 1995, Staniec et al. 2015), it is likely that PG1788 activation mechanism and activities resembles that of Tpr, rather than PG1605. In addition, the N- terminus of PG1788, which is currently denoted as a signal sequence, is expected to be truncated in the mature form of PG1788 as with the 3PW3 homologue. The modelled structure of PG1788 also showed structural conservation to the yeast BH and PepW of L. rhamnosis, which both have a RMS smaller than BDI_2249. Although it was uncertain if PG1788 could be function as an aminopeptidase, the substrate positioning mechanism of PG1788 is unlike that of BHs, which are made of C-terminal arm that is involved in the positioning of peptide substrate for cleavage. Instead of being inserted in close proximity to its active sites, the C- terminus of PG1788 modelled structure was short and exposed to the surrounding, suggesting different mechanisms adopted by PG1788 for substrate positioning and hydrolysis. As many P. gingivalis peptidases remain structurally and functionally unexplored, the protein sequences that showed high amino acid sequence identity to other characterised peptidases can be inferred for their potential roles. In addition, in silico docking of substrate ligands on peptidase of interest could also provide insights into the molecular interactions involved between the interacting partners and peptidase, especially in the active sites or subsites. Regardless, experimental data on molecular, biochemical and kinetic assays with an extensive array of peptide substrates are important to validate the bioinformatic predictions of PG1605 and PG1788 peptidases.

233

Bibiolography (cited in Appendice A)

Cambra, I., Hernández, D., Diaz, I. and Martinez, M. (2012) 'Structural Basis for Specificity of Propeptide-Enzyme Interaction in Barley C1A Cysteine Peptidases', Plos One, 7(5), 1-13.

Ennahar, S., Sashihara, T., Sonomoto, K. and Ishizaki, A. (2000) 'Class IIa bacteriocins: biosynthesis, structure and activity', FEMS Microbiology Reviews, 24(1), 85-106. Madden, T. E., Clark, V. L. and Kuramitsu, H. K. (1995) 'Revised sequence of the Porphyromonas gingivalis prtT cysteine protease/hemagglutinin gene: homology with streptococcal pyrogenic exotoxin B/streptococcal proteinase', Infection and Immunity, 63(1), 238-247. Mata, L., Gripon, J. C. and Mistou, M. Y. (1999) 'Deletion of the four C-terminal residues of PepC converts an aminopeptidase into an oligopeptidase', Protein Engineering, 12(8), 681-686. Mengjin, L., Bayjanov, J. R., Renckens, B., Nauta, A. and Siezen, R. J. (2010) 'The proteolytic system of lactic acid bacteria revisited: a genomic comparison', Bmc Genomics, 11, 1-15. Nelson, D., Potempa, J., Kordula, T. and Travis, J. (1999) 'Purification and characterisation of a novel cysteine proteinase (periodontain) from Porphyromonas gingivalis - Evidence for a role in the inactivation of human alpha(1)-proteinase inhibitor', Journal of Biological Chemistry, 274(18), 12245-12251. O'Farrell, P. A., Gonzalez, F., Zheng, W., Johnston, S. A. and Joshua-Tor, L. (1999) 'Crystal structure of human bleomycin hydrolase, a self-compartmentalizing cysteine protease', Structure with Folding and Design, 7(6), 619-628. Otogoto, J. I. and Kuramitsu, H. K. (1993) 'ISOLATION AND CHARACTERISATION OF THE PORPHYROMONAS-GINGIVALIS PRTT GENE, CODING FOR PROTEASE ACTIVITY', Infection and Immunity, 61(1), 117-123. Rawlings, N. D. and Salvesen, G. (2012) Handbook of proteolytic enzymes, Academic Press.

Rawlings, N. D., Waller, M., Barrett, A. J. and Bateman, A. (2014) 'MEROPS: the database of proteolytic enzymes, their substrates and inhibitors', Nucleic Acids Research, 42(D1), D503-D509. Seo, J. S., Jeon, E. J., Jung, S. H., Park, M. A., Kim, J. W., Kim, K. H., Woo, S. H. and Lee, E. H. (2013) 'Molecular cloning and expression analysis of peptidase genes in the fish-pathogenic scuticociliate Miamiensis avidus', BMC Veterinary Research, 9(1), 1- 10. Staniec, D., Ksiazek, M., Thøgersen, I. B., Enghild, J. J., Sroka, A., Bryzek, D., Bogyo, M., Abrahamson, M. and Potempa, J. (2015) 'Calcium Regulates the Activity and Structural Stability of Tpr, a Bacterial Calpain-like Peptidase', The Journal Of Biological Chemistry, 290(45), 27248-27260.

234

Veith, P. D., Chen, Y.-Y., Gorasia, D. G., Chen, D., Glew, M. D., O'Brien-Simpson, N. M., Cecil, J. D., Holden, J. A. and Reynolds, E. C. (2014) 'Porphyromonas gingivalis outer membrane vesicles exclusively contain outer membrane and periplasmic proteins and carry a cargo enriched with virulence factors', Journal of Proteome Research, 13(5), 2420-2432. Yang, J., Yan, R., Roy, A., Xu, D., Poisson, J. and Zhang, Y. (2015) 'The I-TASSER Suite: protein structure and function prediction', Nature Methods, 12(1), 7-8. Zheng, W., Johnston, S. A. and Joshua-Tor, L. (1998) 'The Unusual Active Site of Gal6/Bleomycin Hydrolase Can Act as a Carboxypeptidase, Aminopeptidase, and Peptide Ligase', Cell, 93(1), 103-109.

235

Minerva Access is the Institutional Repository of The University of Melbourne

Author/s: Kin, Lin Xin

Title: The effect of Treponema denticola on Porphyromonas gingivalis phenotypes and transcriptome

Date: 2019

Persistent Link: http://hdl.handle.net/11343/221829

File Description: The effect of Treponema denticola on Porphyromonas gingivalis phenotypes and transcriptome

Terms and Conditions: Terms and Conditions: Copyright in works deposited in Minerva Access is retained by the copyright owner. The work may not be altered without permission from the copyright owner. Readers may only download, print and save electronic copies of whole works for their own personal non-commercial use. Any use that exceeds these limits requires permission from the copyright owner. Attribution is essential when quoting or paraphrasing from these works.