Repurposing auranofin to treat Clostridium difficile infection by targeting selenium.
by
Christine Roder B. Sci (Hons), M. Biotech
Submitted in fulfilment of the requirements for the degree of
Doctor of Philosophy
Deakin University
June, 2017
[Type here]
Acknowledgements
difficile [dif-i-seel] adjective 1. hard to deal with, satisfy or please. 2. hard to do; difficult
“referring to the unusual difficulty that was encountered in its isolation and study” –Hall & O’Toole, 1935.
Well they certainly weren’t wrong. A PhD is already a difficult mountain to climb and choosing to study an organism that was named for how difficult it is to grow only serves to make that mountain more challenging. As I stand here at the summit and reflect on the arduous uphill battle that was, I feel a sense of accomplishment and relief. I also feel overwhelming gratitude toward everyone who has helped me along the way. Unfortunately there are too many people to thank everyone by name, so I will take this opportunity to thank everyone who has contributed to my completing this thesis. Whether your contribution be small or large, it has meant a lot to me and I thank you for it. There were a number of people whose contribution was so significant they deserve special mention. Firstly I would like to thank my PhD supervisor Eugene Athan. You have provided ongoing support and encouragement throughout this endeavour and seem to have an unwavering confidence in my abilities which was especially important when I did not. It has been an absolute pleasure working with you and I look forward to our future projects together. Thank you to Deakin University for my scholarship, and also to Bellberry Limited for a top up scholarship and NHMRC for an equipment grant awarded to Deakin University, School of Medicine CMMR for the anaerobic cabinet “Annie”. I would also like to thank all the staff and students at CMMR and GCEID for all of their support and guidance. I would particularly like to thank the other members of my Lab group. I would like to acknowledge the contributions made by Dena Lyras and
[Type here]
Melanie Hutton from Monash University in both the donation of bacterial strains, and the assistance and use of facilities for the animal model. I would also like to thank John Stenos and Gemma Vincent from the Australian Rickettsial Reference Laboratory for donating the Vero cell line used in the tissue culture assays. On a personal note, I would like to thank my wonderful family and friends, who have given me a lot of love and support over the years. From coffee dates to wine nights, without these distractions to keep me sane I would not have made it through to the end. Many of you have also completed their PhDs, so to have a support network that understood the trials of PhD life was invaluable. To my parents, Margaret and Craig, the two of you raised me to be strong and independent, and have supported me emotionally and financially throughout my entire tertiary education. You have also read numerous drafts of papers, reports and finally my thesis. A task that was neither easy nor enjoyable, and for which I am eternally grateful. To my two beautiful dogs Charlie and Louie, for being the best fluffy study buddies a girl could ask for. Finally I thank my wonderful husband Kane. For every malteaser run, every time you picked me up in the middle of the night from Uni, for every high and low throughout this whole thing, you were there for me constantly. You are my rock and I would not be who I am without you.
[Type here]
List of publications
1. Roder, C. and Thomson, M.J. Auranofin: Repurposing an old drug for a golden new age. Drugs R & D. 15, 13-20 (2015). 2. Roder, C. et al. Population-based study of ribotypes and toxin-b production in clinical isolates of Clostridium difficile in Victoria, Australia. EC Microbiology. 1, 46-53 (2015). 3. Trubiano, J.A, et al. Australian Society of Infectious Diseases update guidelines for the management of Clostridium difficile infection in adults and children in Australia and New Zealand. Intern Med J. 46, 479-493 (2016).
[Type here]
Table of Contents
Abstract ...... 1 Tables figures and abbreviations ...... 3 List of figures ...... 3
List of tables ...... 4
List of abbreviations ...... 4
1. Introduction ...... 7 1.1. Background and epidemiology ...... 7
1.2. C. difficile infection, pathology, the intestinal microbiota and the immune response ...... 9
1.2.1. The intestinal microbiota ...... 9 1.2.2. Colonization of the intestinal epithelium ...... 10 1.2.3. Toxins and virulence ...... 12 1.2.4. The host immune response ...... 18 1.3. Risk to public health ...... 19
1.3.1. Sources ...... 19 1.3.2. Hypervirulent strains ...... 20 1.3.3. Recurrent disease ...... 20 1.3.4. Antibiotic resistance ...... 21 1.4. Current and emerging therapies ...... 24
1.4.1. Diagnosis and treatment guidelines ...... 24 1.4.2. Antibiotics ...... 24 1.4.3. Faecal microbiota transplant (FMT) ...... 26 1.4.4. Probiotics ...... 28 1.4.5. Vaccines and antibodies ...... 30 1.5. Auranofin ...... 31
1.5.1. Background and overview of pharmacokinetics ...... 31 1.5.2. Mechanisms of action ...... 33 1.5.3. Potential as a treatment for CDI ...... 34 2. Methodology ...... 36
[Type here]
2.1. Identifying selenoproteins in the C. difficile using in silico methods .... 36
2.2. Investigating auranofin as a treatment for CDI using in vitro methods . 38
2.2.1. Routine in vitro methodology ...... 38 2.2.2. Growth Assays ...... 39 2.3. Preliminary testing of auranofin as a treatment for CDI using an in vivo animal model ...... 44
2.3.1. CDI prophylaxis animal model ...... 44 3. Identifying selenoproteins in C. difficile using in silico methods...... 47 3.1. Background ...... 47
3.2. Results ...... 51
3.3. Discussion ...... 58
4. Investigation of auranofin as a treatment for C. difficile infection ...... 63 4.1. Background ...... 63
4.2. Results ...... 68
4.2.1. Minimum inhibitory concentration (MIC) and subinhibitory concentration of auranofin ...... 68 4.2.2. Treatment concentrations of auranofin ...... 75 4.2.3. Comparison of the activity of auranofin with that of vancomycin and metronidazole ...... 80 4.2.4. Preliminary testing of auranofin as a treatment for CDI using an in vivo animal model ...... 87 4.3. Discussion ...... 90
5. Discussion and conclusion ...... 97 6. References ...... 104 7. Appendix 1 ...... 124 Clostridium difficile genomes from NCBI used for analysis of gene conservation...... 124
8. Appendix 2 ...... 128 SECIS elements from Clostridium difficile 630 genome identified using in silico methods...... 128
ABSTRACT
Abstract Clostridium difficile infection (CDI) is a complex infectious disease that is commonly associated with antibiotic use and disruption of the intestinal microbiota. The current therapies, metronidazole or vancomycin, are increasingly associated with high rates of treatment failure and recurrent disease. A proposed alternative to the current therapies is the repurposed drug auranofin. Auranofin was previously approved for the treatment of rheumatoid arthritis, and has recently been investigated as a treatment for a number of diseases including cancer, HIV, parasitic infections and bacterial infections including CDI. The mechanism by which auranofin exerts antimicrobial activity is through the inhibition of selenoprotein synthesis. Auranofin does this by limiting the bioavailability of selenium to C. difficile, forming a stable adduct with selenide. This prevents C. difficile from taking up the selenide, thus blocking the selenium metabolism pathways necessary for selenoprotein synthesis. The aim of this study was to further investigate auranofin as a treatment option for CDI. This study had two main areas of focus: firstly, the number and level of conservation of selenoproteins in the C. difficile genome, and secondly, the in vitro activity of auranofin on C. difficile. Selenoproteins are proteins that contain selenium in the form of selenocysteine. C. difficile uses selenoproteins in the reductive deamination of the amino acids glycine and proline. This is done by the glycine reductase and proline reductase pathways, respectively, both of which are energy conserving. Using in silico methods it was found that there were 86 genes for selenoproteins identified in the C. difficile 630 genome, 87.21% of these showed >95% conservation among 100 C. difficile strains. The proteins identified had a wide variety of putative functions including metabolism, information storage and processing, cellular processes and signalling. The high number of selenoproteins and the high level of conservation among them suggests that selenium and selenoproteins may play an essential role in C. difficile metabolism, information storage and processing, cellular processes and signalling. This provides strong evidence supporting selenium targeted therapy for the treatment of CDI.
1 ABSTRACT
Previous investigations have demonstrated the mechanism by which auranofin inhibits selenium metabolism in C. difficile, and that this in turn inhibits bacterial growth. This study further investigated the growth inhibitory activity of auranofin. We focused on the growth, vegetative cell viability, sporulation and toxin production of C. difficile cultures treated with auranofin after the cultures reached stationary phase. The efficacy of auranofin was compared to that of the current therapies metronidazole and vancomycin. Finally, auranofin was tested in an in vivo model of CDI in a preliminary animal trial. These investigations demonstrated that auranofin has showed bactericidal activity at a concentration as low as 50 µM in vitro effectively reducing bacterial growth, vegetative cell viability and spore viability. Auranofin also prevented the cytotoxic effects of the C. difficile toxin TcdB on Vero cells, and a dose of 200 µg/ml was sufficient to prevent CDI in an in vivo mouse model without adverse side effects. In conclusion, auranofin has many properties that make it suitable to use to treat CDI. Its activity may be considered narrow spectrum and is not directed toward a cellular molecular target but rather an essential nutrient. Auranofin also has the ideal pharmacokinetics for treating a gastrointestinal infection. Furthermore, the in vitro data presented here strongly demonstrated its efficacy against C. difficile. Auranofin has the potential to be an ideal therapeutic option for the treatment of CDI.
2 TABLES, FIGURES AND ABBREVIATIONS
Tables figures and abbreviations List of figures Figure 1.1 Sporulation cycle...... 11 Figure 1.2: Layers of the C. difficile spore...... 12 Figure 1.3: Endosomal mediated entry of toxins into epithelial cells...... 14 Figure 1.4: Positive regulation of gene expression in the pathogenicity locus...... 16 Figure 1.5: Negative regulation of gene expression in the pathogenicity locus...... 17 Figure 1.6: The chemical structure of auranofin...... 32 Figure 2.1: Flow diagram of the in silico method used to identify C. difficile selenoproteins...... 37 Figure 2.2: Bacterial SECIS element consensus model...... 38 Figure 2.3: Overview of the in vitro growth assay used to investigate auranofin as a treatment for CDI...... 41 Figure 2.4: Overview of cell cytotoxicity assay (CCA)...... 44 Figure 2.5: Overview of Monash CDI prophylaxis mouse model...... 46 Figure 3.1: The sel operon...... 48 Figure 3.2: Selenophosphate synthesis...... 48 Figure 3.3: Selenocysteine synthesis...... 49 Figure 3.4: Incorporation of selenocysteine...... 50 Figure 3.5: Proposed new consensus structure for bacterial SECIS elements...... 52 Figure 3.6: Putative functional classes of C. difficile selenoproteins...... 58 Figure 4.1: The glycine reductase (grd) operon...... 63 Figure 4.2: The glycine reductase pathway...... 64 Figure 4.3: The proline reductase (prd) operon...... 65 Figure 4.4: The proline reductase pathway ...... 66 Figure 4.5: Auranofin MIC assay, 100 µM auranofin...... 68 Figure 4.10: Auranofin MIC assay, 3.125 µM auranofin ...... 71 Figure 4.11: Auranofin MIC assay, 0.098 µM auranofin...... 71
3 TABLES, FIGURES AND ABBREVIATIONS
Figure 4.12: Bacterial growth of C. difficile M7404 treated with 0.1 µM auranofin, 6 µM auranofin, 0.1% ethanol (v/v) or no treatment (control). ... 72 Figure 4.13: Vegetative cell viability of C. difficile M7404...... 73 Figure 4.14: Spore counts for C. difficile M7404 cultures...... 74 Figure 4.15: Bacterial growth of C. difficile M7404 treated with 50 µM auranofin, 0.8% ethanol (v/v) or no treatment (control)...... 75 Figure 4.16: Bacterial growth of C. difficile M7404 treated with auranofin at concentrations of 50 µM, 250 µM and 500 µM compared to no treatment (control)...... 76 Figure 4.17: Vegetative cell viability of C. difficile M7404...... 77 Figure 4.18: Spore counts for C. difficile M7404 cultures...... 78 Figure 4.20: Cytotoxic activity of C. difficile TcdB against Vero cells...... 80 Figure 4.21: Auranofin MIC titration assay for C. difficile M7404...... 81 Figure 4.22: Vancomycin MIC titration assay for C. difficile M7404...... 82 Figure 4.23: Metronidazole MIC titration assay for C. difficile M7404...... 82 Figure 4.24: Bacterial growth of C. difficile M7404 ...... 83 Figure 4.25: Vegetative cell viability of C. difficile M7404...... 85 Figure 4.26: Spore viability of C. difficile M7404 cultures...... 86 Figure 4.27: Infected mice treated with 40 µg/ml auranofin...... 88 Figure 4.28 Infected mice treated with 200 µg/ml auranofin...... 88 Figure 4.30: Colonisation fo mice with C. difficile spores...... 89 Figure 4.31: Cytotoxic activity of C. difficile TcdB...... 90
List of tables Table 1.1: Mechanisms of antibiotic resistance in C. difficile...... 22 Table 1.2: Donor screening for faecal microbiota transplant...... 27 Table 2.1: Treatment groups for in vivo trial of auranofin as CDI treatment...... 46 Table 3.1: C. difficile selenoproteins ...... 53
List of abbreviations
ACSQH Australian Commission for Safety and Quality in Healthcare ASID Australian Society for Infectious Diseases
4 TABLES, FIGURES AND ABBREVIATIONS
ATP Adenosine triphosphate BHI Brain heart infusion agar BLAST Basic Local Alignment Search Tool bp Base pair CBD Cell-binding domain CCA Cell cytotoxicity assay CDC Centre for Disease Control and Prevention CDI Clostridium difficile infection CFU Colony forming unit CGT Clostridial glycosylating toxin COG Cluster of orthologous groups CPD Cysteine protease domain CRE Carbapenem-resistant Enterobacteriaceae CROPs Combined repetitive oligopeptides CSPG4 Chondroitin sulphate proteoglycan 4 EF Elongation factor EUCAST The European Committee on Antimicrobial Susceptibility Testing FBS Fetal bovine serum FdhF (fdhF) Formate dehydrogenase FMT Faecal microbiota transplant G Guanine gp96 Glycoprotein 96 grd Glycine reductase operon GrdA (grdA) Glycine reductase complex protein A GrdB (grdB) Glycine reductase complex protein B GTD Glucosyltransferase domain GTP Guanosine triphosphate HI Heart infusion HPLC High performance liquid chromatography IBD Inflammatory bowel disease ICU Intensive care unit IFN Interferon Ig Immunoglobulin IL Interleukin InsP6 Inositol hexakisphosphate LSR Lipolysis-stimulated lipoprotein receptor MEM Minimal essential media MIC Minimal inhibitory concentration MRSA Methicillin-resistant Staphylococcus aureus NCBI National Centre for Biotechnology Information NimB 5-nitromidazole reductase NS Non-stimulated nt Nucleotide OD Optical density PMC Pseudomembranous colitis
5 TABLES, FIGURES AND ABBREVIATIONS
prd Proline reductase operon PrdA (prdA) Proline reductase complex protein A PS Phosphate buffered saline Redox reductive/oxidation ROS Reactive oxygen species Sec Selenocysteine SECIS Selenocysteine insertion sequence SeH Selenol SelA (selA) Selenocysteine synthase SelB (selB) Selenocysteine-specific elongation factor SelC (selC) Selenocysteine-specific tRNASec SelD (selD) Selenide water dikinase SEM Standard error of the mean SH Thiol TcdA (tcdA) Clostridium difficile toxin A TcdB (tcdB) Clostridium difficile toxin B TD Transmembrane domain TNF Tumor necrosis factor TraDIS Transposon-directed insertion site sequence TrxR (trxR) Thioredoxin reductase TY Tryptone yeast extract U Uridine UTI Urinary tract infection VCC Viable cell count VSC Viable spore count
6 INTRODUCTION
1. Introduction
1.1. Background and epidemiology Clostridium difficile is a Gram-positive, spore forming, anaerobic, pathogenic bacterium belonging to the genus clostridia. Other members of this genus include the pathogens C. perfringens, C. botulinum, and C. tetani1. C. difficile was first isolated from the faeces of healthy neonates in 19352 and was first recognised as a human pathogen responsible for diarrhoea and pseudomembranous colitis (PMC) in 19781,3. C. difficile infection (CDI) is the leading cause of hospital acquired gastrointestinal infection. Recent studies suggest that C. difficile is responsible for 25% more nosocomial infections in the U.S. than methicillin-resistant Staphylococcus aureus (MRSA)4. In 2013 the Centre for Disease Control and Prevention (CDC) in the U.S. declared CDI an urgent public health threat due to its capability of causing widespread and difficult to treat disease and its contribution to antimicrobial resistance. There are two other pathogens in this threat level; Carbapenem-resistant Enterobacteriaceae (CRE) and drug resistant Neisseria gonorrhoea5. Symptoms of CDI range from mild, watery diarrhoea to more severe pseudomembranous colitis (colon only) or pseudomembranous enterocolitis (colon and intestine; collectively referred to as PMC)2,6,7. PMC is characterised by yellow grey patches containing inflammatory cells and cellular debris forming a pseudomembrane over the intestinal mucosa2. Other symptoms of CDI include loss of appetite, nausea, fever, leucocytosis and abdominal cramping. Serious complications of CDI are life threatening and include colonic ileus, toxic megacolon, intestinal perforation, sepsis and shock2,6,7. Chronic diarrhoea also puts patients at risk of dehydration, electrolyte disruption, renal failure, hypotension and systemic inflammatory response syndrome2,6. Some people are colonised with C. difficile, but are asymptomatic. While they have no symptoms, they can still shed spores, the transmissible element, thereby effectively spreading the infection. The incidence of CDI has been steadily rising since the year 2000. In the U.S., the number of CDI related hospital admissions has tripled2,8 to 450,000 cases annually4,5,9-11. This adds a significant burden to healthcare systems with the annual
7 INTRODUCTION
cost of CDI in the U.S. estimated to be in excess of USD $1 billion4,5,10,12. There has also been an increase in the severity of disease associated with CDI2,3,6,13-17 resulting in longer hospital stays, increased admission to the intensive care unit (ICU), more emergency interventions such as colectomy and a higher mortality rate8,15,18,19. It has been estimated that the number of CDI related deaths in the U.S. is as high as 15,000 annually11. Until the development of a national surveillance model for C. difficile in 2010 by the Australian Commission for Safety and Quality in Healthcare (ACSQH) there was very little active surveillance of CDI in Australia. The commencement of this surveillance model in 2010 produced the first nationwide profile of CDI in Australia. A 2-year study (2011 to 2012) identified 12,683 cases of CDI by surveying the laboratory diagnosis of CDI in 450 Australian public hospitals. This study also found the rate of CDI increased during the study by 26%11. Infection with C. difficile is facilitated by the ingestion of metabolically dormant spores4,10,12,20-24. If the intestinal microbiota has been altered or depleted i.e. by antibiotic use, then the environment in the gut will be suitable for vegetative cell growth and the spores will germinate. The vegetative form then colonises the gut mucosa, proliferates, producing more spores and the toxins, TcdA and TcdB, which are the primary virulence factors and cause the disease pathology25. Antibiotic use is associated with nearly all cases of CDI, with most patients having received antibiotics within the three months prior to the onset of disease symptoms2. Most antibiotics, in particular broad-spectrum antibiotics, have been implicated in CDI, with clindamycin, cephalosporin, and quinolones such as moxifloxacin posing the greatest risk26. The elderly have an increased risk of symptomatic infection with C. difficile because of their decreased humoral immunity. There is also a higher occurrence of disease in women compared to men which may be due to increased antibiotic prescribing for urinary tract infections (UTIs), particularly in older women27. Other risk factors include the number and severity of underlying conditions including, immunosuppressive therapy, chemotherapy, inflammatory bowel disease (IBD), Crohn’s disease, long hospital stays, admission to ICU, recent surgical procedures particularly gastrointestinal, nasogastric tube procedures, gastric acid suppressors, renal impairment, contact with CDI patients, HIV infection, prophylactic use of
8 INTRODUCTION
antibiotics and the use of non-steroidal anti-inflammatory agents2,3,6,7,13,19,25,28. Although traditionally considered a hospital-associated infection, CDI is increasingly being recognised as a community-associated disease. There is evidence of increasing occurrence of CDI in persons previously considered to be low risk such as children, pregnant women and younger adults with no underlying diseases6,13,29,30.
1.2. C. difficile infection, pathology, the intestinal microbiota and the immune response
1.2.1. The intestinal microbiota The intestinal microbiota refers to the normal microbial inhabitants of the gastrointestinal tract31 and is highly beneficial to human health32,33. At birth the gastrointestinal tract is sterile with colonisation occurring rapidly in the post- partum period. Mode of delivery (i.e. vaginal or caesarean section) and diet (i.e. breast milk or formula) play an important role in microbiota development32,33. There are thousands of individual species that could colonize the human gut and the composition of the intestinal microbiota varies considerably between individuals32-34. The majority of the intestinal microbiota is made up of obligate anaerobes from the phyla Bacteroidetes, Firmicutes and Actinobacteria. Less dominant phyla include Lentisphaera, Fusobacterium and Verrucomicrobia. The microbiota also includes viruses such as bacteriophages, archaea such as methanogens and eukaryotes such as Candida and Blastocystis34. The intestinal microbiota plays a role in the breakdown of polysaccharides into short chain fatty acids, driving oxygen and pH regulation, maintaining intestinal barrier integrity, mediating the host immune response, reducing gut mucosal inflammation and increasing protection against infection by enteric pathogens such as C. difficile3,32,34. The intestinal microbiota is also responsible for the synthesis of several beneficial compounds including the B vitamins, folic acid, biotin and vitamin K, and the metabolism of bile acid and cholesterol as well as potentially harmful substances, such as nitrosamines, heterocyclic acid amines and oxalate32. Colonisation resistance is the protection of the intestinal mucosa from colonisation by enteric pathogens4,32,33,35,36. The intestinal microbiota can prevent colonisation
9 INTRODUCTION
by pathogens by competing for attachment sites on the mucosa, limiting the resources available for pathogens, producing antimicrobial compounds and by stimulating the host’s defences33,35. A disruption of intestinal microbiota, known as intestinal dysbiosis, will limit its effectiveness to protect the host against pathogens, giving opportunists such as C. difficile a chance to colonise, overpopulate the lower gastrointestinal tract and cause disease4,19,32,33,35-38. One of the roles the intestinal microbiota plays in protection against C. difficile is the de-conjugation of the bile acid taurocholate into taurine and cholate. Taurocholate is a germinant for C. difficile spores, which are able to survive passage through the stomach to the intestines, a journey that most vegetative cells do not survive. De-conjugation of the spore germinant protects the gut from this mode of infection23,39. Intestinal dysbiosis is the state where the composition of the intestinal microbiota has been shifted from the normal beneficial state to one that has a negative effect on human health32,33. Intestinal dysbiosis can be attributed to a number of CDI risk factors including inflammatory bowel disease, chemotherapy, use of proton pump inhibitors, advanced age and most notably antibiotic use19,31,35,40,41. Treatment with broad spectrum antibiotics decreases the diversity and richness of the intestinal microbiota, with decreases in the abundance of Firmicutes and Bifidobacteria and increases in Proteobacteria40. Intestinal dysbiosis increases the host’s susceptibility to CDI by altering several aspects of colonization resistance including the production of antimicrobials, competition for nutrients and space and bile acid metabolism. Further to this, intestinal dysbiosis may play a role in the severity of CDI as the intestinal microbiota plays an integral role in the host immune response. Intestinal dysbiosis may contribute to the attenuated immune response that is associated with severe CDI.
1.2.2. Colonization of the intestinal epithelium The disruption of the intestinal microbiota and subsequent reduction in colonisation resistance creates an ideal environment for opportunistic C. difficile. The transmissible element is the spores, which are metabolically dormant, highly resilient and capable of causing widespread disease4,10,20,21,23,42-47. C. difficile produces spores when the environment no longer supports vegetative growth, such
10 INTRODUCTION
as nutrient restriction, exposure to antibiotics or disinfectant4,21,22,43,46,48. The general sporulation cycle is outlined in Figure 1.1.
Figure 1.1 Sporulation cycle. When the environmental conditions do not favour vegetative cell growth, the vegetative cell can undergo cell division to produce a metabolically dormant spore. Adapted from Fimlaid et al., 2013 and Leggett et al., 2012
Once formed and released, the spores are metabolically dormant and contain very few genomic transcripts43. The transcripts that are found in the dormant spores are a mixture of transcripts from late stage sporulation that were trapped in the spore and proteins that are required during the germination process43 including redox proteins that aid in detoxifying the environment during germination to provide the oxygen-sensitive vegetative cells an anaerobic environment43. C. difficile spores have a multilayered structure designed to protect the core, which contains DNA, RNA ribosomes and the previously mentioned transcripts and proteins21,22,43,46,49. The core is relatively dehydrated compared to the vegetative cell, allowing it to remain dormant and resistant to heat and chemicals21,43,46,49. The layers of the spore are outlined in Figure 1.221,43,46,49.
11 INTRODUCTION
Figure 1.2: Layers of the C. difficile spore. The multilayered structure of the C. difficile spore is designed to protect the DNA containing core. Adapted from Leggett et al., 2012.
Once ingested, the spores pass through the stomach to the intestine where, upon detecting favourable conditions such as the presence of taurocholate, they begin the germination process43. Upon germination there is a dramatic change in the regulation of many proteins. Among the upregulated genes are ABC-transporters to import metabolites, genes associated with DNA replication, cell division and stress response, ribosomal proteins, DNA directed RNA-polymerase, transcriptional regulators and protein translocation mechanisms50. Genes associated with amino acid metabolism, the flagella, chemotaxis and toxin are all down regulated indicating that the germinating spore is dedicating its energy and resources to rebuilding itself into a vegetative cell43. Once the germination process is complete, the vegetative cell then upregulates these genes to resume normal cellular functions such as metabolism, movement and toxin production43.
1.2.3. Toxins and virulence The pathology of CDI is caused by the primary virulence factors, the toxins TcdA and TcdB. These toxins belong to the large Clostridial glycosylating toxin (CGT) family which includes C. sordellii lethal toxin (TcsL) and C. novyi α-toxin (TcnA)38,51,52. The two toxins, TcdA and TcdB, have highly similar amino acid
12 INTRODUCTION
sequences, with the exception of the C-terminus. They each have four domains: the glucosyltransferase domain (GTD) at the N-terminus, the hydrophobic transmembrane domain (TD), the cysteine protease domain (CPD), and the cell- binding domain (CBD) at the C-terminus51,52. The CBD is characterised by a repetitive peptide element called the combined repetitive oligopeptides (CROPs), which are solenoid-like structures that facilitate protein-protein and protein-carbohydrate binding to each of the toxins cell surface receptors. TcdA binds to glycoprotein 96 (gp96), a member of the heat shock protein family that is expressed on the apical surface and in the cytoplasm of human colonocytes. TcdB binds to chondroitin sulphate proteoglycan 4 (CSPG4), a high molecular weight melanoma associated antigen, that is highly expressed throughout the intestine51. Binding of the toxins to their respective cell receptors initiates clathrin-dependent endocytosis, which is outlined in Figure 1.3. Acidification of the endosomal compartment causes a conformational change in the toxin structure, exposing the TD which is able to form a pore in the endosomal membrane via cholesterol- dependent mechanism. This allows the GTD to enter the cytosol. The CPD is then activated by the binding of inositol hexakisphosphate (InsP6), resulting in auto- cleavage and release of the GTD51,52. The GTD acts on the guanosine triphosphate (GTP)-binding proteins Rho, Rac and Cdc42. GTD catalyses the attachment of a glucose residue from uridine diphosphate (UDP)-glucose to a conserved threonine amino acid on the target protein. This irreversible monoglucosylation renders the protein inactive leading to disordered cell signalling, disruption of the actin cytoskeleton, opening of tight junctions, depolarisation of the cell, cell rounding, apoptosis, decreased epithelial barrier function, accumulation of fluid, inflammation, and intestinal injury. TcdA is not essential for virulence but does play a role in disease. TcdB has been shown to be essential for virulence and is 10 times more potent than TcdA25,51,52.
13 INTRODUCTION
Figure 1.3: Endosomal mediated entry of toxins into epithelial cells. 1) The CBD at the C- terminus of the toxins TcdA and TcdB bind to surface receptors on the epithelial cells outer membrane. 2) Once these surface receptors have been activated, an endosome forms around the toxins. 3) The toxins undergo a conformation change inside the endosome. Toxin insertion into the endosomal membrane is facilitated by TD. The GTD enters the cytosol and is cleaved off by activation of CPD. GTD is released into the cytosol. 4) The GTD acts on the Rho GTPases through the following interactions: a) Prevents cycling of Rho between the cytosol and the membrane, b) Inhibits Rho activation, c) Glycosylates Rho to block the active conformation of Rho and d) Blocks Rho/effector interaction. 5) The inhibition of the Rho GTPases interferes with a number of downstream processes such as tight junctions, epithelial barrier, immune cell migration, superoxide production, cytokine production, immune cell signalling, phagocytosis and adhesion. Adapted from Carter, Rood and Lyras, 2010 and Papatheodorou et al. 2010.
The genes for TcdA and TcdB (tcdA and tcdB respectively) are located within a 19.6kb region of the genomic chromosome called the Pathogenicity Locus (PaLoc), along with three regulatory genes: tcdR, tcdC and tcdE. The PaLoc and its protein products are outlined in Figures 1.4 and 1.5. The product of tcdE is a small protein (TcdE) predicted to have three transmembrane domains and is likely a member of the class 1 holins. Holins are small membrane proteins that are typically associated with the terminal lysis of phage infected bacteria53. C. difficile TcdE is similar in structure to holins53,54 and plays a role in toxin release by creating a pore in the bacterial cell membrane24,52-58.
14 INTRODUCTION
The positive regulatory gene tcdR is located upstream of tcdA and tcdB55 and encodes the alternate sigma factor TcdR. TcdR regulates toxin transcription by forming a complex with RNA polymerase that binds to the toxin promoter regions45,52,55,59-61. TcdR is repressed during exponential growth phase but is highly expressed during stationary phase to positively regulate the expression of TcdA, TcdB and TcdE as well as its own expression. The negative regulator tcdC transcribes in the opposite direction to the other four genes and encodes the anti-sigma factor TcdC52,59-61. TcdC negatively regulates transcription by destabilising the TcdR-RNA polymerase complex55,56,62-65. It is thought that TcdC sequesters TcdR, possibly by binding to it, preventing it from binding to RNA polymerase65. This prevents DNA polymerase from attaching to the promoter region of tcdA, tcdB and tcdE thereby preventing transcription and release of the toxins. Transcription of TcdC is high in early exponential growth phase and decreases as the cells enter stationary phase55,56,58,62-64.
15 INTRODUCTION
Figure 1.4: Positive regulation of gene expression in the pathogenicity locus. Positive regulation of gene expression by the alternate sigma factor tcdR. By forming a complex with RNA polymerase, tcdR promotes expression of the toxin genes tcdA and tcdB as well as tcdE and itself. Adapted from Carter, Rood and Lyras, 2010; Dupuy et al., 2006; Lyras et al., 2009; and Tan, Wee and Song, 2001
16 INTRODUCTION
Figure 1.5: Negative regulation of gene expression in the pathogenicity locus. Negative regulation of gene expression occurs when tcdC destabilises the connection between RNA polymerase and tcdR. Adapted from Carter, Rood and Lyras, 2010; Dupuy et al., 2006; Lyras et al., 2009; and Tan, Wee and Song, 2001
In addition to TcdA and TcdB, approximately 10% of C. difficile strains produce a third toxin, C. difficile transferase (CDT)66-68. This toxin belongs to the binary actin-ADP-ribosylating toxins, which includes the C. perfringens iota toxin, C. botulinum C2 toxin, C. spirioforme toxin (CS) and B. cereus vegetative insecticidal toxin (VIP). CDT consists of a 48kDa enzymatic protein (CDTa) and a 99kDa binding and translocating protein (CDTb) 66-68. The cell surface receptor for CDTb is lipolysis-stimulated lipoprotein receptor (LSR) which is highly expressed in the intestine. Monomeric CDTb binds to LSR and undergoes oligomerization to form a heptamer of CDTb. The whole receptor-toxin complex is then endocytosed. Acidification of the endosomal compartment causes a structural change of the heptameric CDTb into a pre-pore that can insert itself into the endosomal membrane, creating a pore that mediates entry of CDTa into the cytosol. CDTa modifies actin to inhibit polymerization, leading to a decrease in the
17 INTRODUCTION
cortical actin cytoskeleton resulting in the formation of microtubule protrusions which extend the cell surface. These microtubule protrusions form a dense extracellular network that can enhance bacterial adherence, thus facilitating colonisation66-69.
1.2.4. The host immune response The immune response caused by C. difficile toxins plays an important role in regulating the disease progression of CDI. An insufficient immune response can result in an overwhelming infection, while too much of a response can exacerbate the intestinal injury caused by the toxins70-72. Asymptomatic carriers will experience an early increase in serum immunoglobulin (Ig) G antibodies against TcdA. Although other antigens and toxins elicit an immune response, it is the response to TcdA that plays a role in disease severity and recurrence. Patients who have an initial episode and develop an early increase in IgM antibodies against TcdA are unlikely to have a recurrent episode. Patients who experience multiple recurrences have very low serum levels of both IgG and IgM antibodies against TcdA70. The cellular stress caused by the toxins triggers the secretion of pro-inflammatory cytokines and expression of leukocyte adhesion molecules which facilitate neutrophil infiltration to the site of infection. Neutrophils are the first line of defence against pathogenic bacteria and can release bactericidal reactive oxygen intermediates, and directly phagocytose the bacteria. Neutrophils also release pro- inflammatory cytokines and chemokines leading to an increased neutrophil response. An intense neutrophil response has been associated with increased severity of disease due to off-target effects that lead to tissue damage and pseudomembranous colitis73. It is likely that the intestinal dysbiosis that made the host susceptible to infection also contributes to the unbalanced immune response that contributes to the pathology of CDI33.
18 INTRODUCTION
1.3. Risk to public health
1.3.1. Sources Historically CDI has been predominantly a heath care associated infection. As previously discussed, C. difficile spores are a source of contamination in the hospital environment, able to survive for many months due to their resistance to cleaning products such as hospital disinfectants and alcohols74,75. The most commonly identified sources within the hospital environment include patients with symptomatic C. difficile, contaminated surfaces, carpets and soft furnishings, as well as the hands of healthcare workers74-76. Often overlooked as a source within the hospital environment and in the community, are asymptomatic carriers74,77. These are persons who have been exposed to and colonised by C. difficile but do not display any symptoms of disease. It is estimated that asymptomatic carriage in the community ranges from 0-17% of healthy adults50,78 and can be as high as 50% in hospitals and aged care facilities populations6. It is thought that asymptomatic carriers may be an important reservoir for this disease6,50. The number of CDI cases reported from community sources is growing, with the proportion of community acquired cases ranging from 15-44% of all cases in the U.S. and Europe79. C. difficile has been found in a variety of environmental samples including water, soil and animal faeces, as well as a number of food sources suggesting that C. difficile may be transmitted as a food-borne pathogen78,80. Studies conducted in Europe, Canada, the U.S. and Australia have found C. difficile spores in ground beef, pork, turkey, raw diet for dogs and cats as well as fruit and vegetable sources including ready-to-eat salads7,78,80-84. The contamination of meats with C. difficile could occur from faecal or environmental contamination during processing or while the animal is still alive. C. difficile has been shown to be quite prevalent on pig farms, isolated from pig faeces and the muscle tissue of rodents from these farms85. There have only been a few studies in the diversity of C. difficile in animals and the overlap of these strains with those found in humans86-88. A study in the Netherlands found ribotype 078 was particularly prominent on pig farms, and that rodents could play a role in transmission between farms89. Another study looking at C. difficile in Australian cattle found that 56% of calves less than 7 days old,
19 INTRODUCTION
3.8% of calves aged 2 to 6 months and 1.8% of adult cattle were colonised with C. difficile. The ribotypes found were 002, 014, 033, 056, 064, 087, 103, 126, 127 and 137, along with 11 that could not be identified86. It is interesting that ribotype 064 was isolated as it is known to cause disease in humans but is not typically associated with animals. Keel et al. performed ribotyping on C. difficile isolates from bovine, equine, canine and human sources and found that there was some overlap in human and zoonotic C. difficile. Overlapping ribotypes include 001, 002, 010, 012, 014, 015, 020, 078 and 12679.
1.3.2. Hypervirulent strains The increase in occurrence and severity of CDI is due in part to the emergence of hypervirulent strains such as the increasingly endemic ribotype 027 and the zoonotic ribotype 078. Ribotype 027 has been implicated in a number of hospital outbreaks in Canada, the U.S. and Europe and there have been isolated cases in Asia, Central America and Australia2,26,90. This ribotype is associated with more severe disease, resistance to antibiotics such as fluoroquinolones, a higher rate of treatment failure and increased mortality. Ribotype 027 has a greater potential for toxin production due to mutations in the toxin regulatory genes2,26,91. This increase in toxin production is thought to contribute to the increased virulence of this ribotype. Ribotype 078 has zoonotic associations and is also considered to be hypervirulent. It is commonly found in neonatal pigs88 and although transmission from pigs to humans has not been established, it is possible that contaminated meat products are a source of community acquired CDI, with rodents being a potential vector88.
1.3.3. Recurrent disease The economic burden of CDI is largely due to recurrent disease92. Recurrent CDI is disease that initially responds to antibiotic therapy but recurs after treatment has been completed. This should not be confused with refractory CDI, which is an initial failure to respond to antibiotic therapy93. Recurrent CDI can be very serious and difficult to treat, often requiring prolonged treatment with antibiotics94. Recurrence could indicate ongoing intestinal dysbiosis, that the initial treatment merely suppressed the infection rather than clearing it, or that there are residual spores present in the gastrointestinal tract that are resistant to the antibiotic therapy.
20 INTRODUCTION
After initial infection 15-30% of patients will have recurrent CDI. A patient’s risk of recurrent CDI increases with each episode. After the initial episode, the risk of recurrence is approximately 20%. After the first recurrence, the risk increases to 40%, and then increases to 60% after two or more recurrences92-94. 1.3.4. Antibiotic resistance C. difficile has a variety of mechanisms for antibiotic resistance, including drug inactivation, target modification and efflux pumps. Table 1.1 provides an overview of antimicrobial resistance mechanisms in C. difficile. Resistance to cephalosporins, clindamycin, fluoroquinolones, β-lactams, macrolides, lincosamides and Streptogramin B classes of antibiotics has been well documented95-98. Reports of resistance to vancomycin and metronidazole are cause for concern98-100 since they are the antibiotics currently used to treat CDI. A study in Spain in 2002 reported that 3.1% of their isolates had intermediate resistance to vancomycin and 6.3% had resistance to metronidazole99.
21 INTRODUCTION
Table 1.1: Mechanisms of antibiotic resistance in C. difficile. Adapted from Harmvoravongchai et al. 2017 Type of resistance Antibiotic Mechanism, proteins (genes)
Drug inactivation Β-lactams Hydrolyse β-lactam ring. Β-lactamases • Antibiotic degradation (CD034, CD0458, CD0464, CD0527, CD0655, CD0829, CD1802, CD2742, CD3651)
• Antibiotic Chloramphenicol Adds acetyl group to CHL, prevents modification (CHL) binding to 50S ribosomal subunit. CHL acetyltransferase (catD) Tn4453a and Tn4453b
• Prevention of Metronidazole • Disruption of iron homeostasis antibiotic activation (MTZ) prevents activity of redox proteins, and detoxification prevents reduction of MTZ • Convert MTZ to non-toxic aminomidazole, NimB homologue. Target modification Macrolides, Methylation of 50S ribosomal subunit, • Target methylation lincosamides, ERY ribosomal methylase B (ermB) streptogramin B Tn5398, Tn6194 and Tn6215 Linezolid Methylation of 23S rRNA ribosomal subunit, rRNA methyltransferase (cfr)
• Target mutation Fluoroquinolones Mutation in DNA gyrase results in conformation change of target to prevent drug binding (gyrA or gyrB) Rifamycins Mutation in β-subunit RNA polymerase (RNAP; rpB) Fidaxomicin Mutation in RNAP-promoter complex (rpoB or rpoC)
Fusidic acid Mutation in elongation factor G (EF-G)
Vancomycin* Mutation in N- acetylglucosaminyltransferase (murG) prevents drug binding peptidoglycan precursor
• Target protection Tetracycline Prevention of TET binding to 30S (TET) ribosomal subunit by ribosome protection protein. Efflux pumps: Secondary multidrug transporters • Major facilitator Erythromycin Clostridial Cme protein family (MFS) (ERY)
• Multidrug and toxic Norfloxacin and Sodium-dependent efflux pump (cdeA) compound extrusion ciprofloxacin (MATE)
*Vancomycin resistance mechanisms are not fully determined
22 INTRODUCTION
Although resistance mechanisms against vancomycin have not been fully determined, there have been studies that explored some possibilities. One study proposed an amino acid mutation in N-acetylglucosaminyltransferase (MurG) confers resistance by target mutation98. Another study by Sebaihia et al. reported finding genes in the genome of the strain C. difficile 630 that were very similar to the vancomycin genes encoding VanG-type resistance in Enterococcus faecalis, however C. difficile 630 shows no in vitro resistance to vancomycin63. It is possible that these genes are non-function or inactive in C. difficile 630, but are contributing to vancomycin resistance in other strains. Metronidazole resistance in C. difficile has been observed to be heterogenic, indicating that this type of resistance could be multifactorial98-100. Metronidazole enters the bacterial cell in an inactive form that needs to be reduced into its active form by redox proteins such as pyruvate:ferredoxin oxidoreductases, pyruvate:flavoredoxins or thioredoxin and thioredoxin reductases. Once active, metronidazole binds non-specifically to the bacterial DNA causing strand breakages that lead to cell death. Iron is a vital component of these redox proteins, so disrupting cellular iron homeostasis, such as a reduction in ferric uptake, could reduce the susceptibility of C. difficile to metronidazole98. Another catalyst to bypassing metronidazole activation in C. difficile is a putative 5-nitromidazole reductase (NimB) homologue. NimB converts metronidazole into aminomidazole which is a non-toxic compound. C. difficile strains with resistance to metronidazole have shown up to a 3-fold increase in NimB homologue expression compared to the wild type98. Other factors contributing to decreased susceptibility to metronidazole and vancomycin is the ability C. difficile has to form biofilms and spores. Biofilms are communities of bacteria within a polymeric matrix and are formed during chronic infection. Resistance to vancomycin is 12-fold higher in biofilm communities of C. difficile than it is in planktonic (non-biofilm) C. difficile. Resistance to metronidazole is 100-fold higher in biofilm communities than it is in planktonic C. difficile98. C. difficile spores are intrinsically resistant to antimicrobial therapy, including vancomycin and metronidazole, a factor that contributes to recurrent and refractory CDI98.
23 INTRODUCTION
1.4. Current and emerging therapies
1.4.1. Diagnosis and treatment guidelines The current Australasian Society for Infectious Diseases (ASID) guidelines for the diagnosis of CDI recommend that diagnostic testing should be performed on unformed stools from all patients hospitalised for greater than 72 hours for toxigenic C. difficile. Screening tests for toxigenic C. difficile should be highly sensitive (>90%) and laboratories need to use a testing algorithm that is highly specific. Diagnostic testing should not be performed on patients who are asymptomatic or minimally symptomatic, have formed or soft stools, and who have diarrhoea that is associated with laxative use. Re-testing of patients as a test of cure is also not recommended93. Treatment guidelines still recommend metronidazole for a mild to moderate initial episode CDI at 400 mg three times daily for 10 days, and vancomycin for a refractory initial episode or severe CDI at 125 mg four times daily for 10 days. A vancomycin taper is recommended for treatment of recurrent CDI. Fidaxomicin is not recommended as a first line therapy, instead should be reserved for refractory or recurrent CDI that has not responded to vancomycin. Further recurrence and treatment failure warrants consideration of faecal microbiota transplant (FMT)93.
1.4.2. Antibiotics Antibiotics routinely recommended for CDI treatment are metronidazole and vancomycin, with newer or alternate antibiotics such as fidaxomicin, tigecycline, and rifaximin suggested as options for refractory cases of CDI that have failed to respond to vancomycin93. There are also a number of emerging antibiotics that are currently undergoing clinical and pre-clinical trials including surotomycin, LFF571 and ridinilazole101. Metronidazole is a broad-spectrum, small molecule nitroimidazole antibiotic (Azoles class)102. Once reduced to its active form, metronidazole targets DNA102 and inhibits cellular enzymatic functions103. Metronidazole is highly absorbed (up to 80%) so only low concentrations are reached in the gastrointestinal tract104-106. Metronidazole is preferred for mild-moderate disease as it is inexpensive and has a reduced risk of inducing vancomycin resistant bacteria such as vancomycin resistant enterococci (VRE)93,107. It cannot be used to treat chronic infections as
24 INTRODUCTION
long-term use could result in cumulative neurotoxicity. Metronidazole has been shown to be inferior to vancomycin in the treatment of CDI, particularly severe cases108, and subinhibitory concentrations have been shown to increase toxin production and sporulation109. Vancomycin is a broad-spectrum glycopeptide antibiotic102. It is slow acting110 and kills rapidly growing C. difficile by inhibiting the later stages of peptidoglycan biosynthesis102. Oral vancomycin is recommended for severe CDI as it has superior efficacy, good pharmacokinetics (poorly absorbed) and has less side effects than metronidazole105,108. Vancomycin use does increase the risk of acquiring vancomycin-resistant organisms such as VRE. Vancomycin is able to suppress spore formation110, but is ineffective against already formed spores108 and subinhibitory concentrations increase toxin production109. There is also evidence that subinhibitory concentrations of vancomycin can increase the number of spores produced by some strains of C. difficile111. Vancomycin and metronidazole are becoming less effective in their treatment of CDI. There are continual reports of failure of these therapies, with failure as high as 38%4,112,113. The high failure rates and potential for antibiotic resistance against these current therapies are cause for concern and highlight the need for new effective therapies. Fidaxomicin is a macrocyclic class antibiotic that exhibits narrow-spectrum activity113-116. It inhibits DNA transcription by targeting the bacterial RNA polymerase102,116,117. Fidaxomicin is minimally absorbed and has limited activity against components of the normal gut flora114-116 Clinical cure rates with fidaxomicin are similar to those of vancomycin, but the rate of recurrence is significantly lower. The risk of recurrence when infected with hypervirulent ribotypes such as ribotype 027 is the same as that of vancomycin116. There is evidence to suggest that fidaxomicin can also prevent spore germination108. Tigecycline is a broad-spectrum glycycline class antibiotic that inhibits protein synthesis. It has strong activity against C. difficile, does not increase toxin production and is capable of reducing spore viability109. Rifaximin is a synthetic analogue of rifamycin, a broad-spectrum ansamycin class antibiotic. Rifaximin is primarily used for the treatment of traveller’s diarrhoea due to its effectiveness against several Gram-positive and Gram-negative bacteria. It is poorly absorbed
25 INTRODUCTION
and has no systemic adverse events associated with its use. Rifaximin acts by inhibiting RNA synthesis and has good activity against C. difficile, however it may be too disruptive of the intestinal microbiota to prevent recurrent CDI118,119. Surotomycin is a narrow-spectrum cyclic lipopeptide antibiotic that is currently undergoing Phase III clinical trials. It is minimally absorbed and demonstrated to have dose-dependent bactericidal activity against C. difficile101. Surotomycin is chemically and structurally similar to daptomycin, which targets the bacterial membrane102. Surotomycin acts by depolarising the bacterial membrane which leads to a loss of proton gradient and cell death103. Ridinilazole is a narrow- spectrum, first-in class antibiotic that is poorly absorbed106. It has demonstrated good activity in vitro and in an in vivo hamster model120 and has been shown to have a good safety and tolerability profile in Phase I clinical trials106. Ridinilazole reduces C. difficile vegetative cells and reduces C. difficile TcdA and TcdB production120 and has little antimicrobial activity against intestinal microbiota106. The mechanism of action for this antibiotic are not well understood120. LFF571 is a semisynthetic thiopeptide antibiotic currently undergoing Phase I clinical trials. LFF571 has shown good efficacy in vitro121,122 and in an in vivo golden hamster model, and has been shown to be safe and well tolerated in Phase I clinical trials122. It is poorly absorbed with limited systemic exposure and high faecal concentrations122.
1.4.3. Faecal microbiota transplant (FMT) Faecal microbiota transplant (FMT) treats intestinal dysbiosis, and thereby recurrent CDI, by restoring the intestinal microbiota to a healthy state30,93,123-127. This is done by implanting faecal material, either donated or synthetic, into the distal colon of the gastrointestinal tract123,126,127. Once the intestinal microbiota has been re-established, microbiota functions such as metabolism, immune response modulation and colonization resistance are also restored123. FMT has been proven to be fast and effective at resolving recurrent CDI, with worldwide success rates of up to 90% being reported93,124,125,127-131. FMT is typically performed using fresh donated faeces from a healthy donor119,123-125,127,132. Fresh stool is generally preferred since it is considered to have the most viable organisms, there are also a number of risks that need to be
26 INTRODUCTION
addressed. One of the biggest risks associated with FMT is the introduction of new pathogens47 and donors need to be screened for the pathogens in Table 1.2 below. This screening process is expensive, delays treatment and there is still a risk of transmitting pathogens that are either outside the period of detectability or for which diagnostic tests are unavailable129. Donors should also have had no antibiotics in the past 3 months, no gastrointestinal symptoms, immune disorders or other diseases outlined in Table 1.2, as well as no intravenous drug use, high- risk sexual behaviours, tattoos, body piercings in the past 6 months and no history of incarceration47. There are also ethical considerations that need to be accounted for, such as coercion and confidentiality132. These risks and ethical concerns have prompted studies on the viability of frozen stool banks using anonymous volunteer donors. These studies determined that FMT using stool frozen less than 2 months is as clinically effective as fresh samples for recurrent CDI132,133, and frozen stool stored in 10% glycerol for 2-10 months has a first dose efficacy of 88%, and a second-dose efficacy of 100%132. Frozen stool banks could be an effective and reliable way to standardize donor stool processing, screen for pathogens and allow stool to be on demand for clinicians3,132,133.
Table 1.2: Donor screening for faecal microbiota transplant. Pathogen screening Medical/disease screening • Hepatitis A, B and C • History of diarrhoea • HIV • History of constipation • Syphilis • Inflammatory bowel disease • Clostridium difficile • Irritable bowel syndrome • Giardia • Immune disorders • Rotavirus • Concurrent immunosuppressive agents • Cryptosporidium • Malignancy • Cyclospora • Diabetes mellitus type II • Isopoda • Metabolic syndromes • Ova parasites • Major gastrointestinal surgery • Shiga toxin-producing E. coli • Chronic pain syndromes • Salmonella • Autoimmune disorders • Shigella • Yersinia • Campylobactor • Non-cholera vibrio • Plesiomonas • Aeromonas
There is a lack of standardization of FMT protocols. Delivery of the faecal material can be via colonoscopy, nasogastric tube, retention enema or orally delivered
27 INTRODUCTION
encapsulated stool126. Different methods of delivery have different levels of efficacy, and associated risks. Methods such as colonoscopy and nasogastric tube have an efficacy of 80%-90%129, but have higher risks associated with them including infection, abdominal pain, discomfort, or bleeding95. Colonoscopy also presents the risk of bowel perforation95. Lower risk methods such as retention enema and oral encapsulation FMT have a significantly lower first-dose efficacy of 50%-70%129,134. A study by Khanna et al. found that the low efficacy of oral encapsulated FMT can be improved by using “synthetic” faeces. Named SER-109, this faeces alternative is composed of approximately 50 species of Firmicutes spores, one of the dominant phyla in a healthy intestinal microbiota, which were derived from the stool specimens of healthy donors. This treatment is based on the hypothesis that spore-forming organisms would compete with C. difficile for essential nutrients, and that the ethanol preparation would reduce the risk of transmitting other pathogens. This study found that SER-109 achieved clinical resolution of recurrent CDI in 96.7% of patients and resulted in rapid diversification of the intestinal microbiota129. The risks, strict donor screening requirements, invasiveness of the procedure and the perceived unpleasantness associated with FMT means that it is typically reserved as a treatment for severe recurrent or refractory CDI, rather than as routine therapy. A study investigating patient attitudes to FMT found that although most FMT patients were satisfied with the outcome of FMT, they acknowledge that they only agreed to the procedure because they were desperate for a cure127. Most patients find the procedure unappealing and experience feelings of embarrassment and discomfort when discussing FMT with their doctor and donor.
1.4.4. Probiotics Probiotics are live bacteria or yeast which confer a health benefit to the host when they are consumed in adequate amounts. A number of different approaches have been investigated for the use of probiotics as a therapeutic or preventative method in the management of CDI. Commonly investigated probiotics for prevention of CDI include lactic acid bacteria such as Lactococcus and Lactobacillus spp. and the yeast Sacharomyces boulardii.
28 INTRODUCTION
Lactic acid bacteria play critical roles as commensals in the gastrointestinal tract. They are ideal candidates for probiotics as they are able to survive transit through the stomach, can adhere to the intestinal epithelium, have immunomodulatory properties and are safe to consume in the quantity required to confer health benefits135. There have been a number of investigations into the role of Lactobacillus spp as probiotics. L. delbrueckii, have been shown to inhibit C. difficile adhesion to Caco-2 cells and reduce the cytotoxic effects of TcdA and TcdB136. L. casei, which was shown to suppress interleukin (IL)-8 production in C. difficile stimulated HT-29 cells136. L. rhamnosus, which was also able to inhibit C. difficile adherence to Caco-2 cells137 and suppress IL-8 production136 as well as induce immunocompetent cells to produce cytokines including IL-10, IL-12, tumor necrosis factor (TNF)-α and interferon (IFN)-γ. The cytokines IL-12 and IFN-γ are implicated in innate defence mechanisms in response to bacteria137. Lactococcus lactis has been used as a live delivery vehicle for a C. difficile vaccine. L. lactis modified to express the binding domains of TcdA and TcdB given orally or subcutaneously triggered an immune response against TcdA and TcdB in vivo. This immune response protected the animals from CDI and recurrent CDI135. S. boulardii secretes a protease that interferes with the binding of TcdA and TcdB to the receptor on the surface of human colonic epithelial cells. It has been demonstrated that S. boulardii significantly reduces recurrent episodes of CDI when used in combination with metronidazole or vancomycin136. Another approach to preventing CDI through the use of probiotics is non-toxigenic C. difficile. Colonizing the host with non-toxigenic C. difficile follows the colonization competition hypothesis, where the non-toxigenic C. difficile prevents the toxigenic C. difficile from colonizing the distal colon by occupying the same ecologic niche, thereby preventing disease. Candidates for this type of probiotic therapy include C. difficile strain M3, which has been shown to be well tolerated, able to colonize the gastrointestinal tract and significantly reduce recurrent CDI, as well as C. difficile strain CD37, which is a poor sporulator and provided significant protection against toxigenic C. difficile in a mouse model of CDI. Although these studies were promising, non-toxigenic C. difficile should be approached with caution as there is the possibility of non-toxigenic C. difficile converting to toxigenic strains138. This has not been reported in vivo, but has been demonstrated
29 INTRODUCTION
by Brouwer et al. in vitro, who were able to observe the horizontal transfer of the pathogenicity locus from a toxigenic strain to a non-toxigenic strain139.
1.4.5. Vaccines and antibodies The immune response to TcdA and TcdB has been shown to play a significant role in disease severity and the risk of recurrence140-144. There is evidence that high concentrations of immunoglobulin (Ig) G to TcdA is associated with asymptomatic carriage, while a low immune response is associated with increased risk of recurrent CDI143. Similarly, the presence of high serum IgG to TcdB is capable of neutralizing the cytotoxic activity of TcdB and correlated to clinical recovery from CDI without relapse143. The importance of antibodies against the toxins in protection from CDI has led to the development of immune based therapies primarily aimed at preventing recurrent C. difficile. Karczewski et al. tested the efficacy of a vaccine consisting of two separately expressed recombinant fragments of TcdB combined with full-length, formaldehyde inactivated TcdA in an in vivo hamster challenge model, and found that the hamsters developed a high titre of neutralizing antibodies that last approximately 3 months145. A toxoid vaccine developed by Sheldon et al. is currently in Phase II clinical trials, and has proven to be safe and well tolerated with a strong increase in neutralizing antibodies after 2-3 doses140. A recombinant fusion protein (CLA84) containing portions of the receptor binding domains from TcdA and TcdB has demonstrated to provide protection from CDI in an in vivo hamster challenge model, inducing high levels of serum antibodies142,146. VLA84 has also been shown to be safe and well tolerated in Phase I clinical trials142. A bivalent toxoid vaccine developed by de Bruyn et al. is also undergoing Phase II clinical trials after proving to be safe and well tolerated in Phase I clinical trials, eliciting a strong immune response in both healthy adult and elderly subjects144. Vaccines have great potential for CDI prevention, but are limited in terms of treating CDI due to the time required for the patient’s immune system to respond. This has prompted a number of investigations into monoclonal antibody therapy143. Monoclonal antibody therapies currently undergoing clinical and pre-clinical trials include a fully humanised monoclonal antibody cocktail against TcdA and TcdB. This demonstrated strong neutralizing activity both in vitro and in an in vivo
30 INTRODUCTION
hamster challenge model143 A combination of two monoclonal antibodies actoxumab (TcdA specific) and bezlotoxumab (TcdB specific) are currently in Phase III clinical trials after reducing the rate of recurrence in Phase II clinical trials by 73% among CDI patients treated with standard therapy147. Once again the latter is associated with disease prevention rather than treatment. Another antibody approach is hyperimmune bovine colostrum (HBC). Colostrum is the first milk collected from a lactating mammal after parturition and is rich in IgG. When a gestating dairy cow is repeatedly immunised with recombinant mutants of TcdA and TcdB, she will produce HBC with IgG against the toxins. HBC has shown promise as an effective therapy for CDI, but is still in the early stages of development148.
1.5. Auranofin
1.5.1. Background and overview of pharmacokinetics Auranofin [2,3,4,6-tetra-o-acetyl-1-β-D-glycotranosato-S-(triethyl-phosphine) gold] is a monomeric gold species with phosphine and thiol ligands in the linear arrangement149-152 shown in Figure 1.6 below153. Auranofin was introduced into clinical use in 1985 as an alternative treatment for rheumatoid arthritis to the injectable gold sodium thiomalate. However, the clinical side effects, poor pharmacokinetics, and concerns regarding the immune suppression that may be associated with long term use of auranofin saw a general preference to continue using gold sodium thiomalate154. As an already approved drug, repurposing auranofin has many benefits over novel drug discovery as it has already been proven safe for use in humans. This removes a significant portion of the time and cost constraints associated with drug discovery and development.
31 INTRODUCTION
Figure 1.6: The chemical structure of auranofin. The gold-thiol complex (Au-S) of this compound is responsible for its reactivity with thiol-redox enzymes such as thioredoxin reductase. From http://www.jmfinechemicals.com/product/auranofin/
Only 15-25 % of an oral dose can be detected in the plasma154,155, most of which was absorbed via the gastrointestinal tract within 20 minutes of being administered156. In the plasma auranofin binds predominantly to albumin, reaching a peak concentration of 6-9ug/100ml within 1-2 hours. The plasma half-life is 15-25 days with almost total body elimination after 55-80 days154,156. Auranofin is minimally retained in tissues with 85% of the administered dose excreted in the faeces and only 15% appearing in the urine150. Only 0.4% of the administered dose is concentrated in the kidneys157. This pharmacokinetic profile is ideal for treating CDI since the organism infects the distal colon and a large proportion of the administered dose will reach the site of infection. The most common side effect of treatment with auranofin is the development of soft and loose stools within the first few weeks of treatment. This may occur in up to 40% of patients, with 2-5% reporting watery diarrhoea150,154. It is suspected that this diarrhoea is caused by the effect auranofin has on sodium and water transport in the intestine. Diarrhoea associated with auranofin use can be alleviated by reducing the dose or stopping treatment, depending on severity, and will settle promptly once the drug is stopped. Sometimes dividing the dose or changing the timing will alleviate this problem155. Jackson-Rosario et al. were able to show that auranofin inhibited the growth of C. difficile in vitro at a concentration of 2 µM24. The standard dose of 3-6 mg auranofin for the treatment of rheumatoid arthritis is estimated to result in a concentration of 0.15 mM in the colon150,154. It is therefore reasonable that the required dose of auranofin for the treatment of CDI will be
32 INTRODUCTION
10-fold lower than that needed for rheumatoid arthritis, and is unlikely to induce diarrhoea. Another side effect is a mild rash that occurs within the first year of treatment150,154. This rash is less common with auranofin treatment than it is with injectable gold compounds, where it can occur in up to 20% of patients, with 2-3% having to discontinue therapy due to a severe rash150. Other less common side effects include stomatitis, which affects 1-12% of patients, proteinuria which affects 5% of patients, conjunctivitis which affects 10% of patients, and thrombocytopenia and related bone marrow suppression, which are extremely rare. These less common side effects are associated with long-term auranofin treatment150,154.
1.5.2. Mechanisms of action The main mechanism of action of auranofin is through inhibition of reduction/oxidation (redox) enzymes such as thioredoxin reductase (TrxR). The thiol ligand contained in auranofin has a high affinity for thiol and selenol groups, to which it forms a stable and irreversible adduct24,158. Redox enzymes such as TrxR are essential to many cellular processes; particularly in maintaining the intracellular levels of reactive oxygen species (ROS)158-161. Controlling the level of ROS to prevent the resulting DNA damage, is critical for the survival of all cell types, including cancer cells, parasites, and memory T cells that harbor pro-viral HIV DNA. These particular cell types all overexpress redox enzymes increasing the affinity of auranofin towards these cells158-161. Inhibition of redox enzymes alters the redox state of the cell which can lead to increased production of hydrogen peroxide and reactive oxygen species (ROS) that causes cellular oxidative stress and ultimately intrinsic apoptosis158-161. Auranofin has particularly potent activity for selenoproteins and selenium dependent enzymes because it also has a high affinity towards inorganic selenium in the form of selenide (HSe-). Auranofin and HSe- are able to form a stable adduct through displacement of the sulphur in the auranofin thiol with the Se in HSe-24, resulting in a hydrogen sulphide by-product and the auranofin selenium compound. Selenoproteins such as TrxR in mammalian cells, have been identified as a potential drug target for cancer158, while selenoproteins involved in Stickland
33 INTRODUCTION
reactions in amino acid fermenting bacteria have been identified as a potential drug target for these bacterial infections24,162.
1.5.3. Potential as a treatment for CDI Auranofin is being investigated as a treatment for a number of diseases including cancer, parasitic infections, bacterial infections, HIV and neurodegenerative disorders such as Parkinson’s disease and Alzheimer’s disease163. Auranofin shows potential as a narrow spectrum treatment for CDI as it targets the selenium dependent metabolic pathways used by C. difficile. Inhibiting these pathways also inhibits C. difficile growth, highlighting the dependence of this pathogen on these pathways24. Auranofin is able to block the primary metabolite selenium, rather than target individual metabolic pathways, which will limit the ability of the bacteria to acquire resistance to this therapy162. A computational analysis of bacterial genomes indicates that only 14% of bacteria utilise selenoproteins, so selenium targeted therapy can be considered narrow spectrum, minimising disruption to the intestinal microbiota and reducing the risk of recurrence24,162. Further to this, it has been shown that clostridia spp. that do not use selenoproteins, such as C. perfringens and C. tetani are not affected by auranofin. Clostridia belong to the phyla Firmicutes, one of the dominant phyla present in the intestinal microbiota24. No further work on auranofin as a treatment for CDI has been published, and there is no data available on the effects auranofin has on C. difficile sporulation and toxin production, or intestinal inflammation With increasing rates of treatment failure and recurrent CDI, there is a clear and urgent need for new therapies. There are many treatments currently being investigated including new antibiotics, FMT and immune therapies. These therapies have either shown little efficacy, are expensive, unpleasant and invasive, or are in the early stages of development and therefore will not be used in routine clinical therapy for some time. It is the aim of this study to assess the potential of auranofin as a treatment for CDI by continuing the work performed by Jackson-Rosario et al. The hypothesis for this study is that auranofin will prove to be an effective therapy for CDI that has a lower risk of recurrence than the current treatments due to a smaller impact on the host’s microbiota.
34 INTRODUCTION
In order to achieve this main objective, assessing the validity of auranofin as a treatment for CDI, two aims have been devised. Aim 1 is to use in silico methods to identify potential selenoprotein genes in C. difficile 630 and confirm the conservation of these genes in a panel of clinical isolates. Aim 2 is to test the activity of auranofin on C. difficile using in vitro assays and then in an in vivo mouse model.
35 METHODOLOGY
2. Methodology
2.1. Identifying selenoproteins in the C. difficile using in silico methods The overall process used to identify selenoproteins in the C. difficile genome is outlined in Figure 2.1. For this in silico work, C. difficile 630 (NCBI164; accession number: NC_009089) was used as it is the model genome for C. difficile in silico work. The genome for C. difficile 630 was searched for selenocysteine insertion sequence (SECIS) elements using bSECISearch165 (Figure 2.1, process 1), an internet based program designed by Zhang and Gladyshev. This program identifies bacterial SECIS elements based on a bacterial SECIS consensus model. The key features of this consensus model, shown in Figure 2.2, are an apical loop of 3-14 nucleotides (nt) with at least 1 conserved guanosine (G) among the first 2 nt, an upper stem of 4-16 base pairs (bp) and a spacing of 16-37 nt between the UGA codon and the apical loop165. The resulting sequences from bSECISearch were aligned to the annotated C. difficile 630 genome using a nucleotide BLAST166 (NCBI, Blast+2.4.0; Figure 2.1, process 2) to identify the SECIS elements that were within the open reading frame of a gene. Any SECIS sequences that were not within an open reading frame were discarded. The gene sequences from BLAST were then put back through bSECISearch to confirm they contain a SECIS element (Figure 2.1, process 3). Gene sequences that did not were discarded. The secondary structure of the remaining SECIS elements was determined using RNAfold167-169 (Vienna RNA Webserver) and compared to the consensus SECIS element secondary structure determined by Zhang and Gladyshev (Figure 2.1, process 4). Those that did not conform to the consensus structure were discarded. The final list of potential selenoproteins were aligned with the genomes of 100 C. difficile strains in the Clostridium difficile txid1496 [ORGN] database (listed in Appendix 1) using nucleotide BLAST166 (Figure 2.1, process 5) to determine the level of conservation of these genes.
36 METHODOLOGY
Figure 2.1: Flow diagram of the in silico method used to identify C. difficile selenoproteins.
37 METHODOLOGY
Figure 2.2: Bacterial SECIS element consensus model. From Zhang and Gladyshev 2005.
2.2. Investigating auranofin as a treatment for CDI using in vitro methods
2.2.1. Routine in vitro methodology
Bacterial culture methods All C. difficile isolates were stored in a glycerol stock at -80°C. When removed from storage, isolates were streaked for single colonies on brain heart infusion (BHI; Oxoid, Hampshire, England) agar with 0.1% (w/v) taurocholic acid sodium salt hydrate (Sigma-Aldrich, NSW, VIC) to allow for spore germination. All bacterial work including incubation periods were undertaken in a Don Whitley DG250 Workstation (Don Whitley Scientific, NSW, AUS), which maintained an anaerobic environment (10% H2, 10% CO2 in N2; BOC, NSW, AUS) at 37°C and 75% humidity. For experiments using broth cultures, tryptone yeast extract (TY) broth containing 1% (w/v) tryptone (Oxoid), 0.5% (w/v) yeast extract (Oxoid), and 1% (w/v) sodium chloride (NaCl, Biochemicals, NSW, AUS) in deionised water was used at it is a low nutrient media that supports toxin production and sporulation. All media was anaerobically conditioned prior to use by incubating in anaerobic conditions overnight. The strain used for growth assays and the animal model was C. difficile M7404. This strain belongs to the hypervirulent ribotype 027 and has
38 METHODOLOGY
been responsible for outbreaks in Canada, UK and North America. The M7404 is a commonly used strain for controlled laboratory experiments. This strain was donated by Associate Professor Dena Lyras, Monash University, Clayton, Victoria.
Tissue culture methods Routine culturing of Vero cells in Minimum essential media (MEM-alpha; Gibco, Life Technologies, NY, USA) containing 10% fetal bovine serum (FBS; HyClone, Thermo, Victoria, AUS), 100 units/ml penicillin-streptomycin (HyClone, Thermo) in T25 cell culture flasks (Nunc, Thermo) at 37℃ with 5% CO2 until 80% confluent and then passaged.
Preparation of antibiotic stock solutions Auranofin (Enzo Life Sciences, United Biosciences, NSW, AUS) was prepared by dissolving 100 mg Auranofin in 100% ethanol (Chem-supply, Westlabs, VIC, AUS) to make a stock solution of 5895.36 µM. Metronidazole (Sigma-Aldrich) was prepared by dissolving 1.7 mg metronidazole in deionised water to make a stock solution of 1000 µM. Vancomycin (Sigma-Aldrich) was prepared by dissolving 10 mg of vancomycin hydrochloride in deionised water to make a stock solution of 3449.94 µM. Growth assays
Further dilutions of these antibiotic stock solutions was done using deionised water as necessary. All concentrations in experiments take into account the volume of the media to which the antibiotic is added, and therefore refer to the final concentration of the antibiotic.
2.2.2. Growth Assays
Minimum inhibitory concentration To determine the minimum inhibitory concentration of auranofin, metronidazole and vancomycin required to inhibit C. difficile growth, a growth assay was set up in sterile 96 well, F-bottom cell culture plate (Cellstar, Greiner Bio-one, Interpath, VIC, AUS). The plates were prepared with titrations of auranofin serially diluted 1:2 in anaerobically conditioned sterile TY, with a starting concentration of 500 µM. The final column contained 0 µM. Controls were no treatment, which was
39 METHODOLOGY
sterile TY, and the diluent for auranofin, ethanol, which was titrated 1:2 at a starting concentration of 2% (v/v), which was equivalent to the volume of auranofin added to the highest concentration of the treatment group. Each well was inoculated with 100 µl of an overnight culture of C. difficile M7404 or an additional 100 µl of sterile TY for the no bacteria control. Plates were incubated anaerobically at 37℃ for 24 hours. Bacterial was growth tracked by measuring optical density (OD600nm) every 2 hours from inoculation for 10 hours and again at 24 hours using an Epoch 2 microplate reader (Biotek, Millennium Biosciences, VIC, AUS). MIC was defined as the minimum concentration where there was statistically less bacterial density than the control cultures. This represents inhibition of bacterial growth. Statistical analysis was performed using a 2-tailed unpaired t-test with 95% confidence interval.
These assays were based on the broth diluton method for antimicrobial testing described by Weigand et al.170
Auranofin growth assay A number of growth assays were performed to assess the ability of auranofin to inhibit C. difficile bacterial growth. These were based on experiments described by May et al., Merrigan et al. and Vohra et al.171-173 An overview of the method is provided in Figure 2.3. Concentrations tested include 0.1 µM (subinhibitory), 6 µM (Minimum inhibitory concentration; MIC), 50 µM, 200 µM and 500 µM. The last three concentrations are an equivalent to those used in the preliminary in vivo animal model. Controls in these assays include no treatment and the diluent ethanol, used at an equivalent volume to the auranofin treatment group in the respective assay. For each assay, 40 ml of TY broth (n=3) were inoculated with 1% (v/v) overnight culture of C. difficile M7404 and incubated anaerobically at 37℃ in an anaerobic chamber. Treatments were applied at 12 hours, which is the end of exponential growth. This time point is when the bacteria would be releasing toxins, which would be consistent with the treatment of disease in a clinical setting. OD600nm was measured at 0, 2, 4, 6, 8, 10, 12, 14, 16, 24 and 48 hours using an Epoch 2 microplate reader as a measure of bacterial growth. Samples were taken at 0, 12,
40 METHODOLOGY
24 and 48 hours for viable cell counts (VCC) and viable spore counts (VSC). Additional samples were taken at 48 hours and filtered using a 0.22 µm Millex-GP filter unit (Merck, VIC, AUS) for cell cytotoxicity assays (CCA). Statistical analysis of the growth assays and cell viability was performed using a 2-tailed unpaired t-test with 95% confidence interval.
Figure 2.3: Overview of the in vitro growth assay used to investigate auranofin as a treatment for CDI.
Auranofin vs metronidazole vs vancomycin To compare the activity of auranofin against C. difficile with that of vancomycin and metronidazole MIC assays and growth assays were performed. The MIC assays were carried out as described above (2.2.2) with the addition of metronidazole and vancomycin plates. Each treatment had its own positive and negative controls on its sterile 96 well, F-bottom cell culture plate to eliminate excess variance from being across multiple plates. Growth assays were similar to those described above (2.2.2) with the addition of a set of groups that were treated at inoculation (0 hours) as well as the previously described 12 hour treatment groups. As described above, OD600nm was measured at 0, 2, 4, 6, 8, 10, 12, 14, 16, 24 and 48 hours, VCC and VSC samples taken at 0, 12, 24 and 48 hours. Statistical analysis of the growth assays and cell viability was performed using a 2-tailed unpaired t-test with 95% confidence interval.
Viable cell counts (VCC)
41 METHODOLOGY
Cell viability was assessed by VCC performed from samples taken at 0, 12, 24 and 48 hours from the growth assay described above. Samples taken at each of these time points were serially diluted and 100 µl of each dilution from neat to 10-5 were inoculated onto BHI agar and incubated for 48 hours. Colonies on each plate were then counted and the average colony forming unit (CFU) was calculated using Equation 2.1. These assays were based on experiments described by May et al., Merrigan et al. and Vohra et al.171-173
Equation 2.1: Average colony forming unit (CFU) 퐴푣 푛푢푚푏푒푟 표푓 푐표푙표푛푖푒푠 × 푑푖푙푢푡푖표푛 푓푎푐푡표푟 퐶퐹푈 = 푉표푙푢푚푒 표푓 푖푛표푐푢푙푢푚
Viable spore counts (VSC) Viability of the C. difficile spores was assessed by viable spore counts (VSC) performed from samples taken at 0, 12, 24 and 48 hours from the growth assay described above. Samples taken at each of these time points were heat killed at 75℃ for 30 minutes to kill all vegetative cells. Samples were then serially diluted and 100 µl of each dilution from neat to 10-5 were inoculated onto BHI agar with 0.1% (w/v) taurocholic acid sodium salt hydrate to germinate the spores. Plates were incubated for 48 hours and CFU was calculated as described above using Equation 2.1. Control plates for vegetative cells were also done by spreading the neat heat killed samples onto BHI agar without the germinant. As there was no germinant in this media, any colonies that grow on these plates likely grew from a vegetative cell. These assays were based on experiments described by May et al., Merrigan et al. and Vohra et al.171-173
Cell cytotoxicity assay (CCA) Cytotoxic activity of the toxins was assessed using CCA, where the cytotoxic activity of Toxin B is tested with Vero cells. An overview of the method is provided in Figure 2.4. Samples taken from the growth assays at 48 hours were sterile filtered using a 0.22 µm Millex-GP filter unit and stored at 4°C. Vero cells at approximate 80% confluency were passaged into 2 sterile 96 well, F-bottom cell culture plates at a concentration of 0.1x106 cells/ml and incubated at 37℃ with 5% CO2 until
42 METHODOLOGY
80% confluent. Culture media was removed and the cells were washed with sterile 1X Phosphate buffered saline [PBS; 0.52% (w/v) sodium phosphate monobasic (NaH2PO4; Biochemicals), 2.3% (w/v) sodium phosphate dibasic (Na2HPO4; Biochemicals), and 8.76% NaCl) and 150 µl of sterile MEM was added to each well. Filtered C. difficile samples were added to the wells as follows: wells 1 and 24 of each row were designated non-stimulated (NS) and contained Vero cells in MEM alone, well 2 received 150 µl of filtered C. difficile or media control to make a 1:2 dilution of the original sample which was then serially diluted across wells 3 to 23. The negative control was sterile MEM. An additional set of control samples from the growth assays (C. difficile + no treatment) was set up as described and treated with 0.1 µM and 50 µM of auranofin to directly test the effects of auranofin on the cytotoxic activity of Toxin B. The C. difficile toxin treated Vero cells were incubated for 20 hours at 37℃ with 5% CO2, after which the cells were observed for the lowest titration where all Vero cells were toxin affected (represented by the yellow circles in Figure 2.4). This titration is the cytotoxic end-point titre. These assays were based on the methods described by Just and Gerhard, Lyras et al. Merrigan et al. and Vohra et al.25,61,172,173
43 METHODOLOGY
Figure 2.4: Overview of cell cytotoxicity assay (CCA). Used to test the level of cytotoxic activity of C. difficile TcdB against Vero cells. Yellow circles represent the lowest titration with no viable Vero cells remaining.
2.3. Preliminary testing of auranofin as a treatment for CDI using an in vivo animal model
2.3.1. CDI prophylaxis animal model To assess the potential of auranofin as a treatment for CDI, it was trialled in an in vivo murine model of CDI as a prophylaxis. This work was done with the assistance of Associate Professor Dena Lyras and her research group at Monash University, VIC, Australia. This model is described by Lyon et al. and Carter et al.174,175 All animal handling and experimentation were performed in accordance with institutional guidelines (Monash University AEC no. MARP/2014/142).
44 METHODOLOGY
The in vivo model is outlined in Figure 2.5 below and the treatment and control groups are defined in Table 2.1. Male C57BL/6J mice (Walter and Eliza Hall Institute of Medical Research, VIC, AUS) at 6-7 weeks old were pre-treated with an antibiotic cocktail in their drinking water for 7 days to induce intestinal dysbiosis, making them susceptible to infection from C. difficile spores. The antibiotic cocktail contained 0.4 mg/ml kanamycin (Amresco, Astral Scientific, NSW, AUS), 0.035 mg/ml gentamicin (Sigma-Aldrich), 850 U/ml colistin (Sigma- Aldrich), 0.215 mg/ml metronidazole, 0.045 mg/ml vancomycin, and 0.3 mg/ml cefaclor (Sigma-Aldrich).This was then followed by 2 days on cefaclor alone. Auranofin prophylaxis was started one day before infection at the doses of 400 µg/ml (maximum dose able to be given due to the solubility of auranofin), 200 µg/ml (1:2 maximum dose), and 40 µg/ml (1:10 maximum dose). Auranofin was given in the drinking water. Mice were infected with C. difficile by oral gavage of 106 C. difficile M7404 spores, and were monitored twice daily for signs of disease including weight loss, behavioural and physical changes, and diarrhoea. On the day of infection antibiotic pre-treatments were stopped and mice were given either the auranofin drinking water or plain drinking water. Mice were humanely killed by CO2 asphyxiation when the following endpoints were met: acute weight loss of greater than 10% body weight within any 24 hour period, chronic weight loss of greater than 15% body weight following infection, animals becoming moribund or showing any signs of irregular or laboured breathing, animals reaching a clinical severity score of 2 on activity, alertness, coat, eyes, diarrhoea, nose, dehydration, movement or food consumption. Faeces were collected at 24 hours post infection to enumerate C. difficile spore load by sporulation count. Faecal pellets were resuspended in sterile PBS (100 mg/ml), heat shocked for 30 minutes at 65℃ and plated on heart infusion (HI; Oxoid, Hampshire, England) agar containing 1.5% glucose (Merck), 0.1% (w/v) sodium taurocholate (New Zealand Pharmaceuticals, Palmerston North, NZ), 0.1% (w/v) L-cysteine (Sigma-Aldrich), 250 µg/ml D-cycloserine (Sigma-Aldrich), 8 µg/ml cefoxitin (Fluka, NSW AUS), 10 µg/ml erythromycin (Amresco), 12 µg.ml norfloxacin (Sigma-Aldrich, NSW, AUS), and 32 µg/ml moxalactam (Sigma- Aldrich). Plates for sporulation count were incubated and CFU was calculated as previously described (2.2.2).
45 METHODOLOGY
Figure 2.5: Overview of Monash CDI prophylaxis mouse model.
Table 2.1: Treatment groups for in vivo trial of auranofin as CDI treatment. Treatment group Infected Uninfected Auranofin 400 µg/ml n=5 n=5 Auranofin 200 µg/ml n=5 n=5 Auranofin 40 µg/ml n=5 n=5 Control (no treatment) n=5 -
46 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
3. Identifying selenoproteins in C. difficile using in silico methods.
3.1. Background Selenium is an essential micronutrient for many organisms176. The primary role of selenium in a biological system is as a metal cofactor in a number of proteins known as selenoproteins162,177,178. Selenoproteins are found in all domains of life and primarily function as oxidoreductases179,180, using the selenium containing amino acid selenocysteine at the active site. Estimates on the proportion of bacteria that utilise selenium range from 14%24 to 33.7%181. Currently there are only four selenoproteins identified in C. difficile; selenide water dikinase (SelD), the glycine reductase proteins GrdA and GrdB and the D-proline reductase protein PrdA. SelD is a part of the sel operon, the protein products of which play a role in selenocysteine synthesis and incorporation. Glycine reductase and D-proline reductase reduce the amino acids glycine and D-proline respectively through Stickland reactions for energy. It is not known how many other selenoproteins C. difficile may utilise. It is the aim of this work to identify potential C. difficile selenoproteins. Selenocysteine differs from cysteine by having a selenol (SeH) group instead of a thiol (SH) group182, which is more acidic. This makes selenocysteine ionized at a physiological pH (7.6), giving it increased catalytic activity compared to cysteine. Selenocysteine is an unusual amino acid in that it is encoded by an in-frame UGA or stop codon. Insertion of selenocysteine requires specialised proteins and insertion machinery. In prokaryotes, there are four proteins required for selenocysteine incorporation along with an mRNA stem-loop structure called the selenocysteine insertion sequence (SECIS). These proteins are selenocysteine synthase (SelA), selenocysteine-specific elongation factor (SelB), selenocysteine specific tRNASec (SelC) and SelD183-189. The genes for these proteins are located on the sel operon185-190, shown in Figure 3.1 below.
47 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Figure 3.1: The sel operon. The genes and gene products are shown by the arrows with selD (selenide water dikinase), selA (selenocysteine synthase), selB (selenocysteine-specific elongation factor) and selC (selenocysteine-specific tRNA). Adapted from Gursinsky et al., 2000.
Selenocysteine is synthesised from selenophosphate, which is synthesised in a multistep reaction catalysed by SelD, shown in Figure 3.2191. In this reaction hydrogen selenide is reduced to selenium and then bound directly to phosphate, with adenosine triphosphate (ATP) as the phosphate donor176,182-184,186,192,193. Selenophosphate then acts as the selenium donor in a reaction with SelC that is catalysed by SelA183,184, shown in Figure 3.3. SelC is a specialised tRNASec with the anticodon UCA that recognises the in-frame UGA codon for selenocysteine183,184,187-189,194. Initially charged with a serine176,182,183,185,192,195, it is pre-translationally converted to a selenocysteine by SelA176,183,184,192, which changes the hydroxyl group of the serine to a selenol group176,187-189,195,196. SelA utilises selenophosphate as the selenium donor176,182,186,193.
Figure 3.2: Selenophosphate synthesis. SelD catalyses the synthesis of selenophosphate in a multistep reaction that utilises hydrogen selenide (HSe-) as the selenium donor and ATP as the phosphate donor. Adapted from Veres et al., 1994.
48 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Figure 3.3: Selenocysteine synthesis. SelA converts the serine charged SelC to a selenocysteine charged SelC by converting the hydroxyl group of serine to a selenol group (circled in red). This reaction uses selenophosphate as the selenium donor. Adapted from Forchhammer and Bock 1991; and Veres et al., 1994.
Selenocysteine is cotranslationally incorporated at an in-frame UGA codon with the aid of the SECIS element176,182-184,187-189,195-197 as outlined in Figure 3.4. In prokaryotes, the SECIS is located immediately downstream of the UGA codon176 and interacts with SelB to recruit the tRNASec to the ribosome176,182,185,197. The selenoprotein specific transcription regulator SelB has a similar function to the general transcription regulator elongation factor (EF)-Tu. There is high sequence similarity in the N-terminal of SelB and EF-TU, which binds GTP and tRNA. The C-terminal of SELB has a specialised 17kDa domain called domain IVb, which is responsible for binding to the SECIS element182,185,192,195,197. The simultaneous binding of SelB to tRNASec, GTP and the SECIS element forms the quaternary complex SelB:GTP:selenocysteyl-tRNASec:SECIS, which can then bind to the ribosome. Interaction of the tRNASec anticodon at the ribosomal A site triggers the GTPase activity of SelB, which releases the tRNASec to allow incorporation of selenocysteine into the polypeptide chain. Once tRNASec has been released, SelB dissociates from the SECIS element, allowing translation to continue197.
49 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Figure 3.4: Incorporation of selenocysteine. a) Upon detection of the UGA codon by the ribosome, SelB forms a quaternary complex with tRNASec, GTP and the SECIS element. b) GTP hydrolysis releases tRNASec into the ribosome A site. c) SelB dissociates from the SECIS element. d) Selenocysteine (U) is incorporated into the polypeptide chain and translation continues. Adapted from Allmang and Krol, 2006; Gursinsky et al., 2008; and Zhang et al., 2008 Fourmy 2002.
The SECIS element is unique to selenoprotein mRNA, and thus makes a good target when searching for selenoproteins. As bacterial SECIS differ from eukaryotic and archaic SECIS, a specialised SECIS search program is necessary.
50 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Zhang et al. have developed such a program called bSECISearch, which predicts bSECIS and the open reading frame of bacterial selenoproteins using the consensus structure outlined in the methods chapter (Figure 2.2)165. This program combined with nucleotide BLAST and RNAfold were used to identify potential selenoproteins in the C. difficile 630 genome.
3.2. Results The initial bSECISearch identified 385 SECIS–like structures in the C. difficile 630 genome. The first nucleotide BLAST reduced this number to 283 by eliminating bSECISearch hits that did not correspond to a gene. These 283 genes were put back through bSECISearch individually to confirm the presence of a SECIS element. The secondary structure of the SECIS elements were predicted using RNAfold, and these structures were compared to the consensus structure derived by Zhang et al. After this manual comparison, there were 94 strong candidates for selenoproteins, including SelD, GrdB and PrdA. One gene that did not match the SECIS consensus structure was glycine reductase A (grdA), which is known to be a selenoprotein193,198. The only difference between the SECIS element of grdA and that of the consensus structure was the length of the upper stem. The consensus structure stipulated an upper stem of 4-16 base pairs (bp)165,193,198 while grdA has an upper stem of just three base pairs. Based on the knowledge that GrdA is a selenoprotein165,199,200, a reassessment of genes with a shorter upper stem was done to identify an additional five genes as potential selenoproteins. This change in length is supported by a number of studies that used mutated oligonucleotides of the Escherichia. coli selenoprotein formate dehydrogenase (fdhF) to determine the key features of the mRNA stem loop necessary for SelB binding and subsequent selenocysteine insertion197,199,200. These studies determined that key components of the fdhF SECIS element were the guanosine (G) in the beginning of the apical loop and a bulged uridine (U) at +17 in the upper stem just prior to the apical loop. Further experiments also demonstrated that there was room for variability on the position of U, with interaction with SelB still being observed in oligonucleotides with U at positions 16 and 18 instead of 17. This U also didn’t need to be bulged if the upper stem was U rich with multiple U residues197,199,200. Figure 3.5 shows the proposed new
51 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
consensus structure with a 3-16 bp upper stem that is U rich. The other features of the consensus structure remain unchanged from the Zhang and Gladyshev model. All selenoproteins were reassessed for this new consensus structure, including those that fitted with the one previously proposed by Zhang and Gladyshev. There were 13 genes that were identified using the Zhang and Gladyshev model but rejected by the model proposed here due to a lack of U residues in the upper stem.
Figure 3.5: Proposed new consensus structure for bacterial SECIS elements. The Sec-UGA codon is shown in bold and is underlined. This consensus model includes an apical loop of 3-14 nt and an upper stem of 3-16 bp, at least one guanosine (G) among the first two nucleotides of the apical loop, a U rich region in the upper stem before the apical loop and a spacing of 16-37 nt between the UGA codon and the apical loop.
The 86 potential selenoproteins identified using the new SECIS consensus structure along with their putative functional class201, level of conservation and U richness in the upper stem are presented in Table 3.1 below. Level of conservation was determined by comparing the genes from the C. difficile 630 genome to 100 other C. difficile genomes representing a variety of different ribotypes using NCBI Blast. The list of genomes from NCBI is shown in Appendix 1. The putative function of the protein products was determined using the Cluster of Orthologous groups (COGs) protein database201. Overall 61 (70.93%) of selenoproteins were 100% conserved, with 6 (6.98%) and 8 (9.30%) being very highly or highly conserved, respectively. Moderate to high conservation was found in 3 (3.49%), moderate in 2 (2.33%) and low to moderate also in 2 (2.33%), with just 4 (4.65%) of
52 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
selenoproteins showing low levels of conservation. The average U richness in the upper stem before the apical loop was 43.02% (range: 83.33% to 14.29%). Selenoproteins with SECIS elements that scored 0% in this criterion were rejected. The SECIS element structures for these potential selenoproteins are shown in Appendix 2, including the 13 rejected by the new consensus model.
Table 3.1: C. difficile selenoproteins Selenoproteins from C. difficile 630 (NCBI) that match the new SECIS consensus structure. Putative functional class determined from COG database, level of conservation in 100 C. difficile genomes from NCBI database using NCBI Blast and uridine richness of upper stem as an indication of strength of SelB binding to SECIS element. Uridine Level of richness Selenoprotein name Putative function class conservation of upper stem Translation, ribosomal 30S ribosomal protein S1 (RpsA) 100% 50% structure and biogenesis 4-aminobutyrate aminotransferase Amino acid transport and 100% 33.33% (GabT) metabolism ABC-transporter ATP-binding protein Signal transduction 100% 66% (LolD) mechanisms ABC-transporter sugar-family ATP- Signal transduction 100% 60% binding protein (SalY) mechanisms Chromosomal replication initiation Replication, 100% 40% protein recombination and repair Secondary metabolites Acetolactate synthase large subunit biosynthesis, transport and 100% 50% (IlvB) catabolism Carbohydrate transport Alpha-mannosidase 100% 20% and metabolism Amino acid transport and Aminotransferase 100% 50% metabolism Secondary metabolites Asparagine synthetase (AsnB) biosynthesis, transport and 100% 40% catabolism Amino acid transport and Aspartate aminotransferase (AspB) 100% 66.66% metabolism Diguanylate kinase signalling protein Cell motility 100% 33.33% Replication, DinG family helicase (DinG) 100% 28.57% recombination and repair Replication, DNA mismatch repair protein (MutL) 100% 50% recombination and repair Replication, DNA primase (DnaG) 100% 50% recombination and repair Amino acid transport and D-proline reductase (PrdA) 100% 50% metabolism Electron transport complex protein Energy production and 100% 25% (RnfC) conversion Secondary metabolites Glutaminase (GlsA) biosynthesis, transport and 100% 60% catabolism GTP-binding protein (LepA) Translation 100% 50%
53 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Secondary metabolites HAD-superfamily hydrolase biosynthesis, transport and 100% 57.14% catabolism General function Minor head protein 100% 40% prediction only Cell Iron-sulfur binding protein wall/membrane/envelope 100% 28.57% biogenesis Secondary metabolites Isoleucine-tRNA ligase (IleS) biosynthesis, transport and 100% 20% catabolism K/Mg/Cd/Cu/Zn/Na/Ca/Na/H- Inorganic ion transport 100% 33.33% transporting P-type ATPase (ZntA) and metabolism Signal transduction Lipoprotein signal peptidase (LspA) 100% 30% mechanisms LysR family transcriptional regulator Transcription 100% 16.66% (LysR) Amino acid transport and M24 family peptidase (PepP) 100% 33.33% metabolism MATE family drug/sodium antiporter Defence mechanisms 100% 66.66% (NorM) General function Putative membrane protein 100% 33.33% prediction only Secondary metabolites Methionine gamma-lyase (MdeA) biosynthesis, transport and 100% 83.33% catabolism Secondary metabolites Putative methyltransferase biosynthesis, transport and 100% 42.86% catabolism Multidrug family ABC transporter/ Defence mechanisms 100% 50% ATP-binding protein/permease (MdlB) Inorganic ion transport NhaC family Na(+)/H(+) antiporter 100% 80% and metabolism Secondary metabolites Bifunctional nitrilase/nitrile hydratase biosynthesis, transport and 100% 40% catabolism Secondary metabolites Prolyl-tRNA ligase (ProS2) biosynthesis, transport and 100% 20% catabolism PTS system lactose/cellobiose-family Carbohydrate transport 100% 28.51% IIC component (CelB) and metabolism PTS system, fructose/mannitol-family Carbohydrate transport 100% 40% IIC component and metabolism Amino acid transport and Oligopeptide transporter (OPT) 100% 25% metabolism Secondary metabolites Peptidase, M20D family (AgbB) biosynthesis, transport and 100% 33.33% catabolism Secondary metabolites Peptidase, S9 family (DAP2) biosynthesis, transport and 100% 50% catabolism Pyruvate phosphate dikinase, Carbohydrate transport 100% 40% PEP/pyruvate-binding and metabolism Signal transduction Signalling protein 100% 50% mechanisms Inorganic ion transport Sodium:solute symporter 100% 60% and metabolism
54 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Secondary metabolites Spore cortex-lytic hydrolase (SleB) biosynthesis, transport and 100% 66.66% catabolism Post-translational Thioredoxin-disulfide reductase modification, protein 100% 20% (TrxB1) turnover and chaperones Inorganic ion transport Thiosulfate sulfurtransferase (SseA) 100% 71.43% and metabolism Intracellular trafficking, Type II secretion system protein (PulF) secretion and vesicular 100% 20% transport UDP-glycosyltransferase, MGT Carbohydrate transport 100% 44.44% subfamily and metabolism Replication, UvrABC system protein A 100% 71.43% recombination and repair Replication, Queuine tRNA-ribosyltransferase (Tgt) 100% 57.14% recombination and repair Secondary metabolites Selenide water dikinase (SelD) biosynthesis, transport and 100% 50% catabolism Amino acid transport and Sodium/glutamate symporter (GltC) 100% 60% metabolism Cell Spore maturation protein B (SpmB) wall/membrane/envelope 100% 40% biogenesis Secondary metabolites Threonine synthase (ThrC) biosynthesis, transport and 100% 40% catabolism Transcription antiterminator, PTS Transcription 100% 42.86% operon regulator C4-dicarboxylate anaerobic carrier Inorganic ion transport 100% 83.33% (YfcC) and metabolism Replication, Exonuclease ABC subunit A (UvrA) 100% 37.50% recombination and repair Amino acid transport and Glycine reductase complex component metabolism: Energy 100% 50% A (GrdA) production and conversion Multidrug family ABC transporter Defence mechanisms 100% 50% permease Translation, ribosomal 23s ribosomal RNA (23s-rRNA) 100% 25% structure and biogenesis Two-component sensor histidine kinase Signal transduction 100% 50% (OmpR) mechanisms Two-component response regulator Transcription 100% 83.33% Secondary metabolites Argininosuccinate synthase biosynthesis, transport and Very high 20% catabolism Replication, Site-specific recombinase Very high 25% recombination and repair Inorganic ion transport ABC transporter permease Very high 28.57% and metabolism General function Membrane protein Very high 60% prediction only Transcriptional regulator, Xylose Transcription Very high 33.33% repressor (XylR) Carbohydrate transport Xylose kinase (XylB) Very high 41.66% and metabolism
55 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Amino acid transport and Glycine reductase complex component metabolism; Energy High 42.86% B subunit gamma production and conversion General function Phage tail fiber protein High 42.86% prediction only Secondary metabolites Amidohydrolase biosynthesis, transport and High 25% catabolism Transcriptional regulator, TetR family Transcription High 40% (TetR_N) Putative phage DNA-binding protein* General function High 14.29% (YjcR) prediction only Replication initiation factor Tn916- Replication, High 66.66% like, CTn7-Orf27 recombination and repair Pyrroline-5-carboxylate reductase Amino acid transport and High 37.5% (ProC2) metabolism General function Conjugative transposon protein (TpcC) High 40% prediction only PTS system lichenan-specific IIC Carbohydrate transport Moderate- 50% component (LicC) and metabolism high Inorganic ion transport Moderate- ATPase 14.29% and metabolism high Moderate- Cell surface protein Extracellular structures 36.36% high Cobalt-specific ABC transporter Inorganic ion transport Moderate 40% permease (EcfT) and metabolism Replication, DNA/RNA helicase Moderate 60% recombination and repair Secondary metabolites Moderate- Putative hydrolase biosynthesis, transport and 25% low catabolism General function Moderate- Conjugative transposon protein (Spr) 16.66% prediction only low General function Conjugative transposon protein Low 50% prediction only Secondary metabolites Cell wall hydrolase biosynthesis, transport and Low 25% catabolism Cell cycle control, cell Cell division FtsK/SpoIIIE-family division, and chromosome Low 60% protein partitioning General function Membrane protein Low 16.66% prediction only
Overall, the majority of the potential selenoproteins identified were involved in metabolic pathways. As shown in Figure 3.6, the largest putative functional class was secondary metabolite biosynthesis, transport and catabolism with 18 potential selenoproteins, 14 of which were 100% conserved. Other transport and metabolic functions were amino acid transport and metabolism (10 selenoproteins), carbohydrate transport and metabolism (seven selenoproteins), inorganic ion transport and metabolism (eight selenoproteins) and energy production and conversion (three selenoproteins). Most of the selenoproteins identified in each of
56 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
these COGs were 100% conserved. There were a number of genes that were involved in information storage and processing, the largest proportion of which were involved in replication recombination and repair. In this functional class, there were 10 selenoproteins identified and most of them were 100% conserved. Transcription had seven selenoproteins with conservation ranging from 100% to low, while translation, ribosomal structure and biogenesis only had three selenoproteins, but these were all 100% conserved. Putative functional classes involved in cellular processes and signalling had the least number of selenoproteins overall. The largest putative functional class in this group was signal transduction mechanisms with five selenoproteins, which were 100% conserved. The smallest COGs were cell cycle control, cell division and chromosome partitioning, and cell motility, each with just one selenoprotein. The selenoprotein in cell motility, diguanylate kinase signalling protein, was highly conserved, while the conservation of the Ftsk/SpoIIIE-family protein from cell cycle control, cell division and chromosome partitioning was low. There were nine selenoproteins that could have their general putative function predicted, but not their actual cellular role201.
57 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
Figure 3.6: Putative functional classes of C. difficile selenoproteins. The selenoproteins identified in this study are grouped according to the putative functional class of the protein. These putative functional classes are based on the Cluster of Orthologous Groups classification. The grouping for level of classification (legend) is based on the data in Table 3.1.
3.3. Discussion A good therapeutic target is one that has a high likelihood of being present, and one that is essential for bacterial survival202. Auranofin is able to prevent C. difficile from utilizing selenium to make selenoproteins by forming a stable adduct with inorganic selenium, thereby limiting its bioavailability24. It is clear from the results that selenoproteins in C. difficile are highly conserved, with 61 having 100%
58 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
sequence identity in a panel of 100 C. difficile strains from NCBI, and a further 14 having highly similar (≥ 95%) sequence identity. Further to this there is a wide range of selenoproteins with various functions including metabolism, information storage and processing, as well as cellular processes and signalling. The breakdown of the putative functional classes of selenoproteins showed that more than 50% of the selenoproteins identified were associated with metabolism and energy conservation, indicating a strong reliance on selenium for these processes. This finding is not surprising since many of the known selenoproteins are oxidoreductases involved in metabolic pathways181,195. Dembek et al. used transposon-directed insertion site sequence (TraDIS) to generate a library of C. difficile transposon mutants. This method allowed them to identify 404 genes essential for in vitro growth and 798 genes likely to impact spore production29. Of the genes identified as potential selenoproteins from this work 11 belong to the set identified as essential for in vitro growth, including 30S ribosomal protein S1, DNA mismatch repair protein, three membrane protein, methyltransferase, amidohydrolase and two cell surface proteins. There were also a number of proteins that were not themselves essential for growth but do form a part of essential pathways including ribosomal, metabolic, aminoacyl-tRNA biosynthesis, DNA replication, RNA polymerase, protein export, glutamine and glutamate metabolism54. The genes essential for growth are also likely to impact spore production along with aminotransferase, electron transport-complex protein, lipoprotein signal peptidase, LysR family transcriptional regulator, Na+/H+ antiporter, putative polysaccharide deacetylase, putative sodium:solute symporter, ribonuclease R, selenide water dikinase, argininosuccinate synthase, ABC transporter permease, glycine reductase complex component B subunit gamma, putative phage-tail fiber protein, conjugative transposon proteins, putative hydrolase, cell wall hydrolase, glycine reductase complex component A and spore maturation protein B29. With 10 potential selenoproteins being essential for bacterial growth and even more being involved, selenium metabolism and selenoprotein synthesis should prove to be an effective target for the treatment of CDI. Targeting pathways that are essential for bacterial growth is likely to be an effective method for treating infections. Since these pathways are highly conserved this reduces the risk of C. difficile developing
59 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
resistance to the therapy202. Targeting pathways involved in sporulation has the added advantage of possibly reducing C. difficile transmission. C. difficile is capable of producing highly resilient spores that are able to survive exposure to antibiotics such as vancomycin and metronidazole. This is a contributing factor to the increase in recurrent and refractory CDI. If auranofin is able to prevent C. difficile from producing selenoproteins, then this would reduce bacterial growth and sporulation, thus treating the infection and possibly reducing the risk of recurrence and transmission of disease10,74,104. Zhang and Gladyshev who described the consensus model used by bSECISearch and for the manual analysis described above, determined the consensus structure based on the analysis of the UGA flanking region of 100 selenoprotein mRNA sequence165. By using RNAfold to compare the secondary structures, they were able to identify the common features to construct this consensus model165. Most of the potential selenoproteins described above conformed to this consensus structure with one notable exception. Glycine reductase complex component A, often referred to as glycine reductase selenoprotein A, has been demonstrated to be a selenoprotein193,198. Sliwkowski et al. have been able to demonstrate that glycine reductase activity is lost in selenium-deficient cells, but is regained once selenium is available again198. They were also able to confirm incorporation of selenium into the protein experimentally by using radiolabeled selenium (75Se) and high performance liquid chromatography (HPLC)193. The upper stem of the C. difficile 630 grdA SECIS element was 3 bp in length, which is shorter than that of the 4-16 bp specified by the consensus structure165. With the knowledge that GrdA is a selenoprotein, it is reasonable to investigate the possibility that the consensus structure proposed by Zhang and Gladyshev may not apply to all bacterial selenoproteins. Indeed, Zhang and Gladyshev acknowledged that their SECIS was a loose model to allow greater tolerance for variation165, although clearly GrdA was overlooked in the development of this model. The Zhang and Gladyshev model is also lacking a bulged uridine (U) in the upper stem, a feature that is considered a key component in the well-studied E. coli SECIS elements165,197,199,200. This bulged U was omitted based on the observation that the SECIS elements of many bacterial organisms including Clostridium sticklandii, Clostridium purinolycticum and Eubacterium acidaminophilum had little in
60 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
common with the E.coli SECIS elements165. Studies looking at the key components of the E.coli selenoprotein fdhA SECIS element found that although the bulge at U17 is a key component of the SECIS element, it is not essential for SelB binding, there merely needs to be sufficient U in the upper stem before the apical loop197,199,200. While many of these studies concluded that the upper stem needed to be a minimum of 4-5 bp197,199,200, they did present mutated oligonucleotides with an upper stem of just 3 bp. These oligonucleotides were still capable of binding to SelB, and thus facilitate selenocysteine insertion. While the authors did not address the shorter upper stem, it is highly suggestive that an upper stem of 3 bp, like the one in the grdA SECIS element, would still be functional. Therefore, the new bacterial SECIS consensus model proposed here has a number of features in common with the Zhang and Gladyshev model, namely a distance between the UGA and apical loop of 16-37 nt, an apical loop of 3-14 nt and at least 1 of the first 2 nucleotides in the apical loop being G165, with the addition of a U rich region in the upper stem before the apical loop and an upper stem length of 3-16 bp. One limitation of this work is that it is entirely in silico. Further work would need to be done to confirm that the above proteins are in fact selenoproteins. This could be done by generating a knock-out SelD mutant of C. difficile. SelD is not essential for in vitro growth29 however it is important for selenium metabolism and selenoprotein synthesis, as it is the catalyst for the synthesis of selenophosphate, the selenium donor for selenocysteine. Without active SelD it is likely that the expression of other selenoproteins will be reduced. An analysis of the mRNA transcripts could identify the proteins whose expression is altered as a result of disrupting the selenium metabolism. Additionally, it would be useful to investigate the application of the bacterial SECIS consensus model proposed here on a variety of known selenoproteins from different bacteria to confirm its validity. The reduced constraints on the length of the upper stem may also identify a number of new selenoproteins. A further future addition to this work would be to look for homologous genes in other clostridia, including cysteine containing homologues, as this would provide further supporting evidence. In conclusion, there are a large number of highly conserved selenoproteins in the C. difficile 630 genome, many of which play essential roles in bacterial growth and sporulation. This makes selenium metabolism a strong therapeutic target as there
61 IDENTIFYING SELENOPROTEINS IN C. DIFFICILE USING IN SILICO METHODS
is a reduced chance of C. difficile acquiring resistance or being intrinsically resistant to selenium targeted therapies, such as auranofin.
62 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
4. Investigation of auranofin as a treatment for C. difficile infection
4.1. Background C. difficile utilises stickland reactions for energy conservation in which the amino acids proline and glycine are reduced. In stickland reactions one amino acid acts as the electron donor by undergoing oxidative deamination, while the other acts as the electron acceptor by undergoing reductive deamination185,187-189,196. In Clostridia, proline and glycine reductase are both electron acceptors, being reduced to aminovaleric acid and acetic acid plus ammonia, respectively187-189. Glycine reductase catalyses the reductive deamination of glycine to acetyl phosphate and ammonium178,181,193,196,198. The grd operon (Figure 4.1) contains the genes grdA, grdB, grdC, grdD and grdE185,196. This operon also contains a gene predicted to encode proline aminopeptidase (proP)196 as well as thioredoxin (trx) and thioredoxin reductase (trxR)194. The glycine reductase complex contains three subunits GrdA (grdA), GrdB (grdB and grdE) and GrdC (grdC and grdD). GrdA and GrdB are selenoproteins185,196,203. A selenocysteine functions as the redox centre for the catalytic activity of these proteins180,193,198. Selenium is essential for the activity of the glycine reductase complex18,185,196.
Figure 4.1: The glycine reductase (grd) operon. Adapted from Bouillaut, Self and Sonenshein 2013, and Jackson et al. 2006.
The glycine reductase pathway is outlined in Figure 4.2. Glycine binds to GrdB which forms a Schiff base intermediate via an essential carbonyl moiety195,204-206. GrdA then cleaves this Schiff base, breaking the carbon-nitrogen bond194,195,204-206. This transforms glycine to ammonia and a carboxymethyl selenoether bound to the ionized selenocysteine residue in GrdA194,195,204-206. GrdC cleaves the carboxymethyl group and reduces it to an acetyl group which it then phosphorylates
63 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
into acetyl phosphate195,205,206. The now oxidised GrdA is reduced by NADPH dependent thioredoxin to start a new cycle194,204-206. Meanwhile the acetyl phosphate is converted to ATP by acetate kinase in the presence of ADP186,192,194,206.
Figure 4.2: The glycine reductase pathway. Glycine binds to the GrdB subunit and is reduced by the GrdA subunit. GrdC then phosphorylates it to create acetyl phosphate which is converted into ATP by acetyl kinase. The now oxidised GrdA needs to be reduced to regain activity. This is done by NADPH dependent thioredoxin/thioredoxin reductase system.
Proline reductase reductively cleaves D-proline to 5-aminovalerate185,194,196,207. The prd operon (Figure 4.3) contains the genes prdA, prdB, prdC, prdD, prdE1, prdE2 and prdF185,194. The product of prdA is a proprotein that is cleaved into two subunits, a 23-kDa subunit containing a pyruvoyl group and a 45-kDa subunit194. The product of prdB is a selenoprotein similar to GrdB. PrdB and GrdB are highly conserved around the selenocysteine moiety suggesting a similarity in function194. The product of prdF is a proline racemase which converts L-proline to D-proline, the substrate for proline reductase. This allows C. difficile to utilise both D-proline and L-proline185.
64 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.3: The proline reductase (prd) operon. Adapted from Bouillaut, Self and Sonenshein, 2013 and Jackson et al., 2006.
The proline reductase pathway is outlined in Figure 4.4. D-proline binds to the pyruvoyl group on the PrdA 23-kDa subunit possibly forming a Schiff base intermediate similar to that seen in the glycine reductase pathway. The selenocysteine in PrdB cleaves the carbon-nitrogen bond of the proline ring forming an unknown intermediate compound bound to the selenocysteine of PrdB. The final product, 5-aminovalerate, is formed by hydrolysis activity194,207. The PrdB subunit is oxidised in this reaction and needs to be reduced to regain activity. The electron donor for this redox reaction is unknown194,207. The reduction of D-proline does not yield a molecule with a high group transfer potential like ATP from the glycine reductase pathway194,207. Based on experiments with C. sporogenes, it has been suggested that proline reduction is coupled to the formation of a pH gradient across the cytoplasmic membrane194. It is likely that this chemiostatic mechanism is a key energy-yielding pathway that is critical for anaerobic bacteria specialising in Stickland fermentation of amino acids185,194,207.
65 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.4: The proline reductase pathway Proline binds to the PrdA 23-kDa subunit and is reduced by the PrdB subunit to produce 5-aminovalerate. It is thought that the energy-yielding mechanism in this reaction is the formation of a pH gradient across the cytoplasmic membrane. The electron donor used to reduce PrdB to its active state is currently unknown. Adapted from Kabisch 1999
Auranofin is a monomeric gold complex containing a gold-thiol (Au-S) bond stabilised by a triethylphosphine group. It is a potent inhibitor of selenoproteins such as thioredoxin reductase in mammalian cells and parasites such as Trypanosoma brucei and Schistosoma mansoni24. As previously discussed (Chapter 3), selenoproteins contain selenium in the form of selenocysteine, which is synthesised from selenophosphate. The synthesis of selenophosphate occurs in a reaction catalysed by SelD, where reduced selenium in the form of hydrogen selenide (HSe-) is bound to phosphorus using ATP as the phosphate donor. Talbot et al. demonstrated that auranofin is capable of inhibiting selenoprotein synthesis, and attributed this to disruption of selenium metabolism208. Jackson-Rosario et al. investigated the mechanisms of the interaction between auranofin and selenium metabolism in C. difficile and E. coli, which is a model organism for prokaryotic selenoprotein synthesis. They were able to demonstrate that auranofin reacts with HSe- to form a stable adduct, and that the addition of
66 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
auranofin to C. difficile cultures prevented the uptake of 75Se and subsequent selenoprotein synthesis24. Based on this, Jackson-Rosario et al. then investigated the impact of auranofin on E. coli (MC4100) and a number of C. difficile strains. Auranofin reduced the growth of E. coli at concentrations as low as 25 µM, and they observed that this growth reduction was similar to that seen in selD knockouts24. Auranofin inhibited the growth of C. difficile with no observable growth at a concentration of 2 µM24. These experiments provide strong evidence supporting the theory that auranofin interacts with selenium to prevent selenium metabolism and selenoprotein synthesis in C. difficile. However, in these experiments, the C. difficile cultures were treated with auranofin at inoculation. C. difficile produces its virulence factors, TcdA and TcdB, at the end of exponential growth phase and during stationary growth phase21,62. As these toxins are responsible for the pathology of CDI122, it is necessary to test the activity of auranofin against an established C. difficile culture in stationary growth phase, since this reflects when the infection would be treated in a clinical setting. It was the aim of these experiments to investigate the potential of auranofin as a treatment for CDI. This was done by investigating the activity of auranofin against an established in vitro C. difficile culture, looking at bacterial growth, cell viability, sporulation and toxin production. It was hypothesised that auranofin would inhibit the growth of C. difficile without stimulating an increase in sporulation or toxin production. To further elucidate the potential of auranofin as a treatment for CDI, it is necessary to compare the activity of auranofin against C. difficile to that of the current treatments, vancomycin and metronidazole. This was also done in vitro focusing on bacterial growth, cell viability and sporulation. It was hypothesised that auranofin would be non-inferior to vancomycin and metronidazole. The final component of this study was a preliminary animal trial to determine if auranofin could prevent CDI and its symptoms, as well as determine an effective and safe dose for further animal trials. These trials were performed at Monash University, Clayton VIC, Australia, with the assistance of Associate Professor Dena Lyras and Dr. Melanie Hutton. It was hypothesised that auranofin would
67 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
reduce the level of colonisation and offer protection from the cytotoxic activity of C. difficile toxins TcdA and TcdB.
4.2. Results
4.2.1. Minimum inhibitory concentration (MIC) and subinhibitory concentration of auranofin The minimum inhibitory growth (MIC) assays showed a significant reduction in cell densiy compared to the treatment group when treated with 6.25 µM of auranofin after 2 hours (p=0.0008). This was consistent through all time points up to 24 hours. Auranofin at a concentration of 3.125 µM showed inconsistent inhibition of C. difficile M7404 growth, being significantly less than the untreated groups at some, but not all time points. This indicates that the MIC is between 3.125 and 6.25 µM. For all further experiments, MIC is used to describe a concentration of 6 µM of auranofin. Auranofin at a concentration of 0.098 µM showed no inhibition of bacterial growth at any time point, and is therefore subinhibitory. For all further experiments, a subinhibitory concentration of auranofin refers to a concentration of 0.1 µM. Figures 4.5 to 4.11 show the bacterial growth with treatment concentrations ranging from 100 µM through to 0.098 µM of auranofin.
Figure 4.5: Auranofin MIC assay, 100 µM auranofin. Mean absorbance for cultures treated with 100 µM auranofin or 2% ethanol. Controls are no treatment and sterile TY broth. n=3. Error bars show standard error of the mean (SEM). The p- values are comparing auranofin to the no treatment control using a students t-test: 2 hours p=0.0005, 4 hors p=0.0005, 6 hours p=0.0296, 8 hours p<0.0001, 10 hours p=0.0093 and 24 hours p=0.0001.
68 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.6: Auranofin MIC assay, 50 µM aurnaofin Mean absorbance for cultures treated with 50 µM auranofin or 1% ethanol. Controls are no treatment and sterile TY broth. n=3. Error bars show standard error of the mean (SEM). The p-values are comparing auranofin to the no treatment control using a students t-test: 2 hours p=0.0241, 4 hours p=0.0159, 6 hours p=0.0004, 8 hours p=0.0059 and 10 hours p=0.0035.
Figure 4.7: Auranofin MIC assay, 25 µM auranofin. Mean absorbance for cultures treated with 25 µM auranofin or 0.5% ethanol. Controls are no treatment and sterile TY broth. n=3. Error bars show standard error of the mean (SEM). The p- values are comparing auranofin to the no treatment control using a students t-test: 2 hours p=0.0023, 4 hours p=0.0179, 6 hours p=0.001, 8 hours p=0.0155, 10 hours p=0.003 and 24 hours p=0.0009.
69 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.8: Auranofin MIC assay, 12.5 µM aurnaofin Mean absorbance for cultures treated with 12.5 µM auranofin or 0.25% ethanol. Controls are no treatment and sterile TY broth. n=3. Error bars show standard error of the mean (SEM). The p- values are comparing auranofin to the no treatment control using a students t-test: 2 hours p=0.0128, 4 hours p=0.0006, 6 hours p=0.0074, 8 hours p=0.0023, 10 hours p=0.0072 and 24 hors p=0.0196.
Figure 4.9: Auranofin MIC assay, 6.25 µM auranofin Mean absorbance for cultures treated with 6.25 µM auranofin or 0.125% ethanol. Controls are no treatment and sterile TY broth. n=3. Error bars show standard error of the mean (SEM). The p- values are comparing auranofin to the no treatment control using a students t-test: 2 hours p=0.0008, 4 hours p=0.0006, 6 hours p=0.0328, 8 hours p=0.0.0211, 10 hours p=0.0257 and 24 hours p=0.018.
70 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.6: Auranofin MIC assay, 3.125 µM auranofin Mean absorbance for cultures treated with 3.125 µM auranofin or 0.0625% ethanol. Controls are no treatment and sterile TY broth. n=3. Error bars show standard error of the mean (SEM). The p- values are comparing auranofin to the no treatment control using a students t-test: 2 hours p=0.0319, 6 hours p=0.0372 and 8 hours p=0.0063.
Auranofin 0.098uM
) 0.6
m n
0 Ethanol 0.002%
0 0.5 6
D Control
O 0.4
(
e Media control
c 0.3
n
a b
r 0.2
o
s b
A 0.1 0.0 0 2 4 6 8 10 12 14 16 18 20 22 24 Time (hours)
Figure 4.7: Auranofin MIC assay, 0.098 µM auranofin. Mean absorbance for cultures treated with 0.098 µM auranofin or 0.002% ethanol. Controls are no treatment and sterile TY broth. n=3. Error bars show standard error of the mean (SEM).
Growth assays observing the effect of the MIC and subinhibitory concentrations of auranofin on the growth of an established culture, as well as cell viability and sporulation, were performed. After treatment at 12 hours, all three treatment groups showed a decrease in cell density, with the effect of auranofin being greater in the subinhibitory group than the MIC group, although this was not significant. The subinhibitory group then plateaued to match the control group, while the ethanol
71 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
and MIC groups both continued to decrease in cell density, indicating inhibition of bacterial growth. The MIC group had significantly less bacterial growth than the 0.1% ethanol group at 14 hours (p=0.0258) and 16 hours (p=0.0008), but at the later time points there was no difference between the two. The OD600nm of the four groups across all time points is shown in Figure 4.12.
Figure 4.8: Bacterial growth of C. difficile M7404 treated with 0.1 µM auranofin, 6 µM auranofin, 0.1% ethanol (v/v) or no treatment (control). Mean absorbance for cultures. Dotted line represents the treatment time point, * indicates a p-value of <0.05, n=3. Error bars show SEM.
The cell viability of C. difficile M7404 was greatly reduced by treatment with 6 µM of auranofin and 0.1% ethanol, with the reduction of viability being greater in the auranofin group at 24 and 48 hours. The subinhibitory concentration of auranofin did not significantly reduce viability at 24 hours and there was only a slight decrease in cell viability at 48 hours. The cell viability for all groups at 0, 12, 24 and 48 hours is shown in Figure 4.13.
72 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.9: Vegetative cell viability of C. difficile M7404. Treatment groups 0.1 µM auranofin, 6 µM auranofin, 0.1% ethanol and no treatment (control). n=3. Line represents mean CFU. Error bars show SEM.
The sporulation of C. difficile M7404 was reduced in the MIC auranofin treated culture after 48 hours. There was little difference between any other treatment
73 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
groups at 48 hours. At 0, 12, and 24 hours there was no difference between any of the treatment groups. The spore counts for all groups at 0, 12, 24 and 48 hours is shown in Figure 4.14.
Figure 4.10: Spore counts for C. difficile M7404 cultures. Treatment groups 0.1 µM auranofin, 6 µM auranofin, 0.1% ethanol and no treatment (control). n=3. Line represents mean CFU. Error bars show SEM.
74 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
4.2.2. Treatment concentrations of auranofin Auranofin was then tested at the proposed treatment concentrations of 500 µM, 250 µM and 50 µM. The highest concentration (500 µM) is equivalent to the maximum dose that can be given in an in vivo mouse model due to the solubility properties of auranofin. The middle treatment dose (250 µM) is half of the maximum dose and the lowest concentration (50 µM) showed a significant reduction in bacterial growth in the MIC assay above (4.2.1). Initially the lowest treatment dose was compared to an equivalent concentration of the diluent ethanol (0.8%). Due to the rate at which the bacteria grew, treatment was brought forward to 10 hours. Cell viability and sporulation assays were not performed. There was a significant reduction in the cell density of C. difficile M7404 treated with 50 µM auranofin at 12 hours compared to both the control group (p=0.0001) and ethanol (p=0.0093). This decrease in cell density was even more pronounced at 24 hours (p<0.0001, p=0.0004, respectively). There was no significant difference between the ethanol treated cultures and the control group at any time point. The bacterial growth of the three groups across all time points is shown in Figure 4.15.
Figure 4.11: Bacterial growth of C. difficile M7404 treated with 50 µM auranofin, 0.8% ethanol (v/v) or no treatment (control). Mean absorbance for cultures. Dotted line represents treatment time point, *** indicates a p-value of <0.001, n=3. Error bars show SEM.
Growth assays to test all three treatment cultures were then performed, looking at bacterial growth, cell viability, sporulation and toxin activity. All three auranofin treatment groups (50 µM, 250 µM and 500 µM) showed a significant reduction in
75 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
cell density at 24 hours compared to the control group (p<0.0001, p<0.0001 and p=0.0003, respectively). The bacterial growth of the four groups across all time points is shown in Figure 4.16.
Figure 4.12: Bacterial growth of C. difficile M7404 treated with auranofin at concentrations of 50 µM, 250 µM and 500 µM compared to no treatment (control). Mean absorbance for cultures. Dotted line represents treatment time point, ***/+++/xxx indicates a p-value of <0.001, n=3. Error bars show SEM.
The cell viability in all three treatment groups was significantly reduced compared to the control group (p=0.0017, p=0.0011 and p=0.0017, respectively). The cell viability for all groups at 0, 12, 24 and 48 hours is shown in Figure 4.17. The viability of the bacterial cells in the 50 µM and 250 µM group was significantly higher than that of the 500 µM and control groups. The reason for this is unclear as all groups contained the same media, were inoculated with the same overnight culture and were incubated under the same conditions at the same time.
76 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
0 hour 12 hour
24 hour 48 hour
Figure 4.13: Vegetative cell viability of C. difficile M7404. Treatment groups 50 µM auranofin, 250 µM auranofin, 500 µM auranofin and no treatment (control). */+/x indicates a p-value of <0.05, **/++/xx indicates a p-value of <0.01, n=3. Line represents mean CFU. Error bars show SEM.
There was also a significant reduction in the number of viable spores in all three treatment groups (p=0.0313, p=0.0306 and p=0.0304, respectively). The spore count for all groups at 0, 12, 24 and 48 hours is shown in Figure 4.18. There was also a decrease in the number of viable spores in the control group between 24 and 48 hours. This was unexpected, but is likely due to the highly variable nature of C. difficile, in particular C. difficile spores.
77 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
0 hour 12 hour
48 hour 24 hour
Figure 4.14: Spore counts for C. difficile M7404 cultures. Treatment groups 50 µM auranofin, 250 µM auranofin, 500 µM auranofin and no treatment (control). */+/x indicates a p-value of <0.05, **/++/xx indicates a p-value of <0.01, n=3. Line represents mean CFU. Error bars show SEM.
Cell cytotoxicity assays (CCA) were performed on sterile filtered C. difficile M7404 culture supernatant from the 48 hour time point. All three treatment groups had a significant reduction on the level of cytotoxic activity observed against Vero cells, with both the 250 µM and 500 µM group being the same as non-stimulated Vero cells. The 50 µM group still showed some cytotoxic activity but this was significantly reduced compared to the control group (p=0.0406). The results of the
78 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
cytotoxicity assay are shown in Figure 4.19. When the supernatant from the control C. difficile M7404 group was added to Vero cells with 50 µM of auranofin, there was a complete reduction in the cytotoxic activity observed against Vero cells treated with control C. difficile M7404 supernatant alone, as seen in Figure 4.20.
Figure 4.19: Cytotoxic activity of C. difficile TcdB against Vero cells. Toxins in C. difficile M7404 supernatant treatment with 50 µM, 250 µM, 500 µM auranofin or no treatment (control). Toxin end-point titre represents lowest titration of C. difficile supernatant with 100% affected Vero cells. * indicates p-value of <0.05, n=3. Error bars show SEM.
79 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.15: Cytotoxic activity of C. difficile TcdB against Vero cells. Toxins in C. difficile M7404 control group supernatant were compared to this same supernatant with 50 µM of auranofin added when applied to Vero cells. Toxin end-point titre represents lowest titration of C. difficile supernatant with 100% affected Vero cells. n=3. Error bars show SEM.
4.2.3. Comparison of the activity of auranofin with that of vancomycin and metronidazole The activity of auranofin against C. difficile M7404 was compared to that of vancomycin and metronidazole to confirm that auranofin was non-inferior to the current treatments for CDI. First an MIC assay was performed to determine the minimum inhibitory concentration of each antibiotic. Auranofin inhibited C. difficile M7404 growth at concentrations as low as 1.95 µM at 4 hours, but this was not observed at 8 and 24 hours. Inhibition at 3.91 µM and higher was consistently seen at 4, 8 and 24 hours. To retain consistency with the previous auranofin experiments, an MIC of 6 µM was used for the following experiments. The 0, 4, 8 and 24 hour time points for the auranofin MIC assays are shown in Figure 4.21. Vancomycin inhibited bacterial growth at 3.91 µM, however this was only seen at 24 hours. At 8 hours inhibition was seen as low as 7.81 µM. For the following in vitro experiments an MIC of 8 µM was used for vancomycin. The 0,
80 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
4, 8 and 24 hour time points for the vancomycin MIC assays are shown in Figure 4.22. Metronidazole showed activity as low as 3.91 µM at 4 hours, but this was reduced to 7.81 µM at 8 hours and 15.63 µM at 24 hours. For the following in vitro experiments an MIC of 16 µM was used for metronidazole. The 0, 4, 8 and 24 hour time points for the metronidazole MIC assays are shown in Figure 4.23.
Figure 4.16: Auranofin MIC titration assay for C. difficile M7404. Mean absorbance for cultures. Auranofin compared to no treatment (control). n=4. Error bars show SEM.
81 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.17: Vancomycin MIC titration assay for C. difficile M7404. Mean absorbance for cultures. Vancomycin compared to no treatment (control). n=4. Error bars show SEM.
Figure 4.18: Metronidazole MIC titration assay for C. difficile M7404. Mean absorbance for cultures. Metronidazole compared to no treatment (control). n=4. Error bars show SEM.
Next, growth assays to test and compare the effect auranofin, vancomycin and metronidazole had on C. difficile M7404 were performed, looking at bacterial growth, cell viability and sporulation. In these assays three sets of cultures (n=3) were treated with either auranofin, vancomycin or metronidazole at inoculation.
82 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Another three sets of cultures (n=3) were treated at 12 hours to represent an established infection. Additionally, there was an untreated control group. All groups treated with auranofin, vancomycin or metronidazole at inoculation showed no bacterial growth. Auranofin and vancomycin cultures treated at 12 hours both had significantly reduced growth compared to the control groups at 36 hours (p<0.001), 48 hours (p<0.0001) and 60 hours (p<0.0001). There was no significant difference between the auranofin and vancomycin treated cultures. Cultures treated with metronidazole at 12 hours initially showed a reduction in growth at 36 hours (p<0.001), but subsequently there was increased growth higher than that of the control group (p<0.0001). The bacterial growth of the four groups across all time points is shown in Figure 4.24.
Figure 4.19: Bacterial growth of C. difficile M7404 Mean absorbance for cultures. Treated with 6 µM auranofin, 8 µM vancomycin, 16 µM metronidazole or no treatment (control). ***/+++/xxx indicates a p-value of <0.001, n=9. Error bars show SEM. (a) Groups that were treated at inoculation (0 hour). (b) Groups treated at 12 hours (stationary phase)
83 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
The cell viability of the C. difficile groups treated at inoculation with auranofin, vancomycin and metronidazole decreased immediately and there were no viable cells at 12 hours. The cell viability of the auranofin and metronidazole 12 hour treated groups were reduced compared to the control group at 24 hours. At 48 hours the cell viability of the metronidazole group had increased to greater than the control group. At 48 hours there were no viable cells in the auranofin treated groups. The vancomycin 12 hour treated group had a higher number of viable cells than the control group at 24 hours. This group had decreased 10 fold by 48 hours, although this was still higher than the auranofin. The cell viability for all groups at 0, 12, 24 and 48 hours is shown in Figure 4.25.
84 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.20: Vegetative cell viability of C. difficile M7404. Treated with 6µM auranofin, 8 µM vancomycin or 16 µM metronidazole at either inoculation (0 hours) or beginning of stationary phase (12 hours), compared to no treatment (control). n=9. Line represents mean CFU. Error bars show SEM.
All treatment groups had no spores at 24 hours, while the control group had begun to sporulate. The metronidazole 12 hour treated group began to sporulate between 24 and 48 hours, although the number of viable spores was not as high as the control
85 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
group. All other treatment groups still had no spores at 48 hours. The spore viability for all groups at 0, 12, 24 and 48 hours is shown in Figures 5.26.
Figure 4.21: Spore count of C. difficile M7404 cultures. Treated with 6 µM auranofin, 8 µM vancomycin or 16 µM metronidazole at either inoculation (0 hours) or beginning of stationary phase (12 hours), compared to no treatment (control). n=9. Line represents mean CFU. Error bars show SEM.
86 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
4.2.4. Preliminary testing of auranofin as a treatment for CDI using an in vivo animal model Preliminary animal trials were performed to investigate the potential of auranofin as a treatment for CDI in an in vivo model. The model used was the Monash CDI prophylaxis mouse model (see 2.3.1) which used 6 week old, male C57BL/6J mice. The intestinal microbiota of these mice was disrupted with an antibiotic cocktail given in their drinking water. Prophylaxis with auranofin given in their drinking water was commenced one day prior to infection with 106 C. difficile M7404 spores, this was administered by oral gavage. Mice were monitored for signs of disease including weight loss, behavioural and physical changes, and diarrhoea. The mice that were infected and given no treatment had 0% survival after 2.5 days and the mice that were uninfected and treated with either 400 µg/ml, 200 µg/ml or 40 µg/ml of auranofin had 100% survival at day 4. There was 25% survival in the infected mice treated with the lowest dose of auranofin after 2.5 days (Figure 4.27a), which is only slightly higher than the infected, no treatment control group. The rate of weight loss for this group was similar to the infected, no treatment control group, but didn’t begin until 0.5 days later (Figure 4.27b). This indicates that the lowest dose of 40 µg/ml auranofin was not enough to provide complete protection, but was sufficient to delay the onset of disease symptoms. There was no significant difference in the level of C. difficile colonisation after 48 hours (Figure 4.30) between the infected, 40 µg/ml auranofin mice and the infected no treatment control mice (p=0.09). The level of cytotoxic activity of TcdB against Vero cells was also similar for these two groups (Figure 4.31). There was 100% survival in both the mice that were infected and treated with either 200 µg/ml or 400 µg/ml of auranofin (Figures 4.28a and 4.29a respectively), indicating that both of these doses were able to protect the mice from CDI. Colonization with C. difficile after 48 hours was significantly lower than the infected, no treatment control (p=0.0206 and p=0.0257, respectively; Figure 4.30), which suggests that auranofin was able to reduce the bacterial load but not completely eradicate it. The mice treated with 400 µg/ml auranofin, both infected and uninfected, showed notable weight loss (Figure 4.29b). Although this weight loss was not enough to reach the experimental endpoint, it did indicate that this
87 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
dose is too high and the mice were experiencing adverse side effects from the auranofin. This weight loss was not observed in the mice treated with 200 µg/ml auranofin (Figure 4.28b), indicating that this dose is high enough to protect the mice from CDI, but not too high to cause adverse side effects. This was further supported by the lower level of cytotoxic activity of TcdB against Vero cells compared to the infected, untreated control group (Figure 4.31).
Figure 4.22: Infected mice treated with 40 µg/ml auranofin. a) Survival of C. difficile infected mice treated with auranofin compared to uninfected mice and infected mice with no treatment. n=5 b) Weight loss of C. difficile infected mice treated with auranofin compared to uninfected mice and infected mice with no treatment. Mean weight loss. n=5
Figure 4.23 Infected mice treated with 200 µg/ml auranofin. a) Survival of C. difficile infected mice treated with auranofin compared to uninfected mice and infected mice with no treatment. n=5 b) Weight loss of C. difficile infected mice treated with auranofin compared to uninfected mice and infected mice with no treatment. Mean weight loss n=5
88 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.29: Infected mice treated with 400 µg/ml auranofin. a) Survival of C. difficile infected mice treated with auranofin compared to uninfected mice and infected mice with no treatment. n=5 b) Weight loss of C. difficile infected mice treated with auranofin compared to uninfected mice and infected mice with no treatment. Mean weight loss n=5
Figure 4.24: Colonisation fo mice with C. difficile spores. Colonisation of mice infected with C. difficile spores after 48 hours. Mice given no treatment or auranofin prophylaxis at 40 µg/ml (1:100), 200 µg/ml (1:20) or 400 µg/ml (1:10). Line represents mean CFU, error bars are SEM. n=5
89 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
Figure 4.25: Cytotoxic activity of C. difficile TcdB. Cytotoxic activity of C. difficile toxin TcdB tested against Vero cells. Toxin was from faeces of mice 48 hours after infection with C. difficile spores. Mice received no treatment or auranofin prophylaxis at 40 µg/ml or 200 µg/ml. Error bars are SEM n=5
4.3. Discussion C. difficile reduces the amino acids glycine and proline for energy production in a redox reaction catalysed by the selenoproteins glycine reductase and proline reductase. Synthesis of these proteins is a process that requires inorganic selenium24,185. It has been demonstrated that auranofin is capable of removing the bioavailability of selenium by forming a stable adduct with inorganic selenium through an irreversible reaction24. This prevents the synthesis of the selenoproteins that C. difficile requires for energy production, and thus inhibits the growth of C. difficile. The experiments by Jackson-Rosario et al. were the first step in investigating the potential of auranofin as a treatment for CDI. The work presented here takes the next step by demonstrating the bactericidal activity of auranofin against an established C. difficile culture, by showing non-inferiority of auranofin to the current treatments, vancomycin and metronidazole, and finally by
90 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
demonstrating that auranofin may be capable of preventing CDI in a prophylaxis mouse model. The in vitro model used in these experiments was developed to reflect an established infection. C. difficile produces the toxins TcdA and TcdB at the beginning of stationary growth phase21,62. Since these toxins are responsible for the pathology of the disease25, a patient infected with C. difficile is unlikely to be treated until after the bacteria has started producing toxins, since they will be asymptomatic up until this point. Once the toxins have started to be produced, the patient will show symptoms and seek medical treatment. It is at this point that a potential treatment for CDI needs to be tested to ensure that it is able to treat the infection rather than prevent it. Therefore the C. difficile cultures in these experiments were not treated until they had reached stationary growth phase. All in vitro growth assays were performed in a low nutrient media that promoted toxin production and sporulation. It should be noted that while stationary phase is the most clinically relevant cell stage, an infection is likely to contain different cell phases, including those in exponential growth phase. An assessment of activity against this phase is also necessary. In the Minimum Inhibitory Concentration (MIC) titration assays, cultures were treated at inoculation, as were the cultures in the experiments performed by Jackson-Rosarion et al24. Since an inhibition of bacterial growth was observed in these assays, it is reasonable to suggest that auranofin is active against C. difficile cells in exponential growth phase. MIC titration assays were used to establish the MIC for auranofin, vancomycin and metronidazole to be used in this model. The MIC for vancomycin and metronidazole determined in these assays was slightly higher than those reported by European Committee on Antimicrobial Susceptibility Testing (EUCAST)209. This is likely due to the difference in the method used to obtain these results. The EUCAST MIC distributions are determined using E-test agar diffusion209, however there is no standardised E-test available for auranofin. Since the auranofin MIC was determined by broth microdilution, the vancomycin and metronidazole MICs were determined by the same methods. The effect of auranofin at MIC on the growth of C. difficile was compared to that of ethanol, the diluent for auranofin. The reduction in cell density, indicating inhibition of bacterial growth, when treated with 6 µM auranofin was greater than
91 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
ethanol shortly after treatment, but by 48 hours the two showed the same reduction in bacterial growth. The difference in cell viability between the two at 24 and 48 hours is significantly different, with the MIC of auranofin treated cultures having less viable vegetative C. difficile cells than the ethanol treated cultures. This indicated that OD, while suitable for cell density, should only be used as a guide for culture viability, because non-viable cells may contribute to the optical density of the solution. It is therefore necessary to assess the viability of the bacterial cells using additional methods. When the concentration of auranofin was increased to the lowest of the three treatment concentrations, the difference between auranofin and ethanol was clear. C. difficile cultures treated with auranofin at a concentration of 50 µM showed a rapid decrease in cell density, while the ethanol treated cultures initially showed a slight decrease, before continuing to grow at a rate similar to the control group. This indicated that the inhibited growth observed in auranofin treated C. difficile cultures is due to auranofin and not the diluent ethanol. All three treatment groups showed a significant reduction in bacterial growth after treatment with 50 µM, 250 µM or 500 µM of auranofin. There was also a decline in viable vegetative cells, with no viable cells remaining after 24 hours in the auranofin treated cultures. This indicates that auranofin appears to be bactericidal, capable of killing vegetative C. difficile, rather than bacteriostatic as was initially expected. The bactericidal activity of auranofin makes it a strong candidate as an alternative therapy for CDI. Auranofin at a concentration of 6 µM was compared to the MIC of the current CDI treatments vancomycin and metronidazole. Auranofin was shown to be non- inferior to both of the current treatments causing an almost identical reduction in C. difficile cell density to treatment with vancomycin. Metronidazole also showed an initial decline in cell density until 36 hours, at which point these cultures showed an increase in bacterial growth until the OD was greater than the control groups at 60 hours. This not only highlights the inferiority of metronidazole as a treatment for severe CDI when compared to vancomycin210 but also indicates that metronidazole may be inferior to auranofin. The viability of the vegetative cells treated with metronidazole and auranofin reflect the OD measurements. Metronidazole treated C. difficile cultures had a
92 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
reduction in the number of viable cells between 12 and 24 hours and an increase in the number of viable cells between 24 and 48 hours. At 48 hours, the number of viable cells in the metronidazole treated C. difficile cultures was greater than that in the control group. Auranofin treated C. difficile cultures showed a reduction in the number of viable cells between 12 and 24 hours and at 48 hours there were no viable cells remaining, thus providing further evidence that auranofin is bactericidal against C. difficile. The reduction in C. difficile viability when treated with vancomycin was delayed compared to the OD measurements and was less pronounced than the reduction observed in the auranofin treated C. difficile cultures, indicating vancomycin may be slower at killing the bacteria than auranofin. This is not surprising as vancomycin has been described as slow acting and bacteriostatic115. Overall these results indicate that the bactericidal auranofin is non-inferior to vancomycin as a treatment for CDI. These results also indicate that metronidazole may be inferior to auranofin and vancomycin as a treatment for CDI. Further investigations, including in vivo animal trials and human clinical trials, need to be performed to properly assess the effectiveness of auranofin as a treatment in comparison to vancomycin and metronidazole. C. difficile spores are the transmittable element10, so it is important to assess the impact of a treatment on sporulation and spore viability. Viable spore counts show that subinhibitory concentrations of auranofin do not influence sporulation, further supporting auranofin as a treatment for CDI, as diminished concentrations of auranofin in the gastrointestinal tract will not trigger or increase sporulation as is often the case with other treatments such as vancomycin109. An increase in sporulation would contribute to the spread of disease and increase the risk of recurrent infections10. Since auranofin does not increase sporulation, even at subinhibitory concentrations, this does not appear to be a problem associated with this potential treatment. In fact, MIC and treatment concentrations of auranofin reduced the number of viable spores from C. difficile cultures between 12 hours (treatment) and 24 hours, with a further reduction between 24 and 48 hours. The C. difficile cultures treated with 250 µM and 500 µM auranofin had no viable spores after 48 hours.
93 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
This reduction in spore counts strongly suggests that auranofin is in some way affecting the spores or the germination process. If auranofin were inhibiting the ability of the bacteria to produce spores, or the low spore counts were simply an artefact of the low number of viable cells, then there would be a plateau in the number of viable spores after treatment, rather than a decline. This narrows the possibility to affecting the viability of the spore itself, or affecting the germination process. Auranofin may be sporicidal or sporistatic. This is certainly an area that warrants further investigation, as a bactericidal treatment that is also sporicidal would prove to be very useful as a CDI treatment because it may also prevent recurrent infection and spread of disease. An unfortunate limitation of this work was the capacity of the anaerobic chamber limited the n for these experiments. This, coupled with the highly variable nature of C. difficile, did result in a loss of statistical power. However, as the difference between treated and untreated cultures was so large (greater than 100-fold in some instance) then these results still carry validity and should therefore not be discounted. Additional work using pure spore cultures and larger experimental numbers is required to fully elucidate the true effect of auranofin on C. difficile spores, but these results are certainly encouraging. The pathology of CDI is caused by the glycosylating toxins TcdA and TcdB25. To assess the impact of auranofin on the cytotoxic activity of TcdB and thus formulate a theory on how auranofin will affect disease pathology, cell cytotoxicity assays were performed using Vero cells, which lose their cellular structure when intoxicated with TcdB122. The cytotoxic activity observed from C. difficile cultures treated with 50 µM, 250 µM or 500 µM auranofin was significantly lower than that observed from the untreated C. difficile cultures. This finding was not surprising, as these cultures had a significant reduction in the number of viable cells, and hence less capacity to produce toxins. What was surprising was the finding that auranofin could directly alter the cytotoxic effect TcdB has on Vero cells. When Vero cells were given 50 µM of auranofin at the same time they were exposed to supernatant from control C. difficile cultures, the cytotoxic activity previously observed from this supernatant was completely reversed. Auranofin may be affecting the toxins directly or may affect the endosomal mediated uptake of the toxins by the Vero cells. Further work needs to be done to determine firstly if auranofin has a similar
94 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
effect on TcdA, and secondly if this is directly on the toxins or on the intoxication of the target eukaryotic cells. The preliminary in vivo work was performed using a CDI mouse model designed to replicate antibiotic associated CDI. In this model, the mouse intestinal microbiota is depleted with an antibiotic cocktail, and then the mice are infected with C. difficile spores. In this experiment, auranofin was given prophylactically one day prior to infection with C. difficile to determine if auranofin would have a strong enough effect on disease progression to warrant further investigation as a treatment. The three doses of auranofin used in this experiment, 40 µg/ml, 200 µg/ml and 400 µg/ml, were chosen based on the maximum dose of auranofin able to be given to the mice (400 µg/ml), half the maximum dose (200 µg/ml) and one tenth the maximum dose (40 µg/ml). Controls for this experiment included the infected control and the uninfected control groups, and three additional uninfected mice groups, each receiving one of the three auranofin prophylaxis doses to observe any adverse effects caused by auranofin. The infected control group showed steady weight loss one day post-infection, which continued until the mice reach 15% loss of total body weight, at which point they are euthanized. This typically occurs two to three days post-infection. Mice surviving to four days post-infection are likely to recover from the disease. The survival and weight loss observations revealed that the low dose of 40 µg/ml is too low to prevent CDI, with 75% of the mice in this group reaching the weight loss endpoint of 15% total body weight. The uninfected mice receiving this dose did not display any weight loss, indicating that the weight loss in the infected group was associated with CDI rather than auranofin. The 400 µg/ml dose was able to prevent CDI, with 100% survival in this group. There was substantial weight loss in both the infected and uninfected mice receiving this dose, indicating that the weight loss observed here may be due to treatment with auranofin, rather than the infection. This dose may be therefore too high to be used for the treatment of CDI. The third dose, 200 µg/ml, demonstrated 100% survival without significant weight loss. It therefore appears to be the ideal dose for treatment of CDI as it is high enough to prevent CDI, and not cause apparent adverse side effects. This group had less C. difficile colonisation after 48 hours than the low dose and infected control.
95 INVESTIGATION OF AURANOFIN AS A TREATMENT FOR C. DIFFICILE INFECTION
This supports the theory that auranofin is preventing CDI, as well reducing the level of TcdB cytotoxic activity against Vero cells, which may be due to the low colonisation, or an indication that auranofin may also be capable of preventing the disease caused by the toxins. Further in vivo work is necessary to fully elucidate the role of auranofin as a treatment for CDI. This work will include full scale CDI mouse models testing auranofin as a prophylaxis and then as a treatment, as well as investigations using a recurrent CDI mouse model. Additionally, the mouse gut microbiota needs to be studied to determine the impact of auranofin since this is an indicator for recurrence of CDI32,33 In conclusion, auranofin is able to reduce C. difficile cell viability in vitro at concentrations as low as 6 µM, as well as reduce the cytotoxic activity of the toxin TcdB at concentrations as low as 50 µM. Furthermore, auranofin does not increase sporulation, even at subinhibitory concentrations. Auranofin was non-inferior to vancomycin in vitro, and metronidazole was inferior to both auranofin and vancomycin in vitro. Preliminary in vivo mouse assays demonstrated that auranofin prophylaxis of 200 µg/ml daily can prevent antibiotic associated CDI in mice without adverse effects. The evidence presented here strongly supports investigating treatment of CDI with auranofin as an alternative to the current antibiotics, vancomycin and metronidazole. Further in vitro and in vivo studies are required, as discussed above, before validation in small scale human clinical trials.
96 DISCUSSION AND CONCLUSION
5. Discussion and conclusion Clostridium difficile infection (CDI) is an antibiotic associated infection of the gastrointestinal tract. The current treatments, vancomycin and metronidazole, are associated with high levels of recurrent and refractory disease4,112,113. This is due to their disruption of the intestinal microbiota, which, when in extreme dysbiosis, requires a faecal microbiota transplant (FMT) to replenish the diversity of the microbiota32,33. It is clear that new therapies are needed to treat and prevent recurrent CDI, and while there are some new developments, many of these are expensive, such as fidaxomicin. One potential candidate, auranofin, was presented by Jackson-Rosario et al. 2009. This group investigated the growth inhibiting activity of auranofin on C. difficile, attributing this activity to the inhibition of selenoprotein synthesis in the bacteria24. There have been a number of studies, predominantly by Theresa Stadtman, that have identified selenoproteins utilised by C. difficile for amino acid fermentation, namely components of the glycine reductase and proline reductase pathways186,190,193,198,206,207. The aim of this study was to further investigate the repurposing of the arthritis drug auranofin, as a treatment for CDI. This overall aim was subdivided into two: the first was to investigate the number of potential selenoproteins in the C. difficile genome using in silico methods, and the second, to further investigate the growth inhibitory activity of auranofin on C. difficile using an in vitro model of CDI. The in silico investigation into the number of potential selenoproteins in C. difficile was necessary to assess the strength of selenium metabolism as a therapeutic target. This investigation identified 86 genes as potential selenoproteins in the C. difficile 630 genome, 87.21% of which showed high levels of conservation (>95%) among 100 C. difficile genomes. The protein products of these genes had a variety of putative functions including metabolism, information storage and processing, as well as cellular processes and signalling. The results of this investigation were surprising for a number of reasons. The first was the number of selenoproteins identified. Previously there were four confirmed selenoproteins found in C. difficile, (selenide water dikinase (SelD), glycine reductase complex protein A (GrdA), glycine reductase complex protein B (GrdB),
97 DISCUSSION AND CONCLUSION
and D-proline reductase protein A (PrdA)211. The identification of a possible 82 more was unexpected as there have not been any reports of a selenoproteomes of this size before. Previous reports on the number of selenoproteins in bacteria have identified a total of 25 different selenoproteins and three selenoprotein rich phyla: Deltaproteobacteria (22 selenoproteins), Firmicutes–Clostridia (16 selenoproteins) and Actinobacteria (12 selenoproteins)211. Further in vitro investigations are required to confirm that all the genes identified here produce functional selenoproteins. Even if just a portion of them are true selenoproteins, this would make the C. difficile selenoproteome the largest selenoproteomes ever reported. The second unexpected result was the high level of conservation among these genes. C. difficile has an unusually low level of genome conservation63,212, with reports estimating the core genome to be just 16-23% of the entire genome. This is one of the smallest core genomes of any bacteria212. A total of 61 of the selenoprotein genes identified here were 100% conserved and are therefore likely to play an essential role in C. difficile survival. This represents ~6.1% of the estimated C. difficile core genome. A further 14 were highly conserved. The sheer number of potential selenoproteins identified in C. difficile combined with the high levels of genome conservation and the variety of putative functions suggests that selenium plays an essential role in this organism. Based on these findings it can be concluded that selenium metabolism is a good therapeutic target to treat CDI. Once selenium metabolism was established as a good therapeutic target, it was necessary to test auranofin as a treatment for CDI in an in vitro model. This investigation found that auranofin was able to inhibit C. difficile growth and reduce cell viability at concentrations as low as 6 µM, eliminating all viable cells at a concentration of 50 µM. Auranofin also protected Vero cells from the cytotoxic activity of C. difficile toxins TcdA and TcdB, and did not increase sporulation. The activity of auranofin against C. difficile was more effective than initially hypothesized. It was originally thought that auranofin would be bacteriostatic and only inhibit bacterial growth, however the results strongly indicate bactericidal activity. There was clearly reduced cell viability at a concentration of 6 µM, and concentrations of 50 µM and higher consistently showed no viable bacterial cells after treatment.
98 DISCUSSION AND CONCLUSION
Treatments such as Vancomycin can induce sporulation at subinhibitory concentrations109. Some disinfectants, such as alcohol based hand sanitizer, also promote sporulation4. C. difficile spores are highly resilient, able to survive in the environment for several months and are unaffected by antibiotics such as metronidazole and vancomycin. Spores are the transmissible element, so an increase in sporulation would contribute to the spread of CDI109. Spores also play a role in recurrent CDI94, with speculation that the recurrence of disease is caused by failure to eliminate the spores from the gastrointestinal tract. Upon cessation of antibiotic therapy, the spores would then be able to germinate, thus reviving the infection. An increase of sporulation would also increase the risk of recurrent disease. Auranofin did not induce sporulation, even at subinhibitory concentrations. After treatment with 6 µM or higher of auranofin, there was a reduction in the number of viable spores yielded, indicating that auranofin was in some way affecting the C. difficile spores. This activity could be sporicidal, affecting the viability of the spores, or sporistatic, affecting the germination process. Distinguishing between sporicidal and sporistatic activity is a somewhat difficult process, requiring the elimination of the active molecule, in this case auranofin. This may be either by physical removal of the molecule or by neutralizing its activity213. If auranofin were sporistatic, this may reduce the risk of recurrent CDI, as the remaining spores in the gastrointestinal tract would be unable to germinate, however it would be unlikely to reduce the spread of disease as the spores would be able to germinate inside a new host that has not been treated with auranofin. If auranofin were sporicidal, this could both reduce the risk of recurrence and spread of disease. Further work is required to fully elucidate the activity of auranofin against C. difficile spores. An unexpected outcome of this investigation was the elimination of C. difficile toxin activity by auranofin. It was not surprising that there was less toxin activity from the supernatant of C. difficile cultures treated with auranofin, as one would expect that these cultures would be producing less toxin, particularly since three of the known C. difficile selenoproteins play a role in energy production in this bacteria. However, auranofin was able to neutralize the activity of the toxins produced by the untreated control C. difficile cultures. These cultures normally
99 DISCUSSION AND CONCLUSION
exhibit quite high cytotoxic activity against Vero cells, resulting in the Vero cells rounding and lifting from the surface of their well. The addition of 50 µM of auranofin at the time the C. difficile supernatant was added to the Vero cells media completely protected the Vero cells from the cytotoxic activity of the C. difficile toxins. It remains unclear if auranofin is directly affecting the toxin activity, or is in some way preventing the endosomal mediated uptake of the toxins51 by the Vero cells. While further work is needed to determine the mechanisms at work here, this finding is very significant given that the toxins are responsible for the pathology of CDI. The cytotoxic activity of the toxins TcdA and TcdB results in the apoptosis of the intestinal epithelial cells, which leads to the accumulation of fluid, inflammation and intestinal injury that are clinical symptoms of CDI52. Prevention of the apoptosis of the epithelial cells would lessen the severity of CDI in patients, reducing the likelihood of complications requiring surgical intervention and improve patient outcomes. The strong bactericidal activity and potential sporicidal activity of auranofin is most likely due to the essential nature of selenium that was highlighted in the results of the in silico component of this study. Selenium in the form of selenocysteine is known to be an essential component of the glycine reductase complex (GrdA, GrdB) and the proline reductase complex (PrdA)190,195,207. These proteins are responsible for the reduction of the amino acids glycine and proline, and form energy conservation pathways in C. difficile. By eliminating these energy conservation pathways through the removal of selenium, it was anticipated that bacterial growth would be affected. It is clear that auranofin causes more than just inhibition of C. difficile bacterial growth, indicating a stronger reliance on selenium than previously thought. Selenium utilization in C. difficile begins with the conversion of selenium to selenophosphate, the selenium donor for selenocysteine. The synthesis of selenophosphate is catalysed by the selenoprotein SelD29,181, which is known to be essential for selenium utilization in all domains of life29,181,211,214. It has been previously shown by Dembek et al. that SelD is essential for sporulation in C. difficile29, hence it was not surprising to discover that auranofin affected the viability and/or germination of C. difficile spores. However, Dembek et al. observed that SelD, was not essential for in vitro vegetative cell growth29, which was surprising given its essential role in selenium utilization. This
100 DISCUSSION AND CONCLUSION
study identified a number of potential selenoproteins that are involved in the essential pathways for vegetative cell growth identified by Dembek et al., including ribosomal pathways, metabolic pathways, DNA replication pathways and secondary metabolite biosynthesis pathways29. While SelD may not be essential for C. difficile vegetative cell growth, selenium clearly plays a vital role in C. difficile growth, viability and sporulation. Based on the significant effects exhibited by auranofin on C. difficile in vitro, a preliminary in vivo investigation into auranofin as a treatment for CDI was performed at Monash University with the assistance of Associate Professor Dena Lyras and Dr Melanie Hutton. These experiments showed that a dose of 200 µg/ml was able to protect C57BL/6J mice from the cytotoxic effects of C. difficile. Limitations of this experiment were the small number of mice used (n=5) and the fact that auranofin was tested as a prophylaxis rather than a treatment. Unfortunately, due to time constraints further in vivo experiments were not performed, however there are plans for future experiments that will rectify the limitations discussed here. The results of this preliminary investigation are encouraging, and provide strong evidence supporting further testing of auranofin as a treatment for CDI. Auranofin effectively eliminated the viability, spread and cytotoxic effects of C. difficile by removing the bioavailability of selenium. As described by Jackson-Rosario et al., auranofin forms a stable and irreversible bond with inorganic selenium, thus preventing its use by the bacteria24. In their studies, Jackson-Rosario et al. estimated that just 14% of all bacteria used selenium24, which would make auranofin a narrow-spectrum treatment. A more recent study by Peng et al. estimated this number to be ~33%181, however they took their analysis further and divided the bacteria up based on the environment in which they were found, i.e. terrestrial environments, aquatic environments and host environments. Of the 1754 bacteria identified as having selenium utilization traits, just 14.7% were associated with a host environment. It should be noted that Peng et al. defined evidence of selenium utilization traits as the presence SelD181, so this 14.7% would include bacteria that have SelD, but no other selenium utilization traits, such as the other selenoproteins. Peng et al. speculated that these bacteria with ‘orphan’ SelD may possess selenium pathways that have not yet been
101 DISCUSSION AND CONCLUSION
characterised181. Selenium utilization traits are present in all three domains of life, and evolutionary analyses suggest that most bacterial lineages have lost their ability to use selenium211,214, so it is far more likely that bacteria with orphan SelD have lost their other selenium utilization traits and that SelD is a remnant of their ancestral selenium use. This would make the number of host related bacteria that rely on selenium, and thus would be affected by auranofin, less than 14%, so auranofin may still be considered a narrow spectrum treatment for CDI. Auranofin has been investigated as an antibiotic for other infections recently, including Staphylococcus aureus and Enterococcus faecalis, along with other Gram-positive bacteria. Studies by Harbut et al. and Thangmai et al. demonstrated that auranofin had good in vitro activity against these organisms, and proposed that the mechanism of action was against the redox protein thioredoxin reductase (TrxR)215,216 As discussed previously auranofin is known to have an affinity for redox proteins such as TrxR163, and that in a number of organism TrxR is a selenoprotein211. It is interesting to note that there is evidence both have selenium utilization properties211, so activity by targeting selenium cannot be ruled out. These two studies also investigated that activity of auranofin against Gram- negative bacteria. Both studies found limited activity against Gram-negative bacteria, suggesting that the bacterial membrane was protecting the bacteria from auranofin215,216. This narrows the spectrum of activity to Gram-positive bacteria only. Treatment of CDI is increasingly challenging. Production of spores leads to widespread disease, and emerging resistance to antibiotics contributes to refractory and recurrent infections. These factors have placed C. difficile on the CDCs urgent public health threat list5. It is clear that new antibiotics are needed for this disease. Ideal bacterial therapies need to target essential pathways for efficiency and reduced risk of resistance202, and should be narrow spectrum to reduce the disruption of the microbiota32,33. They also need to be able to infiltrate the site of infection in a high enough concentration to eradicate the pathogen. Additional requirements specific to CDI include a reduction in transmission capability (i.e. spore production and/or viability) and action against the cytotoxic effects of TcdA and TcdB.
102 DISCUSSION AND CONCLUSION
This study has provided evidence supporting the essential nature of selenium as a metabolite for C. difficile, and proposes that using auranofin to limit the bioavailability of selenium to C. difficile is a strong therapeutic approach. The drug does not act on a molecular target on or within the bacteria, thus strongly reducing the likelihood of C. difficile acquiring resistance to this treatment. Auranofin may be able to be considered a narrow spectrum antibiotic, as previous studies have determined that <14% of all host bacteria utilise selenium. Furthermore, auranofin has limited activity against Gram-negative bacteria. Auranofin also has ideal pharmacokinetic properties for CDI treatment, as up to 85% of the administered dose would make it to the distal colon, the site of infection for C. difficile. The in vitro and in vivo aspects to this study have demonstrated that auranofin can effectively eliminate vegetative C. difficile cell viability as well as affect spore production and/or viability, thus eliminating the infection and potentially minimising transmission. Auranofin can also protect Vero cells and mice from the cytotoxic activity of TcdB, thus potentially preventing symptoms such as diarrhoea, intestinal inflammation and intestinal damage that are typically caused by this infection. Furthermore, auranofin is a repurposed drug, meaning it has been approved for another disease and is now off-license, so the cost is significantly reduced. Limitations of this study are primarily due to time constraints. These limitations include no in vitro data to support the in silico identification of selenoproteins, the space limitations within the anaerobic chamber for the in vitro growth assays, as well as only doing a preliminary in vivo investigation. Future studies will address these limitations, as well as further investigate the activity of auranofin on both C. difficile toxins and spores. In conclusion, auranofin is a strong candidate for the treatment of CDI due to its ability to limit the bioavailability of selenium. The data presented here strongly supports further investigations including in vivo animal trials, before proceeding to human clinical trials. With increasing failure rates of the current treatments and the alternatives deemed expensive, invasive and risky, auranofin may be the ideal treatment for CDI as it appears to be effective, easy to administer and cost effective.
103 REFERENCES
6. References 1. Schroeder, M.S. Clostridium difficile-associated diarrhea. American family physician. 71, 921-928 (2005).
2. Kelly, C.P. & LaMont, J.T. Clostridium difficile –more difficult than ever. The New England Journal of Medicine. 359, 1932-1940 (2008).
3. Hamilton, M.J., Weingarden, A.R., Sadowsky, M.J. & Khoruts, A. Standardized frozen preparation for transplantation of fecal microbiota for recurrent Clostridium difficile infection. The American Journal of Gastroenterology. 107, 761-767 (2012).
4. Howerton, A., Patra, M. & Abel-Santos, E. A new strategy for the prevention of Clotridium difficile infections. The Journal of Infectious Diseases. jit068 (2013).
5. CDC. Antibiotic resistance threats in the United States, 2013. https://www.cdc.gov/drugresistance/threat-report-2013/ (accessed 26 November 2014).
6. Riddle, D.J. & Dubberke, E.R. Trends in Clostridium difficile disease: epidemiology and intervention. Infections in Medicine. 26, 211-220 (2009).
7. Songer, J.G. et al. Clostridium difficile in retail meat products, USA, 2007. Emerging Infectious Diseases .15, 819-821 (2009).
8. Cox, A.D. et al. Investigating the candidacy of a lipoteichoic acid-based glycoconjugate as a vaccine to combat Clostridium difficile infection. Glycoconjugate Journal. 30, 1-13 (2013).
9. van Dissel, J.T. et al. Bovine antibody-enriched whey to aid in the prevention of a relapse of Clostridium difficile-associated diarrhoea: preclinical and preliminary clinical data. Journal of Medical Microbiology. 54, 197-205 (2005).
10. Burns, D.A. & Minton, N.P. Sporulation studies in Clostridium difficile. Journal of Microbiological Methods. 87, 133-138 (2011).
11. Jenkin, G.A. Clostridium difficile infection: an Australian clinical perspective. Microbiology Australia. 36, 106-108 (2015).
12. Antunes, A. et al. Global transcriptional control by glucose and carbon regulator CcpA in Clostridium difficile. Nucleic acids research. 40, 10701- 10718, (2012).
104 REFERENCES
13. Lo Vecchio, A., Della Ventura, B. & Nicastro, E. Clostridium difficile antibodies (WO2013028810): a patent evaluation. Expert Opinion on Therapeutic Patents. 23, 1635-1640 (2013).
14. Musher, D.M. et al. Clostridium difficile colitis that fails conventional metronidazole therapy: response to nitazoxanide. Journal of Antimicrobial Chemotherapy. 59, 705-710 (2007).
15. Nagaro, K.J. et al. Nontoxigenic Clostridium difficile Protects Hamsters against challenge with historic and epidemic strains of toxigenic BI/NAP1/027 C. difficile. Antimicrobial agents and chemotherapy. 57, 5266-5270 (2013).
16. Tannock, G.W. et al. A new macrocyclic antibiotic, fidaxomicin (OPT-80), causes less alteration to the bowel microbiota of Clostridium difficile- infected patients than does vancomycin. Microbiology (Reading, England). 156, 3354-3359 (2010).
17. Pepin, J., Valiquette, L., Gagnon, S., Routhier, S. & Brazeau, I. Outcomes of Clostridium difficile-associated disease treated with metronidazole or vancomycin before and after the emergence of NAP1/027. The American journal of gastroenterology 102, 2781-2788 (2007).
18. Buonomo, E.L. et al. Role of IL-23 signaling in Clostridium difficile colitis. The Journal of Infectious Diseases. 208, 917-920 (2013).
19. Manges, A.R. et al. Comparative metagenomic study of alterations to the intestinal microbiota and risk of nosocomial Clostridum difficile-associated disease. Journal of Infectious Diseases. 202, 1877-1884 (2010).
20. Underwood, S. et al. Characterization of the sporulation initiation pathway of Clostridium difficile and its role in toxin production. Journal of Bacteriology. 191, 7296-7305 (2009).
21. Dupuy, B. et al. Regulation of toxin and bacteriocin gene expression in Clostridium by interchangeable RNA polymerase sigma factors. Molecular Microbiology. 60, 1044-1057 (2006).
22. Cheng, A.C. et al. Australasian Society for Infectious Diseases guidelines for the diagnosis and treatment of Clostridium difficile infection. The Medical Journal of Australia. 194, 353-358 (2011).
23. Oka, K. et al. Molecular and microbiological characterization of Clostridium difficile isolates from single, relapse, and reinfection cases. Journal of Clinical Microbiology. 50, 915-921 (2012).
105 REFERENCES
24. Jackson-Rosario, S. et al. Auranofin disrupts selenium metabolism in Clostridium difficile by forming a stable Au-Se adduct. Journal of Biological Inorganic Chemistry. 14, 507-519 (2009).
25. Lyras, D. et al. Toxin B is essential for virulence of Clostridium difficile. Nature. 458, 1176-1179, (2009).
26. Stuart, R. L. & Marshall, C. Clostridium difficile infection: a new threat on our doorstep. The Medical Journal of Australia. 194, 331-332 (2011).
27. Riley, T.V., O'Neill, G.L., Bowman, R.A. & Golledge, C.L. Clostridium difficile-associated diarrhoea: epidemiological data from Western Australia. Epidemiology and Infection. 113, 13-20 (1994).
28. Sougioultzis, S. et al. Clostridium difficile toxoid vaccine in recurrent C. difficile-associated diarrhea. Gastroenterology. 128, 764-770 (2005).
29. Dembek, M., Stabler, R.A., Witney, A.A., Wren, B. W. & Fairweather, N.F. Transcriptional analysis of temporal gene expression in germinating Clostridium difficile 630 endospores. PloS One. 8, e64011, (2013).
30. Walker, A.W. & Lawley, T.D. Therapeutic modulation of intestinal dysbiosis. Pharmacological Research. 69, 75-86, (2013).
31. Cervantes, J. Use your antibiotics wisely. Consequences to the intestinal microbiome. FEMS Microbiology Letters. 363, fnw081 (2016).
32. Willing, B.P., Russell, S.L. & Finlay, B.B. Shifting the balance: antibiotic effects on host-microbiota mutualism. Nature reviews. Microbiology. 9, 233-243 (2011).
33. Willing, B.P., Vacharaksa, A., Croxen, M., Thanachayanont, T. & Finlay, B.B. Altering host resistance to infections through microbial transplantation. PloS One. 6, e26988 (2011).
34. The Human Microbiome Project Consortium. Structure, function and diversity of the healthy human microbiome. Nature. 486, 207-214, (2012).
35. Kelly, C.P. & Lamont, J.T. in Sleisenger and Fordtran's Gastrointestinal and Liver Disease: Pathophysiology, Diagnosis, Management. Vol. 2, Ch. 108, 1889-1903 (Saunders/Elsevier, 2010).
36. Relman, D.A. Restoration of the gut microbial habitat as a disease therapy. Nature Biotechnology. 31, 35-37 (2013).
106 REFERENCES
37. McFarland, L.V. et al. A randomized placebo-controlled trial of Saccharomyces boulardii in combination with standard antibiotics for Clostridium difficile disease. JAMA. 271, 1913-1918 (1994).
38. Thomas, C. & Riley, T.V. Restriction of third generation cephalosporin use reduces the incidence of Clostridium difficile-associated diarrhoea in hospitalised patients. Communicable Diseases Intelligence. 27 Suppl, S28- S31 (2003).
39. Wilson, K.H., Sheagren, J.N. & Freter, R.G. Selective antimicrobial modulation of human microbial flora. The Journal of Infectious Diseases. 145, 427-429 (1982).
40. Ross, C.L., Spinler, J.K. & Savidge, T.C. Structural and functional changes within the gut microbiota and susceptibility to Clostridium difficile infection. Anaerobe. 41, 37-43 (2016).
41. Zidaric, V. et al. Multiclonal presence of Clostridium difficile PCR ribotypes 078, 126 and 033 within a single calf farm is associated with differences in antibiotic resistance and sporulation properties. Applied and Environmental Microbiology. AEM-02185 (2012).
42. Barbut, F. et al. Prospective study of Clostridium difficile infections in Europe with phenotypic and genotypic characterisation of the isolates. Clinical Microbiology and Infection. 13, 1048-1057 (2007).
43. Deakin, L.J. et al. The Clostridium difficile spo0A gene is a persistence and transmission factor. Infection and Immunity. 80, 2704-2711 (2012).
44. Dumyati, G. et al. Community-associated Clostridium difficile infections, Monroe County, New York, USA. Emerging Infectious Diseases 18, 392- 400 (2012).
45. Figueroa, I. et al. Relapse versus reinfection: recurrent Clostridium difficile infection following treatment with fidaxomicin or vancomycin. Clinical Infectious Diseases. 55 Suppl 2, S104-109 (2012).
46. Johnson, S., Schriever, C., Galang, M., Kelly, C.P. & Gerding, D.N. Interruption of recurrent Clostridium difficile-associated diarrhea episodes by serial therapy with vancomycin and rifaximin. Clinical Infectious Diseases. 44, 846-848 (2007).
47. Owens, C., Broussard, E. & Surawicz, C. Fecal microbiota transplantation and donor standardization. Trends in Microbiology. 21, 443-445 (2013).
107 REFERENCES
48. Fimlaid, K.A. et al. Global analysis of the sporulation pathway of Clostridium difficile. PLoS Genetics 9, e1003660 (2013).
49. Barbut, F.P. et al. Clinical features of Clostridium difficile–associated infections and molecular characterization of strains: results of a retrospective study, 2000‐2004. Infection Control and Hospital Epidemiology. 28, 131-139 (2007).
50. Kato, H. et al. Colonisation and transmission of Clostridium difficile in healthy individuals examined by PCR ribotyping and pulsed-field gel electrophoresis. Journal of Medical Microbiology. 50, 720-727 (2001).
51. Papatheodorou, P., Zamboglou, C., Genisyuerek, S., Guttenberg, G. & Aktories, K. Clostridial glucosylating toxins enter cells via clathrin- mediated endocytosis. PloS One 5, e10673 (2010).
52. Carter, G.P., Rood, J.I. & Lyras, D. The role of toxin A and toxin B in Clostridium difficile-associated disease: Past and present perspectives. Gut Microbes. 1, 58-64 (2010).
53. Magasanik, B. Global regulation of gene expression. Proceedings of the National Academy of Sciences. 97, 14044-14045 (2000).
54. Sonenshein, A.L. CodY, a global regulator of stationary phase and virulence in Gram-positive bacteria. Current Opinion in Microbiology. 8, 203-207 (2005).
55. Dineen, S.S., McBride, S. M. & Sonenshein, A.L. Integration of metabolism and virulence by Clostridium difficile CodY. Journal of Bacteriology. 192, 5350-5362 (2010).
56. Cohen, S.H., Tang, Y.J. & Silva, J.Jr., Analysis of the pathogenicity locus in Clostridium difficile strains. The Journal of Infectious Diseases. 181, 659-663, (2000).
57. Tan, K.S., Wee, B.Y. & Song, K.P. Evidence for holin function of tcdE gene in the pathogenicity of Clostridium difficile. Journal of Medical Microbiology 50, 613-619 (2001).
58. Pothoulakis, C. & Lamont, J.T. Microbes and microbial toxins: paradigms for microbial-mucosal interactions II. The integrated response of the intestine to Clostridium difficile toxins. American Journal of Physiology- Gastrointestinal and Liver Physiology. 280, G178-G183 (2001).
108 REFERENCES
59. Just, I. et al. The enterotoxin from Clostridium difficile (ToxA) monoglucosylates the Rho proteins. Journal of Biological Chemistry. 270, 13932-13936 (1995).
60. Just, I. et al. Glucosylation of Rho proteins by Clostridium difficile toxin B. (1995).
61. Just, I. & Gerhard, R in Reviews of Physiology, Biochemistry and Pharmacology. Vol. 152, 23-47 (Springer Berlin Heidelberg, 2005).
62. Dupuy, B. & Sonenshein, A.L. Regulated transcription of Clostridium difficile toxin genes. Molecular Microbiology. 27, 107-120 (1998).
63. Sebaihia, M. et al. The multidrug-resistant human pathogen Clostridium difficile has a highly mobile, mosaic genome. Nature Genetics. 38, 779-786, (2006).
64. Papatheodorou, P. et al. Clostridium difficile binary toxin CDT induces clustering of the lipolysis-stimulated lipoprotein receptor into lipid rafts. mBio. 4, e00244-13 (2013).
65. Puri, B.K., Hakkarainen-Smith, J.S. & Monro, J.A. The potential use of cholestyramine to reduce the risk of developing Clostridium difficile- associated diarrhoea in patients receiving long-term intravenous ceftriaxone. Medical Hypotheses 84, 78-80 (2015).
66. Carter, G. P. et al. Binary toxin production in Clostridium difficile is regulated by CdtR, a LytTR family response regulator. Journal of Bacteriology 189, 7290-7301 (2007).
67. Eckert, C. et al. Prevalence and pathogenicity of binary toxin-positive Clostridium difficile strains that do not produce toxins A and B. New Microbes and New Infections 3, 12-17 (2015).
68. Knapp, O., Benz, R. & Popoff, M.R. Pore-forming activity of clostridial binary toxins. Biochimica et Biophysica Acta. 1858, 512-525 (2015).
69. Rupnik, M., Grabnar, M. & Geric, B. Binary toxin producing Clostridium difficile strains. Anaerobe. 9, 289-294 (2003).
70. Bauer, M.P., Nibbering, P.H., Poxton, I.R., Kuijper, E.J. & van Dissel, J.T. Humoral immune response as predictor of recurrence in Clostridium difficile infection. Clinical Microbiology and Infection. 20, 1323-1328 (2014).
109 REFERENCES
71. Buonomo, E.L. & Petri, W.A., Jr. The microbiota and immune response during Clostridium difficile infection. Anaerobe. 41, 79-84 (2016).
72. Kelly, C.P. Immune response to Clostridium difficile infection. European Journal of Gastroenterology & Hepatology. 8, 1048-1053 (1996).
73. Jose, S. & Madan, R. Neutrophil-mediated inflammation in the pathogenesis of Clostridium difficile infections. Anaerobe. 41, 85-90 (2016).
74. Fawley, W.N., et al. Efficacy of hospital cleaning agents and germicides against epidemic Clostridium difficile strains. Infection Control and Hospital Epidemiology. 28, 920-925 (2007).
75. Stuart, R. L. et al. ASID/AICA position statement - infection control guidelines for patients with Clostridium difficile infection in healthcare settings. Healthcare Infection. 16, 33-39 (2011).
76. McFarland, L.V. et al. Nosocomial acquisition of Clostridium difficile infection. New England Journal of Medicine. 320, 204 (1989).
77. Skoutelis, A.T., Westenfelder, G.O., Beckerdite, M. & Phair, J.P. Hospital carpeting and epidemiology of Clostridium difficile. American Journal of Infection Control. 22, 212-217 (1994).
78. Bakri, M.M., Brown, D.J., Butcher, J.P. & Sutherland, A.D. Clostridium difficile in ready-to-eat salads, Scotland. Emerging Infectious Diseases. 15, 817-818 (2009).
79. Keel, K., Brazier, J.S., Post, K.W., Weese, S. & Songer, J.G. Prevalence of PCR ribotypes among Clostridium difficile isolates from pigs, calves, and other species. Journal of Clinical Microbiology. 45, 1963-1964 (2007).
80. Bouttier, S. et al. Clostridium difficile in ground meat, France. Emerging Infectious Diseases. 16, 733-735 (2010).
81. Weese, J.S., Avery, B.P., Rousseau, J. & Reid-Smith, R.J. Detection and enumeration of Clostridium difficile spores in retail beef and pork. Applied and Environmental Microbiology. 75, 5009-5011 (2009).
82. Rodriguez-Palacios, A., Ilic, S. & LeJeune, J.T. Clostridium difficile with moxifloxacin/clindamycin resistance in vegetables in Ohio, USA, and prevalence meta-analysis. Journal of Pathogens. 2014, 158601 (2014).
83. Rodriguez-Palacios, A. et al. Possible seasonality of Clostridium difficile in retail meat, Canada. Emerging Infectious Diseases. 15, 802-805 (2009).
110 REFERENCES
84. Rodriguez-Palacios, A., Reid-Smith, R.J., Staempfli, H.R. & Weese, J.S. Clostridium difficile survives minimal temperature recommended for cooking ground meats. Anaerobe. 16, 540-542 (2010).
85. Burt, S.A., Siemeling, L., Kuijper, E.J. & Lipman, L.J. Vermin on pig farms are vectors for Clostridium difficile PCR ribotypes 078 and 045. Veterinary Microbiology. 160, 256-258 (2012).
86. Knight, D.R., Thean, S., Putsathit, P., Fenwick, S. & Riley, T.V. Clostridium difficile carriage in Australian cattle: a cross-sectional study reveals high prevalence of non-PCR ribotype 078 strains in veal calves at slaughter. Applied and Environmental Microbiology. AEM-03951 (2013).
87. Knight, D.R., Squire, M.M. & Riley, T.V. Laboratory detection of Clostridium difficile in piglets in Australia. Journal of Clinical Microbiology. 52, 3856-3862 (2014).
88. Knight, D.R., Squire, M.M. & Riley, T.V. Clostridium difficile in Australian neonatal pigs; nationwide surveillance study shows high prevalence and heterogeneity of PCR ribotypes. Applied and Environmental Microbiology. AEM-03032 (2014).
89. Knetsch, C. et al. Whole genome sequencing reveals potential spread of Clostridium difficile between humans and farm animals in the Netherlands, 2002 to 2011. Euro Surveillance : bulletin Europeen sur les maladies transmissibles = European communicable disease bulletin 19 (2014).
90. Clements, A.C.A., Magalhães, R.J.S., Tatem, A.J., Paterson, D.L. & Riley, T.V. Clostridium difficile PCR ribotype 027: assessing the risks of further worldwide spread. The Lancet Infectious Diseases. 10, 395-404 (2010).
91. Dingle, K. E. et al. Clinical Clostridium difficile: clonality and pathogenicity locus diversity. PloS One 6, 1-10 (2011).
92. Dubberke, E.R. et al. Attributable inpatient costs of recurrent Clostridium difficile infections. Infection Control and Hospital Epidemiology. 35, 1400- 1407 (2014).
93. Trubiano, J. A. et al. Australasian Society of Infectious Diseases updated guidelines for the management of Clostridium difficile infection in adults and children in Australia and New Zealand. Internal Medicine Journal. 46, 479-493, (2016).
94. Cole, S.A. & Stahl, T.J. Persistent and recurrent Clostridium difficile colitis. Clinics in Colon and Rectal Ssurgery. 28, 65-69 (2015).
111 REFERENCES
95. Ackermann, G. et al. Antecedent use of fluoroquinolones is associated with resistance to moxifloxacin in Clostridium difficile. Clinical Microbiology and Infection. 9, 526-530 (2003).
96. Dridi, L., Tankovic, J., Burghoffer, B., Barbut, F. & Petit, J.C. gyrA and gyrB mutations are implicated in cross-resistance to ciprofloxacin and moxifloxacin in Clostridium difficile. Antimicrobial Agents and Chemotherapy. 46, 3418-3421 (2002).
97. Freeman, J. et al. Pan-European longitudinal surveillance of antibiotic resistance among prevalent Clostridium difficile ribotypes. Clinical Microbiology and Infection. 21, 248-e9 (2014).
98. Harnvoravongchai, P., Pipatthana, M., Chankhamhaengdecha, S. & Janvilisri, T. Insights into drug resistance mechanisms in Clostridium difficile. Essays In Biochemistry. 61, 81-88 (2017).
99. Lynch, T. et al. Characterization of a stable, metronidazole-resistant Clostridium difficile clinical isolate. PloS One. 8, e53757 (2013).
100. Pelaez, T. et al. Metronidazole resistance in Clostridium difficile is heterogeneous. Journal of Clinical Microbiology. 46, 3028-3032 (2008).
101. Chilton, C. et al. Efficacy of surotomycin in an in vitro gut model of Clostridium difficile infection. Journal of Antimicrobial Chemotherapy. 69, 2426-2433 (2014).
102. Alam, M. Z., Wu, X., Mascio, C., Chesnel, L. & Hurdle, J. G. Mode of action and bactericidal properties of Surotomycin against growing and non- growing Clostridium difficile. Antimicrobial Agents and Chemotherapy. 59, 5165-5170(2015).
103. Bouillaut, L. et al. Effects of surotomycin on Clostridium difficile viability and toxin production in vitro. Antimicrobial Agents and Chemotherapy. 59, 4199-4205 (2015).
104. Alam, M. J., Anu, A., Walk, S. T. & Garey, K. W. Investigation of potentially pathogenic Clostridium difficile contamination in household environs. Anaerobe. 27C, 31-33 (2014).
105. Bolton, R. & Culshaw, M. Faecal metronidazole concentrations during oral and intravenous therapy for antibiotic associated colitis due to Clostridium difficile. Gut. 27, 1169-1172 (1986).
106. Baines, S. D. et al. SMT19969 as a treatment for Clostridium difficile infection: an assessment of antimicrobial activity using conventional
112 REFERENCES
susceptibility testing and an in vitro gut model. The Journal of Antimicrobial Chemotherapy. 70, 182-189 (2014).
107. Al-Nassir, W. N. et al. Both oral metronidazole and oral vancomycin promote persistent overgrowth of vancomycin-resistant enterococci during treatment of Clostridium difficile-associated disease. Antimicrobial Agents and Chemotherapy. 52, 2403-2406 (2008).
108. Baines, S. D., O'Connor, R., Saxton, K., Freeman, J. & Wilcox, M.H. Activity of vancomycin against epidemic Clostridium difficile strains in a human gut model. The Journal of Antimicrobial Chemotherapy. 63, 520- 525 (2009).
109. Aldape, M. J., Heeney, D. D., Bryant, A. E. & Stevens, D.L. Tigecycline suppresses toxin A and B production and sporulation in Clostridium difficile. The Journal of Antimicrobial Chemotherapy. 70, 153-159 (2014).
110. Corbett, D. et al. In vitro susceptibility of Clostridium difficile to SMT19969 and comparators, as well as the killing kinetics and post- antibiotic effects of SMT19969 and comparators against C. difficile. Journal of Antimicrobial Chemotherapy, dkv006 (2015).
111. Garneau, J. R., Valiquette, L. & Fortier, L.C. Prevention of Clostridium difficile spore formation by subinhibitory concentrations of tigecycline and piperacillin/tazobactam. BMC Infectious Diseases. 14, 29 (2014).
112. Musher, D. M. et al. Relatively poor outcome after treatment of Clostridium difficile colitis with metronidazole. Clinical Infectious Diseases. 40, 1586- 1590 (2005).
113. Chilton, C. et al. Successful treatment of simulated Clostridium difficile infection in a human gut model by fidaxomicin first line and after vancomycin or metronidazole failure. Journal of Antimicrobial Chemotherapy, dkt347 (2013).
114. Goldstein, E.J., Babakhani, F. & Citron, D.M. Antimicrobial activities of fidaxomicin. Clinical Infectious Diseases. 55 Suppl 2, S143-148 (2012).
115. Louie, T. J. et al. Fidaxomicin versus vancomycin for Clostridium difficile infection. New England Journal of Medicine. 364, 422-431 (2011).
116. Cornely, O. A. et al. Clinical efficacy of fidaxomicin compared with vancomycin and metronidazole in Clostridium difficile infections: a meta- analysis and indirect treatment comparison. The Journal of Antimicrobial Chemotherapy. dku261 (2014).
113 REFERENCES
117. Artsimovitch, I., Seddon, J. & Sears, P. Fidaxomicin is an inhibitor of the initiation of bacterial RNA synthesis. Clinical Infectious Diseases. 55, S127-131 (2012).
118. Esposito, G. et al. Rifaximin improves Clostridium difficile toxin A- induced toxicity in Caco-2 cells by the PXR-dependent TLR4/MyD88/NF- kappaB pathway. Frontiers in Pharmacology. 7, 120 (2016).
119. Scarpignato, C. & Pelosini, I. Experimental and clinical pharmacology of rifaximin, a gastrointestinal selective antibiotic. Digestion. 73, 13-27 (2006).
120. Basseres, E. et al. Impact on toxin production and cell morphology in Clostridium difficile by ridinilazole (SMT19969), a novel treatment for C. difficile infection. The Journal of Antimicrobial Chemotherapy. dkv498 (2016).
121. Debast, S. B., Bauer, M. P., Sanders, I. M., Wilcox, M.H. & Kuijper, E.J. Antimicrobial activity of LFF571 and three treatment agents against Clostridium difficile isolates collected for a pan-European survey in 2008: clinical and therapeutic implications. The Journal of Antimicrobial Chemotherapy. dkt013 (2013).
122. Bhansali, S. G. et al. Pharmacokinetics of LFF571 and vancomycin in patients with moderate Clostridium difficile infections. Antimicrobial Agents and Chemotherapy. 59, 1441-14465(2014).
123. Agito, M.D., Atreja, A. & Rizk, M.K. Fecal microbiota transplantation for recurrent C difficile infection: Ready for prime time? Cleveland Clinic Journal of Medicine. 80, 101-108 (2013).
124. Broecker, F. et al. Long-term changes of bacterial and viral compositions in the intestine of a recovered Clostridium difficile patient after fecal microbiota transplantation. Cold Spring Harbor Molecular Case Studies 2, a000448 (2016).
125. Broecker, F. et al. Analysis of the intestinal microbiome of a recovered Clostridium difficile patient after fecal transplantation. Digestion. 88, 243- 251 (2013).
126. Pakyz, A.L., Moczygemba, L.R., VanderWielen, L.M. & Edmond, M.B. Fecal microbiota transplantation for recurrent Clostridium difficile infection: The patient experience. American Journal of Infection Control. 44, 554-559 (2016).
114 REFERENCES
127. Zellmer, C., De Wolfe, T.J., Van Hoof, S., Blakney, R. & Safdar, N. Patient perspectives on fecal microbiota transplantation for Clostridium difficile infection. Infectious Diseases and Therapy. 5, 155-164(2016).
128. Brandt, L. J. Editorial commentary: fecal microbiota transplantation: patient and physician attitudes. Clinical Infectious Disease. 55, 1659-1660 (2012).
129. Khanna, S. et al. A novel microbiome therapeutic increases gut microbial diversity and prevents recurrent Clostridium difficile infection. The Journal of Infectious Diseases. 214, 173-181 (2016).
130. Khoruts, A. & Sadowsky, M.J. Therapeutic transplantation of the distal gut microbiota. Mucosal Immunology. 4, 4-7 (2011).
131. Millan, B. et al. Fecal microbial transplants reduce antibiotic-resistant genes in patients with recurrent Clostridium difficile infection. Clinical Infectious Diseases. 62, 1479-1486 (2016).
132. Costello, S.P., Conlon, M.A., Vuaran, M.S., Roberts-Thomson, I.C. & Andrews, J.M. Faecal microbiota transplant for recurrent Clostridium difficile infection using long-term frozen stool is effective: clinical efficacy and bacterial viability data. Alimentary Pharmacology & Therapeutics. 42, 1011-1018 (2015).
133. Lee, C. H. et al. Frozen vs fresh fecal microbiota transplantation and clinical resolution of diarrhea in patients with recurrent Clostridium difficile infection: a randomized clinical trial. JAMA 315, 142-149 (2016).
134. Tian, H. et al. Freeze-dried, capsulized fecal microbiota transplantation for relapsing Clostridium difficile infection. Journal of Clinical Gastroenterology. 49, 537-538 (2015).
135. Guo, S. et al. The recombinant Lactococcus lactis oral vaccine induces protection against C. difficile spore challenge in a mouse model. Vaccine. 33, 1586-1595 (2015).
136. Boonma, P., Spinler, J.K., Venable, S.F., Versalovic, J. & Tumwasorn, S. Lactobacillus rhamnosus L34 and Lactobacillus casei L39 suppress Clostridium difficile-induced IL-8 production by colonic epithelial cells. BMC Microbiology. 14, 177 (2014).
137. Tuo, Y. et al.Study of probiotic potential of four wild Lactobacillus rhamnosus strains. Anaerobe. 21, 22-27 (2013).
115 REFERENCES
138. Zhang, K. et al. The non-toxigenic Clostridium difficile CD37 protects mice against infection with a BI/NAP1/027 type of C.difficile strain. Anaerobe. 36, 49-52 (2015).
139. Brouwer, M. S. et al. Horizontal gene transfer converts non-toxigenic Clostridium difficile strains into toxin producers. Nature Communications. 4 (2013).
140. Sheldon, E. et al. A phase 1, placebo-controlled, randomized study of the safety, tolerability, and immunogenicity of a Clostridium difficile vaccine administered with or without aluminum hydroxide in healthy adults. Vaccine. 34, 2082-2091 (2016).
141. Aboudola, S. et al. Clostridium difficile vaccine and serum immunoglobulin G antibody response to toxin A. Infection and Immunity. 71, 1608-1610 (2003).
142. Bezay, N. et al. Safety, immunogenicity and dose response of VLA84, a new vaccine candidate against Clostridium difficile, in healthy volunteers. Vaccine. 34, 2585-2592 (2016).
143. Anosova, N. G. et al. A combination of three fully-human toxin A- and toxin B-specific monoclonal antibodies protects against challenge with highly virulent epidemic strains of C. difficile in the hamster model. Clinical and Vaccine Immunology. 22, 711-725 (2015).
144. de Bruyn, G. et al. Defining the optimal formulation and schedule of a candidate toxoid vaccine against Clostridium difficile infection: A randomized Phase 2 clinical trial. Vaccine. 34, 2170-2178 (2016).
145. Karczewski, J. et al. Development of a recombinant toxin fragment vaccine for Clostridium difficile infection. Vaccine. 32, 2812-2818 (2014).
146. Tian, J.-H. et al. A novel fusion protein containing the receptor binding domains of C. difficile toxin A and toxin B elicits protective immunity against lethal toxin and spore challenge in preclinical efficacy models. Vaccine. 30, 4249-4258 (2012).
147. Orth, P. et al. Mechanism of action and epitopes of Clostridium difficile toxin B-neutralizing antibody bezlotoxumab revealed by X-ray crystallography. Journal of Biological Chemistry. 289, 18008-18021 (2014).
116 REFERENCES
148. Sponseller, J. K. et al. Hyperimmune bovine colostrum as a novel therapy to combat Clostridium difficile infection. Journal of Infectious Diseases. 211, 1334-1341 (2015).
149. Messori, L. & Marcon, G. Gold complexes in the treatment of rheumatoid arthritis. Metal Ions in Biological Systems. 41, 279-304 (2004).
150. Walz, D.T., DiMartino, M.J., Griswold, D.E., Intoccia, A.P. & Flanagan, T.L. Biologic actions and pharmacokinetic studies of auranofin. The American Journal of Medicine. 75, 90-108 (1983).
151. Sannella, A. R. et al. New uses for old drugs. Auranofin, a clinically established antiarthritic metallodrug, exhibits potent antimalarial effects in vitro: Mechanistic and pharmacological implications. FEBS Letters 582, 844-847 (2008).
152. Madeira, J. M. et al. Novel protective properties of auranofin: Inhibition of human astrocyte cytotoxic secretions and direct neuroprotection. Life Sciences. 92, 1072-1080 (2013).
153. Johnson Matthey Fine Chemical. Auranofin. www.jmfinechemicals.com/product/auranofin/ (accessed 17 May 2017).
154. Kean, W.F. & Kean, I.R. Clinical pharmacology of gold. Inflammopharmacology. 16, 112-125 (2008).
155. Kean, W. F., Hart, L. & Buchanan, W. W. Auranofin. Br J Rheumatol 36, 560-572 (1997).
156. Madeira, J. M., Bajwa, E., Stuart, M.J., Hashioka, S. & Klegeris, A. Gold drug auranofin could reduce neuroinflammation by inhibiting microglia cytotoxic secretions and primed respiratory burst. Journal of Neuroimmunology. 276, 71-79 (2014).
157. Messori, L. & Marcon, G. Gold complexes as antitumor agents. Metal Ions in Biological Systems. 42, 385-424 (2004).
158. Fan, C. et al. Enhancement of auranofin-induced lung cancer cell apoptosis by selenocystine, a natural inhibitor of TrxR1 in vitro and in vivo. Cell Death & Disease. 5, e1191 (2014).
159. Fiskus, W. et al. Auranofin induces lethal oxidative and endoplasmic reticulum stress and exerts potent preclinical activity against chronic lymphocytic leukemia. Cancer Research. 74, 2520-2532 (2014).
117 REFERENCES
160. Pessetto, Z. Y., Weir, S. J., Sethi, G., Broward, M. A. & Godwin, A. K. Drug repurposing for gastrointestinal stromal tumor. Molecular Cancer Therapeutics. 12, 1299-1309(2013).
161. Marzano, C. et al. Inhibition of thioredoxin reductase by auranofin induces apoptosis in cisplatin-resistant human ovarian cancer cells. Free Radical Biology & Medicine. 42, 872-881 (2007).
162. Jackson-Rosario, S.E. & Self, W.T. Targeting selenium metabolism and selenoproteins: novel avenues for drug discovery. Metallomics : Integrated Biometal Science. 2, 112-116 (2010).
163. Roder, C. & Thomson, M.J. Auranofin: repurposing an old drug for a golden new age. Drugs in R&D. 15, 13-20 (2015).
164. U.S. National Library of Medicine. Genome. National Centre for Biotechnology Information. http://www.ncbi.nlm.nih.gov/genome (accessed 16 June 2016).
165. Zhang, Y. & Gladyshev, V.N. An algorithm for identification of bacterial selenocysteine insertion sequence elements and selenoprotein genes. Bioinformatics. 21, 2580-2589 (2005).
166. U.S. National Library of Medicine. Basic Local Alignment Search Tool (BLAST). National Centre for Biotechnology Information. https://blast.ncbi.nlm.nih.gov/Blast.cgi (accessed 16 June 2016).
167. Hofacker, I.L. et al. Fast folding and comparison of RNA secondary structures. Monatshefte für Chemie/Chemical Monthly. 125, 167-188 (1994).
168. Wuchty, S., Fontana, W., Hofacker, I.L. & Schuster, P. Complete suboptimal folding of RNA and the stability of secondary structures. Biopolymers. 49, 145-165 (1999).
169. Hofacker, I.L., Fekete, M. & Stadler, P.F. Secondary structure prediction for aligned RNA sequences. Journal of Molecular Biology. 319, 1059-1066 (2002).
170. Wiegand, I., Hilpert, K. & Hancock, R. E. Agar and broth dilution methods to determine the minimal inhibitory concentration (MIC) of antimicrobial substances. Nature protocols. 3, 163-175 (2008).
171. May, T., Mackie, R., Fahey, G., Cremin, J. & Garleb, K. Effect of fiber source on short-chain fatty acid production and on the growth and toxin
118 REFERENCES
production by Clostridium difficile. Scandinavian journal of gastroenterology. 29, 916-922 (1994).
172. Merrigan, M. et al. Human hypervirulent Clostridium difficile strains exhibit increased sporulation as well as robust toxin production. Journal of bacteriology. 192, 4904-4911 (2010).
173. Vohra, P. & Poxton, I. R. Comparison of toxin and spore production in clinically relevant strains of Clostridium difficile. Microbiology (Reading, England). 157, 1343-1353 (2011).
174. Lyon, S. A., Hutton, M. L., Rood, J. I., Cheung, J. K. & Lyras, D. CdtR regulates TcdA and TcdB production in Clostridium difficile. PLoS pathogens. 12, e1005758 (2016).
175. Carter, G. P. et al. Defining the Roles of TcdA and TcdB in localized gastrointestinal disease, systemic organ damage, and the host response during Clostridium difficile infections. mBio. 6, e00551-00515 (2015).
176. Forchhammer, K., Boesmiller, K. & Bock, A. The function of selenocysteine synthase and SELB in the synthesis and incorporation of selenocysteine. Biochimie. 73, 1481-1486 (1991).
177. Allmang, C. & Krol, A. Selenoprotein synthesis: UGA does not end the story. Biochimie. 88, 1561-1571 (2006).
178. Jackson-Rosario, S. & Self, W.T. Inhibition of selenium metabolism in the oral pathogen Treponema denticola. Journal of Bacteriology. 191, 4035- 4040 (2009).
179. Glass, R. S. & Stadtman, T. C. Selenophosphate. Methods in Enzymology. 252, 309-315 (1995).
180. Barker, H.A. Amino acid degradation by anaerobic bacteria. Annual Review of Biochemistry. 50, 23-40 (1981).
181. Peng, T., Lin, J., Xu, Y.Z. & Zhang, Y. Comparative genomics reveals new evolutionary and ecological patterns of selenium utilization in bacteria. The ISME Journal. 10, 2048-2059 (2016).
182. Leinfelder, W., Forchhammer, K., Veprek, B., Zehelein, E. & Bock, A. In vitro synthesis of selenocysteinyl-tRNA(UCA) from seryl-tRNA(UCA): involvement and characterization of the selD gene product. Proceedings of the National Academy of Sciences of the United States of America. 87, 543- 547 (1990).
119 REFERENCES
183. Gursinsky, T. et al. Factors and selenocysteine insertion sequence requirements for the synthesis of selenoproteins from a gram-positive anaerobe in Escherichia coli. Applied and Environmental Microbiology. 74, 1385-1393 (2008).
184. Forchhammer, K. & Bock, A. Selenocysteine synthase from Escherichia coli. Analysis of the reaction sequence. The Journal of Biological Chemistry. 266, 6324-6328 (1991).
185. Jackson, S., Calos, M., Myers, A. & Self, W.T. Analysis of proline reduction in the nosocomial pathogen Clostridium difficile. Journal of Bacteriology. 188, 8487-8495 (2006).
186. Stadtman, T.C. Clostridial glycine reductase: protein C, the acetyl group acceptor, catalyzes the arsenate-dependent decomposition of acetyl phosphate. Proceedings of the National Academy of Sciences of the United States of America. 86, 7853-7856 (1989).
187. Stickland, L., H. Studies in the metabolism of the strict anaerobes (Genus Clostridium). II. The reduction of proline by Cl. Sporogenes. Biochem J. 29, 288-290 (1935 ).
188. Stickland, L.H. Studies in the metabolism of the strict anaerobes (Genus Clostridium). I. The Chemical reactions by which Cl. Sporogenes obtains its energy. Biochem. J. 28, 1746-1759 (1934).
189. Stickland, L.H. Studies in the metabolsim of the strict anaerobes (Genus Clostridium). III. The oxidation of alanine by Cl. sporogenes. IV. The reduction of glycine by Cl. sporogenes. Biochem. J. 30, 889-898 (1935).
190. Stadtman, T.C. in Methods in Enzymology Vol. 53 (eds Fleischer Sidney & Packer Lester) 373-382 (Academic Press, 1978).
191. Kohlstock, U. M. et al. Cys359 of GrdD is the active-site thiol that catalyses the final step of acetyl phosphate formation by glycine reductase from Eubacterium acidaminophilum. European Journal of Biochemistry. 268, 6417-6425 (2001).
192. Sliwkowski, M.X. & Stadtman, T.C. Selenoprotein A of the clostridial glycine reductase complex: purification and amino acid sequence of the selenocysteine-containing peptide. Proceedings of the National Academy of Sciences of the United States of America. 85, 368-371 (1988).
193. Kabisch, U. C. et al. Identification of D-proline reductase from Clostridium sticklandii as a selenoenzyme and indications for a catalytically active
120 REFERENCES
pyruvoyl group derived from a cysteine residue by cleavage of a proprotein. The Journal of Biological Chemistry. 274, 8445-8454 (1999).
194. Stadtman, T.C. Biosynthesis and function of selenocysteine-containing enzymes. The Journal of Biological Chemistry. 266, 16257-16260 (1991).
195. Bouillaut, L., Self, W.T. & Sonenshein, A.L. Proline-dependent regulation of Clostridium difficile Stickland metabolism. Journal of Bacteriology. 195, 844-854 (2013).
196. Stickland, L.H. Studies in the metabolism of the strict anaerobes (Genus Clostridium). II. The reduction of proline by Cl. Sporogenes. Biochem J. 29, 288-290 (1935 ).
197. Fourmy, D., Guittet, E. & Yoshizawa, S. Structure of prokaryotic SECIS mRNA hairpin and its interaction with elongation factor SelB. Journal of Molecular Biology. 324, 137-150 (2002).
198. Sliwkowski, M.X. & Stadtman, T.C. Purification and immunological studies of selenoprotein A of the clostridial glycine reductase complex. The Journal of Biological Chemistry. 262, 4899-4904 (1987).
199. Liu, Z., Reches, M., Groisman, I. & Engelberg-Kulka, H. The nature of the minimal 'selenocysteine insertion sequence' (SECIS) in Escherichia coli. Nucleic Acids Research. 26, 896-902 (1998).
200. Sandman, K.E., Tardiff, D.F., Neely, L.A. & Noren, C.J. Revised Escherichia coli selenocysteine insertion requirements determined by in vivo screening of combinatorial libraries of SECIS variants. Nucleic Acids Research. 31, 2234-2241 (2003).
201. Tatusov, R.L., Koonin, E.V. & Lipman, D.J. A genomic perspective on protein families. Science. 278, 631-637 (1997).
202. Brown, E.D. & Wright, G.D. New targets and screening approaches in antimicrobial drug discovery. Chem Rev. 105, 759-774 (2005).
203. Andreesen, J.R. Glycine reductase mechanism. Curr Opin Chem Biol. 8, 454-461 (2004).
204. Kreimer, S. & Andreesen, J.R. Glycine reductase of Clostridium litorale. Cloning, sequencing, and molecular analysis of the grdAB operon that contains two in-frame TGA codons for selenium incorporation. European Journal of Biochemistry. 234, 192-199 (1995).
121 REFERENCES
205. Lubbers, M. & Andreesen, J.R. Components of glycine reductase from Eubacterium acidaminophilum. Cloning, sequencing and identification of the genes for thioredoxin reductase, thioredoxin and selenoprotein PA. European Journal of Biochemistry. 217, 791-798 (1993).
206. Stadtman, T.C. & Davis, J.N. Glycine reductase protein C. Properties and characterization of its role in the reductive cleavage of Se-carboxymethyl- selenoprotein A. The Journal of Biological Chemistry. 266, 22147-22153 (1991).
207. Stadtman, T.C. & Elliott, P. Studies on the enzymic reduction of amino acids. II. Purification and properties of D-proline reductase and a proline racemase from Clostridium sticklandii. The Journal of Biological Chemistry. 228, 983-997 (1957).
208. Talbot, S., Nelson, R. & Self, W.T. Arsenic trioxide and auranofin inhibit selenoprotein synthesis: implications for chemotherapy for acute promyelocytic leukaemia. British Journal of Pharmacology. 154, 940-948 (2008).
209. The European Committee on Antimicrobial Susceptibility Testing (EUCAST). Breakpoint tables for interpretation of MICs and zone diameters. Version 7.1, 2017. http://www.eucast.org (accessed 12 May 2017)
210. Zar, F.A., Bakkanagari, S.R., Moorthi, K. & Davis, M.B. A comparison of vancomycin and metronidazole for the treatment of Clostridium difficile– associated diarrhea, stratified by disease severity. Clinical Infectious Diseases. 45, 302-307 (2007).
211. Zhang, Y., Romero, H., Salinas, G. & Gladyshev, V.N. Dynamic evolution of selenocysteine utilization in bacteria: a balance between selenoprotein loss and evolution of selenocysteine from redox active cysteine residues. Genome biology. 7, R94 (2006).
212. Scaria, J. et al. Analysis of ultra low genome conservation in Clostridium difficile. PloS One. 5, e15147 (2010).
213. Fraise, A. P. et al. Development of a sporicidal test method for Clostridium difficile. The Journal of Hospital Infection. 89, 2-15 (2014).
214. Liu, Q. & Jiang, L. in Selenoproteins and Mimics 125-140 (Springer, 2011).
122 REFERENCES
215. Thangamani, S. et al. Antibacterial activity and mechanism of action of auranofin against multi-drug resistant bacterial pathogens. Scientific reports. 6 (2016).
216. Harbut, M. B. et al. Auranofin exerts broad-spectrum bactericidal activities by targeting thiol-redox homeostasis. Proceedings of the National Academy of Sciences. 112, 4453-4458 (2015).
123 APPENDIX 1
7. Appendix 1 Clostridium difficile genomes from NCBI used for analysis of gene conservation.
Genome NCBI Accession Clostridium difficile E28 genomic scaffold, 1851_171, whole genome NZ_HF923795.1 shotgun sequence Clostridium difficile E7 genomic scaffold, 1839_235, whole genome NZ_HF929526.1 shotgun sequence Clostridium difficile E9 genomic scaffold, 1845_204, whole genome NZ_HF929904.1 shotgun sequence Clostridium difficile T17 genomic scaffold, 1843_295, whole genome NZ_HF925065.1 shotgun sequence Clostridium difficile E14 genomic scaffold, 1848_175, whole genome NZ_HF923480.1 shotgun sequence Clostridium difficile T6 genomic scaffold, 1856_217, whole genome NZ_HF922885.1 shotgun sequence Clostridium difficile T15 genomic scaffold, 1842_357, whole genome NZ_HF928152.1 shotgun sequence Clostridium difficile E25 genomic scaffold, 1838_260, whole genome NZ_HF929039.1 shotgun sequence Clostridium difficile T61 genomic scaffold, 2338_59, whole genome NZ_HF930132.1 shotgun sequence Clostridium difficile T19 genomic scaffold, 1853_160, whole genome NZ_HF924058.1 shotgun sequence Clostridium difficile T3 genomic scaffold, 1859_180, whole genome NZ_HF926414.1 shotgun sequence Clostridium difficile T42 genomic scaffold, 2340_70, whole genome NZ_HF933940.1 shotgun sequence Clostridium difficile E24 genomic scaffold, 2339_46, whole genome NZ_HF927455.1 shotgun sequence Clostridium difficile E19 genomic scaffold, 1849_210, whole genome NZ_HF930365.1 shotgun sequence Clostridium difficile T22 genomic scaffold, 1855_190, whole genome NZ_HF924628.1 shotgun sequence Clostridium difficile E16 genomic scaffold, 2337_72, whole genome NZ_HF927756.1 shotgun sequence Clostridium difficile CD002 genomic scaffold, 2335_51, whole genome NZ_HF929271.1 shotgun sequence Clostridium difficile E13 genomic scaffold, 1847_143, whole genome NZ_HF925362.1 shotgun sequence Peptoclostridium difficile 840 gcd840.contig.118, whole genome shotgun NZ_AVIN01000119 sequence .1 Peptoclostridium difficile CD39 gcdCD39.contig.119, whole genome NZ_AVGS0100012 shotgun sequence 0.1 NZ_FAXM0100001 Clostridioides difficile isolate IT1118, whole genome shotgun sequence 2.1 Peptoclostridium difficile P7 gcdP7.contig.161, whole genome shotgun NZ_AVLQ0100015 sequence 6.1 Peptoclostridium difficile Y202 gcdY202.contig.217, whole genome NZ_AVKY0100021 shotgun sequence 6.1 Peptoclostridium difficile Y165 gcdY165.contig.278, whole genome NZ_AVKV0100026 shotgun sequence 7.1
124 APPENDIX 1
Peptoclostridium difficile DA00193 gcdDA00193.contig.49, whole genome NZ_AVJO01000049 shotgun sequence .1 Peptoclostridium difficile DA00149 gcdDA00149.contig.68, whole genome NZ_AVJF01000068 shotgun sequence .1 Peptoclostridium difficile DA00129 gcdDA00129.contig.54, whole genome NZ_AVIY01000054 shotgun sequence .1 Peptoclostridium difficile 842 gcd842.contig.115, whole genome shotgun NZ_AVIO01000116 sequence .1 Peptoclostridium difficile CD200 gcdCD200.contig.179, whole genome NZ_AVIF01000164 shotgun sequence .1 Peptoclostridium difficile CD175 gcdCD175.contig.457, whole genome NZ_AVIB01000444 shotgun sequence .1 Peptoclostridium difficile CD22 gcdCD22.contig.121, whole genome NZ_AVGP0100012 shotgun sequence 2.1 Peptoclostridium difficile CD3 gcdCD3.contig.229, whole genome shotgun NZ_AVGI01000222 sequence .1 Clostridioides difficile strain LIBA-5719 .11940_2_38.2, whole genome NZ_JXBQ01000002 shotgun sequence .1 Clostridioides difficile strain LIBA-5701 .11940_2_22.1, whole genome NZ_JXBP01000001 shotgun sequence .1 Clostridioides difficile strain LIBA-5704 contig.11940_2_24.1, whole NZ_JXCP01000001 genome shotgun sequence .1 Clostridioides difficile strain LIBA-5734 .11940_2_42.2, whole genome NZ_JXBR01000002 shotgun sequence .1 NZ_CCEV0100007 Clostridioides difficile strain F17xCD13A, whole genome shotgun sequence 8.1 Clostridioides difficile strain CII7xCD13A, whole genome shotgun NZ_CCDW0100000 sequence 1.1 Clostridioides difficile strain LIBA-5784 .11940_2_72.3, whole genome NZ_JXBS01000003 shotgun sequence .1 Peptoclostridium difficile P53 gcdP53.contig.30, whole genome shotgun NZ_AWZO0100003 sequence 1.1 Peptoclostridium difficile F665 gcdF665.contig.35, whole genome shotgun NZ_AWZN0100003 sequence 6.1 Peptoclostridium difficile P68 gcdP68.contig.41, whole genome shotgun NZ_AWZM010000 sequence 41.1 Peptoclostridium difficile P64 gcdP64.contig.40, whole genome shotgun NZ_AWZL0100004 sequence 1.1 Peptoclostridium difficile P37 gcdP37.contig.43, whole genome shotgun NZ_AWZK0100004 sequence 4.1 Peptoclostridium difficile P33 gcdP33.contig.34, whole genome shotgun NZ_AWZI0100003 sequence 5.1 Peptoclostridium difficile DA00130 gcdDA00130.contig.93, whole genome NZ_AWZH0100009 shotgun sequence 3.1 Peptoclostridium difficile F601 gcdF601.contig.24, whole genome shotgun NZ_AVNH0100002 sequence 5.1 Peptoclostridium difficile P77 gcdP77.contig.38, whole genome shotgun NZ_AVNA0100003 sequence 9.1 Peptoclostridium difficile P75 gcdP75.contig.36, whole genome shotgun NZ_AVMZ0100003 sequence 7.1 Peptoclostridium difficile P74 gcdP74.contig.25, whole genome shotgun NZ_AVMY0100002 sequence 6.1 Peptoclostridium difficile P73 gcdP73.contig.33, whole genome shotgun NZ_AVMX0100003 sequence 4.1 Peptoclostridium difficile P71 gcdP71.contig.39, whole genome shotgun NZ_AVMW010000 sequence 40.1
125 APPENDIX 1
Peptoclostridium difficile P72 gcdP72.contig.30, whole genome shotgun NZ_AVMV0100003 sequence 1.1 Peptoclostridium difficile P70 gcdP70.contig.34, whole genome shotgun NZ_AVMU0100003 sequence 5.1 Peptoclostridium difficile P69 gcdP69.contig.32, whole genome shotgun NZ_AVMT0100003 sequence 3.1 Peptoclostridium difficile P78 gcdP78.contig.41, whole genome shotgun NZ_AVMS0100004 sequence 2.1 Peptoclostridium difficile P61 gcdP61.contig.39, whole genome shotgun NZ_AVMR0100004 sequence 0.1 Peptoclostridium difficile P59 gcdP59.contig.39, whole genome shotgun NZ_AVMQ0100004 sequence 0.1 Peptoclostridium difficile P49 gcdP49.contig.28, whole genome shotgun NZ_AVMN0100002 sequence 9.1 Peptoclostridium difficile P48 gcdP48.contig.41, whole genome shotgun NZ_AVMM010000 sequence 42.1 Peptoclostridium difficile P46 gcdP46.contig.41, whole genome shotgun NZ_AVML0100004 sequence 1.1 Peptoclostridium difficile P45 gcdP45.contig.33, whole genome shotgun NZ_AVMK0100003 sequence 4.1 Peptoclostridium difficile P42 gcdP42.contig.37, whole genome shotgun NZ_AVMJ0100003 sequence 8.1 Peptoclostridium difficile P32 gcdP32.contig.30, whole genome shotgun NZ_AVMG0100003 sequence 0.1 Peptoclostridium difficile P31 gcdP31.contig.34, whole genome shotgun NZ_AVMF0100003 sequence 5.1 Peptoclostridium difficile P30 gcdP30.contig.34, whole genome shotgun NZ_AVME0100003 sequence 5.1 Peptoclostridium difficile P29 gcdP29.contig.40, whole genome shotgun NZ_AVMD0100004 sequence 1.1 Peptoclostridium difficile P24 gcdP24.contig.44, whole genome shotgun NZ_AVMA0100004 sequence 4.1 Peptoclostridium difficile P23 gcdP23.contig.29, whole genome shotgun NZ_AVLZ0100003 sequence 0.1 Peptoclostridium difficile P21 gcdP21.contig.38, whole genome shotgun NZ_AVLY0100003 sequence 9.1 Peptoclostridium difficile P20 gcdP20.contig.41, whole genome shotgun NZ_AVLX0100004 sequence 2.1 Peptoclostridium difficile P19 gcdP19.contig.40, whole genome shotgun NZ_AVLW0100004 sequence 1.1 Peptoclostridium difficile P13 gcdP13.contig.34, whole genome shotgun NZ_AVLU0100003 sequence 4.1 Peptoclostridium difficile P9 gcdP9.contig.107, whole genome shotgun NZ_AVLS0100010 sequence 7.1 Peptoclostridium difficile P8 gcdP8.contig.138, whole genome shotgun NZ_AVLR0100013 sequence 1.1 Peptoclostridium difficile P5 gcdP5.contig.131, whole genome shotgun NZ_AVLO0100013 sequence 1.1 Peptoclostridium difficile P3 gcdP3.contig.158, whole genome shotgun NZ_AVLN0100012 sequence 7.1 Peptoclostridium difficile P2 gcdP2.contig.143, whole genome shotgun NZ_AVLM0100014 sequence 4.1 Peptoclostridium difficile Y401 gcdY401.contig.137, whole genome NZ_AVLK0100013 shotgun sequence 8.1 Peptoclostridium difficile Y381 gcdY381.contig.43, whole genome shotgun NZ_AVLI01000044 sequence .1
126 APPENDIX 1
Peptoclostridium difficile Y358 gcdY358.contig.63, whole genome shotgun NZ_AVLH0100006 sequence 4.1 Peptoclostridium difficile Y266 gcdY266.contig.132, whole genome NZ_AVLC0100013 shotgun sequence 3.1 Peptoclostridium difficile Y247 gcdY247.contig.25, whole genome shotgun NZ_AVLB0100002 sequence 6.1 Peptoclostridium difficile Y231 gcdY231.contig.31, whole genome shotgun NZ_AVLA0100003 sequence 2.1 Peptoclostridium difficile Y21 gcdY21.contig.139, whole genome shotgun NZ_AVKS0100014 sequence 0.1 Peptoclostridium difficile F253 gcdF253.contig.323, whole genome shotgun NZ_AVKO0100030 sequence 9.1 Peptoclostridium difficile F249 gcdF249.contig.80, whole genome shotgun NZ_AVKN0100008 sequence 1.1 Peptoclostridium difficile F152 gcdF152.contig.13, whole genome shotgun NZ_AVKM0100001 sequence 3.1 Peptoclostridium difficile DA00313 gcdDA00313.contig.82, whole genome NZ_AVKL0100008 shotgun sequence 2.1 Peptoclostridium difficile genome assembly CD630DERM, chromosome : 1 NZ_LN614756.1 Peptoclostridium difficile 630, complete genome NZ_CP010905.1 NZ_CDND0100000 Clostridioides difficile strain G46, whole genome shotgun sequence 1.1 Peptoclostridium difficile P25 gcdP25.contig.17, whole genome shotgun NZ_AVMB0100001 sequence 8.1 Peptoclostridium difficile DA00261 gcdDA00261.contig.11, whole genome NZ_AVKE0100001 shotgun sequence 2.1 Peptoclostridium difficile DA00244 gcdDA00244.contig.9, whole genome NZ_AVKA0100001 shotgun sequence 0.1 Peptoclostridium difficile DA00203 gcdDA00203.contig.10, whole genome NZ_AVJS01000011 shotgun sequence .1 Peptoclostridium difficile strain 08ACD0030, complete genome NZ_CP010888.1 Peptoclostridium difficile genome assembly NCTC13307, chromosome : 1 NZ_LN831030.1 Peptoclostridium difficile CD26A54_S contig6, whole genome shotgun NZ_AMDL0100000 sequence 6.1 Peptoclostridium difficile CD26A54_R contig4, whole genome shotgun NZ_AMDM010000
sequence 04.1
127 APPENDIX 2
8. Appendix 2 SECIS elements from Clostridium difficile 630 genome identified using in silico methods.
Selenoprotein 30S Ribosomal protein S1 name (RpsA) Level of 100% conservation Translation, ribosomal Functional class structure and biogenesis
SECIS length 54 nt
Length UGA to 32 nt apical loop Upper stem 8 bp Upper stem U 50% richness
Apical loop 3 nt Selenoprotein 4-aminobutyrate name aminotransferase (GabT) Level of 100% conservation Amino acid transport and Functional class metabolism SECIS length 45 nt Length UGA to 25 nt apical loop Upper stem 6 bp Upper stem U 33.33% richness Apical loop 6 nt Selenoprotein ABC transporter ATP- name binding protein (LolD)
Level of 100% conservation Signal transduction Functional class mechanisms SECIS length 39 nt Length UGA to 19 nt apical loop Upper stem 4 bp Upper stem U 66% richness
Apical loop 5 nt
128 APPENDIX 2
ABC-transporte sugar- Selenoprotein family ATP-binding name protein (SalY) Level of 100% conservation Signal transduction Functional class mechanisms
SECIS length 63 nt
Length UGA to 23 nt apical loop
Upper stem 5bp
Upper stem U 660% richness Apical loop 3 nt
Selenoprotein Chromosomal replication name initiation protein Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 52 nt
Length UGA to 17 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness Apical loop 5 nt Selenoprotein Acetolactate synthase large name subunit (IlvB) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 42 nt
Length UGA to 22 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 3 nt
129 APPENDIX 2
Selenoprotein Alpha-mannosidase name Level of 100% conservation Carbohydrate transport and Functional class metabolism
SECIS length 51 nt
Length UGA to 18 nt apical loop
Upper stem 5 bp
Upper stem U 20% richness Apical loop 4 nt Selenoprotein Aminotransferase name Level of 100% conservation Amino acid transport and Functional class metabolism
SECIS length 63nt
Length UGA to 31 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 4 nt Selenoprotein Asparagine synthetase name (AsnB) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 39 nt
Length UGA to 19 nt apical loop
Upper stem 4 bp
Upper stem U 40% richness
Apical loop 3 nt
130 APPENDIX 2
Selenoprotein Aspartate aminotransferase name (AspB) Level of 100% conservation Amino acid transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 26 nt apical loop
Upper stem 5 bp
Upper stem U 66.66% richness Apical loop 7 nt
Selenoprotein Diguanylate kinase name signalling protein Level of 100% conservation
Functional class Cell motility
SECIS length 45 nt
Length UGA to 22 nt apical loop
Upper stem 4 bp
Upper stem U 33.33% richness
Apical loop 6 nt Selenoprotein DinG family helicase name (DinG) Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 45 nt
Length UGA to 18 nt apical loop
Upper stem 6 nt
Upper stem U 28.57% richness
Apical loop 6 nt
131 APPENDIX 2
Selenoprotein DNA mismatch repair name protein (MutL) Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 55 nt
Length UGA to 24 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 4 nt
Selenoprotein DNA primase (DnaG) name Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 50 nt
Length UGA to 28 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 4 nt Selenoprotein D-proline reductase (PrdA) name Level of 100% conservation Amino acid transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 20 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 8 nt
132 APPENDIX 2
Selenoprotein Electron transport complex name protein (RnfC) Level of 100% conservation Energy production and Functional class conversion
SECIS length 42 nt
Length UGA to 21 nt apical loop
Upper stem 4 bp
Upper stem U 25% richness
Apical loop 6 nt
Selenoprotein Glutaminase (GlsA) name Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 42 nt
Length UGA to 18 nt apical loop
Upper stem 5 bp
Upper stem U 60% richness
Apical loop 5 nt Selenoprotein GTP-binding protein name (LepA) Level of 100% conservation
Functional class Translation
SECIS length 45 nt
Length UGA to 19 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 6 nt
133 APPENDIX 2
Selenoprotein HAD superfamily name hydrolase Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 66 nt
Length UGA to 30 nt apical loop
Upper stem 7 bp
Upper stem U 57.14% richness
Apical loop 3 nt
Selenoprotein Minor head protein name Level of 100% conservation General function Functional class prediction only
SECIS length 50 nt
Length UGA to 21 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness
Apical loop 6 nt Selenoprotein Iron-sulfur binding protein name Level of 100% conservation Cell Functional class wall/membrane/envelope biogenesis
SECIS length 39 nt
Length UGA to 19 nt apical loop
Upper stem 6 bp
Upper stem U 28.57% richness
Apical loop 4 nt
134 APPENDIX 2
Selenoprotein Isoleucine-tRNA ligase name (IleS) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 54 nt
Length UGA to 30 nt apical loop
Upper stem 5 bp
Upper stem U 20% richness
Apical loop 4 nt K/Mg/Cd/Cu/Zn/Na/Ca/Na Selenoprotein /H-transporting P-type name ATPase (ZntA) Level of 100% conservation Inorganic ion transport and Functional class metabolism SECIS length 39 nt Length UGA to 22 nt apical loop
Upper stem 9 bp
Upper stem U 33.33% richness Apical loop 8 nt
Selenoprotein Lipoprotein signal name peptidase (LspA) Level of 100% conservation Signal transduction Functional class mechanisms
SECIS length 48 nt
Length UGA to 21 nt apical loop
Upper stem 9 bp
Upper stem U 30% richness
Apical loop 3 nt
135 APPENDIX 2
LysR family Selenoprotein transcriptional regulator name (LysR) Level of 100% conservation
Functional class Transcription
SECIS length 51 nt
Length UGA to 30 nt apical loop
Upper stem 4 bp
Upper stem U 16.66% richness
Apical loop 5 nt
Selenoprotein M24 family peptidase name (PepP) Level of 100% conservation Amino acid transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 28 nt apical loop
Upper stem 6 bp
Upper stem U 33.33% richness
Apical loop 5 nt Selenoprotein MATE family drug/sodium name antiporter (NorM) Level of 100% conservation
Functional class Defence mechanisms
SECIS length 48 nt
Length UGA to 30 nt apical loop
Upper stem 8 bp
Upper stem U 66.66% richness
Apical loop 3 nt
136 APPENDIX 2
Selenoprotein Putative membrane protein name Level of 100% conservation General function Functional class prediction only
SECIS length 54 nt
Length UGA to 23 nt apical loop
Upper stem 6 bp
Upper stem U 33.33% richness
Apical loop 5 nt
Selenoprotein Methionine gamma-lysase name (MdeA) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 66 nt
Length UGA to 27 nt apical loop
Upper stem 6 bp
Upper stem U 83.33% richness
Apical loop 10 nt Selenoprotein Putative methyltransferase name Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 57 nt
Length UGA to 29 nt apical loop
Upper stem 7 bp
Upper stem U 42.86% richness
Apical loop 8 nt
137 APPENDIX 2
Multidrug family ABC transporter/ Selenoprotein ATP-binding protein/permease name (MdlB) Level of 100% conservation
Functional class Defence mechanisms
SECIS length 63 nt
Length UGA to 30 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness Apical loop 5 nt Selenoprotein NhaC family Na(+)/H(+) name antiporter Level of 100% conservation Inorganic ion transport and Functional class metabolism
SECIS length 39 nt
Length UGA to 18 nt apical loop
Upper stem 5 bp
Upper stem U 80% richness
Apical loop 4 nt Selenoprotein Bifunctional name nitrilase/nitrile hydralatase Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 39 nt
Length UGA to 22 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness
Apical loop 4 nt
138 APPENDIX 2
Selenoprotein Prolyl-tRNA ligase name (ProS2) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 39 nt
Length UGA to 20 nt apical loop
Upper stem 4 bp
Upper stem U 20% richness
Apical loop 4 nt
PTS system Selenoprotein lactose/cellobiose-family name IIC component (CelB) Level of 100% conservation Carbohydrate transport and Functional class metabolism
SECIS length 51 nt
Length UGA to 25 nt apical loop
Upper stem 7 bp
Upper stem U 28.51% richness Apical loop 3 nt PTS system, Selenoprotein fructose/mannitol-family name IIC component Level of 100% conservation Carbohydrate transport and Functional class metabolism
SECIS length 57 nt
Length UGA to 29 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness
Apical loop 8 nt
139 APPENDIX 2
Selenoprotein Oligopeptide transporter name (OPT) Level of 100% conservation Amino acid transport and Functional class metabolism
SECIS length 56 nt
Length UGA to 18 nt apical loop Upper stem 4 bp Upper stem U 25% richness
Apical loop 4 nt
Selenoprotein Peptidase, M20D family name (AgbB) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 41 nt
Length UGA to 19 nt apical loop
Upper stem 6 bp
Upper stem U 33.33% richness
Apical loop 6 nt Selenoprotein Peptidase, S9 family name (DAP2) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 39 nt
Length UGA to 18 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 4 nt
140 APPENDIX 2
Pyruvate phosphate Selenoprotein dikinase, PEP/pyruvate- name binding Level of 100% conservation Carbohydrate transport and Functional class metabolism
SECIS length 42 nt
Length UGA to 19 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness
Apical loop 4 nt
Selenoprotein Signalling protein name Level of 100% conservation Signal transduction Functional class mechanisms
SECIS length 42 nt
Length UGA to 21 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness Apical loop 3 nt Selenoprotein Sodium:solute symporter name Level of 100% conservation Inorganic ion transport and Functional class metabolism
SECIS length 54 nt
Length UGA to 33 nt apical loop
Upper stem 4 bp
Upper stem U 60% richness
Apical loop 4 nt
141 APPENDIX 2
Selenoprotein Spore cortex-lytic name hydrolase (SleB) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 54 nt
Length UGA to 20 nt apical loop
Upper stem 6 bp
Upper stem U 66.66% richness
Apical loop 8 nt
Selenoprotein Thioredoxin-disulfide name reductase (TrxB1) Level of 100% conservation Post-translational Functional class modification, protein turnover, and chaperones
SECIS length 39 nt
Length UGA to 19 nt apical loop
Upper stem 4 bp
Upper stem U 20% richness
Apical loop 6 nt
Selenoprotein Thiosulfate name sulfurtransferase (SseA) Level of 100% conservation Inorganic ion transport and Functional class metabolism
SECIS length 42 nt
Length UGA to 19 nt apical loop
Upper stem 7 bp
Upper stem U 71.43% richness
Apical loop 8 nt
142 APPENDIX 2
Selenoprotein Type II secretion system name protein (PulF) Level of 100% conservation Intracellular trafficking, Functional class secretion and vesicular transport
SECIS length 51 nt
Length UGA to 18 nt apical loop
Upper stem 5 bp
Upper stem U 20% richness
Apical loop 4 nt
Selenoprotein UDP-glycosyltransferase, name MGT subfamily Level of 100% conservation Carbohydrate transport and Functional class metabolism
SECIS length 57 nt
Length UGA to 28 nt apical loop
Upper stem 9 bp
Upper stem U 44.44% richness Apical loop 5 nt
Selenoprotein UvrABC system protein A name Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 54 nt
Length UGA to 17 nt apical loop
Upper stem 7 bp
Upper stem U 71.43% richness
Apical loop 7 nt
143 APPENDIX 2
Selenoprotein Queuine tRNA- name ribosyltransferase (Tgt) Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 48 nt
Length UGA to 23 nt apical loop
Upper stem 6 bp
Upper stem U 57.14% richness
Apical loop 4 nt
Selenoprotein Selenide water dikinase name (SelD) Level of 100% conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 45 nt
Length UGA to 21 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 3 nt
Selenoprotein Sodium/glutamate name symporter (GltC) Level of 100% conservation Amino acid transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 20 nt apical loop
Upper stem 5 bp
Upper stem U 60% richness
Apical loop 6 nt
144 APPENDIX 2
Selenoprotein Spore maturation protein B name (SpmB) Level of 100% conservation Cell Functional class wall/membrane/envelope biogenesis
SECIS length 54 nt
Length UGA to 26 nt apical loop
Upper stem 4 bp
Upper stem U 40% richness Apical loop 4 nt Selenoprotein Threonine synthase (ThrC) name Level of 100% conservation Secondary metabolites Functional class biosynthesis, transport and catabolism
SECIS length 51 nt
Length UGA to 18 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness
Apical loop 6 nt Transcription Selenoprotein antiterminator, PTS operon name regulator Level of 100% conservation
Functional class Transcription
SECIS length 51 nt Length UGA to 26 nt apical loop Upper stem 5 bp Upper stem U 42.86% richness
Apical loop 3 nt
145 APPENDIX 2
Selenoprotein C4-dicarboxylate name anaerobic carrier (YfcC) Level of 100% conservation Inorganic ion transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 18 nt apical loop
Upper stem 10 bp with bulged U+30
Upper stem U 83.33% richness
Apical loop 8 nt
Selenoprotein Exonuclease ABC subunit name A (UvrA) Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 54 nt
Length UGA to 23 nt apical loop
Upper stem 3 bp
Upper stem U 37.50% richness
Apical loop 10 nt Selenoprotein Glycine reductase complex name component A (GrdA) Level of 100% conservation Amino acid transport and Functional class metabolism: Energy production and conversion
SECIS length 45 nt
Length UGA to 16 nt apical loop
Upper stem 3 bp
Upper stem U 50% richness
Apical loop 6 nt
146 APPENDIX 2
Selenoprotein Multidrug family ABC name transporter permease Level of 100% conservation
Functional class Defence mechanisms
SECIS length 45 nt
Length UGA to 20 nt apical loop
Upper stem 10 bp with bulged U+18
Upper stem U 50% richness
Apical loop 8 nt
Selenoprotein 23s ribosomal RNA (23s- name rRNA) Level of 100% conservation Translation, ribosomal Functional class structure and biogenesis
SECIS length 45 nt
Length UGA to 21 nt apical loop
Upper stem 3 bp
Upper stem U 25% richness
Apical loop 6 nt Selenoprotein Two-component sensor name histidine kinase (OmpR) Level of 100% conservation Signal transduction Functional class mechanisms
SECIS length 51 nt
Length UGA to 26 nt apical loop
Upper stem 8 bp with bulged C+37
Upper stem U 50% richness
147 APPENDIX 2
Apical loop 7 nt
Selenoprotein Two-component response name regulator Level of 100% conservation
Functional class Transcription
SECIS length 51 nt
Length UGA to 23 nt apical loop
Upper stem 11 bp with bulged U+37
Upper stem U 83.33% richness
Apical loop 7 nt
Selenoprotein Argininosuccinate synthase name Level of Very high conservation Secondary metabolites Functional class biosynthesis, transport and catabolism
SECIS length 48 nt
Length UGA to 25 nt apical loop
Upper stem 5 bp
Upper stem U 20% richness Apical loop 6 nt
148 APPENDIX 2
Selenoprotein Site-specific recombinase name Level of Very high conservation Replication, recombination Functional class and repair
SECIS length 47 nt
Length UGA to 28 nt apical loop
Upper stem 4 bp
Upper stem U 25% richness
Apical loop 5 nt
Selenoprotein ABC transporter permease name Level of Very high conservation Inorganic ion transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 21 nt apical loop
Upper stem 7 bp
Upper stem U 28.57% richness
Apical loop 3 nt Selenoprotein Membrane protein name Level of Very high conservation General function Functional class prediction only
SECIS length 39 nt
Length UGA to 22 nt apical loop
Upper stem 5 bp
Upper stem U 60% richness
Apical loop 5 nt
149 APPENDIX 2
Selenoprotein Transcriptional regulator, name Xylose repressor (XylR) Level of Very high conservation
Functional class Transcription
SECIS length 51 nt
Length UGA to 16 nt apical loop
Upper stem 3 bp
Upper stem U 33.33% richness
Apical loop 10 nt Selenoprotein Xylose kinase (XylB) name Level of Very high conservation Carbohydrate transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 25 nt apical loop
Upper stem 12 bp with bulged C+35
Upper stem U 41.66% richness
Apical loop 7 nt Glycine reductase complex Selenoprotein component B subunit name gamma Level of High conservation Amino acid transport and Functional class metabolism; Energy production and conversion
SECIS length 39 nt
Length UGA to 18 nt apical loop
Upper stem 7 bp
Upper stem U 42.86% richness Apical loop 3 nt
150 APPENDIX 2
Selenoprotein Phage tail fiber protein name Level of High conservation General function Functional class prediction only
SECIS length 56 nt
Length UGA to 26 nt apical loop
Upper stem 7 bp
Upper stem U 42.86% richness
Apical loop 3 nt
Selenoprotein Amidohydrolase name Level of High conservation Secondary metabolites Functional class biosynthesis, transport and catabolism
SECIS length 66 nt
Length UGA to 36 nt apical loop
Upper stem 4 bp
Upper stem U 25% richness Apical loop 4 nt Selenoprotein Transcriptional regulator, name TetR family (TetR_N) Level of High conservation
Functional class Transcription
SECIS length 48 nt
Length UGA to 23 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness
Apical loop 3 nt
151 APPENDIX 2
Selenoprotein Putative phage DNA- name binding protein* (YjcR) Level of High conservation General function Functional class prediction only
SECIS length 39 nt
Length UGA to 18 nt apical loop
Upper stem 7 bp
Upper stem U 14.29% richness
Apical loop 5 nt Replication initiation Selenoprotein factor Tn916-like, CTn7- name Orf27 Level of High conservation Replication, recombination Functional class and repair
SECIS length 56 nt
Length UGA to 26 nt apical loop
Upper stem 7 bp
Upper stem U 66.66% richness Apical loop 3 nt Selenoprotein Pyrroline-5-carboxylate name reductase (ProC2) Level of High conservation Amino acid transport and Functional class metabolism
SECIS length 45 nt
Length UGA to 19 nt apical loop
Upper stem 9 bp with bulged G28
Upper stem U 37.5% richness Apical loop 4 nt
152 APPENDIX 2
Selenoprotein Conjugative transposon name protein (TpcC) Level of High conservation General function Functional class prediction only
SECIS length 45 nt
Length UGA to 24 nt apical loop
Upper stem 5 bp
Upper stem U 40% richness
Apical loop 3 nt PTS system lichenan- Selenoprotein specific IIC component name (LicC) Level of Moderate-high conservation Carbohydrate transport and Functional class metabolism
SECIS length 42 nt
Length UGA to 17 nt apical loop
Upper stem 4 bp
Upper stem U 50% richness
Apical loop 4 nt Selenoprotein ATPase name Level of Moderate-high conservation Inorganic ion transport and Functional class metabolism
SECIS length 42 nt
Length UGA to 22 nt apical loop
Upper stem 7 bp
Upper stem U 14.29% richness
Apical loop 6 nt
153 APPENDIX 2
Selenoprotein Cell surface protein name Level of Moderate-high conservation
Functional class Extracellular structures
SECIS length 45 nt
Length UGA to 29 nt apical loop
Upper stem 11 bp with bulged A+26
Upper stem U 36.36% richness
Apical loop 4 nt Cobalt-specific ABC Selenoprotein transporter permease name (EcfT) Level of Moderate conservation Inorganic ion transport and Functional class metabolism
SECIS length 51 nt
Length UGA to 18 nt apical loop
Upper stem 10 bp with bulged C16
Upper stem U 40% richness
Apical loop 7 nt
Selenoprotein DNA/RNA helicase name Level of Moderate conservation Replication, recombination Functional class and repair
SECIS length 48 nt
Length UGA to 20 nt apical loop
Upper stem 5 bp
Upper stem U 60% richness
Apical loop 7 nt
154 APPENDIX 2
Selenoprotein Putative hydrolase name Level of Moderate-low conservation Secondary metabolites biosynthesis, Functional class transport and catabolism
SECIS length 48 nt
Length UGA to 30 nt apical loop
Upper stem 4 bp
Upper stem U 25% richness
Apical loop 5 nt
Selenoprotein Conjugative transposon name protein (Spr) Level of Moderate-low conservation General function Functional class prediction only
SECIS length 54 nt
Length UGA to 21 nt apical loop
Upper stem 6 bp
Upper stem U 16.66% richness
Apical loop 4 nt Selenoprotein Conjugative transposon name protein Level of Low conservation General function Functional class prediction only
SECIS length 39 nt
Length UGA to 21 nt apical loop
Upper stem 8 bp
Upper stem U 50% richness
Apical loop 7 nt
155 APPENDIX 2
Selenoprotein Cell wall hydrolase name Level of Low conservation Secondary metabolites Functional class biosynthesis, transport and catabolism
SECIS length 51 nt
Length UGA to 29 nt apical loop
Upper stem 4 bp
Upper stem U 25% richness
Apical loop 8 nt
Cell division Selenoprotein FtsK/SpoIIIE-family name protein Level of Low conservation Cell cycle control, cell Functional class division and chromosome partitioning
SECIS length 39 nt
Length UGA to 16 nt apical loop Upper stem 9 bp Upper stem U 60% richness Apical loop 6 nt Selenoprotein Membrane protein name Level of Low conservation General function Functional class prediction only
SECIS length 39 nt
Length UGA to 18 nt apical loop
Upper stem 12 bp
Upper stem U 16.66% richness Apical loop 4 nt
156 APPENDIX 2
Selenoprotein Hydrolase beta-lactamase- name like (GloB) Level of 100% conservation Secondary metabolites Functional class biosynthesis, transport and catabolism
SECIS length 57 nt
Length UGA to 23 nt apical loop Upper stem 5 bp Upper stem U 0% richness
Apical loop 5 nt Pyruvate Selenoprotein flavodoxin/ferredoxin name oxidoreductase subunit gamma Level of 100% conservation Carbohydrate transport and Functional class metabolism
SECIS length 48 nt
Length UGA to 30 nt apical loop
Upper stem 6 bp
Upper stem U 0% richness
Apical loop 5 nt Selenoprotein Deoxyribonuclease name (DNase) Level of 100% conservation Replication, recombination Functional class and repair
SECIS length 51 nt
Length UGA to 26 nt apical loop
Upper stem 4 bp
Upper stem U 0% richness
Apical loop 6 nt
157 APPENDIX 2
Selenoprotein Polysaccharide deacetylase name Level of 100% conservation Secondary metabolites Functional class biosynthesis, transport and catabolism
SECIS length 54 nt
Length UGA to 26 nt apical loop
Upper stem 5 bp
Upper stem U 0% richness
Apical loop 5 nt Pyridine nucleotide- Selenoprotein disulfide oxidoreductase name (Lpd) Level of 100% conservation Energy production and Functional class conversion
SECIS length 45 nt
Length UGA to 26 nt apical loop
Upper stem 5 bp
Upper stem U 0% richness Apical loop 8 nt Selenoprotein NADPH-dependent FMN name reductase (WrbA) Level of 100% conservation General function Functional class prediction only
SECIS length 50 nt
Length UGA to 27 nt apical loop
Upper stem 4 bp
Upper stem U 0% richness
Apical loop 8 nt
158 APPENDIX 2
Oxygen-independent Selenoprotein coproporphyrinogen-III name oxidase (HemN) Level of 100% conservation Secondary metabolites Functional class biosynthesis, transport and catabolism
SECIS length 39 nt
Length UGA to 21 nt apical loop
Upper stem 5 bp Upper stem U 0% richness
Apical loop 4 nt
Selenoprotein Ribonuclease R (Rnase R) name Level of 100% conservation Translation, ribosomal Functional class structure and biogenesis
SECIS length 43 nt
Length UGA to 19 nt apical loop
Upper stem 11 bp with bulged A32
Upper stem U 0% richness
Apical loop 8 nt MarR family Selenoprotein transcriptional regulator name (EffR) Level of Moderate-low conservation
Functional class Transcription
SECIS length 48 nt
Length UGA to 19 nt apical loop
Upper stem 8 bp with bulged A13
Upper stem U 0% richness Apical loop 4 nt
159 APPENDIX 2
Selenoprotein Cell surface protein name (Cwp84) Level of High conservation
Functional class Extracellular structures
SECIS length 51 nt
Length UGA to 25 nt apical loop
Upper stem 4 bp
Upper stem U 0% richness
Apical loop 6 nt Selenoprotein Antirestriction protein name (ArdA) Level of Moderate-low conservation General function Functional class prediction only
SECIS length 51 nt
Length UGA to 29 nt apical loop
Upper stem 4 bp
Upper stem U 0% richness
Apical loop 6 nt Selenoprotein Integrase (XerC) name Level of Low conservation
Functional class Transcription
SECIS length 42 nt
Length UGA to 30 nt apical loop
Upper stem 5 bp
Upper stem U 0% richness Apical loop 3 nt
160