CHARACTERIZATION OF A PHASE VARIABLE REGULATORY SYSTEM AND ITS IMPACTS ON CLOSTRIDIOIDES DIFFICILE PHYSIOLOGY

Elizabeth Mathias Garrett

A dissertation submitted to the faculty at the University of North Carolina at Chapel Hill in partial fulfillment of the requirements for the degree of Doctor of Philosophy in the Department of Microbiology and Immunology.

Chapel Hill 2020

Approved by:

Rita Tamayo

Peggy A. Cotter

Virginia Miller

Brian P. Conlon

Elizabeth A. Shank © 2020 Elizabeth Mathias Garrett ALL RIGHTS RESERVED

ii ABSTRACT Elizabeth Mathias Garrett: Characterization of a phase variable regulatory system and its impacts on Clostridioides difficile physiology (Under the direction of Rita Tamayo)

Clostridioides difficile is a Gram-positive, spore-forming, obligate anaerobic bacterium that causes gastrointestinal disease and is the leading cause of nosocomial infections in the U.S. Despite its significant public health burden, key aspects of pathogenesis including the regulatory strategies that benefit C. difficile fitness during infection are not well understood. In this work, we identify and characterize a two- component regulatory system with significant impacts on C. difficile physiology as well as elucidate the complex regulatory networks that govern its expression. We find that expression of cmrRST, a highly conserved operon that encodes two response regulators (CmrR and CmrT) and a histidine kinase (CmrS), is subject to phase variation. Phase variation is a regulatory mechanism used by many bacterial pathogens to introduce phenotypic heterogeneity in a population as a bet-hedging strategy to help ensure the survival of the overall population. cmrRST is phase variably regulated by conservative site-specific recombination, which involves the reversible inversion of a

DNA element containing regulatory information, termed the cmr switch, by the site- specific recombinase RecV. The phase variable production of CmrRST results in two distinct C. difficile subpopulations distinguishable by colony morphology. CmrRST alters cell morphology, promotes a novel form of surface migration, and inhibits swimming

iii motility. Additionally, we provide evidence that CmrRST has a role in virulence and that phase variation occurs in vivo. In addition to phase variable regulation by the cmr switch, cmrRST expression is regulated by c-di-GMP, an intracellular signaling molecule, through a riboswitch. Furthermore, cmrRST is subject to autoregulation by

CmrR. We find that regulation by CmrR and c-di-GMP supersede the effects of the cmr switch orientation. Finally, we begin to identify and characterize the CmrRST regulon and provide evidence that CmrT is the primary regulator of gene expression that results in changes in motility and cell morphology, while CmrR appears to be dedicated to autoregulation of cmrRST transcription. This work identifies an important regulatory system with significant implications for our understanding of C. difficile population heterogeneity and pathogenesis.

iv ACKNOWLEDGMENTS First, I must thank my advisor, Rita Tamayo. Rita is an amazing mentor, and I benefited greatly from her advice and her guidance throughout my time in her lab. She is the type of advisor that every grad student wishes to have. I have always enjoyed our scientific discussions and I admire her ability to turn everything including negative data into a positive step forward. Her mentorship has been invaluable, and I will always be grateful for her part in my development as a person and as a scientist.

I am also grateful to all my lab mates through the years. I appreciate their discussion, ideas, and feedback. I have been lucky to have such friendly and collaborative colleagues. Brandon Anjuwon-Foster and Robert McKee served as great examples to look up to as I started in the lab. I appreciate all the undergraduate students that helped me through the years. I am also grateful to Ognjen Sekulovic for his collaboration, which made so much of this project possible.

Finally, I want to thank my friends and family. I am grateful to my friends in

Chapel Hill for commiserating and also celebrating the peaks and valleys of the graduate student experience, and for also providing an escape. I am so thankful for all the support that my family has given me. Their encouragement carried me forward.

Without the support of my mom and my brother, the road would have been much harder. Finally, I thank my husband Dan. Words cannot express how important his support has been. He has held my hand through good times and bad, and his presence has always been both an encouragement and a comfort. I love you.

v TABLE OF CONTENTS

LIST OF FIGURES...... ix

LIST OF TABLES ...... xi

LIST OF ABBREVIATIONS ...... xii

CHAPTER 1. INTRODUCTION ...... 1

CLOSTRIDIOIDES DIFFICILE EPIDEMIOLOGY, RISK FACTORS, AND TREATMENT ...... 1

C. DIFFICILE TOXINS ...... 6

MOTILITY AND SURFACE BEHAVIORS OF C. DIFFICILE ...... 9

Flagella ...... 9

Type IV pili ...... 11

Biofilm formation ...... 11

PHENOTYPIC HETEROGENEITY AND PHASE VARIATION ...... 13

Slipped strand mispairing ...... 15

Methylation ...... 16

Site-specific recombination ...... 17

Phase variation in C. difficile ...... 20

C-DI-GMP SIGNALING ...... 24

C-di-GMP signaling in C. difficile ...... 30

SUMMARY ...... 33

vi CHAPTER 2: PHASE VARIATION OF A SIGNAL TRANSDUCTION SYSTEM CONTROLS CLOSTRIDIOIDES DIFFICILE COLONY MORPHOLOGY, MOTILITY, AND VIRULENCE ...... 35

INTRODUCTION ...... 35

MATERIALS AND METHODS ...... 38

RESULTS ...... 48

C. difficile strains of diverse ribotypes develop two distinct, phase- variable colony morphotypes ...... 48

Rough and smooth colony isolates exhibit distinct motility behaviors ...... 49

Colony morphology is regulated by a phase variable signal transduction system and c-di-GMP ...... 50

The response regulators CmrR and CmrT regulate colony morphology ...... 53

CmrR and CmrT promote bacterial chaining ...... 56

Phase variable colony morphology via CmrRST impacts C. difficile virulence ...... 58

DISCUSSION ...... 61

FIGURES AND TABLES ...... 68

CHAPTER 3. MULTIPLE REGULATORY INPUTS CONTROL THE EXPRESSION OF cmrRST ENCODING AN ATYPICAL SIGNAL TRANSDUCTION SYSTEM IN CLOSTRIDIOIDES DIFFICILE ...... 93

INTRODUCTION ...... 93

MATERIALS AND METHODS ...... 97

RESULTS ...... 104

Generation and phenotypic characterization of phase locked cmr ON and cmr OFF strains ...... 104

C-di-GMP regulates cmrRST expression independently of the invertible element ...... 106

The cmr switch contains a promoter in the ON orientation ...... 107

vii CmrR positively autoregulates expression of cmrRST ...... 109

cmrRST and its upstream regulatory elements are well conserved in C. difficile ...... 110

DISCUSSION ...... 111

FIGURES AND TABLES ...... 116

CHAPTER 4. TRANSCRIPTOME ANALYSIS OF THE PHASE VARIABLE SIGNAL TRANSDUCTION SYSTEM CmrRST IN CLOSTRIDIOIDES DIFFICILE ...... 134

INTRODUCTION ...... 134

MATERIALS AND METHODS ...... 137

RESULTS ...... 141

CmrR and CmrT regulate many genes of diverse functions ...... 141

Cross-talk between phase variable genes ...... 143

Regulation of swimming motility and colony morphology by CDR20291_1689-1690 ...... 144

Contribution of CmrR and CmrT to regulation ...... 146

DISCUSSION ...... 147

FIGURES AND TABLES ...... 151

CHAPTER 5. DISCUSSION ...... 165

GENE REGULATION BY CmrR AND CmrT ...... 166

SIGNALS FOR THE REGULATION OF cmrRST EXPRESSION ...... 167

REGULATION OF MOTILITY, CELL MORPHOLOGY, AND COLONY MORPHOLOGY BY CmrRST ...... 172

IMPLICATIONS OF CmrRST PHASE VARIATION FOR PATHOGENESIS ...... 175

REFERENCES ...... 182

viii LIST OF FIGURES

Figure 2.1. Formation of rough and smooth colonies by multiple C. difficile strains...... 68

Figure 2.2. Reversible selection of distinct colony morphotypes with opposing motility phenotypes...... 69

Figure 2.3. Expression of a c-di-GMP regulated phosphorelay system promotes rough colony formation in a TFP- and flagellum-independent manner...... 71

Figure 2.4. The putative response regulators CmrR and CmrT promote the development of rough colonies...... 73

Figure 2.5. Growth of mutant and over-expression strains in vitro...... 74

Figure 2.6. CmrT is required for rough colony formation...... 75

Figure 2.7. CmrR and CmrT inversely regulate surface and swimming motility...... 76

Figure 2.8. Amino acid substitutions of the phosphorylation sites in CmrR and CmrT alter activity...... 78

Figure 2.9. CmrR and CmrT do not regulate flagellum or TFP gene expression...... 80

Figure 2.10. CmrR and CmrT promote bacterial chaining...... 81

Figure 2.11. The cmrT mutant is defective in cell elongation and chaining...... 83

Figure 2.12. The cmrR mutant exhibits an increase in biofilm formation...... 84

Figure 2.13. cmrR and cmrT mutants are not defective in sporulation, germination or toxin production...... 85

Figure 2.14. Colony morphology and CmrR impact C. difficile virulence...... 87

Figure 3.1. Generation of genetically phase-locked strains...... 116

Figure 3.2. Phase-locked cmr ON strain phenocopies the WT rough morphotype...... 118

Figure 3.3. Regulation of cmrRST by c-di-GMP occurs independent of cmr switch orientation...... 120

Figure 3.4. C-di-GMP does not cause inversion of the cmr invertible element...... 122

ix Figure 3.5. The cmr switch is phase locked in recV-deficient strains...... 123

Figure 3.6. The cmr switch contains a promoter in the ON orientation ...... 124

Figure 3.7. Identification of multiple transcriptional start sites upstream of cmrRST...... 125

Figure 3.8. CmrR positively regulates cmrRST expression ...... 126

Figure 3.9. cmrRST and regulatory element sequences are highly conserved across C. difficile strains and ribotypes...... 129

Figure 4.1. Orientation of invertible DNA elements varies between strain populations...... 151

Figure 4.2. CmrR and CmrT regulate CDR20291_1689-1690...... 152

Figure 4.3. CDR20291_1689-1690 regulates colony morphology...... 153

Figure 4.4. CDR20291_1689-1690 inhibits swimming motility but not surface motility...... 154

Figure 4.5. Gene regulation by CmrR occurs through CmrT...... 155

Figure 5.1. Model of cmrRST regulation and the role of CmrRST in infection...... 181

x LIST OF TABLES

Table 2.1. Strains and plasmids used in this study...... 89

Table 2.2 Primers used in this study...... 91

Table 3.1. Strains and plasmids used in this study...... 130

Table 3.2. Primers used in this study...... 133

Table 4.1. Strains and plasmids used in this study...... 157

Table 4.2. Primers used in this study...... 158

Table 4.3. Genes differentially regulated by CmrR and/or CmrT ...... 160

xi LIST OF ABBREVIATIONS

5’RACE 5’ rapid amplification of cDNA ends

5’UTR 5’ untranslated region

Amp ampicillin

AP alkaline phosphatase

ATc anhydrotetracycline

BHIS Brain Heart Infusion plus yeast bp base pair

CA-CDI community acquired Clostridioides difficile infection cDNA complementary DNA

CDI Clostridioides difficile infection c-di-GMP 3’-5’-Cyclic diguanylic acid

CFU colony forming units

Cm chloramphenicol cmr colony morphology regulator

CO2 Carbon dioxide cwp cell wall protein

DAR Division of Animal Resources

DGC diguanylate cyclase

EPS exopolysaccharide

FDR False Discovery Rate

xii FESEM field emission scanning electron microscope gDNA genomic DNA

H2 molecular hydrogen

IACUC Institutional Animal Care and Use Committee

IFN type 1 interferon

IR Inverted repeat

Kan kanamycin

LIR left inverted repeat mut mutant

N2 molecular nitrogen

OD optical density

PaLoc pathogenicity locus

PBS phosphate buffered saline

PCA principle component analysis

PCR polymerase chain reaction

PDE Phosphodiesterase qPCR quantitative polymerase chain reaction qRT-PCR quantitative reverse transcription polymerase chain reaction

R rough

RBS ribosomal binding site

REC receiver domain

RIR right inverted repeat

xiii RM restriction-modification

RPKM reads per kilobase of transcript per million mapped reads

RT ribotype

S smooth

SBP solute binding proteins

SEM scanning electron microscope

SIE superinfection exclusion system

SSM slipped strand mispairing

T3SS type III secretion system

T6SS type IV secretion system

TCCFA taurocholate cycloserine cefoxitin fructose agar

TCS two component system

TFP type IV pili

Tm thiamphenicol

TSS transcriptional start site

TY tryptone yeast

UTI urinary tract infection wHTH winged helix turn helix

WT wild type

xiv CHAPTER 1. INTRODUCTION

CLOSTRIDIOIDES DIFFICILE EPIDEMIOLOGY, RISK FACTORS, AND TREATMENT

Clostridioides difficile is a Gram-positive, opportunistic bacterial pathogen. C. difficile infection (CDI) causes mild to severe diarrhea, with a risk of complications including pseudomembranous colitis, toxic megacolon, and sepsis [1-3]. The CDC estimates that there were 233,900 cases of CDI with nearly 13,000 fatalities in 2017 in the U.S [4]. C. difficile causes an estimated 20-30% of antibiotic-associated diarrhea cases [5]. Furthermore, C. difficile is the leading cause of nosocomial infections in the

U.S. One study of a hospital over an 11-month period found that 21% of patients acquired C. difficile during their hospital stay, illustrating the extent of this problem [6].

Additionally, community-acquired CDI (CA-CDI) incidence has increased over the last decade, constituting about a third of cases in the U.S [7-9].

C. difficile is a sporulating bacterium and an obligate anaerobe [2, 3]. Because C. difficile dies in environmental oxygen conditions, the transmission of the spore is the primary means of infection through the fecal-oral route [10]. The spore is a metabolically inactive form of the bacterium in which the genetic material is protected by thick glycoprotein walls. When the spore is ingested by an individual, it passes through gastrointestinal tract until it encounters primary bile acids such as cholate in the intestine. These primary bile acids bind receptors on the spore and cause the spore to germinate into a replicative “vegetative” cell. In this form, the bacterium can colonize the colon and cause disease.

1 The normal gut microbiota in healthy individuals is very protective against CDI.

However, antibiotic use can disturb the gut microbiome, making the individual more susceptible to CDI [11-14]. Much research in recent years has been focused on elucidating the relationship between the gut microbiota and CDI. Antibiotic treatment reduces the number and the diversity of commensal bacteria that appear to prevent CDI

[13, 14]. One mechanism by which commensal gut microbes protect against CDI is through the conversion of primary bile acids necessary for spore germination into secondary bile acids [15, 16]. These secondary bile acids do not stimulate C. difficile spore germination and instead inhibit vegetative cell growth. Mice that have been treated with the antibiotic clindamycin produce more of the primary bile acid taurocholate and less secondary bile acids [11, 15]. Many commensal bacterial families including Lachnospiraceae and Ruminococcaceae are associated with transformation of primary to secondary bile acids [15, 17]. Members of these families have been shown to provide protection from CDI in animal models. Other Clostridia such as Clostridium scindens also can convert primary bile acids into secondary bile acids and are associated with providing resistance to C. difficile colonization [18]. Gut microbes additionally may prevent CDI by direct bacterial competition. A human intestinal derived strain of Bacillus thuringiensis was found to produce a bacteriocin, named Thuricin CD, that specifically targets spore-forming Gram-positive bacteria including C. difficile [19].

Commensal gut bacteria may also protect against C. difficile colonization by competing for resources needed for C. difficile growth. C. difficile has been shown to utilize components of mucin including sialic acid as an energy source [20]. Bacterial communities grown from the cecums of hamsters outcompeted C. difficile when grown

2 with limited nutrient availability; however, this effect was lost when the medium was supplemented with additional carbon sources including components of mucins [21].

Additionally, Ng et al. demonstrated that antibiotic treatment increased free sialic acid in a murine model and that this nutrient source benefitted C. difficile expansion [20]. These studies suggest that commensal microbes prevent C. difficile colonization by limiting nutrient availability. Not all commensal bacteria are protective against CDI. Bacterioides thetaiotamicron is a common gut microbe that produces an enzyme that cleaves mucin, therefore increasing the availability of sialic acid for C. difficile [20]. B. thetaiotamicron also produces high levels of succinate, which C. difficile can utilize for energy through fermentation [22]. In an animal model, co-infection of C. difficile and B. thetaiotamicron increased C. difficile bacterial burden as compared to C. difficile infection alone [20].

Additionally, in a study surveying the effect of clindamycin treatment on the gut flora of healthy individuals over two years, B. thetaiotamicron was much more highly represented than non-antibiotic treated controls [23]. Together, these data suggest that the ability of C. difficile to colonize the gut and cause infection is intimately tied to the commensal gut microbiome and that disruption of the gut microbiome increases the risk of CDI.

The incidence of C. difficile increased significantly across the world in the last 20 years [1]. This increased incidence is associated with outbreaks of epidemic ribotypes of C. difficile, including ribotype (RT) 027. RT 027 is a hypervirulent RT first recognized in 2002 [24, 25]. It quickly became the most prevalent RT isolated from outbreaks, and outbreaks caused by RT 027 strains have been associated with increased morbidity and mortality [26, 27]. Characterization of 027 strains shows that they produce more of the

3 toxins TcdA and TcdB than non-epidemic isolates and are also able to produce the binary toxin CDT [28]. Additionally, 027 strains are more antibiotic resistant. More recently, the number of cases associated with 027 strains has decreased, and it has been replaced as the most prevalent RT by 012-020 in the U.S., with 002 also becoming more common [29]. In addition to the public health burden of nosocomial cases of CDI, the incidence of CA-CDI has increased over the last decade [7-9]. CA-

CDI is more prevalent among females and affects many patient populations previously considered to be low risk, including children and young adults [30]. In Europe, CA-CDI is commonly caused by RT 027 and 014-020, while RT 027 is the most common cause in the U.S [31, 32]. Additionally, new RT are being identified regularly as the cause of both CA-CDI and nosocomial cases. Surveillance in the U.S. has led to 49 unique RT designations from 2010-2014 [29]. An outbreak of severe and recurrent cases of CDI in

2018 was shown to be caused by a new RT, 826 [33]. Therefore, continued surveillance is important to our understanding and control of CDI.

The antibiotics metronidazole, vancomycin, and fidoxomicin are commonly used to treat symptomatic CDI [34]. However, CDI has a high rate of recurrent infection after conventional antibiotic treatment; about 20-30% of patients experience a second episode of CDI, with the chance of subsequent infection increasing with each occurence

[8, 9, 35]. The mechanisms underlying recurrent infection are unclear. More than 80% of first recurrences are caused by the same strain as the initial infection; one explanation for this result is that relapse is occurring [36, 37]. Therefore, alternate mechanisms of treatment are being explored. One strategy under investigation is the use of antibiotics in conjunction with a microbe to repopulate the gut, potentially preventing C. difficile re-

4 colonization. For example, several studies in human patients showed that treatment with antibiotic as well as Saccharomyces boulardii, a species of yeast that colonizes the colon but is not maintained for long, increased the efficacy of treatment versus antibiotic treatment alone [38, 39]. Similar studies have shown mixed results [40]. Several studies have shown that the administration of Lactobacillus strains with antibiotic treatment reduced CDI incidence in patients [41, 42]. Additionally, treatment with C. difficile spores from non-toxigenic strains has been shown to reduce CDI [43]. Other studies investigating the administration of microbes with antibiotcs have shown no effect or inconclusive results [44]. Fecal microbiota transplantation has also been an effective strategy for the treatment of recurrent CDI [34, 45]. In this method, patients with CDI are given a fecal sample from a healthy donor. The microbiota from the donor repopulate the gut, reestablishing microbial diversity and preventing relapse or reinfection. Grehan et al. showed that recipients of fecal transplants were stably colonized by the donor microbiota [46]. Multiple studies have shown the efficacy of this treatment in resolving infection and preventing recurrence, with success rates around 80-90% [47-49].

However, the use of FMT is not without risk, is not well regulated, and there are few rigorous clinical studies [50]. Further efforts have been made towards preventing CDI through the use of probiotics. In vitro studies have identified several bacteria, including

Lactobacillus, Bifidobacterium and Clostridium strains, as potential probiotics that inhibit

C. difficile growth or spore germination [51, 52]. Probiotic therapies have shown mixed results in human patients. In summary, improved therapies for the treatment and prevention of CDI, especially recurrent, are required.

5 C. DIFFICILE TOXINS

C. difficile produces toxins TcdA and TcdB during colonization of the colon [53].

Both TcdA and TcdB target factors in the host cell that result in disassembly of the cytoskeleton, cell lysis, and a robust immune response [54]. While not all strains produce both, symptomatic CDI is associated with the production of at least one of these toxins [43, 55]. Therefore, their role in disease has been thoroughly studied both in vitro and in vivo.

The genes encoding TcdA and TcdB are within the 19.6 kb pathogenicity locus

(PaLoc) of the genome along with tcdE, tcdR, and tcdC [2]. TcdE is suggested to function as a holin for the secretion of toxins TcdA and TcdB [56, 57].TcdR is an RNA polymerase sigma factor that positively regulates the transcription of tcdA and tcdB [58].

TcdR also has additional regulatory targets that promote sporulation [59]. TcdC is an anti-sigma factor that antagonizes TcdR to negatively regulate expression of tcdA and tcdB [60, 61]. RT 027 strains, which are associated with large outbreaks of CDI and hypervirulence, produce high amounts of toxin which has been attributed to an inactivating mutation in tcdC [28]. In vitro data to support this has been inconclusive.

Overexpression of tcdC in 027 strain M7404 reduced toxin production as well as virulence in a mouse model of infection [62]. In contrast, restoration of a functional tcdC allele into 027 strain R20291 was not sufficient to reduce toxin levels or cytotoxicity [63].

Therefore, role of TcdC in infection is unclear.

TcdA and TcdB are glucosylating toxins with four domains: ACDB [54]. The A domain contains glucosyltransferase activity. This domain is required for the cytotoxic effects of these toxins. TcdA and TcdB glucosylate host Rho GTPases, preventing their activation. Rho proteins have a wide range of functions in the cell. Therefore, the loss of

6 their activity results in many cytopathic and cytotoxic effects. Rho inactivation causes a redistribution of the cytoskeleton, resulting in cell rounding [64, 65]. Tight junctions are also disrupted, which alters cell-to-cell contact and causes permeabilization of the epithelial cell layer [66, 67]. TcdA and TcdB also induce apoptosis, activate the inflammasome, and trigger pyroptosis [53, 68]. Together, these cytotoxic effects result in a massive inflammatory response, causing diarrhea and colitis [53, 69]. These toxins are highly potent, and administration of these toxins alone is sufficient to cause disease in an animal model [70].

The B domain is important for the binding of the toxins to host receptors [54]. The

B domains of TcdA and TcdB differ in structure, and data suggest that they have different receptors as well. While many receptors for TcdA and TcdB have been identified, the receptors important for pathogenesis in a human host are still unclear.

TcdA has been shown to bind host cell glycans [71, 72]. TcdA also binds gp96, a glycoprotein that is found on human colonocytes [73]. However, loss of gp96 binding only partially protects from TcdA cytotoxicity, suggesting other factors are at play.

Chondroitin sulfate proteoglycan 4 and the Wnt receptor FZD have been shown to bind

TcdB and promote internalization of the toxin, but their role in pathogenesis is unclear

[74, 75].

Once bound to host epithelial cells, TcdA and TcdB enter the cell by endocytosis mediated by the D domain [76]. The C domain has autoproteolytic activity and cleaves the toxin, releasing the glucosyltransferase domain [77]. This proteolytic activity is stimulated by inositol hexakisphosphate, which is produced by host cells. After

7 cleavage, the glucosyltransferase domain of TcdA or TcdB is free to modify target Rho proteins.

There has been much investigation into the individual roles of TcdA and TcdB in pathogenesis. Several studies have presented data suggesting that while both toxins can cause disease, TcdB is more potent and associated with increased virulence in animal models [78-81]. For example, Carter et al. demonstrated using three different mouse models of CDI that a mutation in tcdB, but not tcdA, caused severe attenuation of a C. difficile RT 027 strain [78]. A study in a hamster model of CDI showed similar data that hamsters infected with a strain of C. difficile only producing TcdA had significantly higher survival than those infected with wild-type or a tcdA mutant strain

[79]. Characterization of clinical isolates from patients with CDI seem to support these results. Only one disease-causing isolate that does not produce TcdB has been identified [82]. In contrast, virulent strains that do not produce TcdA are far more common [83-85].

Some C. difficile RT including epidemic RT 027 produce a third toxin called binary toxin or CDT [86-88]. CDT consists of two subunits, CDTa and CDTb [54]. CDTb binds an unknown receptor on host cells to mediate uptake while CDTa functions as an actin specific ADP-ribosyltransferase. This toxin prevents actin polymerization, causing severe disruptions in the cell cytoskeleton. Treatment of epithelial cells with CDT causes the formation of microtubule-based protrusions which may increase the ability of

C. difficile to adhere to the cell surface [89]. It is estimated that less than a quarter of toxigenic strains produce CDT [90, 91]. While CDT is not required for virulences,

8 several studies have found that CDT production is associated with increased CDI mortality [91]. However, the role of CDT in pathogenesis is still undefined.

MOTILITY AND SURFACE BEHAVIORS OF C. DIFFICILE

Flagella

Flagella serve as important virulence factors in many bacterial pathogens [92]. In other gastrointestinal pathogens including Campylobacter jejuni and Vibrio cholerae, flagella contribute motility, adherence, and colonization of the host [92-94]. Flagellar- mediated swimming motility occurs when the flagellar motor rotates the flagellar filament in a propeller-like manner. C. difficile produces peritrichous flagella that are essential for swimming motility in vitro [95, 96]. Flagellar biosynthetic genes are organized into three major operons called F1, F2, and F3 [97]. F3, also called the early stage flagellar operon, is the largest of these operons with 29 genes. This operon encodes many of the motor and structural components necessary for flagellar assembly. F1, also called the late stage operon, is not expressed until the proteins encoded by F3 have assembled.

F1 encodes proteins including the flagellin, FliC, and the cap protein, FliD. F2 encodes proteins that function in glycosylation of the flagellar filament. The F2 operon is the most divergent of the flagellar operons between C. difficile strains, leading to differences in glycosylation patterns [98-100].

In addition to encoding structural components of the flagella, the early stage flagellar operon encodes the alternate sigma factor SigD [97]. SigD is necessary for the expression of the other flagellar operons, and sigD mutants are non-motile [101-103].

Furthermore, SigD promotes the expression of tcdR [101, 104]. Because TcdR is a positive regulator of tcdA and tcdB expression, SigD promotes toxin production and

9 links this to flagellar biosynthesis. SigD has additional regulatory targets outside the flagellar operon and the PaLoc including genes involved in sporulation [101].

The role of flagella in virulence is still not well defined. Many virulent ribotypes are missing one or more of the flagellar operons, suggesting that they are dispensable for virulence [105]. Flagella contribute to multiple phenotypes in C. difficile in addition to motility. For example, flagella contribute to agglutination. The degree of agglutination is strain dependent; even within the same ribotype, strains have different agglutination habits, which may be due to differences in glycosylation of the flagellar filament [98-

100]. Flagella also affect adherence and colonization in a strain dependent manner.

Loss of the flagellar filament through a mutation in fliC in RT 027 strain R20291 resulted in reduced adherence to mammalian cell lines as well as reduced colonization in a mouse model [95, 96]. Interestingly, RT 027 strain R20291 in which a gene encoding the flagellar motor, motB, had been mutated showed no difference in colonization, suggesting that motility is not required for colonization but the flagellum may function as an adhesin [95]. In contrast, a fliC mutation in the RT 012 strain 630Δerm increased adherence to mammalian cell lines and had no difference in colonization in a mouse model. Two studies in 630Δerm showed that mutant fliC and fliD strains were more virulent in a hamster model of CDI, possibly because flagella are immunogenic or because loss of these genes increased toxin production [96, 102]. In summary, flagella are an important component of C. difficile physiology and have a complex and multi- faceted role in the interaction between the bacterium and the host.

10 Type IV pili

Type IV pili (TFP) are extracellular appendages produced by a wide range of both Gram-positive and Gram-negative bacteria [106, 107]. Smaller and thinner than flagella, TFP similarly can function in motility and adherence. TFP-mediated motility occurs by a grappling hook mechanism in which TFP extend, attach to a surface, and retract, pulling the bacterium along [108]. TFP are a key virulence factor in many bacteria including Clostridium perfringens, where they contribute to a surface motility termed gliding motility, adhesion to host tissues, and biofilm formation [109-112].

Genes encoding components for TFP are considered core genes and are found in all sequenced C. difficile strains [105, 106]. In C. difficile, TFP contribute to surface motility and autoaggregation in vitro [113, 114]. TFP also promote adherence to host cells. McKee et al. showed that TFP-deficient mutants had reduced adherence to mammalian cell lines at 24 hours post inoculation but not at 1 hour, suggesting that TFP are not required for initial attachment but contribute to long term adherence [115].

Similarly, in a mouse model of CDI, TFP mutants behaved similarly to WT at early stages of infection but were cleared more rapidly, indicating a role for TFP not in initial colonization but in persistence. In a hamster model of infection, C. difficile was shown to express TFP genes, and microscopy showed the attachment of bacteria to the host epithelia by structures resembling TFP [116]. However, the significance of TFP- mediated motility versus adherence during infection is unknown.

Biofilm formation

Biofilms are complex bacterial communities that form both in the environment and during infection [117]. Biofilms consist of bacterial aggregates adhered and

11 surrounded by a protective matrix composed of polysaccharides, proteins, and extracellular DNA. For many bacterial pathogens, biofilm formation is a key virulence strategy [117, 118]. Biofilms are highly protective against antimicrobials, environmental stresses, and the host immune response. Additionally, they contribute to bacterial persistence and the recurrence of infection.

C. difficile biofilms were first described in 2012 [119], and the ability of C. difficile to form biofilms in vitro is now well documented [120-123]. Several factors have been shown to contribute to biofilm formation in vitro. Multiple studies have shown using different models of biofilm formation that TFP contribute to biofilm formation in C. difficile [113, 121, 122]. Additionally, genes encoding TFP components are more highly expressed during biofilm growth than planktonic growth [122]. Like that of other bacteria, the C. difficile biofilm matrix consists of DNA, protein, and polysaccharides

[123-125]. Treatment of biofilms with proteinase K or DNAse I prevents biofilm formation and disperses established biofilms, suggesting that these molecules are important structural components of the matrix [123, 124]. Furthermore, flagella and components of the S-layer have been shown to contribute to biofilm formation as well [123, 126, 127].

Less is understood about the importance of C. difficile biofilms in vivo. In a mouse model of infection, Lawley et al. described large aggregates of C. difficile associated with the host epithelium that corresponded to sites of intense inflammation and tissue damage [128]. Similarly, Soavelomandroso et al. showed that some, but not all, C. difficile strains form aggregates on the surface of the mucus layer in the cecum and colon of mice [129]. These aggregates contained polysaccharides and extracellular

DNA consistent with a biofilm matrix. Similar aggregates have been seen in the cecums

12 and colons of hamsters [125]. Further studies have suggested that C. difficile can form polymicrobial biofilms with gut microbiota [130]. While these studies suggest that C. difficile forms biofilms in vivo, their role in infection remains unclear.

Biofilms are linked to recurrent infection in many bacterial pathogens [117, 118].

Therefore, whether C. difficile biofilms contribute to recurrent infection is of great interest. Semenyuk et al. showed that mature biofilms contain a dense concentration of spores [125]. As spores are highly resistant to antimicrobials, biofilms might serve as a reservoir of spores for reinfection. However, other studies have shown conflicting results, possibly due to differences in growth conditions and strains tested [123, 129]. C. difficile in biofilms are more resistant than planktonic cells to metronidazole and vancomycin, antibiotics commonly used to treat CDI [123-125, 131]. This observation suggests that biofilms protect vegetative cells from antibiotics, and these cells could reestablish infection once antibiotic treatment has ceased. Further study on the role of biofilms in recurrent CDI is necessary.

PHENOTYPIC HETEROGENEITY AND PHASE VARIATION

Even within the same environmental conditions, populations of clonal bacteria can display diverse phenotypes [132, 133]. This phenotypic heterogeneity is an important survival strategy for bacteria. If environmental conditions change quickly, bacteria may not be able to sense and develop a response rapidly enough to ensure survival. Phenotypic heterogeneity works as a bet-hedging strategy to help ensure that at least one of a number of phenotypically distinct subpopulations of bacteria have a fitness or survival advantage.

13 Many bacterial pathogens have been demonstrated to generate phenotypic heterogeneity, which can contribute to virulence as a survival strategy that promotes persistence during infection. One well characterized example of phenotypic heterogeneity as a virulence strategy is the development of persisters [132, 134]. Many pathogens including Staphylococcus aureus, Mycobacterium tuberculosis, and

Escherichia coli produce a small subpopulation of slow-growing or dormant cells called persisters. Under favorable conditions, this subpopulation may go unnoticed. However, persisters are highly resistant to antibiotic treatment as well as many host immune responses. Thus, they have a fitness advantage compared to vegetative cells during these stressful conditions. Once in more favorable conditions (such as the cessation of antibiotic treatment), persisters can resume normal growth. Persister cells have been shown to perpetuate chronic and recurrent infections and serve as an important example of how phenotypic heterogeneity can aid pathogens in infection.

Phenotypic heterogeneity can be introduced into a population through many mechanisms including phase variation [135-137]. Phase variation is a reversible change, often a recombination event, which happens stochastically at a low rate per bacterial generation. This reversible change results in a change in gene expression and an ON/OFF phenotypic switch. The subpopulation with the switch in the state that confers the greatest fitness advantage, ON or OFF, may become the dominant subpopulation as a result of selective pressure. Phase variable factors are typically surface components including pili, flagella, and capsule. These factors often provide distinct advantages under specific conditions but may be costly due to energy requirements or immunogenicity, for example. Phase variation is widespread throughout

14 the bacterial kingdom, including commensal and pathogenic bacteria, and can occur through multiple mechanisms [136, 138, 139].

Slipped strand mispairing

Small, repetitive DNA sequences are highly mutagenic due to the tendency for slippage during DNA replication [140, 141]. One mechanism of phase variation called slipped strand mispairing (SSM) has evolved to utilize these “mistakes” in order to regulate gene expression in an ON/OFF manner. Repetitive sequences can cause a misalignment between daughter and parent strands during replication, resulting in an expansion or contraction of the repeated sequence. If SSM occurs in a promoter region, transcription of the downstream may be affected. Alternately, SSM in a coding region can alter the open reading frame, affecting translation. However, because either expansion or contraction can occur in these repetitive sequences, the sequence change is reversible. Sequences involved in SSM have been shown to occur with 1 to 7 nucleotide repeats, though tetrameric repeats are most common [136, 142-144].

SSM has been demonstrated in multiple pathogens including Neisseria meningitidis, Haemophilus influenzae, and Campylobacter jejuni [136]. In H. influenzae, hifA, encoding the major pilin, is phase variably regulated by SSM of nine to eleven dinucleotide repeats [145]. These repeats are found between the -10 and -35 regions of the promoter. Therefore, expansion or contraction of this region allows RNA polymerase to only bind and drive transcription with a permissive number of repeats. HifA pili are a virulence factor that contribute to adherence; therefore, phase variation of this factor could be significant for pathogenesis [146].

15 Methylation

Reversible changes in the DNA methylation can alter gene expression in a phase variable manner [136, 137]. Methylation of important regulatory elements can affect the binding of transcription factors or other DNA-binding proteins. In contrast to other mechanisms of phase variation, this mechanism does not alter the DNA sequence but rather is epigenetic.

Phase variable methylation has been identified in Salmonella enterica

Typhimurium and E. coli [136]. Several loci in E. coli are phase variably regulated by the activity of Dam methyltransferase including pap, which encodes pili that are a key virulence factor contributing to pyelonephritis during urinary tract infections (UTI), and agn43, which encodes the outer membrane protein Ag43. Ag43 promotes autoaggregation as well as biofilm formation, contributing to bacterial virulence, persistence, and recurrent UTI [147-150]. Dam methyltransferase globally recognizes

GATC sequences in the E. coli genome and methylates the adenine [151]. When three

GATC sequences near the -10 site of the ag43 promoter are methylated, the transcriptional regulator OxyR is unable to bind [152, 153]. Upon DNA replication, the genome becomes hemimethylated, allowing OxyR to bind and repress expression of ag43. Both Dam and OxyR are constitutively expressed [151]. Therefore, both compete for the ag43 promoter sequence in its hemi- or nonmethylated state, leading to ON and

OFF phase cells within the bacterial population. Importantly, the methyltransferase itself is not phase variably regulated; only the methylation state and therefore downstream gene expression is phase variable.

In contrast, phase variable methylation in Streptococcus pneumoniae occurs through phase variation of a methyltransferase. Methyltransferases contain an S subunit

16 that is responsible for recognition of DNA sequences and the specificity of binding [154].

S. pneumoniae has three genes encoding S subunits. Only one of these genes encodes a functional protein; however, extensive recombination by inversion occurs between these three genes, altering the sequence of the S subunit and therefore the recognized binding sequences of the methyltransferse [155-157]. Altered methylation patterns result in differential production of many factors including capsular exopolysaccharides, causing phase variable opaque and transparent colonies. Studies using animal models of infection indicate that the transparent variant has improved colonization while the opaque is better adapted for later stages of infection [154].

Site-specific recombination

Another mechanism of phase variation involves site-specific DNA recombination

[136, 137]. In contrast to homologous recombination, which typically requires more than

50 base pair (bp) of sequence homology, site-specific recombination can occur with cognate sequences of less than 30 bp. Site-specific recombination includes inversion, excision, and insertion [158]. While excision and insertion events have been shown to mediate phase variation, inversion of DNA elements is more common. During inversion, also called conservative site-specific recombination, DNA sequences flanked by inverted repeats can undergo strand exchange mediated by a recombinase, resulting in

“flipping” of the region between the inverted repeats. Crucially, no sequence is lost during this event, and the sequence can repeatedly invert without sequence degradation. If the invertible element contains regulatory information, its inversion can regulate expression of surrounding genes. Invertible DNA elements are widespread

17 through bacteria, and advances in the field continue to elucidate the extent to which they are present [138, 139].

One well-characterized example of phase variation mediated by an invertible element occurs in E. coli [136, 159]. The fimS invertible element consists of a 296 bp sequence flanked by 9 bp inverted repeats. Within the invertible element is a promoter.

One orientation of the invertible element, called the ON orientation, contains the promoter in the proper orientation to promote transcription of the downstream gene, fimA; in the OFF orientation, the promoter is divergently positioned, and fimA is not transcribed. FimA is a major component of type 1 fimbriae. Type 1 fimbriae are a key virulence factor in UTIs. Strains lacking type 1 fimbriae are highly attenuated in an animal model as well as in human patients with UTI [160]. Additionally, the ability to phase vary the production of type 1 fimbriae is an important virulence strategy. E. coli in which the fimS is locked in the OFF orientation was shown to be significantly attenuated in a mouse model of UTI [161]. Further studies have shown that during infection of the lower urinary tract, fimS is found predominantly in the ON orientation [162, 163]. The invertible element is predominantly OFF in bacteria recovered from the urine, suggesting that type 1 fimbriae are important for adherence to uroepithelia during infection of the bladder [164]. However, non-fimbriated bacteria become predominant over time during infection of the kidney, possibly because fimbriae are immunogenic; therefore, there may be selection against fimS ON cell in this environment [163, 165,

166]. Together, these studies suggest that phase variable regulation of type I fimbriae is important to virulence and pathogenesis.

18 Site-specific recombination requires the activity of recombinases [158]. Nearly all of these recombinases are either in the tyrosine or serine recombinase families. While these two families of recombinases act through different mechanisms, the end result is the same. Recombinases bind the inverted repeats flanking the invertible DNA sequence. After the DNA-protein complex has been formed, the recombinase cleaves, rearranges, and rejoins the DNA strands such the intervening sequence has been inverted. In the case of the fim system, the tyrosine recombinases FimB and FimE are encoded directly upstream of fimS [136, 159]. FimB mediates inversion of fimS bidirectionally (i.e., ON to OFF and OFF to ON) [167]. In contrast, FimE predominantly inverts this sequence from ON to OFF, and only acts in the opposite direction at a very low rate. This selectivity occurs due to binding affinity of FimE to the left inverted repeat with fimS in the OFF orientation, impeding inversion from this orientation [167, 168].

The amount of a recombinase within the cell affects the rate of inversion [138,

169]. Therefore, factors that regulate FimB and FimE production affect the proportion of

ON/OFF cells within a population. The DNA binding protein H-NS represses expression of both recombinase genes by blocking the promoter of each [170, 171]. Additionally, the orientation of fimS affects fimE expression. In the OFF orientation, the fimE transcript is destabilized by the formation of a Rho-dependent terminator at the 3’ end of the coding sequence [172, 173]. Multiple factors called recombination directionality factors have been shown to affect inversion by changing the interaction of the recombinases with fimS [136, 159, 169]. IHF and Lrp bind the DNA around the invertible element, bending it in such a way that it increases the ability of the recombinase to interact with the inverted repeats [159]. Studies have found that the rate of inversion is

19 decreased in Lrp-deficient cells, while fimS is unable to invert without IHF [174-177].

The expression of the recombination directionality factors can be regulated by environmental factors. For example, the expression of lrp is regulated by nutrient availability [178]. While phase variation is canonically stochastic, the dynamics of inversion are subject to a network of interactions and multiple levels of regulation. The well-characterized fimS invertible element serves as an important model for further understanding of phase variation through conservative site-specific recombination.

Phase variation in C. difficile

Recent work by our lab and others has revealed that C. difficile has the potential to generate extensive phenotypic heterogeneity through phase variation. Several invertible DNA elements have been identified in C. difficile. The first invertible element to be identified, Cdi1, was found by Emerson et al. in a 2009 study that investigated the cell wall protein CwpV [179]. This study found that within a clonal population of bacteria only 5-10% of cells expressed cwpV. They identified a putative invertible element flanked by inverted repeats directly upstream of cwpV and confirmed that this sequence undergoes inversion. Transcriptional analysis revealed that transcription was initiated upstream of the invertible element. The ON orientation permits transcription to continue through the invertible element, allowing cwpV expression. In contrast, in the OFF orientation, the nascent mRNA forms an intrinsic transcriptional terminator that prematurely halts transcription 5’ of the cwpV coding sequence. This post-transcription initiation mechanism of regulation by an invertible element was the first of its kind to be described.

20 The significance of the phase variable regulation of cwpV is unclear. CwpV is a cell wall protein that is antigenically distinct among C. difficile strains [180]. Despite the differences in protein structure, its phase variable regulation seems to be conserved.

Overexpression of cwpV results in autoaggregation of cells in vitro, and CwpV also confers resistance to phage predation by allowing adsorption of phage but preventing

DNA injection [181]. Based on these results, Sekulovic et al. speculated that CwpV functions as a superinfection exclusion (SIE) system, a mechanism of phage defense

[181, 182]. If so, a subpopulation of CwpV-producing cells may provide protection to neighbor cells by binding and disarming phage. However, the significance of this phenomenon to infection remains unknown.

Stabler et al. compared whole genomes of a three C. difficile strains [183]. In addition to multiple mutations, insertions, and excisions, they identified three sequences found in inverted orientations between strains and flanked by inverted repeats. One of these was the previously identified Cdi1 invertible element, regulating cwpV. The other two were named Cdi2 and Cdi3. Both are upstream of genes encoding c-di-GMP phosphodiesterases. They speculated that these sequences are also invertible elements with a role in phase variation but did not characterize them further.

Anjuwon-Foster et al. identified the Cdi4 invertible element, also called the “flg switch”, upstream of the early stage flagellar operon [103]. They verified that the sequence flanked by inverted repeats inverts and regulates expression of the downstream flagellar genes in a phase variable manner. The mechanism of regulation by this invertible element is unknown. The use of transcriptional reporters indicated that the flg invertible element does not contain promoter, nor does it contain an intrinsic

21 terminator as observed for the cwpV invertible element. However, this sequence does stop transcription from the promoter found directly upstream of the flg invertible element, PflgB, suggesting another mechanism of transcriptional termination.

The flg switch regulates the expression of the early stage flagellar operon [103].

As a result, flg ON cells are flagellated while flg OFF cells are aflagellate and do not undergo swimming motility. The early stage flagellar operon encodes SigD, an alternate sigma factor that promotes the expression of the toxin genes tcdA and tcdB. Therefore, the production of toxins is phase variable as well through regulation by SigD. flg ON bacteria produce significantly more toxin than flg OFF bacteria and are more cytotoxic to mammalian cell lines. Together, these results indicate that C. difficile populations consist of motile/toxigenic and non-motile/non-toxigenic subpopulations. Because of the significance of these factors to virulence, the phase variable production of flagella and toxins likely has significant ramifications for infection, though this has not been experimentally supported.

The initial characterization of flg phase variation was done in RT 027 strain

R20291, an epidemic and hypervirulent isolate. RT 012 strain 630, isolated from a patient in 1982, was found to have the invertible element only in the ON orientation

[184]. Additionally, Anjuwon-Foster et al. investigated the flg invertible element in RT

012 clinical and environmental isolates and found that it was only in the OFF orientation.

These data suggest that in contrast to RT 027 strains, RT 012 strains may only invert at an extremely low rate. The underlying reason is unknown but may be related to differences in the sequence of the inverted repeats. Whether that lack of ability to readily invert the flg invertible element affects virulence is unclear. Further investigation

22 of the dynamics of phase variation in vivo might provide important information about C. difficile pathogenesis.

Sekulovic et al. utilized a new method for detecting invertible elements using deep sequencing of whole bacteria genomes [138]. Analysis of the reads from pair-end sequencing shows specific patterns in regions of inversion. Pairs of reads are typically of opposing orientations and at a certain inner-mate distance. In regions of inversion, these read-pairs are in the same orientation and at a longer inner-mate distance. These atypical reads are typically trimmed or discarded during mapping, but they can predict the location of invertible elements. When accounting for such atypical reads in whole genome sequencing of C. difficile strain R20291, the four previously recognized invertible elements were detected and three additional elements were identified.

Sekulovic et al. demonstrated that these predicted invertible elements undergo inversion and at least one of them regulates downstream genes in a manner consistent with phase variation. Application of this method in C. difficile RT 078 and 017 identified one additional invertible element not found in RT 027 [185]. Additionally, Sekulovic et al. found that some invertible elements are widely conserved with a high degree of sequence homology among C. difficile strains, while others are limited to specific lineages. In summary, eight invertible DNA elements have been identified in C. difficile, seven of which are found in strain R20291. To this point, only two (cwpV and flg) have been characterized, and the study of a third invertible element is contained within this dissertation. Together these data indicate that C. difficile is capable of considerable phenotypic heterogeneity through phase variation.

23 Additional work has been focused on the identification of the recombinases responsible for inversion. Emerson et al. investigated the ability of seven conserved recombinases to invert the Cdi1 invertible element and found that only one, designated

RecV, had this ability [179]. RecV was able to invert Cdi1 in a heterologous organism, suggesting that no additional C. difficile-specific factors are required. Deletion of the gene encoding this recombinase locked the invertible element, indicating that RecV is required for inversion [180]. Anjuwon-Foster et al. similarly found that RecV was both sufficient and required for inversion of the flg switch [103]. Further study found that

RecV affected the orientation frequency of five of the seven invertible elements in

R20291 [138]. Typically, invertible elements have a dedicated recombinase encoded in their vicinity, but the regulation of multiple invertible elements by one recombinase has been described [158, 159]. The Mpi recombinase in Bacteroides fragilis regulates the inversion of at least 13 sequences throughout the genome. Most of these loci are associated with the regulation of polysaccharide production, allowing for a high degree of antigenic variability [186]. Similarly, regulation of multiple invertible elements by RecV may suggest some coordination or common function between the regulated loci or that high phenotypic diversity resulting from one recombinase is beneficial under some conditions. Whether RecV regulates these invertible elements in a hierarchical manner is unknown but may provide important insights into the dynamics of phase variation in

C. difficile.

C-DI-GMP SIGNALING

3’-5’-Cyclic diguanylic acid (also cyclic diguanylate, c-di-GMP) is an intracellular dinucleotide signaling molecule. Enzymes for c-di-GMP metabolism have been

24 identified in all major bacterial phyla, evidencing the ubiquity of this molecule [187]. C- di-GMP was first discovered in 1987 as a signal stimulating cellulose production in

Gluconacetobacter xylinus [188]. Since then, c-di-GMP has been shown to have roles in the regulation of many types of motility, polysaccharide production, biofilm formation, toxin production, attachment and adherence, and many other phenotypes.

Most bacteria exist in a surface-associated state [189, 190]. However, the ability to move is important for nutrient acquisition, the avoidance of detrimental conditions, and other essential functions. C-di-GMP has an important role in many bacteria in the transition between sessile, community-associated and motile, free-living lifestyles.

Broadly, c-di-GMP is a negative regulator of factors that promote motility, such as flagella, and is a positive regulator of factors that promote surface attachment, including extracellular matrix components, pili, and other adhesins [189, 191, 192]. C-di-GMP is well-noted as a positive regulator of biofilm formation. C-di-GMP also plays a role in sensing and responding to surface contact. Several studies have shown that intracellular c-di-GMP levels increase with surface contact in several bacteria [113, 193-

195]. Additionally, c-di-GMP regulated factors like flagella and TFP have a role in surface sensing in many bacteria. For example, flagella can mediate transient attachment to a surface in many species including Pseudomonas aeruginosa [196]. This contact via flagella can begin a signal cascade involving c-di-GMP that results in the production of factors including TFP and exopolysaccharides that tether the cell to the surface [193, 197]. Caulobacter crescentus utilizes a similar mechanism of reinforcing transient, flagellar surface attachment with adhesins through a signal cascade involving

25 c-di-GMP [198]. Attachment to a surface is a key step in colonization, tissue invasion, and biofilm formation.

Numerous bacterial pathogens utilize c-di-GMP signaling pathways to control the production of virulence factors [192, 199]. In P. aeruginosa, c-di-GMP is a negative regulator of the type III secretion system (T3SS), a major virulence determinant that delivers toxic effector proteins to host cells [200, 201]. Furthermore, c-di-GMP positively regulates the type VI secretion system (T6SS). The T6SS contributes to interbacterial killing and is associated with chronic infection [202-204]. C-di-GMP also regulates the production of cholera toxin by Vibrio cholerae, which drives pathogenesis [205]; the production of the Dot/ICM type IV secretion system in Legionella pneumophila, which is required for intracellular growth of the bacteria [206]; and T3SS effector production in S. enterica serovar Typhimurium, which contributes to host cell invasion [207, 208].

Together, these examples convey the important role that c-di-GMP plays in the virulence of many bacterial pathogens.

Not only is c-di-GMP a regulator of virulence factors in bacteria, but it can also modulate the host immune response. Karaolis et al. found that the administration of c- di-GMP to mice before infection with Staphylococcus aureus significantly reduced bacterial burden but did not affect bacterial growth in vitro [209]. Administration of c-di-

GMP in the absence of bacteria increased monocyte and granulocyte recruitment and activated proinflammatory signals in dendritic cells and macrophages. McWhirter et al. demonstrated that c-di-GMP is sensed via a cytosolic surveillance pathway [210]. Two host proteins have been shown to bind c-di-GMP and initiate an inflammatory response

[211]. STING, a protein found on the endoplasmic reticulum, binds c-di-GMP and

26 initiates a type 1 interferon (IFN) response. The host protein DDX41 also binds c-di-

GMP and contributes to an IFN response, though this role may be dependent on

STING. Binding of c-di-GMP to DDX41 promotes the interaction of DDX41 with STING, potentially increasing the affinity of STING for c-di-GMP and potentiating the inflammatory response. While the pro-inflammatory effects of c-di-GMP are well established, it is unclear how c-di-GMP becomes extracellular to the bacterium. It is possible that c-di-GMP exits bacteria through non-specific transporters [212].

Alternatively, it may be released as a consequence of bacterial cell death and lysis. The conditions under which this is relevant for infection are unclear.

C-di-GMP is synthesized from two GTP molecules by diguanylate cyclases

(DGCs), which often act as homodimers. DGCs contain a GGDEF domain containing the active site necessary for c-di-GMP synthesis. C-di-GMP is degraded by phosphodiesterases (PDEs), which contain either EAL or HD-GYP domains for hydrolysis. The activity of these enzymes is often regulated through signal input domains, such as PAS or REC response regulator domains. Through these domains, c- di-GMP metabolic enzymes can respond to environmental signals, and the opposing activities of these enzymes modulate intracellular c-di-GMP concentrations [213-216].

For example, the DGC WspR in Pseudomonas aeruginosa becomes phosphorylated at its REC domain in response to surface contact [217]. As a result, phosphorylated WspR oligomerizes at the cell poles and synthesizes c-di-GMP [218]. Proteins can also contain both GGDEF and EAL domains [216, 219]. Often, one of these domains is degraded such that it has lost catalytic activity but can still bind c-di-GMP and regulate activity of the enzyme.

27 C-di-GMP receptors include both proteins and RNA riboswitches. Several c-di-

GMP binding protein domains have been identified [220, 221]. This includes degenerate

GGDEF and EAL domains as previously mentioned, which can regulate the activity of the c-di-GMP receptor. PilZ domains are a well-characterized c-di-GMP binding domain found in many bacteria [222]. PilZ domains canonically contain two conserved motifs: an arginine rich motif (RXXXR) and a (D/N)XSXXG motif within a β-barrel. This domain can be found alone or in conjunction with other domains. Single-domain PilZ proteins are thought to function as adaptor proteins, interacting with and regulating the function of proteins that don’t directly bind c-di-GMP. PilZ domains can also be found alongside a multitude of domains with diverse functions, regulating their function through c-di-

GMP. For example, YcgR-like proteins, found in diverse bacteria including E. coli and

Bacillus subtilis, contain an N-terminal β-barrel domain and C-terminal PilZ domain

[222-225]. In response to c-di-GMP binding, the N-terminal domain interacts with flagellar motor or stator proteins, resulting in the modulation of motility. In several proteins, a binding site for another nucleotide has evolved to bind c-di-GMP, such as

FleQ in P. aeruginosa and Clp in E. coli [220].

C-di-GMP can also bind riboswitches, which are secondary structures that form in the 5’ untranslated region (5’UTR) of RNA transcripts [226, 227]. Riboswitches consist of an aptamer domain, which binds a ligand, and an expression platform, which regulates transcription or translation of the mRNA. Ligand binding to the aptamer causes a change in conformation, leading to regulatory changes through the expression platform. For example, in the absence of a ligand, the expression platform can form a transcriptional terminator. Binding of the ligand to the aptamer can cause a

28 conformational change, abolishing the terminator and allowing transcription to continue.

Alternatively, the expression platform may permit transcription in the absence of ligand, and binding results in the formation of a transcriptional terminator. Riboswitches can also regulate translation by altering the availability of the ribosomal binding site (RBS) through conformational changes in the RNA secondary structure. Lee et al. demonstrated that a c-di-GMP riboswitch regulated the spliced products of an allosteric ribozyme [228]. In the presence of c-di-GMP, the ribozyme splices such that the product contains an RBS and start codon, allowing translation. In the absence of c-di-GMP, an alternate secondary structure forms such that the spliced product is lacking an RBS.

These examples demonstrate that riboswitches can function in many different ways. As detailed in the next section, riboswitches appear to be the primary receptors of c-di-

GMP leading to transcriptional and physiological changes in C. difficile.

The specificity of c-di-GMP signaling is an area of great interest within the field.

Many bacteria that utilize c-di-GMP signaling encode multiple DGCs and PDEs, and these proteins can have specific roles in the regulation of individual factors, even within bacteria encoding multiple c-di-GMP effectors [229]. One mechanism underlying this is differences in effector c-di-GMP binding affinity. C-di-GMP effectors can have vastly different binding affinities such that they effectively become activated at different thresholds of c-di-GMP concentrations. For example, Pultz et al. demonstrated that

PilZ-domain proteins within both S. enterica serovar Typhimurium and P. aeruginosa had vastly different binding affinities, and altering this binding affinity changed the phenotypic response to c-di-GMP [230]. Another mechanism that contributes to c-di-

GMP signaling specificity includes interactions between the DGC or PDE and the

29 effector. For example, a DGC can cause a local increase in c-di-GMP concentrations such that a co-localized effector, but not a distal effector, can respond. For example, the

P. fluorescens DGC GcbC directly binds the effector LapD [231, 232]. LapD contains a degenerate EAL domain that binds c-di-GMP, activating the protein and contributing to biofilm formation [233, 234]. Furthermore, this interaction occurs regardless of c-di-GMP production by GcbC, suggesting that c-di-GMP can be quickly produced and sensed within this preformed complex [231, 232]. These studies indicate that c-di-GMP signaling is dynamic with both global and specific regulatory roles.

C-di-GMP signaling in C. difficile

C-di-GMP is a key regulator in C. difficile [235]. Transcriptome analysis of a strain overexpressing a DGC revealed 166 genes that are differentially regulated in response to elevated c-di-GMP levels [236]. C-di-GMP promotes aggregation, biofilm formation, adherence to mammalian cells, and surface motility in C. difficile [113-115,

237]. These phenotypes may in part be due to the regulation of TFP, which contribute to these behaviors as discussed above. However, TFP-deficient mutants still exhibit increased biofilm formation and adherence, suggesting other c-di-GMP-regulated factors at play [113, 115]. C-di-GMP regulates multiple known or putative adhesins, and several of these have been demonstrated to increase biofilm formation when overexpressed [236]. Additionally, c-di-GMP is a negative regulator of the flagellar biosynthetic operons [101, 104, 237]. Overexpression of a DGC inhibits swimming motility. Because toxin gene expression is linked to flagellar gene expression through the alternate sigma factor SigD, c-di-GMP also negatively regulates the expression of

30 tcdA and tcdB [101, 104]. Therefore, c-di-GMP is an important regulator of known and putative virulence factors of C. difficile.

C. difficile has a large c-di-GMP network including 37 genes encoding predicted

DGCs and PDEs, consisting of nearly 1% of its total genome [238, 239]. Based on two separate studies expressing these genes in heterologous bacteria, 28 DGCs and PDEs have been demonstrated to be enzymatically active. Several of these proteins, including some not found to be active, contain both GGDEF and EAL domains, suggesting that they may be c-di-GMP receptors. However, the regulation and function of specific DGCs and PDEs is largely unknown. The exception is PdcA, an EAL-domain PDE that also contains a PAS domain and a degenerate GGDEF domain [120]. Removal of the PAS and GGDEF domains decreased enzyme activity, suggesting that these domains regulate the EAL domain activity. Expression of pdcA is regulated by CodY, a global transcription factor that represses gene expression during nutrient-rich conditions such as exponential phase. Consistent with this, pdcA is expressed during the stationary phase of growth. Additionally, PdcA activity is regulated by GTP, which is more abundant during stationary phase. A pdcA mutant strain had decreased toxin production and cytotoxicity against a mammalian cell line as well as increased biofilm formation.

However, loss of PdcA had no effect on swimming motility or on total intracellular c-di-

GMP levels. These data indicate that PdcA has a specific role in c-di-GMP signaling, modulating only a subset of c-di-GMP regulated processes through an unknown mechanism. PdcA is only one of many c-di-GMP metabolizing proteins in C. difficile which likely also have specific signaling roles and patterns of activation.

31 No c-di-GMP binding protein effectors have been demonstrated in C. difficile, though TFP ATPase PilB1 has a predicted c-di-GMP binding MshEN domain [240].

However, Soutourina et al. identified and confirmed the expression of 16 c-di-GMP riboswitches in strain 630Δerm [241]. Mckee et al. confirmed that 11 of these are functional and differentially regulate downstream genes in response to altered c-di-GMP levels [236]. Seven riboswitches are class I and negatively regulate gene expression while four are class II and positively regulation expression. Therefore, it seems that riboswitches are the primary mechanism by which c-di-GMP signals in C. difficile.

Several of these riboswitches have been characterized. One riboswitch, Cdi-2-4, regulates the major pilin gene pilA1 and is therefore responsible for the regulation of

TFP by c-di-GMP. Cdi-2-4 positively regulates pilA1; binding of c-di-GMP to this riboswitch abrogates the formation of a transcriptional terminator [113, 114]. The Cdi-1-

3 riboswitch is directly upstream of flgB and the early stage flagellar operon. This riboswitch, when bound to c-di-GMP, negatively regulates expression of flagellar genes as well as toxin genes through SigD. The early stage flagellar operon is also phase variably regulated by an invertible element. Under basal c-di-GMP levels, the Cdi-1-3 riboswitch permits transcription and therefore the invertible element is the determinant of flagellar gene expression. However, when the riboswitch binds c-di-GMP, transcription is prematurely terminated regardless of the orientation of the invertible element. Therefore, the production of flagella and toxins are regulated by both c-di-GMP and phase variation.

32 SUMMARY

Many aspects of C. difficile biology and pathogenesis remain poorly understood.

Several factors contribute to this lack of information. C. difficile has only emerged as a major public health threat in recent decades [1]. Additionally, as suggested by its name,

C. difficile can be difficult to culture as an obligate anaerobe, and many techniques utilized with oxygen-tolerant microbes do not function in anaerobic environments [242,

243]. C. difficile is not naturally competent and was considered to be genetically intractable until recent years; the first plasmids were not conjugated into this organism until 2002 and the first targeted gene disruption was not created until 2006 [244-247].

Advances in the field have allowed new insights into C. difficile physiology, but large gaps in our understanding remain, including many key aspects of virulence and the regulatory strategies that allow adaptation and persistence in infection.

Recent studies have identified that with multiple invertible DNA elements, C. difficile is capable of considerable phenotypic heterogeneity, but the extent and impact of this heterogeneity is unclear [138, 185]. This work began with the goal of characterizing one visible sign of phenotypic heterogeneity, colony morphology. We observed that C. difficile strains from multiple ribotypes form two distinct colony morphologies, the cause and significance of which were unknown. In this work, we characterized the operon colony morphology regulator cmrRST, its role in multiple aspects of C. difficile physiology, and its regulation. This operon encodes a putative two component regulatory system including a sensor kinase (CmrS) and two DNA-binding response regulators (CmrR and CmrT). We show that cmrRST is subject to a complex and hierarchical network of regulation including phase variable regulation by an invertible DNA element as well as regulation c-di-GMP signaling through a riboswitch.

33 We describe the role of CmrRST in colony morphology, swimming motility, biofilm formation, and a previously unknown mechanism of cellular migration. In an animal model of CDI, we find that CmrRST has a role in virulence and that cmrRST undergoes phase variation in vivo. Finally, we present data describing the transcriptional profiles of

CmrR and CmrT, allowing further investigation into the unique roles of each response regulator and the novel mechanisms by which they regulate motility. cmrRST as well as its upstream regulatory elements are highly conserved across C. difficile strains and ribotypes. Therefore, understanding the important role that this phase variable operon plays in C. difficile physiology may provide crucial information about the factors and regulatory strategies that contribute to pathogenesis.

34 CHAPTER 2: PHASE VARIATION OF A SIGNAL TRANSDUCTION SYSTEM CONTROLS CLOSTRIDIOIDES DIFFICILE COLONY MORPHOLOGY, MOTILITY, AND VIRULENCE1

INTRODUCTION

Phenotypic heterogeneity within bacterial populations is a widely established phenomenon that allows the population to survive sudden environmental changes [248-

250]. Heterogeneity serves as a “bet-hedging” strategy such that a subpopulation can persist and propagate an advantageous phenotype. Phase variation is one mechanism that imparts phenotypic heterogeneity, typically by controlling the ON/OFF production of a surface-exposed factor that directly interfaces with the environment [136, 158]. Many human pathogens, including pathogenic Escherichia coli, Neisseria meningitidis, and

Streptococcus pneumoniae, employ phase variation to evade the host immune system and increase host colonization, persistence, and virulence [136, 164, 251-257]. Phase variable phenotypes are heritable yet reversible, allowing the surviving subpopulation to regenerate heterogeneity. In phase variation by conservative site-specific recombination, a serine or tyrosine recombinase mediates the inversion of a DNA sequence adjacent to the regulated gene(s) [158]. The invertible DNA sequences are

1 This chapter previously appeared as an article in the journal PLoS Biology. Figure numbers have been modified. The original citation is as follows: Garrett EM, Sekulovic O, Wetzel D, Jones JB, Edwards AN, Vargas-Cuebas G, McBride SM, Tamayo R. (2019) Phase variation of a signal transduction system controls Clostridioides difficile colony morphology, motility, and virulence. PLoS Biol 17(10): e3000379. I generated all data with the exception of those found in Fig. 2.13A-B and Fig. S2.14A-B. I wrote and edited the original manuscript with feedback from Rita Tamayo as well revised it in response to reviewer comments.

35 flanked by inverted repeats and contain the regulatory information, such as a promoter, that when properly oriented controls gene expression in cis.

Clostridioides difficile (formerly Clostridium difficile) is a gram-positive, spore- forming, obligate anaerobe and a significant public health burden globally, causing gastrointestinal disease ranging from diarrhea to potentially fatal complications such as pseudomembranous colitis, toxic megacolon, bowel perforation, and sepsis. C. difficile pathogenesis is primarily driven by the toxins TcdA and TcdB, which inactivate host

Rho-family GTPases resulting in actin cytoskeleton dysregulation, tight junction disruption, and host cell death; consequently, TcdA and TcdB compromise the epithelial barrier and elicit diarrheal symptoms and inflammation [78, 258, 259]. Many aspects of

C. difficile physiology and pathogenesis remain poorly understood. For example, some

Clostridioides species, including C. difficile, are capable of forming two distinct colony morphologies; one is smooth and circular, and the other is rough and filamentous [103,

260-262]. However, the underlying mechanisms and physiological relevance of this dimorphism are unknown. Many bacterial species develop multiple colony morphologies as a result of the regulated production of surface factors, which can impact diverse physiological traits and behaviors [263-268]. In multiple species, the production of surface factors is subject to phase variation, leading to changes in gross colony morphology. In S. pneumoniae and Acinetobacter baumannii, phase variation of capsule polysaccharides leads to the formation of either opaque or translucent colonies that differ in a multitude of phenotypes including cell morphology, biofilm formation, antibiotic sensitivity, host colonization, and virulence [251, 253, 256, 266, 268, 269].

36 Phase variation of colony morphology is therefore an important adaptive strategy during infection for multiple pathogens.

The biological significance and mechanisms underlying colony morphology development of C. difficile have not been reported. C. difficile encodes multiple factors that are regulated through phase variation. In C. difficile, eight invertible DNA sequences, or “switches”, have been identified, seven of which are present in the epidemic strain R20291 [138, 185]. Two have been characterized: one controlling phase variation of the cell wall protein CwpV and the other controlling flagellar phase variation

[103, 181, 184, 270]. The phase variable flgB flagellar operon encodes the sigma factor

SigD, which coordinates flagellar gene expression and positively regulates the tcdA and tcdB toxin genes [101, 102, 237]. Consequently, the flagellar switch mediates phase variation of flagella and toxin production, highlighting the potential impact of phase variation on C. difficile physiology and virulence.

As in many bacteria, the intracellular signaling molecule c-di-GMP regulates the transition between community-associated and planktonic, motile lifestyles of C. difficile

[113, 114, 120, 235, 237]. This regulation occurs in part through direct control of gene expression by c-di-GMP via riboswitches [236, 241, 271, 272]. For example, a c-di-GMP riboswitch lies in the 5’ leader sequence of the flgB operon and causes transcription termination when c-di-GMP is bound, thus inhibiting swimming motility [104, 236, 271].

Because flagellum and toxin gene expression is linked through SigD, c-di-GMP also inhibits toxin production [101, 104]. Conversely, a c-di-GMP riboswitch upstream of the

Type IV pilus (TFP) locus allows c-di-GMP to positively regulate gene expression and promote TFP-dependent behaviors such as autoaggregation, surface motility, biofilm

37 formation, and colonization of host tissues [104, 113-115, 120, 237]. Additionally, c-di-

GMP regulates the expression and cell wall anchoring of surface proteins and adhesins

[228, 241, 273]. Therefore, c-di-GMP coordinates the expression of multiple surface factors with implications for pathogenesis.

In this study, we characterized rough and smooth colony isolates of C. difficile

R20291 and determined that they exhibit distinct motility behaviors in vitro. Colony morphology and associated motility phenotypes are controlled by both phase variation and c-di-GMP, and colony morphology is independent of TFP and flagella. We identified the c-di-GMP regulated, phase-variable signal transduction system, consisting of a putative histidine kinase and two DNA-binding response regulators, that modulates colony morphology, surface migration, and swimming motility. Finally, we provide evidence using a hamster model of infection for phase variation and a role for the

CmrRST system in virulence. The link between phase variation and c-di-GMP signaling to control CmrRST production suggests a mechanism for switching a global expression program during infection, which appear to have critical implications for disease development.

MATERIALS AND METHODS

Growth and maintenance of bacterial strains

C. difficile strains were maintained in an anaerobic environment of 85% N2, 5% CO2, and 10% H2. C. difficile strains were cultured in Tryptone Yeast (TY; 30 g/liter Bacto tryptone, 20 g/liter yeast extract, 1 g/liter thioglycolate) broth or Brain Heart Infusion plus yeast (BHIS; 37 g/liter Bacto brain heart infusion, 5 g/liter yeast extract) media at 37 °C

38 with static growth. Escherichia coli strains were grown in Luria Bertani medium at 37 °C with aeration. Antibiotics were used where indicated at the following concentrations: chloramphenicol (Cm), 10 µg/mL; thiamphenicol (Tm), 10 µg/mL; kanamycin (Kan), 100

µg/mL; ampicillin (Amp), 100 µg/mL. Table S1 lists descriptions of the strains and plasmids used in this study.

Differentiation of rough and smooth colonies

To differentiate rough and smooth colonies from wild-type C. difficile strains, 5 µL of liquid cultures were spotted onto BHIS agar plates and grown for 48-72 hours. For rough and smooth colonies, all growth was collected and plated in dilutions on BHIS agar. For predominantly rough colonies, growth was collected from the filamentous edges [237, 261]. For predominantly smooth colonies, growth was collected from the center of the spot (our stock of C. difficile R20291 consists primarily of bacteria yielding smooth colonies, so the spot center represents the inoculum). For enumeration of rough and smooth colonies, serial dilutions were plated, rough and total colony forming units

(CFU) were counted, and data were expressed as percent rough CFU.

Microscopy

For whole colony imaging, colonies were grown on BHIS plates for 24 hours prior to imaging. For strains containing plasmids, the BHIS agar medium contained 10 µg/mL

Tm. For strains carrying pDccA, pDccAmut, pEAL, or pMC-Pcpr, 2 µg/mL nisin was included to induce gene expression. For strains with pCmrR, pCmrT, or mutant derivatives, anhydrotetracycline (ATc) was added at the indicated concentrations.

39 Colonies were imaged at 2-8x magnification using a Zeiss Stereo Discovery V8 dissecting microscope with a glass stage and light from the top.

For imaging of motile spots, R20291 WT, ΔcmrR, ΔcmrT, and pilB1-null strains were grown overnight in BHIS and then spotted on BHIS 1.8% agar 1% glucose. After

72 hours of growth, colonies were images using a Keyence BZ-X810 microscope.

For Gram stain imaging, R20291 WT rough or smooth, ΔcmrR, or ΔcmrT colonies were differentiated as previously described and heat fixed on a glass slide.

Cells were Gram-stained (BD Kit 212524) and visualized at 40-80x magnification using an Olympus BX60 compound microscope or a Keyence BZ-X810 microscope. For quantification of cell length, at least two images from two biological replicates were analyzed using ImageJ.

For visualization by scanning electron microscopy, R20291 rough and smooth isolates and the sigD mutant were grown overnight in TY broth and washed once with phosphate buffered saline (PBS). For visualization of colony structure, R20291 was spotted on BHIS agar and grown for 3 days. The colony was excised on agar and fixed.

All samples were fixed in 4% glutaraldehyde in 150 mM sodium phosphate buffer. The samples were dehydrated through increasing concentrations of ethanol and dried using carbon dioxide as the transition solvent in a Samdri-795 critical-point dryer (Tousimis

Research Corporation, Rockville, MD). Coverslips were mounted on aluminum stubs with carbon adhesive tabs, followed by a 5-nm thickness platinum sputter in a Hummer

X sputter coater (Anatech USA, Union City, CA). Images were taken with a working distance of 5 mm and a 130-μm aperture using a Zeiss Supra 25 field emission scanning electron microscope (FESEM) operating at 5 kV (Carl Zeiss SMT Inc.,

40 Peabody, MA). For quantification of cell length, at least seven images from two biological replicates were analyzed using ImageJ.

Quantitative PCR analysis of invertible element orientations

Quantitative PCR (qPCR) was used to quantify the percent of the population with each of the seven invertible elements in a given orientation as described previously [138].

Rough and smooth colonies were collected from a BHIS plate, and genomic DNA was purified. qPCR was performed with SensiMix SYBR (BioLine). Twenty μL reactions with

100 ng of genomic DNA and 100 nM primers were used. Reactions were run on

Lightcycler 96 system (Roche) with the following three-step cycling conditions: 98°C for

2 min, followed by 40 cycles of 98°C for 30 s, 60°C for 1 min, and 72°C for 30 sec.

Quantification was done using the ΔΔCT method, with rpoA as the reference gene and the indicated reference condition [138]. All primers used in this and other experiments are listed in Table S2.

Construction of ∆cmrR and ∆cmrT in C. difficile R20291

Markerless deletions of cmrR (locus tag CDR20291_3128) and cmrT (locus tag

CDR20291_3126) were done using a previously described codA-based allelic exchange method with minor modifications [63]. Briefly, approximately 1100-1300 bp genomic fragments were PCR-amplified upstream and downstream of cmrR and cmrT with the following primers sets: OS158 and OS266 (cmrR, upstream), OS163 and OS267 (cmrR, downstream); OS268 and OS269 (cmrT, upstream), OS271 and OS283 (cmrT, downstream). Complementary overlapping sequences were added to the 5-prime end of

41 primers to allow for accurate fusion of all PCR products into PmeI-linearized pMTL-

SC7215 vector using Gibson Assembly Master Mix (New England BioLabs). The resulting plasmids were then conjugated into C. difficile R20291 strain (GenBank accession: CBE06969.1) using the heat-stimulated conjugation method described elsewhere [274]. Mutants were selected as previously described and screened by colony-PCR for the presence of the correct left and right homology-chromosomal junction and the absence of respective cmrR and cmrT coding sequence [63]. All PCR products were Sanger-sequenced to confirm the correct genetic construct and the absence of any secondary mutations.

Surface and swimming motility assays

For surface motility assays, 5 µL of overnight (16-18hr) cultures were spotted onto

BHIS-1.8% agar supplemented with 1% glucose [235]. For swimming motility assays, 1

µL of overnight cultures was inoculated into 0.5x BHIS-0.3% agar [103]. For plasmid- bearing strains in both assays, the medium was supplemented with 10 µg/mL Tm.

Where indicated, ATc was included at the indicated concentrations to induce expression. At 24 hour intervals the diameters of growth were taken as the average of two perpendicular measurements. For measurements of surface colony diameters, the first measurement corresponds to the widest part of the colony including any tendrils, and a perpendicular measurement regardless of the presence of a tendril. These values were averaged and used as the colony diameter as described previously [113]. Data from at least 8 biologically independent cultures were analyzed using a one-way

42 ANOVA to determine statistical significance. The plates were photographed using a

Syngene G:Box imager.

Biofilm Assay

Biofilm assays were done as previously described [120]. Overnight cultures of C. difficile were diluted 1:100 in BHIS 1% glucose 50 mM sodium phosphate buffer (pH 7.5) in 24- well polystyrene plates. After 24 hours of growth, supernatants were removed, the biofilms were washed once with PBS and then stained for 30 minutes with 0.1% crystal violet. After 30 minutes, the biofilms were washed again with PBS, and the crystal violet was solubilized with ethanol. Absorbance was read at 570 nm. Four independent experiments were performed.

RNA isolation and real-time PCR

For quantitative real-time PCR (qRT-PCR) analysis of flagellar gene expression, C. difficile with vector or cmrR/cmrT expression plasmids was grown for 48 hours in 0.5x

BHIS-0.3% agar to express flagellar genes. Bacteria were recovered and cultured in TY broth with 10 ng/mL ATc for vector and pCmrR; 2 ng/mL ATc for pCmrT. For qRT-PCR analysis of toxin gene expression, C. difficile with vector or cmrR/cmrT expression plasmids were grown overnight in TY broth and diluted into BHIS broth with 10 ng/mL

ATc for vector and pCmrR; 2 ng/mL ATc for pCmrT. For both experiments, samples were collected at mid-exponential phase and preserved in 1:1 ethanol-acetate. For qRT-

PCR analysis of cmrR and cmrT transcript abundance in rough and smooth isolates, rough and smooth colonies were differentiated as described above and streaked on

43 BHIS agar. After 24 hours growth, colonies were collected and preserved in 1:1 ethanol- acetate.

RNA was isolated, treated with DNase I, and reverse transcribed including a no- reverse transcriptase control as previously described [184, 237]. Real-time PCRs were done using 2 ng of cDNA, primers at a final concentration of 500 nM, and SYBR Green

Real-Time qPCR reagents (Thermo Fisher) with an annealing temperature of 55°C.

Primers used were as follows: R856-R857, flgB; R858-R859, flgM; R1063-R1064, fliC;

R930-R931, pilA1; R852-R853, tcdA; R2298-R2299, cmrR; and R2537-R2538, cmrT).

The data were analyzed using the ΔΔCt method with rpoC (R850-R851) as the reference gene [103].

Animal experiments

All animal experimentation was performed under the guidance of veterinarians and trained animal technicians within the Emory University Division of Animal Resources

(DAR). Animal experiments were performed with prior approval from the Emory

University Institutional Animal Care and Use Committee (IACUC). C. difficile spores were collected from 70:30 sporulation broth after 3 days of growth [275]. The spores were purified using a sucrose gradient and stored in PBS with 1% bovine serum albumin as described previously [115, 276]. Prior to inoculation, the spores were enumerated by plating serial dilutions on BHIS-agar containing 0.1% sodium taurocholate.

Male and female Syrian golden hamsters (Mesocricetus auratus, Charles River

Laboratories) were housed individually in sterile cages and given a standard rodent diet

44 and water ad libitum. To render the animals susceptible to C. difficile, one dose of clindamycin (30 mg kg−1 of body weight) was administered by oral gavage seven days prior to inoculation. Hamsters were inoculated with ~5000 spores of a single strain of C. difficile [276, 277]. Uninfected controls treated with clindamycin were included in each experiment. Hamsters were weighed at least daily and monitored for signs of disease

(weight loss, lethargy, diarrhea and wet tail). Fecal samples were collected daily for determination of bacterial burden (see below). Hamsters were considered moribund if they lost 15% or more of their highest weight or if they showed disease symptoms of lethargy, diarrhea and wet tail. Animals meeting either criterion were euthanized by

CO2 asphyxiation and thoracotomy. Immediately following euthanasia, animals were necropsied, and cecal contents were obtained for enumeration of total C. difficile CFU and for subsequent DNA isolation. To enumerate C. difficile CFU, fecal and cecal samples were weighed, suspended in 1X PBS, heated to 60°C for 20 min to minimize growth of other organisms, and plated on taurocholate cycloserine cefoxitin fructose agar (TCCFA; [278, 279]). C. difficile CFU were enumerated after 48 hours. Six animals

(3 male, 3 female) per C. difficile strain were tested, and the data were analyzed using

GraphPad Prism.

Semi-quantitative PCR and qPCR analysis of DNA from hamster cecums

Hamster cecal contents in PBS were treated with lysozyme and subjected to bead beating to lyse spores. DNA was purified by phenol:chloroform:isopropanol extraction and washed with ethanol. For semi-quantitative PCR, 0.5 ng DNA per 50 µL reaction was PCR amplified using 0.5 μM orientation-specific primers to detect the orientation of

45 the cmr switch (Cdi6): RT2378-RT2379 for the ON orientation; RT2378-RT2197 for the

OFF orientation. For qPCR, 4 ng DNA per 20 µL reaction was used with 0.5 μM orientation-specific primers.

Sporulation assays

C. difficile cultures were grown in BHIS medium supplemented with 0.1% taurocholate and 0.2% fructose until mid-exponential phase (i.e., OD600 of 0.5), and 0.25 ml aliquots were plated onto 70:30 agar supplemented with 2 µg/ml thiamphenicol [275]. After 24 h growth, ethanol resistance assays were performed as previously described [280, 281].

Briefly, the cells were removed from plates after 24 h (H24) and suspended in BHIS medium to an OD600 of ~1.0. The number of vegetative cells per milliliter was determined by immediately serially diluting and plating the suspended cells onto BHIS.

Simultaneously, a 0.5-ml aliquot was added to 0.3 ml of 95% ethanol and 0.2 ml of dH2O to achieve a final concentration of 28.5% ethanol, vortexed, and incubated for 15 min to eliminate all vegetative cells; ethanol-treated cells were subsequently serially diluted in

1× PBS with 0.1% taurocholate and applied to BHIS plus 0.1% taurocholate plates to determine the total number of spores. After at least 24 h of growth, CFU were enumerated, and the sporulation frequency was calculated as the total number of spores divided by the total number of viable cells (spores plus vegetative cells).

A nonsporulating mutant (MC310; spo0A::erm) was used as a negative control.

Germination assays

C. difficile strains were grown in 500-ml liquid 70:30 medium and spores were purified for germination studies as previously described, with some modifications [277, 282-

46 284]. Cultures in sporulation medium were removed from the anaerobic chamber after

120 h of anaerobic growth and kept outside of the chamber in atmospheric oxygen overnight. Spore cultures were collected by centrifugation, suspended in water, and frozen for 15 min at -80°C. After thawing, spore suspensions were centrifuged for 10 min at ~3200 x g in a swing bucket rotor, and the supernatant was discarded. Spore pellets were washed two times with water then suspended in 1 ml of a 1x PBS + 1%

BSA solution, applied to a 12 ml 50% sucrose bed volume, and centrifuged at ~3200 x g for 20 min. The supernatant was decanted and the spores were checked by phase- contrast microscopy for purity. Sucrose gradients were repeated until the preparations reached a purity of greater than 95%. Spore pellets were then washed three times with

1x PBS + 1% BSA and suspended to an OD600 = 3.0. Germination assays were carried out as previously described, with slight modifications [282, 283]. After treatment of spores for 30 min at 60°C, spore germination was analyzed in BHIS containing 10 mM taurocholate and 100 mM glycine. The OD600 was determined immediately and at various time points during incubation at room temperature. Results are presented as means and standard errors of the means of three independent biological replicates.

Detection of TcdA by western blot

Rough and smooth isolates as well as cmrR and cmrT mutants were grown for 24 hours in TY broth and normalized to OD600 ~ 1.8 prior to collection. Cells were pelleted, suspended in SDS-PAGE buffer, and boiled for 10 minutes. Samples were then run on

4–20% Mini-PROTEAN TGX Precast Protein Gels (Bio Rad) and transferred to a nitrocellulose membrane. Membranes were stained with Ponceau S stain (Sigma

47 Aldrich) to assess sample loading. TcdA was detected as described previously using a mouse α-TcdA primary antibody (Novus Biologicals) and goat anti-mouse IgG conjugated with IR800 (Thermo Fisher) [42].

RESULTS

C. difficile strains of diverse ribotypes develop two distinct, phase-variable colony morphotypes

The C. difficile strain R20291, a ribotype 027 strain associated with epidemic infections, exhibits two distinct colony morphologies: a “smooth” colony that is round and circular, and a “rough” colony that is flatter and has filamentous edges [103, 261,

262]. In addition to R20291, strains UK1 (ribotype 027), VPI10463 (ribotype 087), 630

(ribotype 012), and ATCC 43598 (ribotype 017) yielded both rough and smooth colonies

(Figure 2.1). Some of these strains (R20291 and UK1) showed both colony morphotypes through routine plating, while others (630) required extended incubation to allow the appearance of bacteria that develop rough colonies. ATCC BAA 1875

(ribotype 078) did not develop smooth colonies under any conditions tested. Therefore, many strains are capable of generating two distinct colony morphologies in vitro.

We and others previously observed that spotting a culture of C. difficile R20291 on an agar surface results in expansion of the colony as dendritic, filamentous growth

[237, 261]. We discovered that this culturing method allows the segregation of the two colony morphologies. Streaking from the center of the colony yielded a population of predominantly smooth colonies, while streaking from the edges yielded a population of predominantly rough colonies (Figure 2.2A). To evaluate the stability of the colony

48 phenotypes, we passaged rough and smooth colonies through multiple growth conditions. Streaking a rough or smooth colony again onto agar medium resulted in overall maintenance of the original morphology (Figure 2.2A,C), indicating that colony morphology is heritable and generally stable under specific growth conditions. Similarly, the starting morphology was maintained when the isolates were passaged in broth

(Figure 2.2C). Because surface growth promoted the development of the rough morphotype, we speculated that conditions favoring swimming motility might yield smooth colonies. Rough and smooth colonies were inoculated into soft (0.3%) agar to allow for swimming motility over 48 hours (Figure 2.2B, panel 4). Growth recovered from the edges of the motile growth solely yielded smooth colonies, regardless of the starting morphology, indicating that colony morphology is reversible (Figure 2.2B,C). These data support that C. difficile colony morphology phase varies and reveal in vitro growth conditions that select for a specific phase variant: expansion of a colony on an agar surface favors bacteria that yield rough colonies, while swimming conditions select for bacteria that yield exclusively smooth colonies.

Rough and smooth colony isolates exhibit distinct motility behaviors

The observation that surface growth favors bacteria that develop rough colonies suggests that the rough morphotype has an advantage over smooth in this growth condition. Conversely, smooth morphotype bacteria may have an advantage over the rough in swimming motility. To test this, we assessed the motility behaviors of rough and smooth colony isolates. Expansion of colonies of “wild-type” (i.e. undifferentiated by colony morphology) R20291 on an agar surface is characterized by initial (~ 24 hours) growth restricted to the site of inoculation, then spreading tendrils of growth between 24

49 and 96 hours [113]. Type IV pili were previously shown to contribute to motility on an agar surface, so the non-piliated pilB1 mutant was included as a control [113, 114]. In this assay, bacteria isolated from rough colonies exhibited greater colony expansion after 48 hours compared to undifferentiated (wild type) R20291 (Figure 2.2D,E).

Conversely, bacteria derived from smooth colonies were deficient in colony expansion compared to the rough isolates and more similar to the pilB1 control after 48 hours

(Figure 2.2D,E), though the smooth isolates remained capable of colony expansion.

Flagellum-dependent swimming motility was assayed by inoculating bacteria into

0.3% agar and measuring migration through the medium. Undifferentiated R20291 and a non-motile sigD mutant served as controls [103]. Rough colony isolates showed significantly decreased swimming motility compared to smooth and undifferentiated populations (Figure 2.2F,G). Bacteria from smooth colonies showed swimming motility comparable to the undifferentiated parent (Figure 2.2F,G), likely because the parental isolate consisted primarily of bacteria that yield smooth colonies. These results indicate that rough and smooth colony morphotypes have distinct motility behaviors.

Colony morphology is regulated by a phase variable signal transduction system and c-di-GMP

Because colony morphology development exhibited characteristics suggestive of phase variation, we sought to identify the underlying mechanism. To date, the only known phase variation mechanism in C. difficile involves site-specific DNA recombination resulting in a reversible DNA inversion that is known or predicted to impact gene expression in cis [138]. We postulated that one of the seven known invertible sequences in R20291 regulates the expression of genes involved in colony

50 morphology development, which would be evident as a correlation between the orientation of the invertible sequence and colony morphology (Figure 2.3A). To test this idea, we differentiated rough and smooth populations and analyzed the orientation of the “switches” in these isolates using qPCR with orientation-specific primers. For six of the switches, no correlation between morphology and sequence orientation was observed (Figure 2.3B). However, the orientation of the invertible element Cdi6, which is upstream of the operon CDR20291_3128-3126, showed a strong correlation with colony morphology [138, 270]. Each of four independently isolated rough populations contained the sequence predominantly in the orientation previously suggested to favor gene expression, while each of the smooth populations contained the sequence predominantly in the inverse orientation [138]. CDR20291_3128-3126 encodes a putative phosphorelay system consisting of two predicted response regulators and a predicted histidine kinase [138]. Accordingly, we named the operon colony morphology regulators RST, where cmrR and cmrT encode the response regulators and cmrS encodes the histidine kinase, and we refer to the Cdi6 invertible element as the “cmr switch”.

Previous work showed that expression of the cmrRST locus is heterogeneous in a population of C. difficile R20291, occurring in an on/off manner consistent with phase variation [138]. We therefore used qRT-PCR to examine the impact of the orientation of the cmr switch on downstream gene expression in rough and smooth isolates.

Expression of cmrR and cmrT was significantly increased in rough isolates relative to smooth (Figure 2.3C). Combined with the results of orientation-specific PCR of these isolates, these results indicate that bacteria that yield rough colonies bear the cmr

51 switch in the “on” orientation, while the bacteria yielding smooth colonies contain the switch in the “off” orientation. The cmrRST locus and its upstream regulatory region containing the cmr switch are present in all 65 complete C. difficile genomes available on NCBI with > 96% identity at the nucleotide level. This high conservation across diverse ribotypes indicates that this regulatory system, as well as the ability to control its production through phase variation, is important for C. difficile fitness.

A c-di-GMP binding riboswitch (Cdi-2-2) lies 5’ of cmrRST and the invertible sequence and positively regulates expression of the operon (Figure 2.3A) [236, 241].

Additionally, rough and smooth colony morphotypes exhibited inverse swimming and surface motility phenotypes, which is characteristic of c-di-GMP regulation [104, 113].

Therefore, we hypothesized that c-di-GMP regulates C. difficile colony morphology.

Using a previously described strategy, intracellular c-di-GMP was increased or decreased by ectopically expressing genes encoding the diguanylate cyclase DccA or the EAL domain from PdcA, respectively [113, 237]. Expression of dccA led to the uniform development of rough colonies, while the expression of a catalytically inactive

DccA (DccAmut) did not, indicating that increased c-di-GMP stimulates rough colony formation (Figure 2.3D). Reducing c-di-GMP through overproduction of the EAL domain did not impact colony morphology, possibly because basal c-di-GMP levels are already low [237]. To determine whether the rough colony effect results from c-di-GMP- mediated inhibition of flagellar motility or activation of TFP-dependent surface motility, dccA was expressed in TFP-null (pilB1::erm) or aflagellate (sigD::erm) backgrounds.

The pilB1 and sigD mutants remained capable of forming both smooth and rough colonies in unmodified and increased c-di-GMP conditions (Figure 2.3E,F). Together,

52 these results indicate dual regulation of the cmrRST locus by both phase variation and c-di-GMP. Furthermore, c-di-GMP regulates colony morphology through a mechanism independent of TFP and flagella. To distinguish the spreading colonies developed by rough isolate bacteria from the surface motility imparted by TFP, we refer to the

CmrRST-mediated phenomenon as “surface migration”.

The response regulators CmrR and CmrT regulate colony morphology

Because activation of cmrRST expression, whether by increased c-di-GMP and/or inversion of the switch to the “on” orientation, promotes the development of rough colonies, we examined the contributions of the response regulators CmrR and

CmrT to this process. The cmrR and cmrT genes were individually expressed under the control of an ATc-inducible promoter in C. difficile R20291 during growth on BHIS-agar.

Both CmrR and CmrT stimulated rough colony formation compared to non-induced controls (Figure 2.4). Expression of cmrT resulted in a rough phenotype at lower levels of induction than cmrR. Interestingly, some levels of induction of cmrT inhibited growth, suggesting that cmrT expression is toxic. Growth inhibition by CmrT, but not CmrR, was observed in broth culture as well (Figure 2.5A,B).

Response regulators typically have a conserved aspartic acid residue in the phosphoreceiver domain that, when phosphorylated, leads to activation [285]. Mutation of this residue to a glutamic acid often mimics phosphorylation [286, 287]. Accordingly, compared to cmrR expression, lower levels of induction of cmrRD52E expression were required for formation of rough colonies (0.5 ng/ml versus 4 ng/ml ATc, respectively;

Figure 2.4). Instead of a conserved aspartic acid residue, CmrT contains a glutamic acid at the expected phosphorylation site, precluding the need for a phosphomimic

53 substitution, and potentially explaining the lower level of inducer needed to yield rough colonies compared to cmrR. Substitution of the phosphorylation site of CmrT with alanine (CmrT-D53A) increased the concentration of inducer needed to obtain rough colonies. However, for the wild type and CmrR-D52A mutant, comparable levels of inducer (4 ng/mL ATc) resulted in rough colonies; this may be due to the need for an activating signal that is absent in these assay conditions. That the inactivating mutations did not abolish CmrR and CmrT function suggests that phosphorylation is not required for the activity of these response regulators if they are expressed at high levels.

Because CmrR and CmrT promoted the rough colony morphotype, we examined the requirement of cmrR and cmrT in rough colony formation. Individual in-frame deletions of cmrR and cmrT were generated in C. difficile R20291. The mutants and undifferentiated wild-type R20291 parent were grown on BHIS-agar to allow selection of rough colony isolates from the edges (as in Figure 2.2). While the parent strain and cmrR mutant formed both rough and smooth colonies, the cmrT mutant did not form rough colonies under the given conditions (Figure 2.6A). Expression of cmrT in trans complemented the cmrT mutation, restoring the capacity to form rough colonies (Figure

2.6B). Importantly, cmrR and cmrT mutants do not differ in growth rate from WT (Figure

2.5C). Thus, while CmrR and CmrT both impact colony morphology development when overproduced, only CmrT is required for rough colony formation under the tested conditions.

CmrR and CmrT inversely regulate swimming motility and surface migration

Because bacteria isolated from rough and smooth colonies differed in surface migration and swimming motility (Figure 2.2), we assessed the roles of CmrR and CmrT

54 in these processes. Ectopic expression of cmrR or cmrT in R20291 significantly increased surface migration compared to uninduced and vector controls (Figures 2.7A,

2.8A,B). Enhanced surface migration was also observed in a TFP-deficient (pilB1::erm) background, again indicating that CmrR and CmrT-mediated surface migration is independent of TFP (Figures 2.7A, 2.8A,B). Conversely, cmrR and cmrT expression inhibited swimming motility of R20291 (Figures 2.7D, 2.8C,D). These data are consistent with the increased surface migration and decreased swimming motility exhibited by rough colony isolates (cmrRST ON), which are the inverse of the motility phenotypes of the smooth colony isolates (cmrRST OFF). Expression of cmrR and cmrT alleles with inactivating alanine substitutions resembled the wild-type allele, suggesting again that high levels of expression overcome the need for phosphorylation

(Figure 2.8). However, as seen with colony morphology (Figure 2.4), the cmrR-D52E allele increased surface migration and decreased swimming motility at lower levels of induction relative to the wild-type allele, suggesting that the amino acid substitution increased activity (Figure 2.8). Expression of cmrR or cmrT did not alter transcript levels of representative flagellum or TFP genes (Figure 2.9), suggesting that the observed differences in surface migration and swimming motility occur through post- transcriptional regulation or through an alternative mechanism.

To further define the role of cmrRST, the cmrR and cmrT mutants and relevant controls were evaluated for surface migration and swimming motility. In both assays, the cmrR mutant behaved comparably to the R20291 parent (Figure 2.7B,C,E,F), indicating that CmrR is dispensable for these processes. However, the cmrT mutant exhibited significantly reduced surface migration, a defect that was complemented by expressing

55 cmrT in trans. The opposite effect was observed for swimming motility; the cmrT mutant showed greater motility compared to the parent strain. CmrT thus appears to be the dominant response regulator of the CmrRST system for control of rough colony formation, surface migration, and swimming motility in vitro.

CmrR and CmrT promote bacterial chaining

Phenotypic analysis of motility and quantification of flagellum and TFP transcripts indicates that CmrRST mediated colony morphology occurs independently of these surface structures. Cell morphology has also been shown to affect gross colony morphology [268, 288, 289], so we investigated whether cell morphology differs between rough and smooth morphotypes. Examination of C. difficile colonies by scanning electron microscopy (SEM) revealed the presence of the expected bacilli towards the center, as well as elongated cells particularly along the edge of the colony

(Figure 2.10A). These elongated cells appeared as organized bundles corresponding to the tendrils apparent in the macrocolony, suggesting a role for the elongated bacteria in expansion of rough colonies. Consistent with this, SEM of bacteria derived from smooth and rough colonies revealed dichotomous cellular morphologies. Smooth colony- derived bacteria appeared as standard bacilli (4.089 µm ± 1.207 in length), while rough colony-derived cells were longer (7.428 µm ± 4.130) and sometimes exhibited an extremely elongated shape, resembling a bacterial filament or chain (Figure 2.10B,C).

To determine whether the elongated cells result from filamentation or cell chaining, cells from rough and smooth colonies were Gram stained. While cells from smooth colonies consisted of the expected single or double rods, cells from rough colonies more commonly appeared in bacterial chains of three or more cells (Figure 2.10E,F). Septa

56 separating cells within chains were clearly visible, differentiating this morphology from filamentation. Differences in processing may explain why cell chains were more common in Gram stained samples than in SEM imaged samples; processing of samples for SEM may have disrupted chains.

Since cmrRST expression favors rough colony development, we examined the roles of CmrR and CmrT in the chaining phenotype. In contrast to uninduced and vector controls, bacteria over-expressing cmrR or cmrT also displayed cell chaining (Figure

2.10G). Based on our surface motility data, we predicted that though CmrR and CmrT both contribute to cell chaining and elongation, only the ΔcmrT mutant would be deficient in these phenotypes. We imaged Gram stained WT, ΔcmrR, and ΔcmrT cells collected from the edges of motile spots (Figure 2.11A,B). The ΔcmrR mutant displayed a slight decrease in cell length as compared to the WT control but still appeared as elongated cells and cell chains. However, ΔcmrT cells were significantly shorter than

WT and appeared only as single or double cells. Together, these results indicate that rough colonies contain bacterial chains whose formation is promoted by CmrRST.

To better understand the contribution of cell chaining to the expansion of a colony on a surface, we visualized WT, ΔcmrR, and ΔcmrT grown on agar. WT, ΔcmrR, and TFP-deficient strains displayed elongated cells in bundled cell chains, and these structures appeared to expand in an ordered manner away from the colony edge

(Figure 2.11C). However, the ΔcmrT strain, which is defective in surface motility, did not display these long cell bundles. Together, these data suggest that CmrT-mediated changes in cell morphology and chaining contribute to the movement of C. difficile across a surface.

57 Bacterial motility, cell morphology, and adherent behaviors are central to biofilm development, and chaining and filamentation can also affect surface adhesion and biofilm formation [290-293]. We therefore examined the role of CmrRST in C. difficile biofilm formation. Bacteria isolated from rough and smooth colonies and the cmrR and cmrT mutants were assayed for biofilm production in rich medium on plastic as described previously [113, 120]. The cmrR mutant showed significantly increased biofilm formation compared to the rough isolate, while the cmrT mutant showed no significant difference (Figure 2.12), suggesting that CmrR negatively regulates biofilm formation.

Phase variable colony morphology via CmrRST impacts C. difficile virulence

Given the broad role of CmrRST in controlling C. difficile physiology and behaviors, we predicted that the cmrR or cmrT mutants would have altered virulence in a hamster model of C. difficile infection. Male and female Syrian golden hamsters were inoculated with spores of ΔcmrR or ΔcmrT, or smooth or rough isolates of R20291, which remain capable of phase varying cmrRST expression. Notably, the strains tested did not differ in sporulation or germination rates in vitro (Figure 2.13). The animals were monitored for disease symptoms and euthanized when moribund as described in the

Materials and Methods. Hamsters infected with ΔcmrR showed increased survival compared to those infected with the rough isolate of the R20291 parent (P = 0.003, log- rank test). Survival of hamsters infected with ΔcmrR was also greater than that of animals infected with the smooth isolate, but the difference did not reach statistical significance (P = 0.098, log-rank test). While the mean times to morbidity were comparable for the animals that did succumb (44.08 hours for cmrR, 48.15 ± 12.87

58 hours for rough, 54.69 ± 22.90 hours for smooth), most of the animals inoculated with

ΔcmrR survived (Figure 2.14A). For animals infected with ΔcmrT, both time to morbidity and percent animal survival were equivalent to those infected with the rough and smooth isolates.

To evaluate intestinal colonization, fecal samples were collected every 24 hours post-inoculation, and dilutions were plated on TCCFA medium (containing the spore germinant taurocholic acid) to enumerate C. difficile colony forming units (CFU). All animals infected with ΔcmrT, rough, and smooth isolates yielded detectable CFU at the time of collection preceding disease development and euthanasia (Figure 2.14B).

However, we observed no correlation between bacterial load and disease onset or severity. The number of ΔcmrR CFU was never above the limit of detection in feces of infected animals (Figure 2.14B), though C. difficile was detected in the cecum of the

ΔcmrR-infected hamster that became moribund (data not shown). These data indicate that CmrR, but not CmrT, is important for the ability of C. difficile to colonize the intestinal tract and cause disease in the hamster model.

To determine whether the attenuation of the ΔcmrR mutant was due to reduced toxin levels, we evaluated the effect of CmrR and CmrT on TcdA production. Western blot analysis showed that ΔcmrR does not produce significantly lower levels of TcdA toxin than rough and smooth isolates in vitro, suggesting that the virulence defect of

ΔcmrR occurs through a toxin-independent mechanism (Figure 2.13C,E). Interestingly,

ΔcmrT produced slightly higher TcdA levels by western blot. However, overexpression of cmrR and cmrT in R20291 did not significantly alter tcdA transcript abundance, suggesting that CmrRST does not transcriptionally regulate toxin gene expression.

59 Because the ΔcmrR strain was attenuated in hamsters, we expected the smooth isolate (CmrRST “off”) to be attenuated compared to the rough isolate (CmrRST “on”).

While hamsters infected with the rough isolate all succumbed to disease, 33% of hamsters inoculated with the smooth isolate survived, suggesting moderately lower virulence of the smooth isolate (Figure 2.14A). However, this difference did not reach statistical significance as the survival times of hamsters that became moribund were not significantly different (P = 0.112, log-rank test), and both isolates were able to colonize

(Figure 2.14A, B).

We considered the possibility that the rough and smooth isolates, though chosen based on distinct colony morphologies, are “wild-type” and capable of phase variation.

Accordingly, selective pressures in the intestinal tract may have resulted in a phenotypic switch during infection, mitigating the observed differences in virulence. To test this, we determined the orientations of the cmrRST switch in the spore preparations and the contents of the cecums collected from infected animals immediately after sacrifice. DNA obtained from cecal contents was subjected to PCR with two primer sets that amplify from the cmrRST switch in a particular orientation. Analysis of the cmrRST switch orientation in the spore inoculums indicate that the rough isolate consisted of bacteria with the switch primarily in the ON orientation, while the smooth isolate contained those primarily in the OFF orientation, consistent with prior in vitro analyses (Figure 2.14C, D).

In the cecal contents of hamsters inoculated with the smooth isolate, C. difficile retained the cmrRST switch predominantly in the OFF orientation (94.56 ± 2.52% OFF; Figure

2.14D). In contrast, cecal contents from rough isolate-infected animals also contained the switch in a mixed or predominantly OFF orientation (75.71 ± 9.44% OFF), indicating

60 that that the rough (ON) isolate inoculum underwent phase variation during infection.

These data suggest that either ON/OFF switch rates are altered in favor of the OFF orientation during acute, short-term infection, and/or there is a selective pressure for the cmrRST OFF state and corresponding phenotypes.

DISCUSSION

Phase variation provides a mechanism for generating phenotypic heterogeneity in bacterial populations to enhance the survival of the population as a whole when encountering detrimental environmental stresses. In this study, we characterized the phase variation of a signal transduction system, CmrRST, which we found broadly impacts C. difficile physiology and behavior. During growth in vitro, C. difficile populations consist of bacterial subpopulations that are CmrRST ON, yielding rough colonies, and CmrRST OFF, yielding smooth colonies. These subtypes display divergent motility phenotypes in vitro, with smooth colony isolates showing increased swimming motility and rough colony isolates displaying enhanced surface motility. The identification of growth conditions that segregate for the respective morphological variants, each with enhanced motility in the corresponding growth condition, indicates a selective advantage for that motility behavior.

Although ectopic expression of either response regulator, CmrR or CmrT, increased surface motility, only CmrT is required for surface migration and rough colony development, suggesting that CmrT is the dominant regulator under these conditions.

Alternatively, CmrR may also play a role but requires an activating signal through the histidine kinase CmrS to be functional. The latter possibility is supported by the

61 observation that a phosphomimic substitution at the predicted phosphorylation site in the receiver domain increases CmrR activity. In contrast, because CmrT contains a glutamic acid at this site, it likely does not require activation by a histidine kinase and instead acts constitutively. For this reason, the receiver domain of CmrT may be considered a pseudo-receiver domain [285, 294].

In the hamster model of acute infection, the cmrR mutant did not detectably colonize and exhibited reduced virulence, indicating a role for this signal transduction system during infection. In contrast with the in vitro motility phenotypes driven primarily by CmrT, the CmrR regulator appears to be critical for virulence. The divergent roles for the regulators suggest that an activating signal for CmrR is present in the host intestine, though it remains possible that CmrT activity is somehow inhibited during infection. A similar mechanistic distinction between the regulators may explain the role of CmrR, but not CmrT, in biofilm formation in vitro. Whether CmrR and CmrT control distinct regulons, act cooperatively, and/or antagonistically is unknown and currently under investigation.

Compared to the strong attenuated virulence of the cmrR mutant, the rough and smooth isolates exhibited modest differences in virulence. The latter result may have arisen from the ability of the isolates to phase vary, resulting in convergent phenotypes due to selective pressures in the intestinal environment. Tracking of the cmr switch orientations in the inoculums and cecal contents at the experimental endpoint indicate that the rough variant (cmrRST ON) switched to populations with the cmr switch in a mixed or predominantly OFF orientation. That phase variation of CmrRST occurred during infection may have resulted in diminished differences in virulence between the

62 rough and smooth variants. In addition, these findings imply the presence of a selective pressure that favors the cmrRST OFF phase variant, at least in the later stages of infection. However, the high conservation of the cmrRST locus and the upstream regulatory region indicates that this regulatory system and its ability to phase vary confer a fitness advantage. The colonization defect of the cmrR mutant suggests that signaling through CmrR contributes to initial colonization, and therefore the ability to generate a CmrRST ON subpopulation may be important at an early stage of infection.

In our study, cecal contents were harvested at the experimental endpoint when disease was fulminant, and the collection of samples after more frequent intervals is limited by the short duration of the infection and low bacterial load prior to 24 hours post- inoculation. This sampling strategy may not capture transient changes in the cmr switch bias during different stages of infection, over a longer infection, or in a different location in the intestinal tract. Testing the impact of cmrR and cmrT mutations in a phase-locked

ON background as well as a double cmrR cmrT mutant may better reveal the role of this system in C. difficile virulence.

Unlike phase variation mechanisms that control the production of a specific surface factor, the phase variation of the CmrRST system is poised to coordinate the expression of multiple pathways. Previous studies have described phase variation of proteins with regulatory function resulting in broad changes in gene expression. A few transcriptional regulators have been reported to phase vary, including the global virulence gene regulators BvgAS in Bordetella and the response regulator FlgR in

Campylobacter jejuni [295-297]. In C. difficile, phase variation of flagellar genes impacts the transcription of the alternative sigma factor gene sigD, which couples the expression

63 of flagellum and toxin genes, linking virulence factor production to the orientation of the flagellar switch [103, 298]. Coordinated regulation of genes resulting in global changes in transcription can also result from phase variation of DNA modification systems, as observed in Haemophilus influenza, Helicobacter pylori, Salmonella enterica, and

Streptococcus strains [143, 299-303]. In Streptococcus pneumoniae, for example, DNA inversions within three genes encoding the sequence recognition factor of a Type I restriction-modification (RM) system allow for changes in genomic methylation patterns, resulting in phase variable regulation of virulence factors including capsule [154, 303].

The CmrR and CmrT response regulators each contain DNA binding domains and have the capacity to regulate gene transcription. Therefore, inversion of the cmr switch likely also indirectly controls the expression of multiple genes in a phase variable manner.

The CmrRST system regulates multiple processes, but the basis for the phenotypic changes are unclear. Colony morphology and motility assays indicate TFP- and flagellum-independent mechanisms, and the transcription of TFP and flagellum genes is not affected by CmrR and CmrT. TFP-independent surface migration has not previously been described in C. difficile. Examination of cellular morphology indicates that CmrRST promotes bacterial chaining. Bacillus subtilis also exhibits cell chaining, and this chaining may contribute to bacterial migration as cells are pushed along the axis of growth and division [304-307]. In support of this hypothesis, the edges of C. difficile motile spots exhibit well-organized bundled chains of elongated cells that extend away from the center and are absent without cmrT. This mechanism appears similar to gliding motility in Clostridium perfringens, although that process requires TFP [110,

308]. Bacterial chaining is observed in a number of bacterial species and can affect

64 flagellar motility, surface adhesion, biofilm formation, susceptibility to phagocytosis, and virulence [289-293, 304, 309-312]. Chaining of C. difficile through regulation by

CmrRST may similarly alter motility behaviors and persistence in the intestine. How

CmrRST mediates changes to cell chaining, colony morphology, and motility is unknown, and ongoing work to define the CmrRST regulon may reveal the underlying mechanisms.

The complexity of cmrRST regulation not only alludes to the importance of controlling CmrRST activity, but also highlights the growing link between phase variation and c-di-GMP signaling in C. difficile. Expression of flagellar genes is also modulated by both a c-di-GMP binding riboswitch and an invertible element [103, 104, 271], as well as two additional phase variable genes in C. difficile that encode c-di-GMP phosphodiesterases [138]. In the case of flagellar gene regulation, c-di-GMP inhibits transcription through a riboswitch that induces transcription termination, and the invertible flagellar switch further controls expression through an uncharacterized post- transcriptional mechanism. For cmrRST expression, c-di-GMP positively regulates transcription, and the cmr switch controls phase variation through an undefined mechanism [138, 236]. The hierarchy between these regulatory elements is unclear. A

σA-dependent promoter lies 5’ of both the riboswitch and cmr switch [241]. The invertible sequence may contain an additional promoter to drive expression when properly oriented [136, 159]. Alternatively, the upstream σA-dependent promoter may be the only site of transcriptional initiation, and the invertible element may contain another regulatory sequence that modulates expression [179].

65 Finally, this and other recent work underscores that genetically clonal C. difficile strains, as well as isogenic mutants, may differ phenotypically due to phase variation.

The seven sites of site-specific DNA recombination in the C. difficile genome can be inverted independently, and this study shows that in vitro culturing conditions can change the biases of switch orientations in the overall population. From a research standpoint, this can introduce significant variability into otherwise controlled experiments. This concern may be especially significant in infection studies, where differences in switch biases in inoculums may skew apparent outcomes. Caution should be taken in interpreting such data, as observed differences may be due to differences in expression of phase variable traits rather than the targeted mutations.

Through phase variation, populations of bacterial pathogens can rapidly adapt to changing environmental pressures, balancing the need to move, adhere, and avoid immune recognition through the course of infection. The phase variation of CmrRST may allow alternating modes of surface or swimming motility in C. difficile as needed during infection –swimming motility (cmrRST OFF) may allow exploration and colonization of new sites, while bacterial chaining (cmrRST ON) may allow the spread of bacteria along the epithelial surface once contact has been made. Further work understanding the role of CmrRST will provide important insights into C. difficile pathogenesis.

Ethics Statement

All animal experimentation was performed under the guidance of veterinarians and trained animal technicians within the Emory University Division of Animal Resources

66 (DAR). Animal experiments were performed with prior approval (approval number

201700396) from the Emory University Institutional Animal Care and Use Committee

(IACUC) under protocol #DAR-2001737-052415BA. Animals considered moribund were euthanized by CO2 asphyxiation followed by thoracotomy in accordance with the Panel on Euthanasia of the American Veterinary Medical Association. The University is in compliance with state and federal Animal Welfare Acts, the standards and policies of the Public Health Service, including documents entitled "Guide for the Care and Use of

Laboratory Animals" - National Academy Press, 2011, "Public Health Service Policy on

Humane Care and Use of Laboratory Animals" - September 1986, and Public Law 89-

544 with subsequent amendments. Emory University is registered with the United

States Department of Agriculture (57-R-003) and has filed an Assurance of Compliance statement with the Office of Laboratory Animal Welfare of the National Institutes of

Health (A3180-01).

Acknowledgments

We thank the University of North Carolina (UNC) Microscopy Services Laboratory for assistance with scanning electron microscopy. The Microscopy Services Laboratory,

Department of Pathology and Laboratory Medicine, is supported in part by P30

CA016086 Cancer Center Core Support Grant to the UNC Lineberger Comprehensive

Cancer Center.

67 FIGURES AND TABLES

Figure 2.1. Formation of rough and smooth colonies by multiple C. difficile strains. C. difficile strains representing five different ribotypes (indicated in parentheses) were grown on BHIS-agar medium for 72 hours to allow differentiation of colony morphology. Growth was recovered and plated on BHIS-agar for visualization of individual colonies. Shown are representative images from two independent experiments. Black triangles indicate “smooth” colonies, and white triangles indicate

“rough” colonies. All images were taken at 2X magnification.

68

Figure 2.2. Reversible selection of distinct colony morphotypes with opposing motility phenotypes. (A) C. difficile R20291 liquid culture was spotted on BHIS-1.8 % agar and grown for 48 hours to allow spreading growth (panel 1). Bacteria collected from the center of a spot yield mostly smooth colonies (panel 2), while bacteria from the edge yield almost exclusively rough colonies (panel 3). (B) Rough and smooth colony isolates obtained as in A were passaged in 0.5X BHIS-0.3% agar (panel 4), and samples were collected from the edge of motile growth. Both rough and smooth colony isolates only gave rise to smooth colonies (panels 5,6). Shown are representative images of colonies from four independent experiments. All images were taken at 2X magnification. (C) Quantification of colony morphology for rough and smooth colony isolates after passage on a BHIS-agar surface (surf) as in A or in motility medium

(swim) as in B, with samples collected from the edges of growth. R, S indicate rough

69 and smooth starting inoculums, respectively. Rough and smooth isolates were also directly passaged through BHIS broth medium or struck on an BHIS agar plate and then enumerated. Symbols indicate individual values from 3-4 independent experiments, and bars indicate means and standard deviations. (D,E) C. difficile R20291, a TFP-null control (pilB1), and rough (R) and smooth (S) colony isolates were assayed for surface motility on BHIS-1.8 % agar. (D) A representative of four experiments is shown. (E)

Surface motility was quantified by measuring the diameter of growth after 48 hours.

(F,G) C. difficile R20291, a non-motile control (sigD), and rough (R) and smooth (S) colony isolates were assayed for swimming motility in 0.5x BHIS-0.3% agar. (F) A representative of four experiments is shown. (G) Swimming motility was quantified by measuring the diameter of growth after 48 hours. (E,G) **p < 0.01, ***p < 0.001, ****p <

0.0001, one-way ANOVA and Tukey’s post-test

70

Figure 2.3. Expression of a c-di-GMP regulated phosphorelay system promotes rough colony formation in a TFP- and flagellum-independent manner. (A) Diagram of seven invertible DNA elements previously identified in C. difficile R20291. Gray triangles represent inverted repeats flanking invertible elements, which are named according to previously defined nomenclature [29]. Downstream regulated genes are shown. Blue rectangles denote c-di-GMP riboswitches, and direction of regulation is indicated with arrows. (B) qPCR analysis of the orientations of the seven invertible DNA sequences in rough (R) and smooth (S) colony isolates. Data are expressed as the percentage of the population with the sequence in the orientation present in the reference genome (FN545816). Symbols indicate values from individual isolates. (C) qRT-PCR analysis for expression of cmrR and cmrT in rough and smooth isolates grown on BHIS agar. Data from four biological replicates was analyzed using the ΔΔCt

71 method with rpoC as the reference gene and was normalized to the smooth condition.

Shown are means with standard deviations. *p < 0.0001, Holm-Sidak multiple t-test. (D-

F) C. difficile R20291, a TFP-null (pilB1) mutant, and an aflagellate (sigD) mutant containing the indicated plasmids for manipulation of intracellular c-di-GMP were grown on BHIS-agar with or without inducer (2 µg/ml nisin). Shown are images representative of the most common morphology yielded by each strain. Values in each panel indicate the percentage of rough colonies of ≥ 100 total colonies from at least two independent experiments. All images were taken at 2X magnification

72

Figure 2.4. The putative response regulators CmrR and CmrT promote the development of rough colonies. C. difficile R20291 with plasmids for expression of cmrR, cmrT, or mutant alleles was evaluated for colony morphology on BHIS-1.8 % agar with ATc to induce expression (ATc concentrations indicated in ng/mL). Expression of cmrT inhibited growth in the presence of 4 ng/µL ATc.

73

Figure 2.5. Growth of mutant and over-expression strains in vitro. C. difficile strains were grown in BHIS broth with ATc at 0, 2 or 10 ng/mL for induction. Optical densities

(600 nm) over time for R20291 with plasmids for expression of (A) cmrR and mutant alleles or (B) cmrT and mutant alleles. (C) Optical densities over time for R20291 WT, cmrR, and cmrT mutants. Shown are means and standard deviations.

74

Figure 2.6. CmrT is required for rough colony formation. (A) Colony morphology of

C. difficile R20291, ΔcmrR, ΔcmrT with vector after 24 hours on BHIS-1.8% agar. (B)

Complementation of ΔcmrR and ΔcmrT mutants with expression of the respective genes in trans induced with 2 ng/mL ATc, imaged after 24 hours. All images were taken at 2X magnification. Representative images from at least two independent experiments are shown.

75

Figure 2.7. CmrR and CmrT inversely regulate surface and swimming motility. C. difficile strains were assayed for surface motility on BHIS-1.8% agar 1% glucose after

72 hours (A-C) and for swimming motility through 0.5X BHIS-0.3% agar after 48 hours

(D-F). (A) C. difficile R20291 (WT) and a TFP-null (pilB1) mutant, each with plasmids for expression of cmrR, cmrT, or a vector control, assayed for surface motility with or without ATc (10 ng/µL for pCmrR, 2 ng/µL for pCmrT). (B) Surface motility of WT,

ΔcmrR, ΔcmrT strains with vector or respective expression plasmids for

76 complementation. (C) Representative image of surface motility. (D) C. difficile R20291

(WT) and a non-motile sigD mutant with pCmrR, pCmrT, or vector as indicated, assayed for swimming motility with or without ATc (10 ng/µL for pCmrR, 2 ng/µL for pCmrT). (E) Swimming motility of WT, ΔcmrR, ΔcmrT strains with vector or complementation plasmids, in the presence of 0.2 ng/µL. A non-motile sigD mutant was included as a control. (F) Representative image of swimming motility. Shown are the means and standard deviations of the diameters of motility growth. *p < 0.05, ****p <

0.0001, two-way ANOVA and Tukey’s post-test comparing +/- ATc (A,D) or one-way

ANOVA with Dunnett’s post-test comparing values to WT/vector (B,E).

77

Figure 2.8. Amino acid substitutions of the phosphorylation sites in CmrR and

CmrT alter activity. C. difficile with plasmids for expression of cmrR, cmrT, or mutant alleles, as well as a vector control, were assayed for surface migration on BHIS-1.8% agar 1% glucose (A,B) and for swimming motility through 0.5X BHIS-0.3% agar

(C,D). TFP-null (pilB1) mutant with vector or cmrR/cmrT expression plasmids were included in the surface migration assay (A,B). A non-motile sigD mutant was used as a control for swimming motility experiments (C,D). The media contained ATc at 0, 2, or 10 ng/ml (white, grey, and black bars, respectively) to induce gene expression.

(B,D) Expression of cmrT inhibited growth at 10 ng/ml and was not included. Shown are

78 the means and standard deviations of the diameters of motile growth after 48 (C,D) or

72 (A,B) hours. **p < 0.005, # p < 0.0005, + p < 0.0001, two-way ANOVA and Tukey’s post-test. These data are representative of four independent experiments.

79

Figure 2.9. CmrR and CmrT do not regulate flagellum or TFP gene expression. C. difficile with vector or cmrR/cmrT expression plasmids were grown for 48 hours in 0.5x

BHIS-0.3% agar to express flagellar genes. Bacteria were recovered and cultured in TY broth with inducer (10 ng/mL ATc for vector and pCmrR; 2 ng/mL ATc for pCmrT).

Samples were collected at mid-exponential phase for RNA extraction and qRT-PCR analysis. The data were analyzed using the ΔΔCt method with rpoC as the reference gene and no ATc as the control condition. Shown are the means and standard deviations of three biological replicates.

80

Figure 2.10. CmrR and CmrT promote bacterial chaining. A) SEM images of a whole

R20291 colony after 3 days of growth on BHIS agar. (A) A colony excised on an agar slice, at 1 x 103 magnification. The arrow denotes the orientation of the image with respect to the colony. The dotted line marks the apparent transition from the colony edge and center. (B) 2.5 x 103 magnification images of cells at the colony center and edge. (C) 4.5 x 103 magnification of rough and smooth cell suspensions. Samples were

81 grown in broth culture prior to fixation. Shown are representative images from two biological replicates. (D) Quantification of cell lengths obtained from SEM images using

ImageJ. Cells were measured from at least seven images from two biological replicates

(rough n = 574, smooth n = 457). ****p < 0.0001 by unpaired t-test. (E, F) Gram stain of

C. difficile samples taken directly from the center and edge of a colony (E) or from rough and smooth colonies (F) at 60X magnification. G) Gram stain of C. difficile with plasmids for expression of cmrR or cmrT, or a vector control were grown on BHIS-agar with and without inducer (10 ng/mL ATc for vector and pCmrR; 2 ng/mL ATc for pCmrT).

Magnifications are indicated. Shown are representative images from two independent experiments.

82

Figure 2.11. The cmrT mutant is defective in cell elongation and chaining. (A)

R20291 WT, cmrR, and cmrT cultures were spotted and grown on BHIS 1.8% agar

1% glucose for 72 hours. Cells from the colony edge were collected, Gram stained, and imaged at 60x magnification. Shown are representative images. (B) Quantification of cell lengths in Gram stain images from (A). At least 2 images from 2 biological replicates were used. The lengths of more than 514 cells per strain were measured using ImageJ and normalized to the average WT cell length. Means and standard deviations are shown. *p < 0.0001, one-way ANOVA. (C) Representative images of the colony edges of WT, cmrR, cmrT, and pilB1::ermB. Cultures were spotted and grown on BHIS

1.8% agar 1% glucose for 72 hours and imaged at 2x (top) and 20x (bottom) magnification.

83

Figure 2.12. The cmrR mutant exhibits an increase in biofilm formation. R20291 smooth and rough isolates and the cmrR, and cmrT mutants were grown in

BHIS 1% glucose- 50 mM sodium phosphate buffer for 24 hours in 24-well polystyrene plates. Adhered biofilms were washed and quantified using a crystal violet staining assay. The means of 4-5 technical replicates were normalized to values for the R20291 rough isolate and combined from two independent experiments. * p < 0.05, one-way

ANOVA with Dunnett’s post-test.

84

Figure 2.13. cmrR and cmrT mutants are not defective in sporulation, germination or toxin production. (A) Sporulation of R20291 rough and smooth isolates and the cmrR and cmrT mutants after 24 hours on 70:30 agar. Sporulation is expressed as a percentage of viable spores versus total cells, then normalized to values obtained for the rough isolate. (B) Germination of spores over time after addition of the germinants taurocholic acid and glycine. TAG = taurocholic acid and glycine germinants. (A,B) No statistically significant differences were observed using a one-way ANOVA, N = 3 biological replicates. (C) TcdA levels in bacterial lysates were assessed after 24 hours of growth in TY medium by western blot. Ponceau S staining was used to determine

85 equal sample loading. (D) Quantification of TcdA western blots for 4 biological replicates. Intensity of the TcdA bands for each was normalized to intensity of Ponceau

S staining per lane. Values were then normalized to the intensity for the smooth isolate.

Shown is a representative image. *p < 0.05, one-way ANOVA with Tukey’s post-test. (E) qRT-PCR analysis tcdA mRNA levels in strains overexpressing cmrR, cmrT, or vector control. Bacteria were cultured in BHIS broth with inducer (10 ng/mL ATc for vector and pCmrR; 2 ng/mL ATc for pCmrT). The data were analyzed using the ΔΔCt method with rpoC as the reference gene and R20291 with vector as the control condition. Shown are the means and standard deviations of 4-6 biological replicates. No statistically significant differences were observed using a one-way ANOVA.

86

Figure 2.14. Colony morphology and CmrR impact C. difficile virulence. Male and female Syrian golden hamsters with were inoculated with 5000 spores of the indicated

C. difficile strains/isolates. (A) Kaplan-Meier survival data showing time of morbidity and euthanasia. Log-rank test: rough v. ΔcmrR, P = 0.003; smooth v. ΔcmrR, P = 0.098, log- rank test. (B) CFU enumerated in feces collected at 24 hour intervals post-inoculation.

Symbols indicate CFU from individual animals, and bars indicate the means. (C)

Orientation of the cmr switch determined by PCR for the spore inoculums and cecal contents of moribund hamsters (rough R1-6, smooth S1-4) with detectable C. difficile.

(D) Quantification of cmr switch orientation by qPCR in spore inoculums and cecal contents of moribund hamsters. Data are expressed as the percentage of the population with the sequence in the OFF orientation. Symbols indicate values from individual

87 hamsters, and horizontal bars indicate the median. Values of rough and smooth cecal contents do not significantly from each other ( p > 0.44, one-way ANOVA).

88 Table 2.1. Strains and plasmids used in this study.

Lab Strain/Plasmid Name Description Reference Notation Escherichia coli DH5α F- φ80lacZΔM15 Δ(lacZYA-argF)U169 Invitrogen recA1 endA1 hsdR17(rκ -, mκ+) phoA [313] supE44 thi-1 gyrA96 relA1 λ- tonA RT270 Escherichia coli E. coli used in conjugations with C. [314] HB101(pRK24) difficile, ApR, CmR RT1124 C. difficile 630 Ribotype 012 strain (Genbank Accession # AM180355) RT273 C. difficile R20291 Ribotype 027 strain (Genbank [99] Accession # FN545816) RT1065 C. difficile UK1 Ribotype 027 strain [315, 316] RT1125 C. difficile VPI 10463 Ribotype 003 strain [317] RT1357 C. difficile ATCC BAA- Ribotype 078 strain ATCC 1875 RT1358 C. difficile ATCC 43598 Ribotype 017 strain ATCC, [318] RT526 R20291 pMC-Pcpr R20291 with pMC-Pcpr, nisin inducible [113] RT527 R20291 pDccA R20291 with pMC-Pcpr::dccA, nisin [113] inducible RT528 R20291 pPdcA-EAL R20291 with pMC-Pcpr::EAL, nisin [237] inducible RT539 R20291 pDccAmut R20291 with pMC-Pcpr::dccAmut, nisin [113] inducible RT2196 R20291 pRPF185 R20291 with pRPF185, ATc inducible This work RT2085 R20291 pCmrR R20291 pRPF185::cmrR, ATc This work inducible RT2086 R20291 pCmrR-D52E R20291 pRPF185::cmrR-D52E, ATc This work inducible RT2087 R20291 pCmrR-D52A R20291 pRPF185::cmrR-D52A, ATc This work inducible RT2107 R20291 pCmrT R20291 pRPF185::cmrT, ATc inducible This work RT2201 R20291 pCmrT-D53A R20291 pRPF185::cmrT-D53A, ATc This work inducible RT1566 R20291 sigD R20291 sigD::ermB, Targetron [103] insertion RT2111 R20291 sigD pMC-Pcpr R20291 sigD::ermB with pMC-Pcpr This work RT2112 R20291 sigD pMC- R20291 sigD::ermB with pMC- This work pDccA Pcpr::dccA RT2113 R20291 sigD pPdcA-EAL R20291 sigD::ermB with pMC- This work Pcpr::EAL RT947 R20291 pilB R20291 pilB::ermB, Targetron insertion [113] RT2177 R20291 pilB pMC-Pcpr R20291 pilB::ermB with pMC-Pcpr This work RT2178 R20291 pilB pDccA R20291 pilB::ermB with pMC- This work Pcpr::dccA RT2179 R20291 pilB pMC-Pcpr R20291 pilB::ermB with pMC- This work Pcpr::EAL

89 RT2180 R20291 pilB pCmrR R20291 pilB::ermB with This work pRPF185::cmrR RT2197 R20291 pilB pRPF185 R20291 pilB::ermB with pRPF185 This work RT2204 R20291 pilB pCmrT R20291 pilB::ermB with This work pRPF185::cmrT RT2256 R20291 ΔcmrR R20291 with in-frame deletion of cmrR This work RT2257 R20291 ΔcmrT R20291 with in-frame deletion of cmrT This work RT2267 R20291 ΔcmrR R20291 ΔcmrR with pRPF185 This work pRPF185 RT2268 R20291 ΔcmrR pCmrR R20291 ΔcmrR with pRPF185::cmrR This work RT2269 R20291 ΔcmrT pRPF185 R20291 ΔcmrT with pRPF185 This work RT2270 R20291 ΔcmrT pCmrT R20291 ΔcmrT with pRPF185::cmrT This work MC310 R20291 spo0A R20291 spo0A::ermB [319] RT399 pMC-Pcpr pMC123 with nisin-inducible cpr [237] promoter RT402 pDccA pMC-Pcpr::dccA (CD630_14200), [237] encodes DGC RT529 pDccAmut pMC-Pcpr::dccAmut (AADEF), encodes [237] inactive DGC RT404 pPdcA-EAL pMC-Pcpr::pdcA-EAL [120] RT709 pRPF185 Contains ATc-inducible Ptet promoter [320] RT2073 pCmrR pRPF185::cmrR This work RT2074 pCmrR-D52E pRPF185::cmrR-D52E This work RT2075 pCmrR-D52A pRPF185::cmrR-D52A This work RT2106 pCmrT pRPF185::cmrT This work RT2200 pCmrT-D53A pRPF185::cmrT-D53A This work

90 Table 2.2 Primers used in this study.

Lab Primer name Sequence (5’ to 3’)a Notation R2175 qPCR_FlgSwit- GTTTTCTTACCAAAGTGATACATTATTATATTAATG ON R2176 qPCR_FlgSwit- CATTAATATAATAATGTATCACTTTGGTAAGAAAAC OFF R2177 qPCR_FlgSwit- GCTATTGTCTGACTTCTTAAATTAGTTGCAT REV R2213 LCF801 GGTAAGTTTGATTTTTATGTTAATGAATTG R2214 LCF714 CAGTTTGTGCACTAGCTATGCCTGC R2215 LCF796 CGCAATTATTTGTTTTTCATATGGATAAAATTGG R2216 LCF797 GATTTTTATGTTAATGAATTGTTATAAAAAACATGG R2261 CDR20291_0685 GTTAAAAATTTAAGATATCTTTTCAGTATAATGGA -IEq1 R2262 CDR20291_0685 CATTTCTAAGAAATATCCTAACATAAAAACAAAA -IEq2 R2263 CDR20291_0685 CGATTACACTACAGAATTAGAATGTCAATG -IEq3 R2264 CDR20291_0963 GTAAATTAAGATGTATTTCATTTCTCAAAAATATCCT -IEq1 R2265 CDR20291_0963 GTAAAGTTTATAAAATCTGAAAAGCTCAAGA -IEq2 R2266 CDR20291_0963 GCTTTTATCGCAAGTTTGTTTTAAATGAC -IEq3 R2270 CDR20291_3128 GGAGATATATGGAGTTAGTGGTGCAA -IEq1 R2271 CDR20291_3128 CTAGCCAATAGACAAGTTTCTAGAAAAATA -IEq2 R2272 CDR20291_3128 GAACAATTCTTGAATATTGTATTGAACATTAAGA -IEq3 R2273 RpoAqF TCATTACCAGGTGTAGCAGTGAATGC R2274 RpoAqR GATAGAGCATGGTCCTTGAGCTTCT R2378 OS 196 GTACAGAAGTTACCCAGAAGCTTGT R2379 OS 197 TCCCCGCAATGGATGTTTTTTAATTCATC R2380 OS 198 TCCCAATTTAAATGTAGAGGTCATCAAT R2527 3128_F TACGAGCTCCTTGAGATTATGATTAAAATACCTTTG R2528 3128_R TACGGATCCCAGTATCTCACTTATGGTACAAACTTATAT R2529 3128pm_glu_F GTATAATTTTGGAAATTTCATTGCC R2530 3128pm_glu_R GGCAATGAAATTTCCAAAATTATAC R2531 3128pm_ala_F GTATAATTTTGGCTATTTCATTGCC R2532 3128pm_ala_R GGCAATGAAATAGCCAAAATTATAC R2533 3126_F TACGAGCTCTAATATAAAAGAATAATGATATTTGGGAGTG R2534 3126_R TACGGATCCCGTTAGCATTTCACATTTATAAC R2535 3126pm_ala_F TTTAGCAATAATTCTAACTGATGGTG R2536 3126pm_ala_R CACCATCAGTTAGAATTATTGCTAAA OS158 GTGTTTTTTGTTACCCTAAGTTTAGTAATAGTATCAAGAG AAGAAG

91 OS163 AGATTATCAAAAAGGAGTTTCCACATCTGCCAAGAATTTT TTAC OS266 AGTATCTCACTACACCACTCCATTCAAAG OS267 GAGTGGTGTAGTGAGATACTGTAATTAATAAATAGTTCTT G OS268 TTTTTTGTTACCCTAAGTTTGCCATCCATGTATCCTATC OS269 TTAATATATTTATATAATCACTCCCAAATATCATTATTC OS271 AGATTATCAAAAAGGAGTTTTGTCAAATCTTGTAACCAAC OS283 AGTGATTATAGAAAATTATAACAATAAGAGGAGC R850 rpoCqF CTAGCTGCTCCTATGTCTCACATC R851 rpoCqR CCAGTCTCTCCTGGATCAACTA R856 CD0245qF GCAACTAATCTAAGAAGTCAGACAATAGC R857 CD0245qR AGGCATAGCATCATTTAGTGTTTCTTC R858 CD0229qF CAAGTGTATCAAATATGAGCGATGAA R859 CD0229qR TTATCCTCGCATCTCCTCTATCATT R1063 fliCqF TACAAGTTGGAGCAAGTTATGGAAC R1064 fliCqR GTTGTTATACCAGCTGAAGCCATTA R930 qRTpilA1F TGGCAGTTCCAGCTTTATTTAGTAAT R931 qRTpilA1R AAGATAATGCTGCACTCTTAACTGA R852 CD0663qF GGAGAAGTCAGTGATATTGCTCTTG R853 CD0663qR CAGTGGTAGAAGATTCAACTATAGCC R2298 CDR_3128qF1 AAGAAAAAGTTCGGGGATTTTTAGC R2299 CDR_3128qR1 CGCTGAAAACTTTAACACATTAGGA R2537 3126_qF GACAAGGATAATTGCC R2538 3126_qR CCATCACCATCAGTTAG a underlined sequences denotes restriction site

92 CHAPTER 3. MULTIPLE REGULATORY INPUTS CONTROL THE EXPRESSION OF cmrRST ENCODING AN ATYPICAL SIGNAL TRANSDUCTION SYSTEM IN CLOSTRIDIOIDES DIFFICILE2

INTRODUCTION

The ability to adapt to environmental changes is critical to the survival of bacterial pathogens, which can face rapidly changing conditions and stresses during infection

[321, 322]. Bacteria have therefore developed diverse strategies for adaptation [249,

250, 323]. Sense-and-respond strategies often involve two-component systems (TCS) that consist of a sensor kinase and a cognate response regulator [324]. In response to an activating signal, such as binding of a ligand or a pH change, the sensor kinase autophosphorylates and activates a response regulator through transfer of the phosphoryl group [324, 325]. Response regulators have a wide range of functions, but many bind DNA to effect changes in transcription when activated [285]. These transcriptional changes contribute to adaptation of the bacterium to the environmental stimuli sensed by the sensor kinase. Signal transduction can also occur via signaling by intracellular small molecules such as cyclic diguanylate (c-di-GMP) [326, 327]. The intracellular level of c-di-GMP is modulated by the opposing activities of diguanylate cyclases and phosphodiesterases that synthesize and degrade c-di-GMP, respectively.

2 This chapter is a manuscript in progress. Authors: Elizabeth M. Garrett, Ognjen Sekulovic, Anchal Mehra, John Shook, Rita Tamayo. I conducted the experiments in all figures except for Figure 3.7. I also wrote and revised the manuscript with feedback from Rita Tamayo.

93 The production and function of these enzymes is controlled by largely undefined environmental signals [120, 326]. C-di-GMP is recognized by specific protein or RNA receptors (riboswitches), which then mediate the adaptive response. Binding of c-di-

GMP to the riboswitch changes the RNA secondary structure to alter transcript stability, termination, or translation initiation, resulting in gene expression changes that contribute to bacterial adaptation [226, 328].

In contrast, generation of phenotypic heterogeneity in a bacterial population serves as a bet-hedging strategy to help ensure survival of the population as a whole.

The development of phenotypically distinct variants improves the odds that a subpopulation survives a sudden stress [248, 249]. Phase variation, a mechanism of generating phenotypic heterogeneity, occurs through reversible genetic changes that typically cause an ON/OFF phenotypic “switch” [135, 329]. Several mechanisms of phase variation have been described including conservative site-specific recombination, in which a sequence-specific recombinase binds inverted repeats and mediates inversion of the intervening DNA [158]. The invertible DNA element contains regulatory information, such as a promoter, that impacts the expression of adjacent genes. In a well-characterized example in Escherichia coli, phase variation of fimbriae production is mediated by the fimS invertible element, which contains a promoter that drives transcription of the fimbrial genes when properly oriented [330-332]. In contrast, an invertible element in Clostridioides difficile results in the formation of an intrinsic terminator in the mRNA transcript when in the OFF orientation, preventing expression of the downstream gene cwpV; the intrinsic terminator does not form in the mRNA with the invertible element in the ON orientation [179].

94 Clostridioides difficile is an intestinal pathogen and a leading cause of nosocomial infections in the U.S [3, 333]. C. difficile infection (CDI) can result in mild to severe diarrhea and potentially fatal complications such as pseudomembranous colitis, toxic megacolon, and sepsis. Recent work has shown that C. difficile contains multiple invertible DNA elements, indicating a considerable capacity for phenotypic heterogeneity through phase variation [138, 185]. Three of these invertible elements have been demonstrated to regulate downstream genes and related phenotypes in a phase variable manner. The Cdi4 invertible element modulates expression of the early stage flagellar operon resulting in phase variation of flagella [103]. This operon encodes the sigma factor SigD, which promotes the transcription of flagellar genes as well as transcription of tcdR, which encodes a direct activator of the C. difficile toxin genes tcdA and tcdB [101, 103, 104, 184, 262]. Accordingly, the production of these toxins is also phase variable. The Cdi1 invertible element mediates phase variation of CwpV, a cell wall protein that contributes to phage resistance [179-181, 270]. Finally, the Cdi6 invertible element, here called the cmr switch, regulates the expression of cmrRST in a phase variable manner [138, 334]. The cmrRST operon encodes a putative non- canonical TCS with two DNA-binding response regulators (CmrR and CmrT) and a sensor kinase (CmrS) [334]. Phase variation of CmrRST allows C. difficile to switch between two colony morphologies, termed rough and smooth, that differ in several physiological characteristics. Through unknown mechanisms, CmrRST positively regulates type IV pili-independent surface motility and cell elongation, and it negatively regulates swimming motility and biofilm formation. Furthermore, a cmrR mutant strain is

95 deficient for colonization and shows attenuated virulence in a hamster model of infection, indicating a role for this regulatory system in disease.

In addition to regulation by the cmr switch, cmrRST expression is regulated by c- di-GMP via a riboswitch [236, 241, 334]. C-di-GMP riboswitches are widespread in the

C. difficile genome, with 11 functional riboswitches in strain 630Δerm [236, 241]. These riboswitches appear to be the primary mechanism of c-di-GMP regulation in C. difficile

[236]. C-di-GMP riboswitches are classified as class I or class II; in C. difficile, the classes act as “off” and “on” switches, respectively [228, 271]. Upstream of the cmr switch are encoded a σA-dependent promoter followed by a class II c-di-GMP riboswitch that positively regulates cmrRST transcription (Figure 3.1A) [236, 241]. Accordingly, high c-di-GMP levels result in the formation of the rough colony morphology, consistent with increased cmrRST expression [334].

The respective contributions of the c-di-GMP riboswitch and the cmr switch to controlling expression of cmrRST, and therefore to the phenotypes controlled by this system, are unknown. One challenge to the study of phase variation is its stochastic nature, which adds uncontrolled variation into an otherwise controlled experiment.

Phase locked strains in which the invertible element is prevented from inverting can be a useful tool with which to study phase variable systems [103, 184, 335]. Our previous work on CmrRST relied primarily on characterization of wild-type (WT) C. difficile rough and smooth colony isolates, which have a strong bias for the ON and OFF cmr switch orientations, respectively, but remain capable of switch inversion and phenotypic switching [334]. In this study, we characterized cmrRST regulation through the interplay of c-di-GMP and the invertible element by generating phase locked cmr OFF and cmr

96 ON strains. Through phenotypic characterization of these phase locked strains and analysis of the effects of c-di-GMP and cmr switch orientation, we found that that these regulatory features act independently. Moreover, c-di-GMP control through the riboswitch supersedes the effects of invertible element orientation. We determined that phase variable regulation occurs through inversion of a promoter encoded in the cmr switch. In addition, we demonstrate that CmrR positively autoregulates expression of cmrRST. The results of this study indicate that cmrRST expression is subject to regulation both through sense-and-respond mechanisms and phase variation, highlighting the potential importance of this system to C. difficile physiology through the variety of activating signals.

MATERIALS AND METHODS

Growth and maintenance of bacterial strains

C. difficile R20291 (NCBI accession number FN545816.1) and derivative strains were maintained in an anaerobic environment of 85% N2, 5% CO2, and 10% H2. C. difficile strains were grown statically at 37 °C in BHIS (37 g/L Bacto brain heart infusion,

5 g/L yeast extract) or in Tryptone Yeast (TY; 30 g/L Bacto tryptone, 20 g/L yeast extract, 1 g/L thioglycolate) broth as indicated. E. coli strains were grown in Luria

Bertani medium at 37 °C with aeration. Antibiotics were used where indicated at the following concentrations: chloramphenicol (Cm), 10 μg/mL; thiamphenicol (Tm), 10

μg/mL; kanamycin (Kan), 100 μg/mL; ampicillin (Amp), 100 μg/mL. Table 3.1 lists strains and plasmids used in this study.

97 Construction of plasmids and strains

Table 3.2 lists primers used in this study. To construct pPtet-DccA (pRT1587) and pPtet-EAL (pRT2444), dccA and the catalytic EAL domain of pdcA were PCR amplified from R20291 genomic DNA (dccA, primers R1907 and R1908; EAL, primers R2689 and

R2690). Purified PCR products were digested with SacI and BamHI (New England

Biolabs) and ligated into a similarly digested pRPF185 [320].

To construct pMC123::5'UTR-phoZ (pRT2497), pMC123::5'UTRcmrOFF-phoZ

(pRT2514), pMC123::5'UTRcmrON-phoZ (pRT2515), inserts were PCR amplified from

R20291 genomic DNA using the following primers: pRT2497, R2754 + R2342; pRT2514 and pRT2515, R2337 + R2342. PCR products were digested with EcoRI and BamHI

(New England Biolabs) and ligated into similarly digested pMC123::phoZ (pRT1343) to produce a transcriptional fusion with phoZ. Ligations were transformed into DH5α, and

Cm-resistant colonies were screened by PCR with the above primers. Plasmids were confirmed by PCR and sequencing. Plasmids were transferred to C. difficile R20291 by conjugation using E. coli HB101(pRK24) as previously described [274, 314].

R20291 ΔcmrRΔcmrT (RT2296) was made by creating a markerless deletion of cmrT in R20291 ΔcmrR (RT2256) as previously described [334]. Regions up- and downstream of cmrT were PCR amplified (upstream, primers OS268 + OS269; downstream, primers OS271 + OS288) and joined by Gibson assembly (New England

Biolabs) into PmeI-linearized pMTL-SC7215 [63]. The plasmid was introduced into

R20291 ΔcmrR by conjugation with E. coli HB101(pRK24). Mutants were identified as described previously and verified by PCR and sequencing [63, 334].

98 The cmr switch was locked either in the OFF or the ON configuration in C. difficile R20291 using a modified version of the previously described codA-based allelic exchange method [63]. Briefly, primer pairs OS383 + OS384 and OS385 + OS386 were used to PCR-amplify flanking homology regions in order to lock the Cdi6 switch in the

OFF configuration. In addition, the primers were designed to delete three nucleotides at position 3,736,176 – 3,736,178 inclusively. The homology arms were then cloned by

Gibson Assembly (New England BioLabs) into PmeI-linearized pMTL-SC7215 vector in which the codA negative-selection marker was replaced by the holin-endolysin module from a prophage in the C. difficile CD630 strain (locus tags CD630_28941 and

CD630_28940, NCBI accession number CP010905.2) under the control of anhydrotetracycline (ATc)-inducible promoter from the pRPF185 plasmid [320]. The resulting plasmid was then conjugated into C. difficile R20291 strain as described previously [274]. The clones were selected on BHI agar containing 15 mg/ml Tm and 12 mg/ml norfloxacin. Single-crossover integrant clones were purified by three rounds of streaking onto BHI agar with 15 mg/ml Tm and 12 mg/ml norfloxacin followed by double-crossover selection on BHI containing 100 ng/ml ATc. Colonies were screened for the desired mutation with primers OS387 + OS137. Genomic DNA from PCR- positive clones was extracted, and the Cdi6 region was PCR-amplified using primers

OS109 + OS111 following by Sanger-sequencing of the amplicon for verification. The same strategy was used to lock the Cdi6 switch in the ON configuration using primer pairs OS388 + OS383 and OS389 + OS386 to PCR-amplify homology arms and OS390

+ OS391 for PCR-screening of presumed double-crossover clones.

99 R20291 recV cmr ON (RT2520) was derived from R20291 recV cmr OFF by ectopically expressing recV (pRecV) to stimulate cmr switch inversion, then screening resulting isolates for the cmr ON switch orientation. Specifically, RT1697 was grown in the presence of 20 ng/mL ATc to induce recV expression, then plated on BHIS without antibiotic selection to allow loss of pRecV. Tm-sensitive colonies, indicating loss of the plasmid pRecV, were identified by replica plating, and cmr ON isolates were identified by PCR with orientation-specific primers R2270 + R2271.

Orientation specific PCR

Orientation specific PCR was done as previously described [103, 138]. C. difficile strains were grown on BHIS-agar. Colonies were boiled to produce lysate for PCR using primers R2270 + R2271 to detect the ON orientation (140 bp product) and

RT2271 + RT2272 to detect the OFF orientation (240 bp product). PCR products were separated on a TAE-1.5% agarose gel.

Quantitative reverse-transcriptase PCR

For analysis of cmrR and cmrT expression in C. difficile R20291 (WT), RIRmut cmr

ON and cmr OFF, these strains were grown overnight (16 h) in TY medium, and 5 L were spotted on BHIS plates. After 24 hours, growth was collected, suspended in 1:1 ethanol:acetone, and stored at -80 ºC for subsequent RNA isolation.

For analysis of expression in R20291 strains carrying pCmrR, pCmrT or vector, these strains were grown overnight in TY-Tm medium. Cultures were diluted 1:30 in

BHIS-Tm broth. After two hours growth, ATc was added to induce gene expression (WT

100 with vector and pCmrR, 10 ng/mL; WT with pCmrT, 2 ng/mL). Samples were collected at mid-exponential phase and saved in 1:1 ethanol:acetone at -80 ºC.

After RNA was isolated and purified, samples were treated with TURBO DNA- free™ Kit (Life Technologies), and cDNA was synthesized using High-Capacity cDNA

Reverse Transcription Kit (Applied Biosystems)using the manufacturer’s protocols [184,

237]. Real-time PCR was performed using SensiFAST SYBR & Fluorescein Kit (Bioline) as previously described. Data were analyzed using the ΔΔCt method, with rpoC as the reference gene and the indicated control condition [103]. Primers are as follows: rpoC,

R850 + R851; cmrR, R2298 + R2299; cmrT, R2537 + R2538; cmrS, R2539 + 2540;

TSS1, R2745 + R2746; TSS2/3, R2751 + R2804; TSS4, R2803 + R2716.

Quantification of switch orientation by quantitative PCR

WT pMC-Pcpr (RT526) and WT pDccA (RT527) were grown overnight in TY-Tm broth then diluted 1:30 in BHIS-Tm containing 1 µg/mL nisin to induce dccA expression.

After growth to late exponential phase, genomic DNA was purified by phenol:chloroform:isopropanol extraction and washed with ethanol. qPCR was performed as previously described with 100 ng of DNA per 20 µL reaction and 100 nM primers [334]. Data were analyzed using the ΔΔCt method, with rpoA as the reference gene and the indicated control condition [138]. Primers are as follows: ON orientation,

R2270 + R2271; OFF orientation, R2271 + R2272; and rpoA, R2273 + R2274.

Microscopy

To image whole colonies, strains were grown on BHIS-Tm agar for 24 hours. For

mut RIR cmr ON/OFF with pPtet-DccA, pPtet-EAL, or vector, ATc (20 ng/mL) was included in the agar. For WT, ΔcmrT, and ΔcmrRΔcmrT carrying pCmrR, pCmrT or vector

101 control, ATc was included at 10 ng/mL for strains with vector or pCmrR and at 2 ng/mL for pCmrT. Colonies were imaged using a Keyence BZ-X810 microscope or Zeiss

Stereo Discovery V8 dissecting microscope with a glass stage.

Fluorescence microscopy of SNAP-labelled strains was performed as previously described [138]. R20291 cmrR::SNAP pPtet-DccA or vector control were grown overnight in TY-Tm broth. Cultures were diluted 1:30 in BHIS-Tm. After two hours of growth, ATc (20 ng/mL) was added for induction. Mid-exponential phase samples were pelleted, washed, and incubated with SNAP-Cell TMR-Star (New England Biolabs) at

37 ºC for 30 minutes. Samples were further washed and mounted on 1% agarose pads.

Imaging was done using bright field and red channel on a Keyence BZ-X810 microscope. Cells were counted on at least three fields from two biological replicates for each strain and condition.

Motility assays

Motility assays were done as previously described [103, 113, 334]. For surface motility, 5 µL of overnight culture was spotted onto BHIS-1% glucose-1.8% agar. For swimming motility, 1 µL was inoculated into 0.5x BHIS-0.3% agar. For strains carrying plasmids, Tm and ATc induction was included (10 ng/mL for vector and pCmrR; 2 ng/mL for pCmrT). After 48 hours (swimming) or 72 hours (surface), the diameter of the motile spots were measured at the widest point and perpendicular to the first measurement. These measurements were averaged to reflect overall diameter. Surface motility was imaged after 72 hours using a Syngene G:Box imager.

102 Biofilm assays

Biofilm assays were done as described previously [120, 334]. Cultures were grown overnight in TY broth and diluted 1:100 in BHIS-1% glucose-50 mM sodium phosphate buffer in a 24-well polystyrene plate. Plates were grown statically for 24 hours. Supernatant were removed from each well. The adherent biofilms were washed once with PBS then stained for 30 minutes with 0.1% crystal violet. After the crystal violet was removed, the biofilms were washed once more with PBS and the remaining crystal violet was solubilized with 1 mL ethanol. Absorbance at 570 nm was measured.

Alkaline phosphatase assays

Strains carrying phoZ reporters were grown in BHIS medium to OD600 ~ 1.0, then

1.5 mL of culture was collected, pelleted and frozen. Samples were thawed, and alkaline phosphatase activity using the substrate p-nitrophenyl phosphate was measured as previously described [103, 336].

5’ rapid amplification of cDNA ends (RACE)

R20291 recV cmr OFF (RT1693), recV cmr OFF pDccA (RT2184), and recV cmr ON (RT2520) were grown in BHIS to mid-exponential phase. RT2184 was grown with Tm and induced with 1

µg/mL nisin. Samples were collected and frozen in 1:1 ethanol:acetone. RNA was extracted and purified as described above. cDNA was synthesized with the 5' RACE System for Rapid

Amplification of cDNA Ends kit (Thermofisher) using the manufacturer’s protocol. Briefly, cDNA was synthesized using primer R2792 and SuperScript II reverse transcriptase. The samples were treated with RNase Mix then column purified and treated with TdT. The products were

103 then PCR amplified with the provided anchor primer and primer R2793, then subjected to

Sanger sequencing.

RESULTS

Generation and phenotypic characterization of phase locked cmr ON and cmr OFF strains

To characterize the dual regulation of the cmrRST system by phase variation and c-di-GMP, we first generated strains phase locked with the cmr invertible element irreversibly in the ON and OFF orientations. Phase locking can be achieved in multiple ways. One strategy is to delete the site-specific recombinase responsible for inversion.

In C. difficile R20291, the RecV recombinase is required for inversion of the cmr switch

[138]. However, RecV mediates inversion of multiple sequences, thus a mutation in recV has pleiotropic effects [103, 138, 179]. Instead, we chose to phase lock the cmr switch by mutating an inverted repeat, preventing site-specific recombination from occurring [158]. The exact nucleotides at which recombination and inversion of the cmr switch occurs were previously identified [138]. We used allelic exchange to delete the nucleotide at the site of inversion in the right inverted repeat (RIR) (position 3,736,177) in the R20291 genome, as well as one additional nucleotide on each side (Figure 3.1B).

We recovered independent mutants with the cmr switch in either orientation and designated them RIRmut cmr ON and RIRmut cmr OFF. These mutants were subjected to orientation specific PCR to confirm that they are genotypically phase locked. In contrast to the WT parent that yielded both cmr ON and OFF products, RIRmut cmr ON exclusively yielded a PCR product corresponding to the ON orientation, and RIRmut cmr

OFF only yielded the OFF orientation product (Figure 3.1C).

104 Previous work showed that R20291 rough and smooth colony variants correlate with cmr ON and OFF gene expression and phenotypes, respectively [334]. Specifically, the cmr ON, rough colony variant exhibits greater surface growth and reduced swimming motility compared to the cmr OFF, smooth colony variant. Furthermore, ectopic expression of cmrR and cmrT individually stimulated rough colony formation and surface motility while inhibiting flagellum-mediated swimming motility, although only cmrT was required for rough colony development and surface motility. To conclusively determine the effect of cmr switch orientation on gene expression, we measured the abundance of the cmrR and cmrT transcripts in the RIRmut cmr ON and RIRmut cmr OFF mutants. Non-locked rough and smooth isolates were included as controls [334]. Both cmrR and cmrT transcripts were 7- to 10-fold more abundant in C. difficile with the cmr switch in the ON orientation, whether in the naturally arising rough colony variant or in

RIRmut cmr ON (Figure 3.1D). Therefore, the three nucleotides at the site of RecV- mediated recombination are required for cmr switch inversion, and the ON orientation of the invertible element is sufficient to promote expression of cmrRST.

The RIRmut phase-locked mutants were then evaluated for the phenotypes previously associated with cmrRST expression. As a control we included WT R20291, which is capable of CmrRST phase variation and generation of both rough and smooth colonies, as well as a ΔcmrRΔcmrT double mutant that forms only smooth colonies

(Figure 3.2A). The RIRmut cmr OFF mutant formed exclusively smooth colonies, similar to ΔcmrRΔcmrT. RIRmut cmr ON formed rough colonies that resemble those in WT, though they appeared to have less defined topology (Figure 3.2A). Also consistent with increased cmrRST expression, RIRmut cmr ON displayed greater surface motility than

105 WT, while RIRmut cmr OFF had decreased surface motility more similar to that of

ΔcmrRΔcmrT (Figure 3.2B, C). Finally, RIRmut cmr ON showed decreased swimming motility and biofilm formation compared to all other strains (Figure 3.2D, E). Together, these results reinforce the model that the cmr switch orientation modulates expression of cmrRST and therefore regulates colony morphology, motility, and biofilm formation.

C-di-GMP regulates cmrRST expression independently of the invertible element

Upstream of the cmr switch is a σA-dependent promoter followed by a c-di-GMP riboswitch sequence, adding another regulatory layer to the transcription of cmrRST

[236, 241, 334]. The two regulatory features could act independently with one regulatory element dominant over the other, or act coordinately such that both the riboswitch and cmr switch must be in the “on” state to allow cmrRST transcription. To distinguish between these possibilities, we modulated c-di-GMP levels in the phase-locked RIRmut cmr ON and OFF strains by overexpressing the diguanylate cyclase dccA (pPtet-DccA) to increase intracellular c-di-GMP or the catalytic EAL domain of the phosphodiesterase pdcA, which hydrolyzes c-di-GMP, to reduce c-di-GMP (pPtet-EAL) [113, 120, 237].

Increasing c-di-GMP through overexpression of dccA resulted in rough colony formation in RIRmut cmr OFF, which normally forms smooth colonies (Figure 3.3A). Additionally, increasing c-di-GMP enhanced the rough colony appearance of RIRmut cmr ON. In contrast, reduction of c-di-GMP through overexpression of the EAL domain did not visibly affect the morphology of RIRmut cmr ON or OFF, suggesting that c-di-GMP levels are insufficient to drive expression of cmrRST under these conditions. Overexpression of dccA led to a significant 2- to 3-fold increase in the abundance of the cmrS transcript,

106 which served as a marker for cmrRST expression, in both RIRmut cmr ON and OFF, consistent with the colony morphology phenotypes observed (Figure 3.3B).

These methods reflect changes in cmrRST expression as a population average.

To assess the effects of the c-di-GMP riboswitch and the cmr switch on the heterogeneity of cmrRST expression among individual bacteria, we used a strain in which cmrR was replaced by a codon-optimized SNAP-tag gene reporter [138]. The pDccA plasmid and vector control were introduced into this strain. In the ΔcmrR::SNAP strain with vector and in the uninduced control, approximately 1% of the population fluoresced (Figure 3.3C). This result is similar to prior analyses showing that a minority of WT cells express cmrRST [138]. Induction of dccA increased fluorescence-positive cells to > 95%. This result is not due to a shift in cmr switch orientation. By qPCR, the orientation of the cmr switch was not significantly different between WT overexpressing dccA and the vector control (Figure 3.4). Therefore, the higher percentage of fluorescent cells reflects cmrRST expression in response to c-di-GMP. Together, these data indicate that expression of cmrRST from the upstream σA-dependent promoter and c-di-GMP riboswitch is not dependent on orientation of the cmr switch. Rather, high c-di-

GMP levels increase cmrRST expression regardless of orientation.

The cmr switch contains a promoter in the ON orientation

The best characterized phase variation regulatory mechanisms involve a promoter encoded within the invertible element that drives expression of adjacent genes when in the proper orientation [139, 332]. Our data suggest that cmrRST transcription may originate from multiple sites in a manner dependent on internal c-di-GMP concentrations and the orientation of the cmr switch. Therefore, we sought to map the

107 transcriptional start sites (TSS) in strains reflecting these different states. Instead of using RIRmut phase locked strains, which are missing three nucleotides of unknown significance to transcription, we used strains in which the cmr switch is locked by a mutation in recV, designated recV cmr ON and recV cmr OFF (Figure 3.5). Additionally, dccA was overexpressed in recV cmr OFF to identify any TSS associated with increased c-di-GMP. Using 5’ RACE, four total TSS were identified in recV cmr ON, recV cmr OFF, and recV cmr OFF pDccA. In recV cmr OFF with and without dccA expression, a TSS designated TSS1 was identified 699 nt upstream of the cmrR start codon, upstream of the c-di-GMP riboswitch (Figure 3.6A, Figure 3.7). TSS1 was not detected in recV cmr ON, but this may be due to lower relative abundance of the TSS1 transcript compared to transcripts from other initiation sites. Another TSS, designated

TSS2, was found in recV cmr ON, located 336 nt upstream of cmrR and mapping within cmr switch in the ON orientation. A third TSS, designated TSS3, was identified 343 nt upstream of the cmrR start codon in both recV cmr OFF strains. This position maps

TSS3 to within the cmr switch, appearing only in transcripts from the cmr OFF strains. In all three strains, TSS4 was identified 90 nt upstream from the cmrR start codon.

These results suggest that expression of cmrRST can originate from multiple sites and confirm an expected TSS upstream of the c-di-GMP riboswitch. TSS2 within the cmr switch in the ON orientation was identified; however, we unexpectedly identified

TSS3 within the cmr switch in the OFF orientation. To elucidate how phase variable cmrRST expression is mediated, we examined the functionality of these newly identified

TSS. We created a series of transcriptional reporter fusions of the alkaline phosphatase

(AP) gene phoZ to regions immediately upstream of cmrRST (Figure 3.6B) [336]. Two

108 of the fusions contain the upstream region from the LIR to the start of the cmrR coding sequence. These fusions differ in the orientation of the invertible element: ON

(pMC123::5'UTRcmrON-phoZ) encompasses TSS2, and OFF

(pMC123::5'UTRcmrOFF-phoZ) encompasses TSS3. Both fusions also include TSS4.

Additionally, we included a reporter with the region from the 3’ end of the RIR to the start of the cmrR coding sequence (pMC123::5'UTR-phoZ), which contains TSS4, as well as a promoterless control (vector). To prevent inversion of the invertible element present in the reporters once in C. difficile, we assayed AP activity in a recV mutant lacking the recombinase necessary for inversion [138]. C. difficile with pMC123::5'UTR- phoZ (Figure 3.6B, construct 1) or pMC123::5'UTRcmrOFF-phoZ (construct 2) showed no significant difference in activity compared to the promoterless control (Figure 3.6C).

In contrast, C. difficile with pMC123::5'UTRcmrON-phoZ (construct 3) exhibited more than 30-fold higher AP activity (Figure 3.6C). These results indicate the presence of a functional promoter in the ON orientation of the cmr switch that leads to the TSS2 transcript and suggest that the TSS3 and TSS4 transcripts are minor products.

CmrR positively autoregulates expression of cmrRST

Response regulators often have the property of autoregulating their expression

[337]. To determine whether CmrR and/or CmrT have the ability to transcriptionally regulate cmrRST, we overexpressed cmrR (pCmrR) and cmrT (pCmrT) in C. difficile and analyzed cmrRST expression by qRT-PCR. Overexpression of cmrR resulted in a significant 7-fold increase in transcript abundance of cmrS and cmrT as compared to the vector control (Figure 3.8A). In contrast, overexpression of cmrT had no effect on

109 cmrR or cmrS transcript abundance. These results indicate the CmrR, but not CmrT, transcriptionally regulates the cmrRST operon.

To assess whether CmrR-mediated regulation is dependent on the orientation of the invertible element, we overexpressed cmrR in the phase-locked recV mutants.

Overexpression of cmrR significantly increased cmrS transcript abundance compared to the respective vector control in both recV cmr OFF and cmr ON strains (Figure 3.8B).

While overexpression of cmrR in recV cmr OFF resulted in over a 20-fold increase in cmrS transcript abundance relative to vector control, cmrS transcript increased only about 2-fold in recV cmr ON pCmrR relative to vector control. We considered the possibility that CmrR regulates cmrRST expression from one of the previously identified

TSS. To address this possibility, we used qRT-PCR to assess the transcript abundance of sequences immediately downstream of TSS1, the cmr switch containing TSS2 and

TSS3, and TSS4 (Figure 3.8C). Transcript abundance did not increase following TSS1 but did increase following the cmr switch with cmrR overexpression in both recV cmr

OFF and cmr ON relative to the respective vector controls (Figure 3.8D). This result suggests that CmrR autoregulates cmrRST expression from either orientation of the cmr switch, though most strongly from the OFF orientation. cmrRST and its upstream regulatory elements are well conserved in C. difficile

C. difficile strains from multiple ribotypes form rough and smooth colonies associated with phase variable cmrRST expression [334]. To examine the potential scope of CmrRST phase variation and function in C. difficile, we used NCBI BLAST to determine the extent of conservation of cmrRST and its upstream regulatory region using R20291 as the reference sequence. The analysis examined the 895 bases 5’ of

110 cmrR, which contains a c-di-GMP riboswitch and the cmr switch (Figure 3.1A), and the cmrRST coding sequences. Among the 71 available C. difficile whole genome sequences available in NCBI (taxonomic ID 1496), C. difficile strains shared high percent sequence identity to the reference for both the riboswitch (98.9% ± 1.2) and the cmr invertible element (98.1% ± 1.4) (Figure 3.9). This similarity is comparable to the average percent sequence identity for the cmrRST operon (98.7% ± 1.1). These results indicate that the c-di-GMP riboswitch and invertible element upstream of cmrRST are highly conserved across many C. difficile strains from divergent ribotypes and underlines the importance of this operon and its regulation to C. difficile physiology.

DISCUSSION

In this study, we demonstrate that the regulatory systems controlling the expression of cmrRST are complicated and multilayered. Our data support a model in which under basal c-di-GMP conditions, the orientation of the invertible element determines the expression level of cmrRST. However, under high c-di-GMP conditions, cmrRST is expressed and the orientation of the invertible element is secondary. Our data also suggest that the expression of cmrRST is subject to autoregulation through

CmrR. Therefore, cmrRST expression may be influenced by multiple environmental signals through c-di-GMP, phase variation, and the response regulator CmrR. Because

CmrRST has important roles in C. difficile motility, biofilm formation, and virulence, these results suggest there are multiple environmental criteria that need to be integrated to appropriately control cmrRST expression.

111 To disentangle the roles of c-di-GMP and phase variation in cmrRST transcription, we phase locked the cmr switch by deleting only three nucleotides out of

28 in the RIR, one of which was identified as the exact site of recombination during inversion of the invertible element by using pair-end sequencing [138]. Phenotypic analysis of phase locked R20291 RIRmut cmr ON and cmr OFF strains showed that they behave similarly to WT rough and smooth populations, respectively. cmrRST expression in the locked ON strain was equivalent to that of a WT rough population, as was the expression in RIRmut cmr OFF compared to WT smooth population. Consistent with these results, RIRmut cmr ON exhibited rough colonies, increased surface motility and decreased swimming motility and biofilm formation as compared to smooth WT, locked OFF, and ΔcmrRΔcmrT strains. These data indicate that orientation of the cmr switch is sufficient to drive expression of cmrRST and further confirms the role of cmrRST in these phenotypes. Interestingly, while RIRmut cmr ON forms rough colonies, they are not identical to those formed in WT populations, and there is still some heterogeneity of colony morphology. This observation may reflect differences in activity of CmrR or CmrT or that other factors contribute to colony morphology.

We found that c-di-GMP regulates cmrRST expression regardless of the orientation of the invertible element. This suggests that the OFF orientation does not contain a terminator sufficient to completely prevent transcription from an upstream promoter as described for other invertible DNA elements in C. difficile [103, 179]. RIRmut cmr OFF exhibited surface motility intermediate between WT and ΔcmrRΔcmrT. This may be because in the locked OFF strain, cmrRST can still be expressed from the promoter upstream of the invertible element if c-di-GMP levels are permissive. C-di-

112 GMP levels increase in C. difficile with growth on a surface [113]. Therefore, c-di-GMP likely promotes CmrRST-mediated surface motility regardless of switch orientation.

Multiple TSS were identified upstream of cmrRST including TSS in the ON and

OFF orientations of the cmr switch (TSS2 and TSS3, respectively) and between the switch and the translational start of cmrR (TSS4). However, only the ON orientation of the switch showed significant promoter activity through transcriptional reporters. This suggests that a promoter in the ON orientation of the cmr switch is responsible for the phase variable expression of cmrRST. The physiological significance of TSS3 and

TSS4 is unclear. These may represent the start sites of low abundance transcripts that do not significantly contribute to cmrRST transcription under the tested conditions.

In other bacteria, the inversion of a promoter is common mechanism by which invertible elements regulated downstream genes [139]. Interestingly, the cmr switch is the first invertible element shown to use this mechanism of regulation in C. difficile. The

Cdi1 invertible element, which regulates the expression of downstream gene cwpV, forms an intrinsic transcriptional terminator in the OFF orientation [179]. In the ON orientation, the Cdi1 invertible element simply permits expression originating from an upstream promoter. While the mechanism by which mechanism by which the flg (Cdi-4) invertible element regulates expression is unknown, evidence suggests that there is not a promoter in the ON orientation and that the OFF orientation terminates transcription through an undefined mechanism [103]. Like the cmr switch, there is a promoter with a c-di-GMP riboswitch upstream of the flg invertible element [103, 104, 236]. However, this riboswitch inhibits transcription in response c-di-GMP binding. Therefore, in spite of

113 having similar regulatory elements, the flg and cmr operons are regulated by those elements in opposing manners.

We found that CmrR positively regulates expression of cmrRST. It is not uncommon for response regulators to regulate their own expression [337]. For example, in Salmonella enterica, the response regulator PhoP autoregulates expression of phoPQ, which regulates the composition of the bacterial enveloped and contributes to pathogenesis [338]. Further examples of autoregulation by response regulators include the BvgAS system in Bordetella pertussis and the VanRS system in Enterococcus faecium, both of which are key virulence determinants [339, 340]. Interestingly, CmrT does not regulate the expression of cmrRST, suggesting that despite significant homology between CmrR and CmrT, they may not share DNA binding motifs and therefore may have distinct regulons. We found that CmrR positively regulates expression of cmrRST with the cmr switch in both orientations and that CmrR-promoted transcription may originate from the cmr switch, possibly corresponding to the TSS2 and

TSS3 sites. This suggests that cmrRST transcription may be subject to a positive feedback loop independent of switch orientation through CmrR.

In summary, this work demonstrates that cmrRST expression is subject to multilayered regulation with multiple potential inputs from environmental signals. The complexity of this regulatory network suggests that cmrRST expression, and therefore its transcriptional targets, requires careful control. Further work that defines the signals which promote cmrRST expression will provide important insights into the role of this

TCS in C. difficile physiology and pathogenesis.

114 Acknowledgements

This work was supported by NIH award R01-AI107029 and R01-AI143638 to R.T. We thank Danielle Fortune for construction of RT2184.

115 FIGURES AND TABLES

Figure 3.1. Generation of genetically phase-locked strains. A) Sequence of the right inverted repeat (RIR) of the cmr switch. Gray shading indicates sequence identity between the imperfect inverted repeats of the RIR in each orientation. Three nucleotides (shown in red) were deleted from the RIR to lock the cmr switch in the ON and OFF orientations. The asterisk (*) indicates the nucleotide at the site of recombination. B) Orientation-specific PCR to detect each orientation of the cmr switch in WT, RIRmut cmr OFF, and RIRmut cmr ON strains. C) qRT-PCR analysis for the cmrR and cmrT transcripts in WT rough and smooth isolates, RIRmut cmr OFF and RIRmut cmr

ON. Data from four biological replicates were analyzed using the ΔΔCt method with

116 rpoC as the reference gene and normalization to the WT smooth samples . Shown are means with standard deviations. + p < 0.0001, two-way ANOVA with Sidak’s multiple comparisons post-test.

117

Figure 3.2. Phase-locked cmr ON strain phenocopies the WT rough morphotype.

A) Colony morphology of WT rough and smooth isolates, RIRmut cmr OFF, RIRmut cmr

ON, and ΔcmrRΔcmrT. Shown are representative images after 24 hours growth on

BHIS agar. B) Surface motility of WT, RIRmut cmr OFF, RIRmut cmr ON, and

ΔcmrRΔcmrT. Surface motility was quantified by measuring the diameter of growth after

72 hours. Shown are means with standard deviations of eight biological replicates. C)

Representative images of surface motility after 72 hours of growth. D) Swimming motility of WT, RIRmut cmr OFF, RIRmut cmr ON, and ΔcmrRΔcmrT mutant. Motility was quantified by measuring the diameter of growth in 0.5x BHIS 0.3% agar after 48 hours.

118 Shown are means with standard deviations of eight biological replicates from two independent experiments. E) Biofilm formation after 24 h in BHIS-1% glucose-50 mM sodium phosphate, quantified by crystal violet staining. Shown are means and standard deviations of 5-6 biological replicates normalized to values of WT. * p < 0.05, **** p <

0.0001, one-way ANOVA with Tukey’s multiple comparisons post-test.

119

Figure 3.3. Regulation of cmrRST by c-di-GMP occurs independent of cmr switch orientation. A) Colony morphology of RIRmut cmr OFF and RIRmut cmr ON containing either vector, pPtet-DccA, or pPtet-EAL manipulation of intracellular c-di-GMP. Strains were grown on BHIS-agar with or without ATc to induce DGC and EAL gene expression. B) qRT-PCR analysis for expression of cmrR in RIRmut cmr OFF and RIRmut cmr ON strains carrying pDccA or vector control. Strains were grown in BHIS broth with

ATc induction. Data were analyzed using the ΔΔCt method with rpoC as the reference gene and normalization to cmr OFF-vector. Shown are means with standard deviations of 6 biological replicates from two experiments. * p < 0.05, *** p < 0.001, one-way

ANOVA with Tukey’s multiple comparisons post-test. C) ∆cmrR::SNAP carrying pPtet-

DccA or vector control was grown in BHIS broth with 20 ng/ml ATc. Samples were imaged at 60x magnification by light and fluorescence microscopy following staining

120 with SNAP-Cell TMR-Star substrate. Shown are representative images with percentages of cells that are fluorescence-positive. Cells from at least three fields each from two biological replicates were counted. Numbers indicate the means and standard deviations of the percent of total cells that were fluorescent.

121

Figure 3.4. C-di-GMP does not cause inversion of the cmr invertible element. WT pDccA and vector control were grown in BHIS with nisin induction until late exponential phase, and gDNA was collected for qPCR analysis of cmr switch orientation. Data are expressed as the percent OFF orientation. Shown are the means and standard deviations of six biological replicates from two independent experiments. n.s. not significantly different, unpaired t-test.

122

Figure 3.5. The cmr switch is phase locked in recV-deficient strains. Orientation- specific PCR to detect each orientation of the cmr switch in WT, recV cmr OFF and recV cmr ON.

123

Figure 3.6. The cmr switch contains a promoter in the ON orientation. A) Diagram of TSS identified by 5' RACE B) Representation of the alkaline phosphatase reporters used in this experiment. Brackets indicate the regions from the native cmr locus that were used to make transcriptional reporters. Construct 1 contains the region between the 3’ end of the right inverted repeat and the 5’ end of cmrR coding sequence

(pMC123::5'UTR-phoZ). Construct 2 includes the region in in construct 1 plus the cmr switch in the OFF orientation (pMC123::5'UTRcmrOFF-phoZ). Construct 3 is identical to construct 2 but with the cmr switch in the ON orientation (pMC123::5'UTRcmrON-phoZ).

B) Activity of constructs 1, 2, and 3 as well as promoterless control, in the recV cmr OFF strain background. Shown are the means and standard deviations data from nine biological replicates from three independent experiments. ****p < 0.0001, one-way

ANOVA with Tukey’s multiple comparisons post-test.

124

Figure 3.7. Identification of multiple transcriptional start sites upstream of cmrRST. Map of TSS identified by 5’ RACE in recV cmr ON, recV cmr OFF and recV cmr OFF pDccA. Gray highlight indicates the riboswitch sequence. Red and blue text denote the OFF and ON orientations of the cmr switch, respectively. Bold text indicates the inverted repeats. Green highlight mark TSS. Underline text indicates predicated -

10/-35 sites if identifiable. Yellow highlight marks the cmrR start codon.

125

Figure 3.8. CmrR positively regulates cmrRST expression. qRT-PCR measurement of cmrRST transcript abundance. Strains were grown with ATc induction (10 ng/mL for vector and pCmrR, 2 ng/mL for pCmrT) prior to RNA isolation. A) WT with pCmrR, pCmrT, and vector control. n.d., not determined; cmrR and cmrT were not measured in

126 strains overexpressing those genes. Shown are means and standard deviations of 4-6 biological replicates. For cmrS measurements, ****p < 0.0001, one-way ANOVA with

Tukey’s multiple comparisons post-test. B) qRT-PCR analysis of cmrS transcript abundance of recV cmr OFF and cmr ON, each with pCmrR or vector control. cmr OFF vector serves as the reference condition for ΔΔCt analysis. Shown are means and standard deviations of 4 biological replicates. ****p < 0.0001 relative to respective vector control, one-way ANOVA with Sidak’s multiple comparisons post-test. C-D) qRT-PCR analysis of sequence immediately downstream of the indicated TSS (shown in C). cmrS transcript abundance data is also shown in B. cmr OFF vector serves as the reference condition for ΔΔCt analysis. ***p < 0.001, ****p < 0.0001 relative to respective vector control, Holm-Sidak t-test for multiple comparisons.

127

128 Figure 3.9. cmrRST and regulatory element sequences are highly conserved across C. difficile strains and ribotypes. Heat map showing percent sequence identity of 71 sequenced C. difficile strains in the NCBI database compared to R20291 reference sequence (NCBI accession number FN545816.1). If the strain has been ribotyped, the ribotype is indicated on the left. For the cmr switch, both orientations were used as the reference and the highest percent sequence identity is shown. The following nucleotides were used as the reference sequence from R20291: riboswitch,

3736512-3736863; cmr switch, 3735969-3736513; cmrRST, 3732474-3735968.

129 Table 3.1. Strains and plasmids used in this study.

Lab Notation Strain/Plasmid Name Description Reference AC472 Escherichia coli DH5a F- φ80lacZΔM15 Δ(lacZYA-argF)U169 Invitrogen recA1 endA1 hsdR17(rκ -, mκ+) phoA [313] supE44 thi-1 gyrA96 relA1 λ- tonA RT270 Escherichia coli E. coli used in conjugations with C. difficile, [314] HB101(pRK24) ApR, CmR RT273 C. difficile R20291 Ribotype 027 strain (Genbank Accession # [99] FN545816) RT2395 R20291 RIRmut cmr OFF R20291 with the cmr invertible element This work locked in the OFF orientation due to the deletion of three nucleotides in the right inverted repeat RT2406 R20291 RIRmut cmr ON R20291 with the cmr invertible element This work locked in the ON orientation due to the deletion of three nucleotides in the right inverted repeat RT2256 R20291 ΔcmrR R20291 with in-frame deletion of cmrR [334] RT2296 R20291 ΔcmrRΔcmrT R20291 with in-frame deletions of cmrR This work and cmrT RT2435 R20291 RIRmut cmr ON R20291 cmr locked ON with pRT1611 This work vector RT2436 R20291 RIRmut cmr ON R20291 cmr locked ON with pRT1587 This work pPtet-DccA RT2437 R20291 RIRmut cmr ON R20291 cmr locked ON with pRT2444 This work pPtet-EAL RT2438 R20291 RIRmut cmr OFF R20291 cmr locked OFF with pRT1611 This work vector RT2439 R20291 RIRmut cmr OFF R20291 cmr locked OFF with pRT1587 This work pPtet-DccA RT2440 R20291 RIRmut cmr OFF R20291 cmr locked OFF with pRT2444 This work pPtet-EAL RT2187 R20291 cmrR::SNAP R20291 with cmrR replaced by allelic [138] exchange with a SNAP-tag coding sequence RT2500 R20291 cmrR::SNAP R20291 cmrR::SNAP with pRT1611 This work vector RT2501 R20291 cmrR::SNAP R20291 cmrR::SNAP with pRT1587 This work pPtet-DccA RT1693 R20291 recV cmr OFF R20291 with an insertional mutation in [181] recV (recV::ermB); cmr locked OFF RT2502 R20291 recV R20291 recV::ermB cmr OFF with This work pMC123::phoZ pMC123::phoZ RT2507 R20291 recV R20291 recV::ermB cmr OFF with This work pMC123::5'UTR-phoZ pRT2497 RT2516 R20291 recV R20291 recV::ermB cmr OFF with This work pMC123::5'UTRcmrOFF pRT2514 -phoZ

130 RT2517 R20291 recV R20291 recV::ermB cmr OFF with This work pMC123::5'UTRcmrON- pRT2515 phoZ RT1615 R20291 vector R20291 with pRT1611 [103] RT2085 R20291 pCmrR R20291 with pCmrR (pRT2073) [334] RT2107 R20291 pCmrT R20291 with pCmrT (pRT2106) [334] RT2463 R20291 RIRmut cmr OFF R20291 cmr locked OFF pCmrR This work pCmrR RT2465 R20291 RIRmut cmr ON R20291 cmr locked ON pCmrT This work pCmrR RT526 R20291 pMC-Pcpr R20291 with pMC-Pcpr containing a nisin [113] inducible promoter RT527 R20291 pDccA R20291 with pMC-Pcpr::dccA [113] RT2269 R20291 ΔcmrT vector R20291 ΔcmrT with pRT1611 [334] RT2402 R20291 ΔcmrT pCmrR R20291 ΔcmrT with pRT2073 This work RT2270 R20291 ΔcmrT pCmrT R20291 ΔcmrT with pRT2106 [334] RT2183 R20291 recV cmr OFF R20291 recV::ermB cmr OFF with pMC- This work pMC-Pcpr Pcpr RT2184 R20291 recV cmr OFF R20291 recV::ermB cmr OFF with pMC- This work pDccA Pcpr::dccA RT1697 R20291 recV cmr OFF R20291 recV::ermB cmr OFF with pPtet- [262] pRecV RecV RT2520 R20291 recV cmr ON R20291 recV::ermB cmr locked ON This work RT2198 R20291 recV cmr OFF R20291 recV::ermB cmr OFF with This work vector pRT1611 RT2543 R20291 recV cmr OFF R20291 recV::ermB cmr OFF with This work pCmrR pRT2073 RT2544 R20291 recV cmr ON R20291 recV::ermB cmr ON with pRT1611 This work vector RT2545 R20291 recV cmr ON R20291 recV::ermB cmr ON with pRT2073 This work pCmrR RT1077 pMTL-SC7215 Vector for allelic exchange in C. difficile [63] R20291 RT709 pRPF185 Contains ATc-inducible Ptet promoter with [320] gusA RT1611 pRT1611 Derivative of pRPF185 with gusA removed [103]

RT1587 pPtet-DccA pRPF185 with gusA replaced by dccA This work

RT2444 pPtet-EAL pRPF185 with gusA replaced by pdcA-EAL This work RT2073 pCmrR pRPF185 with gusA replaced by cmrR [334] RT2106 pCmrT pRPF185 with gusA replaced by cmrT [334] RT1343 pMC123::phoZ pMC123 with Enterococcus faecalis phoZ [103] RT2497 pMC123::5'UTR-phoZ phoZ transcriptional reporter including the This work region between cmrR and right inverted repeat of the cmr invertible element RT2514 pMC123::5'UTRcmrOFF phoZ transcriptional reporter including the This work -phoZ region from the cmr invertible element (OFF) to cmrR

131 RT2515 pMC123::5'UTRcmrON- phoZ transcriptional reporter including the This work phoZ region from the cmr invertible element (ON) to cmrR

132 Table 3.2. Primers used in this study.

Primer Name Primer Sequence (5’ to 3’) OS383 gttttttgttaccctaagtttAGTAATAGTATCAAGAGAAGAAGG OS384 agtctccgGTTGGAGAATGGAGCTTAAAG OS385 tctccaacCGGAGACTTTAATTAGTATTTAATAATAAATC OS386 gattatcaaaaaggagtttCCATTACTTGCTCTAATAGAC OS387 GACTTTAAGCTCCATTCTCCAACAAT OS388 agtctccgGTTGGAAAAGGGAAATTTTTTAAAAAG OS389 ttttccaacCGGAGACTTTAATTAGTATTTAATAATAAATC OS390 CTTTTTAAAAAATTTCCCTTTTCCAACAAT OS391 CACCACTCCATTCAAAGGTATTTTAATC OS109 GGAGATATATGGAGTTAGTGGTGCAA OS111 CGCTCTACTATATCCATAGCATCTTT OS137 GAACAATTCTTGAATATTGTATTGAACATTAAGA OS268 ttttttgttaccctaagtttGCCATCCATGTATCCTATC OS269 ttaatatatttaTATAATCACTCCCAAATATCATTATTC OS271 agattatcaaaaaggagtttTGTCAAATCTTGTAACCAAC OS283 agtgattataGAAAATTATAACAATAAGAGGAGC R1907 CAGAGCTCCTAGTACAAAGTATTTTATTTTGGAG R1908 GACGGATCCCAGTACTTAATGTCAATATCTTGTATAG R2689 CAGAGCTCCTATGGAGGAGATAAGTATATGGATTTAAAGGCATCAATTAATGA TG R2690 CAGGATCCATACTATTCCCAATTTAACATCC R2754 ATGCAGAATTCACTTTAATTAGTATTTAATAATAAATCTTAATG R2342 ATGCAGGATCCCATTACACCACTCCATTCAAAGG R2337 ATGCAGAATTCGGTGAAATTTTGGCTTTTAAAGTAGCC R2270 GGAGATATATGGAGTTAGTGGTGCAA R2271 CTAGCCAATAGACAAGTTTCTAGAAAAATA R2272 GAACAATTCTTGAATATTGTATTGAACATTAAGA R850 CTAGCTGCTCCTATGTCTCACATC R851 CCAGTCTCTCCTGGATCAACTA R2298 AAGAAAAAGTTCGGGGATTTTTAGC R2299 CGCTGAAAACTTTAACACATTAGGA R2537 GACAAGGATAATTGCC R2538 CCATCACCATCAGTTAG R2539 GATAGATGACTGGGAG R2540 CGATAAGTAGCATTCCC R2273 TCATTACCAGGTGTAGCAGTGAATGC R2274 GATAGAGCATGGTCCTTGAGCTTCT R2792 GTCTAAAACTATACGCTC R2793 CCATAGCATCTTTAGCAGTCTCTGATGTATATA

133 CHAPTER 4. TRANSCRIPTOME ANALYSIS OF THE PHASE VARIABLE SIGNAL TRANSDUCTION SYSTEM CmrRST IN CLOSTRIDIOIDES DIFFICILE3

INTRODUCTION

The ability to sense and respond to environmental signals is key to bacterial adaptation and survival. Two component systems (TCS) are an important mechanism by which bacteria sense environmental signals and respond in order to adapt to quickly changing environments and survive environmental stresses. TCS consist of a sensor kinase, typically a histidine kinase found in the bacterial membrane, and a cognate response regulator [341, 342]. In response to an environmental signal, the sensor kinase autophosphorylates and transfers the phosphoryl group to the receiver domain of the response regulator. Histidine kinases can also have phosphatase activity and dephosphorylate the response regulator when its signal is not sensed [343, 344]. When phosphorylated, the response regulator becomes active and mediates the intracellular response to the extracellular signal through its effector domain [341, 342, 345].

Response regulators can have a variety of effector domains. The majority of known response regulators contain a DNA binding domain, allowing the response regulator to alter gene transcription through direct binding of a consensus DNA sequence [341].

3 This chapter is a manuscript in progress. Authors: Elizabeth M. Garrett, Derrick K. Cooper, Jilarie A. Santos-Santiago, John Shook, Rita Tamayo. I contributed data to all figures except for Figures 4.3B and 4.5B,C, and experiments in those figures were done by students under my supervision. I also wrote and revised the manuscript with feedback from Rita Tamayo.

134 Therefore, TCS allow the bacterium to couple a discrete environmental signal to changes in gene expression.

Many bacteria produce multiple TCS, allowing the bacterium to sense and respond to many diverse environmental signals [341, 342]. For example, Escherichia coli encodes 62 conserved TCS genes while Myxococcus xanthus encodes 263 [346].

The number of TCS genes correlates not only with genome size but also the diversity of the bacterium’s natural environment, indicating the importance of TCS for adaptation

[347]. TCS play an important role in virulence by affecting bacterial survival within the host. TCS have been described as key virulence determinants in multiple bacterial pathogens including Escherichia coli, Bordetella pertussis, Pseudomonas aeruginosa, and Staphylococcus aureus [348-351]. For example, the SaeRS system in S. aureus regulates toxin production as well as innate immune evasion [352-354]. In P. aeruginosa, a leading cause of chronic infection in patients with cystic fibrosis, the

GacAS system regulates the production of factors associated with chronic infection such as the type VI secretion system and biofilm formation [201, 350, 355]. The

EnvZ/OmpR system is produced by multiple intestinal pathogens, including E. coli and

Salmonella Typhimurium, and is important for survival under osmotic and acid stress as encountered during growth in the host gastrointestinal system [356, 357]. Several TCS contribute to antibiotic resistance as well. For example, VanSR is a TCS produced by multiple pathogens including S. aureus and Enterococcus faecium that provides vancomycin resistance by regulating the production of factors that result in cell wall remodeling [358]. Because of their significance to pathogenesis and antibiotic resistance, efforts have been made to target TCS as an antimicrobial therapy [359].

135 Clostridioides difficile is an intestinal pathogen and a leading cause of nosocomial infections in the U.S., Europe, and Australia [1-3]. C. difficile infection causes mild to severe diarrheal symptoms with potentially fatal complications including pseudomembranous colitis, toxic megacolon, and sepsis. C. difficile produces many

TCS; strain CD196 (ribotype 027) encodes 53 putative histidine kinases and 54 response regulators [360]. Many of these TCS remain uncharacterized. Recent work has identified a TCS, CmrRST, with an important role in C. difficile physiology [334].

CmrRST consists of two response regulators (CmrR and CmrT) and a histidine kinase

(CmrS). CmrRST promotes type IV pilus-independent surface motility and the formation of cell chains while inhibiting flagella-dependent swimming motility and biofilm formation. High cmrRST expression also results in rough colony morphology, while low cmrRST expression results in smooth colonies. In a hamster model of infection, a

ΔcmrR strain was attenuated, indicating a role for CmrRST in virulence. Expression of cmrRST is regulated by c-di-GMP and by phase variation by site-specific recombination.

The mechanisms by which CmrRST exerts its effects on colony morphology, surface motility, swimming motility, cell chaining, and virulence are unknown. CmrR and

CmrT are both OmpR-family response regulators with winged helix-turn-helix (wHTH) domains, indicating that they likely function as transcription factors. Consistent with this role, CmrR promotes cmrRST transcription (Chapter 3). However, other targets of CmrR and CmrT regulation are unknown. The goal of this study was to define the transcriptional regulon of CmrRST. We determined that CmrR and CmrT regulate transcription of numerous genes encoding products of varied functions including cell wall proteins and proteins with functions in metabolism. We have begun to evaluate

136 CmrR- and CmrT-regulated genes for roles in CmrRST-regulated phenotypes. One

CmrR- and CmrT-regulated locus, CDR20291_1689-1690, was found to impact C. difficile colony morphology and swimming motility, but not surface motility.

MATERIALS AND METHODS

Growth and maintenance of bacterial strains

C. difficile R20291 (NCBI accession number FN545816.1) and derivative strains were grown statically at 37 °C in BHIS (37 g/L Bacto brain heart infusion, 5 g/L yeast extract) or in TY (30 g/L Bacto tryptone, 20 g/L yeast extract, 1 g/L thioglycolate) broth as indicated and were maintained in an anaerobic environment of 85% N2, 5% CO2, and

10% H2. E. coli strains were grown with aeration in Luria Bertani medium at 37 °C.

Antibiotics were used where indicated at the following concentrations: chloramphenicol

(Cm), 10 μg/mL; thiamphenicol (Tm). Table 4.1 lists strains and plasmids used in this study.

Construction of plasmids and strains

Table 4.2 lists primers used in this study. To construct p1689-1690,

CDR20291_1689-1690 were PCR amplified from R20291 using primers R2755 +

R2756, then spliced into SacI and BamHI-digested pRPF185 by Gibson Assembly (New

England Biolabs) [320]. Plasmids were confirmed by PCR and sequencing. p1689-1690 was transferred to C. difficile R20291 by conjugation using E. coli HB101(pRK24) as previously described [274, 314].

137 RNA-seq

R20291 (WT) with plasmids pRPF185 (vector; RT1615), pCmrR (RT2085), or pCmrT (RT2107) were grown overnight in TY-Tm [334]. Cultures were then diluted in

BHIS-Tm10, and anhydrotetracline (ATc; 2 ng/µL for vector and pCmrT; 10 ng/µL for vector and pCmrR) was added after two hours of growth. Cultures were grown to mid- exponential phase, collected, and stored at -80 °C in 1:1 ethanol:acetone. RNA was extracted by phenol-chloroform as previously described and purified using the RNeasy kit (Qiagen) [184, 237]. The samples were treated RNase-Free DNase Set (Qiagen).

RNA sample QC, additional DNase treatment, library preparations and sequencing reactions were conducted at GENEWIZ, LLC. (South Plainfield, NJ, USA).

RNA samples were quantified using a Qubit 2.0 Fluorometer (Life Technologies), and

RNA integrity was assessed with a 4200 TapeStation (Agilent Technologies). rRNA depletion was performed using the Ribozero rRNA Removal Kit (Illumina). RNA sequencing library preparation used the NEBNext Ultra RNA Library Prep Kit for

Illumina by following the manufacturer’s recommendations (New England Biolabs).

Briefly, enriched RNAs were fragmented for 15 minutes at 94 °C. First strand and second strand cDNA were subsequently synthesized. cDNA fragments were end repaired and adenylated at 3’ends, and a universal adapter was ligated to the cDNA fragments, followed by index addition and library enrichment with limited cycle PCR.

Sequencing libraries were validated using the Agilent Tapestation 4200 and quantified by using a Qubit 2.0 Fluorometer as well as by quantitative PCR (qPCR; Applied

Biosystems).The sequencing libraries were multiplexed and clustered on one lane of a flowcell and loaded on the Illumina HiSeq instrument according to manufacturer’s

138 instructions. The samples were sequenced using a 2x150 Paired End (PE) configuration. Image analysis and base calling were conducted by the HiSeq Control

Software (HCS). Raw sequence data (.bcl files) generated from Illumina HiSeq were converted into FASTQ files and de-multiplexed using Illumina's bcl2fastq 2.17 software.

One mismatch was allowed for index sequence identification.

Data was analyzed using CLC Genomics Workbench v. 11 (Qiagen). Reads were mapped to the C. difficile R20291 genome (FN545816.1) using the software’s default scoring penalties for mismatch, deletion, and insertion differences. Principle component analysis (PCA) indicated that WT vector treated with ATc 2 ng/mL was not significantly different than WT vector with ATc 10 ng/mL, so data from these samples were combined for further analysis. Transcript reads for each gene were normalized to the total number of reads and gene length (expressed as reads per kilobase of transcript per million mapped reads [RPKM]). Fold changes greater than 2 relative to

WT vector with a False Discovery Rate (FDR) p < 0.05 were considered to be statistically significant.

Quantification of switch orientation by quantitative PCR

After growth in TY overnight, genomic DNA was purified by phenol:chloroform:isopropanol extraction and washed with ethanol. qPCR was performed as previously described with 100 ng of DNA per 20 µL reaction and 100 nM primers [334]. Data were analyzed using the ΔΔCt method, with rpoA as the reference gene and the indicated control condition [138]. Primers are listed in Table 4.2.

139 Quantitative reverse-transcriptase PCR (qRT-PCR)

Strains were grown overnight in TY-Tm medium. Cultures were diluted 1:30 in

BHIS-Tm broth. After two hours of growth, ATc was added to induce gene expression

(vector and pCmrR, 10 ng/mL; WT with pCmrT, 2 ng/mL). Samples were collected at mid-exponential phase and saved in 1:1 ethanol:acetone at -80 ºC.

RNA was isolated and purified as previously described using phenol-chloroform extraction [184, 237]. The samples were treated with TURBO DNA-free™ Kit (Life

Technologies). cDNA was synthesized using the High-Capacity cDNA Reverse

Transcription Kit (Applied Biosystems) using the manufacturer’s protocols. Real-time

PCR was performed using SensiFAST SYBR & Fluorescein Kit (Bioline) as previously described [334]. Data were analyzed using the ΔΔCt method and the indicated control condition, with rpoC as the reference gene [138]. Primers are listed in Table 4.2.

Microscopy

To image whole colonies, strains were grown for 24 hours on BHIS-Tm agar for

24 hours with ATc included at the indicated concentration. Images were taken using a

Keyence BZ-X810 microscope at 2x magnification. For Gram staining, strains were grown on BHIS-Tm agar with ATc for 48 hours. Colonies were heat fixed on glass slides and Gram stained (BD Kit 212524). Slides were imaged at 60x magnification using a

Keyence BZ-X810 microscope.

Surface and swimming motility assays

Motility assays were done as previously described [103, 113, 334]. For surface motility, 5 µL of overnight culture was spotted onto BHIS-Tm 1% glucose-1.8% agar.

140 For swimming motility, 1 µL was inoculated into 0.5x BHIS-Tm 0.3% agar. ATc was included at the indicated concentrations. After 48 hours (swimming) or 72 hours

(surface), the diameters of the motile spots were measured at the widest point and perpendicular to the first measurement, and the measurements were averaged to reflect overall diameter.

RESULTS

CmrR and CmrT regulate many genes of diverse functions

To identify regulatory targets of CmrRST, we used RNA-seq to characterize the transcriptomes of the response regulators CmrR and CmrT. In C. difficile R20291, cmrR and cmrT genes were overexpressed individually from plasmids with ATc induction.

Because overexpression of cmrT was previously found to be toxic, a lower concentration of ATc (2 ng/mL) was used for the R20291 pCmrT samples compared to the concentration of ATc (10 ng/mL) used for R20291 pCmrR samples (n = 3 each)

[334]. Accordingly, we included vector-control cultures treated with either 2 or 10 ng/uL

ATc (n = 2 each). For the vector-control samples, the difference in ATc concentration did not significantly affect the transcriptomes of R20291 vector samples based on PCA analysis. Therefore, the vector-control data sets were combined to serve as the reference condition (n = 4). Gene transcripts that showed more than a two-fold change in reads per kilobase of transcript per million mapped reads (RPKM) and False

Discovery Rate (FDR) p <0.05 were considered significant. Table 4.3 contains a complete list of genes that were differentially expressed in strains overexpressing cmrR or cmrT relative to vector.

141 A total of 105 genes were differentially regulated by CmrR, with 43 negatively regulated and 62 positively regulated. CmrT differentially regulated 40 genes, 20 of which were positively regulated. Consistent with previous work, cmrS and cmrT were significantly upregulated about 7-fold by CmrR, indicating that CmrR positively autoregulates cmrRST expression. Therefore, we would expect genes regulated by

CmrT to also be regulated by CmrR as well. Twenty-six genes were differentially regulated by both CmrR and CmrT. Due to CmrR autoregulation of cmrRST, we cannot distinguish between direct targets of both CmrR and CmrT versus genes that are regulated indirectly by CmrR through CmrT. Of the 14 regulated only by CmrT, 12 genes were within 0.6-fold of the two-fold cutoff; therefore, these genes may not have met significance due to the lower CmrT abundance in pCmrR samples compared to pCmrT.

The largest category of genes regulated by CmrR and CmrT include those from the flagellar operons, nearly all of which are about two-fold upregulated. CmrR and

CmrT were previously shown to inhibit flagellar swimming motility, therefore inhibition of swimming motility is not due to altered flagellar gene transcription. The next largest functional category of genes includes those that encode cell surface proteins. The gene with the highest fold change in R20291 pCmrR was cwp28 (37.2-fold,

CDR20291_1911), which encodes a cell wall protein of unknown function. Genes encoding cell wall proteins Cwp10, Cwp11 and Cwp12 were also differentially regulated, though their function is also unknown. Additionally, CDR20291_2503-2508, which may constitute an operon containing three genes with functions in cell division regulation as well as a gene annotated as encoding a membrane protein, is significantly

142 downregulated. CmrR also differentially fapR (-2.7-fold, CDR20291_1015) encoding a fatty acid/phospholipid biosynthesis regulator with functions in the regulation of membrane lipids in other bacteria [361]. These results may indicate that CmrRST has a role in remodeling the cell surface, which perhaps relates to its effects on cell elongation and chaining. CmrR and CmrT also significantly altered expression of several genes involved in transport and metabolism, including genes encoding a putative osmoprotectant transporter (>-6.9-fold, CDR20291_3074-3075). CmrR significantly downregulated an operon involved in riboflavin biosynthesis (-5.3 to -7.9-fold,

CDR20291_1595-1598) as well as a riboflavin transporter (-3.7-fold, CDR20291_0146).

These targets suggest that CmrRST impacts metabolic pathways within the bacterium as well.

Cross-talk between phase variable genes

The C. difficile R20291 genome contains seven invertible DNA elements, flanked by inverted repeats, that are known or putative targets of site-specific recombination.

Three of these elements, including the cmr switch, have been demonstrated to regulate the expression of neighboring genes in a phase variable manner [103, 179, 334]. Our

RNA-seq results indicate that two genes known to be phase variable (flgB and cwpV) and two putative phase variable genes (CDR20291_0685 and CDR20291_1514) were differentially expressed in R20291 pCmrR and pCmrT compared to R20291 with vector.

These transcriptional differences may be due to disparities in the orientation of the invertible element between conditions, transcriptional regulation of the genes by CmrR and CmrT, or both. To distinguish between these possibilities, we quantitatively determined the orientation of these invertible elements in R20291 with vector, pCmrR,

143 or pCmrT without induction by qPCR. The invertible element upstream of

CDR20291_0685 in R20291 pCmrR showed a significant difference in orientation from

R20291 vector and pCmrT; while R20291 vector and pCmrT possessed the switch in the orientation matching the sequence in the R20291 reference genome, the sequence was inverted in R20291 pCmrR (Figure 4.1). These results correlate with the RNA-seq data showing differential expression of these genes in pCmrR compared to pCmrT and the vector control. Differences in CDR20291_0685 transcript abundance is likely attributable to differences in the orientation of invertible elements and not direct regulation by CmrR and CmrT. Similarly, the invertible element upstream of

CDR20291_1514 showed a bias toward the reference genome orientation in pCmrR and pCmrT, but the opposite bias in the vector control (Figure 4.1). Accordingly,

CDR20291_1514 transcript abundance differed in the vector control compared to pCmrT and pCmrT. In contrast, no significant difference in switch orientation bias was observed for the flg and cwpV switches, suggesting that CmrR and CmrT are direct or indirect regulators of cwpV and flagellar gene transcription.

Regulation of swimming motility and colony morphology by CDR20291_1689- 1690

Having identified regulatory targets of CmrR and CmrT, we sought to identify which genes are responsible for phenotypes known to be regulated by CmrRST.

CDR20291_1689 and CDR20291_1690 were both positively regulated by CmrR and

CmrT (pCmrR, 27.5- and 2.5-fold, respectively; pCmrT, 19.1- and 2.4-fold, respectively), and CDR20291_1689 was one of the most highly upregulated genes that was identified

(Figure 4.2A). This result was validated by qRT-PCR. Overexpression of either cmrR or

144 cmrT increased the abundance of CDR20291_1689 transcript more than 40-fold (Figure

4.2B). CDR20291_1690 transcript also increased more than 3-fold but did not reach statistical significance.

cmrRST expression results in a rough colony morphology, increased surface motility, decreasing swimming motility, and cell chaining and elongation [334]. To assess the contribution of CDR20291_1689-1690 to these phenotypes, both genes were overexpressed in tandem in R20291, under the control of an ATc-inducible promoter. Increasing levels of induction showed a progressive change in colony morphology resembling, but not identical, to the rough colony morphology (Figure

4.3A)[334]. Similar to rough colonies or those formed with the overexpression of CmrT,

R20291 p1689-1690 formed colonies that were irregular and exhibited tendrils; in contrast, these colonies still have a rounded and smooth texture [334]. We also Gram stained bacteria overexpressing CDR20291_1689-1690 as well as vector and pCmrT controls to determine whether these genes contribute to cell elongation or chaining.

CDR20291_1689-1690 overexpression did not alter cell morphology (Figure 4.3B).

CmrRST positively regulates surface motility and negatively regulates swimming motility [334]. Therefore, we also examined the ability of CDR20291_1689-1690 to regulate motility. Overexpression of CDR20291_1689-1690 did not significantly increase surface motility relative to vector control (Figure 4.4A). However,

CDR20291_1689-1690 overexpression significantly reduced swimming motility about two-fold (Figure 4.4B). In contrast, overexpression of cmrT inhibits swimming entirely.

These results suggest that CDR20291_1689-1690 contribute to, but do not fully underly, the CmrRST-mediated inhibition of swimming motility.

145 Contribution of CmrR and CmrT to regulation

CmrR regulates the expression of cmrT (Table 4.3). Therefore, our RNA-seq results cannot distinguish whether gene targets are regulated directly by CmrR or if they are regulated indirectly by CmrR through CmrT. To evaluate the contribution of CmrR to regulation of gene expression, we overexpressed cmrR and cmrT individually in a

ΔcmrRΔcmrT and used qRT-PCR to assess differential regulation of genes identified in our RNA-seq analysis to be regulation by both CmrR and CmrT (CDR20291_1689,

1690, 1911, 1913, 1914, 3074, 3075) or just by CmrR (CDR20291_1595 and 2122), focusing on those with the highest fold changes in transcript abundance.

Overexpression of cmrR in a ΔcmrRΔcmrT background did not significantly affect the expression of any of the investigated genes, including in CDR20291_1595 and 2122

(Figure 4.5A). cmrT overexpression did affect CDR20291_1595 expression, in contrast to our RNA-seq analysis. CDR20291_2122 expression was not affected by overexpression by either cmrR or cmrT, and may instead be regulated by differences in the expression of phase variable genes as described above (Figure 4.1). cmrT overexpression resulted in the differential regulation of all tested genes identified by

RNA-seq to be regulated by CmrR and CmrT, including CDR20291_1689-1690 (Figure

4.5A).

CmrR and CmrT both promote the rough colony morphology and surface motility when overexpressed [334]. However, only a cmrT mutant is deficient in these phenotypes as compared to WT [334]. To determine the respective contributions of

CmrR and CmrT to colony morphology, we overexpressed cmrR in ΔcmrT and

ΔcmrRΔcmrT backgrounds and examined colony morphology and surface motility.

Consistent with prior work showing a requirement for CmrT in rough colony

146 development, the ΔcmrT and ΔcmrRΔcmrT strains exclusively formed smooth colonies

(Figure 4.5B). Overexpression of cmrT in these strains as well as in WT restored the rough colony morphology. In contrast, while overexpression of cmrR causes rough colony formation in WT, cmrR does not restore the rough colony morphology in ΔcmrT and ΔcmrRΔcmrT. The same patterns arose when examining surface motility.

Overexpression of cmrT significantly increased surface motility in all strain backgrounds

(Figure 4.5C). However, while cmrR overexpression increased surface motility in WT, it was unable to promote surface motility to ΔcmrT or ΔcmrRΔcmrT strains. These data indicate that CmrT is essential for the rough colony morphology and surface motility, and CmrR cannot compensate for the loss of cmrT. Together, our results suggest that

CmrR primarily regulates motility and colony morphology through regulation of cmrT expression.

DISCUSSION

In this study, we characterized the CmrRST regulon. CmrRST regulates more than a hundred genes of diverse functions, indicating the significance of this TCS to C. difficile physiology. Because CmrRST is phase variable, its regulon is also subject to regulation by phase variation. Therefore, defining the transcriptional impact of CmrRST is important to understand the extent to which phase variation affects C. difficile physiology and pathogenesis as well as the scope of phenotypic heterogeneity within the bacterial population.

RNA-seq analysis indicates that CmrRST has wide ranging transcriptional effects. However, our analysis is complicated by several factors. First, CmrR regulates

147 cmrRST expression. Therefore, our experimental design cannot distinguish between targets of CmrR regulation versus those regulated by CmrT. Additionally, we found evidence of phase variation in other loci; the strains, and in some cases replicates of the same strain, differed in bias of some invertible DNA elements orientation within the populations. Differences in population-wide invertible element orientation is likely a result of stochastic differences between colonies saved as bacterial stocks and then propagated, or may have arisen during culturing for experimentation. For these genes, the observed differential expression may be due to phase variation rather than transcriptional regulation by CmrR or CmrT. Importantly, phase variation is a complication not just in this study but in all such studies of C. difficile and underscores the importance of accounting for phase variation in experiments.

CDR20291_1689 is one of the most highly upregulated genes by both CmrR and

CmrT (27.9- and 19.1-fold, respectively). Overexpression of this gene and its neighbor

CDR20291_1690, which together may constitute an operon, affected some but not all phenotypes known to be affected by CmrRST. A strain overexpressing

CDR20291_1689-1690 exhibited altered colony morphology and decreased swimming motility relative to a vector control but showed no difference in cell morphology or surface motility. These results indicate that CmrRST regulation of colony morphology and motility is the result of multiple independent factors and that these are discrete phenotypes that are not mediated by a single factor. Additionally, this result demonstrates that cell elongation and chaining are not required for the observed reduction in swimming motility. Further work will identify the factors that contribute to

148 CmrRST regulation of motility, cell morphology, and colony morphology as well as the interdependency of these behaviors.

CDR20291_1689 and CDR20291_1690 are highly conserved and found in all C. difficile genomes available through NCBI. Both genes are predicted to encode small proteins (81 and 87 amino acids, respectively). No homology to proteins with a defined function was detectable using BLASTp, and prediction of protein structure using Phyre was uninformative. These small proteins may act as post-translational regulators of other proteins, and it remains possible that the loci do not encode proteins. Future studies will characterize these gene products and their mechanisms of action in the regulation of swimming motility and colony morphology.

CmrR is an autoregulator of cmrRST expression. We found that eight of nine tested genes, including those with the highest fold changes found through RNA-seq, were regulated by CmrR through CmrT. Our data also indicate that CmrR is unable to restore colony morphology and surface motility phenotypes in a cmrT mutant. This data does not preclude the possibility that CmrR also regulates that production of factors that contribute to the rough morphology and surface motility. In previous work, a cmrR mutant, but not a cmrT mutant, was shown to have increased biofilm formation and was attenuated in a hamster model of infection. This indicates that CmrR may regulate important factors. However, our data do suggest that one mechanism by which CmrR positively regulates colony morphology and surface motility is through increasing cmrT expression.

In summary, this study begins to characterize the regulon of the putative TCS

CmrRST. Many questions remain, including the conditions that favor cmrRST

149 expression as well as the function and significance of the CmrR and CmrT regulated genes. Our results suggest that many genes may be regulated in a phase variable manner through CmrRST, leading to phase variation-mediated global regulatory changes and broad phenotypic heterogeneity. Phase variation of CmrRST likely has a vast impact C. difficile physiology, and elucidating the underlying mechanisms is essential to understanding the contribution of CmrRST phase variation to C. difficile pathogenesis.

150 FIGURES AND TABLES

Figure 4.1. Orientation of invertible DNA elements varies between strain populations. WT vector, pCmrR and pCmrT were grown overnight in broth without induction. qPCR was performed using gDNA and orientation specific primers to quantify percent invertible element orientation. Reference orientation refers to the orientation found in the published R20291 genome (FN545816.1). Show are individual values with mean of from 7-8 biological replicates over two independent experiments. ****, p <

0.0001, two-way ANOVA with Dunnett’s multiple comparisons post-test.

151

Figure 4.2. CmrR and CmrT regulate CDR20291_1689-1690. A) Diagram of

CDR20291_1689-1690 locus. B) qRT-PCR measurement of CDR20291_1689 and

CDR20291_1690 transcript abundance in WT with pCmrR, pCmrT, and vector control.

Strains were grown with ATc induction (10 ng/mL for vector and pCmrR, 2 ng/mL for pCmrT) prior to RNA isolation. Shown are means and standard deviations of 4-6 biological replicates from two independent experiments. **** p < 0.0001, one-way

ANOVA with Tukey’s multiple comparisons post-test.

152

Figure 4.3. CDR20291_1689-1690 regulates colony morphology. A) WT vector and pCDR20291_1689-1690 were grown on BHIS-Tm with the indicate concentrations of

ATc. Representative images of colony morphology after 24 hours of growth. B) WT vector, pCmrT and CDR20291_1689-1690 were grown on BHIS-Tm with ATc (vector and CDR20291_1689-1690, 50 ng/mL; pCmrT, 2 ng/mL). Colonies were Gram stained and imaged at 60x magnification.

153

Figure 4.4. CDR20291_1689-1690 inhibits swimming motility but not surface motility. Motility of WT vector, pCmrT and CDR20291_1689-1690 grown with the indicated concentration of ATc. n.d., no data; pCmrT growth is inhibited with more than

2 ng/mL ATc. Shown are means with standard deviations of eight biological replicates from two independent experiments A) Surface motility was quantified by measuring the diameter of growth after 72 hours. D) Swimming motility was quantified by measuring the diameter of growth in 0.5x BHIS 0.3% agar after 48 hours. ** p < 0.01, *** p < 0.001,

**** p < 0.0001, two-way ANOVA with Dunnett’s multiple comparisons post-test.

154

Figure 4.5. Gene regulation by CmrR occurs through CmrT. A) qRT-PCR analysis of transcript abundance for cmrRST targets in ΔcmrRΔcmrT strains overexpressing cmrR, cmrT, or vector control. Shown are means and standard deviations of 4-5 biological replicates. * p < 0.05, ** p < 0.01, **** p < 0.0001, Holm-Sidak t-test for multiple comparisons. B) Representative images of colony morphology of WT, ΔcmrT, and ΔcmrRΔcmrT, each carrying pCmrR, pCmrT or vector control. Colonies were grown

155 on BHIS-agar with ATc induction (10 ng/mL for vector, pCmrR; 2 ng/mL for pCmrT). C)

Surface motility after 72 hours of growth. Shown are means and standard deviations of eight biological replicates from two independent experiments. ***p < 0.001, ****p <

0.0001, one-way ANOVA with Tukey’s multiple comparisons post-test.

156 Table 4.1. Strains and plasmids used in this study.

Lab Notation Strain/Plasmid Name Description Reference AC472 Escherichia coli DH5α F- φ80lacZΔM15 Δ(lacZYA- Invitrogen argF)U169 recA1 endA1 hsdR17(rκ [313] -, mκ+) phoA supE44 thi-1 gyrA96 relA1 λ- tonA RT270 Escherichia coli E. coli used in conjugations with C. [314] HB101(pRK24) difficile, ApR, CmR RT273 C. difficile R20291 Ribotype 027 strain (Genbank [99] Accession # FN545816) RT1615 R20291 vector R20291 with pRT1611 [103] RT2085 R20291 pCmrR R20291 with pCmrR (pRT2073) [334] RT2107 R20291 pCmrT R20291 with pCmrT (pRT2106) [334] RT2518 R20291 p1689-1690 R20291 with p1689-1690 (RT2520) This work RT2307 R20291 ΔcmrRΔcmrT R20291 with deletions of cmrR and This work vector cmrT, contains plasmid RT1611 RT2308 R20291 ΔcmrRΔcmrT R20291 ΔcmrRΔcmrT with pCmrR This work pCmrR (RT2073) RT2309 R20291 ΔcmrRΔcmrT R20291 ΔcmrRΔcmrT with pCmrT This work pCmrT (RT2106) RT709 pRPF185 Contains ATc-inducible Ptet promoter [320] with gusA RT1611 pRT1611 Derivative of pRPF185 with gusA [103] removed RT2073 pCmrR pRPF185 with gusA replaced by [334] cmrR RT2106 pCmrT pRPF185 with gusA replaced by [334] cmrT RT2520 p1689-1690 pRPF185 with gusA replaced by This work p1689-1690

157 Table 4.2. Primers used in this study.

Primer Lab Notation Primer Sequence (5’ to 3’) Name R2175 qPCR_FlgSwit-ON GTTTTCTTACCAAAGTGATACATTATTATATTAATG R2176 qPCR_FlgSwit-OFF CATTAATATAATAATGTATCACTTTGGTAAGAAAAC R2177 qPCR_FlgSwit-REV GCTATTGTCTGACTTCTTAAATTAGTTGCAT R2213 LCF801 GGTAAGTTTGATTTTTATGTTAATGAATTG R2214 LCF714 CAGTTTGTGCACTAGCTATGCCTGC R2215 LCF796 CGCAATTATTTGTTTTTCATATGGATAAAATTGG R2216 LCF797 GATTTTTATGTTAATGAATTGTTATAAAAAACATGG R2261 CDR20291_0685- IEq1 GTTAAAAATTTAAGATATCTTTTCAGTATAATGGA R2262 CDR20291_0685- IEq2 CATTTCTAAGAAATATCCTAACATAAAAACAAAA R2263 CDR20291_0685- IEq3 CGATTACACTACAGAATTAGAATGTCAATG R2267 CDR20291_1514- IEq1 GATTTGTCGAAACCATTGTAATAAGA R2268 CDR20291_1514- IEq2 CAATAGTTAAGACAATGAATATGCTACATTCT R2269 CDR20291_1514- IEq3 GTAAATTCCTCATAAAAATTTCCTCCCA R2273 RpoAqF TCATTACCAGGTGTAGCAGTGAATGC R2274 RpoAqR GATAGAGCATGGTCCTTGAGCTTCT R2755 CDR_1689-1690F cgtagcgttaacagatctgagctcGAGATTGGAGGAATTGCTTTG R2756 CDR_1689-1690R gttttattaaaacttataggatccTCCAATATTCTAAAATTAATTTTG R2805 CDR20291_1689qF GAAGGGTATATATTAACATGTAAAATATCTG R2806 CDR20291_1689qR GCAACAACTTCTCCATCCCATAC R2807 CDR20291_1690qF GTAGAGGTAAATTTAGAGAAACAAGCC R2808 CDR20291_1690qR CTTGATTTACTACTAGATACCATGTTCC R2816 CDR20291_1911qF CAATTAGTAGATGCAATATCATCTGTAG R2817 CDR20291_1911qR GCACTTAATCCATCTGCAATAGAG R2818 CDR20291_1913qF GGGATAGCTATATCTTTTCTACTTATAATCAG R2819 CDR20291_1913qR CCTACTTTTATGTCTACTTTTAGATTTAATAG R2820 CDR20291_1914qF TTGGCAAATCTAGCATCAATTTTAAAAC R2821 CDR20291_1914qR CTCTCAACTTCTGCCCATATTCC R2822 CDR20291_3074qF GGTCTGATTTTGCCAGATGGAGG R2823 CDR20291_3074qR CAGTCATATGAGGAAAAAGAACACTTCC R2824 CDR20291_3075qF CTATAATAATTGCGATTGTTTTGGGTGG R2825 CDR20291_3075qR CAGAAATCCTAGCATAGAAATAGATGG R2828 CDR20291_1595qF CTTTTATCAGGAGCTGAAGACTGTC

158 R2829 CDR20291_1595qR ATAGCATCATACCTACCACTTTTAGCC R2830 CDR20291_2122qF ATATGGTTGGAGTATCGCCTTGG R2831 CDR20291_2122qF GGAAGTAATATCTTTTATAAGAACATCTGC

159 Table 4.3. Genes differentially regulated by CmrR and/or CmrT

Locus Tag (Gene Function pCmrR/vector pCmrT/vector Name) Fold Change Fold Change

CDR20291_3126 cmrT response regulator 6.892 14.498 CDR20291_3127 cmrS histidine kinase 7.053 1.402 CDR20291_3128 cmrR response regulator 17.696 1.506 Flagella CDR20291_0224 glucose-1-P 2.144 1.79 thymidylyltransferase CDR20291_0225 dtdp-4-dehydrorhamnose 2.623 1.868 3,5-epimerase CDR20291_0226 dtdp-glucose 4,6- 2.418 1.706 dehydratase CDR20291_0227 putative transglycosylase 2.551 1.668 CDR20291_0228 conserved hypothetical 2.288 1.736 CDR20291_0229 fliN_1 flagellar motor switch protein 2.338 1.662 CDR20291_0230 flgM anti-SigD factor 2.199 1.9 CDR20291_0231 putative flagellar 2.505 1.948 biosynthesis protein CDR20291_0232 flgK putative flagellar hook- 2.418 1.707 associated protein CDR20291_0233 flgL flagellar hook associated 2.475 1.725 protein CDR20291_0234 fliW flagellar assembly factor 2.539 1.68 CDR20291_0235 csrA carbon storage regulator 2.741 1.874 CDR20291_0236 fliS1 flagellar protein FliS 2.537 1.791 CDR20291_0237 fliS2 flagellar protein FliS 2.485 1.757 CDR20291_0238 fliD flagellar cap protein 2.487 1.687 CDR20291_0239 conserved hypothetical 2.532 1.768 protein CDR20291_0240 fliC flagellin 2.631 2.21 CDR20291_0241 putative glycosyltransferase 3.076 2.043 CDR20291_0242 glycosyl transferase family 2 2.913 1.941 CDR20291_0243 glycosyl transferase family 2 3.114 1.875 CDR20291_0244 putative uncharacterized 2.971 1.67 protein CDR20291_0245 putative carbamoyl- 2.82 1.59 phosphate-synthetase CDR20291_0246 putative ornithine 2.156 1.343 cyclodeaminase CDR20291_0247 putative uncharacterized 2.09 1.445 protein

160 CDR20291_0248 flgB flagellar basal body rod 2.312 1.553 protein CDR20291_0249 flgC flagellar basal body rod 2.349 1.538 protein CDR20291_0250 fliE flagellar hook-basal body 2.143 1.545 complex protein CDR20291_0251 fliF flagellar M-ring protein 2.112 1.611 CDR20291_0252 fliG flagellar motor switch protein 1.896 1.79 CDR20291_0253 fliH flagellar assembly protein 2.203 1.911 CDR20291_0254 fliI flagellum-specific ATP 2.014 1.892 synthase CDR20291_0255 fliJ flagellar protein 2.298 1.993 CDR20291_0256 fliK putative flagellar hook length 2.219 2.089 control protein CDR20291_0257 flgD putative basal body rod 2.157 1.946 modification protein CDR20291_0258 flgE flagellar hook protein 1.966 1.947 CDR20291_0259 putative flagellar protein 1.773 1.728 CDR20291_0260 motA flagellar motor protein, 1.915 2 chemotaxis protein CDR20291_0261 motB flagellar motor protein, 2.016 1.995 chemotaxis protein CDR20291_0262 fliL flagellar basal body 2.248 2.137 associated protein CDR20291_0263 putative flagellar protein 1.877 2.136 CDR20291_0264 fliP flagellar biosynthesis protein 1.849 2.093 CDR20291_0265 fliQ flagellar export protein 2.143 2.034 CDR20291_0266 flhB flagellar export protein 1.8 1.998 CDR20291_0267 flhA flagellar export protein 1.795 1.909 CDR20291_0268 flhF flagellar biosynthesis protein 1.808 1.914 CDR20291_0269 fleN flagellar number regulator 1.859 1.852 CDR20291_0270 sigD flagellar sigma factor 1.931 1.879 (SigD/FliA) CDR20291_0271 putative exported protein 1.86 1.833 CDR20291_0272 flgG flagellar basal body rod 1.871 1.86 protein CDR20291_0273 putative flagellar basal body 2.143 2.146 rod protein CDR20291_0274 fliM putative flagellar motor 2.368 2.273 switch protein CDR20291_0275 fliN_2 putative flagellar motor 2.324 2.244 switch protein CDR20291_0276 conserved hypothetical 2.665 2.244 protein Cell wall and membrane

161 CDR20291_0008 processive 1,2-diacylglycerol -2.231 1.095 beta-glucosyltransferase CDR20291_0440 cwpV cell wall protein -6.406 -7.821 CDR20291_0971 Putative cell wall hydrolase -2.778 -1.719 CDR20291_1145 acd endo-beta-N- -3.114 -2.031 acetylglucosamidase CDR20291_1911 cwp28 cell surface protein 37.172 11.158 CDR20291_1913 putative membrane protein 5.024 4.201 CDR20291_1914 hypothetical protein 18.311 10.767 CDR20291_2099 putative cell surface protein -2.47 1.212 CDR20291_2505 putative membrane protein -2.112 -1.478 CDR20291_2677 putative cell wall teichoic -2.797 -1.545 acid glycosylation protein CDR20291_2679 hypothetical protein -2.457 -1.864 CDR20291_2680 cwp2 cell surface protein (S-layer -2.25 -1.541 protein precursor) CDR20291_2683 cwp12 cell surface protein 2.39 1.863 CDR20291_2684 cwp11 cell surface protein 2.892 2.172 CDR20291_2685 cwp10 cell surface protein -2.087 -1.723 Energy and metabolism CDR20291_0176 anaerobic carbon-monoxide -2.093 -2.815 dehydrogenase iron sulfur subunit CDR20291_0177 putative oxidoreductase, -2.111 -3.317 NAD/FAD binding subunit CDR20291_0517 putative membrane protein -1.104 -2.012 CDR20291_0970 gloA glyoxylase I -2.745 -1.584 CDR20291_0972 putative membrane protein -2.429 -1.417 CDR20291_1016 plsX fatty acid/phospholipid -2.102 1.214 synthesis protein CDR20291_1477 hom2 homoserine dehydrogenase 2.308 1.342 CDR20291_1493 cysA serine acetyltransferase 2.901 1.49 CDR20291_1595 ribH riboflavin synthase beta -5.286 1.589 chain CDR20291_1596 ribA riboflavin biosynthesis -6.636 1.433 protein CDR20291_1597 ribB riboflavin synthase alpha -7.918 1.333 chain CDR20291_1598 ribD riboflavin biosynthesis -7.112 1.368 protein CDR20291_1829 pduU ethanolamine utilization -4.218 -1.661 protein CDR20291_2074 hcp hydroxylamine reductase 1.231 -5.357 CDR20291_2075 Iron-sulfur binding protein 1.164 -5.247 CDR20291_2507 putative alanine racemase -2.379 -1.614

162 CDR20291_2732 putative amidohydrolase 1.658 2.057 CDR20291_2829 nrdF ribonucleoside-diphosphate -1.654 -2.222 reductase beta-chain CDR20291_2838 uxaA altronate dehydratase large 2.126 1.386 subunit CDR20291_2839 altronate dehydratase small 3.502 1.365 subunit CDR20291_2934 bglA1 1-phospho-beta-glucosidase -2.007 -1.681 CDR20291_3142 proC2 pyrroline-5-carboxylate 2.174 1.109 reductase CDR20291_3143 pflD putative formate 2.45 1.025 acetyltransferase CDR20291_3144 pflE putative pyruvate formate- 2.428 -1.034 lyase 3 activating enzyme CDR20291_3226 cls cardiolipin synthase 2.135 1.745 Transporters CDR20291_0146 riboflavin transport -3.665 1.031 CDR20291_0516 putative cation transporting -1.374 -2.306 ATPase CDR20291_0946 zupT zinc transporter -1.446 -2.221 CDR20291_1328 feoB1 ferrous iron transport protein -1.349 -2.009 B CDR20291_1548 putative iron compound ABC -1.317 -2.567 transporter, substrate binding protein CDR20291_1718 P-type Ca2+ transporter -2.059 -2.164 ATPase CDR20291_2535 sbp sulfate ABC transporter -2.497 -1.789 binding protein CDR20291_2840 kdgT2 putative carbohydrate 2.719 1.491 permease CDR20291_2935 PTS beta-glucoside specific -2.045 -1.24 IIA component CDR20291_3074 putative osmoprotectant acid -8.448 -7.299 ABC transporter ATP binding protein CDR20291_3075 osmoprotectant transport -8.316 -6.946 system substrate binding protein CDR20291_3236 sn-glycerol 3-phosphate -2.535 -1.509 transport system substrate- binding protein Regulators CDR20291_0205 putative signaling protein, -2.346 -1.181 GGDEF-EAL CDR20291_0685 putative signaling protein, 24.903 1.114 GGDEF-EAL

163 CDR20291_1015 fapR fatty acid/phospholipid -2.371 1.278 biosynthesis regulator (DeoR family) CDR20291_1229 GntR family transcriptional -2.09 -1.715 regulator CDR20291_1514 putative signaling protein, -8.57 -13.188 GGDEF-EAL CDR20291_2040 putative signaling protein, 8.02 2.992 GGDEF-EAL CDR20291_2122 sinR’ putative regulatory protein 3.586 1.234 CDR20291_2503 cell division initiation protein -2.073 -1.452 CDR20291_2504 putative RNA binding cell -2.002 -1.578 division protein CDR20291_2506 sepF cell division inhibitor -2.455 -1.661 CDR20291_3012 conserved hypothetical 2.057 1.107 protein Miscellaneous CDR20291_0522 cotJB1 spore coat protein -2.917 -3.948 CDR20291_0684 putative reductase -2.165 -1.77 CDR20291_1329 putative exported protein -2.196 -2.298 CDR20291_1689 hypothetical protein 27.52 19.109 CDR20291_1690 conserved hypothetical 2.518 2.382 protein CDR20291_1929 hypothetical protein -3.89 -1.828 CDR20291_2116 putative GTP binding protein -2.695 -1.533 CDR20291_2508 putative exported protein -2.103 -1.627 CDR20291_2828 conserved hypothetical -1.852 -2.163 protein CDR20291_3190 conserved hypothetical -1.909 -2.368 protein CDR20291_3417 conserved hypothetical -2.444 -1.933 protein

Grey text, less than 2-fold change Green text, >2-fold up-regulated by CmrR/CmrT Red text, >2-fold down-regulated by CmrR/CmrT Italics, not significant (FDR p value > 0.05)

164 CHAPTER 5. DISCUSSION

This project began with a simple observation: Clostridioides difficile forms two distinct colony morphologies. Further investigation into this observation revealed the regulatory mechanisms underlying these colony morphologies as well as their significance to C. difficile physiology. In Chapter 2, we describe our discovery that these colony morphologies, called rough and smooth, gain their morphology from the phase variable expression of three genes, cmrRST, encoding a putative two-component system (TCS) with two response regulators and a histidine kinase. The response regulators CmrR and CmrT positively regulate type IV pili (TFP)-independent surface motility, which has not previously been described in C. difficile. Additionally, CmrR and

CmrT influence swimming motility, cell morphology, and biofilm formation. Using an animal model, we found that CmrRST has a role in virulence and that phase variation occurs in vivo. In Chapter 3, we characterized the regulation of cmrRST. In addition to regulation by site-specific recombination of an invertible DNA element, termed the cmr switch, cmrRST is transcriptionally regulated by c-di-GMP through a riboswitch.

Regulation by this riboswitch occurs whether the cmr switch is ON or OFF, and the two regulatory mechanisms occur independently of each other. We also found that CmrR, but not CmrT, regulates cmrRST expression. In Chapter 4, we defined the CmrRST regulon through RNA-seq. We identified multiple gene targets of CmrT including one that contributes to the CmrRST-mediated regulation of swimming motility and colony

165 morphology. Together, our data indicate that cmrRST expression is subject to a complex network of regulation with extensive and important impacts on multiple aspects of C. difficile physiology (Fig 5.1).

GENE REGULATION BY CmrR AND CmrT

TCS response regulators and sensor kinases typically exist as cognate pairs; thus, the presence of two response regulator encoding genes with a single kinase gene within one operon is unusual [362]. CmrR and CmrT are similar in structure, with both containing a receiver (REC) domain and a winged helix-turn-helix domain indicating an ability to bind DNA and regulate gene expression. CmrR and the family-defining response regulator OmpR share significant homology. However, key residues within the receiver domain of CmrT, including the conserved residue at the phosphorylation site, are divergent from the OmpR consensus sequence. OmpR-family response regulators are typically phosphorylated at a conserved aspartate residue [285]. CmrR contains this residue, D52, suggesting that CmrR is regulated though phosphorylation, presumably by CmrS. Consistent with the role of this conserved residue in signaling, we found that a

D52A mutation decreased CmrR activity. In contrast, CmrT has a glutamate at this position (E53). Glutamate substitutions at this site can function as a phosphomimic, removing the requirement for phosphorylation and resulting in a constitutively active response regulator [286, 287]. Therefore, while CmrR may require phosphorylation for full activity, CmrT may instead be constitutively active.

Our data show that CmrR and CmrT are not redundant. Overexpression of cmrR or cmrT result in similar phenotypic changes. However, CmrR is not able to regulate

166 these phenotypes in the absence of CmrT. Similarly, we identified genes that were differentially regulated with the overexpression of cmrR and cmrT. Further analysis of a few of these targets found that CmrR was unable to regulate expression upon deletion of cmrT, suggesting that CmrT is the primary regulator of genes involved in CmrRST- regulated phenotypes. CmrR, on the other hand, autoregulates cmrRST expression.

Based on these data, we hypothesize that the primary role of CmrR is to promote transcription of cmrT and that CmrT is a constitutively active transcriptional regulator that, once produced, regulates genes outside the cmrRST operon. Whether CmrR has any direct targets other than cmrRST is unclear; RNA-seq revealed several genes that were differentially regulated by overexpression of cmrR but not cmrT, but we were unable to validate these results for the two targets that were tested. Further analysis of our RNA-seq data set as well as the mechanisms underlying CmrRST-dependent phenotypes may reveal additional roles for CmrR.

SIGNALS FOR THE REGULATION OF cmrRST EXPRESSION

cmrRST expression is subject to regulation through three mechanisms: c-di-GMP signaling through a riboswitch, phase variation through site-specific DNA recombination, and autoregulation by the TCS (Fig. 5.1). In Chapter 3, we characterized how c-di-GMP, phase variation, and CmrR interact to control cmrRST expression. We found that orientation of the cmr switch is independent of regulation by c-di-GMP or CmrR. C-di-

GMP promotes cmrRST expression from a promoter upstream of the cmr switch, regardless of its orientation. Similarly, CmrR promotes expression from both orientations of the cmr switch, though the degree (fold-change) of activation is highest

167 from the OFF orientation. These overlapping regulatory strategies suggest that cmrRST expression can be activated in response to multiple signals and that coordinated activation of cmrRST is critical. Understanding the signals that regulate cmrRST expression may elucidate the conditions under which CmrRST provides a benefit to bacterial fitness.

C-di-GMP-modulating diguanylate cyclases (DGCs) and phosphodiesterase

(PDEs) are often activated in response to environmental signals through sensor domains; therefore, these environmental signals would indirectly contribute to regulation to cmrRST [213-216]. C. difficile has a large c-di-GMP network, with 37 genes in strain

630 encoding predicted DGCs and PDEs, consisting of nearly 1% of its total genome

[238, 239]. Based on two separate studies expressing these genes in heterologous bacteria, 28 DGCs and PDEs have been demonstrated to be enzymatically active [238,

239]. The majority of these DGCs and PDEs remain uncharacterized; their expression patterns as well as activating signals are unknown. However, most of these enzymes contain sensor domains that may be tied to their activity. Six DGC/PDE proteins contain one or more PAS domains [238]. PAS domains are sensor domains that sense a wide range of signals including ligand binding and can contribute to activation and dimerization or proteins [363]. One DGC contains a REC domain, indicating that the activity of this enzyme may be regulated by phosphorylation [345]. Two PDEs contain

Cache domains, which are often extracellular domains that contribute to regulation of protein activity in response to small molecule binding [364]. One PDE contains a PTS

EIIC domain, indicating that it may bind a carbohydrate [365]. Finally, a DGC and a PDE each contain solute-binding protein (SBP) family 3 domains, which participate in signal

168 transduction through the binding of small molecules including polar amino acids [366].

Together, the number of DGCs and PDEs and the diversity of their accessory domains suggest that c-di-GMP signaling in C. difficile is subject to regulation by many signals, potentially affecting cmrRST.

The significance of sensory domains can be illustrated through the regulation of the C. difficile PDE PdcA, which contains a PAS domain and a degenerate GGDEF domain [120]. Removal of the PAS and GGDEF domains decreased enzyme activity.

PdcA activity is regulated by GTP, the intracellular level of which is affected by nutrient availability and growth phase. Additionally, pdcA expression is regulated by CodY, a global transcription factor that represses gene transcription during growth in nutrient-rich conditions such as exponential phase. While the signals that regulate the activity and expression of other C. difficile DGC and PDE are unknown, Purcell et al. showed that c- di-GMP increases with surface growth versus liquid [113]. These examples lead us to propose that other DGC and PDE are subject to similar regulation, potentially tying cmrRST expression to the same environmental signals. One open question is which c- di-GMP enzymes can regulate cmrRST expression and whether any are dedicated to this regulation.

Rather than a sense-and-respond mechanism, phase variation relies on environmental selective pressures on the population as a whole. Accordingly, environmental pressures determine whether cmr ON or OFF cells become the predominant subpopulation. We found that motile spots on a hard agar surface originating from a cmr OFF inoculum develop cmr ON cells on the edges while remaining cmr OFF in the center. One explanation for this is that this growth condition

169 applies selective pressure, perhaps nutrient availability, for the ON orientation.

Inversely, in soft agar, a cmr ON inoculum develops into a predominantly cmr OFF population that is able to swim. These data indicate that the orientation of the cmr switch that allows for the appropriate motility and optimal nutrient acquisition mechanisms the greatest fitness benefit in the given environment. While selective pressures in vitro may be simple, they are undoubtedly more complex and multivariate in vivo. In an acute model of infection, hamsters were inoculated with cmr ON bacteria, and analysis of DNA recovered from ceca at the end point of infection revealed that the cmr switch was predominantly OFF. This result indicates that there is selective pressure against cmr ON cells during infection, at least in this animal model and at the end point of disease. That the cmrRST locus is highly conserved and considered part of the C. difficile core genome implies a selective advantage to maintenance of the locus for survival in the large bowel, the only environment in which C. difficile is known to grow naturally. Future studies may define the environmental factors that contribute to this selective pressure and provide clues as to the function of CmrRST in infection.

CmrR regulates the expression of cmrRST. While many response regulators can bind DNA without phosphorylation, phosphorylation can increase activity and binding affinity [367, 368]. As a response regulator, CmrR may be activated by the sensor kinase CmrS. Key domains for kinase activity and phosphotransfer are well conserved in CmrS, suggesting that it is an active histidine kinase [369]. Additionally, CmrS contains a transmembrane domain typical of sensor kinase with extracellular sensory domains. However, the putative sensory domain does not share significant homology with any characterized proteins; histidine kinase sensory domains are extremely diverse

170 and the activating signal is difficult to predict through sequence alone [369, 370].

Additionally, TCS can be regulated by multiple factors; one study estimated that 15-25% of cognate response regulator/sensor kinase pairs undergo regulatory crosstalk with other TCS [362]. The B. subtilis response regulator Spo0F is phosphorylated by five distinct kinases [371], and CmrR similarly may be activated by multiple kinases. CmrS activity also may be regulated by other membrane proteins. For example, the sensor kinase GacS is regulated by interactions with two other sensor kinases, each of which is regulated by additional protein interactions as well as environmental signals, illustrating how complex sensor kinase activation can be [350]. Together, these studies indicate interactions between TCS components can be subject to multiple interactions and signals. Further work characterizing the phosphorylation requirement of CmrR, the conditions under which CmrR is active, and the signals to which CmrS responds will provide important details about the regulation of CmrRST.

Because CmrR is an autoregulator, it may initiate a positive feedback loop. This mechanism would require low amounts of CmrR and possibly CmrS to be present in the cell to initiate autoactivation. For example, in S. enterica, the TCS-encoding genes phoPQ are expressed from a constitutive promoter, leading to low levels of PhoP and

PhoQ [338, 372, 373]. If this TCS is activated by changes in Mg2+, PhoP promotes phoPQ expression in a positive feedback loop. We identified a transcriptional start site, designated TSS4, between the cmr switch and the start of cmrR, and transcriptional reporter experiments indicate the TSS4 promoter has weak activity. TSS4 thus may mark the location of a constitutive promoter, allowing low levels of cmrRST transcription which can be amplified upon activation of CmrS/R. Alternatively, cmrR expression could

171 originate from the promoter upstream of the c-di-GMP riboswitch under permissive conditions, similarly perpetuating cmrRST expression in a positive feedback loop.

These overlapping mechanisms of regulation raise questions about the relative contributions of these regulatory mechanisms to cmrRST expression and under what conditions they are relevant (Fig. 5.1). The cmr switch likely inverts stochastically at a low rate per generation, with cmr ON and OFF subpopulations coexisting as a bet- hedging strategy against sudden environmental changes that provide a fitness benefit to one subpopulation over the other. In contrast to this passive form of regulation, c-di-

GMP signaling and CmrR regulation serve as sense-and-respond mechanisms that allow each bacterium to act individually in response to its environment. Notably, the phenotypic heterogeneity provided by phase variation provides an existent advantage to subpopulations, which the c-di-GMP and CmrR autoregulatory mechanisms follow some lag as the bacteria modify gene expression. That multiple discrete, independent mechanisms impact cmrRST transcription implies the need of C. difficile to failsafe expression of this signaling system and its regulatory targets. Further experiments will investigate the conditions under which each of these transcriptional regulatory mechanisms benefits C. difficile fitness.

REGULATION OF MOTILITY, CELL MORPHOLOGY, AND COLONY MORPHOLOGY BY CmrRST

CmrRST negatively regulates swimming motility, but the underlying mechanism is unknown. qRT-PCR analysis as well as RNA-seq indicated that overexpression of cmrR and cmrT did not decrease transcript abundance of flagellar genes. Additionally,

172 scanning electron microscopy (SEM) revealed structures resembling flagella on the surface of cmr ON cells [334]. These data suggest that CmrRST may regulate swimming through the post-translational regulation of flagella. There are several mechanisms through which post-translational regulation of flagella could occur. C. difficile flagella are glycosylated, and loss of glycosylation has been shown to inhibit swimming motility [98, 100]. While genes encoding flagellar glycosyltransferases are not downregulated by CmrR or CmrT, their activity could be. Alternatively, many bacteria demonstrate regulation of flagellar motor or stator activity [374]. For example, YcgR is a c-di-GMP binding protein produced by E. coli and S. Typhimurium that inhibits swimming motility by directly binding MotA, a component of the flagellar motor, and decoupling it from other components of the motor [223, 375]. Flagella can also be sterically hindered by extracellular factors such as exopolysaccharides. Cellulose production was found to contribute to swimming inhibition in addition to YcgR in S.

Typhimurium [376]. Overexpression of CDR20291_1689 partially reduced swimming motility, which may indicate that multiple factors contribute to CmrRST-mediated inhibition of motility. This will be addressed in future studies.

Type IV pili (TFP)-mediated surface motility has been well demonstrated in C. difficile [113, 114, 237]. However, CmrRST positively regulates surface motility in a

TFP-deficient strain. Additionally, CmrRST does not transcriptionally regulate TFP genes. Together, these results indicate that C. difficile is capable of an additional TFP- independent surface motility that has not been previously identified. Given that CmrRST negatively regulates flagella-mediated swimming motility, this novel mechanism of surface motility is likely independent of flagella as well. Several mechanisms of pili- and

173 flagella-independent motility are exhibited by other bacteria [377-379]. During sliding motility, bacteria migrate through the pushing force of cell division [377]. This can require additional factors. Several bacteria, including Pseudomonas species and

Serratia marcescens, produce surfactants that facilitate movement [377]. Sliding motility in Bacillus subtilis requires the production of exopolysaccharide (EPS) as well as surfactant [304]. Cell-to-cell contact can also contribute to sliding motility. B. subtilis forms bundles of cells connected pole-to-pole that, through cell division, expand in a loop facilitated by surfactant that results in migration [304]. Mycobacteria such as M. smegmatis also exhibit sliding motility. These bacteria produce glycopeptidolipids in the cell envelope that allow sliding also facilitated by cell contact and the formation of pseudofilaments [380]. C. difficile may undergo motility through similar mechanisms.

CmrRST positively regulates cell length and the formation of cell chains, consisting of cells joined pole-to-pole. The mechanisms underlying these phenotypes are unclear but may reflect altered cell division. Alterations in the cell cycle and the regulators of cell division can increase cell length [381]. Additionally, cell chaining can be a results of incomplete daughter cell separation during division [311, 382, 383]. RNA- seq analysis of CmrRST targets identified several genes whose products are annotated as inhibitors of cell division (CDR20291_2503-2508). Further study will determine the contribution of these genes to cell morphology. Alternatively, cell chaining can be the result of extracellular factors that bind cells pole-to-pole. In E. coli, the phase variable factor Ag43 contributes to cell chaining in biofilms [291]. CmrRST regulates the expression of several genes encoding cell wall proteins that may function in cell adhesion.

174 One unanswered question is whether the changes in surface motility, swimming motility, and cell morphology are interdependent. As described above, cell chaining is important to surface motility for many bacteria. The arrangement of cells during B. subtilis sliding motility resembles that of cells during C. difficile CmrRST-mediated surface motility [304]. Cell length and chaining could also affect the ability of the bacterium to swim by affecting its physical properties and slowing its ability to move through viscous substances [290]. Alternatively, production of an EPS or other matrix component could explain the regulation of both surface and swimming motility.

Overexpression of CDR20291_1689 partially inhibited swimming motility and altered colony morphology but did not affect surface motility. However, differences between these results versus those from the overexpression of cmrT indicate that

CDR20291_1689 is not fully responsible for CmrRST-mediated regulation of these phenotypes, and therefore it remains possible that surface motility, swimming motility and cell morphology are interconnected. Further analysis of CmrRST targets will allow us to disentangle these phenotypes.

IMPLICATIONS OF CmrRST PHASE VARIATION FOR PATHOGENESIS

An important part of understanding the role of CmrRST in infection is to understand the contributions of cell morphology and motility to pathogenesis. Cell morphology can be a key virulence determinant for bacterial pathogens. Cell morphology physically affects the way a bacterium interacts with its environment, neighboring bacteria, and the host [384]. The increased surface area of elongated or chained cells can affect targeting by the immune system. Several studies have found

175 that filamentous cells are resistant to phagocytosis by macrophages [385, 386]. In contrast, Salamaga et al. showed that Enterococcus faecalis strains that form long cell chains were severely attenuated in an animal model and were more susceptible to phagocytosis [383]. Filamentous cells can be more susceptible to targeting by the complement system due to increased surface area. Increased chain length in

Streptococcus pneumoniae was shown to increase opsonization and phagocytosis by neutrophils [387]. Furthermore, short-chain S. pneumoniae were more virulent than long-chain cells in a model of systemic infection. Together these studies indicate that cell morphology plays an important role in bacterial interactions with immune cells. Cell chaining and length can also affect the interaction of bacteria with surfaces including host cells. S. pneumoniae chains were shown to increase adherence to human epithelial cells, and strains with increased chain length outcompeted short-chain strains in a model of respiratory colonization [388]. Prashar et al. demonstrated that filamentation of Legionella pneumophila modifies its interaction with host epithelial cells in order to promote invasion, after which it differentiates into short rods [389]. These studies underscore that the consequences of cell morphology can be context dependent. Many pathogens exhibit heterogeneity of cell morphology in order to balance benefits and costs in response to host stresses. Therefore, these studies highlight why the phase variation of cell morphology as demonstrated in C. difficile could allow the bacterium to bet-hedge in order to increase fitness.

Motility is also a key virulence determinant and an essential aspect of pathogenesis for many bacteria, and therefore many are capable of multiple types of motility specialized for planktonic and surface-associated states [390-392]. The

176 transition between these states and these motility patterns is key to bacterial fitness. For example, Proteus mirabilis is capable of both swimming motility and swarming motility, a flagellum- and surfactant-dependent form of surface motility [393]. P. mirabilis reversibly differentiates into swarmer cells in response to inhibition of flagellar rotation, surface contact, and cell-to-cell contact, initiating changes not only in motility but in the regulation of virulence factors [393, 394]. Swarmer cells have been shown to allow P. mirabilis to migrate across catheters to establish urinary tract infections [395, 396].

Taylor et al. demonstrated the importance of having multiple motility strategies in

Pseudomonas aeruginosa [397]. This study used different surface conditions to evolve isogenic strains that exhibited increased surface or swimming motility, then showed that these strains exhibited trade-offs in motility and virulence. The ability to switch between motility types has been shown to be an important aspect of P. aeruginosa colonization.

Laventie et al. demonstrated that surface contact of the bacterium increased c-di-GMP, promoting TFP-mediated surface attachment and motility [193]. Through an asymmetric distribution of c-di-GMP following cell division, surface associated populations release swimming cells that can travel farther distances to establish new sites of colonization.

Therefore, both surface-associated and swimming cells are coordinated to maximize bacterial colonization. These examples demonstrate the ability to adapt with different motility strategies, and not just motility itself, can contribute to bacterial pathogenesis.

Characterization of the role of CmrRST in virulence will be central to future work.

In a hamster model of acute infection, a cmrR mutant strain was highly attenuated, suggesting an important role for CmrRST in virulence. The cmrR mutant was not recovered from the feces at any point during infection, which may indicate that CmrRST

177 contributes to colonization. Additionally, we found evidence that phase variation of this system occurs in vivo. We inoculated hamsters with bacterial populations in which the cmr switch was predominantly ON and OFF, individually. Regardless of the starting orientation, DNA from the hamster cecum at the time of sacrifice showed that the cmr switch was predominantly in the OFF orientation, indicating selective pressure against cmr ON bacteria in the late stages of infection. These cecal samples do suggest that a minority cmr ON population remains. Many open questions remain regarding the role of

CmrRST in infection. Together, our data suggest that CmrRST may be important for colonization and the early stages of infection and detrimental to overall fitness during late infection. Future studies will assess the spatial and temporal dynamics of cmr phase variation over the course of infection. Competition experiments with phase locked and unlocked strains may provide details about the significance of phase variation over time. Additionally, cmr switch orientation may affect the spatial organization of bacteria and whether they are motile or surface associated. Finally, understanding the signals regulating cmrRST expression and activity may provide important information about the significance c-di-GMP signaling and CmrR autoregulation during infection.

Based on all our findings, I propose the following model: C. difficile populations consist of cmr ON and cmr OFF subpopulations (Fig. 5.1). During infection, cmr ON subpopulations form bacterial chains that are associated with the surface of the host epithelia or perhaps mucus layer. CmrRST-mediated surface motility allows migration across surfaces, potentially aiding access to nutrients. CmrR may become activated by

CmrS, promoting cmrRST expression and surface migration in response to nutrient cues. Regulation by c-di-GMP, which increases with surface contact, may reinforce

178 cmrRST expression in surface associated cells and activate cmrRST expression in phase OFF cells that become surface associated. The cmr OFF cells are capable of swimming motility which may allow rapid movement to more distant areas, propagating the colonization of new sites. The fitness benefits of cmr ON surface associated cells versus cmr OFF planktonic cells likely changes over the course of infection as the epithelial surface is disrupted by toxins and the host immune response develops.

Elongated cmr ON cell chains may be more susceptible to targeting by immune cells, consistent with our data showing bacterial populations to be predominantly phase OFF at the endpoint of infection. Many open questions remain, and much work will be needed to address multiple parts of this proposed model. Continued investigation of

CmrRST will provide important insights into C. difficile pathogenesis including the identification of important environmental signals to which the bacteria respond, characterization of behaviors significant through the course of infection, and understanding of the fitness benefits of phase variation and phenotypic heterogeneity in this critical human pathogen.

179

180 Figure 5.1. Model of cmrRST regulation and the role of CmrRST in infection. PDE, phosphodiesterase; DGC, diguanylate cyclase; R, CmrR; S, CmrS; T, CmrT; P, phosphorylation; Lightning bolt represents an activating signal. 1) With the cmr switch in the OFF orientation and basal c-di-GMP levels, cmrRST is not expressed. cmrRST expression can be activated through three mechanisms: 2) increased c-di-GMP levels through activity of DGC; 3) inversion of the cmr switch to the ON orientation; 4) activation of CmrR by CmrS. 5) During infection, C. difficile may consist of cmr OFF, swimming cells and cmr ON, surface associated cells.

181 REFERENCES

[1] Martin JS, Monaghan TM, Wilcox MH. Clostridium difficile infection: epidemiology, diagnosis and understanding transmission. Nat Rev Gastroenterol Hepatol. 2016;13(4):206-16.

[2] Karen C C, John G B. Biology of Clostridium difficile: implications for epidemiology and diagnosis. J Annu Rev of Microbiol. 2011;65:501-21.

[3] Smits WK, Lyras D, Lacy DB, Wilcox MH, Kuijper EJ. Clostridium difficile infection. Nat Rev Dis Primers. 2016;2:16020.

[4] CDC. Antibiotic Resistance Threats in the United States, 2019. 2019.

[5] McFarland LV, Ozen M, Dinleyici EC, Goh S. Comparison of pediatric and adult antibiotic-associated diarrhea and Clostridium difficile infections. World J Gastroenterol. 2016;22(11):3078-104.

[6] McFarland LV, Mulligan ME, Kwok RY, Stamm WE. Nosocomial acquisition of Clostridium difficile infection. N Engl J Med. 1989;320(4):204-10.

[7] Khanna S, Pardi DS, Aronson SL, Kammer PP, Orenstein R, St Sauver JL, et al. The epidemiology of community-acquired Clostridium difficile infection: a population-based study. Am J Gastroenterol. 2012;107(1):89-95.

[8] Ofori E, Ramai D, Dhawan M, Mustafa F, Gasperino J, Reddy M. Community- acquired Clostridium difficile: epidemiology, ribotype, risk factors, hospital and intensive care unit outcomes, and current and emerging therapies. J Hosp Infect. 2018;99(4):436-42.

[9] Lessa FC, Mu Y, Bamberg WM, Beldavs ZG, Dumyati GK, Dunn JR, et al. Burden of Clostridium difficile infection in the United States. N Eng J Med. 2015;372(9):825-34.

[10] Paredes-Sabja D, Shen A, Sorg JA. Clostridium difficile spore biology: sporulation, germination, and spore structural proteins. Trends Microbiol. 2014;22(7):406-16.

[11] Theriot CM, Koenigsknecht MJ, Carlson PE, Jr., Hatton GE, Nelson AM, Li B, et al. Antibiotic-induced shifts in the mouse gut microbiome and metabolome

182 increase susceptibility to Clostridium difficile infection. Nat Commun. 2014;5:3114.

[12] Theriot CM, Young VB. Interactions between the gastrointestinal microbiome and Clostridium difficile. Annu Rev Microbiol. 2015;69:445-61.

[13] Schubert AM, Sinani H, Schloss PD. Antibiotic-induced alterations of the murine gut microbiota and subsequent effects on colonization resistance against Clostridium difficile. mBio. 2015;6(4):e00974.

[14] Buffie CG, Jarchum I, Equinda M, Lipuma L, Gobourne A, Viale A, et al. Profound alterations of intestinal microbiota following a single dose of clindamycin results in sustained susceptibility to Clostridium difficile-induced colitis. Infect Immun. 2012;80(1):62-73.

[15] Theriot CM, Bowman AA, Young VB. Antibiotic-induced alterations of the gut microbiota alter secondary bile acid production and allow for Clostridium difficile spore germination and outgrowth in the large intestine. mSphere. 2016;1(1).

[16] Winston JA, Theriot CM. Impact of microbial derived secondary bile acids on colonization resistance against Clostridium difficile in the gastrointestinal tract. Anaerobe. 2016;41:44-50.

[17] Reeves AE, Koenigsknecht MJ, Bergin IL, Young VB. Suppression of Clostridium difficile in the gastrointestinal tracts of germfree mice inoculated with a murine isolate from the family Lachnospiraceae. Infect Immun. 2012;80(11):3786-94.

[18] Buffie CG, Bucci V, Stein RR, McKenney PT, Ling L, Gobourne A, et al. Precision microbiome reconstitution restores bile acid mediated resistance to Clostridium difficile. Nature. 2015;517(7533):205-8.

[19] Rea MC, Sit CS, Clayton E, O'Connor PM, Whittal RM, Zheng J, et al. Thuricin CD, a posttranslationally modified bacteriocin with a narrow spectrum of activity against Clostridium difficile. PNAS. 2010;107(20):9352-7.

[20] Ng KM, Ferreyra JA, Higginbottom SK, Lynch JB, Kashyap PC, Gopinath S, et al. Microbiota-liberated host sugars facilitate post-antibiotic expansion of enteric pathogens. Nature. 2013;502(7469):96-9.

183 [21] Wilson KH, Perini F. Role of competition for nutrients in suppression of Clostridium difficile by the colonic microflora. Infect Immun. 1988;56(10):2610-4.

[22] Ferreyra JA, Wu KJ, Hryckowian AJ, Bouley DM, Weimer BC, Sonnenburg JL. Gut microbiota-produced succinate promotes C. difficile infection after antibiotic treatment or motility disturbance. Cell Host Microbe. 2014;16(6):770-7.

[23] Löfmark S, Jernberg C, Jansson JK, Edlund C. Clindamycin-induced enrichment and long-term persistence of resistant Bacteroides spp. and resistance genes. J Antimicrob Chemoth. 2006;58(6):1160-7.

[24] He M, Miyajima F, Roberts P, Ellison L, Pickard DJ, Martin MJ, et al. Emergence and global spread of epidemic healthcare-associated Clostridium difficile. Nat Genet. 2013;45(1):109-13.

[25] Freeman J, Bauer MP, Baines SD, Corver J, Fawley WN, Goorhuis B, et al. The changing epidemiology of Clostridium difficile infections. Clin Microbiol Rev. 2010;23(3):529-49.

[26] Rao K, Micic D, Natarajan M, Winters S, Kiel MJ, Walk ST, et al. Clostridium difficile ribotype 027: relationship to age, detectability of toxins A or B in stool with rapid testing, severe infection, and mortality. Clin Infect Dis. 2015;61(2):233-41.

[27] Giancola SE, Williams RJ, 2nd, Gentry CA. Prevalence of the Clostridium difficile BI/NAP1/027 strain across the United States Veterans Health Administration. Clin Microbiol Infect. 2018;24(8):877-81.

[28] Warny M, Pepin J, Fang A, Killgore G, Thompson A, Brazier J, et al. Toxin production by an emerging strain of Clostridium difficile associated with outbreaks of severe disease in North America and Europe. Lancet. 2005;366(9491):1079-84.

[29] Snydman D, McDermott L, Jenkins S, Goldstein E, Patel R, Forbes B, et al. Epidemiologic trends in Clostridium difficile Isolate Ribotypes in United States from 2010 to 2014. Open Forum Infect Dis. 2017;4:S391-S.

[30] CDC. Severe Clostridium difficile-associated disease in populations previously at low risk--four states, 2005. 2005;54(47):1201.

[31] Couturier J, Davies K, Gateau C, Barbut F. Ribotypes and new virulent strains across Europe. In: Mastrantonio P, Rupnik M, editors. Updates on Clostridium

184 difficile in Europe: Advances in Microbiology, Infectious Diseases and Public Health Volume 8. Cham: Springer International Publishing; 2018. p. 45-58.

[32] Chitnis AS, Holzbauer SM, Belflower RM, Winston LG, Bamberg WM, Lyons C, et al. Epidemiology of community-associated Clostridium difficile infection, 2009 through 2011. JAMA Intern Med. 2013;173(14):1359-67.

[33] Crobach MJT, Voor In 't Holt AF, Knetsch CW, van Dorp SM, Bras W, Harmanus C, et al. An outbreak of Clostridium difficile infections due to new PCR ribotype 826: epidemiologic and microbiologic analyses. Clin Microbiol Infect. 2018;24(3):309.e1-.e4.

[34] Czepiel J, Drozdz M, Pituch H, Kuijper EJ, Perucki W, Mielimonka A, et al. Clostridium difficile infection: review. Eur J Clin Microbiol Infect Dis. 2019;38(7):1211-21.

[35] Fekety R, McFarland LV, Surawicz CM, Greenberg RN, Elmer GW, Mulligan ME. Recurrent Clostridium difficile diarrhea: characteristics of and risk factors for patients enrolled in a prospective, randomized, double-blinded trial. Clin Infect Dis. 1997;24(3):324-33.

[36] Figueroa I, Johnson S, Sambol SP, Goldstein EJ, Citron DM, Gerding DN. Relapse versus reinfection: recurrent Clostridium difficile infection following treatment with fidaxomicin or vancomycin. Clin Infect Dis. 2012;55 Suppl 2:S104- 9.

[37] Kamboj M, Khosa P, Kaltsas A, Babady NE, Son C, Sepkowitz KA. Relapse versus reinfection: surveillance of Clostridium difficile infection. J Clin Infect Dis. 2011;53(10):1003-6.

[38] McFarland LV, Surawicz CM, Greenberg RN, Fekety R, Elmer GW, Moyer KA, et al. A randomized placebo-controlled trial of Saccharomyces boulardii in combination with standard antibiotics for Clostridium difficile disease. JAMA. 1994;271(24):1913-8.

[39] Surawicz CM, McFarland LV, Greenberg RN, Rubin M, Fekety R, Mulligan ME, et al. The search for a better treatment for recurrent Clostridium difficile Disease: use of high-dose vancomycin combined with Saccharomyces boulardii. Clin Infect Dis. 2000;31(4):1012-7.

185 [40] Tung JM, Dolovich LR, Lee CH. Prevention of Clostridium difficile infection with Saccharomyces boulardii: a systematic review. Can J Gastroenterol. 2009;23(12):817-21.

[41] Pattani R, Palda VA, Hwang SW, Shah PS. Probiotics for the prevention of antibiotic-associated diarrhea and Clostridium difficile infection among hospitalized patients: systematic review and meta-analysis. Open Med. 2013;7(2):e56-67.

[42] Saltzman T, Fazzari M, Chung S, Cunha BA, Blum S. The effect of probiotics on the incidence of Clostridioides difficile: Retrospective cohort analysis. Am J Infect Control. 2020;48(2):184-8.

[43] Gerding DN, Meyer T, Lee C, Cohen SH, Murthy UK, Poirier A, et al. Administration of spores of nontoxigenic Clostridium difficile strain M3 for prevention of recurrent C. difficile infection: a randomized clinical trial. JAMA. 2015;313(17):1719-27.

[44] Williams D, Adcock L. Probiotics for antibiotic-associated diarrhea and Clostridium difficile infection: a review of clinical effectiveness CADTH Rapid Response Reports. Ottawa (ON): Canadian Agency for Drugs and Technologies in Health 2018.

[45] Borody TJ, Khoruts A. Fecal microbiota transplantation and emerging applications. Nat Rev Gastroenterol Hepatol. 2011;9(2):88-96.

[46] Grehan MJ, Borody TJ, Leis SM, Campbell J, Mitchell H, Wettstein A. Durable alteration of the colonic microbiota by the administration of donor fecal flora. J Clin Gastroenterol. 2010;44(8):551-61.

[47] Reigadas E, Olmedo M, Valerio M, Vazquez-Cuesta S, Alcala L, Marin M, et al. Fecal microbiota transplantation for recurrent Clostridium difficile infection: Experience, protocol, and results. Rev Esp Quimioter. 2018;31(5):411-8.

[48] van Nood E, Vrieze A, Nieuwdorp M, Fuentes S, Zoetendal EG, de Vos WM, et al. Duodenal infusion of donor feces for recurrent Clostridium difficile. N Engl J Med. 2013;368(5):407-15.

[49] Tariq R, Pardi DS, Bartlett MG, Khanna S. Low cure rates in controlled trials of fecal microbiota transplantation for recurrent Clostridium difficile infection: a systematic review and meta-analysis. Clin Infect Dis. 2019;68(8):1351-8.

186 [50] Tan X, Johnson S. Fecal microbiota transplantation (FMT) for C. difficile infection, just say 'No'. Anaerobe. 2019;60:102092.

[51] Li X, Chu Q, Huang Y, Xiao Y, Song L, Zhu S, et al. Consortium of probiotics attenuates colonization of Clostridioides difficile. Front Microbiol. 2019;10(2871).

[52] Mills JP, Rao K, Young VB. Probiotics for prevention of Clostridium difficile infection. Curr Opin Gastroenterol. 2018;34(1):3-10.

[53] Kelly CP, Kyne L. The host immune response to Clostridium difficile. J Med Microbiol. 2011;60(Pt 8):1070-9.

[54] Aktories K, Schwan C, Jank T. Clostridium difficile Toxin Biology. Annu Rev Microbiol. 2017;71(1):281-307.

[55] Shim JK, Johnson S, Samore MH, Bliss DZ, Gerding DN. Primary symptomless colonisation by Clostridium difficile and decreased risk of subsequent diarrhoea. Lancet. 1998;351(9103):633-6.

[56] Govind R, Dupuy B. Secretion of Clostridium difficile toxins A and B requires the holin-like protein TcdE. PLoS pathogens. 2012;8(6):e1002727.

[57] Govind R, Fitzwater L, Nichols R. Observations on the Role of TcdE Isoforms in Clostridium difficile Toxin Secretion. J Bacteriol. 2015;197(15):2600-9.

[58] Mani N, Dupuy B. Regulation of toxin synthesis in Clostridium difficile by an alternative RNA polymerase sigma factor. PNAS. 2001;98(10):5844-9.

[59] Girinathan BP, Monot M, Boyle D, McAllister KN, Sorg JA, Dupuy B, et al. Effect of tcdR mutation on sporulation in the epidemic Clostridium difficile strain R20291. mSphere. 2017;2(1).

[60] Dupuy B, Govind R, Antunes A, Matamouros S. Clostridium difficile toxin synthesis is negatively regulated by TcdC. J Med Microbiol. 2008;57(Pt 6):685-9.

[61] Matamouros S, England P, Dupuy B. Clostridium difficile toxin expression is inhibited by the novel regulator TcdC. Mol Microbiol. 2007;64(5):1274-88.

[62] Carter GP, Douce GR, Govind R, Howarth PM, Mackin KE, Spencer J, et al. The anti-sigma factor TcdC modulates hypervirulence in an epidemic

187 BI/NAP1/027 clinical isolate of Clostridium difficile. PLoS Pathog. 2011;7(10):e1002317.

[63] Cartman ST, Kelly ML, Heeg D, Heap JT, Minton NP. Precise manipulation of the Clostridium difficile chromosome reveals a lack of association between the tcdC genotype and toxin production. Appl Environ Microb. 2012;78(13):4683-90.

[64] Hecht G, Pothoulakis C, LaMont JT, Madara JL. Clostridium difficile toxin A perturbs cytoskeletal structure and tight junction permeability of cultured human intestinal epithelial monolayers. J Clin Invest. 1988;82(5):1516-24.

[65] Hecht G, Koutsouris A, Pothoulakis C, LaMont JT, Madara JL. Clostridium difficile toxin B disrupts the barrier function of T84 monolayers. Gastroenterology. 1992;102(2):416-23.

[66] Nusrat A, von Eichel-Streiber C, Turner JR, Verkade P, Madara JL, Parkos CA. Clostridium difficile toxins disrupt epithelial barrier function by altering membrane microdomain localization of tight junction proteins. Infect Immun. 2001;69(3):1329-36.

[67] Kasendra M, Barrile R, Leuzzi R, Soriani M. Clostridium difficile toxins facilitate bacterial colonization by modulating the fence and gate function of colonic epithelium. J Infect Dis. 2013;209(7):1095-104.

[68] Gerhard R, Nottrott S, Schoentaube J, Tatge H, Olling A, Just I. Glucosylation of Rho GTPases by Clostridium difficile toxin A triggers apoptosis in intestinal epithelial cells. J Med Microbiol. 2008;57(6):765-70.

[69] Just I, Wilm M, Selzer J, Rex G, von Eichel-Streiber C, Mann M, et al. The enterotoxin from Clostridium difficile (ToxA) monoglucosylates the Rho proteins. J Biol Chem. 1995;270(23):13932-6.

[70] Zhang Y, Yang Z, Gao S, Hamza T, Yfantis HG, Lipsky M, et al. The role of purified Clostridium difficile glucosylating toxins in disease pathogenesis utilizing a murine cecum injection model. Anaerobe. 2017;48:249-56.

[71] Dingle T, Wee S, Mulvey GL, Greco A, Kitova EN, Sun J, et al. Functional properties of the carboxy-terminal host cell-binding domains of the two toxins, TcdA and TcdB, expressed by Clostridium difficile. Glycobiology. 2008;18(9):698- 706.

188 [72] El-Hawiet A, Kitova EN, Kitov PI, Eugenio L, Ng KK, Mulvey GL, et al. Binding of Clostridium difficile toxins to human milk oligosaccharides. Glycobiology. 2011;21(9):1217-27.

[73] Na X, Kim H, Moyer MP, Pothoulakis C, LaMont JT. gp96 is a human colonocyte plasma membrane binding protein for Clostridium difficile toxin A. Infect Immun. 2008;76(7):2862-71.

[74] Yuan P, Zhang H, Cai C, Zhu S, Zhou Y, Yang X, et al. Chondroitin sulfate proteoglycan 4 functions as the cellular receptor for Clostridium difficile toxin B. Cell Res. 2015;25(2):157-68.

[75] Tao L, Zhang J, Meraner P, Tovaglieri A, Wu X, Gerhard R, et al. Frizzled proteins are colonic epithelial receptors for C. difficile toxin B. Nature. 2016;538(7625):350-5.

[76] Papatheodorou P, Zamboglou C, Genisyuerek S, Guttenberg G, Aktories K. Clostridial glucosylating toxins enter cells via clathrin-mediated endocytosis. PLOS ONE. 2010;5(5):e10673.

[77] Kreimeyer I, Euler F, Marckscheffel A, Tatge H, Pich A, Olling A, et al. Autoproteolytic cleavage mediates cytotoxicity of Clostridium difficile toxin A. N-S Arch Pharmaco. 2011;383(3):253-62.

[78] Carter GP, Chakravorty A, Pham Nguyen TA, Mileto S, Schreiber F, Li L, 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. 2015;6(3):e00551-15.

[79] Kuehne SA, Collery MM, Kelly ML, Cartman ST, Cockayne A, Minton NP. Importance of toxin A, toxin B, and CDT in virulence of an epidemic Clostridium difficile strain. J Infect Dis. 2014;209(1):83-6.

[80] Kuehne SA, Cartman ST, Minton NP. Both toxin A and toxin B are important in Clostridium difficile infection. Gut Microbes. 2011;2(4):252-5.

[81] Lyras D, O'Connor JR, Howarth PM, Sambol SP, Carter GP, Phumoonna T, et al. Toxin B is essential for virulence of Clostridium difficile. Nature. 2009;458(7242):1176-9.

189 [82] Monot M, Eckert C, Lemire A, Hamiot A, Dubois T, Tessier C, et al. Clostridium difficile: new insights into the evolution of the pathogenicity locus. Sci Rep. 2015;5:15023.

[83] Shin B-M, Kuak EY, Yoo SJ, Shin WC, Yoo HM. Emerging toxin A−B+ variant strain of Clostridium difficile responsible for pseudomembranous colitis at a tertiary care hospital in Korea. Diagn Micr Infec Dis. 2008;60(4):333-7.

[84] Wang B, Peng W, Zhang P, Su J. The characteristics of Clostridium difficile ST81, a new PCR ribotype of toxin A- B+ strain with high-level fluoroquinolones resistance and higher sporulation ability than ST37/PCR ribotype 017. FEMS Microbiol Lett. 2018;365(17).

[85] van den Berg RJ, Claas EC, Oyib DH, Klaassen CH, Dijkshoorn L, Brazier JS, et al. Characterization of toxin A-negative, toxin B-positive Clostridium difficile isolates from outbreaks in different countries by amplified fragment length polymorphism and PCR ribotyping. J Clin Microbiol. 2004;42(3):1035-41.

[86] Gonçalves C, Decré D, Barbut F, Burghoffer B, Petit J-C. Prevalence and characterization of a binary toxin (actin-specific ADP-ribosyltransferase) from Clostridium difficile. J Clin Microbiol. 2004;42(5):1933-9.

[87] Perelle S, Gibert M, Bourlioux P, Corthier G, Popoff MR. Production of a complete binary toxin (actin-specific ADP-ribosyltransferase) by Clostridium difficile CD196. Infect Immun. 1997;65(4):1402-7.

[88] Stare BG, Delmee M, Rupnik M. Variant forms of the binary toxin CDT locus and tcdC gene in Clostridium difficile strains. J Med Microbiol. 2007;56(Pt 3):329- 35.

[89] Schwan C, Stecher B, Tzivelekidis T, van Ham M, Rohde M, Hardt WD, et al. Clostridium difficile toxin CDT induces formation of microtubule-based protrusions and increases adherence of bacteria. PLoS Pathog. 2009;5(10):e1000626.

[90] Stubbs S, Rupnik M, Gibert M, Brazier J, Duerden B, Popoff M. Production of actin-specific ADP-ribosyltransferase (binary toxin) by strains of Clostridium difficile. FEMS Microbiol Lett. 2000;186(2):307-12.

190 [91] Gerding DN, Johnson S, Rupnik M, Aktories K. Clostridium difficile binary toxin CDT: mechanism, epidemiology, and potential clinical importance. Gut Microbes. 2014;5(1):15-27.

[92] Chaban B, Hughes HV, Beeby M. The flagellum in bacterial pathogens: For motility and a whole lot more. Semin Cell Dev Biol. 2015;46:91-103.

[93] Echazarreta MA, Klose KE. Vibrio Flagellar Synthesis. Front Cell Infect Microbiol. 2019;9:131.

[94] Bolton DJ. Campylobacter virulence and survival factors. Food Microbiol. 2015;48:99-108.

[95] Baban ST, Kuehne SA, Barketi-Klai A, Cartman ST, Kelly ML, Hardie KR, et al. The role of flagella in Clostridium difficile pathogenesis: comparison between a non-epidemic and an epidemic strain. PLOS ONE. 2013;8(9):e73026.

[96] Dingle TC, Mulvey GL, Armstrong GD. Mutagenic analysis of the Clostridium difficile flagellar proteins, FliC and FliD, and their contribution to virulence in hamsters. Infect Immun. 2011;79(10):4061-7.

[97] Stevenson E, Minton NP, Kuehne SA. The role of flagella in Clostridium difficile pathogenicity. Trends Microbiol. 2015;23(5):275-82.

[98] Valiente E, Bouche L, Hitchen P, Faulds-Pain A, Songane M, Dawson LF, et al. Role of glycosyltransferases modifying type B flagellin of emerging hypervirulent Clostridium difficile lineages and their impact on motility and biofilm formation. J Biol Chem. 2016;291(49):25450-61.

[99] Stabler RA, He M, Dawson L, Martin M, Valiente E, Corton C, et al. Comparative genome and phenotypic analysis of Clostridium difficile 027 strains provides insight into the evolution of a hypervirulent bacterium. Genome Biol. 2009;10(9):R102.

[100] Faulds-Pain A, Twine SM, Vinogradov E, Strong PC, Dell A, Buckley AM, et al. The post-translational modification of the Clostridium difficile flagellin affects motility, cell surface properties and virulence. Mol Microbiol. 2014;94(2):272-89.

[101] El Meouche I, Peltier J, Monot M, Soutourina O, Pestel-Caron M, Dupuy B, et al. Characterization of the SigD regulon of C. difficile and its positive control of toxin production through the regulation of tcdR. PLOS ONE. 2013;8(12):e83748.

191 [102] Aubry A, Hussack G, Chen W, Kuolee R, Twine SM, Fulton KM, et al. Modulation of toxin production by the flagellar regulon in Clostridium difficile. Infect Immun. 2012;80(10):3521-32.

[103] Anjuwon-Foster BR, Tamayo R. A genetic switch controls the production of flagella and toxins in Clostridium difficile. PLoS Genet. 2017;13(3):e1006701.

[104] McKee RW, Mangalea MR, Purcell EB, Borchardt EK, Tamayo R. The second messenger cyclic di-GMP regulates Clostridium difficile toxin production by controlling expression of sigD. J Bacteriol. 2013;195(22):5174-85.

[105] Stabler RA, Gerding DN, Songer JG, Drudy D, Brazier JS, Trinh HT, et al. Comparative phylogenomics of Clostridium difficile reveals clade specificity and microevolution of hypervirulent strains. J Bacteriol. 2006;188(20):7297-305.

[106] Melville S, Craig L. Type IV pili in Gram-positive bacteria. Microbiol Mol Biol Rev. 2013;77(3):323-41.

[107] Craig L, Li J. Type IV pili: paradoxes in form and function. Curr Opin Struct Biol. 2008;18(2):267-77.

[108] Conrad JC. Physics of bacterial near-surface motility using flagella and type IV pili: implications for biofilm formation. Res Microbiol. 2012;163(9):619-29.

[109] Varga JJ, Therit B, Melville SB. Type IV pili and the CcpA protein are needed for maximal biofilm formation by the gram-positive anaerobic pathogen Clostridium perfringens. Infect Immun. 2008;76(11):4944-51.

[110] Varga JJ, Nguyen V, O'Brien DK, Rodgers K, Walker RA, Melville SB. Type IV pili-dependent gliding motility in the Gram-positive pathogen Clostridium perfringens and other Clostridia. Mol Microbiol. 2006;62(3):680-94.

[111] Rodgers K, Arvidson CG, Melville S. Expression of a Clostridium perfringens type IV pilin by Neisseria gonorrhoeae mediates adherence to muscle cells. Infect Immun. 2011;79(8):3096-105.

[112] Liu S, Wang J, Zhang J, Wang T, Zhou Y, Lv Q, et al. Tectorigenin reduces type IV pilus-dependent cell adherence in Clostridium perfringens. FEMS Microbiol Lett. 2019;366(10).

192 [113] Purcell EB, McKee RW, Bordeleau E, Burrus V, Tamayo R. Regulation of type IV pili contributes to surface behaviors of historical and epidemic strains of Clostridium difficile. J Bacteriol. 2015;198(3):565-77.

[114] Bordeleau E, Purcell EB, Lafontaine DA, Fortier LC, Tamayo R, Burrus V. Cyclic di-GMP riboswitch-regulated type IV pili contribute to aggregation of Clostridium difficile. J Bacteriol. 2015;197(5):819-32.

[115] McKee RW, Aleksanyan N, Garrett EM, Tamayo R. Type IV pili promote Clostridium difficile adherence and persistence in a mouse model of infection. Infect Immun. 2018;86(5):pii: e00943-17.

[116] Goulding D, Thompson H, Emerson J, Fairweather NF, Dougan G, Douce GR. Distinctive profiles of infection and pathology in hamsters infected with Clostridium difficile strains 630 and B1. Infect Immun. 2009;77(12):5478-85.

[117] Flemming H-C, Wingender J, Szewzyk U, Steinberg P, Rice SA, Kjelleberg S. Biofilms: an emergent form of bacterial life. Nature reviews Microbiology. 2016;14(9):563-75.

[118] Venkatesan N, Perumal G, Doble M. Bacterial resistance in biofilm-associated bacteria. Future Microbiol. 2015;10(11):1743-50.

[119] Donelli G, Vuotto C, Cardines R, Mastrantonio P. Biofilm-growing intestinal anaerobic bacteria. FEMS Immunol Med Microbiol. 2012;65(2):318-25.

[120] Purcell EB, McKee RW, Courson DS, Garrett EM, McBride SM, Cheney RE, et al. A nutrient-regulated cyclic diguanylate phosphodiesterase controls Clostridium difficile biofilm and toxin production during stationary phase. Infect Immun. 2017;85(9):pii: e00347-17.

[121] Maldarelli GA, Piepenbrink KH, Scott AJ, Freiberg JA, Song Y, Achermann Y, et al. Type IV pili promote early biofilm formation by Clostridium difficile. Pathog Dis. 2016;74(6):pii: ftw061.

[122] Poquet I, Saujet L, Canette A, Monot M, Maihajlovic J, Ghigo JM, et al. Clostridium difficile biofilm: remodeling metabolism and cell surface to build a sparse and heterogeneously aggregated architecture. Front Microbiol. 2018;9.

193 [123] Dapa T, Leuzzi R, Ng YK, Baban ST, Adamo R, Kuehne SA, et al. Multiple factors modulate biofilm formation by the anaerobic pathogen Clostridium difficile. J Bacteriol. 2013;195(3):545-55.

[124] Dubois T, Tremblay YDN, Hamiot A, Martin-Verstraete I, Deschamps J, Monot M, et al. A microbiota-generated bile salt induces biofilm formation in Clostridium difficile. NPJ Biofilms Microbiomes. 2019;5:14.

[125] Semenyuk EG, Laning ML, Foley J, Johnston PF, Knight KL, Gerding DN, et al. Spore formation and toxin production in Clostridium difficile biofilms. PLOS ONE. 2014;9(1):e87757.

[126] Chu M, Mallozzi MJG, Roxas BP, Bertolo L, Monteiro MA, Agellon A, et al. A Clostridium difficile cell wall glycopolymer locus influences bacterial shape, polysaccharide production and virulence. PLoS Pathog. 2016;12(10):e1005946.

[127] Pantaleon V, Soavelomandroso AP, Bouttier S, Briandet R, Roxas B, Chu M, et al. The Clostridium difficile protease Cwp84 modulates both biofilm formation and cell-surface properties. PLOS ONE. 2015;10(4):e0124971.

[128] Lawley TD, Clare S, Walker AW, Goulding D, Stabler RA, Croucher N, et al. Antibiotic treatment of Clostridium difficile carrier mice triggers a supershedder state, spore-mediated transmission, and severe disease in immunocompromised hosts. Infect Immun. 2009;77(9):3661-9.

[129] Soavelomandroso AP, Gaudin F, Hoys S, Nicolas V, Vedantam G, Janoir C, et al. Biofilm Structures in a Mono-Associated Mouse Model of Clostridium difficile Infection. Front Microbiol. 2017;8:2086.

[130] Semenyuk EG, Poroyko VA, Johnston PF, Jones SE, Knight KL, Gerding DN, et al. Analysis of bacterial communities during Clostridium difficile infection in the mouse. Infect Immun. 2015;83(11):4383-91.

[131] James GA, Chesnel L, Boegli L, deLancey Pulcini E, Fisher S, Stewart PS. Analysis of Clostridium difficile biofilms: imaging and antimicrobial treatment. J Antimicrob Chemother. 2018;73(1):102-8.

[132] Weigel WA, Dersch P. Phenotypic heterogeneity: a bacterial virulence strategy. Microbes Infect. 2018;20(9-10):570-7.

194 [133] Schroter L, Dersch P. Phenotypic diversification of microbial pathogens- cooperating and preparing for the future. J Mol Biol. 2019;431(23):4645-55.

[134] Fisher RA, Gollan B, Helaine S. Persistent bacterial infections and persister cells. Nature reviews Microbiology. 2017;15(8):453-64.

[135] van der Woude MW. Phase variation: how to create and coordinate population diversity. Clin Microbiol Rev. 2011;14(2):205-11.

[136] van der Woude MW, Baumler AJ. Phase and antigenic variation in bacteria. Clin Microbiol Rev. 2004;17(3):581-611, table of contents.

[137] Bayliss CD. Determinants of phase variation rate and the fitness implications of differing rates for bacterial pathogens and commensals. FEMS Microbiol Rev. 2009;33(3):504-20.

[138] Sekulovic O, Mathias Garrett E, Bourgeois J, Tamayo R, Shen A, Camilli A. Genome-wide detection of conservative site-specific recombination in bacteria. PLoS Genet. 2018;14(4):e1007332.

[139] Jiang X, Hall AB, Arthur TD, Plichta DR, Covington CT, Poyet M, et al. Invertible promoters mediate bacterial phase variation, antibiotic resistance, and host adaptation in the gut. Science. 2019;363(6423):181-7.

[140] van Belkum A, Scherer S, van Alphen L, Verbrugh H. Short-Sequence DNA repeats in prokaryotic genomes. Microbiol Mol Biol Rev. 1998;62(2):275-93.

[141] van Belkum A, van Leeuwen W, Scherer S, Verbrugh H. Occurrence and structure-function relationship of pentameric short sequence repeats in microbial genomes. Res Microbiol. 1999;150(9-10):617-26.

[142] Dawid S, Barenkamp SJ, St. Geme JW. Variation in expression of the Haemophilus influenzae HMW adhesins: A prokaryotic system reminiscent of eukaryotes. PNAS. 1999;96(3):1077-82.

[143] De Bolle X, Bayliss CD, Field D, van de Ven T, Saunders NJ, Hood DW, et al. The length of a tetranucleotide repeat tract in Haemophilus influenzae determines the phase variation rate of a gene with homology to type III DNA methyltransferases. Mol Microbiol. 2000;35(1):211-22.

195 [144] Martin P, Sun L, Hood DW, Moxon ER. Involvement of genes of genome maintenance in the regulation of phase variation frequencies in Neisseria meningitidis. Microbiology. 2004;150(9):3001-12.

[145] van Ham SM, van Alphen L, Mooi FR, van Putten JP. Phase variation of H. influenzae fimbriae: transcriptional control of two divergent genes through a variable combined promoter region. Cell. 1993;73(6):1187-96.

[146] Kubiet M, Ramphal R, Weber A, Smith A. Pilus-mediated adherence of Haemophilus influenzae to human respiratory mucins. Infect Immun. 2000;68(6):3362-7.

[147] Anderson GG, Palermo JJ, Schilling JD, Roth R, Heuser J, Hultgren SJ. Intracellular bacterial biofilm-like pods in urinary tract infections. Science. 2003;301(5629):105-7.

[148] Danese PN, Pratt LA, Dove SL, Kolter R. The outer membrane protein, antigen 43, mediates cell-to-cell interactions within Escherichia coli biofilms. Mol Microbiol. 2000;37(2):424-32.

[149] Diderichsen B. flu, a metastable gene controlling surface properties of Escherichia coli. J Bacteriol. 1980;141(2):858-67.

[150] Ulett GC, Valle J, Beloin C, Sherlock O, Ghigo JM, Schembri MA. Functional analysis of antigen 43 in uropathogenic Escherichia coli reveals a role in long- term persistence in the urinary tract. Infect Immun. 2007;75(7):3233-44.

[151] van der Woude MW, Henderson IR. Regulation and function of Ag43 (flu). Annu Rev Microbiol. 2008;62:153-69.

[152] Wallecha A, Munster V, Correnti J, Chan T, van der Woude M. Dam- and OxyR-dependent phase variation of agn43: essential elements and evidence for a new role of DNA methylation. J Bacteriol. 2002;184(12):3338-47.

[153] Henderson IR, Meehan M, Owen P. A novel regulatory mechanism for a novel phase-variable outer membrane protein of Escherichia coli. Adv Exp Med Biol. 1997;412:349-55.

[154] Li J, Zhang JR. Phase variation of Streptococcus pneumoniae. Microbiol Spectr. 2019;7(1).

196 [155] Bai Y, Yang J, Zarrella TM, Zhang Y, Metzger DW, Bai G. Cyclic di-AMP impairs potassium uptake mediated by a cyclic di-AMP binding protein in Streptococcus pneumoniae. J Bacteriol. 2014;196(3):614-23.

[156] Li J, Li J-W, Feng Z, Wang J, An H, Liu Y, et al. Epigenetic Switch Driven by DNA Inversions Dictates Phase Variation in Streptococcus pneumoniae. PLoS pathogens. 2016;12(7):e1005762-e.

[157] Li JW, Li J, Wang J, Li C, Zhang JR. Molecular mechanisms of hsdS inversions in the cod locus of Streptococcus pneumoniae. J Bacteriol. 2019;201(6).

[158] Grindley ND, Whiteson KL, Rice PA. Mechanisms of site-specific recombination. Annu Rev Biochem. 2006;75:567-605.

[159] Schwan WR. Regulation of fim genes in uropathogenic Escherichia coli. World J Clin Infect Dis. 2011;1(1):17-25.

[160] Connell I, Agace W, Klemm P, Schembri M, Marild S, Svanborg C. Type 1 fimbrial expression enhances Escherichia coli virulence for the urinary tract. PNAS. 1996;93(18):9827-32.

[161] Gunther NWt, Snyder JA, Lockatell V, Blomfield I, Johnson DE, Mobley HLT. Assessment of virulence of uropathogenic Escherichia coli type 1 fimbrial mutants in which the invertible element is phase-locked on or off. Infect Immun. 2002;70(7):3344-54.

[162] Struve C, Krogfelt KA. In vivo detection of Escherichia coli type 1 fimbrial expression and phase variation during experimental urinary tract infection. Microbiology. 1999;145 ( Pt 10):2683-90.

[163] Hultgren SJ, Porter TN, Schaeffer AJ, Duncan JL. Role of type 1 pili and effects of phase variation on lower urinary tract infections produced by Escherichia coli. Infect Immun. 1985;50(2):370-7.

[164] Lim JK, Gunther NWt, Zhao H, Johnson DE, Keay SK, Mobley HL. In vivo phase variation of Escherichia coli type 1 fimbrial genes in women with urinary tract infection. Infect Immun. 1998;66(7):3303-10.

[165] Schaeffer AJ, Schwan WR, Hultgren SJ, Duncan JL. Relationship of type 1 pilus expression in Escherichia coli to ascending urinary tract infections in mice. Infect Immun. 1987;55(2):373-80.

197 [166] Silverblatt FJ, Dreyer JS, Schauer S. Effect of pili on susceptibility of Escherichia coli to phagocytosis. Infect Immun. 1979;24(1):218-23.

[167] Gally DL, Leathart J, Blomfield IC. Interaction of FimB and FimE with the fim switch that controls the phase variation of type 1 fimbriae in Escherichia coli K- 12. Mol Microbiol. 1996;21(4):725-38.

[168] Kulasekara HD, Blomfield IC. The molecular basis for the specificity of fimE in the phase variation of type 1 fimbriae of Escherichia coli K-12. Mol Microbiol. 1999;31(4):1171-81.

[169] McCusker MP, Turner EC, Dorman CJ. DNA sequence heterogeneity in Fim tyrosine-integrase recombinase-binding elements and functional motif asymmetries determine the directionality of the fim genetic switch in Escherichia coli K-12. Mol Microbiol. 2008;67(1):171-87.

[170] Olsen PB, Schembri MA, Gally DL, Klemm P. Differential temperature modulation by H-NS of the fimB and fimE recombinase genes which control the orientation of the type 1 fimbrial phase switch. FEMS Microbiol Lett. 1998;162(1):17-23.

[171] O′gara JP, Dorman CJ. Effects of local transcription and H-NS on inversion of the fim switch of Escherichia coli. Mol Microbiol. 2000;36(2):457-66.

[172] Joyce SA, Dorman CJ. A Rho-dependent phase-variable transcription terminator controls expression of the FimE recombinase in Escherichia coli. Mol Microbiol. 2002;45(4):1107-17.

[173] Hinde P, Deighan P, Dorman CJ. Characterization of the detachable Rho- dependent transcription terminator of the fimE gene in Escherichia coli K-12. J Bacteriol. 2005;187(24):8256-66.

[174] Blomfield IC, Calie PJ, Eberhardt KJ, McClain MS, Eisenstein BI. Lrp stimulates phase variation of type 1 fimbriation in Escherichia coli K-12. J Bacteriol. 1993;175(1):27-36.

[175] Blomfield IC, Kulasekara DH, Eisenstein BI. Integration host factor stimulates both FimB- and FimE-mediated site-specific DNA inversion that controls phase variation of type 1 fimbriae expression in Escherichia coli. Mol Microbiol. 1997;23(4):705-17.

198 [176] Kelly A, Conway C, T OC, Smith SG, Dorman CJ. DNA supercoiling and the Lrp protein determine the directionality of fim switch DNA inversion in Escherichia coli K-12. J Bacteriol. 2006;188(15):5356-63.

[177] Dorman CJ, Higgins CF. Fimbrial phase variation in Escherichia coli: dependence on integration host factor and homologies with other site-specific recombinases. J Bacteriol. 1987;169(8):3840-3.

[178] Chen CF, Lan J, Korovine M, Shao ZQ, Tao L, Zhang J, et al. Metabolic regulation of lrp gene expression in Escherichia coli K-12. Microbiology. 1997;143 ( Pt 6):2079-84.

[179] Emerson JE, Reynolds CB, Fagan RP, Shaw HA, Goulding D, Fairweather NF. A novel genetic switch controls phase variable expression of CwpV, a Clostridium difficile cell wall protein. Mol Microbiol. 2009;74(3):541-56.

[180] Reynolds CB, Emerson JE, de la Riva L, Fagan RP, Fairweather NF. The Clostridium difficile cell wall protein CwpV is antigenically variable between strains, but exhibits conserved aggregation-promoting function. PLoS pathogens. 2011;7(4):e1002024.

[181] Sekulovic O, Ospina Bedoya M, Fivian-Hughes AS, Fairweather NF, Fortier LC. The Clostridium difficile cell wall protein CwpV confers phase-variable phage resistance. Mol Microbiol. 2015;98(2):329-42.

[182] Rostol JT, Marraffini L. (Ph)ighting phages: how bacteria resist their parasites. Cell Host Microbe. 2019;25(2):184-94.

[183] Stabler RA, Valiente E, Dawson LF, He M, Parkhill J, Wren BW. In-depth genetic analysis of Clostridium difficile PCR-ribotype 027 strains reveals high genome fluidity including point mutations and inversions. Gut Microbes. 2010;1(4):269-76.

[184] Anjuwon-Foster BR, Maldonado-Vazquez N, Tamayo R. Characterization of flagellar and toxin phase variation in Clostridiodes difficile ribotype 012 isolates. J Bacteriol. 2018.

[185] Sekulovic O, Bourgeois J, Shen A, Camilli A. Expanding the repertoire of conservative site-specific recombination in Clostridioides difficile. Anaerobe. 2019:102073.

199 [186] Coyne MJ, Weinacht KG, Krinos CM, Comstock LE. Mpi recombinase globally modulates the surface architecture of a human commensal bacterium. PNAS. 2003;100(18):10446-51.

[187] Romling U, Galperin MY, Gomelsky M. Cyclic di-GMP: the first 25 years of a universal bacterial second messenger. Microbiol Mol Biol Rev. 2013;77(1):1-52.

[188] Ross P, Weinhouse H, Aloni Y, Michaeli D, Weinberger-Ohana P, Mayer R, et al. Regulation of cellulose synthesis in Acetobacter xylinum by cyclic diguanylic acid. Nature. 1987;325(15):279-81.

[189] Hall-Stoodley L, Costerton JW, Stoodley P. Bacterial biofilms: from the natural environment to infectious diseases. Nature reviews Microbiology. 2004;2(2):95- 108.

[190] Costerton JW, Lewandowski Z, Caldwell DE, Korber DR, Lappin-Scott HM. Microbial biofilms. Annu Rev Microbiol. 1995;49:711-45.

[191] Valentini M, Filloux A. Biofilms and cyclic di-GMP (c-di-GMP) signaling: lessons from Pseudomonas aeruginosa and other bacteria. J Biol Chem. 2016;291(24):12547-55.

[192] Tamayo R, Pratt JT, Camilli A. Roles of cyclic diguanylate in the regulation of bacterial pathogenesis. Annu Rev Microbiol. 2007;61:131-48.

[193] Laventie B-J, Sangermani M, Estermann F, Manfredi P, Planes R, Hug I, et al. A surface-induced asymmetric program promotes tissue colonization by Pseudomonas aeruginosa. Cell Host Microbe. 2019;25(1):140-52.e6.

[194] Kuchma SL, Griffin EF, O'Toole GA. Minor pilins of the type IV pilus system participate in the negative regulation of swarming motility. J Bacteriol. 2012;194(19):5388-403.

[195] Güvener ZT, Harwood CS. Subcellular location characteristics of the Pseudomonas aeruginosa GGDEF protein, WspR, indicate that it produces cyclic-di-GMP in response to growth on surfaces. Mol Microbiol. 2007;66(6):1459-73.

[196] Alexandre G. Chemotaxis control of transient cell aggregation. J Bacteriol. 2015;197(20):3230-7.

200 [197] Schniederberend M, Williams JF, Shine E, Shen C, Jain R, Emonet T, et al. Modulation of flagellar rotation in surface-attached bacteria: A pathway for rapid surface-sensing after flagellar attachment. PLoS pathogens. 2019;15(11):e1008149.

[198] Jain R, Kazmierczak BI. Should I stay or should I go? Pseudomonas just can't decide. Cell Host Microbe. 2019;25(1):5-7.

[199] Hall CL, Lee VT. Cyclic-di-GMP regulation of virulence in bacterial pathogens. Wiley Interdiscip Rev RNA. 2018;9(1).

[200] Klockgether J, Tümmler B. Recent advances in understanding Pseudomonas aeruginosa as a pathogen. F1000Rese. 2017;6:1261-.

[201] Moscoso JA, Mikkelsen H, Heeb S, Williams P, Filloux A. The Pseudomonas aeruginosa sensor RetS switches Type III and Type VI secretion via c-di-GMP signalling. Environ Microbiol. 2011;13(12):3128-38.

[202] Basler M, Ho BT, Mekalanos JJ. Tit-for-tat: type VI secretion system counterattack during bacterial cell-cell interactions. Cell. 2013;152(4):884-94.

[203] Potvin E, Lehoux DE, Kukavica-Ibrulj I, Richard KL, Sanschagrin F, Lau GW, et al. In vivo functional genomics of Pseudomonas aeruginosa for high- throughput screening of new virulence factors and antibacterial targets. Environ Microbiol. 2003;5(12):1294-308.

[204] Mougous JD, Cuff ME, Raunser S, Shen A, Zhou M, Gifford CA, et al. A virulence locus of Pseudomonas aeruginosa encodes a protein secretion apparatus. Science. 2006;312(5779):1526-30.

[205] Tischler AD, Camilli A. Cyclic diguanylate regulates Vibrio cholerae virulence gene expression. Infect Immun. 2005;73(9):5873-82.

[206] Allombert J, Lazzaroni JC, Bailo N, Gilbert C, Charpentier X, Doublet P, et al. Three antagonistic cyclic di-GMP-catabolizing enzymes promote differential Dot/Icm effector delivery and intracellular survival at the early steps of Legionella pneumophila infection. Infect Immun. 2014;82(3):1222-33.

[207] Lamprokostopoulou A, Monteiro C, Rhen M, Romling U. Cyclic di-GMP signalling controls virulence properties of Salmonella enterica serovar Typhimurium at the mucosal lining. Environ Microbiol. 2010;12(1):40-53.

201 [208] Ahmad I, Lamprokostopoulou A, Le Guyon S, Streck E, Barthel M, Peters V, et al. Complex c-di-GMP signaling networks mediate transition between virulence properties and biofilm formation in Salmonella enterica serovar Typhimurium. PLOS ONE. 2011;6(12):e28351.

[209] Karaolis DKR, Means TK, Yang D, Takahashi M, Yoshimura T, Muraille E, et al. Bacterial c-di-GMP Is an immunostimulatory molecule. J Immunol. 2007;178(4):2171-81.

[210] McWhirter SM, Barbalat R, Monroe KM, Fontana MF, Hyodo M, Joncker NT, et al. A host type I interferon response is induced by cytosolic sensing of the bacterial second messenger cyclic-di-GMP. J Exp Med. 2009;206(9):1899-911.

[211] Cui T, Cang H, Yang B, He ZG. Cyclic dimeric guanosine monophosphate: activation and inhibition of innate immune response. J Innate Immun. 2019;11(3):242-8.

[212] Woodward JJ, Iavarone AT, Portnoy DA. c-di-AMP secreted by intracellular Listeria monocytogenes activates a host type I interferon response. Science. 2010;328(5986):1703-5.

[213] Zahringer F, Lacanna E, Jenal U, Schirmer T, Boehm A. Structure and signaling mechanism of a zinc-sensory diguanylate cyclase. Structure. 2013;21(7):1149-57.

[214] Tuckerman JR, Gonzalez G, Sousa EH, Wan X, Saito JA, Alam M, et al. An oxygen-sensing diguanylate cyclase and phosphodiesterase couple for c-di-GMP control. Biochemistry. 2009;48(41):9764-74.

[215] Chang AL, Tuckerman JR, Gonzalez G, Mayer R, Weinhouse H, Volman G, et al. Phosphodiesterase A1, a regulator of cellulose synthesis in Acetobacter xylinum, is a heme-based sensor. Biochemistry. 2001;40(12):3420-6. [216] Schirmer T. C-di-GMP synthesis: structural aspects of evolution, catalysis and regulation. J Mol Biol. 2016;428(19):3683-701.

[217] Hickman JW, Tifrea DF, Harwood CS. A chemosensory system that regulates biofilm formation through modulation of cyclic diguanylate levels. PNAS. 2005;102(40):14422-7.

202 [218] Huangyutitham V, Güvener ZT, Harwood CS. Subcellular clustering of the phosphorylated WspR response regulator protein stimulates its diguanylate cyclase activity. mBio. 2013;4(3):e00242.

[219] Schirmer T, Jenal U. Structural and mechanistic determinants of c-di-GMP signalling. Nature reviews Microbiology. 2009;7(10):724-35.

[220] Krasteva PV, Giglio KM, Sondermann H. Sensing the messenger: the diverse ways that bacteria signal through c-di-GMP. Protein Sci. 2012;21(7):929-48.

[221] Mills E, Pultz IS, Kulasekara HD, Miller SI. The bacterial second messenger c‐ di‐GMP: mechanisms of signalling. Cell Microbiol. 2011;13(8):1122-9.

[222] Cheang QW, Xin L, Chea RYF, Liang ZX. Emerging paradigms for PilZ domain- mediated c-di-GMP signaling. Biochem Soc Trans. 2019;47(1):381-8.

[223] Ryjenkov DA, Simm R, Romling U, Gomelsky M. The PilZ domain is a receptor for the second messenger c-di-GMP: the PilZ domain protein YcgR controls motility in enterobacteria. J Biol Chem. 2006;281(41):30310-4.

[224] Paul K, Nieto V, Carlquist WC, Blair DF, Harshey RM. The c-di-GMP binding protein YcgR controls flagellar motor direction and speed to affect chemotaxis by a "backstop brake" mechanism. Mol cell. 2010;38(1):128-39.

[225] Subramanian S, Gao X, Dann CE, 3rd, Kearns DB. MotI (DgrA) acts as a molecular clutch on the flagellar stator protein MotA in Bacillus subtilis. PNAS. 2017;114(51):13537-42.

[226] Winkler WC, Breaker RR. Regulation of bacterial gene expression by riboswitches. Annu Rev Microbiol. 2005;59:487-517.

[227] Serganov A, Nudler E. A decade of riboswitches. Cell. 2013;152(1-2):17-24.

[228] Lee ER, Baker JL, Weinberg Z, Sudarsan N, Breaker RR. An allosteric self- splicing ribozyme triggered by a bacterial second messenger. Science. 2010;329(5993):845-8.

[229] Dahlstrom KM, O'Toole GA. A symphony of cyclases: specificity in diguanylate cyclase signaling. Annu Rev Microbiol. 2017;71:179-95.

203 [230] Pultz IS, Christen M, Kulasekara HD, Kennard A, Kulasekara B, Miller SI. The response threshold of Salmonella PilZ domain proteins is determined by their binding affinities for c-di-GMP. Mol Microbiol. 2012;86(6):1424-40.

[231] Dahlstrom KM, Giglio KM, Collins AJ, Sondermann H, O'Toole GA. Contribution of physical interactions to signaling specificity between a diguanylate cyclase and its effector. mBio. 2015;6(6):e01978.

[232] Dahlstrom KM, Giglio KM, Sondermann H, O'Toole GA. The inhibitory site of a diguanylate cyclase is a necessary element for interaction and signaling with an effector protein. J Bacteriol. 2016;198(11):1595-603.

[233] Newell PD, Monds RD, O'Toole GA. LapD is a bis-(3',5')-cyclic dimeric GMP- binding protein that regulates surface attachment by Pseudomonas fluorescens Pf0-1. PNAS. 2009;106(9):3461-6.

[234] Newell PD, Boyd CD, Sondermann H, O'Toole GA. A c-di-GMP effector system controls cell adhesion by inside-out signaling and surface protein cleavage. PLoS biology. 2011;9(2):e1000587.

[235] Purcell EB, Tamayo R. Cyclic diguanylate signaling in Gram-positive bacteria. FEMS Microbiol Rev. 2016;40(5):753-73.

[236] McKee RW, Harvest CK, Tamayo R. Cyclic diguanylate regulates virulence factor genes via multiple riboswitches in Clostridium difficile. mSphere. 2018;3(5):e00423-18.

[237] Purcell EB, McKee RW, McBride SM, Waters CM, Tamayo R. Cyclic diguanylate inversely regulates motility and aggregation in Clostridium difficile. J Bacteriol. 2012;194(13):3307-16.

[238] Bordeleau E, Fortier LC, Malouin F, Burrus V. c-di-GMP turn-over in Clostridium difficile is controlled by a plethora of diguanylate cyclases and phosphodiesterases. PLoS Genet. 2011;7(3):e1002039.

[239] Gao X, Dong X, Subramanian S, Matthews PM, Cooper CA, Kearns DB, et al. Engineering of Bacillus subtilis strains to allow rapid characterization of heterologous diguanylate cyclases and phosphodiesterases. Appl Environ Microb. 2014;80(19):6167-74.

204 [240] Wang YC, Chin KH, Tu ZL, He J, Jones CJ, Sanchez DZ, et al. Nucleotide binding by the widespread high-affinity cyclic di-GMP receptor MshEN domain. Nat Commun. 2016;7:12481.

[241] Soutourina OA, Monot M, Boudry P, Saujet L, Pichon C, Sismeiro O, et al. Genome-wide identification of regulatory RNAs in the human pathogen Clostridium difficile. PLoS Genet. 2013;9(5):e1003493.

[242] Hall IC, O'Toole E. Intestinal flora in new-born infants: with a description of a new pathogenic anaerobe, Bacillus difficilis. Am J Dis of Child. 1935;49(2):390- 402.

[243] Ransom EM, Ellermeier CD, Weiss DS. Use of mCherry Red fluorescent protein for studies of protein localization and gene expression in Clostridium difficile. Appl Environ Microb. 2015;81(5):1652-60.

[244] Shen A. Expanding the Clostridioides difficile genetics toolbox. J Bacteriol. 2019;201(14):e00089-19.

[245] Purdy D, O'Keeffe TA, Elmore M, Herbert M, McLeod A, Bokori-Brown M, et al. Conjugative transfer of clostridial shuttle vectors from Escherichia coli to Clostridium difficile through circumvention of the restriction barrier. Mol Microbiol. 2002;46(2):439-52.

[246] Mani N, Lyras D, Barroso L, Howarth P, Wilkins T, Rood JI, et al. Environmental response and autoregulation of Clostridium difficile TxeR, a sigma factor for toxin gene expression. J Bacteriol. 2002;184(21):5971-8.

[247] O'Connor JR, Lyras D, Farrow KA, Adams V, Powell DR, Hinds J, et al. Construction and analysis of chromosomal Clostridium difficile mutants. Mol Microbiol. 2006;61(5):1335-51.

[248] Ryall B, Eydallin G, Ferenci T. Culture history and population heterogeneity as determinants of bacterial adaptation: the adaptomics of a single environmental transition. Microbiol Mol Biol Rev. 2012;76(3):597-625.

[249] Magdanova LA, Golyasnaya NVJM. Heterogeneity as an adaptive trait of microbial populations. Microbiology. 2013;82(1):1-10.

[250] Casadesús J, Low DA. Programmed heterogeneity: epigenetic mechanisms in bacteria. J Biol Chem. 2013;288(20):13929-35.

205 [251] Weiser JN, Austrian R, Sreenivasan PK, Masure HR. Phase variation in pneumococcal opacity: relationship between colonial morphology and nasopharyngeal colonization. Infect Immun. 1994;62(6):2582-9.

[252] Li X, Lockatell CV, Johnson DE, Mobley HL. Identification of MrpI as the sole recombinase that regulates the phase variation of MR/P fimbria, a bladder colonization factor of uropathogenic Proteus mirabilis. J Mol Microbiol. 2002;45(3):865-74.

[253] Briles DE, Novak L, Hotomi M, van Ginkel FW, King J. Nasal colonization with Streptococcus pneumoniae includes subpopulations of surface and invasive pneumococci. Infect Immun. 2005;73(10):6945-51.

[254] Tauseef I, Ali YM, Bayliss CD. Phase variation of PorA, a major outer membrane protein, mediates escape of bactericidal antibodies by Neisseria meningitidis. Infect Immun. 2013;81(4):1374-80.

[255] Alamro M, Bidmos FA, Chan H, Oldfield NJ, Newton E, Bai X, et al. Phase variation mediates reductions in expression of surface proteins during persistent meningococcal carriage. Infect Immun. 2014;82(6):2472-84.

[256] Oliver MB, Roy AB, Kumar R, Lefkowitz EJ, Swords WE. Streptococcus pneumoniae TIGR4 phase-locked opacity variants differ in virulence phenotypes. mSphere. 2017;2(6):e00386-17.

[257] Arai J, Hotomi M, Hollingshead SK, Ueno Y, Briles DE, Yamanaka N. Streptococcus pneumoniae isolates from middle ear fluid and nasopharynx of children with acute otitis media exhibit phase variation. J Clin Microbiol. 2011;49(4):1646-9.

[258] Kuehne SA, Cartman ST, Heap JT, Kelly ML, Cockayne A, Minton NP. The role of toxin A and toxin B in Clostridium difficile infection. Nature. 2010;467(7316):711-3.

[259] Shen A. Clostridium difficile toxins: mediators of inflammation. J Innate Immun. 2012;4(2):149-58.

[260] Morris GN, Winter J, Cato EP, Ritchie AE, Bokkenheuser VD. Clostridium scindens sp. nov., a human intestinal bacterium with desmolytic activity on corticoids. Int J Syst Bacteriol. 1985;35(4):478-81.

206 [261] Lipovsek S, Leitinger G, Rupnik M. Ultrastructure of Clostridium difficile colonies. Anaerobe. 2013;24:66-70.

[262] Anjuwon-Foster BR, Tamayo R. Phase variation of Clostridium difficile virulence factors. Gut Microbes. 2018;9(1):76-83.

[263] Simpson LM, White VK, Zane SF, Oliver JD. Correlation between virulence and colony morphology in Vibrio vulnificus. Infect Immun. 1987;55(1):269-72.

[264] Kansal RG, Gomez-Flores R, Mehta RT. Change in colony morphology influences the virulence as well as the biochemical properties of the Mycobacterium avium complex. J Microb Pathog. 1998;25(4):203-14.

[265] Lenz LL, Portnoy DA. Identification of a second Listeria secA gene associated with protein secretion and the rough phenotype. Mol Microbiol. 2002;45(4):1043- 56.

[266] Tipton KA, Dimitrova D, Rather PN. Phase-variable control of multiple phenotypes in Acinetobacter baumannii strain AB5075. J Bacteriol. 2015;197(15):2593-9.

[267] Clary G, Sasindran SJ, Nesbitt N, Mason L, Cole S, Azad A, et al. Mycobacterium abscessus smooth and rough morphotypes form antimicrobial- tolerant biofilm phenotypes but are killed by acetic acid. Antimicrob Agents Ch. 2018;62(3):e01782-17.

[268] Chin CY, Tipton KA, Farokhyfar M, Burd EM, Weiss DS, Rather PN. A high- frequency phenotypic switch links bacterial virulence and environmental survival in Acinetobacter baumannii. Nat Microbiol. 2018;3(5):563-9.

[269] Dennis EA, Coats MT, Griffin S, Pang B, Briles DE, Crain MJ, et al. Hyperencapsulated mucoid pneumococcal isolates from patients with cystic fibrosis have increased biofilm density and persistence in vivo. Pathog Dis. 2018;76(7):fty073.

[270] Sekulovic O, Fortier LC. Global transcriptional response of Clostridium difficile carrying the CD38 prophage. Appl Environ Microb. 2015;81(4):1364-74.

[271] Sudarsan N, Lee ER, Weinberg Z, Moy RH, Kim JN, Link KH, et al. Riboswitches in eubacteria sense the second messenger cyclic di-GMP. Science. 2008;321(5887):411-3.

207 [272] Tamayo R. Cyclic diguanylate riboswitches control bacterial pathogenesis mechanisms. PLoS pathogens. 2019;15(2):e1007529.

[273] Peltier J, Shaw HA, Couchman EC, Dawson LF, Yu L, Choudhary JS, et al. Cyclic-di-GMP regulates production of sortase substrates of Clostridium difficile and their surface exposure through ZmpI protease-mediated cleavage. J Biol Chem. 2015;290(40):24453-69.

[274] Kirk JA, Fagan RP. Heat shock increases conjugation efficiency in Clostridium difficile. Anaerobe. 2016;42:1-5.

[275] Putnam EE, Nock AM, Lawley TD, Shen A. SpoIVA and SipL are Clostridium difficile spore morphogenetic proteins. J Bacteriol. 2013;195(6):1214-25.

[276] Edwards AN, Tamayo R, McBride SM. A novel regulator controls Clostridium difficile sporulation, motility and toxin production. Mol Microbiol. 2016;100(6):954- 71.

[277] Nawrocki KL, Edwards AN, Daou N, Bouillaut L, McBride SM. CodY- dependent regulation of sporulation in Clostridium difficile. J Bacteriol. 2016;198(15):2113-30.

[278] George W, Sutter V, Citron D, Finegold S. Selective and differential medium for isolation of Clostridium difficile. J Clin Microbiol. 1979;9(2):214-9.

[279] Wilson KH, Kennedy MJ, Fekety FR. Use of sodium taurocholate to enhance spore recovery on a medium selective for Clostridium difficile. J Clin Microbiol. 1982;15(3):443-6.

[280] Edwards AN, McBride SM. Isolating and purifying Clostridium difficile spores. Method Mol Biol. 2016;1476:117-28.

[281] Childress KO, Edwards AN, Nawrocki KL, Anderson SE, Woods EC, McBride SM. The phosphotransfer protein CD1492 represses sporulation initiation in Clostridium difficile. Infect Immun. 2016;84(12):3434-44.

[282] Sorg JA, Sonenshein AL. Bile salts and glycine as cogerminants for Clostridium difficile spores. J Bacteriol. 2008;190(7):2505-12.

208 [283] Sorg JA, Sonenshein AL. Chenodeoxycholate is an inhibitor of Clostridium difficile spore germination. J Bacteriol. 2009;191(3):1115-7.

[284] Francis MB, Allen CA, Sorg JA. Spore cortex hydrolysis precedes dipicolinic acid release during Clostridium difficile spore germination. J Bacteriol. 2015;197(14):2276-83.

[285] Bourret RB. Receiver domain structure and function in response regulator proteins. Curr Opin Microbiol. 2010;13(2):142-9.

[286] Lan C-Y, Igo MM. Differential expression of the OmpF and OmpC porin proteins in Escherichia coli K-12 depends upon the level of active OmpR. J Bacteriol. 1998;180(1):171-4.

[287] Siam R, Marczynski GT. Glutamate at the phosphorylation site of response regulator CtrA provides essential activities without increasing DNA binding. Nucleic Acids Res. 2003;31(6):1775-9.

[288] Kuhn M, Goebel W. Identification of an extracellular protein of Listeria monocytogenes possibly involved in intracellular uptake by mammalian cells. Infect Immun. 1989;57(1):55-61.

[289] Rowan NJ, Candlish AA, Bubert A, Anderson JG, Kramer K, McLauchlin J. Virulent rough filaments of Listeria monocytogenes from clinical and food samples secreting wild-type levels of cell-free p60 protein. J Clin Microbiol. 2000;38(7):2643-8.

[290] Young KD. The selective value of bacterial shape. Microbiol Mol Biol Rev. 2006;70(3):660-703.

[291] Vejborg RM, Klemm P. Cellular chain formation in Escherichia coli biofilms. J Microbiol. 2009;155(5):1407-17.

[292] Kan A, Del Valle I, Rudge T, Federici F, Haseloff J. Intercellular adhesion promotes clonal mixing in growing bacterial populations. J R Soc Interface. 2018;15(146):20180406.

[293] Branda SS, González-Pastor JE, Ben-Yehuda S, Losick R, Kolter R. Fruiting body formation by Bacillus subtilis. PNAS. 2001;98(20):11621-6.

209 [294] Maule AF, Wright DP, Weiner JJ, Han L, Peterson FC, Volkman BF, et al. The aspartate-less receiver (ALR) domains: distribution, structure and function. PLoS pathogens. 2015;11(4):e1004795.

[295] Stibitz S, Aaronson W, Monack D, Falkow S. Phase variation in Bordetella pertussis by frameshift mutation in a gene for a novel two-component system. Nature. 1989;338(6212):266-9.

[296] Decker KB, James TD, Stibitz S, Hinton DM. The Bordetella pertussis model of exquisite gene control by the global transcription factor BvgA. J Microbiol. 2012;158(Pt 7):1665.

[297] Hendrixson DR. A phase-variable mechanism controlling the Campylobacter jejuni FlgR response regulator influences commensalism. Mol Microbiol. 2006;61(6):1646-59.

[298] Anjuwon-Foster BR, Tamayo R. Phase Variation of Clostridium difficile Virulence Factors. Gut Microbes. 2017:0.

[299] de Vries N, Duinsbergen D, Kuipers EJ, Pot RG, Wiesenekker P, Penn CW, et al. Transcriptional phase variation of a type III restriction-modification system in Helicobacter pylori. J Bacteriol. 2002;184(23):6615-23.

[300] Srikhanta YN, Maguire TL, Stacey KJ, Grimmond SM, Jennings MP. The phasevarion: a genetic system controlling coordinated, random switching of expression of multiple genes. PNAS. 2005;102(15):5547-51.

[301] Broadbent SE, Davies MR, van der Woude MW. Phase variation controls expression of Salmonella lipopolysaccharide modification genes by a DNA methylation-dependent mechanism. Mol Microbiol. 2010;77(2):337-53.

[302] Atack JM, Weinert LA, Tucker AW, Husna AU, Wileman TM, N FH, et al. Streptococcus suis contains multiple phase-variable methyltransferases that show a discrete lineage distribution. Nucleic Acids Res. 2018;46(21):11466-76.

[303] Manso AS, Chai MH, Atack JM, Furi L, De Ste Croix M, Haigh R, et al. A random six-phase switch regulates pneumococcal virulence via global epigenetic changes. Nat Commun. 2014;5:5055.

210 [304] van Gestel J, Vlamakis H, Kolter R. From cell differentiation to cell collectives: Bacillus subtilis uses division of labor to migrate. PLoS Biol. 2015;13(4):e1002141.

[305] Kearns DB, Losick R. Cell population heterogeneity during growth of Bacillus subtilis. J Genes. 2005;19(24):3083-94.

[306] Chen R, Guttenplan SB, Blair KM, Kearns DB. Role of the σD-dependent autolysins in Bacillus subtilis population heterogeneity. J Bacteriol. 2009;191(18):5775-84.

[307] Chai Y, Norman T, Kolter R, Losick R. An epigenetic switch governing daughter cell separation in Bacillus subtilis. J Genes. 2010;24(8):754-65.

[308] Liu H, Bouillaut L, Sonenshein AL, Melville SB. Use of a mariner-based transposon mutagenesis system to isolate Clostridium perfringens mutants deficient in gliding motility. J Bacteriol. 2013;195(3):629-36.

[309] Wuenscher MD, Köhler S, Bubert A, Gerike U, Goebel W. The iap gene of Listeria monocytogenes is essential for cell viability, and its gene product, p60, has bacteriolytic activity. J Bacteriol. 1993;175(11):3491-501.

[310] Gioppo NM, Elias Jr WP, Vidotto MC, Linhares RE, Saridakis HO, Gomes TA, et al. Prevalence of HEp-2 cell-adherent Escherichia coli and characterisation of enteroaggregative E. coli and chain-like adherent E. coli isolated from children with and without diarrhoea, in Londrina, Brazil. FEMS Microbiol Lett. 2000;190(2):293-8.

[311] Mercier C, Durrieu C, Briandet R, Domakova E, Tremblay J, Buist G, et al. Positive role of peptidoglycan breaks in lactococcal biofilm formation. Mol Microbiol. 2002;46(1):235-43.

[312] Champion JA, Mitragotri S. Role of target geometry in phagocytosis. PNAS. 2006;103(13):4930-4.

[313] Hanahan D. Studies on transformation of Escherichia coli with plasmids. J Mol Biol. 1983;166(4):557-80.

[314] McBride SM, Sonenshein AL. Identification of a genetic locus responsible for antimicrobial peptide resistance in Clostridium difficile. Infect Immun. 2011;79(1):167-76.

211 [315] Killgore G, Thompson A, Johnson S, Brazier J, Kuijper E, Pepin J, et al. Comparison of seven techniques for typing international epidemic strains of Clostridium difficile: restriction endonuclease analysis, pulsed-field gel electrophoresis, PCR-ribotyping, multilocus sequence typing, multilocus variable- number tandem-repeat analysis, amplified fragment length polymorphism, and surface layer protein A gene sequence typing. J Clin Microbiol. 2008;46(2):431-7.

[316] Sorg JA, Sonenshein AL. Inhibiting the initiation of Clostridium difficile spore germination using analogs of chenodeoxycholic acid, a bile acid. J Bacteriol. 2010;192(19):4983-90.

[317] Sullivan NM, Pellett S, Wilkins TD. Purification and characterization of toxins A and B of Clostridium difficile. Infect Immun. 1982;35(3):1032-40.

[318] Depitre C, Delmee M, Avesani V, L'Haridon R, Roels A, Popoff M, et al. Serogroup F strains of Clostridium difficile produce toxin B but not toxin A. J Med Microbiol. 1993;38(6):434-41.

[319] Edwards AN, Nawrocki KL, McBride SM. Conserved oligopeptide permeases modulate sporulation initiation in Clostridium difficile. Infect Immun. 2014;82(10):4276-91.

[320] Fagan RP, Fairweather NF. Clostridium difficile has two parallel and essential Sec secretion systems. J Biol Chem. 2011;286(31):27483-93.

[321] Trastoy R, Manso T, Fernandez-Garcia L, Blasco L, Ambroa A, Del Molino MP, et al. Mechanisms of bacterial tolerance and persistence in the gastrointestinal and respiratory environments. J Clin Microbiol Rev. 2018;31(4):e00023-18.

[322] Moradali MF, Ghods S, Rehm BH. Pseudomonas aeruginosa lifestyle: a paradigm for adaptation, survival, and persistence. J Front Cell Infect Microbiol. 2017;7:39.

[323] Fang FC, Frawley ER, Tapscott T, Vazquez-Torres A. Discrimination and integration of stress signals by pathogenic bacteria. Cell Host Microbe. 2016;20(2):144-53.

[324] Bourret RB, Silversmith RE. Two-component signal transduction. Curr Opin Microbiol. 2010;13(2):113-5.

212 [325] Krell T, Lacal J, Busch A, Silva-Jimenez H, Guazzaroni ME, Ramos JL. Bacterial sensor kinases: diversity in the recognition of environmental signals. Annu Rev Microbiol. 2010;64:539-59.

[326] Hengge R, Grundling A, Jenal U, Ryan R, Yildiz F. Bacterial signal transduction by cyclic di-GMP and other nucleotide second messengers. J Bacteriol. 2016;198(1):15-26.

[327] Hengge R. Principles of c-di-GMP signalling in bacteria. Nature reviews Microbiology. 2009;7(4):263-73.

[328] Sherwood AV, Henkin TM. Riboswitch-mediated gene regulation: novel RNA architectures dictate gene expression responses. Annu Rev Microbiol. 2016;70:361-74.

[329] van der Woude MW. Re-examining the role and random nature of phase variation. FEMS Microbiol Lett. 2006;254(2):190-7.

[330] Abraham JM, Freitag CS, Clements JR, Eisenstein BI. An invertible element of DNA controls phase variation of type 1 fimbriae of Escherichia coli. PNAS. 1985;82(17):5724-7.

[331] Klemm P. Two regulatory fim genes, fimB and fimE, control the phase variation of type 1 fimbriae in Escherichia coli. EMBO J. 1986;5(6):1389-93.

[332] Blomfield I, van der Woude M. Regulation of Fimbrial Expression. EcoSal Plus. 2007;2(2):10.1128/ecosal.2.4.2.2.

[333] Cartman ST, Heap JT, Kuehne SA, Cockayne A, Minton NP. The emergence of 'hypervirulence' in Clostridium difficile. IJMM. 2010;300(6):387-95.

[334] Garrett EM, Sekulovic O, Wetzel D, Jones JB, Edwards AN, Vargas-Cuebas G, et al. Phase variation of a signal transduction system controls Clostridioides difficile colony morphology, motility, and virulence. PLoS Biol. 2019;17(10).

[335] Snyder JA, Lloyd AL, Lockatell CV, Johnson DE, Mobley HL. Role of phase variation of type 1 fimbriae in a uropathogenic Escherichia coli cystitis isolate during urinary tract infection. Infect Immun. 2006;74(2):1387-93.

213 [336] Edwards AN, Pascual RA, Childress KO, Nawrocki KL, Woods EC, McBride SM. An alkaline phosphatase reporter for use in Clostridium difficile. Anaerobe. 2015;32:98-104.

[337] Groisman EA. Feedback Control of Two-Component Regulatory Systems. Annu Rev Microbiol. 2016;70:103-24.

[338] Soncini FC, Véscovi EG, Groisman EA. Transcriptional autoregulation of the Salmonella typhimurium phoPQ operon. J Bacteriol. 1995;177(15):4364-71.

[339] Williams CL, Cotter PA. Autoregulation is essential for precise temporal and steady-state regulation by the Bordetella BvgAS phosphorelay. J Bacteriol. 2007;189(5):1974-82.

[340] Arthur M, Depardieu F, Gerbaud G, Galimand M, Leclercq R, Courvalin P. The VanS sensor negatively controls VanR-mediated transcriptional activation of glycopeptide resistance genes of Tn1546 and related elements in the absence of induction. J Bacteriol. 1997;179(1):97-106.

[341] Gao R, Mack TR, Stock AM. Bacterial response regulators: versatile regulatory strategies from common domains. Trends Biochem Sci. 2007;32(5):225-34.

[342] Zschiedrich CP, Keidel V, Szurmant H. Molecular mechanisms of two- component signal transduction. J Mol Biol. 2016;428(19):3752-75.

[343] Zhu Y, Qin L, Yoshida T, Inouye M. Phosphatase activity of histidine kinase EnvZ without kinase catalytic domain. PNAS. 2000;97(14):7808-13.

[344] Carmany DO, Hollingsworth K, McCleary WR. Genetic and biochemical studies of phosphatase sctivity of PhoR. J Bacteriol. 2003;185(3):1112-5.

[345] Gao R, Bouillet S, Stock AM. Structural Basis of Response Regulator Function. Annu Rev Microbiol. 2019;73:175-97.

[346] Capra EJ, Laub MT. Evolution of two-component signal transduction systems. Annu Rev Microbiol. 2012;66:325-47.

[347] Galperin MY. A census of membrane-bound and intracellular signal transduction proteins in bacteria: bacterial IQ, extroverts and introverts. BMC microbiology. 2005;5:35.

214 [348] Tobe T. The roles of two-component systems in virulence of pathogenic Escherichia coli and Shigella spp. Adv Exp Med Biol. 2008;631:189-99.

[349] Chen Q, Stibitz S. The BvgASR virulence regulon of Bordetella pertussis. Curr Opin Microbiol. 2019;47:74-81.

[350] Francis VI, Stevenson EC, Porter SL. Two-component systems required for virulence in Pseudomonas aeruginosa. FEMS Microbiol Lett. 2017;364(11).

[351] Jenul C, Horswill AR. Regulation of Staphylococcus aureus virulence. Microbiol Spectr. 2018;6(1).

[352] Voyich JM, Vuong C, DeWald M, Nygaard TK, Kocianova S, Griffith S, et al. The SaeR/S gene regulatory system is essential for innate immune evasion by Staphylococcus aureus. J Infect Dis. 2009;199(11):1698-706.

[353] Baroja ML, Herfst CA, Kasper KJ, Xu SX, Gillett DA, Li J, et al. The SaeRS two-component system is a direct and dominant transcriptional activator of toxic shock syndrome toxin 1 in Staphylococcus aureus. J Bacteriol. 2016;198(19):2732-42.

[354] Olson ME, Nygaard TK, Ackermann L, Watkins RL, Zurek OW, Pallister KB, et al. Staphylococcus aureus nuclease is an SaeRS-dependent virulence factor. Infect Immun. 2013;81(4):1316-24.

[355] Moscoso JA, Jaeger T, Valentini M, Hui K, Jenal U, Filloux A. The diguanylate cyclase SadC is a central player in Gac/Rsm-mediated biofilm formation in Pseudomonas aeruginosa. J Bacteriol. 2014;196(23):4081-8.

[356] Chakraborty S, Kenney LJ. A new role of OmpR in acid and osmotic stress in Salmonella and E. coli. Front Microbiol. 2018;9(2656).

[357] Kenney LJ. The role of acid stress in Salmonella pathogenesis. Curr Opin Microbiol. 2019;47:45-51.

[358] Hong H-J, Hutchings MI, Buttner MJ. Vancomycin Resistance VanS/VanR Two-Component Systems. In: Utsumi R, editor. Bacterial Signal Transduction: Networks and Drug Targets. New York, NY: Springer New York; 2008. p. 200-13.

215 [359] Tiwari S, Jamal SB, Hassan SS, Carvalho P, Almeida S, Barh D, et al. Two- component signal transduction systems of pathogenic bacteria as targets for antimicrobial therapy: an overview. Front Microbiol. 2017;8:1878.

[360] Gumerov Vadim M, Ortega Davi R, Adebali O, Ulrich Luke E, Zhulin Igor B. MiST 3.0: an updated microbial signal transduction database with an emphasis on chemosensory systems. Nucleic Acids Res. 2019;48(D1):D459-D64.

[361] Schujman GE, Paoletti L, Grossman AD, de Mendoza D. FapR, a bacterial transcription factor involved in global regulation of membrane lipid biosynthesis. Dev Cell. 2003;4(5):663-72.

[362] Procaccini A, Lunt B, Szurmant H, Hwa T, Weigt M. Dissecting the specificity of protein-protein interaction in bacterial two-component signaling: orphans and crosstalks. PLOS ONE. 2011;6(5):e19729-e.

[363] Möglich A, Ayers RA, Moffat K. Structure and signaling mechanism of Per- ARNT-Sim domains. Structure. 2009;17(10):1282-94.

[364] Upadhyay AA, Fleetwood AD, Adebali O, Finn RD, Zhulin IB. Cache domains that are homologous to, but different from PAS domains comprise the largest superfamily of extracellular sensors in prokaryotes. PLOS Comput Biol. 2016;12(4):e1004862.

[365] McCoy JG, Levin EJ, Zhou M. Structural insight into the PTS sugar transporter EIIC. Biochimica et biophysica acta. 2015;1850(3):577-85.

[366] Tam R, Saier MH, Jr. Structural, functional, and evolutionary relationships among extracellular solute-binding receptors of bacteria. Microbiological reviews. 1993;57(2):320-46.

[367] Liu W, Hulett FM. Bacillus subtilis PhoP binds to the phoB tandem promoter exclusively within the phosphate starvation-inducible promoter. J Bacteriol. 1997;179(20):6302-10.

[368] Head CG, Tardy A, Kenney LJ. Relative binding affinities of OmpR and OmpR- phosphate at the ompF and ompC regulatory sites. J Mol Biol. 1998;281(5):857- 70.

[369] Galperin MY, Nikolskaya AN. Identification of sensory and signal-transducing domains in two-component signaling systems. Method Enzymol. 2007;422:47-74.

216 [370] Wolanin PM, Thomason PA, Stock JB. Histidine protein kinases: key signal transducers outside the animal kingdom. Genome Biol. 2002;3(10):Reviews3013.

[371] Jiang M, Shao W, Perego M, Hoch JA. Multiple histidine kinases regulate entry into stationary phase and sporulation in Bacillus subtilis. Mol Microbiol. 2000;38(3):535-42.

[372] Garcia Vescovi E, Soncini FC, Groisman EA. Mg2+ as an extracellular signal: environmental regulation of Salmonella virulence. Cell. 1996;84(1):165-74.

[373] Groisman EA. The pleiotropic two-component regulatory system PhoP-PhoQ. J Bacteriol. 2001;183(6):1835-42.

[374] Baker AE, O'Toole GA. Bacteria, rev your engines: stator dynamics regulate flagellar motility. J Bacteriol. 2017;199(12): e00088-17.

[375] Hou YJ, Yang WS, Hong Y, Zhang Y, Wang DC, Li DF. Structural insights into the mechanism of c-di-GMP-bound YcgR regulating flagellar motility in Escherichia coli. J Biol Chem. 2020;295(3):808-21.

[376] Zorraquino V, García B, Latasa C, Echeverz M, Toledo-Arana A, Valle J, et al. Coordinated cyclic-di-GMP repression of Salmonella motility through YcgR and Cellulose. J Bacteriol. 2013;195(3):417-28.

[377] Hölscher T, Kovács ÁT. Sliding on the surface: bacterial spreading without an active motor. Environ Microbiol. 2017;19(7):2537-45.

[378] Jarrell KF, McBride MJ. The surprisingly diverse ways that prokaryotes move. Nature reviews Microbiology. 2008;6(6):466-76.

[379] Nan B, Zusman DR. Novel mechanisms power bacterial gliding motility. Mol Microbiol. 2016;101(2):186-93.

[380] Martínez A, Torello S, Kolter R. Sliding motility in Mycobacteria. J Bacteriol. 1999;181(23):7331-8.

[381] Wang JD, Levin PA. Metabolism, cell growth and the bacterial cell cycle. Nature reviews Microbiology. 2009;7(11):822-7.

217 [382] Heidrich C, Templin MF, Ursinus A, Merdanovic M, Berger J, Schwarz H, et al. Involvement of N-acetylmuramyl-l-alanine amidases in cell separation and antibiotic-induced autolysis of Escherichia coli. Mol Microbiol. 2001;41(1):167-78.

[383] Salamaga B, Prajsnar TK, Jareño-Martinez A, Willemse J, Bewley MA, Chau F, et al. Bacterial size matters: Multiple mechanisms controlling septum cleavage and diplococcus formation are critical for the virulence of the opportunistic pathogen Enterococcus faecalis. PLoS pathogens. 2017;13(7):e1006526-e.

[384] Yang DC, Blair KM, Salama NR. Staying in shape: the impact of cell shape on bacterial survival in diverse environments. Microbiol Mol Biol Rev. 2016;80(1):187-203.

[385] Horvath DJ, Jr., Li B, Casper T, Partida-Sanchez S, Hunstad DA, Hultgren SJ, et al. Morphological plasticity promotes resistance to phagocyte killing of uropathogenic Escherichia coli. Microbes Infect. 2011;13(5):426-37.

[386] Prashar A, Bhatia S, Gigliozzi D, Martin T, Duncan C, Guyard C, et al. Filamentous morphology of bacteria delays the timing of phagosome morphogenesis in macrophages. The Journal of cell biology. 2013;203(6):1081- 97.

[387] Dalia AB, Weiser JN. Minimization of bacterial size allows for complement evasion and is overcome by the agglutinating effect of antibody. Cell Host Microbe. 2011;10(5):486-96.

[388] Rodriguez JL, Dalia AB, Weiser JN. Increased chain length promotes pneumococcal adherence and colonization. Infect Immun. 2012;80(10):3454-9.

[389] Prashar A, Bhatia S, Tabatabaeiyazdi Z, Duncan C, Garduno RA, Tang P, et al. Mechanism of invasion of lung epithelial cells by filamentous Legionella pneumophila. Cell Microbiol. 2012;14(10):1632-55.

[390] Josenhans C, Suerbaum S. The role of motility as a virulence factor in bacteria. IJMM. 2002;291(8):605-14.

[391] Rütschlin S, Böttcher T. Inhibitors of bacterial swarming behavior. Chemistry (Weinheim an der Bergstrasse, Germany). 2020;26(5):964-79.

218 [392] Lane MC, Alteri CJ, Smith SN, Mobley HL. Expression of flagella is coincident with uropathogenic Escherichia coli ascension to the upper urinary tract. PNAS. 2007;104(42):16669-74.

[393] Rather PN. Swarmer cell differentiation in Proteus mirabilis. Environ Microbiol. 2005;7(8):1065-73.

[394] Belas R, Suvanasuthi R. The ability of Proteus mirabilis to sense surfaces and regulate virulence gene expression involves FliL, a flagellar basal body protein. J Bacteriol. 2005;187(19):6789-803.

[395] Jones BV, Young R, Mahenthiralingam E, Stickler DJ. Ultrastructure of Proteus mirabilis swarmer cell rafts and role of swarming in catheter-associated urinary tract infection. Infect Immun. 2004;72(7):3941-50.

[396] Sabbuba N, Hughes G, Stickler DJ. The migration of Proteus mirabilis and other urinary tract pathogens over Foley catheters. BJU Int. 2002;89(1):55-60.

[397] Taylor TB, Buckling A. Selection experiments reveal trade-offs between swimming and twitching motilities in Pseudomonas aeruginosa. Evolution. 2011;65(11):3060-9.

219