Regulatory transcription factors of Clostridioides difficile pathogenesis with a focus on toxin regulation

Abstract

Clostridioides difficile (CD), a nosocomial gut pathogen, produces two major exotoxins, TcdA and TcdB, which disrupt the gut epithelial barrier and induce inflammatory/immune responses, leading to symptoms ranging from mild diarrhoea to pseudomembranous colitis and potentially to death. The expression of toxins is regulated by various transcription factors (TFs) which are induced in response to CD physiological life stages, nutritional availability, and host environment. This review summarises our current understanding on the regulation of toxin expression by TFs that interconnect with pathways of flagellar synthesis, quorum sensing, motility, biofilm formation, sporulation, and phase variation. The pleiotropic roles of some key TFs suggest that toxin production is tightly linked to other cellular processes of the CD physiology.

SUMMARY

This review summarises the current knowledge of the transcription factors involved in regulation of toxin production, which is affected by C. difficile physiological life stages, nutritional availability, and host environment in the gut.

Introduction

Clostridioides difficile (CD), a toxin-producing and spore-forming anaerobic bacterium, causes nosocomial antibiotic-induced diarrhoea, which may progress to pseudomembranous colitis and potential death (Lawson et al. 2016; Oren and Garrity 2018). In the United States, CD caused an estimated 223,900 infections and 12,800 deaths in 2017 with a loss of $1 billion according to the Centres for Disease Control and Prevention’s 2019 report (2019) (CDC 2019).

CD pathogenesis is caused mainly by the expression and secretion of two major toxins TcdA and TcdB (Zhu et al. 2018; Cohen et al. 2000). The two toxin genes (tcdA and tcdB) and the associated genes tcdR, tcdC and tcdE are located within a 19.6-kb pathogenicity locus (PaLoc) in CD genome. The gene tcdR encodes an alternative RNA polymerase sigma factor that activates the expression of both toxin genes, whereas tcdC encodes an anti-sigma factor whose function in relations to regulation of TcdR activity remains debateable. tcdE encodes a holin-like protein involved in the secretions of the toxins by a phage-like system (Cohen et al. 2000; Govind and Dupuy 2012; Zhu et al. 2018). In addition to TcdA and TcdB, about 20% of CD isolates produce a third toxin, binary toxin (CDT), which is an AB-type toxin encoded by two genes cdtA and cdtB in the CdtLoc locus (Carter et al. 2007; Gerding et al. 2014). Genes cdtA and cdtB are regulated by CdtR a LytTR type of response regulator, which directly activates these genes by binding to the upstream promoter region of the cdtA gene (Carter et al. 2007; Gerding et al. 2014).

Toxin synthesis is highly dynamic and influenced by growth conditions and phases, antibiotic stresses, and nutritional signals. A nutrient-rich medium with rapidly metabolisable sugars and amino acids represses toxin synthesis, while a nutrient-limited condition activates toxin synthesis (Hundsberger et al. 1997; Dupuy and Sonenshein 1998; Drummond et al. 2003; Bouillaut et al. 2013). The role of TcdR as a sigma factor has been established while the role of TcdC as an anti-sigma factor repressing toxin expression is currently debateable. During stationary phase, the PaLoc genes (tcdR, tcdB, tcdE and tcdA) are coordinately expressed, with the level of tcdA transcription greater than that of tcdB transcription, with robust expression in hypervirulent strains which is consistent with the abundance of TcdA present in the CD culture supernatant. In this study, tcdC expression did not diminish in stationary phase, suggesting its modulatory role rather than repressive (Merrigan et al. 2010). In another study using tcdC mutant in C. difficile 630Δerm, total toxin levels did not significantly differ in the wild type and the mutant strain. These results in C. difficile 630Δerm rejects the notion that the TcdC is the major regulator of toxin expression (Bakker et al. 2012).

Production of TcdA and TcdB is also tightly linked to critical physiology, regulatory systems and virulence of CD such as sporulation, quorum sensing, biofilm formation, motility and colonisation, which are under the control of a highly regulated network of regulatory proteins known as transcription factors (TFs) (Martin-Verstraete et al. 2016; Shen et al. 2019). Transcriptional regulation by DNA binding proteins is the most common mechanism of gene regulation in bacteria and these TFs are activated in response to environmental/nutritional fluctuations during CDI (Martin-Verstraete et al. 2016; Zhu et al. 2018). Various CD TFs that control regulation of toxin production, initiation of sporulation, quorum-sensing and biofilm formation have been identified (Gerding et al. 2014; Martin-Verstraete et al. 2016; Gil et al. 2017). The highlights of this review are presented in Table 1. In this review, we summarise the current understanding on how these TFs (Table 2) that control the key steps in the pathogenesis of CDI influence on toxin production and the critically unanswered question that needs to be addressed (Table 3).

Table 1.

Highlights of this review.

• TcdA and TcdB are the major exotoxin of C. difficile and are encoded by genes in the PaLoc region. Expression of toxin genes is directly controlled by a sigma factor TcdR. The expression of tcdR in turn is regulated by various TFs affected by nutritional availability such as CodY, CcpA, Lrp SigL, YpdA/YpdB, PrdR, Rex, TreR and Mfd.

• Regulators of C. difficile sporulation such as Spo0A, RstA, SinR/SinR’ and CodY are linked to toxin production.

• Quorum sensing in C. difficile linked with toxin production, is mediated by LuxSCD auto-inducer (AI-2) and the agr operon expression that produces cyclic AIP. LuxSCD is controlled by the transcriptional regulator RolA, while the AgrA2 transcription factor is activated by AgrC2 phosphorylation via cyclic AIP binding.

• Regulators of biofilm formation in C. difficile such as Spo0A, LexA and c-di-GMP cross talk with the expression of the toxin genes.

• Regulation of a c-di-GMP operated flagellar operon synergizes with toxin production via sigD expression that is mediated by a flagellar switch.

Table 2.

Known transcriptional regulators in C. difficile involved in pathogenesis.

Known transcriptional regulators in C. difficile
Transcriptional regulators	Locus tag	Effect on virulence traits	References
TcdR	CD0659	Positively regulates toxin synthesis; Increases sporulation	(Mani and Dupuy 2001; Girinathan et al. 2017)
CdtR	CD630_RS13955	Induces toxin production	(Matamouros et al. 2007; Lyon et al. 2016; Bilverstone et al. 2019)
TcdC	CD630_06640	Anti-sigma factor inhibits toxin synthesis	(Matamouros et al. 2007)
RstA	CD3668	Enhances sporulation, Inhibits toxin production	(Edwards et al. 2016, 2019)
SinR	CD2214	Inhibits sporulation, enhances toxin and motility	(Girinathan et al. 2018)
SinR’	CD2215	Inhibits SinR activity	(Ciftci et al. 2019)
CodY	CD1275	Inhibits toxin synthesis and sporulation	(Dineen et al. 2007)
Lrp	CD3544	Inhibits toxin production and increases sporulation	(Chen et al. 2019)
YpdA/YpdB (TCS)	CD2602/CD2601	Mediates toxin synthesis in response to pyruvate	(Dubois et al. 2016)
CdsR	CD630_32330	Activates CdsB protein that represses toxin synthesis	(Gu et al. 2017)
SigL	CD3176	Activates CdsB protein that represses toxin synthesis	(Gu et al. 2017)
SigD	CD0266	Represses toxin genes with the help of RstA and TcdR expression	(Edwards et al. 2019)
CcpA	CD1064	Represses toxin synthesis	(Antunes et al. 2012)
Mfd	CD630_2775	Help de-repression of toxin genes	(Willing et al. 2015)
Rol A / Rol B (TCS)	CDIF630_03922/CDIF630_03921	Regulate LuxS activity that produces auto-inducer AI-2 has role in quorum sensing and biofilm formation	(Carter et al. 2005; Slater et al. 2019)
AgrA2	CDIF27147_03515	Quorum sensing and toxin synthesis	(Darkoh et al. 2015)
Spo0A	CD1214	Master regulator of sporulation initiation	(Dawson et al. 2012)
LexA	CD1931	Increase Biofilm formation and sporulation under stress	(Walter et al. 2015)
Rho factor	CDR20291_3324	affects phase variation in CD virulence genes	(Trzilova et al. 2020)
SigH	CD0057	Early phase of sporulation initiation and upregulates Spo0A	(Rosenbusch et al. 2012)
SigA	CD630_RS08205	Activates σH expression, functions in sporulation initiation	(Al-Hinai et al. 2015)
CsiA	CD25890	Controls sporulation initiation and downregulates Spo0A	(Martins et al. 2020)
SpoIIID	CD630_RS01145	Regulates σK expression	(Girinathan et al. 2017)
SigF	CD0772	communication between FC and MC and activates σG	(LaBell et al. 1987; Zhu et al. 2018)
SigE	CD2643	Regulates cspBAC operon, increases spore germination, regulated FC engulfment, cortex formation and coat assembly	(Fimlaid et al. 2013; Saujet et al. 2013)
SigG	CD2642	Controls gene with function in importing DPA, cell wall and cortex synthesis	(Saujet et al. 2013)
SigK	CD630_12300	specific to the MC, function in coat layer assembly, spore maturation and MC lysis, regulates SpoIIID expression	(Saujet et al. 2013; Pereira et al. 2013)
CamA	CD2758	Epigenetic regulator of sporulation, biofilm, and host colonisation	(Oliveira et al. 2020)
TreR	CDIF27147_03229	TreR represses treA that metabolises trehalose	(Collins et al. 2018)
RgaR	CD3255	Positively regulates agrBD locus	(O’Connor et al. 2006)
PrdR	CDIF27147_03424	PrdR represses toxin synthesis	(Bouillaut et al. 2013)

Table 3.

Critical unanswered questions.

• The role of TcdC in toxin regulation is currently not well established and needs further investigation.

• How nitrogen regulated response (Ntr), under nitrogen limitation, impact toxin expression and virulence in C. difficile has not been studied. Therefore, identification of global regulators of nitrogen regulated response (Ntr) genes might open new avenues for research in this direction.

• How TFs regulate C. difficile antibiotics resistance is largely unknown and warrants investigation due to the prevalence and emergence of C. difficile strains with multiple antibiotic resistance.

• How TFs regulate C. difficile germination is currently unknown. Research in this direction will unravel the regulation of spore germination.

• Further investigation is needed about how LexA controls biofilm formation by antibiotic induced stress conditions.

• The epigenetic regulator CamA influences key processes in CDI (Oliveira et al. 2020). It will be interesting to investigate how the common targets of the TFs and the CamA are turned on and off because of cross talk between the epigenetic regulators and the TFs.

• Eight flagellar switches have been identified in C. difficile, and only two have been characterised till date (Anjuwon-Foster et al. 2018). Therefore, characterisation of these putative flagellar switches might reveal other key players in toxin regulation.

• How different TFs are activated in response to the environmental signals such as nutrient availability, spore germination conditions, antibiotic stress, and DNA damage warrants further investigation using cutting edge tools of immunoproteomics and transcriptomics.

• Identification of all the C. difficile TFs using transcription factor data mining tools such as p2tf (http://www.p2tf.org/… ) and Predicted Prokaryotic Regulatory Proteins analysis tools (http://www.p2rp.org… ) (Ortet et al. 2012; Barakat et al. 2013), will open new research line in CD pathogenicity, which might unravel how pathways of C. difficile spore germination, motility, colonisation, biofilm formation, quorum-sensing and sporulation are interlinked with toxin gene expression.

Transcriptional regulators of tcdA and tcdB

The expression of tcdA and tcdB is positively regulated by TcdR (Figure 1). TcdR interacts with the RNA polymerase (RNAP) and directs the core RNAP to recognise the promoters of tcdA and tcdB (Mani and Dupuy 2001; Mani et al. 2002). The expression of tcdR is regulated by several other sigma factors and TFs. The region upstream of tcdR contains four different independent promoter elements: one σ^A-dependent promoter, one σ^D-dependent promoter and two σ^TcdR-dependent promoters (Mani et al. 2002; Edwards et al. 2020).

Figure 1.

Regulation of toxin genes. (A) Transcriptional regulation of TcdA and TcdB: During sufficient nutrition availability, the positive regulators of toxin genes, TcdR and SigD, are unable to direct the RNAP to start the transcription due to CodY and CcpA mediated blockage at the promoter sites and by TcdC interference of TcdR-RNAP leading to transcriptional repression. In contrast, during limited nutrition condition, CodY and CcpA mediated repression is relieved. (B) Co-transcriptional regulations of flagellar and toxin genes: flgB operon genes encode many proteins involved in flagellar assembly including SigD that directly regulates toxin gene of the PaLoc region. This C-di-GMP regulated flgB operon is also regulated by a genetic switch that regulates the flgB operon by flipping the switch in either direction of ON or OFF lock mode with the help of RecV enzymes that catalyses the inversion of the switch from both directions. A Rho factor can discriminate the ON and OFF mode and helps the early termination of mRNA transcript within the 5’ leader sequence. Regulation of the flagellar switch influences flagellar gene expression and toxin synthesis that leads to phase variation heterogeneity. Oval shapes denote TFs; Square shape denotes key proteins involved in phase variation.

Under nutrient-rich conditions, tcdR expression is low due to the binding of various regulatory factors that these promoter elements (Figure 1). The low expression of the toxin genes have been linked to TcdC expression but TcdC does not bind to the promoter of the toxin genes and its interaction either with TcdR or the TcdR-RNAP complex or both, that are still contested (Figure 1 and Table 3) (Matamouros et al. 2007). However, if the TcdR-RNAP complex is preformed, TcdC cannot prevent transcription initiation of the toxins (Matamouros et al. 2007). A recent study showed that the C-terminal domain of TcdC, which is responsible for intracellular target binding, is located extracellularly, and is therefore incompatible as an anti-sigma factor (Oliveira Paiva et al. 2020).

Positive regulation of toxin gene expression by SigD occurs through the induction of tcdR gene expression via the σ^D-dependent promoter present upstream of the tcdR gene (Figure 1) (El Meouche et al. 2013).

Toxin production is also tightly regulated by several other TFs and may be strain-dependent (Edwards et al. 2020) (Figure 2). Recently, a study with two ribotype 027 (RT027) strains demonstrated that CdtR, a response regulator of the CDT genes, also regulates the production of TcdA and TcdB (Lyon et al. 2016). Additionally, CdtR is activated by phosphorylation of Asp61 (Bilverstone et al. 2019).

Figure 2.

Nutritional regulators of Toxin expression. (A) Strong CodY binding to tcdR promoter in the presence of GTP and BCAs (solid arrow) and weak affinity for toxin promoters (dashed arrow). CdTR positively induces toxin synthesis. CcpA and RstA mediates repression by directly binding to toxin gene promoters. Mfd aids in CcpA and Cody mediated toxin repression. Dashed bold line show unknown mechanism in the case of TcdC and Mfd. (B) Several TFs such as, a phage-encoded protein RepR, SinR’/SinR and Spo0A negatively regulate toxin gene expression. TreR inhibits toxin genes by metabolising trehalose.PrdR represses toxin by indirectly modulating the activity of Rex. LRP inhibits toxin production under limiting nutrition condition. Cysteine disulphidase B (CdsB) represses toxin that is induced by SigL and Cysteine disulphidase R (CdsR). YpdA/B TCS partly mediate toxin production in the presence of pyruvate. Transcription factors and other factors in the oval boxes indirectly repress the toxin expression (round dotted arrows).

Toxin production is also influenced by the presence of prophages in CD. The phage-encoded protein RepR of phage CD119, downregulates all the genes of the PaLoc region. The effect of RepR on toxin repression is mediated by repressing the transcription of tcdR gene that has RepR upstream binding sites in its upstream regions (Govind et al. 2009) (Figure 2). In contrast, the lysogenic phage CD38-2, with a host range of 11 ribotypes, enhances the toxin levels with increased transcription of all the PaLoc genes (Sekulovic et al. 2011). Similarly, other phages have been reported to influence toxin synthesis (Goh et al. 2005). Thus, further investigation of regulation of toxin synthesis by phages may provide another opportunity to understand the pathogenicity of new emerging ribotypes.

Regulators involved in nutritional regulation of TcdA and TcdB

Nutritional signalling is one of the important environmental cues that regulate toxin synthesis. These include certain amino acids, biotins, glucose availability and nutrient-rich or nutrient-poor conditions can activate TFs and two-component signalling systems that may directly or indirectly influence toxin synthesis through binding within promoter regions of the toxin genes or to the tcdR promoter sites (Figure 2).

One such nutrient monitoring system is the global regulator, CodY, a dimeric protein that has a winged helix-turn-helix (HTH) motif at its C-terminus. In CD, CodY has greater than 500 genes under its control with 495 being negatively regulated and 57 being positively regulated (Dineen et al. 2010). Compared to the wild-type strain, a CD codY mutant showed significant upregulation in all the five genes in the PaLoc locus during mid-exponential and early stationary phase when grown in the enriched media (Dineen et al. 2007). CodY-mediated toxin repression was demonstrated by its strong binding affinity for the tcdR promoter region, but weak affinity for the toxin promoters, suggesting that CodY regulates the toxin synthesis through interaction with tcdR promoter. Additionally, binding of CodY to the tcdR promoter increased in the presence of GTP and branched chain amino acids (isoleucine, leucine, and valine) (Figure 3) (Dineen et al. 2007). Moreover, DNase I protection assays demonstrated that CodY has three binding regions within the tcdR promoter region: region I (position −348 to −382), region II (position −281 to −309) and region III (position −40 to −58) (Dineen et al. 2007). In a recent study, a codY null mutant showed increased pathogenicity compared to the wild type (Daou et al. 2019). In addition, the mutant showed increased expression of sporulation genes even during exponential growth (Daou et al. 2019). GTP molecules that activate the binding of CodY to the tcdR promoter to suppress toxin synthesis are a source of c-di-GMP, a key mediator of biofilm formation and flagellar gene expression. CodY regulates many genes that are involved in the synthesis and degradation of c-di-GMP, and therefore changes in the levels of c-di-GMP, GTP might influence the activity of the CodY, leading to repression or derepression of toxin genes (Purcell et al. 2017). Thus, it may be argued that when the pathogen encounters the nutrient-limiting conditions during its colonisation in the gut, release of CodY-mediated repression allows for toxin synthesis that helps liberate essential nutrients upon damage to the epithelial cells.

Figure 3.

QS and biofilm formation regulators influence toxin synthesis. (A) LuxS activity controlled by RolA transcriptional regulator, catalyses formation of QS molecule AI-2 which induces toxin production and biofilm production. Expression of DccA induces c-di-GMP that upregulates biofilm formation and toxin synthesis, while diguanylate phosphodiesterase PdcA inhibits biofilm formation and is induced or inhibited by CodY under low and high GTP concentrations, respectively. Biofilm is also positively controlled by the master regulator Spo0A, a negatively regulator of toxin synthesis. The epigenetic regulator CamA also positively modulates biofilm formation. The transcriptional repressor LexA that promotes biofilm formation under antibiotic stress conditions inhibits the synthesis of TcdA in the presence of antibiotics such as levofloxacin by binding at the tcdA promoter. (B) Agr mediated QS in C. difficile. In CD, there are three Agr loci, Agr2 locus contains the full set genes (agrB2D2 and agrC2A2) of the QS pathway, while Agr1 is incomplete (agrB1 and agrD1) regulate toxin genes. Agr3 lacks the AgrA response regulator (Not shown here). The agrD2 encodes the pre-AIP and is processed by AgrB2 to form the cyclic- auto- induce peptide (AIP). Genes agrC2 and agrA2 form a two-component signal transduction system, in which the sensor histidine kinase AgrC2, a transmembrane protein is phosphorylated by binding with AIP and the response regulator AgrA2. The response regulator is then activated by AgrC2 mediated phosphorylation. Phosphorylated AgrA2 then regulates transcription of many target genes including toxin genes. Oval shapes denote TFs; square shape denotes proteins involved in c-di-GMP regulation; Green circles denote QS molecule AI-2; doughnut shape circles show cyclic AIP.

Leucine responsive regulatory protein (Lrp) is another global nutritional regulator that senses the concentrations of leucine and alanine in the cell and is involved in the regulation of toxin genes when the cell enters the stationary phase of growth (Figure 2). Lrp belongs to the Lrp/asnC family of proteins and has an N-terminal helix-turn-helix (HTH) motif and a βαββαβ-fold (αβ-sandwich) at the C-terminus. The HTH motif of CD Lrp also contains a conserved lysine residue that is required for the DNA-binding activity (Chen et al. 2019). Disruption of the lrp gene in CD increased transcription of the tcdR, tcdA and tcdB genes, while it decreased the transcription of the codY gene. In addition, the lrp mutant showed enhanced sporulation efficiency via increased expression of spo0A, the master transcriptional regulator of sporulation (Chen et al. 2019).

The availability of cysteine influences CD toxin synthesis. In the presence of cysteine, transcription of tcdA, tcdB and tcdR decreases significantly, indicating that cysteine metabolism mediates toxin repression through an unknown mechanism. Cysteine, a sulphur containing amino acid and a donor of S to many Fe-S cluster-containing enzymes, is required for the synthesis of methionine and other co-enzymes (Gu et al. 2017; Dubois et al. 2016). In CD, pyruvate and sulphides are produced as by-products in the presence of cysteine as a result of cysteine metabolism by cysteine desulfhydrases. A cysteine desulfidase, CdsB, in CD has been identified, and inactivation of the cdsB gene in CD showed cysteine-dependent de-repression of toxin production. Expression of cdsB was induced by SigL and CdsR (CD630_32330 gene) in response to cysteine (Figure 1(B)) (Gu et al. 2017). Additionally, inactivation of sigL results in de-repression of toxin genes in the presence of cysteine with a decrease in the production of pyruvate and H₂S as compared to the wild-type strain (Dubois et al. 2016). Further supplementation with pyruvate for a short interval reduced the expression of the toxin genes in the sigL mutant, suggesting that the accumulated by-products play a role in cysteine-dependent repression. Additionally, the generated pyruvate is further utilised by a nutrient scavenging two-component system (TCS). In CD, a TCS (i.e., CD2602/CD2601) similar to the E. coli YpdA/YpdB TCS system, was inactivated and tested in response to pyruvate availability. In this mutant, the addition of pyruvate significantly reduced the transcript levels of tcdA, tcdB and tcdR as compared to the wild-type, suggesting that this TCS partly mediated the regulation of tcdA, tcdB and tcdR regulon in response to pyruvate availability (Dubois et al. 2016).

PrdR, a proline-responsive regulatory protein activates the synthesis of proline reductase (PR) and glycine reductase (GR) enzymes that generate NAD⁺ from NADH in Stickland metabolism where proline and glycine act as inducers and substrates for PR and GR, respectively. Proline and glycine suppress toxin synthesis, and it has been demonstrated that PrdR also represses tcdA expression in a proline-dependent manner (Karlsson et al. 1999; Bouillaut et al. 2013). The global redox regulator Rex responds to the NAD⁺/NADH ratio and regulates many metabolic pathways that lead to the production of glycine, succinate, and butyrate. Recently, it was demonstrated that PrdR represses the production of butyrate, a known inducer of toxin, by indirectly modulating the activity of Rex (Figure 2) (Bouillaut et al. 2019). PrdR suppress toxin synthesis in a proline dependent manner and suppresses production of butyrate, however nothing is known how PrdR controls toxin synthesis. Considering the involvement of multiple pathways further experimental investigation is necessary to unravel the direct or indirect control of toxin production by PrdR.

Glucose and other rapidly metabolisable carbon sources that suppress toxin synthesis in CD are transported by Phosphoenolpyruvate-dependent phosphotransferase system (PTS). PTS systems are composed of enzymes I and II for carbohydrate specific transporters and histidine-containing phosphocarrier protein (HPr). The PTS genes undergo carbon catabolite repression (CCR) (Dupuy and Sonenshein 1998). In CD, about 18% of the cell’s genes are regulated in response to glucose through CCR, which is mediated by the activity of transcription factor, Carbon catabolite control protein A (CcpA). CcpA is a member of LacI/GalR repressor family and is a pleotropic regulator that controls many genes in response to carbohydrate catabolism. The genes in the ccpA regulon take part in sugar uptake, amino acid metabolism and fermentation processes(Antunes et al. 2012). The CcpA activity is enhanced in the presence of a phosphorylated form of HPr protein, phosphorylated at serine residues during glucose metabolism by HPr kinase, a component of PTS (Martin-Verstraete et al. 2016). In other bacteria, interaction of CcpA with phosphorylated HPr (serine-containing phosphocarrier protein) leads to the specific recognition of the catabolite responsive elements (Cre) in the promoter regions of the target genes. However in CD, CcpA binds directly to the regulatory regions of the tcdA and tcdB genes, and not through Cre, and mediates glucose-dependent repression of CD toxin production (Antunes et al. 2011). Moreover, CcpA directly targets toxin genes and their regulators (tcdR and tcdC) (Figure 1) (Antunes et al. 2012). Additionally, phosphorylated HPr was found not necessary for repression of the toxin genes (Antunes et al. 2011). CcpA plays a key role as it regulates many other regulators, like CodY and Rex, which mediate production of butyric acid, a known inducer of toxin production (Antunes et al. 2012).

In a recent study, trehalose, a disaccharide, was shown to increase virulence of CD RT027 in a mouse model of CDI with increased production of TcdB (Collins et al. 2018). CD has a mechanism to utilise low concentrations of trehalose. It was found that the CD genome has a gene treA that encodes a phospho-trehalase enzyme that is regulated by a transcriptional repressor encoded by gene treR. TreR binds to the promoter region of treA in the absence of trehalose. TreA metabolises trehalose-6-phosphate (T6P) into glucose and glucose-6-phosphate. Sufficient T6P build-up enables binding of T6P to TreR, causing de-repression of treA (Collins et al. 2018, 2019). It was shown that observed disease severity was due to a conserved single nucleotide polymorphism (SNP) within the treR gene of all RT027 strains that caused enhanced expression of the treA gene. Inactivation of the treA gene showed reduced levels of the TcdB as compared to the wildtype strain (Collins et al. 2018).

Repression of tcdA and tcdB by CcpA and CodY is further aided by the mutation frequency decline (Mfd), a transcription-repair coupling factor (TRCF) in bacteria that is involved in DNA repair mechanisms and transcription elongation processes that removes stalled RNAP at DNA lesions or at transcriptional roadblocks created by these repressor proteins. Mfd helps the stalled RNAP dissociate at transcriptional repression sites blocked by the binding of CcpA and CodY regulators (Robleto et al. 2012; Willing et al. 2015). An mfd mutant in CD630Δerm showed enhanced levels of TcdA and TcdB, compared to the wild-type, suggesting the role of Mfd in enhancing the efficiency of toxin repression (Figure 2) (Willing et al. 2015).

Regulators of sporulation are linked to tcdA and tcdB expression

Regulation of tcdA and tcdB expression is tightly linked to other cellular processes such as sporulation. Spo0A functions as a critical regulator for sporulation by initiating the cascade of sporulation-specific RNA polymerase sigma factors expression and activation, especially sigE, sigF, sigG, and sigK. These sigma factors activate compartment-specific transcriptional regulation during B. subtilis sporulation and are also conserved in Clostridium species. SigE and SigK are mother cell-specific, and SigF and SigG are specific to the developing forespore (Fimlaid et al. 2013).

Sporulation in CD begins with phosphorylation of Spo0A, whose expression is dependent on the alternative SigH. Five orphan histidine kinases (CD1492, CD2492, CD1579, CD1352, and CD1949) have been identified in CD genomes (Underwood et al. 2009; Zhu et al. 2018). Of these, CD1492 is a negative regulator of sporulation and CD1579 can phosphorylate Spo0A in vitro, and inactivation of the CD2492 gene in the CD630Δerm strain led to a decrease in sporulation efficiency while function of CD1352 (CprK) in sporulation is unlikely as it encodes for a lantibiotic sensor histidine Kinase. Spontaneous mutations in CprK at S230Y and W235C resulted in nisin lantibiotic resistance and were associated with increased expression of cprABC operon (CD1349 to CD1351). (Underwood et al. 2009; McBride and Sonenshein 2011; Suárez et al. 2013; Childress et al. 2016; Edwards et al. 2021). A recent study showed that the regulation of Spo0A by a protein, CsiA (CD25890) negatively regulated sporulation by affecting the expression of the spo0A gene. A csiA mutant showed increased sporulation with enhanced levels of Spo0A expression. Further, the mutant also showed enhanced expression of the spoIIG operon (Martins et al. 2020). The spo0A mutant produced more TcdA and TcdB, suggesting that Spo0A negatively regulates toxin production (Figure 2) (Deakin et al. 2012). Similar observations were also made in different CD strains, such as M7404 and R20291 of ribotype 027 and JGS6133 of ribotype 078 (Mackin et al. 2013).

A novel transcriptional factor, regulator of sporulation and toxin (RstA) that positively regulates sporulation has been shown to repress CD toxin production and motility. CD RstA promotes sporulation and represses toxin production. In CD, rstA (CD3668) and CD2123 also show similarity with RNPP (Rap/NprR/PlcR/PrgX) family of proteins. RNPP family proteins form quorum-sensing systems in Gram-positive bacteria. In CD630Δerm, an rstA mutant showed reduced sporulation capacity, increased transcription of the tcdA, tcdB, tcdR and sigD genes and was more virulent in a hamster infection model (Edwards et al. 2016). Recently, the same group showed that RstA directly regulates the toxins by binding to the promoters of tcdA and tcdB (Figure 2) and further demonstrated that RstA also binds to the promoters of sigD operon (comprising sigD and flgB gene) and tcdR. Additionally, RstA was shown to mediate toxin repression independent of SigD regulation. This study revealed that repressive activity of the RstA was mediated by its helix-turn-helix DNA binding domain (Edwards et al. 2019). A decrease in the transcription of CD1492, CD1579 and CD2492 kinases that impact sporulation was observed in the rstA mutant. In addition, SigE transcription decreased in the rstA mutant, demonstrating the critical role of RstA as a regulator of CD sporulation. Recently, a group reported a strain-dependent sporulation effect of RstA, where RstA affected sporulation and toxin production more robustly in CD R20291 than in CD 630Δerm (Edwards et al. 2016; Edwards et al. 2020).

In CD, there are two transcriptional regulators in the sin locus (Sporulation inhibitory) that encodes two proteins SinR and SinR’ (renamed as SinI). SinR positively influences toxin production and sporulation along with motility, whereas SinI negatively regulates toxin production, sporulation and motility (Figure 2). Elevated levels of TcdA and TcdB, spore production and motility were observed in sinI’ mutants in the R20291 strain (Girinathan et al. 2018). A recent study complemented the sinI’ mutant of R20291 either with the HTH or multimerization domain (MD) and showed that SinI regulates sporulation, toxin production and motility by interacting through its MD domain to inhibit SinR (20). Although these transcriptional regulators contain the HTH DNA binding domain, how exactly they regulate the expression of toxin genes has not been demonstrated yet. Further studies showed that CodY binds to the upstream DNA sequences of the sin locus to transcriptionally repress sinRI expression. The same group recently showed that sinRI operon is also repressed by Spo0A by its binding to the promoter region of the sin locus (Dhungel and Govind 2020).

Additional evidence that connects toxin production and spore formation has been revealed by studying the tcdR mutant in R20291 (Girinathan et al. 2017). The R20291 tcdR mutant showed weakly assembled exosporium and had ~50% decrease in SigE- and SigK-dependent gene transcription. RNAseq analysis revealed reduced transcripts levels of spoIIR and spoIIID in the R20291 tcdR mutant. SpoIIR is critical for SigE activation, whereas SpoIIID is a transcriptional regulator that regulates the transcription of the sigK gene (Saujet et al. 2013; Girinathan et al. 2017).

CD pathogenesis also seems to be epigenetically regulated by DNA methylation of the 5′-CAAAAA-3′ motif by CamA, an adenine methyltransferase that methylates DNA at N6-methyladenine (6 mA) containing the methylation motif CAAAAA (Oliveira et al. 2020). DNA methylation system is an epigenetic factor that regulates several biological processes via turning on/off regulatory gene expression. Inactivation of camA impacted not only sporulation but also biofilm formation. In addition, the mutant affected cell length and host colonisation. About 50% of the 120 genes that are expressed during sporulation were downregulated in the camA mutant (e.g., SpoIIQ and SpoIVA of SigF & SigE regulon genes) (Oliveira et al. 2020). Analysis of the CD genome revealed that the non-methylated 5′-CAAAAA-3′ motifs on/off switch overlapped with TFs binding sites and the transcription start sites (e.g., CodY and XylR, a transcriptional regulator of xylose utilisation pathway) (Oliveira et al. 2020). Although no direct link has been established of CamA role in the regulation of toxin synthesis, however further research in this field may reveal a potential indirect effect between toxin gene expression and DNA methylation by CamA via regulation of different transcription factors.

Quorum sensing (QS) regulates tcdA and tcdB expression

Quorum sensing (QS) is a mechanism by which bacterial population coordinately control gene expression by communicating with each other. In QS, bacteria release a diffusible QS signal molecule (autoinducer) which act as auto-inducers (AI) that is sensed by the bacterial population as a whole, and coordinate the production of virulence genes that respond to the surrounding environment. QS AI signalling in CD630 involves a LuxS homologue LuxS^CD (15 kDa), which converts SAH (S-adenosyl homocysteine) to the auto-inducer AI-2, leading to increased expression in the reporter system (Carter et al. 2005). LuxS^CD is regulated by the expression of two genes, rolA and rolB that are situated upstream of LuxS^CD gene, which act as a transcriptional regulator and a sensor kinase of LuxS^CD, respectively (Figure 3(A)). It was demonstrated in a toxigenic strain CD CCUG19126 that AI-2 increased the transcription of tcdA, tcdB and tcdE genes during mid-log-phase growth (Lee and Song 2005). A recent study showed that LuxS^CD plays a role in biofilm formation as well suggesting a link between QS, biofilm formation and toxin synthesis (Slater et al. 2019).

Recently a TF, accessory gene regulator A (AgrA) encoded by a gene in the agr operon mediated QS (Darkoh et al. 2015). The agr is a quorum sensing operon that is best-studied in Staphylococcus aureus and contains agrA, agrB, agrC, and agrD. agrD gene encodes the pre autoinducing peptide, agrB encodes a product that cleaves the preform resulting in the formation of the active autoinducer peptide (AIP). Interaction of AIP with the AgrC histidine kinase activates the AgrA response regulator by phosphorylation, which then regulates the transcription of many target genes (Tan et al. 2018).

In CD R20291, the importance of QS was demonstrated using purified thiolactone which acted as an autoinducer and resulted in the increased transcription of tcdA and tcdB genes (Darkoh et al. 2015). Furthermore, an agrA mutant produced the thiolactone (TI) signal, but was defective in toxin production. Complementation of this agr mutant with the functional wild-type copy of the agrA gene restored toxin production (Darkoh et al. 2015). The hypothesis that the thiolactone signal is an agrD gene product, alignment model of this thiolactone of CDs showed identity in amino acid sequence and have a conserved Cys-28 residue which matches the sequence of agrD gene. Like AIPs of the Staphylococcus aureus, these conserved Cys-28 residues form a thioester bond with amino acid residue 32 to create a 5-amino-acid cyclic peptide with a tail. Furthermore, hydroxylamine treatment caused the loss of the purified C. difficile TI signal activity due to disruption the thioester bond. These data further add support for the cysteine-containing AgrD1 prepeptide as the source of the TI signal. Therefore, it can be concluded that the agrD gene encodes the AIP signal in CD (Darkoh et al. 2015). Bioinformatics analysis of the Agr system in CD showed three agr loci, agr1, agr2 and agr3 (Figure 3(B)). The agr1 locus encodes an incomplete set of QS genes (agrB1 and agrD1) and is involved only in quorum signal generation. The agr2 locus encodes the full complement of genes (agrB2D2 and agrC2A2) of the QS pathway (Darkoh et al. 2015). Interestingly, the agr1 locus is present only in non-hypervirulent strains, while all sequenced hypervirulent strains encode both of the loci as reported by the comparative genomic and phenotypic analysis of in a recent epidemic and hypervirulent 027 (R20291) (Stabler et al. 2009). Additionally, a third Agr system has been found in CD ribotype 078, which is present in the CD bacteriophage phiCDHM1. However, the Agr3 system lacks the AgrA response regulator (Hargreaves et al. 2014). Deletion of agrB1 in CD630 showed increased expression of both the toxin genes and the tcdR gene that correlated with the accumulation of the QS peptide AgrD1. Furthermore, the mutant showed reduced sporulation when either agrB1 and agrD1 or both were inactivated. However, motility was affected when both the genes were inactivated (Ahmed et al. 2020). Transcriptomic analysis of the agrA mutant in R20291 revealed about 75 differentially expressed genes involved in flagellar assembly function, c-di-GMP synthesis and tcdA expression. Moreover, the agrA mutant in R20291 showed poor flagellation and colonisation in a murine model of infection (Martin et al. 2013). Interestingly, CD has two putative response regulators, RgaR and RgbR that have homology to the toxin regulator VirR in C. perfringens. Transcriptomic analyses revealed that RgaR positively regulates agrBD locus genes in CD 630 and purified RgaR protein was shown to bind to two of its promoter elements (O’Connor et al. 2006). Further investigation of the role of RgaR and RgbR might reveal their pathogenic potential in CD. These findings provide direct evidence of the role of different QS systems based regulation of C. difficile toxin synthesis and paves way for the rapid development of QS based therapies by targeting key signalling molecules of this pathway to combat CDI.

Regulators of toxin production are linked to biofilm formation

Biofilm plays a central role in the pathogenesis of CDI by dramatically contributing to drug resistance. In vitro culturing of CD biofilms on polycarbonate filter surfaces demonstrated that the mature biofilm contained vegetative cells, spores and dead cells that accumulated TcdA and TcdB, and were highly resistant to the antibiotic metronidazole (Figure 3(A)) (Semenyuk et al. 2014). Spo0A, a master regulator of sporulation and a negative regulator of TcdA and TcdB synthesis, plays a key role in matrix formation, persistent infection and sustained biofilm formation and dispersion under stress conditions (Deakin et al. 2012; Dawson et al. 2012; Maikova et al. 2018). Inactivation of spo0A reduced biofilm formation in R20291 and complementation with the functional wild-type copy of the spo0A gene restored biofilm formation activity (Dawson et al. 2012).

Biofilm formation in CD is also regulated by stressors found in the environment (e.g., the presence of sub-inhibitory concentrations of antibiotics, such as metronidazole and vancomycin, or the microbially-generated secondary bile salt, deoxycholate) (Vuotto et al. 2016; Dubois et al. 2019). The transcriptional regulator involved in stress induction is the SOS response regulator, LexA, a transcriptional repressor whose activity is in turn repressed by an autocleavage event in the presence of the RecA/ssDNA protein complex (Figure 3(A)). Bioinformatics analysis showed that LexA in CD regulates at least 29 loci and inactivation of lexA in R20291 resulted in enhanced biofilm formation with decreased sporulation and motility (Walter et al. 2015). Furthermore, in the presence of sub-inhibitory concentrations of levofloxacin, the lexA mutant showed enhanced levels of TcdA, but levels of TcdB remained unchanged. Regulation of tcdA by LexA is mediated by binding to the tcdA promoter at the LexA binding motif (Figure 3(A)) (Walter et al. 2015). How exactly Biofilm and antibiotic resistance in CD is mediated is currently unknown but there may be other TF factors that may play important role in antibiotic resistance and biofilm formation. Discovery of such TF will help find novel therapeutics against the antibiotic resistant CD (Table 3).

Expression of tcdA and tcdB is co-regulated with motility and biofilm formation by c-di-GMP

The secondary messenger signalling molecule c-di-GMP plays an important role in the physiology of CD (McKee et al. 2013). Intracellular fluctuations in the levels of c-di-GMP greatly influence toxin genes, motility and biofilm formation. Enhanced levels of c-di-GMP represses the expression of tcdA, tcdB and tcdR and promotes biofilm formation, while low levels of c-di-GMP influence motility (McKee et al. 2013). Furthermore, c-di-GMP levels influence flagellar synthesis and toxin production. Higher levels of c-di-GMP also repressed the transcription of the sigD gene whose expression in turn positively regulates the transcription of tcdA and tcdB through recognition of the tcdR promoter by SigD directed RNAP (El Meouche et al. 2013).

c-di-GMP is a circular RNA dinucleotide that is formed by two molecules of guanosine-5′-triphosphate by diguanylate cyclases (DGCs) and is degraded by phosphodiesterases (PDEs), which are triggered by the environmental stimuli. The changing concentrations of c-di-GMP are sensed by riboswitches, present upstream of the target genes. (Sudarsan et al. 2008; Soutourina et al. 2013). Recently functionality of 11 c-di-GMP riboswitches in C. difficile were demonstrated which functions as “off” or ON switches in response to c-di-GMP (McKee et al. 2018). Addtionally, it is also possible that additional unidentified c-di-GMP protein effectors mediate the response to intracellular c-di-GMP levels.

In strain CD630, 37 putative genes encoding either DGCs or PDEs enzymes are involved in synthesis and degradation of c-di-GMP, respectively (Bordeleau et al. 2011). Expression of CD1420 (dccA) of strain CD630 led to an increase in the intracellular concentration of c-di-GMP with increased biofilm formation (Bordeleau et al. 2011) (Figure 3(A)). In addition, overexpression of dccA caused increased expression of type IV pili operon genes with enhanced phenotypic expression of pili on cell surfaces which are important in biofilm formation (Bordeleau et al. 2015). On the other hand, degradation of the c-di-GMP by phospho-di-esterase pdcA (CD1515) suppressed biofilm formation (Purcell et al. 2017). Furthermore, CodY controlled the expression of pdcA under high and low GTP concentrations. In a recent study, CD2214/CD2215 encoding SinR/SinR’ regulator in strain CD630Δerm played a positive role in regulation of dccA and pilA1 during biofilm formation. Furthermore, overexpression of dccA in CD630Δerm enhanced biofilm formation (Poquet et al. 2018).

In CD, a pilA mutant had reduced biofilm formation compared to the wildtype strain (Maldarelli et al. 2016; Purcell et al. 2016). At least 11 pilin genes in CD encode type IV pili (T4P) composed of different pilin units (Maldarelli et al. 2014). Expression of pilin genes have been linked to the secondary messenger signalling molecule c-di-GMP that in turn binds to c-di-GMP riboswitches. The riboswitch Cdi2_4 that is upstream of the pilA1 operon, turns the pilA genes on by relieving the Rho-independent transcription terminator upon binding with c-di-GMP (Bordeleau et al. 2015).

Regulation of flagellar genes synergize with toxin production

In CD, flagella are important for motility and colonisation, and their synthesis is controlled by four flagellar operons that are linked to toxin synthesis. Functional inactivation of several flagellar genes, such as fliC, CD0240, fliF, fliG, fliM, and flhB-fliR showed significant changes in transcriptional expression of four PaLoc genes (tcdR, tcdA, tcdB, tcdE). The fliC mutant had enhanced levels of toxin production. In contrast, mutants of CD0240, fliF, fliG, fliM, and flhB-fliR showed reduced transcription of the four tcd genes with significant reduction in toxin production (Aubry et al. 2012). Another study showed a ten-fold decrease in the expression of the tcdA gene in a flgE mutant (Baban et al. 2013). SigD, which positively regulates toxin expression via tcdR, is also expressed from the flgB operon, which is regulated by a c-di-GMP binding riboswitch Cdi1_3, located 496 bp upstream of the flgB start codon (Figure 1) (Purcell et al. 2012; El Meouche et al. 2013; Soutourina et al. 2013). This operon is repressed upon binding by high concentrations of c-di-GMP at the riboswitch. Therefore, during biofilm formation, flagellar synthesis is halted along with toxin synthesis as sigD expression is linked to this operon. (Faulds-Pain et al. 2014; Valiente et al. 2016).

Pathogenic CD strains also exhibit phase variation that causes phenotypic heterogeneity in bacterial populations under sudden environmental stress conditions, which help survival of the bacteria. These phenotypic changes are expressed in terms of changes in the colony morphology and surface structures such as flagella and toxin production. Genetic switches mediate phase variation and about eight genetic switches have been predicted in CD. One of the phase variation switches controls the flgB operon (Figure 1) (Anjuwon-Foster et al. 2018). This gene expression switch occurs at promoter elements located upstream of flgB operon in a 154-bp invertible sequence flanked by 21-bp inverted repeats. Orientation of this genetic switch sets the phase variation in either the ON or OFF mode, which is facilitated by a tyrosine recombinase RecV that catalyses the inversion of this switch in both directions. In the ON mode, the promoter elements are in an orientation that allows expression of flagella and toxin, whereas the OFF mode represses the flagellar and toxin gene expression (Anjuwon-Foster and Tamayo 2017). Further work demonstrated that phase variation may be strain-dependent where some strains have the ability to switch the phases while many others remain in a phase off position due to a mutation in the RecV enzyme (Anjuwon-Foster et al. 2018). Recently, a transcription terminator Rho has been identified which can differentiate the orientation of the genetic switch and preferentially terminate the transcription of flgBOFF mRNA within the 5′ leader sequence thus contributing to the phase variation heterogeneity. Besides phase variation, a bimodel expression of toxin synthesis can also be controlled by bistability. Recent studies in this direction have described bistability of TcdR promoter, where σ^TcdR governs the decision between toxin-ON and OFF status (Ransom et al. 2018).

CD flagellar expression is also controlled by several other regulators such as SinR, RstA, Lrp, and AgrA that influence toxin synthesis. SinR activates flagellar gene expression while RstA negatively regulates flagellar gene expression (Edwards et al. 2016; Girinathan et al. 2018; Ciftci et al. 2019; Edwards et al. 2019). The QS agrA locus also influences flagellar expressions, i.e., an agrA mutant exhibited poor flagellation in CD (Martin et al. 2013). The Lrp response regulator negatively regulated sporulation and positively regulated motility and the TcdA and TcdB (Chen et al. 2019).

Concluding remarks and critical unanswered questions

The complexed transcriptional regulation of the tcdA and tcdB genes under nutritional stress conditions and physiological stages such as vegetative growth phase, QS, biofilm formation and sporulation suggest the adaptive strategies employed by the CD pathogen to survive and thrive in the intestine and cause diseases. The CD transcriptional machinery that regulates toxin production is tightly linked to sporulation, quorum sensing, biofilm formation and flagellar motility which is regulated by various TFs (Table 2). Although recent studies have shed lights on linking regulation of toxins with various physiological stages of CD, detailed mechanism of regulation need to be elucidated. Thus, further studies are necessary to understand how coordination of different regulatory networks are fine-tuned with the regulation of tcdA and tcdB expression, and some outstanding questions are summarised in Table 3. Specifically, how the nitrogen regulated response (Ntr) impact toxin expression and virulence in C. difficile. Further, Many flagellar switches identified in CD needs to be characterised. A thorough knowledge of the mechanistic details on the regulation of TFs and their specific target genes that regulate key steps involved in CD pathology such as biofilm formation, antibiotic resistance and toxin production will potentially speedup the development of novel therapeutics for specific drug targets and vaccine development.