Department of Microbiology and Immunology, Emory University School of Medicine, Atlanta, GA, USA
Correspondence: Shonna M. McBride, Department of Microbiology and Immunology, Emory University School of Medicine, 1510 Clifton Road, Atlanta, GA 30322, USA. Tel.: +1 (404) 727 6192; fax: +1 (404) 727 8250; e-mail: firstname.lastname@example.org
The formation of dormant endospores is a complex morphological process that permits long-term survival in inhospitable environments for many Gram-positive bacteria. Sporulation for the anaerobic gastrointestinal pathogen Clostridium difficile is necessary for survival outside of the gastrointestinal tract of its host. While the developmental stages of spore formation are largely conserved among endospore-forming bacteria, the genus Clostridium appears to be missing a number of conserved regulators required for efficient sporulation in other spore-forming bacteria. Several recent studies have discovered novel mechanisms and distinct regulatory pathways that control the initiation of sporulation and early-sporulation-specific gene expression. These differences in regulating the decision to undergo sporulation reflects the unique ecological niche and environmental conditions that C. difficile inhabits and encounters within the mammalian host.
Clostridium difficile is a significant gastrointestinal pathogen that infects humans and other animals and is the primary cause of antibiotic associated diarrhea (AAD). Clostridium difficile infection (CDI) is typically precipitated by the use of antibiotics, which disrupts the native gut microbiota, providing a niche for C. difficile overgrowth and toxin production. CDI and AAD have recently become an increasing health problem in hospital and nursing home settings (O'Brien et al., 2007; Bouza, 2012; Dubberke & Olsen, 2012; Murphy et al., 2012). In addition, C. difficile has recently been recognized as an emerging pathogen and urgent public health threat by the CDC (CDC, 2013). Although C. difficile is an important pathogen, its growth and life cycle within the host remain poorly understood.
Clostridium difficile is a strict anaerobe that forms metabolically inactive spores within the gastrointestinal tract of mammals (Fig. 1). These dormant spores are naturally resistant to a variety of environmental and chemical insults, including exposure to oxygen, disinfectants, desiccation, and extreme temperatures (Lawley et al., 2009). The ability for C. difficile to form spores is critical for survival outside of the host and for transmission from host to host (Deakin et al., 2012). Spore formation is a factor in C. difficile resistance to traditional antibiotic therapies and also contributes to recurrent infection (Deakin et al., 2012). Further, sporulation-specific gene expression is upregulated early in the mouse model of CDI (Janoir et al., 2013). Despite the importance of sporulation in the pathogenesis of C. difficile, the triggers and molecular mechanisms that govern the initiation of spore formation are not well understood.
Our limited understanding of C. difficile sporulation is primarily based on comparisons with other spore-forming bacteria. In the model organism Bacillus subtilis, sporulation is a complex morphological event that is tightly controlled by multiple regulators, checkpoints, and feedback loops. The sporulation process is demarcated into several stages based on physiological landmarks. The physiological changes within the cell are accomplished through compartmentalized transcription programs within the mother cell and the developing endospore compartments. These transcription programs are orchestrated via the four sporulation-specific sigma factors: σF, σE, σG, and σK. The C. difficile genome encodes the four conserved sporulation-specific sigma factors and readily forms heat-resistant spores within mammalian intestinal tracts (Fimlaid et al., 2013; Pereira et al., 2013; Saujet et al., 2013). Although the sporulation-specific sigma factors are highly conserved in C. difficile, recent global studies revealed that divergent regulatory mechanisms mediate the timing of expression and activation of these sigma factors compared with other Clostridium and Bacillus species (Fimlaid et al., 2013; Pereira et al., 2013; Saujet et al., 2013).
In B. subtilis, spore formation begins as the cells transition from exponential to stationary growth (Stage 0). At Stage 0, gene expression shifts to support spore formation, rather than vegetative growth. Stage 0 is defined by the post-translational activation of the transcription factor, Spo0A, which upregulates sporulation-specific gene expression and serves as the master regulator of sporulation. Because spore formation is an energy-intensive process and is irreversible at specific points in spore development, the decision to initiate sporulation involves the integration of multiple environmental signals to determine whether conditions are unfavorable to support further vegetative cell growth. In B. subtilis and other Bacillus species, Spo0A activity is tightly regulated through a phosphorylation-mediated signal transduction pathway in response to nutrient availability, cell density, and other signals (Sonenshein, 2000). Importantly, the C. difficile genome lacks the B. subtilis phosphorelay orthologs, and instead, Spo0A activity appears to be controlled through at least three orphan histidine kinases and potentially other unidentified factors (Paredes et al., 2005; Underwood et al., 2009).
Genomic analyses have revealed that C. difficile does not possess many of the conserved early-stage regulatory components required for Spo0A activation and efficient sporulation in other spore-forming bacteria (Paredes et al., 2005; Galperin et al., 2012). In the last few years, researchers have begun to elucidate the functions of C. difficile early-sporulation orthologs in controlling initiation through the regulation of spo0A transcription and Spo0A phosphorylation. Understanding the genetic pathways and environmental conditions that lead to Spo0A activation in C. difficile is the key for designing targeted therapeutics to inhibit spore formation and thus, limiting the spread of the disease.
Spo0A: the master regulator of sporulation
Several studies characterizing C. difficile sporulation have determined that some of the most conserved global regulators of sporulation play similar roles as found in other spore formers (Saujet et al., 2011; Antunes et al., 2012; Deakin et al., 2012). In particular, studies of Spo0A-dependent regulation have revealed that while the C. difficile Spo0A regulon has significant overlap with B. subtilis, there are differences in early-sporulation gene regulation and in the role Spo0A may play in sigma-factor activation and late-stage gene expression (Fimlaid et al., 2013; Pereira et al., 2013; Saujet et al., 2013; Pettit et al., 2014). All sequenced C. difficile genomes contain the highly conserved Spo0A transcription factor and possess both the N-terminal phosphorylation and dimerization domain and the C-terminal DNA-binding domain (Table 1). As in Bacillus species, C. difficile spo0A mutants are asporogenous and fail to activate stationary-phase and early-sporulation gene transcription (Heap et al., 2007; Deakin et al., 2012; Rosenbusch et al., 2012; Fimlaid et al., 2013; Pettit et al., 2014). The C. difficile Spo0A protein recognizes a consensus sequence similar to the 0A box defined in B. subtilis and directly binds with high affinity and specificity to the promoter regions of itself (spo0A), sigH (σH) and the early-sporulation genes that control σF and σE activation (spoIIAA, spoIIE, and spoIIGA; Rosenbusch et al., 2012). Global transcriptional analyses have further defined the C. difficile Spo0A regulon to include additional conserved early-sporulation genes, such as those encoding σF and σE, and a negative regulator of Spo0A activity, sinR (Pettit et al., 2014; Fig. 2). The Spo0A regulon also contains genes that are unique to C. difficile, such as CD1492 and CD1579, which encode the orphan histidine kinases that directly or indirectly influence Spo0A phosphorylation (see below; Underwood et al., 2009; Pettit et al., 2014; Fig. 2). Unlike B. subtilis, the conditions that facilitate synchronous in vitro sporulation of C. difficile have not been identified, making it difficult to detect small changes in gene transcription during spore development (Fimlaid et al., 2013; Putnam et al., 2013; Saujet et al., 2013). As a result, there are likely many more transcripts in the Spo0A regulon that remain to be identified.
Table 1. Known and predicted genes involved in early sporulation in Clostridium difficile
Spo0A also regulates physiological processes other than sporulation, including biofilm formation (Dawson et al., 2012; Dapa & Unnikrishnan, 2013), motility (Pettit et al., 2014), carbon metabolism (e.g. butyrate biosynthesis; Pettit et al., 2014) and, in some cases, toxin A (TcdA) and toxin B (TcdB) production (Deakin et al., 2012; Mackin et al., 2013; Pettit et al., 2014). Interestingly, Spo0A can repress toxin production in C. difficile, but this regulation appears to occur primarily in the ribotype 027 epidemic strains and not in other evolutionarily divergent strains, such as ribotype 012 (630Δerm) or 078 (JGS6133; Deakin et al., 2012; Rosenbusch et al., 2012; Mackin et al., 2013). Conversely, two studies report Spo0A-dependent regulation of tcdA expression in 630Δerm (Underwood et al., 2009; Pettit et al., 2014), but this effect is likely indirect as no Spo0A consensus sequences are identified upstream of tcdA (Rosenbusch et al., 2012). The inconsistencies observed in Spo0A-mediated regulation of toxin production can be partly attributed to differences in the growth medium used and the experimental conditions tested. These differences underscore the need for standard assays and growth conditions for the study of C. difficile sporulation and highlight the need for complementation studies in this research.
Global regulators of stationary phase control expression of early-sporulation-specific genes
Along with Spo0A, the transition-phase sigma factor, σH, shares responsibility for upregulating expression of sporulation-related genes as well as mediating the transcriptional changes that occur during the switch from exponential-phase to stationary-phase growth. Clostridium difficile SigH recognizes a similar consensus sequence to that in B. subtilis (Saujet et al., 2011). SigH positively regulates transcription of some early-sporulation genes in C. difficile, including spo0A, a putative Spo0A histidine kinase, the putative sinRI operon as well as spo0J and soj, whose gene products are likely involved in chromosomal segregation during asymmetric division (Saujet et al., 2011). These data suggest that SigH positively contributes to sporulation initiation through multiple pathways by directly inducing spo0A gene expression and positively influencing Spo0A activity through increased transcription of the Spo0A-associated kinase, CD2492. In addition, SigH has a negative effect on expression of the app operon, which encodes a predicted oligopeptide permease (see below; Pereira et al., 2013). Finally, the reciprocal control between SigH and Spo0A creates a positive feed-forward loop, with SigH activating spo0A transcription and Spo0A upregulating sigH gene expression. This regulatory circuitry reinforces the global transcriptional changes necessary to initiate sporulation.
There are at least two additional conserved global regulators that influence sporulation in Bacillus and Clostridium species: CcpA and CodY (Table 1). Both of these transcriptional regulators share a number of regulatory targets in C. difficile, including the regulation of toxin synthesis (Dineen et al., 2007, 2010; Antunes et al., 2012). Catabolite control protein A (CcpA) is a DNA-binding transcriptional regulator that governs the global response to carbon availability in low-G+C Gram-positive organisms. CcpA controls expression of genes involved in sugar uptake, fermentation, and amino acid metabolism in the presence of preferred carbon sources such as glucose and other phosphotransferase system (PTS) sugars (Antunes et al., 2011, 2012). In B. subtilis, CcpA activity is linked to carbon availability through the direct interaction with a corepressor, HPr-Ser-P, the phosphorylated form of HPr (Fujita et al., 1995). HPr phosphorylation occurs in the presence of fructose-1,6-biphosphate (FBP) which triggers HPr serine kinase/phosphatase activity of the HprK/P protein (Deutscher & Saier, 1983). In contrast to B. subtilis, the DNA-binding affinity of the C. difficile CcpA protein is enhanced by FBP in vitro, but not by the presence of HPr-Ser-P (Antunes et al., 2011). Although the molecular mechanism of this interaction is not yet understood, CcpA function in C. difficile appears to be regulated by the same carbon metabolite as in B. subtilis.
In the absence of glucose, CcpA is required for efficient sporulation in C. perfringens (Varga et al., 2004), but CcpA downregulates sporulation in C. difficile (Antunes et al., 2012), demonstrating that unique regulatory processes can control sporulation initiation in different Clostridium species. In C. difficile, CcpA binds to conserved recognition sequences known as catabolite-responsive elements (creCD) and directly represses transcription of many stationary-phase transcripts, including spo0A and the opp operon, which encodes a predicted oligopeptide transporter system (Antunes et al., 2012). CcpA also represses expression of the putative Spo0A histidine kinase CD1579 and the sinR transcriptional regulator, but these interactions are likely indirect as no creCD motifs are apparent in the respective promoter regions (Antunes et al., 2012; Fig. 2). Transcription of genes involved in later stages of sporulation are also repressed by CcpA (Antunes et al., 2012). It is important to note that while CcpA-mediated catabolite repression of sporulation has been observed in Bacilli and Clostridia, glucose-mediated repression of sporulation genes occurs independently of CcpA in B. subtilis, C. difficile, and C. perfringens (Moreno et al., 2001; Varga et al., 2004; Antunes et al., 2012). In B. subtilis, the primary mediator of this physiological response appears to be the transcriptional regulator ScoC, which represses sporulation gene expression in a glucose-dependent manner (see below; Shafikhani et al., 2003).
CodY is another conserved global regulator that responds to nutrient availability in C. difficile and other Gram-positive bacteria (Sonenshein, 2005; Dineen et al., 2007, 2010). CodY, along with the cofactors GTP and branched-chain amino acids (BCAAs), represses gene expression in high-nutrient conditions (Dineen et al., 2007, 2010). When nutrients become limiting, such as during the entry into stationary phase, the intracellular concentrations of GTP and BCAAs decreases, thereby relieving CodY-mediated repression (Sonenshein, 2007). A global analysis that combined both in vivo and in vitro methods to define the CodY regulon revealed that CodY directly binds to and regulates expression of the CD2492 histidine kinase, a Rap phosphatase ortholog (CD2123) and the opp operon (Dineen et al., 2010; Fig. 2). While it is clear that both CcpA and CodY affect sporulation of C. difficile by regulating expression of sporulation-related genes, the intricate regulatory networks that connect CcpA and CodY with the initiation of sporulation includes hundreds of directly regulated transcripts and many more indirect targets.
Post-translational activation of Spo0A
In the genus Bacillus, the regulatory pathway controlling Spo0A phosphorylation is known as the sporulation phosphorelay. The phosphorelay is a variant of a two-component signal transduction system (TCS) comprised of five sensor kinases (KinA-E), which respond to specific ligands, allowing multiple signals to converge and influence Spo0A activity (Perego et al., 1989; Burbulys et al., 1991; Kobayashi et al., 1995; LeDeaux & Grossman, 1995; LeDeaux et al., 1995; Jiang et al., 2000b). KinA-E directly phosphorylate Spo0F which subsequently transfers the phosphate to Spo0A through an additional phosphotransferase, Spo0B (Burbulys et al., 1991). Despite the conservation of Spo0A and the similarities between Spo0A targets in Bacillus and Clostridium species, the components of the Bacillus phosphorelay do not appear to be encoded within C. difficile or the other clostridial genomes. Based on the lack of an apparent phosphorelay, the favored hypothesis is that the sporulation initiation pathway in Clostridium species functions more similarly to a traditional TCS in which sporulation-associated sensor kinases recognize specific internal or environmental signals and directly phosphorylate Spo0A (Worner et al., 2006; Steiner et al., 2011). Supporting this hypothesis, three of the five orphan histidine kinases (CD1492, CD1579 and CD2492) present in the C. difficile genome share some sequence identity with KinA-E (Underwood et al., 2009; Table 1). Based on in vitro biochemical assays and in vivo sporulation studies, the CD1579 and CD2492 kinases are anticipated to directly phosphorylate Spo0A (Underwood et al., 2009). CD1579 is predicted to be a cytosolic protein and contains a degenerate heme-oxygen-sensing PAS domain ((Underwood et al., 2009; A.N. Edwards, E. Weinert & S.M. McBride, pers. commun.). The CD1579 histidine kinase was shown to directly phosphorylate Spo0A in vitro, while a CD2492 mutant exhibits a threefold decrease in sporulation in vivo (Underwood et al., 2009). These results strongly suggest that both of these histidine kinases influence C. difficile sporulation (Fig. 2). CD1492 and CD2492 encode conserved autophosphorylation domains and are integral membrane proteins, suggesting that these histidine kinases may autophosphorylate upon contact with specific extracellular signals. However, no characterization of CD1492 has been published. Of the remaining two orphan histidine kinases, CD1352 has been shown to regulate a lantibiotic resistance mechanism (McBride & Sonenshein, 2011; Suarez et al., 2013), and CD1949 has no known function.
Control of sporulation initiation by limiting accumulation of Spo0A~P
In B. subtilis, Spo0A phosphorylation is further controlled through multiple phosphatases and their respective regulators. One mechanism employed by B. subtilis to restrict sporulation initiation is the dephosphorylation of the components within the phosphorelay, which in turn limits the accumulation of phosphorylated Spo0A. The Spo0E, YnzD and YisI aspartyl-phosphate phosphatases act directly on Spo0A (Perego & Hoch, 1991; Ohlsen et al., 1994; Perego, 2001), while the RapA, RapB and RapE phosphatases bind to and dephosphorylate Spo0F (Perego et al., 1994; Jiang et al., 2000a). Additionally, a histidine kinase inhibitor, KipI, directly blocks the catalytic domain of KinA, preventing ATPase activity and subsequent autophosphorylation of KinA (Wang et al., 1997). Adding further complexity to these regulatory pathways, Rap activity is inhibited by specific small quorum-sensing peptides, known as Phr peptides (Magnuson et al., 1994; Lazazzera et al., 1997; Perego, 1997), while KipI antikinase activity is inhibited by KipA (Wang et al., 1997). Clostridium difficile encodes two orthologs to the Rap phosphatases and orthologs of KipI and its antagonist KipA, but no apparent Phr peptide or Spo0E, YnZD or YisI phosphatase orthologs are present (Paredes et al., 2005; Galperin et al., 2012; Table 1). The Rap orthologs in C. difficile have low homology to the B. subtilis Rap proteins. The N-terminal protein-binding and phosphatase domains are not conserved in C. difficile. Rather, the conserved domains present in the C. difficile orthologs are the tetratricopeptide (TPR) repeats which are responsible for direct interaction with the inhibitory Phr peptides (Diaz et al., 2012). Clostridium difficile is also missing an apparent ortholog of sda, which encodes another histidine kinase inhibitor that prevents KinA autophosphorylation in response to defects in DNA replication initiation (Burkholder et al., 2001; Rowland et al., 2004; Paredes et al., 2005). It is not known whether the Rap or Kip orthologs influence spore formation in C. difficile. As their conserved targets within the phosphorelay are absent in C. difficile, the Rap and Kip orthologs may act on the putative histidine kinases or directly on Spo0A.
Bacillus species initiate sporulation in a cell density-dependent manner through the synthesis, export, and uptake of small quorum-sensing peptides known as Phr peptides. These five to six amino acid peptides are recognized and imported by two oligopeptide permeases, Spo0K (Opp) and App (Perego, 1997). Once imported, Phr peptides directly bind to and inhibit the phosphatase activity of the Rap proteins previously mentioned, initiating sporulation (Magnuson et al., 1994; Lazazzera et al., 1997; Perego, 1997). Clostridium difficile encodes orthologs of the Opp and App transporters but is missing orthologs to the Phr peptides (Edwards et al., 2014; Table 1). In contrast to B. subtilis, an opp app mutant in C. difficile increases sporulation-specific gene expression and exhibits a hypersporulation phenotype (Edwards et al., 2014). Increased sporulation appears to occur in response to a decrease in general peptide uptake mediated by these oligopeptide transporters, which corresponds with an overall decrease in CcpA and CodY activity in transporter mutants. Hence, Opp and App likely indirectly influence sporulation initiation by signaling nutrient availability through CodY- and CcpA-mediated gene regulation (Edwards et al., 2014). Because the standard methods of inducing sporulation in other spore formers, such as B. subtilis, do not increase the sporulation rate of C. difficile, it has been unclear whether C. difficile sporulates in response to nutrient starvation (Durre & Hollergschwandner, 2004). The observation that Opp and App function inhibits sporulation in C. difficile is evidence that this organism initiates sporulation in nutrient-limiting conditions.
Auxiliary regulators of sporulation initiation
There are several additional regulatory factors involved in sporulation initiation in B. subtilis that may also play a role in controlling sporulation in C. difficile. ScoC, a negative transcriptional regulator of sporulation in B. subtilis, downregulates both opp and app transcriptions (Koide et al., 1999). Overexpression of a putative scoC ortholog (CD0852; Table 1), divergently transcribed from the opp operon, decreases sporulation efficiency but does not influence gene expression of the opp or app operons in C. difficile, suggesting that CD0852 may negatively influence sporulation through a unique regulatory pathway (Edwards et al., 2014). Another gene, CD3409, was suggested to function as the ScoC ortholog in C. difficile (Pettit et al., 2014); however, this gene encodes for the HPr kinase/phosphorylase mentioned previously. These genes are often confused because in B. subtilis, the hpr gene encodes ScoC, while ptsH and hprK encode HPr and HPrK/P, respectively.
In B. subtilis, the AbrB protein functions as a transition-state regulator and represses stationary-phase and sporulation genes during exponential growth. AbrB can repress its own transcription as well as expression of both sigH (σH) and the Spo0A-phosphatase, spo0E, and acts as an activator of hpr expression (encoding ScoC; Zuber & Losick, 1987; Strauch et al., 1989; Perego & Hoch, 1991). Expression of abrB is controlled by negative autoregulation and is repressed by activated Spo0A upon entry into transition phase (Perego et al., 1988). There are two AbrB paralogs present in B. subtilis, Abh and SpoVT, which function similarly to AbrB as DNA-binding regulators although their regulons do not fully overlap (Bagyan et al., 1996; Dong et al., 2004; Yao & Strauch, 2005). The C. difficile genome encodes two putative abrB orthologs (CD1859A and CD3120) as well as a SpoVT ortholog (CD3499; Table 1). Unfortunately, no experimental evidence for any of the C. difficile AbrB orthologs is available, so the functions of these factors remain a mystery. However, recent work demonstrated that SpoVT functions slightly differently in C. difficile as compared to B. subtilis and is necessary for the formation of mature, heat-resistant spores in C. difficile (Saujet et al., 2013). Similar to B. subtilis, SpoVT does not appear to influence sporulation initiation as a C. difficile spoVT mutant forms phase dark spores (Saujet et al., 2013).
The tetrameric DNA-binding regulator SinR provides another layer of regulatory control in sporulation initiation by directly repressing spo0A gene expression (Mandic-Mulec et al., 1995). SinR repression is disrupted by the production of the antagonist, SinI, which is encoded in the same operon as sinR and forms an inactive heterodimeric complex with SinR (Bai et al., 1993). Regulation of sinIR gene expression in B. subtilis is complex as AbrB, ScoC, and Spo0A all play roles in up- or downregulating transcription (Shafikhani et al., 2002). The sinRI genes are encoded in the C. difficile genome (CD2214 and CD2215; Table 1), and although no characterization of these has been reported, there may be some regulatory circuitry conserved in C. difficile as Spo0A represses sinR gene expression (Pettit et al., 2014).
Finally, fidelity of DNA replication and segregation is a regulatory checkpoint for initiating sporulation in other spore formers. Together, the Soj and Spo0J proteins repress sporulation until chromosomal segregation has occurred. Soj and Spo0J are similar to the ParA/B family of proteins involved in plasmid partitioning and are required for proper chromosomal replication and segregation during vegetative growth, as well as before asymmetric division occurs during sporulation (Ireton et al., 1994). Soj prevents Spo0A-dependent transcription while Spo0J inhibits Soj activity (Ireton et al., 1994; Cervin et al., 1998; Quisel & Grossman, 2000). Clostridium difficile contains both soj (CD3672) and spo0J (CD3671) orthologs along with an additional Spo0J-like ortholog (CD3673) encoded immediately downstream (Table 1). The roles that these proteins play in DNA replication and sporulation in C. difficile have not been confirmed.
The altered roles of conserved regulatory components and the apparent absence of several key factors (e.g. the phosphorelay transferases, Sda, Spo0E, YnzD, YisI and the Phr peptides) involved in early sporulation in the C. difficile genome indicate that this organism has evolved to induce sporulation specifically to survive harsh conditions within its gastrointestinal niche and outside of its host. Because of the specialized ecological niche of C. difficile, there are likely unknown unique regulators and molecular mechanisms controlling Spo0A activation. Compared to other characterized spore formers, C. difficile incorporates different environmental and intracellular signals and utilizes a novel regulatory pathway to mediate Spo0A phosphorylation and subsequent sporulation-specific gene expression. This apparently simplified phosphorylation cascade suggests that C. difficile may possess fewer checkpoints before initiating sporulation compared with Bacillus species, but it is equally plausible that there are alternative pathways and checkpoints that have yet to be discovered. With the recent advances in C. difficile genetics and whole transcriptomic and proteomic sequencing, genetic screens and global studies in this notoriously challenging organism have become feasible and are expected reveal a wealth of information about C. difficile in the coming years.
We give special thanks to Charles Moran, Kathryn Nawrocki, and members of the McBride lab for helpful criticism of this manuscript. This work was supported by the US National Institutes of Health through research Grants DK087763 and DK101870 to S.M.M.