Strain-Dependent RstA Regulation of Clostridioides difficile Toxin Production and Sporulation

Two critical traits of Clostridioides difficile pathogenesis are toxin production, which causes disease symptoms, and spore formation, which permits survival outside the gastrointestinal tract. The multifunctional regulator RstA promotes sporulation and prevents toxin production in the historical strain 630Δerm. Here, we show that RstA exhibits stronger effects on these phenotypes in an epidemic isolate, R20291, and additional strain-specific effects on toxin and rstA expression are evident. Our data demonstrate that sequence-specific differences within the promoter for the toxin regulator TcdR contribute to the regulation of toxin production by RstA and CodY. These sequence differences account for some of the variability in toxin production among isolates and may allow strains to differentially control toxin production in response to a variety of signals.

ABSTRACT

The anaerobic spore former Clostridioides difficile causes significant diarrheal disease in humans and other mammals. Infection begins with the ingestion of dormant spores, which subsequently germinate within the host gastrointestinal tract. There, the vegetative cells proliferate and secrete two exotoxins, TcdA and TcdB, which cause disease symptoms. Although spore formation and toxin production are critical for C. difficile pathogenesis, the regulatory links between these two physiological processes are not well understood and are strain dependent. Previously, we identified a conserved C. difficile regulator, RstA, that promotes sporulation initiation through an unknown mechanism and directly and indirectly represses toxin and motility gene transcription in the historical isolate 630Δerm. To test whether perceived strain-dependent differences in toxin production and sporulation are mediated by RstA, we created an rstA mutant in the epidemic ribotype 027 strain R20291. RstA affected sporulation and toxin gene expression similarly but more robustly in R20291 than in 630Δerm. In contrast, no effect on motility gene expression was observed in R20291. Reporter assays measuring transcriptional regulation of tcdR, the sigma factor gene essential for toxin gene expression, identified sequence-dependent effects influencing repression by RstA and CodY, a global nutritional sensor, in four diverse C. difficile strains. Finally, sequence- and strain-dependent differences were evident in RstA negative autoregulation of rstA transcription. Altogether, our data suggest that strain-dependent differences in RstA regulation contribute to the sporulation and toxin phenotypes observed in R20291. Our data establish RstA as an important regulator of C. difficile virulence traits.

IMPORTANCE Two critical traits of Clostridioides difficile pathogenesis are toxin production, which causes disease symptoms, and spore formation, which permits survival outside the gastrointestinal tract. The multifunctional regulator RstA promotes sporulation and prevents toxin production in the historical strain 630Δerm. Here, we show that RstA exhibits stronger effects on these phenotypes in an epidemic isolate, R20291, and additional strain-specific effects on toxin and rstA expression are evident. Our data demonstrate that sequence-specific differences within the promoter for the toxin regulator TcdR contribute to the regulation of toxin production by RstA and CodY. These sequence differences account for some of the variability in toxin production among isolates and may allow strains to differentially control toxin production in response to a variety of signals.

INTRODUCTION

As the leading cause of antibiotic-associated diarrhea, Clostridioides difficile resides in the mammalian gastrointestinal tract, where disease symptoms are mediated by the production of two large, glucosylating exotoxins, toxin A (TcdA) and toxin B (TcdB) (1). TcdA and TcdB target the Rho and Ras families of small GTPases (2, 3), ultimately disrupting host cell function and triggering apoptotic and/or necrotic cell death (4). TcdA and TcdB are encoded within the 19.6-kb pathogenicity locus (PaLoc), which also contains tcdR, encoding the toxin-specific sigma factor that is required for toxin gene expression; tcdE, encoding a holin that likely allows for toxin efflux; and tcdC, encoding a putative negative regulator of toxin production whose function is debated (5–10). Because the synthesis of these large toxins is energy intensive, C. difficile toxin gene expression is directly repressed by multiple regulatory factors to ensure that toxin production occurs only under conditions in which the function of the toxins contributes to the survival of the bacterium within the host (11–13).

Additionally, as a strict anaerobe, C. difficile relies on morphological transformation into a dormant spore to survive the subsequent exodus from the gastrointestinal tract and efficient transmission to a new host (14). While the characteristic morphological stages of sporulation are conserved, the regulatory network that controls C. difficile sporulation initiation and, thus, the activation of Spo0A, the master regulator of sporulation, is divergent from those of other spore formers and is poorly mapped out (15). The three transcriptional repressors CodY, CcpA, and RstA, which directly repress toxin gene expression in C. difficile, are also involved in the regulation of early-stage sporulation events; however, the genetic pathways through which these regulators control spore formation have not been fully delineated (16–18). CodY and CcpA are largely transcriptional repressors that coordinate adaptation to environmental conditions in response to nutritional signals (17, 19). RstA is a member of the multifunctional Rgg/Rap/NprR/PlcR/PrgX (RRNPP) family of proteins and controls toxin production and sporulation through separate domains (13, 18), indicating that RstA utilizes two different molecular mechanisms to regulate these phenotypes. RstA function has been studied only in the laboratory isolate 630Δerm, and its contribution to sporulation and toxin production has not been evaluated in other C. difficile strains.

As new C. difficile PCR ribotypes emerge and prevail in the clinical population, the toxin and sporulation phenotypes of these isolates are often characterized to determine which traits allow these strains to exhibit increased virulence and circulate persistently (20–25). The variability in tcdA and tcdB gene sequences has led to the established method of toxinotyping C. difficile strains using PCR-restriction fragment length polymorphisms (RFLPs) (reviewed in reference 26), although single nucleotide polymorphisms (SNPs) and small insertions and deletions located within the promoter regions and open reading frames of tcdR, tcdE, and tcdC also purportedly contribute to toxin gene expression, production, and secretion. Some of these changes have been documented in the literature, including deletions and frameshift mutations within the tcdC putative negative regulator (27, 28) and alternate TcdE isoforms that influence toxin secretion (29). Although there are a few nucleotide changes among C. difficile strains within the tcdA and tcdB promoter regions, none of these overlap the σ^TcdR-dependent promoters essential for their transcription. However, numerous point mutations are located within the tcdR promoter region, many of which overlap the consensus sequences of the σ^A- and σ^D-dependent promoters and the RstA and CodY binding sites. We hypothesized that the point mutations within the tcdR promoters affect transcription initiation and influence RstA- and CodY-dependent repression, both of which may account for some of the variable, strain-specific toxin expression phenotypes observed.

To determine the impact of RstA on sporulation and toxin production in clinically relevant C. difficile strains, a null rstA mutant was created in R20291, an epidemic isolate that emerged in the mid-2000s (30). We demonstrate that RstA is a regulator of critical virulence factors in this epidemic background and reveal strain-dependent differences that result in robust regulation of R20291 sporulation and toxin production. Surprisingly, RstA does not affect R20291 motility gene expression, which consequently alters the regulatory pathway through which RstA inhibits toxin gene expression in this background. Furthermore, we dissect the conserved regulatory and promoter features that control transcription of the bistable toxin regulator tcdR and define the regulatory contributions of RstA and CodY to toxin production in four important C. difficile strains. The observation that strain-specific nucleotide substitutions in the tcdR promoter alter the regulation of toxin gene expression in vitro illuminates another molecular mechanism by which more virulent strains exhibit altered toxin levels during C. difficile infection.

RESULTS

RstA positively influences R20291 sporulation.

To assess the contribution of RstA to controlling sporulation and toxin production in phenotypically diverse C. difficile strains, we attempted to create null rstA mutations in R20291, an epidemic ribotype 012 (RT012) isolate; VPI 10463 (ATCC 43255), a high-toxin-producing RT087 strain; and 5325, representing the nonmotile, agriculturally associated ribotype 078. To accomplish this, we used constructs that were previously successful in the 630 background to generate either a nonpolar rstA mutation via the CRISPR/Cas9 system or an insertional disruption within the coding region of rstA (13, 18). An rstA mutant, in which the entire open reading frame was deleted, was generated in R20291 and confirmed by PCR (see Fig. S1 and S2A in the supplemental material).

We successfully conjugated exogenous DNA into VPI 10463 using a previously reported heat shock method (31); however, no detectable rstA mutants were identified. Although we obtained multiple 5325 rstA mutants, this mutation was not stable after a single passage under any tested condition. Both VPI 10463 and 5325 were used throughout the rest of this study, as VPI 10463 is frequently used in laboratory settings and in the mouse model of C. difficile infection (32–36) and 5325 represents an emerging, clinically relevant ribotype with unique characteristics (37, 38).

The amino acid sequences of RstA are identical in 630Δerm, R20291, and VPI 10463, and only a single-residue substitution is present in 5325 (V371I), suggesting that RstA function is highly conserved in these C. difficile strains. As RstA promotes the initiation of sporulation in 630Δerm (18), we asked whether RstA exerts the same effect on strain R20291. Sporulation was measured after 24 h of growth on 70:30 sporulation agar (a mixture of 70% SMC medium and 30% BHIS medium) by enumerating ethanol-resistant spores and viable vegetative cells and by phase-contrast microscopy. As expected, the sporulation frequency was reduced ∼27-fold in the R20291 rstA mutant compared to the R20291 parent (Fig. 1A), indicating that RstA retains similar regulatory functions in sporulation in 630Δerm and R20291. The rstA mutation was complemented by integrating the rstA allele driven by its native promoter onto the chromosome via Tn916 (39, 40) (Fig. S2B), and the sporulation frequency in this strain was increased to greater-than-wild-type levels (Fig. S2C). VPI 10463 exhibited a relatively low sporulation rate of 6.3%, while we observed a robust sporulation frequency of 79.4% in 5325 (Fig. 1A).

FIG 1.

RstA promotes sporulation and Spo0A-dependent gene expression. (A) Representative phase-contrast micrographs of 630Δerm, 630Δerm rstA (MC1118), R20291, R20291 rstA (MC1402), VPI 10463, and 5325 grown on 70:30 sporulation agar at H₂₄. The numbers represent the percentages of ethanol-resistant spores compared to the total number of viable cells, at H₂₄, and the standard errors of the means from at least three biological replicates. Bars, 5 μm. (B) qRT-PCR analysis of sigE transcript levels in 630Δerm, 630Δerm rstA (MC1118), R20291, R20291 rstA (MC1402), VPI 10463, and 5325 grown on 70:30 sporulation agar at H₁₂, relative to 630Δerm. The means and standard errors of the means for four biological replicates are shown, and markers represent the independent values within each mean. *, P < 0.05 using Student’s t test comparing each rstA mutant to its isogenic parent.

To further probe the effect of RstA on sporulation initiation in C. difficile, we measured transcript levels of the Spo0A-dependent gene sigE, which encodes an early-stage sporulation-specific sigma factor, using reverse transcription-quantitative PCR (qRT-PCR) (Fig. 1B). The relative expression level of sigE in each strain mirrored the sporulation frequencies observed, indicating that RstA exerts its regulatory effect on the early stages of R20291 sporulation, as previously observed in 630Δerm (18).

RstA inhibits toxin gene expression and production in R20291.

RstA directly represses tcdA, tcdB, tcdR, and sigD transcription in 630Δerm (13). The direct inhibition of two toxin gene activators, TcdR and SigD, as well as the direct repression of the toxin genes results in a robust regulatory network that tightly controls toxin production (13). To determine whether RstA regulates toxin production in R20291, we first measured TcdA and TcdB protein levels present in tryptone-yeast extract (TY) culture supernatants after 24 h of growth via an enzyme-linked immunosorbent assay (ELISA). As anticipated, toxin levels were increased approximately 11-fold in the 630Δerm rstA mutant compared to its isogenic parent and were approximately 15-fold higher in the R20291 rstA mutant than in its parent strain (Fig. 2A). Supernatant toxin levels in the R20291 rstA Tn916::rstA complemented strain returned to wild-type levels (Fig. S2D). Unsurprisingly, VPI 10463 exhibited large quantities of TcdA and TcdB in the culture supernatant, while the 5325 supernatant contained toxin levels comparable to those of 630Δerm (Fig. 2A).

FIG 2.

RstA inhibits C. difficile toxin production. (A) Levels of TcdA and TcdB present in the supernatants of 630Δerm, 630Δerm rstA (MC1118), R20291, R20291 rstA (MC1402), VPI 10463, and 5325 grown in TY medium (pH 7.4) at H₂₄ were quantified by an ELISA, as detailed in Materials and Methods, and read as the absorbance at 450 nm normalized to the cell density (OD₆₀₀). (B) Representative TcdA Western blot analysis of 630Δerm, 630Δerm rstA (MC1118), R20291, R20291 rstA (MC1402), VPI 10463, and 5325 cells grown in TY medium (pH 7.4) at H₂₄. Densitometry analysis of full-length TcdA is included below. (C) qRT-PCR analysis of tcdA, tcdB, tcdR, and tcdE transcript levels in 630Δerm, 630Δerm rstA (MC1118), R20291, R20291 rstA (MC1402), VPI 10463, and 5325 grown in TY medium (pH 7.4) at T₄ (defined as 4 h after entry into stationary phase). The means and standard errors of the means for at least three biological replicates are shown, and markers represent the independent values within each mean. *, P < 0.05 using Student’s t test comparing each rstA mutant to its isogenic parent.

To examine the amount of toxin within cells, Western blot analyses were performed, and TcdA protein levels were measured after 24 h of growth in TY medium. The Western blot results indicated that the 630Δerm rstA mutant had higher levels of TcdA than the parent strain, as previously observed (13, 18). R20291 cells contained approximately one-half of the total amount of toxin that was found within 630Δerm cells (Fig. 2B). The R20291 rstA mutant had a modest increase in TcdA toxin levels compared to R20291, and this effect was restored to wild-type levels in the complemented strain (Fig. S2E). These data suggest that the majority of TcdA synthesized by R20291 is secreted into the supernatant, whereas a lower ratio of total 630Δerm TcdA is located extracellularly. Although we did not test TcdB levels in whole-cell lysates, it is reasonable to hypothesize that TcdB is exported from the cells through similar mechanisms and with a similar efficiency as TcdA, as these proteins share considerable identity, and no difference between TcdA and TcdB secretion has been reported (8, 41). There remains the possibility, albeit unlikely, that a significantly greater portion of TcdB is retained within R20291 cells than in 630Δerm cells, which could explain the differences observed between the strains. There were ∼2.6-fold-higher TcdA protein levels observed in VPI 10463 cells, while 5325 cells had ∼4-fold-lower TcdA protein levels, than those in strain 630Δerm (Fig. 2B).

To verify that RstA affected toxin production through regulation of toxin gene expression, we measured tcdA, tcdB, and tcdR transcript levels from cultures grown in TY medium at T₄, which corresponds to 4 h after cultures enter stationary phase, using qRT-PCR (Fig. 2C). As previously noted, tcdA, tcdB, and tcdR transcript levels were increased ∼15- to 30-fold in the 630Δerm rstA mutant compared to the 630Δerm parent (13) (Fig. 2C). Similar to the TcdA/TcdB ELISA results, the 630Δerm rstA mutant and the R20291 strain had comparable levels of tcdA transcript; however, R20291 had higher transcript levels of tcdB and tcdR. Transcript levels of all three tcd genes were increased ∼9- to 12-fold in the R20291 rstA mutant compared to its isogenic parent, although more variability was observed in the R20291 rstA mutant than in all other strains. VPI 10463 exhibited higher toxin gene expression levels than R20291, as previously observed under other conditions (22, 23). Of note, the R20291 rstA mutant had higher toxin gene expression levels than VPI 10463 (Fig. 2C), suggesting that R20291 is capable of producing high levels of toxin if RstA-dependent toxin inhibition is removed.

Because of the differences observed for toxin levels in the supernatant versus whole-cell lysates, we asked whether RstA influenced the expression of the toxin holin-encoding gene tcdE, which is purported to play a role in toxin secretion in a strain-dependent manner (8, 29, 41). tcdE is the third gene located within the PaLoc (tcdR-tcdB-tcdE-tcdA), and its transcription is driven by its own, unmapped promoter (29). Transcript levels of tcdE showed a pattern of expression similar to those of the other toxin genes, with increased levels in both the 630Δerm rstA and R20291 rstA strains compared to those of their isogenic parents (Fig. 2C). To determine whether the tcdE sequences in these isolates match the annotated GenBank sequences and encode a functional TcdE protein, primers flanking the tcdE open reading frame were used to amplify this region, and the resulting PCR products were purified and sequenced (data not shown). The 630Δerm, R20291, and VPI 10463 tcdE sequences were identical to the annotated sequences. These tcdE sequences differ only in a nonsynonymous mutation in the R20291 sequence, resulting in a conserved I151V amino acid substitution. Analysis of M120, a sequenced RT078 reference genome, revealed the absence of the M1 start codon, leaving only the M25 and M27 start codons present. The tcdE open reading frame is capable of producing three isoforms, with TcdE₁₄₂, initiating from M25, as the predominant isoform produced (29). TcdE₁₄₂ also exhibits the greatest holin activity (29). TcdE₁₆₆, the full-length isoform initiated from M1, is not efficiently translated in vivo, and the degeneration of the M1 start codon in the RT078 strains supports the hypothesis that this isoform is not essential for C. difficile toxin secretion. Altogether, these data suggest that even though RstA inhibits tcdE gene expression, RstA does not contribute to the strain-dependent differences observed for TcdE-mediated toxin secretion.

In strain 630Δerm, RstA inhibits tcdA and tcdB transcription directly by binding their respective promoters (13). To determine whether RstA inhibits tcdA and tcdB transcription in R20291, alkaline phosphatase (AP) reporters were constructed by fusing the promoter regions of these genes to phoZ and integrating the reporters into the chromosomes of R20291 and R20291 rstA. PtcdA::phoZ and PtcdB::phoZ reporter activities were increased approximately 13- and 10-fold, respectively, in the R20291 rstA background (Fig. 3), indicating that RstA represses tcdA and tcdB transcription. The σ^TcdR-dependent promoters of tcdA and tcdB and the putative RstA binding sites, which overlap the −35 consensus sequences of these promoters, are identical in 630Δerm and R20291, strongly suggesting that RstA directly represses toxin gene transcription similarly in both strains. Additionally, as observed in the qRT-PCR analysis of these same genes (Fig. 2C), reporter activity from the toxin gene promoter fusions in the R20291 rstA background was highly variable (Fig. 3), suggesting that the absence of RstA enhances the variability of tcdR, tcdA, and tcdB transcription (see below).

FIG 3.

RstA inhibits tcdA and tcdB transcription in R20291. Shown are alkaline phosphatase activities of PtcdA::phoZ (A) and PtcdB::phoZ (B) integrated into the chromosome using Tn916 in R20291 (MC1451 and MC1453) or R20291 rstA (MC1452 and MC1454). The means and standard errors of the means are shown for five biological replicates. *, P < 0.05 using Student’s t test.

The expression of tcdR is driven by several promoters (Fig. 4A), one of which requires the motility-specific sigma factor SigD (42, 43). In 630Δerm, RstA inhibits the expression of the sigD gene, the fourth gene located within the flgB operon, by directly binding to the flgB promoter (13). Interestingly, neither flgB nor sigD transcript levels are altered in the R20291 rstA mutant compared to the R20291 parent (Fig. S3). SigD activity is slightly greater in the R20291 rstA mutant, as evidenced by increased transcript levels of the SigD-dependent gene fliC compared to those in the R20291 parent strain (Fig. S3), although this regulatory effect is not statistically significant. Strain 5325 was omitted from this analysis, as its genome lacks the flagellar genes. Previously, reporter fusions revealed that RstA represses PflgB^(630Δerm) and PflgB^(R20291) activity ∼1.5-fold between the 630Δerm and 630Δerm rstA strains (13). Furthermore, biotin-labeled DNA pulldown assays with 630Δerm-derived lysates expressing a recombinantly tagged RstA protein demonstrated that RstA directly binds to DNA fragments containing either the 630Δerm or the R20291 flgB promoter (13), confirming that RstA^(630Δerm) directly represses both 630Δerm and R20291 flgB transcription. However, those previous studies were all performed using the 630Δerm background. Our current data demonstrate no significant effect of RstA on motility gene transcription in the R20291 background, suggesting that there are significant strain-specific differences in RstA-dependent regulation of flgB transcription between the 630Δerm and R20291 backgrounds. RstA may control toxin production in R20291 by influencing SigD activity, but there do not appear to be strong, direct effects of RstA on sigD gene expression in this strain.

FIG 4.

Strain-specific nucleotide changes overlap key regulatory recognition sites within the tcdR promoter region. (A) Schematic depicting the mapped transcriptional start sites (bent arrows) (the nucleotide [nt] position is marked below, and the specific sigma factor that recognizes the promoter is indicated above) and the RstA, CodY, and CcpA binding sites (red boxes) of 630Δerm (11–13). The sequence compositions of the three reporter fusions created for each strain are indicated as gray boxes below. (B) Alignment and annotation of nucleotide changes within the σ^D- and σ^A-dependent tcdR promoter regions for four diverse C. difficile strains, including 630Δerm, R20291, VPI 10463, and 5325. Specific sequence features are identified, and nonconserved nucleotides are marked in red.

RstA and CodY regulate tcdR transcription in concert.

Our previous work revealed that RstA controls 630Δerm tcdR transcription by directly binding to the σ^A- and σ^D-dependent promoters directly upstream of tcdR (13) (Fig. 4A). PtcdR(σ^D) features an imperfect inverted repeat immediately upstream of the conserved −35 sequence that likely serves as the RstA binding site (13) (Fig. 4A). Notably, the RstA box and CodY III box of PtcdR(σ^A) perfectly overlap each other, suggesting that inhibiting PtcdR(σ^A) transcription is important for toxin regulation. Alignment of the PtcdR regions from 630Δerm, R20291, VPI 10463, and 5325 revealed multiple single nucleotide substitutions throughout the conserved regulatory features (Fig. 4B). The R20291 and VPI 10463 PtcdR(σ^A) sequences each revealed distinct single nucleotide changes within the RstA box and CodY III binding site of PtcdR(σ^A). Additionally, the 5325 sequence contains eight individual nucleotide changes within this region (five mismatches within the CodY III binding site and an additional three extending through the RstA box). As 5325 is nonmotile and the flagellar genes, including sigD, are absent from its genome, we accordingly noted significant sequence degeneration in the region that aligns with 630Δerm PtcdR(σ^D) and hypothesized that the 5325 PtcdR region does not rely on SigD for toxin gene regulation. Furthermore, we hypothesized that the nucleotide changes in each strain alter RstA and CodY regulation of tcdR expression. To test these hypotheses, the full-length tcdR promoter and the fragments corresponding to the σ^A- and σ^D-dependent promoters from R20291, VPI 10463, and 5325 were amplified and cloned upstream of the phoZ gene in a multicopy plasmid. We did not include an individual PtcdR(σ^D)^{(VPI 10463)} reporter construct, as the 630Δerm and VPI 10463 sequences spanning this region are identical (Fig. 4B). All of the reporter constructs were expressed in the 630Δerm, 630Δerm rstA, and 630Δerm codY backgrounds to eliminate additional strain-dependent regulatory effects. Alkaline phosphatase reporter assays were performed on cells containing promoter-phoZ fusions from samples collected after 24 h of growth in TY medium.

The reporter activities for the full-length tcdR promoters expressed in the 630Δerm background were equivalent, except that the full-length PtcdR^(R20291) fusion exhibited ∼1.7-fold-lower activity than the others (Fig. 5A). Interestingly, the activities of the R20291, VPI 10463, and 5325 PtcdR(σ^A) reporters in the 630Δerm background were ∼2.4- to 3.6-fold higher than the activity from the PtcdR(σ^A)^(630Δerm) fusion (Fig. 5A), suggesting that CodY and RstA do not repress basal transcription from these promoters as efficiently. These data indicate that the CodY III and RstA consensus sequences in the 630Δerm promoter allow for the greatest repression by RstA and CodY and that nucleotide changes within this region reduce the repressive effect by RstA and/or CodY.

FIG 5.

Strain-specific nucleotide changes within the tcdR promoter affect RstA- and CodY-dependent regulation. Shown are alkaline phosphatase activities of a series of PtcdR::phoZ fusions, expressed from a plasmid in 630Δerm (A), 630Δerm rstA (B), and 630Δerm codY::erm (C), comprised of the full-length tcdR promoter from 630Δerm (MC1088/MC1330/MC1552), R20291 (MC1458/MC1459/MC1554), VPI 10463 (MC1563/MC1564/MC1558), and 5325 (MC1425/MC1426/MC1556); the σ^A-dependent tcdR promoter from 630Δerm (MC1285/MC1331/MC1553), R20291 (MC1460/MC1461/MC1555), VPI 10463 (MC1565/MC1566/MC1559), and 5325 (MC1427/MC1428/MC1557); and the σ^D-dependent tcdR promoter from 630Δerm (MC1145/MC1332), R20291 (MC1462/MC1463), and 5325 (MC1429/MC1430). As the nucleotide sequences of the 630Δerm and VPI 10463 tcdR(σ^D) promoters are identical, an individual PtcdR(σ^D) fusion for VPI 10463 was not constructed (ND). The promoterless phoZ reporter carried by 630Δerm (MC448) was included as a negative control. Strains were grown in TY medium (pH 7.4) supplemented with 2 μg ml⁻¹ thiamphenicol, and samples assayed for alkaline phosphatase activity were collected at H₂₄. The means and standard errors of the means from at least three biological replicates are shown. *, P < 0.05 using one-way ANOVA followed by Dunnett’s multiple-comparison test compared to the PtcdR^(630Δerm)-derived promoter within the same set. The activity of each reporter in the 630Δerm rstA background or the 630Δerm codY background compared to 630Δerm, with the exception of the PtcdR(σ^D)⁽⁵³²⁵⁾ reporter, was statistically significant (P < 0.05) using one-way ANOVA followed by Dunnett’s multiple-comparison test (not shown in the figure).

The reporter activities from the 630Δerm-derived promoter fusions expressed in the rstA deletion mutant were similar to those previously reported (13), with the strongest RstA-dependent effects (∼3.5-fold) observed for the PtcdR(σ^A) reporter and ∼1.6- to 1.7-fold changes in reporter activity for the full-length and the PtcdR(σ^D) reporters (Table 1; Fig. 5A and B). The full-length tcdR and PtcdR(σ^A) reporters from R20291, VPI 10463, and 5325 and the PtcdR(σ^D)^(R20291) reporter showed ∼2- to 3-fold-higher activity when expressed in the 630Δerm rstA mutant than in the parent strain (Table 1; Fig. 5B). These data demonstrate RstA-dependent repression of all strains’ tcdR promoters. Minor changes in activity were observed for the four different tcdR promoters, but none of the sequence changes fully relieved RstA-mediated repression, suggesting that RstA inhibition of tcdR transcription is conserved in C. difficile. As expected, the fusion containing the unconserved PtcdR(σ^D)⁽⁵³²⁵⁾ reported exhibited no reporter activity above background levels, consistent with the degenerative sequence.

TABLE 1.

Fold changes in PtcdR reporter activity expressed in the 630Δerm, 630Δerm rstA, or 630Δerm codY background

Reporter fusion	Fold change in reporter activitya
	rstA/630Δerm	codY/630Δerm
PtcdR(630Δerm)::phoZ	1.61	29.5
PtcdR(R20291)::phoZ	2.01	32.2
PtcdR(VPI 10463)::phoZ	2.85	30.2
PtcdR(5325)::phoZ	3.09	14.0
PtcdR(σA)(630Δerm)::phoZ	3.59	63.0
PtcdR(σA)(R20291)::phoZ	3.20	17.6
PtcdR(σA)(VPI 10463)::phoZ	2.65	21.2
PtcdR(σA)(5325)::phoZ	3.61	13.0
PtcdR(σD)(630Δerm)::phoZ	1.74	ND
PtcdR(σD)(R20291)::phoZ	2.35	ND
PtcdR(σD)(5325)::phoZ	0.75	ND

^aFold change of the average reporter activity (presented in Fig. 5) in the 630Δerm rstA or 630Δerm codY strain divided by the activity in the 630Δerm parent strain. ND, not determined; PtcdR(σ^D) reporter activity was not tested in 630Δerm codY as no CodY binding site is located within the tcdR σ^D-dependent promoter region (11).

To assess the specific repression that CodY exerts on tcdR transcription, we compared the activities of the full-length tcdR and the σ^A-dependent promoter reporters in 630Δerm and 630Δerm codY. Promoter activities from the 630Δerm, R20291, and VPI 10463 full-length PtcdR reporters were similarly elevated in the codY mutant, relative to the parent strain (∼30-fold) (Table 1). However, the PtcdR⁽⁵³²⁵⁾ reporter exhibited approximately one-half of the activity in the absence of CodY (14-fold change) (Table 1; Fig. 5C). Our data suggest that CodY derepression of tcdR expression is weaker in 5325, which corresponds to the greater number of mismatches within the CodY III box of the 5325 sequence (Fig. 4B). We also observed that the activities of the 630Δerm, R20291, VPI 10463, and 5325 PtcdR(σ^A) reporters were similar in the 630Δerm codY mutant (Fig. 5C). But the lower activity from the PtcdR(σ^A)^(630Δerm) reporter in the 630Δerm background resulted in a 63-fold change in activity when measured in the 630Δerm codY mutant (Table 1), emphasizing that CodY most efficiently represses transcription from the 630Δerm-derived σ^A-dependent promoter, as it contains the most conserved CodY binding site. Overall, relief of CodY repression results in greater activity from the PtcdR reporters than in the absence of RstA (Fig. 5B versus Fig. 5C). However, as these regulators respond to different cofactors, RstA and CodY repression likely prevents full tcdR transcription unless DNA binding by both regulators is relieved.

RstA autoregulates its transcription in R20291.

RstA directly represses its own transcription in 630Δerm by binding to an imperfect inverted repeat that overlaps the −10 consensus sequence and start of transcription for the σ^A-dependent rstA promoter (13, 18) (Fig. S4). This direct negative autoregulation results in constitutive expression of rstA throughout growth and is typical of the RRNPP family of proteins (18, 44, 45). Because the rstA mutations that we generated were clean deletions and eliminated the ability to detect rstA transcripts (Fig. S5A and B), we constructed reporter fusions comprising the 500 bp upstream of the rstA translational start from both 630Δerm and R20291, to test whether RstA represses rstA transcription in the R20291 background [referred to here as PrstA^(630Δerm)::phoZ and PrstA^(R20291)::phoZ] (Fig. S4). This upstream region includes the 5ʹ end of the open reading frame of the divergently transcribed upstream gene CD3669, and this region contains 3 nucleotide substitutions that are located in the R20291 sequence compared to 630Δerm (Fig. S4). To determine how RstA impacts its own transcription in these two strains, the fusions were integrated into their respective parent and rstA mutant chromosomes, and the PrstA^(R20291)::phoZ reporter was also integrated into the 630Δerm and 630Δerm rstA mutant strains to allow for direct comparisons.

After 8 h of growth on 70:30 sporulation agar, we found that promoter activity from the PrstA^(630Δerm)::phoZ reporter was ∼1.3-fold greater in the 630Δerm rstA background than in the parent strain (Fig. 6), which correlates with the fold changes seen from the same fusion on a plasmid expressed in 630Δerm and an isogenic rstA::erm mutant (13, 18). Surprisingly, PrstA^(R20291)::phoZ exhibited identical activity in both the 630Δerm and 630Δerm rstA backgrounds (Fig. 6), indicating that there are modest sequence-dependent effects influencing expression from the 630Δerm and R20291 rstA promoters. Finally, the PrstA^(R20291)::phoZ reporter exhibited a significant reduction in activity in the R20291 parent background compared to the 630Δerm parent background and an ∼2.7-fold increase in activity in the rstA mutant compared to the R20291 parent (Fig. 6). We observed significantly higher rstA transcript levels in the R20291 strain than in 630Δerm in two different growth media (Fig. S5A and B), suggesting that the basal expression level of rstA is higher in R20291, which may explain some of the strain-dependent differences observed with RstA regulation. Altogether, these data indicate that RstA strongly represses its own expression in the R20291 background and reveal the presence of strain-dependent differences that influence RstA autoregulation.

FIG 6.

RstA negatively autoregulates its own expression in R20291. Shown are alkaline phosphatase activities of the rstA promoter from 630Δerm [PrstA^(630Δerm)::phoZ] integrated into 630Δerm (MC1640) and 630Δerm rstA (MC1641) or the rstA promoter from R20291 [PrstA^(R20291)::phoZ] integrated into 630Δerm (MC1668), 630Δerm rstA (MC1669), R20291 (MC1642), and R20291 rstA (MC1643). Strains were grown on 70:30 sporulation agar, and samples assayed for alkaline phosphatase activity were collected at H₈. The means and standard errors of the means from four biological replicates are shown. *, P < 0.05 using Student’s t test for the PrstA^(630Δerm)::phoZ comparison or one-way ANOVA followed by the Tukey’s multiple-comparison test for the PrstA^(R20291)::phoZ comparisons.

DISCUSSION

C. difficile toxin production is the mediator of significant gastrointestinal disease throughout the course of infection in the host (1). The production of toxin has been attributed to nutritional availability, temperature control, the activity of the flagellum-specific sigma factor SigD, and the synthesis of secondary messengers (42, 43, 46–50; reviewed in reference 51), yet strain-specific differences have been extensively noted for most aspects of toxin gene expression, synthesis, and secretion (9, 10, 22, 23, 25, 29, 52). Furthermore, many of these environmental conditions also affect another important aspect of C. difficile pathogenesis: sporulation. The multifunctional regulator RstA supports early-stage sporulation events and directly inhibits toxin gene expression in response to a putative cofactor in the historical strain 630Δerm (13, 18). In this study, we determined the regulatory effects of RstA on toxin production, motility gene expression, and sporulation in the epidemic isolate R20291 and assessed the effects of RstA and CodY inhibition of the tcdR promoters from four diverse C. difficile strains.

As in the 630 background, we found that RstA induces R20291 sporulation and inhibits tcdR, tcdA, and tcdB expression. Surprisingly, motility gene expression was not strongly affected, suggesting that the effect of RstA on SigD-dependent toxin gene expression is significantly less pronounced in the R20291 background. Unfortunately, attempts to assess the impact of RstA on the individual R20291 PtcdR promoters were unsuccessful due to the undetectable level of alkaline phosphatase activity from the single-copy reporter in the R20291 and R20291 rstA backgrounds; however, the remarkably low expression level of the tcdR promoter is well documented (13, 43, 47, 53, 54). Further studies to elucidate the regulatory effects of RstA-dependent toxin gene repression in R20291 and additional C. difficile strains will illuminate the seemingly minor differences in the toxin regulatory network between strains that may significantly amplify and impact total toxin production and the resulting virulence.

One apparent difference in the regulation of toxin by RstA between the R20291 and 630Δerm strains is the inherent variability in toxin gene expression (Fig. 2C and Fig. 3). The increased variability of toxin gene expression in the R20291 rstA mutant may be a result of bistable toxin production mediated by TcdR, in which individual cells differentially express toxin genes as either toxin-on or toxin-off (54). Although our data are representative of the population, the higher incidences of variability in tcdA, tcdB, tcdR, and tcdE transcript levels and in tcdA::phoZ and tcdB::phoZ reporter activities in the R20291 rstA mutant implicate RstA as a mediator of bimodal toxin gene expression by biasing the population in the toxin-off state, as is the case for CodY (54). This variability in toxin gene expression is not observed in the 630Δerm rstA mutant, corroborating previous results showing that a higher proportion of cells are toxin-on in 630Δerm, likely due to higher sigD expression levels (54, 55).

The regulatory effects of RstA appeared more marked in the R20291 strain than in the previously studied 630Δerm background, with a more severe sporulation defect than in 630Δerm (7-fold in 630Δerm versus 27-fold in R20291) (Fig. 1). In addition, greater fold changes in activity were observed with the PrstA, PtcdA, and PtcdB reporters in the R20291 and R20291 rstA strains than in the corresponding 630Δerm constructs (Fig. 3 and 6). The rstA transcript levels were also higher in R20291 than in 630Δerm (see Fig. S5 in the supplemental material), suggesting that higher levels of RstA influence RstA activity and its regulatory effects. Furthermore, our observations that VPI 10463 is a poor spore former yet a high-toxin producer and that 5325 produces prolific spores and low levels of toxin suggest that RstA regulation may be involved in these opposing phenotypes.

The dual repressive effects of RstA and CodY on PtcdR(σ^A)-dependent transcription highlight the importance of tightly regulating toxin production in C. difficile. Nucleotide substitutions within the RstA and CodY consensus sequences resulted in minor differences in PtcdR reporter activity (Fig. 4 and 5; Table 1), but none of the changes completely abolished repression by these regulators. Our data suggest that RstA- and CodY-mediated repression of toxin gene expression occurs in the R20291, VPI 10463, and 5325 strains, although caution in extrapolating these data is warranted, as the differences observed in PrstA^(R20291)::phoZ activity between its native R20291 background and the 630Δerm heterologous background were striking. These findings further underscore the importance of differentiating between sequence-specific and strain-specific regulatory impacts. Identifying the cofactor that regulates RstA DNA binding activity and deciphering the molecular mechanisms by which RstA promotes sporulation will aid in dissecting these sequence-dependent and strain-dependent regulatory differences.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

The bacterial strains and plasmids used in this study are shown in Table 2. C. difficile strains were cultured in BHIS (brain heart infusion supplemented with yeast extract) medium (56) or TY medium (pH 7.4) (57) in the presence of 2 to 5 μg/ml thiamphenicol or 2 to 5 μg/ml erythromycin, as indicated, in a 37°C anaerobic chamber containing an atmosphere of 10% H₂, 5% CO₂, and 85% N₂, as previously detailed (58). Cultures of C. difficile grown overnight contained 0.1% taurocholate, to germinate spores, and 0.2% fructose, to prevent the formation of spores (57, 59). Escherichia coli and Bacillus subtilis strains were grown in LB (60), supplemented with 100 μg/ml ampicillin and/or 20 μg/ml chloramphenicol, as indicated, for E. coli or with 1 μg/ml erythromycin, as indicated, for B. subtilis at 37°C. After conjugation with C. difficile, E. coli and B. subtilis were counterselected using 50 to 100 μg/ml kanamycin (61).

TABLE 2.

Bacterial strains and plasmids

Strain or plasmid	Relevant genotype or feature(s)	Source and/or reference(s)
Strains
E. coli HB101	F− mcrB mrr hsdS20(rB− mB−) recA13 leuB6 ara-14 proA2 lacY1 galK2 xyl-5 mtl-1 rpsL20	B. Dupuy
C. difficile
630Δerm	Erms derivative of strain 630	Nigel Minton; 68
R20291	Clinical isolate	30
VPI 10463	ATCC 43255	ATCC
5325	BAA-1875	ATCC
MC364	630Δerm codY::erm	16
MC448	630Δerm/pMC358	67
MC1088	630Δerm/pMC713	13
MC1118	630Δerm ΔrstA	13
MC1145	630Δerm/pMC753	13
MC1285	630Δerm/pMC812	13
MC1330	630Δerm ΔrstA/pMC713	This study
MC1331	630Δerm ΔrstA/pMC812	This study
MC1332	630Δerm ΔrstA/pMC753	This study
MC1335	R20291::phoZ	This study
MC1402	R20291 ΔrstA	This study
MC1425	630Δerm/pMC878	This study
MC1426	630Δerm ΔrstA/pMC878	This study
MC1427	630Δerm/pMC879	This study
MC1428	630Δerm ΔrstA/pMC879	This study
MC1429	630Δerm/pMC880	This study
MC1430	630Δerm ΔrstA/pMC880	This study
MC1451	R20291 PtcdA::phoZ	This study
MC1452	R20291 rstA PtcdA::phoZ	This study
MC1453	R20291 PtcdB::phoZ	This study
MC1454	R20291 rstA PtcdB::phoZ	This study
MC1458	630Δerm/pMC893	This study
MC1459	630Δerm ΔrstA/pMC893	This study
MC1460	630Δerm/pMC894	This study
MC1461	630Δerm ΔrstA/pMC894	This study
MC1462	630Δerm/pMC895	This study
MC1463	630Δerm ΔrstA/pMC895	This study
MC1514	R20291 ΔrstA rstA::Tn916	This study
MC1552	630Δerm codY::erm/pMC713	This study
MC1553	630Δerm codY::erm/pMC812	This study
MC1554	630Δerm codY::erm/pMC893	This study
MC1555	630Δerm codY::erm/pMC894	This study
MC1556	630Δerm codY::erm/pMC878	This study
MC1557	630Δerm codY::erm/pMC879	This study
MC1558	630Δerm codY::erm/pMC947	This study
MC1559	630Δerm codY::erm/pMC948	This study
MC1563	630Δerm/pMC947	This study
MC1564	630Δerm ΔrstA/pMC947	This study
MC1565	630Δerm/pMC948	This study
MC1566	630Δerm ΔrstA/pMC948	This study
MC1640	630Δerm PrstA(630Δerm)::phoZ	This study
MC1641	630Δerm ΔrstA PrstA(630Δerm)::phoZ	This study
MC1642	R20291 PrstA(R20291)::phoZ	This study
MC1643	R20291 ΔrstA PrstA(R20291)::phoZ	This study
MC1668	630Δerm PrstA(R20291)::phoZ	This study
MC1669	630Δerm ΔrstA PrstA(R20291)::phoZ	This study
B. subtilis
BS49	CU2189::Tn916	P. Mullany
MC472	BS49 pMC370::Tn916	67
MC532	BS49 pMC394::Tn916	This study
MC1443	BS49 pMC886::Tn916	This study
MC1444	BS49 pMC887::Tn916	This study
MC1632	BS49 pMC973::Tn916	This study
MC1633	BS49 pMC974::Tn916	This study
Plasmids
pRK24	Tra+ Mob+ bla tet	69
pCR2.1	bla kan	Invitrogen
pSMB47	Tn916 integrational vector; Cmr Ermr	70
pMC123	E. coli-C. difficile shuttle vector; bla catP	65
pMC358	pMC123::phoZ	67
pMC370	pSMB47::phoZ	67
pMC389	pCR2.1 rstA	This study
pMC394	pSMB47 rstA	This study
pMC713	pMC123 PtcdR::phoZ from 630Δerm	13
pMC729	pJK02 with CRISPR/Cas9-rstA	13, 71
pMC753	pMC123 PtcdR(σD)::phoZ from 630Δerm	13
pMC812	pMC123 PtcdR(σA-76 bp)::phoZ from 630Δerm	13
pMC878	pMC123 PtcdR::phoZ from 5325	This study
pMC879	pMC123 PtcdR(σA-76 bp)::phoZ from 5325	This study
pMC880	pMC123 PtcdR(σD)::phoZ from 5325	This study
pMC886	pSMB47 PtcdA::phoZ from R20291	This study
pMC887	pSMB47 PtcdB::phoZ from R20291	This study
pMC893	pMC123 PtcdR::phoZ from R20291	This study
pMC894	pMC123 PtcdR(σA-76 bp)::phoZ from R20291	This study
pMC895	pMC123 PtcdR(σD)::phoZ from R20291	This study
pMC947	pMC123 PtcdR::phoZ from VPI 10463	This study
pMC948	pMC123 PtcdR(σA-76 bp)::phoZ from VPI 10463	This study
pMC973	pSMB47 PrstA::phoZ from 630Δerm	This study
pMC974	pSMB47 PrstA::phoZ from R20291	This study

Strain and plasmid construction.

C. difficile strains 630 (RT012) (GenBank accession no. NC_009089.1), R20291 (RT027) (GenBank accession no. FN545816.1), VPI 10643 (ATCC 43255) (RT087) (GenBank accession no. NZ_CM000604.1), and M120 (RT078) (GenBank accession no. NZ_FRES01000002.1) were used as the templates for primer design and for PCR amplification, with the exception that 5325 (ATCC 1875) (RT078) (unsequenced) genomic DNA was the template for PCR amplification, as 5325 was used as the RT078 representative strain in all of the experiments. Sequencing of cloned PCR fragments was performed by Eurofins Genomics (Louisville, KY), and all M120 and 5325 sequences used were identical. Sequencing of the C. difficile tcdE open reading frames was performed by GenScript (Piscataway, NJ). Oligonucleotides used for PCR and qRT-PCR analyses are listed in Table 3. Details of strain and plasmid construction are shown in Fig. S1 in the supplemental material.

TABLE 3.

Oligonucleotides^a

Primer	Sequence (5ʹ→3ʹ)	Use (reference)
oMC44	CTAGCTGCTCCTATGTCTCACATC	Forward primer for rpoC qPCR (65)
oMC45	CCAGTCTCTCCTGGATCAACTA	Reverse primer for rpoC qPCR (65)
oMC189	TGCCTCTTGTAAAGAGTATAGCA	Forward primer for sigD qPCR (18)
oMC190	GCATCAATCAATCCAATGACTCCAC	Reverse primer for sigD qPCR (18)
oMC339	GGGCAAATATACTTCCTCCTCCAT	Forward primer for sigE qPCR (64)
oMC340	TGACTTTACACTTTCATCTGTTTCTAGC	Reverse primer for sigE qPCR (64)
oMC352	GGAGTAGGTTTAGCTTTGTTATTAGGAACC	PCR verification of ΔrstA
oMC355	CTGTTGGAATATCTAGGCGATAAGC	PCR verification of Tn916::rstA (18)
oMC356	TGGTCCTCAGCCTTGTTTAATTC	PCR verification of Tn916::rstA (18)
oMC547	TGGATAGGTGGAGAAGTCAGT	Forward primer for tcdA qPCR (64)
oMC548	GCTGTAATGCTTCAGTGGTAGA	Reverse primer for tcdA qPCR (64)
oMC960	GCCGGATCCGCTGATTGAGCTTTAGTTTCTTCTT	Forward primer for amplification of rstA and its native promoter for complementation
oMC984	GACGCATGCGCTTGTTATAGATTGTTTCTATACCCTTAT	Forward primer for amplification of rstA and its native promoter for complementation
oMC1204	TTCCACAACTTGCTGTTATTTCTC	PCR verification of ΔrstA (18)
oMC1645	GGCGAATTCGGTTTCTAGATTTCATAAAAGATACTA	Forward primer for amplification of 630Δerm/R20291/VPI 10463 PtcdR (13)
oMC1646	GCCGGATCCAAAATCATCCTCTCTTATATTTATAATG	Reverse primer for amplification of 630Δerm/R20291 PtcdR (13)
oMC1769	TTCCTCCTTCATATCTACCCATACATTGACGGATCCAAAATCATCCTCTCTTATATTTATAATG	Reverse primer for amplification of R20291 PtcdR(σA) (13)
oMC1771	TTCCTCCTTCATATCTACCCATACATTGACGGATCCCATTATAATTATATAATCGGCAAATAAATT	Reverse primer for amplification of 630Δerm/R20291 PtcdR(σD) (13)
oMC1984	CACGACGTTGTAAAACGACGGCCAGTATGAGAATTCTAATGACTGATTTAATTCCAATGTTG	Forward primer for amplification of 630Δerm/R20291 PtcdR(σA) (13)
oMC2155	CACGACGTTGTAAAACGACGGCCAGTATGAGAATTCGGTATCTAGATTTCATAAAAAATATTATTT	Forward primer for amplification of 5325 PtcdR
oMC2156	TTCCTCCTTCATATCTACCCATACATTGACGGATCCAAAATCATCCTCTCTTATTTTATAATG	Reverse primer for amplification of 5325 PtcdR
oMC2157	CACGACGTTGTAAAACGACGGCCAGTATGAGAATTCTAATGACTGATTTAATGCATATGTTG	Forward primer for amplification of 5325 PtcdR(σA)
oMC2159	CACGACGTTGTAAAACGACGGCCAGTATGAGAATTCGCTAAAATACTTTATTTATTAGGATAAGATTA	Forward primer for amplification of 5325 PtcdR(σD)
oMC2160	TTCCTCCTTCATATCTACCCATACATTGACGGATCCCATTATAATTGTATGATTAGAAAATAAATTAA	Reverse primer for amplification of 5325 PtcdR(σD)
oMC2163	GACCACACCCGTCCTGTGGATCCTGGTCAGTTGGTAAAATCTATTAAG	Forward primer for amplification of R20291 PtcdA
oMC2164	CCTCCTTCATATCTACCCATACATTGACGGATCCAAAAACCTCCTAGTACTATTATTTTTGAT	Reverse primer for amplification of R20291 PtcdA
oMC2165	GACCACACCCGTCCTGTGGATCCGTCTGTTTTTGAGGAAGATATTTG	Forward primer for amplification of R20291 PtcdB
oMC2166	CCTCCTTCATATCTACCCATACATTGACGGATCCCATCTAAATGCTAAAACTCTTTTATATATC	Reverse primer for amplification of R20291 PtcdB
oMC2333	CACGACGTTGTAAAACGACGGCCAGTATGAGAATTCGCTAAAATACTTTATTTATTAGAAAAAAATTA	Forward primer for amplification of R20291 PtcdR(σD)
oMC2467	GCCGGATCCAAAATCACCCTCTCTTATATTTATAATG	Reverse primer for amplification of VPI 10463 PtcdR
oMC2468	CACGACGTTGTAAAACGACGGCCAGTATGAGAATTCTAATTACTGATTTAATTCCAATGTTG	Forward primer for amplification of VPI 10463 PtcdR(σA)
oMC2555	GACCACACCCGTCCTGTGGATCCAAGAGTAAATAGTAGCTGATTGAG	Forward primer for amplification of 630Δerm/R20291 PrstA
oMC2556	CCTCCTTCATATCTACCCATACATTGACGGATCCACTATTCCCACCTTTTGAAGAC	Reverse primer for amplification of 630Δerm/R20291 PrstA
oMC2561	GCACAAATAAGCCATATAGAGAGGGATAAA	Forward primer for rstA qPCR
oMC2562	TTTTTGATTGCATTTCCTTGGTCTCTAATAA	Reverse primer for rstA qPCR
oMC2571	GATACACTACATAAAGTGTTCTATC	Forward primer for amplification of tcdE
oMC2572	CAGCTATTCTTATTTGGATAACAC	Reverse primer for amplification of tcdE
oMC2573	ATAAACCTAGGAGGCGTTATGAATATGA	Forward primer for tcdE qPCR
oMC2574	TTATTGCACTTAAACATCCTAATAATGTATCAAA	Reverse primer for tcdE qPCR
tcdRqF	AGCAAGAAATAACTCAGTAGATGATT	Forward primer for tcdR qPCR (43)
tcdRqR	TTATTAAATCTGTTTCTCCCTCTTCA	Reverse primer for tcdR qPCR (43)

^aAll primers were acquired from IDT (Coralville, IA). Restriction enzyme sequences are underlined. qPCR, quantitative PCR.

Sporulation assays.

Sporulation assays were performed as previously documented (62, 63). Briefly, C. difficile was cultured in BHIS medium supplemented with 0.1% taurocholate with 0.2% fructose until mid-exponential phase was reached, and 0.25 ml of the culture was plated onto fresh 70:30 sporulation agar. After 24 h, the cells were removed from the plate and suspended in BHIS medium to an optical density at 600 nm (OD₆₀₀) of ∼1.0. The total number of vegetative cells was determined by immediately plating serial dilutions of the suspended cells onto BHIS plates. To enumerate the total number of spores at the same time, an aliquot of suspended cells was mixed and incubated for 15 min with 28.5% ethanol to eliminate all vegetative cells and subsequently serially diluted in 1× phosphate-buffered saline (PBS) plus 0.1% taurocholate and plated onto BHIS medium supplemented with 0.1% taurocholate. CFU were enumerated after at least 24 h of growth, and the sporulation frequency was calculated as the total number of ethanol-resistant spores divided by the total number of viable cells (spores plus vegetative cells). A spo0A mutant (MC310) was used as a negative sporulation control. Statistical significance was determined using two-tailed Student’s t test comparing each mutant to its isogenic parent.

Phase-contrast microscopy.

At the same time that cells were harvested for sporulation assays, as described above, an aliquot was removed from the anaerobic chamber, pelleted, and suspended in ∼10 μl of the supernatant. As detailed previously (64), the concentration culture was applied to a 0.7% agarose pad on the surface of a slide and imaged using a 100× Ph3 oil immersion objective on a Nikon Eclipse Ci-L microscope with a DS-Fi2 camera.

qRT-PCR analysis.

C. difficile was grown either in TY medium (pH 7.4) or on 70:30 sporulation agar, as described above. Cells were harvested from TY medium at T₄, defined as 4 h after entry into stationary phase, or from 70:30 sporulation agar at H₁₂, defined as 12 h after spotting onto the plates. Cells harvested from TY medium were either mixed with 3 ml ice-cold ethanol-acetone (1:1) or scraped from plates and suspended in 6 ml ice-cold water-ethanol-acetone (3:1.5:1.5) and stored at −80°C. RNA was purified and DNase I treated (Ambion) as previously detailed (19, 64, 65), and cDNA was synthesized using random hexamers (Bioline) (64). Reverse transcription-quantitative PCR (qRT-PCR) was performed in technical triplicate with 50 ng cDNA, using the SensiFAST SYBR and fluorescein kit (Bioline), on a Roche LightCycler 96 instrument for four biological replicates. To confirm the absence of contaminating genomic DNA, cDNA synthesis reaction mixtures containing no reverse transcriptase were included for all samples. Results were calculated using the comparative cycle threshold method (66), normalizing expression to the internal control transcript rpoC. Either two-tailed Student’s t test, to compare the activity in the rstA mutant to that in the parent strain, or one-way analysis of variance (ANOVA) followed by Tukey’s multiple-comparison test was used, as indicated in the figure legends.

Western blot analysis.

Briefly, for Western blot analysis, C. difficile strains were grown in TY medium (pH 7.4) and harvested at H₂₄, defined as 24 h after inoculation (13). Lysates were prepared, and total protein was quantitated using the Pierce Micro bicinchoninic acid (BCA) protein assay kit, as previously described (62). Total protein (3 μg) was separated by electrophoresis on a precast 4 to 15% TGX stain-free gradient gel (Bio-Rad), imaged using the ChemiDoc system (Bio-Rad), and transferred to a 0.45-μm nitrocellulose membrane. Western blot analysis was conducted with mouse anti-TcdA (Novus Biologicals), followed by goat anti-mouse Alexa Fluor 488 (Life Technologies)-conjugated secondary antibody. Imaging and densitometry of full-length TcdA protein were performed with the ChemiDoc system and Image Lab software (Bio-Rad) and analyzed using two-tailed Student’s t test, comparing each mutant to its isogenic parent. A minimum of three biological replicates were analyzed for each strain, and a representative Western blot image is shown.

Enzyme-linked immunosorbent assay.

To quantify C. difficile TcdA and TcdB in the supernatants of C. difficile cultures, aliquots of C. difficile strains grown in TY medium (pH 7.4) were collected at H₂₄ and pelleted, and supernatants diluted in the provided dilution buffer were assayed using a kit for the simultaneous detection of C. difficile toxins A and B from TGCbiomics (catalog no. TGC-E001-1), according to the manufacturer’s instructions. Technical duplicates were averaged and normalized to the OD₆₀₀ of the respective cultures at H₂₄, and the results are provided as the means and standard errors of the means from three biological replicates. Statistical significance was determined using two-tailed Student’s t test, comparing each mutant to its isogenic parent.

Alkaline phosphatase activity assays.

Briefly, for alkaline phosphatase (AP) activity assays, C. difficile strains were grown in TY medium (pH 7.4) and harvested at H₂₄ for the PtcdR fusions or grown on 70:30 sporulation agar and harvested at H₈ for the PrstA fusions (13). AP assays were performed as previously detailed (67), in the absence of chloroform. Technical duplicates were averaged, and the results are provided as the means and standard errors of the means from at least three biological replicates. Either two-tailed Student’s t test was used to compare the activity in the rstA mutant to that in the parent strain or one-way ANOVA followed by Dunnett’s or Tukey’s multiple-comparison test was performed, as indicated in the figure legends.