CodY, a global regulator of gene expression in low G + C Gram-positive bacteria, was found to repress toxin gene expression in Clostridium difficile. Inactivation of the codY gene resulted in derepression of all five genes of the C. difficile pathogenicity locus during exponential growth and stationary phase. CodY was found to bind with high affinity to a DNA fragment containing the promoter region of the tcdR gene, which encodes a sigma factor that permits RNA polymerase to recognize promoters of the two major toxin genes as well as its own promoter. CodY also bound, but with low affinity, to the toxin gene promoters, suggesting that the regulation of toxin gene expression by CodY occurs primarily through direct control of tcdR gene expression. Binding of CodY to the tcdR promoter region was enhanced in the presence of GTP and branched-chain amino acids, suggesting a link between nutrient limitation and the expression of C. difficile toxin genes.
Clostridium difficile, a Gram-positive, anaerobic, spore-forming bacterium, is the principal agent of antibiotic-associated diarrhoea (AAD). In some cases, AAD leads to pseudomembranous colitis, a potentially lethal disease (Kelly and LaMont, 1998). In the typical patient, antibiotic therapy leads to the disruption of the intestinal microflora, allowing colonization of the intestinal tract by C. difficile (Kelly and LaMont, 1998). The primary virulence factors of C. difficile are thought to be two large protein toxins, TcdA (308 kDa) and TcdB (270 kDa) (Eichel-Streiber et al., 1996). Both toxins act by glycosylating members of the Rho family of small GTPases in host cells (Just et al., 1995a,b). TcdA is known to function as an enterotoxin, causing gastrointestinal damage (Lyerly et al., 1985b). TcdB is a potent cytotoxin in vitro, but its role in C. difficile-associated disease (CDAD) is not clear (Voth and Ballard, 2005). A recent study showed that TcdB has cardiotoxic activity in a zebrafish embryo model, suggesting that TcdB may have systemic effects in the progression of CDAD (Hamm et al., 2006).
Clostridium difficile toxins A and B are encoded by the genes tcdA (2711 codons) and tcdB (2367 codons), which lie within a 19.6 kb pathogenicity locus (Hammond and Johnson, 1995; Braun et al., 1996). The toxin genes are separated by a small open reading frame, tcdE, thought to encode a holin, that is, a protein whose activity would potentially allow the release of the toxins from the cell (Tan et al., 2001). The tcdR gene, which lies upstream of tcdB in the pathogenicity locus, encodes an RNA polymerase sigma factor that directs transcription from the toxin promoters and from its own promoter (Mani and Dupuy, 2001; Mani et al., 2002). Transcriptional analysis of the pathogenicity locus indicates that the four upstream genes of the locus (tcdR-tcdB-tcdE-tcdA) can be cotranscribed, presumably from the tcdR promoter, and that the tcdA and tcdB genes are also transcribed from gene-specific promoters (Hammond et al., 1997; Hundsberger et al., 1997; Dupuy and Sonenshein, 1998). Expression of these four genes is induced as C. difficile broth cultures enter the stationary phase of growth (Hundsberger et al., 1997; Dupuy and Sonenshein, 1998). In contrast, the furthest downstream gene of the locus, the convergently transcribed tcdC gene, is maximally expressed during exponential phase; expression greatly diminishes as cells enter stationary phase (Hundsberger et al., 1997). This finding led to the speculation that TcdC is a negative regulator of toxin gene expression. Recent work of Dupuy and colleagues (Matamouros et al., 2007) has shown that TcdC is an antagonist of TcdR. Interestingly, mutations in the tcdC gene have been found in high-toxin-producing C. difficile isolates associated with recent epidemics (Warny et al., 2005; Curry et al., 2007).
In the work described here, we have assessed the role of the global regulator CodY in toxin gene regulation. In Bacillus subtilis, the bacterium in which it was first discovered (Slack et al., 1995), CodY represses expression of over 100 genes during rapid growth in nutrient-rich conditions (Molle et al., 2003). These genes are subsequently induced when nutrients become limiting and cells enter stationary phase. In a B. subtilis codY-null mutant, CodY target genes are derepressed during exponential phase (Molle et al., 2003). CodY appears to act by monitoring the nutrient sufficiency of the environment, repressing genes that are not needed in nutrient-rich conditions, and releasing this repression when nutrient conditions become limiting. In accord with this proposed nutrient monitoring mechanism, GTP and the branched-chain amino acids (BCAAs) isoleucine and valine act as co-repressors by increasing the affinity of CodY for its target DNA (Ratnayake-Lecamwasam et al., 2001; Shivers and Sonenshein, 2004). C. difficile encodes a protein of 261 amino acids with strong similarity to B. subtilis CodY, as do many other low G + C Gram-positive bacteria (e.g. Staphylococcus, Streptococcus, Listeria, Lactococcus and Enterococcus) (Sonenshein, 2005). In Lactococcus lactis, CodY is a repressor of genes for amino acid biosynthesis, peptide transport and peptidases (Chambellon and Yvon, 2003; den Hengst et al., 2005; Guedon et al., 2005). CodY also appears to be involved in virulence regulation in pathogenic bacteria belonging to this group (Kloosterman et al., 2006; Malke et al., 2006; Bennett et al., 2007).
We report here that CodY is a repressor of toxin gene expression in C. difficile. A codY mutant was strongly derepressed for toxin gene expression during exponential growth phase. Toxin gene expression was also higher in the codY mutant compared with wild-type during stationary phase. CodY was found to interact strongly with the tcdR promoter region in vitro, suggesting that regulation of C. difficile toxin synthesis by CodY occurs primarily through this interaction.
Construction of a C. difficile codY mutant
To test whether CodY plays a regulatory role in toxin synthesis, a C. difficile codY-null mutant was constructed using a strategy reported by O'Connor et al. (2006). These authors discovered that a mobilizable Escherichia coli–Clostridium perfringens shuttle vector, pJIR1456 (Lyras and Rood, 1998), requires continuous selection to be maintained in C. difficile after transfer from E. coli by conjugation. When a DNA fragment with homology to the C. difficile chromosome was inserted into pJIR1456 and the engineered plasmid was introduced into C. difficile, the plasmid sequence could be maintained by integration into the chromosome by a homology-driven single cross-over event (O'Connor et al., 2006). They utilized this finding to create C. difficile insertion mutants with defects in two putative response regulator genes. To follow this strategy, an internal region of the codY gene (620 bp) from C. difficile strain 630, corresponding to the sequence starting immediately downstream of the ATG start codon and ending just upstream of the region encoding the recognition helix of the conserved helix–turn–helix (HTH) motif responsible for CodY binding to DNA (Joseph et al., 2005; Levdikov et al., 2006), was cloned in pJIR1456 (Lyras and Rood, 1998). The resulting plasmid, pSD21, which confers thiamphenicol resistance, was transferred to C. difficile strain JIR8094 (an erythromycin-sensitive derivative of strain 630) (O'Connor et al., 2006) by conjugation with E. coli. Four independent C. difficile thiamphenicol-resistant transconjugants were isolated. To allow for the loss of non-integrated pSD21, the transconjugants were passaged three times in BHIS medium lacking thiamphenicol and then plated on BHIS containing thiamphenicol to select, in principle, for clones in which the plasmid had integrated. Colonies that appeared were subcultured in BHIS or TY medium with thiamphenicol and lysates were prepared for polymerase chain reaction (PCR) screening. The PCR screen consisted of four primer pairs (Fig. 1A): one primer pair (A/B) designed to amplify the intact chromosomal codY gene; two primer pairs (A/D and B/C) designed to produce PCR products only if pSD21 had integrated into the chromosome by single cross-over at the codY locus; and one primer pair (C/D) designed to amplify the internal codY fragment in pSD21. The C. difficile clones displayed two different types of PCR profiles. One type, designated JIR8094::pSD21 and representing the progeny of one of the four original transconjugants, yielded a weak product for the intact chromosomal codY gene compared with the parent strain JIR8094 and strong products for the pSD21 integration event (Fig. 1B). The progeny of the three other transconjugants were designated JIR8094(pSD21). These strains yielded a strong PCR product for the uninterrupted chromosomal codY gene, similar to the parent strain JIR8094, and weak products for the pSD21 integration event compared with JIR8094::pSD21 (Fig. 1B). The JIR8094(pSD21) strains also yielded a stronger PCR product for the internal codY fragment in pSD21 than did JIR8094::pSD21 (Fig. 1B). The PCR screening results indicated that both types of strains are mixed populations. In JIR8094::pSD21, most of the cells contained integrated pSD21, while in the JIR8094(pSD21) strains, most of the plasmid was extrachromosomal. Further passaging of JIR8094(pSD21) strains in BHIS without thiamphenicol did not yield a stably integrated pSD21.
To determine whether the C. difficile strains were altered in accumulation of CodY protein, sonicated lysates of overnight cultures were subjected to Western blotting using antibodies raised against B. subtilis CodY (Ratnayake-Lecamwasam et al., 2001). Strain JIR8094::pSD21 did not contain any protein that reacted with the antibodies (Fig. 2). Integration of pSD21 into the codY gene should create a construct in which the codY gene that is driven by its native promoter is truncated at the 3′ end (Fig. 1A). This truncation would lead to expression of a CodY protein of 208 amino acids that lacks the recognition helix of the conserved C-terminal HTH domain. In other work, we have found that our polyclonal antibody reacts primarily with C-terminal epitopes of CodY (unpublished results). Thus, the truncated CodY protein of strain JIR8094::pSD21 would not be detectable even if it were synthesized and stable. Unlike strain JIR8094::pSD21, strain JIR8094(pSD21) synthesized full-length CodY at levels similar to that of the parent strain JIR8094 (Fig. 2). As pSD21 did not stably integrate into the codY gene of strain JIR8094(pSD21), CodY synthesis was not significantly affected.
CodY negatively regulates C. difficile pathogenicity locus gene expression
To determine whether a codY mutation alters pathogenicity locus gene expression, semiquantitative reverse transcription PCR (RT-PCR) analysis was performed on RNA isolated from cells grown in TY medium during mid-exponential phase and early stationary phase. Transcript levels for the tcdR, tcdB, tcdE and tcdA genes were found to be significantly higher in the codY mutant (JIR8094::pSD21) compared with the strains that expressed wild-type levels of CodY [JIR8094 and JIR8094(pSD21)] at both time points (Fig. 3). Expression of tcdC was also derepressed in the codY mutant compared with the other strains, but not to the same extent as the four other genes of the pathogenicity locus. To monitor the expression of a gene located outside of the C. difficile pathogenicity locus, the ackA gene (encoding acetate kinase) was used as a control for semiquantitative RT-PCR analysis. A codY mutation had no effect on ackA expression in mid-exponential phase, but ackA expression was greater in the codY mutant compared with the other strains in early stationary phase (Fig. 3). When grown in TY medium, the codY mutant displayed a growth defect upon entry into stationary phase (Fig. 4). The codY mutant grew at a similar rate compared to the strains that expressed wild-type levels of CodY through mid-exponential growth, but while the other strains continued growing exponentially, the codY mutant abruptly entered stationary phase (Fig. 4). As a result, the cell density of the codY mutant culture in stationary phase was much lower than that of strains that expressed wild-type levels of CodY. The differences observed for ackA expression in early stationary phase may be due to indirect effects of the codY mutation on growth. We do not know the cause of the growth defect, but note that the defect was suppressed when 1% glucose was added to TY medium (data not shown).
To verify that the phenotype of the codY mutant is due to the codY mutation alone, we would have preferred to complement the phenotype with a full-length codY gene in trans. However, complementation of the codY mutant strain JIR8094::pSD21 would be technically difficult to achieve due to the lethality of CodY to E. coli. Cloning the complete codY gene in E. coli typically results in the accumulation of mutations within the gene unless expression of the gene is suppressed. To avoid these complications, an alternative strategy was used to confirm that integration of pSD21 into the codY gene is responsible for the codY mutant phenotype. Strain JIR8094::pSD21 was passaged three times in BHIS without thiamphenicol and plated on BHIS plates. The resulting colonies were screened for thiamphenicol sensitivity, and a thiamphenicol-sensitive revertant was designated as JIR8094 RVT. A JIR8094 RVT culture lysate was screened by PCR, as described earlier, and was found to display a PCR profile similar to that of the parent strain JIR8094, indicating that pSD21 had excised from the chromosome of strain JIR8094::pSD21 and was lost in the absence of selection (Fig. 1B). In accordance with this finding, strain JIR8094 RVT produced wild-type levels of CodY protein as determined by immunoblotting (Fig. 2). Strain JIR8094 RVT also displayed a gene expression profile similar to that of strains JIR8094 and JIR8094(pSD21) for the genes tested by RT-PCR (Fig. 3), suggesting that disruption of codY by integration of pSD21 in strain JIR8094::pSD21 was responsible for the strain's mutant phenotype.
To confirm that the observed mutant phenotype was not due to polar effects of plasmid integration on the expression of genes downstream of codY, we constructed a null mutant of CD1276, the gene immediately downstream of codY, by plasmid integration within its coding sequence. CD1276 encodes a putative ATPase. An internal region of the CD1276 gene (647 bp) from strain 630, corresponding to the sequence starting 15 bp downstream of the ATG start codon and ending 632 bp upstream of the TAA stop codon, was cloned in pJIR1456. The resulting plasmid, pSD29, was transferred to strain JIR8094 by conjugation with E. coli. Multiple transconjugants were obtained, and clones in which the plasmid had integrated were selected for and screened by PCR as described above for the codY mutant. One clone, designated JIR8094::pSD29, yielded a weak product for the intact chromosomal CD1276 gene (primer pair OSD155/OSD156) compared with the parent strain JIR8094 and strong products for the pSD29 integration event (primer pairs OSD155/D and OSD156/C) (Fig. 5A). This result suggests that in strain JIR8094::pSD29, as in the codY mutant (JIR8094::pSD21), plasmid integration led to significant disruption of the targeted gene. Strain JIR8094::pSD29, as expected, retained the wild-type codY gene, as shown by the strong PCR product obtained using primer pair A/B (Fig. 5A). To determine the effect of disruption of the CD1276 gene on gene expression, semiquantitative RT-PCR analysis was performed on RNA isolated from cells grown in TY medium during mid-exponential phase. As expected, transcript levels for the intact CD1276 gene were found to be significantly lower in strain JIR8094::pSD29 compared with parent strain JIR8094 (Fig. 5B). CD1276 transcript levels in the codY mutant were also found to be lower than in strain JIR8094 (Fig. 5B), indicating that integration of pSD21 at the codY locus does have a negative polar effect on CD1276 gene expression. However, tcdR and tcdA expression was not derepressed in strain JIR8094::pSD29 compared with strain JIR8094 (Fig. 5B), indicating that disruption of the codY gene, and not reduced expression of genes downstream of codY, is responsible for derepression of pathogenicity locus gene expression.
TcdA accumulation in the codY mutant
Immunoblotting was used to determine the levels of TcdA protein in the culture fluids of C. difficile strains grown overnight in TY medium. In accordance with the RT-PCR results, the TcdA level in the supernatant fluid of the codY mutant culture was significantly higher than that for the strains that expressed wild-type levels of CodY (Fig. 6). These results provide further evidence that CodY is a repressor of toxin synthesis in C. difficile.
Previous studies have shown that expression of tcdR, tcdA and tcdB is strongly repressed when glucose is present in the growth medium (Dupuy and Sonenshein, 1998; Mani et al., 2002). When strain JIR8094::pSD21 was grown in TY medium supplemented with 1% glucose, TcdA accumulation was significantly lower in the culture supernatant for this strain compared with the same strain grown in TY medium alone (Fig. 6). These results suggest that glucose inhibition of toxin synthesis is at least partially independent of the effect of CodY as a repressor.
CodY interacts with the tcdR promoter region
To determine whether CodY plays a direct role in the repression of C. difficile toxin synthesis, in vitro DNA binding experiments were performed. The codY gene from C. difficile strain 630, with a C-terminal His6 tag, was overexpressed after cloning in E. coli and purified by metal affinity chromatography. In gel mobility shift assays, C. difficile CodY bound to a 542 bp DNA segment that includes the tcdR promoter region, and this binding was enhanced in the presence of added GTP or a mixture of BCAAs or both (Fig. 7). The presence of both effectors had an additive effect on binding compared with each effector alone. The apparent KD for binding (corresponding to the concentration of CodY that shifted 50% of the DNA) in the presence of both effectors was estimated at 20–50 nm CodY. As GTP and BCAAs enhanced binding, it appears that C. difficile and B. subtilis CodY proteins respond to similar effectors (Ratnayake-Lecamwasam et al., 2001; Shivers and Sonenshein, 2004). C. difficile CodY was also shown to interact with high affinity with the B. subtilis ilvB promoter region, a known target of CodY in that organism (data not shown). Only weak interactions (apparent KD > 200 nM) were observed, however, between C. difficile CodY and the tcdA and tcdB promoter regions even in the presence of GTP and BCAAs (data not shown). These results suggest that CodY regulates toxin synthesis primarily through interaction with the tcdR promoter region.
To identify the specific binding sites for CodY, a DNase I footprinting assay was used to analyse the interaction of CodY with the tcdR promoter region. In the presence of GTP and BCAAs, three regions (I–III) of DNA within the tcdR promoter region were protected by CodY from DNase I digestion (Fig. 8). Region II, which spans the sequence from positions −281 to −309 relative to the tcdR ATG start codon, showed the highest affinity for CodY, with protection beginning to appear at 50 nM CodY. Protection of region I (positions −348 to −382) and region III (−40 to −58) started at 100 nM CodY. It is difficult to predict with confidence the effect of binding of CodY to these sites, because we have been unable to reliably map the transcription start site of the tcdR gene. Region I is of particular interest because it overlaps a putative TcdR-dependent promoter predicted by Mani et al. (2002).
In this study, we show that CodY is a negative regulator of toxin gene expression in C. difficile. The key to this discovery was our ability to construct a C. difficile codY mutant using a method recently reported by O'Connor et al. (2006). An internal region of the codY gene was inserted into a mobilizable E. coli–C. perfringens shuttle vector that is unstable in C. difficile. Introduction of this plasmid into C. difficile resulted in two types of strains, one in which the plasmid integrated rather stably at the codY locus, and a second type in which the plasmid remained primarily extrachromosomal. Relatively stable integration of the plasmid resulted in a codY mutant phenotype, as expected, but the strains in which the plasmid did not stably integrate synthesized wild-type levels of CodY. This second type of strain was not found by O'Connor et al. (2006). They were able to inactivate genes for two putative response regulators, rgaR and rgbR, and cited unpublished results in which they had also constructed tcdA and tcdB mutants using their strategy. O'Connor et al. (2006) also found that integration of their suicide vectors was stable in the absence of selective pressure. We made similar observations when constructing the CD1276 mutant for this study. For the C. difficile codY mutant, however, stable integration of pSD21 could be maintained only when thiamphenicol was present in the medium. As the mutant grows more slowly than the wild-type, rare recombination events that restore the wild-type sequence would be selected for in the absence of thiamphenicol. We speculate that initial integration of pSD21 into the chromosome of the codY mutant occurred early in the conjugation process between E. coli and C. difficile, possibly at a point at which the codY mutation did not have a negative effect on the stability of the integrated state. A second possibility that may explain the lack of stable integration in all strains is that the recombination frequency at the codY locus may be very low, resulting in the low frequency of plasmid integration observed. A third possibility is that the codY mutant acquired a secondary mutation that stabilized plasmid integration. The finding that the codY mutant could revert back to wild-type in terms of CodY accumulation and toxin expression indicates that, if this secondary mutation did occur, it likely had no effect on expression of pathogenicity locus genes.
Our results revealed that all the genes of the C. difficile pathogenicity locus are derepressed in the codY mutant during exponential growth and stationary phase in TY medium. The extent of derepression was greatest for the first four genes of the pathogenicity locus, tcdR, tcdB, tcdE and tcdA. Gel mobility shift and DNase I footprinting experiments showed that CodY interacts with high affinity with the tcdR promoter region and only weakly with the tcdA and tcdB promoter regions. Thus, CodY is likely to act in vivo as a direct repressor of tcdR gene expression, resulting in downstream effects on tcdB, tcdE and tcdA gene expression. The tcdA and tcdB genes have TcdR-dependent promoters; the tcdE gene does not appear to have a TcdR-dependent promoter, but is expressed as part of a polycistronic message initiating at the tcdR or tcdB promoters (Hammond et al., 1997). It was surprising to find that tcdC expression was also derepressed in the codY mutant. In gel mobility shift experiments, we have observed high-affinity binding of CodY to a DNA fragment that likely contains the tcdC promoter region (data not shown). If CodY is a direct regulator of tcdC expression, it is unclear why CodY would repress synthesis of both the sigma factor for toxin gene expression and its antagonist.
We were originally drawn to the possibility that CodY might be a regulator of toxin gene expression by the finding for C. difficile strain VPI 10463 that toxin gene expression is significantly induced upon entry into stationary phase (Hundsberger et al., 1997; Dupuy and Sonenshein, 1998). While working with strain 630 (and its erythromycin-sensitive derivative), which, unlike VPI 10463, can act as a recipient in conjugation, we discovered that this strain's toxin expression profile is different from that of strain VPI 10463. Strain 630 synthesizes significantly fewer tcdR transcripts and produces significantly less TcdA in TY medium cultures than does VPI 10463 (data not shown). In addition, toxin gene expression is not significantly induced upon entry into stationary phase in strain 630 (data not shown). However, a codY mutation in strain 630 resulted in derepression of toxin synthesis during both exponential growth and stationary phase. Thus, CodY is a major factor that limits toxin gene expression in strain 630. Although it would be very interesting to test the effect of a codY mutation in a high-toxin-producing strain, such as VPI 10463, there have been no reports of successful introduction of DNA into such strains.
The binding of CodY to the tcdR promoter region was enhanced in the presence of GTP and BCAAs. It appears that GTP and the BCAAs are co-repressors of C. difficile CodY, as they are for B. subtilis CodY (Ratnayake-Lecamwasam et al., 2001; Shivers and Sonenshein, 2004). In fact, it is known that a nine-amino-acid mixture that contains the BCAAs suppresses toxin expression when added to peptone–yeast extract medium (Karlsson et al., 1999). Thus, regulation by CodY may provide a nutritional link to the pathogenicity of C. difficile. Upon colonization of the intestinal tract, the bacteria may experience conditions in which certain nutrients become limiting. Under these conditions, repression of toxin expression by CodY may be relieved, resulting in increased synthesis of toxins A and B, lysis of intestinal epithelial cells, and liberation of potential nutrients.
The discovery that CodY is a repressor of toxin gene expression may aid in the development of therapeutic tools to combat C. difficile-associated disease. It is possible that a compound could be identified that acts as a co-repressor of CodY and keeps CodY in its high-affinity DNA binding conformation. This compound could be potentially used as a therapeutic to keep toxin expression low in a colonizing C. difficile population. In addition, disruption of the ability to sense excess or limitation of nutrients might compromise the ability of C. difficile to proliferate in the intestine.
Bacterial strains and growth conditions
Bacterial strains and plasmids used in this study are described in Table 1. C. difficile strains were grown in BHIS medium (Smith et al., 1981), supplemented with 0.1% l-cysteine HCl and 0.375% glucose, in TY medium (Dupuy and Sonenshein, 1998), or in TY medium supplemented with 1% glucose (TYG). Media for growth of C. difficile were supplemented with 8 µg cefoxitin ml−1, 250 µg d-cycloserine ml−1, or 10 µg thiamphenicol ml−1 as needed. C. difficile strains were grown in an atmosphere of 10% H2, 5% CO2 and 85% N2 in an anaerobic chamber (Coy Laboratory Products) at 37°C. E. coli strains were grown at 37°C in L or BHI medium (Difco) supplemented with 30 µg chloramphenicol ml−1 or 50 µg ampicillin ml−1, as needed.
C. difficile codY with 6× histidine-tag cloned in pBAD30, ApR
E. coli cloning vector, ApR
540 bp DNA fragment containing the tcdR promoter region cloned into pCR II, ApR
Construction of C. difficile codY and CD1276 mutants
Internal regions of codY (620 bp) or CD1276 (647 bp) were amplified by PCR using primers OSD56 and OSD57 or OSD151 and OSD152 (Table 2) respectively, and chromosomal DNA of C. difficile strain 630 as template (Sebaihia et al., 2006). The PCR products were digested with SphI and XbaI and cloned in the E. coli–C. perfringens shuttle vector pJIR1456 (Lyras and Rood, 1998), resulting in plasmids pSD21 and pSD29. Plasmids pSD21 and pSD29 were introduced by transformation into E. coli strain HB101(pRK24). pRK24 is a derivative of the broad-host-range plasmid RP4 that is capable of mobilizing IncP oriT plasmids, such as pJIR1456. The resulting strains were used as donors to transfer pSD21 and pSD29 to C. difficile strain JIR8094 by conjugation as follows. A total of 4 ml of an exponential-phase culture of the E. coli donor strain grown aerobically in BHI supplemented with ampicillin and chloramphenicol was centrifuged, and the pellet was resuspended in 400 µl of a late-exponential-phase culture of JIR8094 grown anaerobically in BHIS medium. The suspension was spotted on nitrocellulose filters (Millipore Corporation) on a BHIS agar plate and incubated overnight at 37°C under anaerobic conditions. The growth on the plate was then resuspended in 2 ml of BHIS, and the suspension was spread on 10 BHIS plates supplemented with cefoxitin, d-cycloserine and thiamphenicol, and incubated anaerobically for a minimum of 48 h. Putative transconjugant colonies were substreaked on the same medium for isolation. To confirm the presence of pSD21 or pSD29 in the putative transconjugants, cell lysates (Furrer et al., 1991) of cultures grown anaerobically overnight at 37°C in BHIS supplemented with thiamphenicol were screened by PCR using primers OSD17 and OSD18 to amplify the catP gene (Table 2). Confirmed transconjugants were passaged three times in 5 ml of BHIS broth lacking thiamphenicol and then plated on BHIS containing thiamphenicol. Colonies that appeared were subcultured in BHIS or TY with thiamphenicol and lysates were prepared for PCR screening using primer pairs A/B, C/D, B/C and A/D (pSD21 integration), or OSD155/OSD156, OSD155/D and OSD156/C (pSD29 integration) (Table 2) as described in Results.
Table 2. Oligonucleotide primers.
The first 22 and 23 nucleotides of OSD17 and OSD18 respectively are not homologous to the pSD21 DNA sequence.
Overnight cultures of C. difficile grown in TYG were diluted to an OD600 of 0.05 in TY or TYG medium and incubated anaerobically at 37°C. Samples were removed for RNA isolation at an OD600 of 0.4 (exponential phase) and early during stationary phase (3 h after the exponential-phase time point), diluted in an equal volume of ice-cold 1:1 acetone : ethanol and stored at −80°C. The cell suspensions were centrifuged at 4°C; the pellets were air-dried, washed twice with 500 µl of TE (10 mM Tris, 1 mM EDTA, pH 7.6), and resuspended in 1 ml of Buffer RLT from the Qiagen RNeasy kit. Silica-glass beads (0.1 mm) were added and cells were disrupted using a Mini-BeadBeater (BioSpecs Products). Silica beads were removed by centrifugation, and total RNA was isolated from the supernatants using the Qiagen RNeasy kit. DNA contaminating the RNA samples was removed using the Ambion TURBO DNA-free kit. Independent samples of each culture were used to provide DNA templates to verify by PCR the maintenance of the structure of the codY locus.
Reverse transcription was performed using SuperScript II reverse transcriptase (Invitrogen) with total RNA (50 or 200 ng for tcdR, 50 ng for all other genes) and 2 pmol of gene-specific primer (Table 2) according to the manufacturer's protocol. Two microlitres of samples of the reverse transcription reactions was used as templates for PCR with gene-specific primers (Table 2). PCR reactions were performed using 22–30 cycles of denaturation (94°C, 30 s), annealing (48°C, 1 min) and extension (72°C, 1 min). For each gene tested, the amount of RNA used and the number of cycles chosen for the experiment of Figs 3 and 5B were below saturation of either the reverse transcriptions or the PCRs. As a control for chromosomal DNA contamination, RNA was used directly for PCR amplification.
For CodY detection, 5 ml of samples of the same cultures used for RNA extraction was centrifuged at 4°C, and the cells were resuspended in 1 ml of 50 mM Tris-HCl (pH 7.5) containing 1 mM phenylmethylsulfonyl fluoride and subjected to three 30 s cycles of sonication with 30 s rests between cycles. Unbroken cells and cell debris were removed by centrifugation at 4°C, and the total protein concentration in the supernatant fluid was determined using the Bio-Rad protein assay reagent. Five micrograms of total soluble protein was subjected to electrophoresis on 12% SDS-polyacrylamide gels and electroblotted onto Immobilon-P membranes (Millipore). CodY was detected on the membranes using rabbit polyclonal antibody to B. subtilis CodY (Ratnayake-Lecamwasam et al., 2001) and goat anti-rabbit IgG conjugated with alkaline phosphatase.
For TcdA detection, 5 ml of cultures of C. difficile was centrifuged at 4°C, and the protein concentration in the cell-free supernatant fluid was determined. Total supernatant proteins (2.2 µg) were separated by electrophoresis on 6% SDS-polyacrylamide gels and electroblotted. TcdA was detected using mouse monoclonal antibody PCG-4 (Lyerly et al., 1985a) in conjunction with PhoA-conjugated goat anti-mouse IgG.
Purification of CodY and gel mobility shift assays
To create a C-terminal, 6× histidine-tagged version of CodY (CodY-His6), the codY gene of C. difficile strain 630 was amplified by PCR using primer A, which includes an EcoRI site, the codY ribosome binding site, and the codY start codon, and primer B, which includes the last 10 codons of codY, an additional 6 histidine codons, a stop codon, and an SphI site. The PCR product was digested with EcoRI and SphI and cloned downstream of the araBAD promoter in plasmid pBAD30 (Guzman et al., 1995), creating pEAV1. E. coli strain KS272 (Strauch and Beckwith, 1988) carrying pEAV1 was grown in L broth supplemented with ampicillin at 37°C until the OD600 reached 0.6, arabinose was then added to a final concentration of 0.2%, and growth was continued for 4–5 h. Cells were pelleted by centrifugation and broken by sonication, and CodY-His6 was purified by cobalt affinity chromatography, as previously described (Kim et al., 2003), using 75 mM imidazole for elution.
A 540 bp DNA fragment containing the tcdR gene promoter region was amplified by PCR from C. difficile strain VPI 10463 chromosomal DNA using primers OBD15 and OBD16 (Table 2) and cloned in plasmid pCR II (Invitrogen), yielding pCD24. For gel mobility shift assays, the tcdR promoter region was amplified from pCD24 template DNA using primers OBD15 and OBD16, agarose gel-purified, and labelled with [γ-32P]-ATP using T4 polynucleotide kinase (Invitrogen) as described by the manufacturer. Labelled DNA was mixed with increasing amounts of CodY protein in 10 µl of reactions that contained 20 mM Tris-HCl (pH 8.0), 50 mM sodium glutamate, 10 mM MgCl2, 5 mM EDTA, 0.05% (v/v) Nonidet P-40 (Igepal; Sigma Chemical), 5% (v/v) glycerol, and 250 ng of calf thymus DNA. Where indicated, 10 mM each of isoleucine, leucine and valine (BCAA), or 2 mM GTP or both were added. After incubation for 30 min at room temperature, binding reactions were loaded on a 12% non-denaturing polyacrylamide gel prepared in Tris-glycine buffer, and electrophoresis was carried out in 35 mM HEPES-43 mM imidazole buffer (pH 7.4), as described previously (Shivers and Sonenshein, 2004). When 10 mM BCAAs were present in the binding reaction, BCAAs at the same concentration were also added to the electrophoresis buffer. Gels were dried under vacuum and exposed to a phosphorimager screen before detection with a Molecular Dynamics Storm 860 Imager. Gel images were analysed using ImageQuant version 1.2 Macintosh software.
DNase I footprinting assays
Primer OBD16 was radioactively labelled with [γ-32P]-ATP using T4 polynucleotide kinase (Invitrogen) as described previously (Kim et al., 2002), and used in conjunction with OBD15 to amplify the tcdR promoter region by PCR using pCD24 as template. Binding reactions were prepared as described for the gel mobility shift assays except that the reaction volume was 20 µl. All reactions contained 10 mM each of isoleucine, leucine and valine, and 2 mM GTP. After incubation for 30 min at room temperature, 6 mM MgCl2, 6 mM CaCl2 and 0.125 U of RQ1 DNase (Promega) were added to each reaction, and incubation was continued for 1 min at room temperature. The reactions were stopped by the addition of 1 µl of 0.5 M EDTA and transferred to an ice bath. Samples were extracted with phenol-chloroform, and the DNA was ethanol-precipitated and resuspended in 4 µl sequencing gel loading buffer (Sambrook et al., 1989). The samples were heated to 80°C for 5 min and then subjected to electrophoresis in a 6% urea-polyacrylamide gel. Dideoxy sequencing reactions were performed with primer OBD16 and pCD24 template DNA using a Sequenase kit (US Biochemical) and [α-35S]-dATP.
We thank J. O'Connor, G. Carter, D. Lyras, J. Rood, S. Matamouros, B. Dupuy and B. Belitsky for many helpful suggestions and discussions during the course of this work; B. Belitsky, B. Dupuy and J. Sorg for helpful criticism of the manuscript; B. Dupuy for strain HB101(pRK24); and J. Rood for strain JIR8094 and plasmid pJIR1456. This work was supported by a research grant (R01 AI057637) from the US Public Health Service. S.S.D. was a National Research Service Award postdoctoral trainee (T32 DK007542).