CdtR regulates production of CDT, TcdA and TcdB in two epidemic, ribotype 027 strains of C. difficile

To investigate the role of CdtR in the regulation of CDT production in the epidemic ribotype 027 strains, we constructed two independent cdtR mutants in the Canadian isolate M7404 and a cdtR mutant in the UK isolate R20291 and confirmed their genotype by Southern hybridisation (S1 Fig). Western blot analysis showed that the cdtR mutants produced less CDTa and CDTb compared to the wild type and that complementation with cdtR in trans resulted in over-expression of both toxin subunits (Fig 2A and 2B). Consistent with these results, ADP-ribosyltransferase assays demonstrated that the cdtR mutants had significantly reduced levels of CDT activity compared to the wild type, while the complemented cdtR mutants showed CDT activity greater than that of the wild type strain (Fig 2D and 2E). Overall, these data show, for the first time, that CdtR is important for regulating CDT production in epidemic ribotype 027 C. difficile strains.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Analysis of CDT production.

a–c, Western immunoblot using CDTa-specific and cross-reactive Ib-specific antibodies and precipitated supernatants from the strains indicated. CD37 (non-toxigenic), V = vector control, R+ = cdtR complemented. The arrows indicate the 48 kDa CDTa and 99 kDa CDTb proteins. d–f, CDT activity assessed by ADP-ribosyltransferase assay. Samples were separated by SDS-PAGE and biotinylated (ADP-ribosylated) actin detected by HRP-streptavidin. Relative CDT activity was assessed by densitometry compared to the non-toxigenic control strain CD37. A = actin, A-A = ADP-ribosylated actin. Data represent the mean ± SEM (n = 3). *, p ≤ 0.05.

https://doi.org/10.1371/journal.ppat.1005758.g002

Unexpectedly, Western blots performed using TcdA- and TcdB-specific antibodies showed that the cdtR mutants produced less TcdA and TcdB than the wild type, while the complemented cdtR derivatives expressed high levels of both toxins (Fig 3A and 3C). Cytotoxicity assays were performed using HT29 and Vero cells to measure the activity of TcdA and TcdB, respectively, in the culture supernatants from the isogenic panel of M7404 and R20291 strains. The TcdA and TcdB activities of all of strains increased over time and, consistent with the Western blot results, showed lower activity in supernatants from the cdtR mutants compared to the wild type, confirming that the cdtR mutants were less cytotoxic in vitro (Fig 3B and 3D). TcdA and TcdB activity of the complemented cdtR mutants was consistently higher than the cdtR mutants and the wild type across all time points (Fig 3B and 3D). These data show that CdtR regulates the production of TcdA and TcdB in both M7404 and R20291, which is the first demonstration of a common regulator modulating the expression of all three toxins in C. difficile.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Analysis of TcdA and TcdB production.

a, c, e, Western immunoblot using TcdA-specific and TcdB-specific antibodies with precipitated supernatant from the strains indicated. CD37 (non-toxigenic), V = vector control, R+ = cdtR complemented. Arrows indicate the 308 kDa TcdA and 270 kDa TcdB proteins. Supernatants were collected at 12, 24, 48 and 72 hours post inoculation and assayed by doubling dilution cytotoxicity assays. b, The panel of M7404 strains were assayed using HT29 cells and Vero cells. d, R20291 panel of strains assayed using HT29 cells and Vero cells. Data represent the mean ± SEM (n = 3–5). *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.

https://doi.org/10.1371/journal.ppat.1005758.g003

CdtR positively regulates expression of genes from the pathogenicity locus (PaLoc) of C. difficile

To investigate the molecular mechanism of regulation, we determined if CdtR controlled toxin production at the transcriptional level. Using the isogenic panel of M7404 strains, reverse-transcription droplet digital PCR (RT-ddPCR) analysis was employed to quantitatively compare the level of expression of each of the toxin encoding genes (tcdA, tcdB, cdtA) and the PaLoc-encoded toxin regulators (tcdR, tcdC). The relative transcription of all three toxin genes and tcdR was significantly decreased in both cdtR mutants compared to the wild type (Fig 4A–4D). Over-expression of cdtR in the complemented strains resulted in a dramatic over-expression of tcdA, tcdB, cdtA and tcdR (Fig 4A–4D). Although TcdC is predicted to be non-functional in the 027 strains [12], we analysed the expression of tcdC and found it to be similar in the isogenic panel of strains (Fig 4E). Previous work in strain 630Δerm, a derivative of the non-027 historical strain 630 which belongs to ribotype 012, showed that the flagella synthesis regulator, SigD, is an important regulator of TcdA and TcdB production via the regulation of tcdR expression [24]. Analysis of our isogenic panel of M7404 strains showed no change in sigD transcription, suggesting that CdtR does not influence the expression of sigD in this strain background and that therefore the modulation of tcdA and tcdB expression does not occur via SigD (Fig 4F). Collectively, these results indicate that regulation of all three toxins by CdtR in M7404 occurs at a transcriptional level and that the regulation of tcdA and tcdB occurs via the upregulation of tcdR transcription.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Transcriptional analysis of M7404 cdtR mutant and complemented strains compared to wild-type.

RNA was isolated from strains for analysis of (a) tcdA, (b) tcdB, (c) cdtA, (d) tcdR, (e) tcdC and (f) sigD expression. Levels of gene expression were normalised to rpoA. Data represent the mean normalised gene expression ± SEM from five independent biological replicates. *, p ≤ 0.05; **, p ≤ 0.01; ***, p ≤ 0.001.

https://doi.org/10.1371/journal.ppat.1005758.g004

CdtR is important for modulating C. difficile virulence

Our observation that CdtR increases toxin production in ribotype 027 strains has important implications for the virulence capacity of these clinically important strains. To determine if the modulatory effect of CdtR on toxin production influences C. difficile disease, we examined whether cdtR inactivation altered virulence in our mouse infection model [2]. It was previously shown that infection with a cdtA mutant of M7404 resulted in disease that was indistinguishable from the parent strain [2]. This mutant no longer produced CDT, but had an intact cdtR gene and continued to produce TcdA and TcdB at wild type levels. A reduction in CDT levels is therefore not likely to have a major effect on disease in our animal model. All of the mice infected with the isogenic cdtR-series of M7404 derivatives were colonised with C. difficile at similar levels (S2 Fig). Wild type-infected mice rapidly lost weight and the majority were euthanized 40 to 48 hours post infection in accordance with animal ethics guidelines, with a mean time to death of 48 ± 5.1 hours and a survival rate of 13% (Fig 5). Mice infected with either of the two independent cdtR mutants had significantly higher survival rates (Mantel-Cox log rank test, P < 0.0001) of 100% and 96% for cdtR1(V) and cdtR2(V), respectively, and showed no overt signs of disease nor significant weight loss (Fig 5). By comparison, mice infected with either complemented mutant, cdtR1(R⁺) or cdtR2(R⁺), had a wild-type virulence phenotype, with marked weight loss, and a mean time to death of 31.5 ± 2.5 hours and 57.6 ± 7.5 hours, respectively, reflected in survival rates of 0% and 33% (Fig 5).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Virulence of M7404 wild-type, cdtR mutant and complemented strains in mice.

a, Kaplan-Meier survival curve showing time from infection to euthanasia of mice infected with different strains of C. difficile in hours. (n = 15). b, Time from inoculation of mice to death in hours. Data represents mean ± S.E.M. (n = 15). Data represent the mean ± SEM ****, p ≤ 0.0001.

https://doi.org/10.1371/journal.ppat.1005758.g005

Damage to the colon and caecum of C. difficile-infected mice results from the production of TcdA and TcdB [2]. We therefore performed histopathological analysis to assess the damage to colonic and caecal tissues collected from the groups of infected and uninfected mice under study here. All tissues were de-identified and independently scored using a previously defined set of scoring parameters that included overall tissue damage, polymorphonucleocyte (PMN) influx, crypt damage and oedema [2]. Tissues of uninfected mice only had minimal surface damage to the intestinal epithelia resulting from the disruption of microbiota by antibiotic pre-treatment and tissue processing (Fig 6A) with low colon and caecum damage scores of 4.7 and 3.8, respectively (Fig 6B and 6C). By comparison, wild type-infected mice had severely inflamed tissues, with extensive damage to the epithelial surface, crypt branching and hyperplasia, goblet cell loss, significant PMN influx and mucosal and sub-mucosal oedema (Fig 6A). These histopathologies were reflected in the high damage scores of 12.9 and 13.6 for their colonic and caecal tissues, respectively (Fig 6B and 6C).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Histopathology of C. difficile infected tissues.

Representative images of sections of colon and caeca collected from uninfected mice or mice infected with different strains of C. difficile, fixed and strained with PAS-Alcian blue. Red brackets ([) indicate crypt hyperplasia, arrow heads (▲) represent surface epithelial damage and asterisks (*) represent oedema and inflammation. Scale bars (200 μm) are shown in yellow. Histopathology damage scores from uninfected or infected colons (b) and caeca (c). Data represent the mean ± SEM ****, p ≤ 0.0001.

https://doi.org/10.1371/journal.ppat.1005758.g006

Mice infected with either cdtR mutant had tissue architecture similar to that seen with the uninfected mice. Very little colonic and caecal damage was observed, with some surface epithelial damage to crypts occurring and no apparent crypt hyperplasia, and little PMN influx into the mucosa (Fig 6A), with correspondingly low colonic (6.0 and 5.3) and caecal (6.6 and 7.1) damage scores (Fig 6B and 6C). However, mice infected with either cdtR-complemented mutant had similar levels of tissue damage to mice infected with the wild-type strain. Severe crypt damage was observed in the majority of these tissues, particularly the caeca, with crypt hyperplasia, loss of goblet cells, PMN influx and severe oedema in the mucosa and sub-mucosa (Fig 6A) and high histopathological damage scores were determined for both the colon (11.8 and 9.6) and the caecum (14.5 and 11.4) (Fig 6B and 6C).

To confirm that the reduced virulence of the cdtR mutants can be attributed to reduced levels of TcdA and TcdB production in vivo, the cytotoxicity of the intestinal contents collected from mice infected with the panel of C. difficile strains was assessed. No cytotoxicity was observed against HT29 or Vero cells by faecal samples collected from uninfected mice whereas mice infected with the wild type strain showed high levels of cytotoxicity against HT29 and Vero cells, indicating the production of TcdA and TcdB in vivo (S4 Fig). By comparison, samples collected from mice infected with the two cdtR mutants (cdtR1(V) and cdtR2(V)) showed reduced cytotoxicity against HT29 or Vero cells, indicating decreased levels of TcdA and TcdB production in vivo compared to the wild type strain (S4 Fig). Intestinal contents collected from mice infected with the two complemented cdtR mutants ((cdtR1(R⁺) and cdtR2(R⁺)) showed increased levels of cytotoxicity against HT29 and Vero cells, indicating restored in vivo TcdA and TcdB production. Collectively, the survival data, histopathological and in vivo cytotoxicity analysis support the hypothesis that CdtR modulates the virulence of the C. difficile ribotype 027 strain M7404 due to the role that it plays in regulating TcdA and TcdB production.

CdtR differentially regulates toxin production in a ribotype 078 strain of C. difficile

To investigate whether TcdA and TcdB regulation by CdtR occurs in other strains, we assessed this phenotype in a ribotype 078 animal isolate, JGS6133. Ribotype 078 strains are commonly isolated from animals, but have also been implicated in severe human infections [25,26]. Our sequencing analysis revealed that the cdtR gene in JGS6133 contains a naturally occurring stop codon mutation at codon 322, which has been described previously in other 078 strains [27], and results in a 142 amino acid truncation that is likely to result in a non-functional protein (Fig 1B). Western blots showed that JGS6133 complemented with the full length cdtR gene in trans, and hence producing functional CdtR, had significantly more CDTa and CDTb than the wild-type and vector control strains, which produced similar amounts of both proteins (Fig 2C). Increased CDT production by the JGS6133 cdtR⁺ derivative was reflected in increased toxin activity as assessed by ADP-ribosyltransferase assays (Fig 2F). By contrast, production of TcdA and TcdB was not altered in the JGS6133 cdtR⁺ strain compared to the wild type (Fig 3E). This result suggests that CdtR-mediated regulation of TcdA and TcdB production is not conserved within all strains of C. difficile.

CdtR does not regulate TcdA and TcdB production in strain 630

Derivatives of a ribotype 012 strain, designated 630, have been routinely used for the analysis of gene regulation and other phenotypes in C. difficile due to the relative ease of their genetic manipulation. Many strains of C. difficile, including strain 630, contain cdtAB pseudogenes within the CdtLoc (Fig 1C). These genes contain multiple frameshift mutations and stop codons and do not encode a functional CDT, but still encode a full length CdtR protein [28]. The possibility that CdtR functions to regulate other important processes, such as TcdA and TcdB production, may provide a rationale for the retention of functional CdtR in strains that carry the cdtAB pseudogenes. We were therefore interested in investigating CdtR-mediated toxin regulation in a strain with these pseudogenes.

To determine whether CdtR regulates TcdA and TcdB production in strains of C. difficile with cdtAB pseudogenes we transferred the cdtR complementation vector, pJIR4218, into strain 630. Western blot analysis showed that derivatives of 630 over-expressing cdtR had similar levels of TcdA and TcdB production as the isogenic vector control strains (S3A and S3B Fig). RT-ddPCR analysis confirmed that cdtR was over-expressed in these strains relative to the wild type and the vector control (S3C Fig). Taken together these results suggest that CdtR is not important for the regulation of TcdA and TcdB production in strain 630. Since CdtR is also not important for TcdA and TcdB regulation in a ribotype 078 strain, regulation of these toxins by CdtR may be specific to the ribotype 027 strains.

Bacterial strains and growth conditions

All bacterial strains are defined in Table 1. Culture media were from Oxoid or Becton Dickinson (BD) and all antibiotics and supplements used are from Sigma-Aldrich, Amresco or Merck unless otherwise stated. E. coli and B. subtilis strains were cultured at 37°C in 2xYT media [41] supplemented with either chloramphenicol (25 µg/ml for E. coli; 5 μg/ml for B. subtilis) or tetracycline (10 µg/ml). C. difficile strains were cultured in HIS broth[42] or on HIS agar supplemented with 0.1% (w/v) L-cysteine and 0.375% (w/v) glucose or TY broth[1] with D-cycloserine (250 µg/ml), thiamphenicol (10 µg/ml), lincomycin (50 µg/ml) or anhydrous tetracycline (50 ng/ml), as required. C. difficile cultures were grown in a Don Whitley A35 Anaerobic Workstation in an atmosphere of 10% (v/v) H₂, 10% (v/v) CO₂ and 80% (v/v) N₂ at 37°C.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Bacterial strains and plasmids.

https://doi.org/10.1371/journal.ppat.1005758.t001

Polymerase Chain Reaction (PCR)

All oligonucleotide primers are listed in Table 2. PCR cycling conditions (unless otherwise stated) were as follows: initial denaturation step at 94°C for 4 min, followed by 30 cycles of denaturation at 94°C for 30 sec, an annealing step at 50°C for 30 sec and an extension step at 72°C for 1 min per 1 kb. A final extension step was performed at 72°C for 10 minutes. PCRs were performed with Phusion DNA polymerase (New England Biolabs) and 2x Failsafe PCR buffer E (Epicentre Biotechnology). Splice-overlap extension (SOE)-PCR to re-target the Targetron was performed as described in the TargeTron Gene Knockout System users guide (Sigma-Aldrich) with modifications as previously described [43].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Oligonucleotide primers.

https://doi.org/10.1371/journal.ppat.1005758.t002

Isolation and manipulation of nucleic acids

Plasmid DNA was isolated from E. coli, B. subtilis and C. difficile using QIAprep spin miniprep columns (Qiagen) following the manufacturer’s instructions. Genomic DNA was isolated from C. difficile as previously described [44]. Standard methods of DNA digestion, modification and ligation were used. DNA sequencing was carried out using BigDye Terminator v3.1 Ready Reaction Mix (Applied Biosystems) following the manufacturer’s instructions. Sequencing reactions were resolved on an Applied Biosystems 3730 DNA Analyzer. Sequences were analysed using ContigExpress (Invitrogen).

Construction of recombinant plasmids

All plasmids are outlined in Table 1. Construction of the cdtR TargeTron plasmid was performed as previously described, with some modifications [43]. Briefly, the group II intron from pDLL45 was retargeted by SOE-PCR to insert between nucleotides 288 and 289 of the cdtR gene using the primer pairs JRP5448 and JRP3867 and JRP5449 and JRP5450 (Table 2) to generate a 350 bp product, which was digested with BsrGI and HindIII and cloned into the corresponding sites of pDLL45, resulting in pJIR4135. A StuI-HindIII fragment was then sub-cloned from pJIR4135 into the corresponding sites of pDLL55, resulting in pJIR4153.

The cdtR complementation plasmid was constructed by PCR amplifying the cdtR gene and approximately 300 bp of its promoter region from C. difficile M7404 using the primers JRP5632 and JIR5633 (Table 2). The resulting 1.1 kb fragment was purified using a PCR purification kit (Qiagen) following the manufacturer’s instructions, digested with BamHI and PstI and cloned into the corresponding sites of pDLL24, resulting in pJIR4218.

Transfer of plasmid DNA into C. difficile by conjugation

Plasmid DNA was introduced into the B. subtilis conjugative donor strain BS34A as previously described [45]. The resulting strain was used as the donor for the conjugative transfer of plasmid DNA into C. difficile strains as before [11].

Isolation of cdtR mutants

C. difficile cdtR mutants were isolated using the method previously described [11] and confirmed by PCR and Southern hybridisation analysis. Complementation of the mutation was achieved using the cdtR complementation plasmid, pJIR4218. The cloning vector, pDLL24, was transferred into the cdtR mutant and the wild-type strain to construct vector (v) controls.

Detection of TcdA, TcdB, CDTa and CDTb by western blotting

Toxins were partially purified and concentrated eight-fold from 72 hour C. difficile TY culture supernatants by methanol-chloroform precipitation [11]. Protein concentrations were determined using the BCA protein assay kit (Pierce) as per the manufacturer’s instructions. Concentrated supernatant proteins (10 µg) were separated by 10% (v/v) sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) [46] and transferred onto a nitrocellulose membrane (Whatman). Membranes were analysed as previously described [43]. TcdA and TcdB were detected using TcdA-specific monoclonal and TcdB-specific polyclonal antibodies (tgcBIOMICS), respectively. CDTa and CDTb were detected, respectively, using a CDTa-specific antibody and C. perfringens Ib-specific antibody that is cross reactive with CDTb [47]. CDTa, CDTb and TcdB-bound antibodies were detected using horseradish peroxidase conjugated anti-rabbit goat antibodies (Millipore) and TcdA-bound antibodies were detected using anti-mouse goat antibodies (Millipore). The Western Lightning Chemiluminescence reagent kit (Perkin-Elmer) was used to detect the bands, which were visualised by exposure to X-ray film or on a BioRad ChemiDoc XRS⁺ system.

ADP-ribosyltransferase assays

Toxins were partially purified from culture supernatants by precipitation with 70% ammonium sulphate as described previously [6]. ADP ribosyltransferase assays were performed as previously described [48]. Briefly, precipitated supernatant protein (50 µg) was incubated for 60 minutes at 37°C with 10 µg of actin in assay buffer (20 mM Tris-HCl pH 7.5, 1 µM dithiothreitol (DTT), 40 µM ATP, 40 µM CaCl₂, 5 µM MgCl₂) and 10 µM of biotinylated NAD⁺ (Trevigen). The reaction was heat inactivated at 95°C for 5 minutes in 4x SDS sample buffer (240 mM Tris-Cl (pH 6.8), 40% glycerol (v/v), 8% SDS (w/v), 5% (v/v) 2-mercaptoethanol, 0.05% (v/v) bromophenol blue and separated by 10% SDS-PAGE. Proteins were transferred onto a nitrocellulose membrane and biotinylated proteins were detected with horseradish peroxidase-conjugated streptavidin (GE Healthcare Life Sciences) and the Western Lightning Chemiluminescence reagent kit (Perkin-Elmer), following the manufacturer’s instructions. Relative band intensities were determined by densitometry using Image Lab Software (Bio-Rad). Data were analysed using GraphPad Prism 6 and statistical significance assessed using an unpaired t-test with a 95% confidence interval.

Vero and HT29 cell cytotoxicity assays

C. difficile strains were grown overnight in 20 ml of HIS broth with thiamphenicol and lincomycin, as required. The cultures then were used to inoculate 50 ml of TY broth with selection, such that each culture had a starting OD₆₀₀ of approximately 0.05. Aliquots (5 ml) were taken at 12, 24, 48 and 72 hours, pelleted by centrifugation (10,000 g, 10 min, room temperature) and the supernatants filter sterilised through 0.45 µM and 0.2 µM filters (Sartorius). Supernatants were stored on ice until use. Vero cell and HT29 cell cytotoxicity assays were performed using the filtered C. difficile supernatants as previously described [43]. The levels of TcdA and TcdB produced by the C. difficile strains in vivo was assessed by determining the cytotoxicity of the intestinal contents collected 24 hours post infection against HT29 and Vero cells. Intestinal samples were resuspended in 100 mg/ml in PBS, diluted one in eight, filter sterilised and applied to Vero and HT29 cells, as described previously [2]. The endpoint (toxin titre) was scored as the last dilution with 100% cytopathic effect (CPE). Data were analysed using GraphPad Prism 6 and statistical significance assessed using an unpaired t-test with a 95% confidence interval.

RT-ddPCR analysis of C. difficile gene expression

Total RNA was extracted using TRIzol® (Life Technologies) following the manufacturer’s instructions. Forty ml of C. difficile TY broth cultures with an OD₆₀₀ of approximately 0.3 for tcdC and cdtA expression analysis, and 10 ml of TY broth culture grown to OD₆₀₀ of approximately 1.8 for tcdA, tcdB and tcdR expression analysis, were used. A total of 200 ng of RNA was converted to cDNA using SuperScript III Reverse Transcriptase (Life Technologies), following the manufacturer’s instructions. Transcript levels were quantified using the QX200 Droplet Digital PCR System (BioRad) using QX200 ddPCR EvaGreen Supermix and 0.1–5 µg of total cDNA and specific primers (Table 2) at a concentration of 200 nM. Transcription levels of each gene was normalised to transcription levels of the housekeeping gene rpoA. Data were analysed using GraphPad Prism 6 and statistical significance assessed using a Mann Whitney U test.

C. difficile virulence trials

Groups of five male six to eight week old C57BL/6 mice were used in C. difficile virulence trials as previously described [2], except mice were switched back to plain drinking water on the day of infection. Mice were administered 10⁶ C. difficile spores by oral gavage and were humanely euthanised at the onset of severe disease or at the end of the experiment, as previously defined [2]. Animal handling and experimentation were performed in accordance with institutional guidelines (Monash University animal ethics committee numbers MARP/2014/135 and SOBSB/M/2010/25). Faecal samples were taken daily to monitor C. difficile shedding using HIS agar supplemented with 0.1% (w/v) cysteine, 0.1% (w/v) taurocholate, 0.375% (w/v) glucose, 250 µg/ml D-cycloserine, 8 µg/ml cefoxitin, 10 µg/ml erythromycin, 12 µg/ml norfloxacin, 32 µg/ml moxalactam. Data were analysed using GraphPad Prism 6 and statistical significance assessed using a log-rank (Mantel-Cox) test.

Histology

The entire colon and caecum were collected from each mouse and Swiss-rolled [49] prior to fixation to allow for cross-sectional examination of the entire length of the colon. Tissues were stained with PAS-Alcian blue and histopathological assessment of damage and scoring of tissues was performed blind by independent observers using a previously defined set of parameters [2].