Fig 1.
The pdcB switch (Cdi2) is invertible in C. difficile.
(A) Diagram of the pdcB locus indicating the relative positions of the upstream invertible element (orange) and inverted repeats (black arrows). LIR–left inverted repeat, RIR–right inverted repeat. PdcB contains a PAS domain (green), a degenerate GGDEF domain (blue), and an EAL domain (purple). (B) Diagram of the OS-PCR strategy used in (C) to selectively amplify the orientation of the pdcB switch present in the published genome of R20291 (Pub) or the inverse (Inv) orientations. Primer positions are indicated by half-arrows, and the genomic region indicated (-1147 to -760) is relative to the predicted pdcB start codon. (C) OS-PCR for the pdcB switch in six different C. difficile strains, with the previously determined ribotype indicated in parenthesis.
Fig 2.
Mutation of the right inverted repeat results in phase-locked strains.
(A) OS-qPCR was performed using genomic DNA from R20291 (WT), pdcBΔ3-ON, and pdcBΔ3-OFF. Data are expressed as the percentage of the invertible element in orientation present in the reference genome (Pub/OFF). (B) qRT-PCR was performed using cDNA derived from R20291 (WT), pdcBΔ3-ON, and pdcBΔ3-OFF. Data are expressed as the ratio of the pdcB transcript abundance to that of the rpoC reference gene. (C) Three nucleotides in the RIR were deleted to create strains with the invertible element locked in the ON and OFF orientations. Asterisk indicates the site of recombination [10]. Italics indicate a portion of the pdcB switch sequence that undergoes inversion. (A, B) Symbols represent values from independent samples. For WT, triangles and squares distinguish isolates with reference and inverse sequence orientations, respectively. Indicated are the medians. **p < 0.01 by Kruskal-Wallis test and Dunn’s post-test.
Fig 3.
The pdcB switch contains an invertible promoter.
(A) Diagrams of the pdcB switch in the OFF and ON orientations with TSS1 and TSS2 positions indicated relative to the pdcB start codon. Positions of primers used to construct the fusions to phoZ in (B) and (E) are indicated by half arrows; numbers in parentheses match the primers used with the corresponding plasmid construct. (B) Alkaline phosphatase (AP) assay using C. difficile strains with plasmid-borne transcriptional fusions to phoZ. Promoterless phoZ was used as control. (C,D) Chromatographs of the Sanger sequencing results obtained by 5’ RACE from (C) wild-type R20291 and (D) R20291 carrying the Cdi2-ON::phoZ reporter. TSS1 and TSS2 were identified as the first nucleotide adjacent to the poly-C tail added to the 5’ end of cDNA; the position relative to the pdcB start codon is indicated. (E) AP assay using C. difficile strains with plasmid-borne transcriptional fusions to phoZ. Red asterisks indicate the presence of a 3-nucleotide deletion in the RIR. (B,E) Means and standard deviations from 3 independent experiments are shown. *p < 0.05, **p < 0.01, ****p < 0.0001 by one-way ANOVA and Tukey’s post-test. No significant differences between strains marked #. The nucleotide positions relative to the pdcB predicted start codon for each end (5’, 3’) of the C. difficile sequences fused to phoZ are as follows: (1) -1,024 to -829, (2) -1,023 to -829, (3) -974 to -829, (4) -915 to -829, (5) -1,214 to -829, (6) -1,214 to -829, and (7) -1,214 to -957.
Fig 4.
Phase variation of PdcB modulates c-di-GMP at the population and single-cell levels.
(A) Measurement of c-di-GMP using the pPRS::mCherry reporter in R20291 (WT), pdcBΔ3-ON, and pdcBΔ3-OFF. Relative Fluorescence Units (RFU) calculated from arbitrary fluorescence units after 180 minutes RFP maturation normalized to OD600 at time 0. Kinetic analysis of fluorescence and data for the pPRSA70G::mCherry controls are in S2 Fig. (B) Representative micrographs showing heterogeneity in fluorescence in the above strains. (C) Number of mCherryOpt-expressing cells determined by flow cytometry expressed as a percentage of total Syto-9 stained cells. (A,C) Bars indicate means and standard deviations, with circles indicating values from independent biological samples. *p< 0.05, **p< 0.01, ****p<0.0001 by one-way ANOVA and Tukey’s post-test.
Fig 5.
Expression of pdcB affects known c-di-GMP regulated processes.
Phenotypic analysis of C. difficile R20291 (WT), ΔpdcB, pdcBΔ3-ON (ON), and pdcBΔ3-OFF (OFF). (A) Swimming motility in 0.5x BHIS-0.3% agar after 48 hours. A non-motile mutant (ΔsigD) served as a negative control. (B) Surface motility on BHIS-1.8% agar-1% glucose after 6 days. A ΔcmrT mutant served as a negative control. (C) Biofilm formation measured by crystal violet staining after 24 hours of growth in BHIS broth. (A-C) Means and error are shown. Symbols indicate values from independent biological samples. *p < 0.05, ***p < 0.001, ****p < 0.0001 by one-way ANOVA and Tukey’s post-test; #p < 0.05 compared to all other strains.
Fig 6.
PdcC contributes to inhibition of biofilm formation, but not swimming or surface motility.
(A) Swimming motility in 0.5x BHIS-0.3% agar after 48 hrs. A non-motile mutant (ΔsigD) served as a negative control. (B) Surface motility on BHIS-1.8% agar-1% glucose after 6 days. A ΔcmrT mutant served as a negative control. (C) Biofilm formation was measured by crystal violet staining after 24 hours of growth in BHIS broth. (A-C) Means and standard error are shown. Symbols indicate values from independent biological samples. *p < 0.05 by Mann-Whitney test.
Fig 7.
Orientations of the other invertible sequences do not correlate with the phenotypes of the pdcB mutants.
Genomic DNA from WT, ΔpdcB, pdcBΔ3-ON, and pdcBΔ3-OFF grown on BHIS agar was subjected to OS-qPCR for the invertible elements pdcB (A), pdcC (B), flg (C), cmr (D), cwpV (E), CDR20291_0963 (F), and CDR20291_3417 (G). The data are presented as the percentage of the invertible element in the orientation of the reference R20291 genome.
Fig 8.
Model of PdcB and PdcC phase variation in the context of c-di-GMP signaling in C. difficile.
Within the range of possible c-di-GMP concentrations established by c-di-GMP biosynthetic and hydrolytic enzymes, represented by the blue wedge, phase variation of PdcB and PdcC may shift c-di-GMP levels in a bacterial cell. A combination of pdcB ON and pdcC OFF or vice versa would maintain intermediate levels; switching of both to OFF would lead to accumulation of c-di-GMP; and switching of both to ON would deplete c-di-GMP. These switch combinations result in heterogeneity in intracellular c-di-GMP levels in the bacterial population and influence the phenotypes exhibited. Higher c-di-GMP concentrations favor surface-associated behaviors such as biofilm formation and surface migration which are in part mediated by type IV pili, CmrRST, and potentially putative adhesins. Lower c-di-GMP favors planktonic behaviors such as swimming motility, toxin production, sporulation, and the ZmpI metalloprotease proposed to cleave adhesins to release C. difficile from a surface/biofilm [31,34,50]. Not shown, additional DGCs and c-di-GMP PDEs influence c-di-GMP levels. Green stars indicate factors whose transcription is directly up-regulated by class II c-di-GMP riboswitches; red stars, factors whose transcription is directly down-regulated by class I c-di-GMP riboswitches. Purple triangles denote factors that independently phase vary.