CamA promotes the expression of the early-acting and critical sporulation gene, spoIIE

Our previous RNA-Seq analyses indicated that the earliest sporulation stage altered in a ∆camA mutant relative to WT is the activation of σ^F because Spo0A regulon genes (like spo0A and the sigF-spoIIAB-spoIIAB operon) are similarly expressed in a ∆camA mutant relative to WT, whereas σ^F regulon genes (like gpr, spoIIQ, and spoIIP) are under-expressed [10] (Fig 1B). Thus, to identify sporulation genes epigenetically regulated by CamA, we focused on genes involved in regulating σ^F activity. σ^F activity is constrained to the forespore even though it is produced throughout the predivisional cell. In B. subtilis, the compartment-specific activation of σ^F is controlled by a signaling cascade consisting of the anti-sigma-factor SpoIIAB, an anti-anti-sigma factor SpoIIAA, and the multi-functional SpoIIE phosphatase [18–21] (Fig 1A). Since the genes encoding these regulatory factors are conserved in C. difficile [17], we compared their expression levels using RT-qPCR. While the sigF-spoIIAA-spoIIAB operon was similarly transcribed in WT and ∆camA, spoIIE was under-expressed ~2-fold in ∆camA relative to WT (Fig 1C), even though both the sigF operon and spoIIE gene are direct targets of Spo0A [13]. Intriguingly, the spoIIE promoter region is enriched in CamA recognition motifs (CAAAAA; the methylated base is bolded): the 160 bp region upstream of the spoIIE transcription start site (TSS) contains three methylation sites (Fig 1D). This represents a significant enrichment compared to equally sized regulatory regions (p < 10^-5), and the presence of these sites is strictly conserved across 135 C. difficile genomes, despite the high degree of genetic variation between isolates (S1 Fig). Furthermore, the Me3 methylation site is directly adjacent to a predicted Spo0A binding site in the spoIIE promoter. While Spo0A binding sites have not been empirically identified in this region, they were inferred based on the consensus C. difficile Spo0A recognition motif [24], allowing up to two mismatches provided that at least two critical guanine/cytosine residues are retained; sites meeting these criteria are found in a similar position relative to the empirically identified Spo0A binding sites in the B. subtilis spoIIE promoter [25] (S2 Fig). Collectively, these analyses suggest an evolutionary pressure to maintain the CamA methylation sites in the spoIIE promoter region.

Although spoIIE expression was only reduced ~2-fold in the bulk RNA-Seq and RT-qPCR analyses (Fig 1B, 1C), these population-wide measurements can obscure underlying cell-to-cell variability in transcription levels. Since DNA methylation often influences the proportion of cells exhibiting a given transcriptional state [9], we hypothesized that CamA affects the frequency of spoIIE-expressing cells rather than the amplitude of gene expression. To test this, we generated a bicistronic spoIIE-RBS-mScarlet-I3 transcriptional reporter by integrating an mScarlet-I3 gene, carrying a spoIIE RBS, downstream of the spoIIE gene in its native locus. Analysis of spoIIE expression at the single-cell level after 9 hours of growth on sporulation-inducing medium revealed that ~2-fold fewer ∆camA cells activated the P_spoIIE transcriptional reporter compared to WT (4.7 ± 3.9% versus 11.0 ± 3.1% of cells, respectively) (Fig 1E,1F). The ∆camA mutant also induced spoIIE at lower levels, with the median fluorescence intensity of spoIIE-positive ∆camA cells being 1262 ± 605 a.u. compared to 1356 ± 754 a.u. for WT (Fig 1G). Together, these results suggest that CamA increases the frequency of cells that induce spoIIE expression, as well as the magnitude of gene expression at the single-cell level.

To determine the functional consequence of altering spoIIE expression levels in C. difficile, we first constructed and characterized a ∆spoIIE mutant. Similar to observations in B. subtilis, a C. difficile∆spoIIE mutant failed to activate σ^F based on fluorescence microscopy analyses of a σ^F-activity reporter (S3A Fig), and they did not form heat-resistant spores (S3B Fig). In addition, ∆spoIIE cells failed to progress beyond asymmetric division and their polar septa were thicker relative to WT in transmission electron microscopy analyses (S3C, S3D Fig), consistent with prior work in B. subtilis [26,27]. This indicates that C. difficile SpoIIE is essential for proper asymmetric division and σ^F activation, and it shares conserved functions with previously studied spore formers [27–29]. Taken together, these data indicate that CamA-mediated DNA methylation increases the proportion of cells expressing a gene that encodes a critical early-acting factor in the transcriptional program controlling sporulation.

Methylation of a single site in the spoIIE promoter region enhances C. difficile sporulation levels

Since camA expression appears to be constitutive across growth conditions [10], we reasoned that changes to spoIIE transcription are due to local regulatory effects of methylation. We sought to test this hypothesis by mutating the CamA methylation sites in the spoIIE promoter region. To this end, we individually mutated the three methylation sites found upstream of the spoIIE TSS by first generating a deletion of this upstream region and then restoring the deleted region with the wild-type sequence (WT*) or sequences carrying mutations in the individual CamA recognition sites (CAAAAA → CAAATA, Me1*, Me2*, and Me3*) (Fig 2A). Importantly, the mutations alter the CamA recognition sequence without altering the fifth adenine methylated by CamA [10], and the Me3* mutation does not affect the predicted Spo0A recognition motif close to the -35 site [28]. We then measured the sporulation efficiency of the resulting mutants using a heat resistance assay. While preventing the methylation of the Me1 or Me2 sites did not affect sporulation levels, preventing the methylation of the promoter-proximal CamA site (Me3*), which is adjacent to a predicted Spo0A binding site (Fig 1D), phenocopied the sporulation defect observed in the ∆camA mutant (Fig 2B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. A single methylation site in the spoIIE promoter enhances sporulation.

(a) Schematic of the individual methylation site mutations made in the spoIIE promoter region of C. difficile 630∆erm. The region upstream of spoIIE was deleted, and DNA fragments containing the WT sequence with a watermark (WT*) or individual methylation site mutations (Me1*, Me2*, Me3*) were reintroduced by allelic exchange. (b) Heat resistance assay comparing the sporulation efficiencies of the methylation site mutants. Approximately 22 hours after sporulation was initiated, cells were heat-treated and then plated to enumerate the viable spore count relative to the untreated sample. ∆spo0A and ∆spoIIE were used as negative controls, since these strains cannot complete sporulation [30]. Results were analyzed using one-way ANOVA and Tukey’s multiple-comparisons test. *, p < 0.033. **, p < 0.002. Data represent a minimum of nine biological replicates. (c) Percentage of spoIIE-ON cells. Cells were classified as spoIIE-OFF or spoIIE-ON based on the bimodal distribution of mean fluorescence intensity observed in violin plots. The cutoff was set at the valley between the peaks and validated against fluorescence levels of cells barely detectable in microscopy images. Statistical significance was determined using a one-way ANOVA and Tukey’s multiple comparisons test. *, p < 0.033. (d) Histogram showing the frequency distribution of cells at different levels of mean fluorescence intensity for images of strains shown in S4 Fig. The data represent 7,500 cells counted across three biological replicates. Data for WT, ∆spo0A and ∆camA shown in Fig 1G are included here to allow for direct comparison with WT* and Me3*.

https://doi.org/10.1371/journal.ppat.1013845.g002

To determine the effect of the Me3* methylation site mutation on spoIIE transcription, we introduced the bicistronic spoIIE-RBS-mScarlet-I3 transcriptional reporter construct into the WT* and Me3* strains and analyzed spoIIE transcription using fluorescence microscopy. These analyses revealed that mutation of the single Me3 site was sufficient to phenocopy the decrease in spoIIE expression observed in ∆camA cells: fewer Me3* cells expressed spoIIE at detectable levels and the magnitude of their expression was reduced compared to WT* (Figs 2C, 2D, S4).

Since the Me3 motif partially overlaps with a putative Spo0A recognition motif (Fig 1D), we considered the possibility that mutation of the Me3 site affects the innate binding affinity of Spo0A for the P_spoIIE promoter, even though the mutation is not predicted to alter the Spo0A recognition sequence [24,31]. Thus, we compared Spo0A binding to the Me3* and WT promoter regions using electromobility shift and fluorescence polarization assays. These assays both revealed that the DNA-binding domain of Spo0A binds to WT and Me3* P_spoIIE DNA probes with similar affinity (S5 Fig). Thus, the reduced spoIIE expression and sporulation levels observed in ∆camA and Me3* mutants can be attributed to the loss of methylation, rather than a decrease in the intrinsic affinity of Spo0A for its binding sites. Together, these findings reveal that, of the nearly 8,000 CamA recognition motifs in C. difficile’s genome, methylation of the single Me3 site in spoIIE’s promoter region is sufficient to enhance C. difficile sporulation.

Methylation of P_spoIIE increases the frequency of σ^F activation in the forespore, as well as the predivisional cell

Since SpoIIE is predicted to be a positive regulator of σ^F activation (Fig 1A), we next asked whether the reduced spoIIE expression in the ∆camA and Me3* mutants is sufficient to reduce σ^F activation at the single-cell level. Analyses of σ^F activation using the σ^F-dependent P_gpr::SNAP transcriptional reporter (Fig 3A) after 15 hours of growth on sporulation-inducing medium revealed that 34% of WT cells activate σ^F in the forespore, while 17% of ∆camA and Me3* cells activate this sigma factor in this compartment (Fig 3B). These results are consistent with the ~ 2-fold decrease in sporulation measured for cells that cannot methylate the Me3 site (∆camA and Me3*, Fig 2B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Methylation of P_spoIIE enhances σ^F activation but impairs the compartmentalization of σ^F activity to the forespore.

(a) Images of cells harboring a σ^F activity reporter construct (P_gpr::SNAP) grown on sporulation-inducing medium for 15 hours. gpr is a member of the σ^F regulon, so its transcription serves as a reporter for σ^F activation. Cell membrane staining (FM 4-64) indicates that cells that have activated σ^F throughout the entire cell (labeled “whole cell”) have not completed asymmetric division. ∆spo0A was used as a negative control, since this strain cannot initiate sporulation [30]. Scale bar, 5 µm. (b) Percent of cells activating σ^F exclusively in the forespore or throughout the whole cell for images shown in panel a. Data represent the mean ± standard deviation of at least 1,500 cells counted across three biological replicates. Statistical significance was determined using a one-way ANOVA and Tukey’s multiple-comparisons test. *, p < 0.033; **, p < 0.002.

https://doi.org/10.1371/journal.ppat.1013845.g003

Unexpectedly, the σ^F activity reporter revealed that a sizable proportion of visibly sporulating WT C. difficile cells activate σ^F throughout the entire cell, rather than exclusively in the forespore compartment. Specifically, 17% of WT cells activated σ^F in the entire cell compared to 3% and 7% of ∆camA and Me3* mutant cells, respectively (Fig 3C). Membrane staining revealed that cells with mislocalized σ^F activity throughout the cell have not yet completed asymmetric division, indicating that σ^F is being prematurely activated in predivisional cells (Fig 3A). In contrast, in B. subtilis, only 0.5-2% of sporulating cells activate σ^F in the predivisional cell [18, 32–34]. Notably, in B. subtilis, the compartment-specific activation of σ^F in the forespore following asymmetric division is critical for the genetically identical mother and forespore cells to establish the disparate transcriptional programs that allow for proper spore formation. Moreover, B. subtilis mutant cells that prematurely activate σ^F in the predivisional cell undergo lysis [19,23], so the high proportion of σ^F activation in predivisional C. difficile cells was surprising.

Given the critical importance of properly compartmentalizing σ^F activity during B. subtilis sporulation, we sought to understand the basis and consequence of prematurely activating σ^F in predivisional C. difficile cells. We first considered whether the high frequency of premature σ^F activation was an artifact of the SNAP-tag labeling system. To test this, we visualized σ^F activation using a P_spoIIQ::mScarlet transcriptional reporter, which is driven by a different σ^F-dependent promoter and uses mScarlet as the visualizable reporter. With this system, we again found that 21% of WT cells activate σ^F in the predivisional cell (S6 Fig). Next, we wondered whether σ^F activation in the predivisional cells might be due to leakage of the reporter protein into the cytoplasm in cells that have retracted their polar septum, which can occur when the peptidoglycan synthesis or hydrolysis machinery is dysregulated during sporulation [35–37]. To assess this, we examined σ^F activation in ∆spoIIP∆spoIID cells, which lack the septal hydrolysis machinery that mediates forespore cytoplasmic leakage in B. subtilis [35–37]. Whole cell σ^F activation was still observed in ∆spoIIP-∆spoIID cells (S7 Fig), indicating that the σ^F activity is indeed observed in predivisional cells rather than in cells that have retracted their septum. Taken together, these data indicate that σ^F activity is less tightly compartmentalized in C. difficile than in B. subtilis and that DNA methylation at the Me3 site of P_spoIIE increases the frequency of premature σ^F activation in predivisional cells.

Elevated spoIIE expression correlates with the loss of σ^F compartmentalization

Since sporulating WT cells express spoIIE at higher levels and activate σ^F prematurely in predivisional cells more frequently compared to the ∆camA and Me3* mutants (Fig 3), we hypothesized that elevated SpoIIE levels lead to the ectopic activation of σ^F prior to division. This hypothesis is based on observations in B. subtilis, where cells engineered to overexpress the spoIIE gene [18] or accumulate the SpoIIE protein [28,38] exhibit premature activation of this sigma factor. To test this possibility, we constructed a dual reporter system where (i) P_spoIIE expression was detected using our bicistronic spoIIE-RBS-mScarlet transcriptional reporter integrated into the native locus and (ii) σ^F was monitored using a P_gpr::SNAP transcriptional reporter construct integrated into the ectopic pyrE locus (Fig 4A). The induction of these reporters was analyzed after 12 hours of growth on sporulation-inducing medium, when both the spoIIE and σ^F activation reporters were visible. Indeed, these analyses revealed that, across all strains, mean spoIIE transcription is lower in cells with forespore compartmentalized σ^F activation (860–4680 a.u., median = 1946 a.u.), whereas uncompartmentalized σ^F activation in predivisional cells occurs at a higher range of spoIIE expression (1600–5447 a.u., median = 2971 a.u.) (Figs 4B, S8). These data strongly suggest that σ^F activation is sensitive to SpoIIE levels and that elevated levels of spoIIE transcription induces the premature activation of σ^F prior to the completion of asymmetric division in C. difficile, similar to B. subtilis cells artificially engineered to overexpress spoIIE [18]. We next attempted to directly correlate SpoIIE levels and  localization to the polar septum with σ^F activation, by generating strains encoding SpoIIE-mScarlet-I3 protein fusions with different linkers. Unfortunately, the mScarlet-I3 variant was cleaved off all the fusions tested, so we were unable to assess the relationship between SpoIIE localization and σ^F activation.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. Elevated spoIIE expression correlates with the loss of σ^F compartmentalization.

(a) Images of cells harboring a σ^F activity reporter (P_gpr::SNAP) in the ectopic pyrE locus and a bicistronic spoIIE-RBS-mScarlet-I3 transcriptional reporter integrated into the native spoIIE locus. Cells were grown on sporulation-inducing medium for 12 hours before labeling and fixation. ∆spo0A was used as a negative control, since this strain cannot initiate sporulation [11]. Green arrows indicate cells that have activated σ^F in the forespore; yellow arrows indicate cells that have activated σ^F across the whole cell. Scale bar, 5 µm. (b) Mean fluorescence intensities of cells grouped by their σ^F activation pattern (i.e., P_gpr::SNAP expression): no σ^F activity, σ^F activity localized to the forespore, or σ^F activity observed across the whole cell. Due to day-to-day variability in the fluorescent signal from the mScarlet-I3 reporter, the data shown are derived from a single biological replicate (1000 cells analyzed) that is representative of the trends observed for three independent replicates. The additional replicates and the resulting statistical analyses are shown in S8 Fig.

https://doi.org/10.1371/journal.ppat.1013845.g004

In B. subtilis, the coordinated regulation of SpoIIE’s localization, stability, and phosphatase activity tightly restricts σ^F activation to the forespore, even though all σ^F regulatory factors are present in the predivisional cell [18,20,39,40]. To probe how elevated SpoIIE levels promote σ^F activation in the predivisional cell, we first asked whether the systems that regulate SpoIIE in B. subtilis are conserved in C. difficile. In B. subtilis, the scaffolding protein DivIVA biases SpoIIE to the forespore face of the polar septum, which contributes to compartment-specific σ^F activation [32,33,41,42]. If similar DivIVA-dependent mechanisms regulate SpoIIE localization in C. difficile, deletion of divIVA should disrupt the architecture of septal proteins and impair σ^F compartmentalization. To test this hypothesis, we generated an in-frame deletion of divIVA and analyzed the activation of σ^F using the P_gpr::SNAP reporter. In contrast with B. subtilis [33,41], loss of DivIVA in C. difficile did not impact the ability to activate σ^F in the forespore (Figs 5A, S9) or cause a sporulation defect. Instead, the sporulation efficiency of our ∆divIVA mutant was higher than WT, although this may be an artifact of the ∆divIVA chaining phenotype reducing the accuracy of cell counts and complicating the heat-resistant CFU measurements (Fig 5B). Complementation of the ∆divIVA strain with an ectopic copy of divIVA reversed the chaining and elevated sporulation phenotype (Fig 5B).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. Forespore-specific σ^F activation in C. difficile is not dependent on canonical mechanisms for ensuring the compartmentalization of SpoIIE activity in B. subtilis.

(a, b) Forespore-specific σ^F activity is independent of the structural protein DivIVA. (a) Representative images of cells harboring a σ^F activity reporter construct (P_gpr::SNAP) in the pyrE locus grown on sporulation-inducing medium for 14-15 hours. Cell membrane staining (FM 4-64) indicates that the ∆divIVA mutant has a cell separation defect, but it can still compartmentalize σ^F activity to the forespore. Scale bar, 5 µm. (b) Phase-contrast images of WT and mutant strains grown on sporulation-inducing medium for 24-26 hours show that ∆divIVA cells sporulate efficiently. H.R. refers to the sporulation efficiency of mutants relative to WT based on their heat resistance properties. Sporulating cells were either heat-treated or left untreated prior to plating the samples on rich medium containing taurocholate germinant. The viable spore count was determined relative to the untreated sample; the resulting ratio was compared to the ratio determined for WT. ∆spo0A was used as a negative control, since this strain cannot initiate sporulation [11]. The data shown represent the mean ± standard deviation of at least three biological replicates. (c) Western blot analyses of SpoIIE stability following the arrest of transcription with rifampicin (100 µg/mL) and translation with chloramphenicol (100 µg/mL). Growth curves indicating C. difficile’s susceptibility to these antibiotics are shown in S11 Fig. Samples were collected at the indicated times after antibiotic addition and probed with anti-SpoIIE and anti-GDH (loading control) antibodies. (d) Quantification of western blot shown in panel c. SpoIIE levels were quantified and normalized to the GDH loading control for each sample, and the fold-change in SpoIIE was calculated relative to WT at 0 minutes of transcription and translation arrest. Data are representative of three biological replicates.

https://doi.org/10.1371/journal.ppat.1013845.g005

Another mechanism that promotes SpoIIE activity specifically in the forespore is its selective degradation by FtsH in the mother cell [32]. Notably, SpoIIE and FtsH are encoded next to each other in the genome of most spore formers, including B. subtilis, and a conserved N-terminal sequence in B. subtilis SpoIIE targets it for degradation [28]. However, C. difficile spoIIE lacks synteny with the ftsH gene and the N-terminus of C. difficile SpoIIE does not share conserved features with the FtsH-targeted degradation tag of B. subtilis (S10 Fig). Consistent with the absence of these features, we found that C. difficile SpoIIE remains stable for at least 120 minutes after both transcription and translation are arrested during sporulation (Figs 5C, 5D, S11). Taken together, these findings suggest that the forespore-specific activation of σ^F is not dependent on canonical mechanisms for compartmentalizing SpoIIE activity in B. subtilis, with mechanisms beyond SpoIIE stability or DivIVA-directed localization constraining C. difficile SpoIIE activity to the forespore.

Activation of σ^F in the predivisional cell does not inhibit vegetative cell growth

In B. subtilis, the activation of σ^F is a critical developmental decision. If a sporulating culture is returned to nutrient-rich conditions, cells can revert to vegetative growth if they have not activated σ^F in the forespore. If they have activated σ^F in the forespore, the cell is irreversibly committed to completing this differentiation process regardless of the environmental conditions encountered [22,43]. Conversely, if a predivisional B. subtilis cell prematurely activates σ^F, the cell will lyse due to mechanisms that remain poorly understood [18,19,23]. Given that σ^F is prematurely activated in a third of sporulating C. difficile cells (Fig 3), we wondered how this ectopic activation affects the fate of these cells and whether cells that have activated σ^F in the forespore are committed to completing sporulation, as has been documented in B. subtilis [22]. To address the latter question, we used time-lapse microscopy to study the fate of sporulating WT, sigF ^–, and sigE^– cultures after they were returned to nutrient-rich conditions using an anaerobic imaging system we recently optimized for C. difficile [44]. Consistent with studies in B. subtilis, cells capable of activating σ^F (WT and sigE^–) were unable to resume vegetative growth upon return to nutrient-rich conditions, provided that they had completed asymmetric division, whereas cells deficient for σ^F (sigF^–) rapidly resumed vegetative growth despite having divided asymmetrically (S12 Fig, S1-S6 Movies). Thus, σ^F is critical for committing C. difficile cells to completing sporulation.

We next examined the consequence of prematurely activating σ^F in predivisional C. difficile cells using time-lapse microscopy. To identify sporulating cells that had activated σ^F, we used WT and ∆camA strains harboring the σ^F activity reporter P_gpr::SNAP. Sporulating cells that had activated σ^F were first labeled with the SNAP-tag reagent under anaerobic conditions [45] and then inoculated onto nutrient-rich medium, and growth was analyzed over time. These analyses revealed that predivisional cells that had prematurely activated σ^F throughout the cell were competent to resume vegetative growth, whereas cells that properly compartmentalized σ^F activity in the forespore were not (Fig 6A, S7-S8 Movies). To determine whether cells that activate σ^F in the predivisional cell were capable of activating σ^E, the second sigma factor in the sporulation cascade, we introduced a σ^E-dependent transcriptional reporter (sipL-RBS-mScarlet-I3) into a WT strain carrying the σ^F-dependent transcriptional reporter (P_gpr::SNAP). Analyses of this dual reporter strain demonstrated that cells that activate σ^F in the forespore activate σ^E in the mother cell, while cells that prematurely activate σ^F in the predivisional cell fail to activate σ^E, indicating they do not proceed with the sporulation program (Fig 6B). Taken together, our data reveal that, despite the irreversible nature of σ^F activation in the forespore (S12 Fig), σ^F activation in the predivisional cell permits exit from the sporulation program and resumption of vegetative growth. This developmental flexibility presumably enhances the ability of C. difficile cells to adapt to fluctuating environments. Our analyses further suggest that CamA-mediated increases in spoIIE expression increase the frequency of cells that activate σ^F in both the forespore and the predivisional cell, allowing CamA to promote sporulation without compromising a population’s developmental plasticity.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Cells that activate σ^F in the predivisional cell can resume vegetative growth.

(a) Time-lapse microscopy analyzing the fate of sporulating WT cells that have activated σ^F in the forespore relative to the whole cell upon being transferred to nutrient-rich conditions. A WT strain carrying a σ^F activity transcriptional reporter (P_gpr::SNAP) integrated into the pyrE locus was grown for 14 hours on sporulation-inducing medium. The cells were then labeled with SNAP-substrate to visualize cells that have activated σ^F. The labeled cells were then spotted onto nutrient-rich agarose pads supplemented with cysteine. Agarose pads were prepared in the anaerobic chamber using gas-tight Gene Frames. After inoculating the agarose pad with the sporulating cultures, the growth chamber was sealed with a coverslip and then removed from the chamber for imaging. Cells were imaged every 2.5 min for approximately 4 hours. Green arrows indicate cells that have activated σ^F in the forespore; yellow arrows indicate cells that have activated σ^F across the whole cell. Corresponding movies are provided as S7 and S8 Movies. Images are representative of three independent replicates. (b) Images of WT cells harboring a σ^F activity reporter (P_gpr::SNAP) integrated into the ectopic pyrE locus and a bicistronic sipL-RBS-mScarlet-I3 transcriptional reporter (reflecting σ^E activity) integrated into the native sipL locus. Cells were grown on sporulation-inducing media for 15 hours before labeling and fixation. Green arrows indicate cells that have activated σ^F in the forespore; yellow arrows indicate cells that have activated σ^F across the whole cell; purple arrows indicate cells that have activated σ^E. Scale bar, 5 µm. Images are representative of three independent replicates.

https://doi.org/10.1371/journal.ppat.1013845.g006

CamA confers a fitness advantage during infection that is largely independent of its effect on sporulation

Although the ability to sporulate does not influence C. difficile’s ability to cause gastrointestinal disease, it is essential for transmission between hosts [11] and may contribute to persistence and disease recurrence [12]. Given the potential for mice to reinoculate themselves with spores through coprophagy, we sought to determine the extent to which ∆camA’s sporulation defect contributes to its persistence defect. To this end, we compared the relative fitness of the Me3* strain, which specifically recapitulates the sporulation defect of the ∆camA mutant, to WT and ∆camA in competition experiments. After sensitizing mice to infection with an antibiotic treatment regimen, mice were inoculated with a 1:1 mixture of spores consisting of pairwise combinations of WT, ∆camA, and Me3* (Fig 7A). To differentiate between strains, one member of each pair harbored a gene encoding spectinomycin resistance (spec^R). To control for potential bias due to the presence of the spec^R cassette, we swapped the cassette between strains across replicates.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 7. DNA methylation confers a multifactorial fitness advantage in a mouse model of infection.

(a) Overview of the competitive infection model in mice. Female C57BL/6 mice were provided with water containing cefoperazone (0.5 mg/mL) for 10 days prior to an IP injection of clindamycin (10 mg/kg). Mice were inoculated via oral gavage with a 1:1 ratio of competing strains for a total dose of 10⁵ C. difficile spores. Bacterial CFU, mouse body weight, and body condition score were recorded at the indicated time points post-infection. The relative fitness of (b) WT/WT (2 experiments, 8 mice total), (c) ∆camA/WT (4 experiments, 18 mice total), (d) Me3*/WT (3 experiments, 12 mice total), and (e) ∆camA/Me3* (3 experiments, 12 mice total). The competitive index was calculated as (CFU_{output, Strain A}/ CFU_{output, Strain B}) ÷ (CFU_{input, Strain A}/ CFU_{input, Strain B}). (f) The competitive index of each pairwise combination on the final day of coinfection was determined, and statistical significance was assessed using a one-way ANOVA and Tukey’s multiple-comparisons test.

https://doi.org/10.1371/journal.ppat.1013845.g007

As expected for infections with C. difficile 630∆erm [46], the mice exhibited mild weight loss and diarrhea at 2 days post-infection, but then regained the lost weight by Day 4 (S13 Fig). The mice remained stably colonized with C. difficile after their symptoms resolved and until the experiment was concluded (S13 Fig). To validate our method of strain differentiation using spectinomycin resistance, we competed WT with the spec^R-marked WT strain (WT/spec^R); the strains exhibited equivalent fitness over the 20-day infection period (Fig 7B). However, by Day 14, WT had outcompeted ∆camA nearly to the limit of detection, and at the conclusion of the experiment, the mean competitive index (CI) was 0.002 (Fig 7C). These data suggest that DNA methylation provides a fitness advantage for C. difficile, particularly during long-term colonization. To determine the extent to which ∆camA’s persistence defect is due to its impaired sporulation, we assessed the relative fitness of the Me3* mutant compared to WT and ∆camA. The Me3* mutant displayed a relatively mild fitness defect when co-infected with WT; this difference was not statistically significant (CI = 0.5) (Fig 7C). In contrast, Me3* outcompeted ∆camA (CI = 0.07) (Fig 7C). This suggests that the ability of CamA to enhance sporulation levels promotes C. difficile’s stable engraftment in a host, but it is not the primary driver of CamA’s fitness advantage. Together, these findings indicate that CamA-mediated DNA methylation regulates processes beyond sporulation that contribute to C. difficile’s ability to persist during murine infection.

C. difficile strain construction and growth conditions

All C. difficile strains used in this study are listed in S2 Table and are derivatives of 630∆erm. Mutant strains were constructed in a 630∆erm∆pyrE strain using the pyrE-based allele-coupled exchange, as described previously [65].

Strains were grown from frozen glycerol stocks on brain heart infusion medium supplemented with yeast extract (0.5% w/v) (BHIS) and taurocholate (0.1% w/v), thiamphenicol (10–15 µg/mL), kanamycin (50µg/mL), or cefoxitin (8 µg/mL) as needed. C. difficile defined medium (CDDM), supplemented with 5-fluoroorotic acid (2mg/mL) and uracil (5µg/mL) as needed, was used for constructing mutant strains. Sporulation analyses were carried out on cells cultured on 70:30 medium (70% BHIS and 30% SMC) as described previously [66]. For time-lapse experiments, strains were diluted into tryptone yeast extract broth supplemented with L-cysteine (0.1% w/v) (TYC) prior to inoculation onto 1.5% agarose pads supplemented with TYC medium. Resuspended and diluted fecal pellets collected during mouse experiments were plated on TCCFA plates with spectinomycin (500µg/mL) as needed. Cultures were grown at 37 °C under anaerobic conditions using a gas mixture containing 85% N₂, 10% H₂, and 5% CO₂.

E. coli strain construction and growth conditions

All E. coli strains used in this study are listed in S3 Table, with links to plasmid maps containing primer sequences used for cloning. Plasmids were cloned using Gibson assembly and subsequently transformed into E. coli DH5ɑ. Isolated plasmids were sequence confirmed using Oxford Nanopore Technology before being transformed into E. coli HB101 for conjugation with C. difficile or E. coli BL21(DE3) for protein expression.

E. coli strains were grown in Luria-Bertani (LB) broth at 37 °C with 225 rpm shaking, supplemented with chloramphenicol (20 µg/mL), ampicillin (100 µg/mL), or kanamycin (30 µg/mL) as needed. For protein production, E. coli BL21(DE3) strains were grown in Terrific Broth (Thermo Fisher) supplemented with glycerol (0.5%), glucose (0.05%), and ɑ-lactose monohydrate (0.1%) at 20 °C for 60 hours with 225 rpm shaking.

B. subtilis strain construction and growth conditions

All B. subtilis strains used in this study are listed in S4 Table, with links to plasmid maps containing primer sequences used for cloning. Plasmids were linearized before transformation into B. subtilis. Positive integrants were confirmed by marker exchange.

B. subtilis strains were grown on LB medium at 37°C, supplemented with chloramphenicol (5 µg/mL) or spectinomycin (100µg/mL) as needed. Sporulation was induced by nutrient exhaustion by resuspension according to the Sterlini and Mandelstam method, as previously described [67,68].

Sporulation induction

Starter cultures were grown in BHIS broth until they reached early stationary phase, then back-diluted to an OD₆₀₀ of 0.05 and grown until they reached an OD₆₀₀ of 0.35 to 0.75. 120 μL of these cultures were spread on 70:30 plates (40 mL media per plate).

Heat resistance assay

After 22–24 hours of growth on 70:30 plates, sporulating cells were resuspended in phosphate-buffered saline (PBS), after which the culture was split into two. One sample was heat-treated at 60 °C for 30 minutes, while the other was plated. Both heat-treated and untreated samples were serially diluted and plated for viable count. Heat-resistance efficacies represent the average ratio of heat-resistant CFU to total CFU for a given strain relative to the ratio for the WT strain.

Spore purification

After a minimum of 64 hours of growth on 70:30 plates, cells were resuspended in ice-cold sterile water, washed 5–8 times, and incubated on ice overnight. Samples were then treated with DNase I (New England Biolabs) for 1 hour at 37°C, washed an additional 2 times in ice-cold water, and purified with a 20% - 50% Histodenz (Sigma Aldrich) gradient. After 2 final washes with water, spore purity was assessed using phase-contrast microscopy, and cells were enumerated by viable count.

RNA isolation

After 9–11 hours of growth on 70:30 plates, RNA was harvested from sporulating cells as previously described [69]. Briefly, samples were processed using the FastRNA ProBlue Kit (MP Biomedical) and a FastPrep automated homogenizer (three cycles of 40 seconds at setting 6.0; MP Biomedical). Contaminating genomic DNA was depleted by two sequential in-solution DNase treatments (New England Biolabs), followed by a column-bound DNase treatment with an RNeasy Kit (Qiagen). The DNase-treated RNA was enriched for mRNAs using a MICROBExpress Bacterial mRNA Enrichment Kit (Invitrogen). Finally, the SuperScript First-Strand Synthesis System (Invitrogen) was used for reverse transcription of the enriched RNA to generate cDNA using random hexamer primers.

RT-qPCR analysis

RT-qPCR was performed as previously described [10], using the iTaq Universal SYBR Green supermix (BioRad) in a Mx3005P qPCR instrument (Stratagene). The following cycling conditions were used: 95 °C for 2 minutes, 40 cycles of 95 °C for 15 seconds, and 60 °C for 1 minute. Transcript levels were normalized to the housekeeping gene rpoB using the standard curve method. Gene-specific primer pairs are listed in S5 Table.

Membrane, and SNAP labeling

FM4-64 (1 µg/mL; Invitrogen) was added to agarose pads to stain cell membranes. For labeling of strains encoding SNAP-tag fusions, cells were washed once with 0.5% Bovine Serum Albumin (BSA) in PBS and incubated with SNAP-tag 505-star or TMR-star (10 µM; New England Biolabs) at 37 °C in the dark. After 30 minutes, cells were washed three times in PBS and resuspended in PBS. All cell labeling was performed prior to fixation.

Cell fixation

Cells were fixed as previously described [70]. Briefly, a 5X fixation solution (containing 20 µL 16% paraformaldehyde and 100 µL 1M NaPO₄ buffer) was added to 500µL culture. Samples were incubated for 30 minutes at room temperature, followed by 30 minutes on ice. Fixed cells were washed three times in PBS and imaged within 48 hours of fixation.

Microscope hardware

All samples were imaged on agarose pads made with TopVision Low Melting Point agarose diluted in PBS (1.5%; Thermo Fisher) and sealed with a coverslip. For phase-contrast and fluorescent micrographs, images were acquired using a Leica DMi8 inverted microscope equipped with a HC plan apochromat 64x 1.5 NA oil immersion phase contrast objective, as previously described [44,71]. Excitation light was generated with a Lumencor Spectra-X multi-LED light source; for all fluorescent proteins aside for SNAP-tag 505-star, this was coupled with an XLED-QP quadruple-band dichroic beam splitter and an external emission filter wheel (Leica). FM4–64 was excited at 550/38 nm, and the emitted light was filtered using a 705/72 nm emission filter; mScarlet, mScarlet-I3 and SNAP-tag TMR-star were excited at 550/38 nm, and the emitted light was filtered using a 590/50 nm emission filter. Images of SNAP-tag 505-star were captured with a YFP filter set (Chroma), equipped with a 500/20 nm excitation filter, a 515-nm dichroic filter, and a 525/30 nm emission filter. 1–2 µm z-stacks were taken when needed. All imaging was carried out a 37 °C using a microscope incubation system (Pecon).

Phase-contrast microscopy without fluorescence was performed using a Zeiss Axioskop upright microscope with a 100x Plan-NEOFLAUR oil-immersion phase-contrast objective and a Hamamatsu C4742-95 Orca 100 CCD Camera.

Time-lapse microscopy

Time-lapse imaging was performed as previously described [44]. Briefly, an anaerobic imaging chamber was constructed by layering two gas-impermeable 125 µL Gene Frames (Thermo Fisher) on a glass slide. Inside the anaerobic chamber, the Gene Frames were filled with pre-reduced TopVision Low Melting Point agarose (1.5%; Thermo Fisher) and TY media containing L-cysteine (0.1% w/v); a second glass slide was placed on the liquid media to create a flat surface; the slide was placed on a frozen freezer block until the agarose solidified. The second glass slide was removed, and the agarose pad was dried for 10 minutes prior to loading the cells and sealing the imaging chamber with a coverslip. The samples were imaged at 37˚ C at 2.5-minute intervals until they reached confluency in the field of view.

Imaging analysis and quantification

After image acquisition, images were exported and processed in FIJI. Images were cropped to remove any out-of-focus cells, and the best-focused z-plane for each channel was selected to correct for chromatic aberration. To improve display, brightness and contrast settings were scaled and applied equally to all images shown in a single plane. At least three images were captured per replicate, and every strain was analyzed with three biological replicates. For cell segmentation and quantification of fluorescence intensities, images were additionally processed. First, Instant Computational Clearing (ICC) was performed (LASX software; Leica) to avoid bleed-through of fluorescent signal into neighboring cells. The adaptive strategy was run with the feature scale set to 2683 nm and 98% strength. Following ICC, images were processed as indicated above to remove all out-of-focus cells. To quantify the mean fluorescence intensity per cell, the MATLAB-based image analysis pipeline SuperSegger [72] was used, with the supplied “60x E. coli” settings. For analysis of dual reporters presented in Fig 4, images were first processed via the SuperSegger image analysis pipeline, after which cells were manually categorized by σ^F-activation status.

Transmission electron microscopy

After 10 hours of growth on 70:30 plates, sporulating cells were fixed and processed for electron microscopy by the University of Vermont Microscopy Center as previously described [73]. Briefly, cells were fixed in 2% paraformaldehyde and 2% glutaraldehyde in 0.1 M sodium cacodylate buffer for 2 hours at 4°C, then washed with 0.1 M sodium cacodylate buffer. Samples were embedded in 2% agarose and cross-linked with 1% paraformaldehyde and 2.5% glutaraldehyde in 0.1 M cacodylate buffer, then washed with 0.1 M cacodylate buffer. After being minced into 1mm³ pieces, samples were dehydrated in a graded ethanol series (35%, 50%, 70% 85%, 95% and 100%) and cleared twice in 100% propylene oxide. Samples were infiltrated with Spurr’s epoxy resin in 100% polypropylene oxide in increasing ratios, then embedded in 100% Spurr’s resin and polymerized at 70°C. First, semi-thin sections (1 µm) were sliced on a Reichart Ultracut Microtome and stained with methylene blue-azure II; next, ultra-thin sections were cut with a diamond knife, retrieved on 200-mesh thin-bar nickel grids, and contrasted with uranyl acetate (2% in 50% ethanol) and Reynolds’ lead citrate. Images were captured on a JEOL 1400 Transmission Electron Microscope (Jeol USA).

Protein purification

E. coli BL21(DE3) encoding lactose-inducible, His₆-tagged proteins of interest (SpoIIE_∆1–905, Spo0A DNA binding domain (DBD), full-length Spo0A and CamA) were grown to stationary phase, back-diluted 1:1000 in 1 L Terrific Broth (Thermo Fisher) supplemented with glycerol (0.5%), glucose (0.05%) and ɑ-lactose monohydrate (0.1%). Cell cultures were grown at 20 °C with 225 rpm shaking; after 60 hours, cultures were pelleted, resuspended in 25 mL low imidazole buffer (LIB; 500 mM NaCl, 50 mM TRIS-HCl pH 7.5, 15 mM imidazole, 10% glycerol, 2 mM β-mercaptoethanol), and flash frozen in liquid nitrogen. Once thawed, cells underwent three cycles of probe sonication consisting of 45 seconds at 40% amplitude followed by 5 minutes on ice. Samples were pelleted at 1000 rpm for 45 minutes at 4 °C, and tagged proteins were affinity-purified from cleared lysates using Ni-NTA agarose beads with gentle rocking at 4 °C for 2 hours. Beads were washed three times with LIB and protein was eluted with high imidazole buffer (HIB; 500 mM NaCl, 50 mM TRIS-HCl pH 7.5, 200 mM imidazole, 10% glycerol, 2 mM β-mercaptoethanol). Beads were washed 6 times with HIB, pelleted, and the supernatant containing the eluted protein was collected and analyzed using Coomassie staining.

Affinity-purified proteins were then concentrated using an Amicon Ultra-15 10 kDa cutoff centrifugal filter (Millipore Sigma) and further purified by size exclusion chromatography (SEC) using a Superdex 200 Increase 10/300 GL column and 200 mM NaCl, 10 mM Tris pH 7.5, and on an AKTA pure protein liquid chromatography instrument. Fractions were collected every 0.5 mL, and fractions of interest were reconcentrated, aliquoted, and flash-frozen in liquid nitrogen.

Fluorescence polarization assay

25-30 bp double-stranded DNA (dsDNA) was generated by incubating equal molar ratios of 6-Carboxyfluorescein-labeled oligonucleotides with their unlabeled complementary counterpart in annealing buffer (10 mM TRIS-HCL pH 7.5, 1 mM EDTA, 50 mM NaCl) at 95 °C for 15 minutes and gradually cooling the solution to room temperature. All oligonucleotide sequences are provided in S4 Table. 1 nM dsDNA was mixed with serially diluted, purified Spo0A in fresh binding buffer (50 mM Tris-HCl pH 8, 100 mM KCl, 2.5 mM MgCl₂, 0.2 mM DTT, 10% glycerol, and 2 µg salmon sperm DNA in H₂O). After 10 minutes of incubation at room temperature, fluorescence polarization was read using a Synergy H1 plate reader (Agilent BioTek). 6-Carboxyfluorescein was excited at 485/20 nm, and emission was detected at 528/20 nm.

Electromobility shift assays

250 bp DNA fragments encompassing the spoIIE promoter were amplified from purified C. difficile 630∆erm genomic DNA using IRDye800-conjugated or regular primers (Integrated DNA Technologies). P_spoIIE-specific pairs are provided in S5 Table. The resulting labeled and cold competitor DNA probes were PCR- and gel-purified prior to use. 20 fmol labeled DNA (or 20 fmol labeled with 1000 fmol cold competitor DNA) was mixed with purified Spo0A-DBD-His₆ in binding buffer (10 mM Tris-HCL, pH 7.6, 1 mM EDTA, 50 mM NaCl, 1 mM DTT, 5% glycerol) for 30 minutes at 37 ˚C. 6X loading dye (40% w/v sucrose, 0.25% bromophenol blue, and 0.25% xylene cyanol FF) was added to 1X before samples were loaded and run on an 8% native polyacrylamide gel at 80 V at 4 ˚C in the dark.

Methylation of P_spoIIE

The enzymatic activity of purified CamA-His₆ was confirmed using the MTase-Glo Methyltransferase Assay Kit [74,75] (Promega). Briefly, 60 bp dsDNA were generated as described in the above section; all oligonucleotide sequences are provided in S5 Table. In white assay plates (CORNING), 5 µM dsDNA was added to a mixture of 40 µM S-adenosyl-_L-methionine and CamA at the indicated concentrations in a reaction buffer (80 mM Tris-Buffer, pH 8.0, 200 mM NaCl, 4 mM EDTA, 12 mM MgCl₂, 0.4 mg/mL BSA, 4 mM DTT). After 30 minutes of incubation at room temperature, the protocol was completed per the manufacturer’s instructions. Luminescence signal was measured by a Synergy H1 plate reader (Agilent BioTek). To methylate DNA probes for electromobility shift assays, 200 nM DNA fragments, encompassing the spoIIE promoter, were incubated with 1 µM affinity- and SEC-purified CamA-His₆ for 30 minutes at room temperature and PCR purified.

Protein degradation and western blot

After 11 hours of growth on 70:30 media, sporulating cells were resuspended in BHIS broth. Translation was inhibited with the addition of chloramphenicol (100 µg/mL) and transcription was inhibited by the addition of rifampicin (100 µg/mL); samples were removed at the indicated time points, pelleted at 4°C, and frozen at -80 °C. Samples were prepared for immunoblotting as previously described [69]. Briefly, they underwent three freeze-thaw cycles followed by the addition of EBB buffer (9 M urea, 2 M thiourea, 4% SDS, 2 mM beta-mercaptoethanol) and then were incubated at 95 °C with periodic vortexing. Pellets were resuspended, and bromophenol blue was added to 0.01% (w/v) to visualize the samples. Samples were vortexed vigorously, boiled, and pelleted again, immediately prior to loading on a gel. Proteins were resolved on a 12% SDS-polyacrylamide gel electrophoresis (SDS-PAGE) gel.

Proteins were transferred to polyvinylidene fluoride membranes (PVFD), which was washed and blocked with Odyssey blocking buffer (LiCor) for 30 minutes. PVDF membranes were then probed with rabbit anti-SpoIIE (1:1000 dilution; generated by CoCalico Biologicals against SpoIIE_906–2370 and further purified using affinity subtraction) and chicken anti-GDH (1:5000 dilution; ThermoFisher) polyclonal primary antibodies. After washing in PBS with Tween, PVFD membranes were probed with goat anti-rabbit IRDye680 and donkey anti-chicken IRDye800 secondary antibody (1:12,000 dilution; LiCor Biosciences).

Growth curve assays

Starter cultures were grown in BHIS broth until they reached early stationary phase and then back-diluted to an OD₆₀₀ of 0.05. Once grown to mid-log phase, cultures were normalized to a starting OD₆₀₀ of 0.5. Cells were diluted 1:30 in BHIS, with antibiotics supplemented as indicated, and 150 µL was distributed into wells in technical triplicate. Plates were incubated in an Epoch plate reader (Agilent BioTek) at 37 °C in the anaerobic chamber with linear shaking every 2 minutes. OD₆₀₀ values were recorded every 15 minutes.

Mouse infection experiment

Groups of seven-week-old female C57BL/6 mice (purchased from Jackson Laboratory) were housed together in a large, sterile rat cage to normalize their microbiota. After 10 days, cefoperazone (0.5mg/mL) was administered in their drinking water, which was provided ad libitum for 10 days and replaced every 2 days. Mice were returned to regular drinking water for 2 days before receiving a single intraperitoneal injection of clindamycin (10 mg/kg in 200 μL PBS). After 24 hours, groups of 4 mice were transferred to smaller cages and inoculated via oral gavage with a total of 1 × 10⁵ purified spores in 200 µL PBS. The spore solution contained a 1:1 mixture of strains of interest, with one strain harboring the aad9 gene encoding spectinomycin resistance integrated into the pyrE locus. Weight and body condition scores were recorded, and fecal pellets were collected at the indicated times. Fecal pellets were weighed prior to being resuspended in 1 mL PBS, and 10-fold dilutions were plated on TCCFA agar in the presence and absence of spectinomycin (500 µg/mL). Colony-forming units were counted and normalized to the mass of the original fecal pellet. The competitive index was calculated as (Output_SpecS/Output_SpecR)÷ (Input_SpecS/Input_SpecR), or vice versa. Cage changes were performed every 2 days, and mice were fed irradiated Lab Diet 2918 throughout. At the experimental endpoint, mice were euthanized with CO₂ asphyxiation followed by cervical dislocation.

Data visualization and statistics

EMSAs and western blots were visualized using a LiCOR Odyssey CLx Imager. All graphs and subsequent statistical tests were generated using GraphPad Prism.