YabG is a cysteine protease with a CheY-like fold

In C. difficile, yabG (annotated as CD630_35690 in the genome of strain 630), is followed by a homologue of the B. subtilis veg gene, itself just upstream of sipL, coding for a protein essential for assembly of the coat/exosporium layers of spores [53,54] (Fig 1A). In B. subtilis, veg, which is also downstream of yabG (S1A Fig), was shown to have a role in biofilm formation and is expressed in both vegetative and sporulating cells [55,56]. Recent work showed that B. subtilis YabG is a cysteine protease and suggested that it uses a catalytic dyad formed by Cys218 and His172 in an auto-proteolytic activity that causes its rapid degradation (Fig 1B) [52]. B. subtilis YabG, together with the C. difficile protein and 202 other orthologues from sporeformers, was included in a new family of clan CD in the MEROPS peptidase database [52,57].

Download:

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

Fig 1. The yabG region of the C. difficile chromosome and organization of YabG.

A: Schematic representation of the yabG region of the C. difficile chromosome. A σ^E/K-dependent promoter in front of the yabG gene is represented. The lines below the gene represent the regions coding to distinct regions of the YabG protein (as in panel C). B: The alignments show highly conserved or invariant residues (grey) in the vicinity of Cys residues at positions 119 (blue) and 207 of YabG (brown) and around His161 and Asp162 (brown). C: An AlphaFold2 model of C. difficile YabG. The model highlights a N-terminal SH3-like domain (A, orange; residues 1–57) and a C-terminal CheY-like domain (B, green; residues 99–286) connected by a linker (L, blue, residues 58–98). A putative active site cleft is indicated at the interface between A and B. D: Expansion of the active site region to show the relative positions of Cys207, His161, Asp162 and Cys119 with estimated distances shown in Å. E and F: The WT and the indicated forms of YabG and CspBA were produced in E. coli. In F, the indicated forms of YabG were produced together with or in the absence of CspBA The proteins in whole cell extracts were resolved by SDS-PAGE and the gels stained with Coomassie or subject to immunoblotting (in F) with anti-CspB antibodies. In E and F, the black arrowheads show the position of YabG, YabG^C207A or CspBA. The green arrowhead in F shows the position of CspB released from CspBA, independently of YabG, while the red arrowhead points to CspB released from CspBA through the action of YabG. The asterisk refers to an unspecific band.

https://doi.org/10.1371/journal.ppat.1011741.g001

An alignment of the residues around Cys218 of the B. subtilis YabG protein with the corresponding region from selected orthologues, reveals a highly conserved region with an invariant Cys residue, C207, in C. difficile YabG (Fig 1B). Likewise, His161 in the C. difficile protein, homologous to His172 in B. subtilis YabG, is also invariant among YabG orthologues. Thus, the two catalytic residues suggested for B. subtilis YabG are conserved. We noticed, however, that an Asp residue, at position 162 in the C. difficile protein, is also strictly conserved (Fig 1B). Cysteine proteases share an acylation-deacylation mechanism involving a nucleophilic cysteine thiol that is part of a Cys-His dyad or a Cys-His-Asp/Asn catalytic triad [58,59]. Thus, Asp162 could also be important for activity.

To gain more insight into the structure and function of YabG, we generated and refined a structural model using the machine learning approach to protein structure prediction implemented through AlphaFold2 [60]. The model generated for C. difficile YabG suggests that the protein consists of two independent structural domains, A (Fig 1C, residues 1–57, in orange) and B (residues 99–286, in green) connected by a linker, L (residues 58–98, in blue) (see also S2 Fig). A search using the Fold Seek server [61] indicates structural similarity of domain A to proteins with an Src homology domain 3 (SH3), including for example the cyanobacterial protein PetP and Escherichia coli HspQ. PetP is a subunit of cytochrome b6f [62], while HspQ is both a substrate and a specificity and allosteric enhancing factor for the Lon protease [63,64] (S3 Fig). SH3 domains are small, 55–70 residues-long protein-protein interaction modules [65,66].

Fold Seek also revealed structural similarity of B to the receiver domain of response regulators including, among the top hitters, the KdpE protein of E. coli (pdb identifier: 4I85), MicA from Streptococcus pneumoniae (7m0s), MtrA from Mycoacterium tuberculosis (3nhz) and CheY from Vibrio cholerae (4hnr) and Thermotoga maritima (4tmy). All of these proteins share the fold of the archetypal CheY response regulator from E. coli [67–69]. The superimposition of the model obtained for domain B of YabG with the crystal structures of KdpE and CheY highlights the similarity to the CheY fold (S4 Fig). KdpE, MicA and MtrA have a receiver domain and a C-terminal DNA binding domain and function as transcription factors [70–72]. CheY, in turn, is a single domain response regulator which upon phosphorylation of an Asp residue, binds directly to the flagellar motor [67]. Most response regulators function through phosphorylation of a conserved aspartate residue [67–69] but others function independently of phosphorylation [68,69]. The phospho-acceptor residue in CheY is D57; this residue is not conserved in YabG with the homologous position occupied by Thr159 in C. difficile YabG and by Thr164 in the B. subtilis protein (Fig 1B). Thus, YabG may function independently of phosphorylation.

Residues of YabG involved in auto-proteolysis

The AlphaFold2 model suggests that the catalytic Cys207 in B is at the bottom of a cleft formed between the A domain and the top of B (Fig 1C). The distances estimated between Cys207 and His161 from the model, and between the latter and Asp162 (Fig 1D) are longer than in other Cys proteases and the side chain of His161 is not oriented towards Cys207 [58,59]. However, of the 204 YabG sequences in the MEROPS U57 family, where YabG used to be included [57], 96.1% have His, Asp and Cys residues conserved at equivalent positions. A second Cys residue at position 119 in the C. difficile protein is also invariant among YabG orthologues (Fig 1B).

To test whether C. difficile YabG showed auto-proteolytic activity and if so, what residues of the putative active site were involved, we overproduced YabG^WT in E. coli, using an auto-induction regime [73], as a N-terminal His₁₀ fusion. His₁₀-YabG^WT did not accumulate in whole cell extracts, as assessed by Coomassie staining (Fig 1E). In contrast, B. subtilis YabG accumulated but underwent auto-proteolytic degradation over time [52]. Why the B. subtilis appears more stable than its C. difficile counterpart is presently unknown. To test whether Cys207, His161 and Asp162 were required for YabG auto-proteolysis, YabG^C207A, YabG^H161A and YabG^D162A were also overproduced in E. coli. As assessed by Coomassie staining, the YabG^C207A, YabG^H161A and YabG^D162A forms of the protein accumulated in the extracts as species of about 34 kDa, consistent with the predicted size of the protein (34.7 kDa), indicating that the substitutions impaired auto-proteolysis (Fig 1E). In contrast, another single alanine substitution, D248A in domain B, did not cause the protein to accumulate in extracts (Fig 1E). Only substitutions that impair protease activity allow YabG to accumulate. Thus, YabG residues Cys207 and His161, homologous to the dyad described for the B. subtilis protein, and Asp162, are required for auto-proteolysis. We also tested whether a variant with Cys119 changed to Ala accumulated in extracts. In contrast to YabG^C207A, the YabG^C119A form did not accumulate in whole cell extracts (S5A Fig). Thus, Cys119 is not required for the auto-proteolytic activity of YabG.

Auto-proteolysis of B. subtilis YabG leads to the accumulation of transiently stable fragments through cleavage after Arg5, Arg17, Arg49 and Arg93 [52]. Of these, Arg17 and Arg49 are conserved in C. difficile YabG (Arg9 and Arg41, respectively); the position homologous to Arg93 is occupied by a Lys in the C. difficile protein but there are two Arg´s in the vicinity of this residue (S5B Fig). This suggests that the C. difficile protein may also be cleaved at least after Arg9 and Arg41. That C. difficile YabG is likely to be specific for Arg at the P1 position is in line with cleavage of the SleC and CspBA after Arg residues [43].

C207 is required for YabG-dependent CspBA interdomain processing

CspBA accumulates in spores of yabG mutants [42,43,74]. Release of the CspB domain from CspA is necessary for the activation of the cortex hydrolase SleC, required for germination completion [47,51]. YabG was found to be involved in CspBA processing [42,43]. To assess the role of Cys207 in the reaction, YabG^WT or YabG^C207A were co-produced in E. coli together with CspBA. Production of CspBA alone resulted in the accumulation of a species of about 125 kDa, consistent with the expected size of CspBA (MW 124.6 kDa), detected with anti-CspB antibodies (Fig 1F, black arrowhead in the bottom panel). A species of about 60 kDa was also detected (Fig 1F, green arrowhead). This species, termed CspB*, results from alternative processing that occured in a YabG-independent manner [42]. In this experiment, and as described above, only YabG^C207A was detected (Fig 1F). However, even though YabG^WT was not detected, its co-production with CspBA resulted in the disappearance of CspBA and the accumulation of a species of about 55 kDa, the size expected for the CspB moiety, detected with anti-CspB antibodies (Fig 1F, red arrowhead). Accumulation of this protein required the activity of YabG: the co-production of CspBA together with YabG^C207A did not lead to depletion of CspBA or to the accumulation of CspB, whereas the YabG-independent CspB* still accumulated (Fig 1F, bottom panel).

We conclude that Cys207 is required for CspBA interdomain processing, consistent its predicted role in catalysis and with the accumulation of the precursor protein in yabG spores [42,43].

yabG is expressed in the mother cell during late stages of spore morphogenesis

Synthesis of the proteins that form the spore coat and exosporium layers is driven by σ^E and σ^K, which are mainly active before and after engulfment completion, respectively [15,75]. In B. subtilis, yabG is under the control of σ^K [39] and genome-wide transcriptional profiling studies of C. difficile sporulating cells have suggested that yabG is also under the control of σ^K in this organism [30,31,35]. Consistent with these studies, putative -10 and -35 promoter elements that match the consensus for σ^K recognition are present in the yabG regulatory region (S6A Fig) [30,76]. To determine the time of yabG expression relative to the stages of spore morphogenesis, we made use of a transcriptional fusion between the yabG promoter region and the SNAP^Cd reporter [29,30,77]. The P_yabG-SNAP^Cd fusion was introduced into the WT strain 630Δerm as well as into congenic sigE::erm and sigK::erm mutants [29]. The resulting strains were inoculated on 70:30 agar plates [53] and imaged by phase contrast and fluorescence microscopy, following labelling with the SNAP substrate TMR-Star, 14 and 20 hours thereafter. In the WT, expression of P_yabG-SNAP^Cd at hour 14 (S6B Fig, top panels) and at hour 20 (bottom panels) was confined to the mother cell. A fluorescence signal from P_yabG-SNAP^Cd was undetected in the sigE::erm mutant, but still detected in the mother cell of a sigK::erm mutant (S6B Fig). This suggests dual control of yabG expression by σ^E and σ^K. The consensus for σ^E recognition is also included in S6A Fig; the highly conserved ATA motif for σ^E recognition in the -35 region is absent, but the putative -10 region conforms better to the consensus for σ^E binding than to the consensus for σ^K binding [76].

To determine the main period of yabG expression, we measured the expression of P_yabG-SNAP^Cd during spore morphogenesis. P_yabG-SNAP^Cd expression was detected at 20 hours of growth in 45% of the cells during asymmetric division and engulfment, in 46% of the sporangia of phase-dark forespores, in 45% of the sporangia of phase-grey forespores, and in 53% of sporangia of phase-bright forespores (S6B Fig, yellow arrowheads). The average fluorescence intensity from P_yabG-SNAP^Cd in sporangia of phase-bright forespores at hour 20 was higher than that of sporangia of phase-dark or phase-grey spores or cells during asymmetric division and engulfment (S6C Fig). At hour 14 the percentage of sporangia with a signal was highest for sporangia of phase-dark spores and the intensity of the fluorescence signal per cell remained relatively uniform regardless of the developmental stage (S6B and S6C Fig). While no expression was detected in a sigE mutant, the intensity of the fluorescence signal for P_yabG-SNAP^Cd in the sigK::erm mutant was decreased relative to the WT, particularly at hour 14 (S6C Fig). Nevertheless, 90 percent of the sporangia of the sigK mutant at hour 14 and 74 percent at hour 20 showed a fluorescence signal presumably because of persistent activity of σ^E. Together, these results indicate that the onset of yabG expression in the mother cell occurs soon after asymmetric division, under the control of σ^E, when the mother cell starts engulfing the forespore; it then increases towards the final stages of sporulation, with the contribution of σ^K, when the forespore transitions from phase-dark/grey to phase-bright (S6 Fig)

The activity of YabG is required for proper spore germination

Amino acids such as alanine or glycine act as co-germinants during spore germination triggered by cholic acid derivatives as for instance taurocholate (TA) (reviewed in [78]). Previous work has shown that deletion of or point mutations in yabG allowed spore germination in response to TA alone in epidemic strain R20291 [43]. To examine the role of YabG in the germination of C. difficile 630Δerm spores, we constructed a congenic yabG in-frame deletion mutant, ΔyabG, using a CRISPR-Cas9 system (S7A and S7B Fig). We constructed two additional strains in the ΔyabG background, with either a copy of the WT yabG gene (yabG^C) or yabG^C207A at the non-essential pyrE locus (S7C and S7D Fig). In the background of strain 630Δerm we found that spores of the yabG mutants germinated slower than WT spores in response to TA in a rich medium (S8A, S8B and S9 Figs). Previous work has shown that the efficiency of germination for ΔyabG spores, as assessed by plating spores exposed to TA onto plates of a rich medium containing the germinant was 0.8 of the WT [42]. Using the same assay, we obtained a plating efficiency of 0.71±0.13 for ΔyabG spores, 0.89±0.13 for yabG^C207A spores and 0.96±0.18 for yabG^C spores (S8C Fig, top panel). Germination involves release of dipicolinic acid (DPA) from the spore core. DPA accumulated and/or was retained by the spores can also influence germination and thus the impaired germination could result from reduced extrusion of DPA. To test this point, we measured the DPA content of the various spores. We found the DPA content, normalized to the WT, of ΔyabG (125.5% ± 11) and yabG^C207A (118.8% ± 10) spores to be not significantly different than that of WT (100%) and yabG^C (116.7% ± 2) spores (S8C Fig, middle panel). Moreover, WT spores released 0.4 ± 0.04 DPA (expressed as a ratio of the OD₂₇₀/OD₆₀₀) during germination, yabG^C spores released 0.3 ± 0.02 DPA, while ΔyabG and yabG^C207A spores released respectively 0.3 ± 0.01 and 0.4 ± 0.09 DPA (S8C Fig, bottom panel). Thus, 1 hour after induction of germination, the DPA released from yabG mutant spores is not significantly reduced compared to WT and yabG^C spores.

Only CspB (55 kDa) was detected in extracts from WT or yabG^C spores; in contrast, both CspBA and CspB* (see above) were detected in ΔyabG and yabG^C207A spores (S8D Fig) [42]. Pre-pro-SleC (47 kDa) was processed to its Pro form (34 kDa) in WT and yabG^C spores, but only full-length SleC was detected in ΔyabG or yabG^C207A spores (S8D Fig, middle panel). Finally, the levels of CspC did not differ significantly between WT and ΔyabG spores or between yabG^C and yabG^C207A spores (S8D Fig, bottom panel). Thus, the partial germination defect of ΔyabG and yabG^C207A spores may be due, at least in part, to loss of YabG activity, which in turn impairs the release of CspB from CspBA and the production of pro-SleC. Because germination is also influenced by the status of the spore surface layers, we next wanted to investigate a possible role of YabG in the assembly of the coat and exosporium.

YabG controls the assembly of various coat and exosporium proteins

We first examined the collection of coat/exosporium proteins that could be extracted from purified WT, ΔyabG, yabG^C and yabG^C207A spores. Spores of the four strains were purified on gradients of metrizoic acid, and the proteins extracted in a buffer with SDS and reducing agents; the resulting spores were then treated with lysozyme and proteins again extracted (see the Material and Methods). This produced a cortex/coat/exosporium fraction and a core/cortex fraction (as explained below). Proteins in the two fractions were then analysed by SDS-PAGE and immunoblotting. In Coomassie-stained gels, several proteins were more extractable or more abundant in the cortex/coat/exosporium fractions from ΔyabG or yabG^C207A spores (Fig 2A, red arrowheads). Mass spectrometry analysis identified SpoIVA in the band at around 70 kDa (the expected mass of SpoIVA), CotE in a band at around 81 kDa (close to the expected mass of the full-length protein, 81.3 kDa) (Fig 2A, red arrowheads) [23,53,54]. Also, the abundance or extractability of a form of CdeC was also slightly increased in the ΔyabG mutant relative to the WT (Fig 2A, band close to the 50 kDa marker in the Coomassie-stained gel and the immunoblot). Also note that for the ΔyabG and yabG^C207A mutants, a form of CdeC that migrates above the 75 kDa marker is also more abundant or extractible as detected by immunoblotting (Fig 2A). The increased extractability of SpoIVA from ΔyabG spores has been observed before [42]. CotE, found in both the coat and exosporium, is a bi-functional enzyme with a peroxiredoxin N-terminal domain and a C-terminal chitinase domain [23,79,80]. CotE is detected in extracts from WT spores as species of 81 kDa, 40 kDa and a minor species of 38 kDa [23,79]. The accumulation of the 81 kDa species in the yabG mutants, suggests that YabG is required for CotE interdomain processing through which the peroxiredoxin and chitinase moieties are separated. In contrast, CdeM and CotA, were absent or much less extractable from both ΔyabG or yabG^C207A spores as assessed by Coomassie staining (Fig 2A). CdeM, a major component of the exosporium of C. difficile spores, is detected by immunoblotting in the cortex/coat/exosporium fraction from WT and yabG^C spores as an abundant species at around 17 kDa [24,25] (Fig 2A). CotA is a surface exposed protein required for the assembly of the outer spore surface layers [79,80]. The protein was detected in the cortex/coat/exosporium fraction of WT and yabG^C spores as a main species in the 34 kDa region of the gel, consistent with its predicted mass (34 kDa) but also as a species just above the 25 kDa marker (Fig 2A); both forms were previously observed [80]. CotA and CdeM remained undetected in the core/cortex fraction (Fig 2A). YabG itself (32 kDa) could only be detected in the cortex/coat/exosporium extracts prepared from yabG^C207A spores (Fig 2A, green arrowhead) presumably because the auto-proteolytic activity of YabG^WT limits its accumulation. Alternatively, YabG^WT may be much less extractable than YabG^C207A (see also below). As expected, the spore core protein GPR was detected by immunoblotting only in the core/cortex fraction (Fig 2A). SleC, however, shown before to be associated with the cortex [74] was detected in the cortex/coat/exosporium fraction of WT or yabG^C spores, mostly in its processed form (Fig 2A). Trace amounts of Pre-pro-SleC and not the processed form were detected in the core/cortex fraction of WT spores; this may reflect different extractability of the protein from spores of 630Δerm relative to strain R20291 [74]. However, the processed form was clearly visible in this fraction of yabG^C spores (Fig 2A). Only the Pre-pro-SleC protein was detected in yabG mutants, in either fraction (Fig 2A). While enriching for coat/exosporium proteins, which are not detected in the core/cortex fraction, our extraction procedure also releases cortex-associated proteins, at least SleC. That proteins known to be associated with the cortex are extracted with the coat/exosporium has been reported before [81].

Download:

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

Fig 2. yabG mutations affect the extractability of coat and exosporium proteins and result in abnormal spore surface layers.

A: Spores of the WT, ΔyabG, yabG^C (complementation strain) and yabG^C207A strains were purified on density gradients, fractionated [107] and the cortex/coat/exosporium and cortex/core proteins extracted. The proteins were resolved by SDS-PAGE and the gels stained with Coomassie (top panel) or subjected to immunoblot analysis (lower panels) with anti-YabG, anti-SleC, anti-GPR, anti-CdeC, anti-CotA and anti-CotM antibodies, as indicated. The red arrowheads indicate proteins that appear to be more extractable from ΔyabG and yabG^C207A spores while black arrowheads show proteins with reduced extractability. The position of YabG^C207A is indicated by a green arrowhead; asterisks show the position of non-specific species. Proteins in the indicated bands in the Coomassie-stained gel were identified by mass spectrometry. B: Spores of the indicated strains imaged by scanning electron microscopy. The yellow arrowheads point to the polar regions in yabG and yabG^C207A spores. The red arrowhead points to material that seems to peel-off the pole of a yabG^C207A spore. C: Spores of the WT (630Δerm) and yabG^C207A mutant were analysed by thin sectioning transmission electron microscopy (TEM). The regions within the red and yellow circles in the two left panels are magnified on the right panels. In WT spores, the yellow arrowheads point to the coat region and the blue arrowheads to the exosporium region. In yabG^C207A spores, the red arrowheads point to the lamellae seen in the appendage region, the green arrowheads to electron dense coat or exosporium material loosely attached to yabG^C207A spores and the black arrowheads to unstructured material present between the cortex and the coat layers. Also note the material peeling off from the appendage (brown arrowheads). Cr, spore core; Cx, cortex; Ap, appendage region. The numbers refer to the percentage of spores in which the coat is detached from the cortex (red) or with a long polar appendage with a lamellar structure (blue); 60–95 spores were counted for each strain. Scale bars: 3 μm in B; 500 nm (left column) or 100 nm (all other panels) in C. In B and C, the spores were purified on density gradients. See also S12 Fig.

https://doi.org/10.1371/journal.ppat.1011741.g002

In all, the ΔyabG or yabG^C207A mutations affected the extractability of proteins important for coat (SpoIVA, CotE, CotA) and exosporium morphogenesis (CdeM) and processing of Pre-pro-SleC.

yabG spores are more permeable to lysozyme

Since the integrity of the surface layers is important to prevent access of peptidoglycan-breaking enzymes to the spore cortex, we tested the resistance of spores to lysozyme treatment. Density-gradient purified spores were plated in a rich medium in the presence of the germinant taurocholate, before or after treatment with lysozyme. Survival was of 98.3% for WT spores, 69% for ΔyabG spores 72.3% for yabG^C207A spores and 90.6% for yabG^C spores. These numbers are close to the efficiency of plating of spores onto plates containing taurocholate without lysozyme treatment (see above) and were not statistically significant as assessed by one-way ANOVA and Tukey’s multiple comparison tests, suggesting that lysozyme had no effect on spore survival. Nevertheless, upon exposure to lysozyme, 46% of the ΔyabG spores and 41% of the yabG^C207A spores become phase-dark (6% for WT spores and 15% of the yabG^C spores) (S10 Fig). Thus, the alterations in assembly of the coat/exosporium layers appear sufficient to allow access of lysozyme to the spore cortex.

YabG is required for proper spore morphogenesis

Because spores formed by the ΔyabG and yabG^C207A mutants show alterations in the levels or extractability of several coat/exosporium proteins and are more permeable to lysozyme, we characterized the morphology of the spores produced by the two mutants. Purified spores were first imaged using phase contrast microscopy. Spores of C. difficile often possess an appendage at one of the spore poles, which is continuous with the exosporium [25]. About 12% of spores produced by the WT (15% for the yabG^C strain) possess a well-developed polar appendage, a number close to that (20%) previously reported [25] (S11A and S11B Fig). Strikingly, the ΔyabG or yabG^C207A alleles increased the fraction of spores with a visible appendage to 26% and 24%, respectively (S11A and S11B Fig). Two features, however, distinguished the appendage region in WT or yabG spores: firstly, in the ΔyabG or yabG^C207A mutants, the appendage tended to be more square in shape (S11A Fig, red arrowheads); secondly, while in WT spores, the appendage does not stain with FM4-64, it does so in the two mutants (S11C Fig). Consistent with this result, the fluorescence intensity signal for FM4-64 in the spore polar appendage region, is higher in spores of yabG mutants (S11D Fig). Furthermore the intensity signal for FM4-64 in yabG^C207A spores was higher than in ΔyabG spores(S11D Fig), indicating a possible difference in the composition and/or morphology between yabG^C207A and ΔyabG spores.

Prominent polar appendages, occasionally present at both spore poles, were also seen by scanning electron microscopy (SEM) (Fig 2B, yellow arrowheads). Moreover, material often appeared to detach from this region of the spore, suggesting misassembly of the appendage (Fig 2B, red arrowheads). To characterize the ultrastructure of spores, we used thin sectioning transmission electron microscopy (TEM). The main features previously documented for spores of the WT strain 630Δerm were observed, including the compact core, the cortex, and the lamellar coat surrounded by a thin, more electron dense, exosporium (Fig 2C, blue arrowheads). The appendage itself shows a compact organization and has a well-defined coat/appendage transition zone at its base [25]; (Fig 2C, yellow arrowheads). Strikingly, the structure of the appendage is markedly different in the mutants, consistent with the FM4-64 staining results (above). The transition zone at the basis of the appendage is less defined (Figs 2C and S12, yellow arrowheads) and consistent with the SEM analysis (above), coat/exosporium material seems to peel of the spore poles in both yabG^C207A (Fig 2C, brown arrowheads) and ΔyabG spores (S12 Fig.). Importantly, in about 35% of the yabG^C207A spores with appendages, and in 32% of the ΔyabG spores, the appendage appears as a series of closely juxtaposed lamellae (Fig 2C, red arrowheads in the lower set of panels and S12 Fig). This feature is clearly distinct from the compact appearance of the appendage in WT or yabG^C spores (Figs 2C and S12). Only 2% of the WT spores and 9% of the yabG^C spores showed the coat detached from the cortex, and no spores with a lamellar structure of the appendage could be seen for either strain (Figs 2C and S12). Nevertheless, In WT or yabG^C spores, hints of a lamellar organization are occasionally seen at the edges of the appendage (Fig 2C and S12) [81,82]. Possibly, the appendage has an underlying lamellar structure upon which proteins such as CdeM, are deposited to form the compact, electron dense appendage. In ΔyabG or yabG^C207A spores, the absence of these proteins reveals the underlying structural organization of the appendage, which may be shared by the rest of the exosporium. Consistent with the absence of CdeM from the extracts of yabG spores (above, [24]), the exosporium-like layer is often absent in spores of the ΔyabG or yabG^C207A spores (Figs 2C and S12). Also, in 68% of the spores of a yabG^C207A mutant, and in 85% of the ΔyabG spores, the coat/exosporium did not adhere to the underlying cortex (Fig 2C, yellow circle in the lower set of panels; see also S12 Fig) and disorganized material was seen in the region between the cortex and the coat/exosporium (Fig 2C, black arrowheads in the bottom set of panels; S12 Fig). The impaired adherence of the coat/exosporium to the cortex may explain why in the phase contrast images, the percentage of spores with a recognizable appendage increases in the yabG mutants, as they are easier to recognize (S12 Fig). Finally, in ΔyabG and yabG^C207A spores, the coat/exosporium shows regions of discontinuity along the periphery of the spore (Fig 2C, region within the yellow circle in the lower set of panels, and green arrowheads in the magnified image). These discontinuities likely contribute to the increased accessibility of lysozyme to the cortex layer in yabG spores.

YabG is required for the expression of cotA and cdeM, involved in coat and exosporium assembly

One explanation for the absence of CotA and CdeM from the coat/exosporium extracts of yabG spores, is that somehow YabG could affect production of these proteins. To test this possibility, we first analysed the accumulation of CotA, CdeM and CdeC, for reference, by immunoblotting, in whole cell extracts prepared from sporulating cells harvested after 14 and 20 hours of growth on 70:30 agar plates. CotA and CdeM were detected in the extracts prepared from the WT at 14 and 20 hours of growth but not in the extracts prepared from the yabG mutants; in contrast, CdeC was detected in the WT and the mutants at both time points (Fig 3A and 3B). Thus, the absence of CdeM and CotA from the coat/exosporium extracts of yabG and yabG^C207A spores appears to be a consequence of reduced synthesis or accumulation of the two proteins.

Download:

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

Fig 3. YabG affects the expression of genes coding for coat and exosporium components.

A and B: Coomassie stained gel and immunoblotting analysis of sporulating cells of the WT, ΔyabG, yabG^C and yabG^C207A 14 (A) and 20 hours (B) after inoculation in 70:30 agar plates [53]. Proteins in whole cell extracts were resolved by SDS-PAGE and the gels subjected to immunoblotting with anti-CotA, anti-CdeC, and anti-CdeM antibodies. The red arrowheads point to the various forms of CotA, CdeC and CdeM. C: Quantification of the expression of the indicated genes (cdeC, cdeM and cotA) by qRT-PCR. Total RNA was extracted from C. difficile 630Δerm and ΔyabG strains grown in 70:30 agar plates for 14 and 20 hours. The graph shows the fold-change in the expression of cdeC, cdeM and cotA between the ΔyabG and the WT. Error bars correspond to the standard deviation derived from three biological replicates. Statistical analysis used a Student’s t-test: * p<0.01; **p<0.001.

https://doi.org/10.1371/journal.ppat.1011741.g003

We next measured the transcript levels of cotA, cdeM and cdeC, in sporulating cells of the WT and the ΔyabG mutant using qRT-PCR. These experiments showed decreased levels of cotA and cdeM transcripts in ΔyabG cells compared to the WT both at 14 h (expression ratio ΔyabG/WT of 0.060 for cotA and 0.216 for cdeM) and at 20h of growth (expression ratio ΔyabG/WT of 0.098 and 0.126, respectively) (Fig 3C). In contrast, cdeC expression increased from a ΔyabG/WT ratio of 1.873 at 14 hours to 7.139 at hour 20 (Fig 3C). While not excluding a direct role for YabG in the assembly of CotA and CdeM (but see section below), these results show that yabG is required for the expression of cotA and cdeM. Moreover, since both the ΔyabG and yabG^C207A mutations strongly reduce the levels of the two proteins in spores, we infer that the proteolytic activity of YabG is required for the expression of cotA and cdeM. One possibility is that the activity of YabG is required for the removal of a negative regulator of cotA and cdeM expression; the increased expression rate of cdeC observed at 20 h of growth might be an indirect effect of the absence of YabG activity at late stages of sporulation (see also the Discussion).

Bypass of YabG for expression of cotA and cdeM and assembly of CdeM and CotA

That both the ΔyabG and yabG^C207A alleles reduced the levels of the cotA and cdeM transcripts raised the possibility that the activity of YabG is somehow required to antagonize a transcriptional repressor or to activate a factor required for transcription of both genes. If so, then replacing the cotA and cdeM promoters by a yabG-independent promoter, should bypass the need for yabG for CotA and CdeM production. Both genes were placed under the control of the cotE promotor; cotE also codes for a late coat protein, produced under the control of σ^K; the cotE promoter, however, does not appear to be YabG-dependent, since CotE^FL accumulates in yabG spores (Fig 2A). Strains bearing the P_cotE-cotA and P_cotE-cdeM fusions in the WT, yabG and yabG^C207A backgrounds were grown under sporulation conditions, spores were purified and coat/exosporium extracts prepared and analysed. CdeM is was detected in Coomassie-stained gels [25], whereas the presence/absence of CotA in spores extracts required verification by immunoblotting. Expression of cdeM or cotA from the cotE promotor restored the presence of CdeM (Fig 4A) and CotA (Fig 4B, bottom panel) in extracts from ΔyabG or yabG^C207A spores.

Download:

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

Fig 4. YabG-independent expression of cotA and cdeM restores assembly of CotA and CdeM to ΔyabG and yabG^C207A spores.

Coomassie stained SDS-PAGE gel of the proteins extracted from purified spores of the WT, ΔyabG and yabG^C207A mutants and derivatives expressing P_cotE- cdeM (A) or P_cotE- cotA (B). In A, the bottom panel shows the immunoblot analysis of the corresponding gel using an anti-CotA antibody. The position of the main forms of CdeM (in A) or CotA (in B) is shown by red arrowheads. C: spores produced by the yabG^C207A mutant and by derivatives expressing P_cotE-cdeM or P_cotE-cotA were imaged by TEM. The polar region of yabG^C207A/P_cotE-cotA and yabG^C207A/P_cotE-cdeM spores is magnified in the panels to the right. Cr, core; Cx, cortex; Ap, spore appendage region. Green arrowheads, coat and exosporium material peeling off the spore; blue arrowheads, regions with an exposed cortex. Scale bars: 100 nm for the magnified images, 500 nm for all other panels.

https://doi.org/10.1371/journal.ppat.1011741.g004

To determine if and to what extent expression of cdeM or cotA corrected the phenotype of yabG^C207A spores we used TEM. In spores of the yabG^C207A mutant expressing of P_cotE-cdeM the lamellar appearance of the polar appendage region was lost (Fig 4C, middle panel); instead, the appendage region appeared compact and electrondense (Fig 4C, Ap), consistent with the accumulation of CdeM and its role in formation of the polar appendage [25]. Other features of yabG spores, however, such as the peeling off of significant sections of the coat/exosporium were maintained (Fig 4C, blue arrowheads). Expression of P_cotE-cotA in the yabG^C207A background resulted in spores with visible juxtaposed sheets in the polar appendage region (Fig 4C, panels in the right), similar to yabG spores and most likely due to the absence of CdeM (see also above).

These results show that YabG acts at the level of the cotA and cdeM promoters to influence transcription of these genes. In addition, since the expression of cdeM and cotA from P_cotE result in the detection of CdeM and CotA in coat/exosporium extracts prepared from yabG spores, we infer that YabG is not a strict requirement for the localization of CotA or CdeM.

Auto-regulatory assembly of YabG

We then wanted to monitor the sub-cellular localization of YabG. To this end, a translational fusion of the WT protein to the SNAP^Cd tag, YabG^WT-SNAP^Cd, was constructed and introduced into the WT strain. As detailed below, the YabG^WT-SNAP^Cd fusion is largely functional. Cells were grown in 70:30 agar plates and imaged by phase contrast and fluorescence microscopy 14 and 20 hours after inoculation. At hour 20 in the WT background, i.e., in the presence of the yabG^WT allele, YabG^WT-SNAP^Cd was detected around 10% of the phase-dark forespores, in 42% of the phase-grey forespores and in 81% of the phase-bright forespores (Fig 5A, yellow arrowheads). A similar localization pattern was observed for YabG^WT-SNAP^Cd at hour 14 (S13A Fig, yellow arrowheads on the left set of panels) and at the two time points in a ΔyabG mutant (Figs 5A and S13A). Thus, consistent with an association of the protein with the coat and/or exosporium layers YabG^WT-SNAP^Cd localizes to the forespore after engulfment completion and remains associated with the developing spore at late stages in morphogenesis. We note that for both YabG^WT-SNAP^Cd and YabG^C207A-SNAP^Cd, in either the WT or ΔyabG background, a haze of fluorescence is detected in the mother cell cytoplasm (Figs 5A and S13A, white arrowheads). This signal may result from release of the SNAP^Cd moiety through proteolysis or otherwise indicate that some of the fusion proteins remain in the mother cell (see below).

Download:

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

Fig 5. Localization of YabG-SNAP^Cd in sporulating cells.

A: Localization of YabG^WT-SNAP^Cd and YabG^C207A-SNAP^Cd in C. difficile 630Δerm (WT) and ΔyabG strains. Cells were collected after 20h of growth in 70:30 agar plates [53], stained with the SNAP substrate TMR-Star and examined by phase contrast and fluorescence microscopy (red channel for TMR signal and green channel for autofluorescence signal). The numbers refer to the percentage of cells at the represented stage showing SNAP fluorescence. Yellow and white arrowheads point to the position of the forespore and the mother cell respectively. At least 150 cells were analysed for each strain, in three independent experiments. Scale bar, 1 μm. B: Accumulation of YabG-SNAP^Cd and YabG^C207A-SNAP^Cd in sporulating cells of strains 630Δerm (WT) and ΔyabG at 20h of growth in 70:30. Proteins in whole cell extracts were resolved by SDS-PAGE and the gel subject to immunoblot analysis with anti-SNAP antibodies. Samples collected from the WT and the ΔyabG mutant bearing no SNAP^Cd fusion were used to control for antibody specificity. The Coomassie-stained gel is included as a loading control. Red arrowheads point to the position of YabG^C207A-SNAP (52 kDa). Asterisks denote possible degradation products that include the SNAP moiety (~19.4 kDa). The black arrowhead shows the position of a cross-reactive species.

https://doi.org/10.1371/journal.ppat.1011741.g005

To determine whether the activity of YabG was involved in the association of the protein with the developing spore, we monitored the localization of the catalytically-inactive YabG^C207A-SNAP^Cd fusion. At 20 hours of growth, YabG^C207A-SNAP^Cd localized as 2 caps at the mother cell proximal and distal poles in 47% of phase-dark spores (Fig 5A). YabG^C207A-SNAP^Cd formed a ring of fluorescence around 70% of phase-grey forespores and around 94% of phase bright-forespores (Fig 5A, yellow arrowheads). A similar pattern of localization was observed at hour 14, except that YabG^C207A-SNAP^Cd was detected even earlier, as a single cap of fluorescence in 6% of the sporangia during engulfment (S13A Fig., yellow arrowheads). The localization of YabG^C207A-SNAP^Cd in cells during engulfment and the higher percentages of localization of the fusion protein in sporangia of phase-dark, phase-grey and phase-bright spores suggests increased stability of the catalytically inactive protein even in the presence of the WT yabG allele. This observation suggests that the auto-proteolytic activity of YabG, detected for both the B. subtilis [52] and the C. difficile proteins (Fig 1E) controls the accumulation and localization of the protein. If so, the localization of the fusion proteins could increase in cells of a ΔyabG mutant. In comparison to the WT background, however, the localization of YabG^WT-SNAP^Cd in ΔyabG sporangia only increased slightly around phase-dark forespores (25% as opposed to 10% in the WT at hour 20; Fig 5A) and for phase-grey forespores (65% as opposed to 55% at hour 14; S13A Fig). For the localization of YabG^C207A-SNAP^Cd in the ΔyabG background, the main difference relative to the WT background was the increase in the single cap pattern in sporangia during engulfment (from 0 to 11% at hour 20 and from 6 to 12% at hour 14; Figs 5A and S13A). The immunoblot analysis of whole cell extracts is in good agreement with the microscopy results. YabG^C207A-SNAP^Cd is detected with an anti-SNAP monoclonal antibody at higher levels than YabG^WT-SNAP^Cd (both proteins run as 52 kDa species) in both the WT and the ΔyabG background with the highest accumulation corresponding to the catalytically inactive fusion in the ΔyabG mutant (red arrowheads in Figs 5B and S13B.). Bands just above and below the 20 kDa marker are likely to result from cleavage of the fusion protein close to the C-terminus of YabG (asterisks in Figs 5B and S13). The increased accumulation of YabG^WT-SNAP^Cd or YabG^C207A-SNAP^Cd in yabG mutants does not seem to result from augmented transcription of yabG: control experiments show that the transcription of a P_yabG-SNAP^Cd fusion increases only slightly in ΔyabG sporangia (S14 Fig).

Thus, YabG^WT-SNAP^Cd appears to degrade itself; since no major difference was detected for YabG^C207A-SNAP^Cd in ΔyabG sporangia in comparison to the WT, it may be that the degradation of YabG requires at least one in cis cleavage event. In any event, since YabG^C207A-SNAP^Cd localizes earlier than the WT, the catalytic activity of YabG controls, at least in part, the localization of the protein during sporulation. In that sense, assembly of YabG is auto-regulatory.

YabG localizes asymmetrically in mature spores

To gain insight onto the localization of YabG in mature spores, we used Super Resolution Structured Illumination microscopy (SR-SIM), in which the lateral resolution is increased to about 110 nm, as compared with the diffraction limit of 250 nm of conventional light microscopy [83]. Spores of strains producing either YabG^WT- or YabG^C207A-SNAP^Cd in the WT or ΔyabG backgrounds were labelled with MTG, which decorates the spore body and with TMR-Star prior to SR-SIM imaging. Both YabG^WT- and YabG^C207A-SNAP^Cd localized around the entire contour of the spore (Fig 6A, white arrowheads; see also S15 Fig), in confirmation of the conventional fluorescence microscopy data where the signal for YabG-SNAP was detected around the forespore in sporangia of phase-bright forespores (above). Strikingly, in the WT, YabG^WT-SNAP^Cd localized to and showed a strong signal overlapping the spore polar appendage in 63% of the spores showing this structure, and YabG^C207A-SNAP^Cd localized in 85% of those spores (Fig 6A, blue arrowheads; see also S15 Fig). In the ΔyabG mutant, YabG^WT-SNAP^Cd localized in 63% of the appendage-bearing spores, and YabG^C207A-SNAP^Cd localized in 46% of those spores (Fig 6A, blue arrowheads; S15 Fig). The impaired localization of the YabG^C207A fusion in the ΔyabG background suggests that the WT protein facilitates the localization of the mutant form to free spores and/or its maintenance. The localization of the fusion proteins to the spore polar appendage is consistent with the role of YabG in the morphogenesis of this structure (see above).

Download:

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

Fig 6. Localization of YabG^WT- or YabG^C207A-SNAP^Cd in mature spores.

A: Localization of YabG^WT- or YabG^C207A-SNAP^Cd in mature spores using SR-SIM, in either the WT or ΔyabG backgrounds. The spores were stained with the membrane dye MTG (green) and with TMR-Star (red) prior to imaging. The blue arrows point to the SNAP signal at the spore poles and the white arrows to the signal along the side of the spore (see also S15 Fig). The distribution of the fluorescence signal (in arbitrary units, AU) in three dimensional intensity graphs is shown below the microscopy images. Scale bar, 500 nm. B: Coomassie stained SDS-PAGE gel of the proteins extracted from the cortex/coat/exosporium and core/cortex fractions of purified spores of the WT strain, the ΔyabG mutant, and of strains producing YabG^WT-SNAP^Cd or YabG^C207A-SNAP^Cd in either the WT or ΔyabG backgrounds. The gel was subjected to immunoblot analysis with anti-SNAP, anti-YabG, anti-SleC, anti-CdeC, anti-CotA and anti-CdeM antibodies. The arrowheads points to the position of the relevant proteins. YabG-SNAP denotes the position of either full-length WT or the C207A variant fused to the SNAP tag. Asterisks denote possible degradation products or cross-reactive species.

https://doi.org/10.1371/journal.ppat.1011741.g006

Proteins in a cortex/coat/exosporium and a core/cortex fraction were resolved by SDS-PAGE and analysed by immunoblotting. Full-length YabG^WT-SNAP^Cd, with an expected size of about 52 kDa, was barely detected in the coat/exosporium fraction of WT or ΔyabG spores with anti-SNAP or anti-YabG antibodies and was not detected in the core/cortex fraction of either strain (Fig 6B, red arrowhead). In contrast, YabG^C207A-SNAP^Cd was detected in the cortex/coat/exosporium fraction of both WT and ΔyabG spores, and in trace amounts in the core/cortex fraction of ΔyabG spores (Fig 6B). Since full-length YabG^WT-SNAP^Cd does not accumulate even in the ΔyabG mutant, it seems that the protein undergoes auto-proteolysis and thus, that the fusion protein is largely functional with respect to this activity (see also below). Auto-proteolysis may occur, at least in part, in cis, since full-length YabG^C207A-SNAP^Cd was detected in the WT but at significantly higher levels in the ΔyabG mutant (Fig 6B, middle panel, red arrowhead). Because the fluorescence signal from YabG^WT- or YabG^C207A-SNAP^Cd in the SR-SIM images is comparable (Fig 6A), we infer that YabG^WT-SNAP^Cd, although accessible to the TMR-Star substrate, is less extractable than YabG^C207A-SNAP^Cd.

Bands below the 20 kDa marker, detected in the WT for YabG^WT-SNAP^Cd and for YabG^C207A-SNAP^Cd, may result from cleavage of the fusion to release the SNAP moiety or fragments of it, since the C-terminally located reporter has a predicted molecular mass of 19.4 kDa (Fig 6B). As above (Fig 2A), SleC was only detected in the cortex/coat/exosporium fraction [74], whereas GPR was only detected in the core/cortex fraction (Fig 6B). Importantly, the analysis of the cortex/coat/exosporium extracts by Coomassie staining and immunoblotting showed that the YabG^WT-SNAP^Cd fusion, but not YabG^C207A-SNAP^Cd, restored both CotA and CdeM assembly and significant processing of Pre-pro-SleC to spores of a ΔyabG mutant and no other major differences to WT spores were noticed (Fig 6B). We infer that with respect to the assembly of the spore surface, YabG^WT-SNAP^Cd is largely functional.

YabG is required for colonization in a hamster model

We then wanted to test whether YabG played a role in colonization or infection in an hamster model. To induce susceptibility to C. difficile infection, female Syrian golden hamsters, individually housed in cages were first gavaged with clindamycin [84,85] and after 5 days, gavaged with 10³ WT or ΔyabG spores. Signs of disease (lethargy, poor fur coat and wet tail) were monitored along time. In parallel, fecal samples were collected daily, plated onto TCCFA plates to enumerate both C. difficile vegetative cells and spores (see also the Material and Methods section). The percentage of hamsters that survived 96 hours post infection was similar for the WT and ΔyabG spores, indicating that the two strains are equally virulent (Fig 7A).

Download:

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

Fig 7. YabG is required for colonization of the hamster colon.

A: Kaplan-Meier curve for hamster challenged with spores of the WT (red) and ΔyabG mutant (blue). Syrian hamsters were first gavaged with clindamycin to induce susceptibility to infection and challenged with 10³ WT or ΔyabG spores five days after. Hamsters showing signs of disease were euthanized. B: Enumeration of C. difficile cells and spores in faecal material following gavage of hamsters with spores produced by the WT (pink panel) and ΔyabG mutant (light blue panel) (as in B). Faecal samples were collected daily, and both C. difficile vegetative cells and spores were enumerated by plating. The data was set to the median values and the limit of detection (dashed line) is shown.

https://doi.org/10.1371/journal.ppat.1011741.g007

The total CFU/g counts from day 1 to day 6 post-infection, however, were considerably reduced for ΔyabG as compared to the WT (Fig 7B). While for the WT, the number of C. difficile cells and spores in the fecal material was between 10⁶ and 10⁸ CFU/g at day 1, this number was below 10⁴ CFU/g for the mutant (Fig 7B). From day 1 to day 2 post infection, the number of CFU/g increased for the WT to between 10⁷ to 10⁹ CFU/g and it remained between 10⁷ and 10⁸ CFU/g up to day 6; in contrast, the number of CFU/g remained low, around 10⁵ CFU/g for the ΔyabG mutant (Fig 7B). We note that the number of animals infected with ΔyabG spores is lower that for the WT because some of those animals were sick and did not produce a fecal sample; for this reason they were not included in the analysis. In any case, yabG seems to have an important role during colonization of the hamster colon.

Domain organization of YabG

An AlphaFold2 model indicates that YabG has an N-terminal domain separated from the catalytic domain by a linker (Fig 1C). The larger, C-terminal domain B most likely has the fold of receiver domains of response regulators such as CheY [67]. The conserved Asp that is the site of phosphorylation in CheY and other response regulators, however, is not conserved. Thus, YabG appears to function independently of phosphorylation [68,69]. A significant number of receiver domains, so called single domain response regulators, are not attached to an independently folded output domain and some, such as CheY function by interacting with other proteins [68,69,86]. We do not know whether YabG functions via interactions with proteins other than its substrates, but since the yabG^C207A catalytic inactive mutant phenocopies a yabG deletion mutant, this seems unlikely.

As found for the B. subtilis protein [52], C. difficile YabG also shows an auto-proteolytic activity (Fig 1E). Auto-proteolysis of B. subtilis YabG uses a C218/H172 catalytic dyad [52]. These residues are homologous to C207 and H161 of the C. difficile protein and may be directly involved in catalysis. Asp162, is also required for the activity of the C. difficile protein and is conserved among YabG ortologues but the role of the homologous residue in the B. subtilis protein, Asp173, has not been tested. In any event, C207, H161 and D162 are all located in the CheY-like domain B (Fig 1C and 1D). A CheY-like domain showing proteolysis activity is, to our knowledge, unique.

The position of the SH3-like domain A, on top of the putative catalytic center (Fig 1C), suggests that it has to be removed for activation [58]. An alternative view is that the SH3-like domain A is involved in interactions with substrate proteins and is thus required in activity. In favour of this last view, since one of the point mutations found in yabG that result in poor processing of Pre-pro-SleC and CspBA and in spores that no longer require co-germinants is located in the A domain [43].

Substrates of YabG

In C. difficile, as also seen in B. subtilis, the morphogenetic ATPase SpoIVA is present at higher levels in coat/exosporium extracts prepared from ΔyabG spores [38,42]. Thus, SpoIVA may a YabG substrate. None of the other known B. subtilis YabG substrates have orthologues in C. difficile and conversely, neither CspBA, Pre-pro-SleC, or CotE is found in B. subtilis or closely related organisms [18,34,44]. It is likely that at least in C. difficile, YabG has additional substrates. For example, in a yabG insertional mutant SleC is activated in response to TA alone, suggesting the involvement of YabG in the processing of an as yet unknown protein presumably controlling germinant specificity [43]. Since YabG is conserved across spore formers, it follows that in spite of a high degree of sequence conservation, YabG must bear structural determinants that allow utilization of specific substrates in different organisms.

Auto-regulatory assembly of YabG

The catalytically inactive YabG^C207A localizes to the forespore surface slightly earlier than the WT protein. Thus, the auto-processing activity translates into delayed assembly of YabG. Localization of YabG^WT is first detected when the forespore becomes phase-dark and it increases when the forespore becomes phase-bright, a stage at which expression of yabG also increases (S6 and S14 Figs). It thus seems possible that YabG is only assembled when a threshold level of the protein is reached. This threshold may also depend on increased synthesis of its substrates, which may compete for and reduce auto-proteolysis of YabG (Fig 8). Although YabG localizes to the cortex/coat/exosporium its exact location within these structures is not known. SpoIVA is most likely close to the forespore outer membrane and CspB, CspA and Pre-pro-SleC are associated with the cortex possibly in a complex [48,78,87,88]. Possibly, YabG is also located in the innermost layers of the coat.

Download:

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

Fig 8. Model for the role of YabG in the assembly of the spore surface layers.

A: yabG expression is first detected in the mother cell compartment after asymmetric division under the control of σ^E but persists in this cell under σ^K control [30]. σ^K also directs transcription of cdeC (which additionally requires SpoIIID); transcription of cdeC and the σ^K-dependent period of yabG transcription begins coincidently with the appearance of phase-dark forespores and CdeC is probably recruited to the forespore surface at this stage. The assembly of YabG, however, is auto-regulatory and self-limiting in that YabG undergoes self-degradation in either the mother cell cytoplasm or at the spore surface. The stability of YabG increases when the spore turns phase-bright, and may require production of the YabG substrates and/or increased production of the protease. At this stage, YabG accumulates at the spore surface and is also required in the mother cell cytoplasm for the degradation of an as yet unidentified repressor (R) or for the activation of a activator (A) of transcription of a class of σ^K-dependent genes that includes cdeM and cotA. The encoded proteins are produced and assembled when repression is relieved or activation is triggered. Full-length YabG is represented, as it unknown whether domain A is removed. B: YabG is involved in processing of CspBA and SleC^FL to produce CspB and Pro-SleC. and possibly also of CotE, to separate the N-terminal chitinase (CotE^N) and the C-terminal peroxiredoxin (CotE^C) domains. YabG is also required for the degradation of SpoIVA, and possibly of CdeC. It is not known where are these proteins processed. Both the role of YabG in enforcing the proper structure of the spore and in germination are thought to contribute to host colonization. Finally, YabG is required for the transcription of at least cdeM and cotA (red arrow).

https://doi.org/10.1371/journal.ppat.1011741.g008

Role of YabG in coat/exosporium assembly: Morphological features of yabG spores

YabG has two main effects on the assembly of the spore surface layers. Firstly, it is required for attachment of the coat to the underlying cortex layer (Figs 2C and S12). The basis for this phenotype is unclear. Since SpoIVA localizes close to or at the forespore outer membrane and is required for proper assembly of the cortex and coat [53,54], proper levels and/or processing of this morphogenetic ATPase may be important for coat/cortex attachment. Secondly, YabG is required for assembly of the exosporium. In the absence of CdeM, the exosporium is thin, lacks electrondensity and the appendage is short and disorganized [24,25]. Therefore, the structural defects seen at the level of the exosporium in yabG mutants may be explained, at least in part, by the absence of CdeM. Strikingly, in yabG spores, the appendage region lacks the compact, electrondense structure seen in WT spores, and instead shows a lamellar organization (Figs 2C and S12). We suspect that the spore polar appendage has an underlying lamellar structure, possibly formed by exosporium proteins such as CdeC, which is revealed in the absence of CdeM. While the localization of YabG around the entire spore is consistent with a role in assembly of the exosporium, its enrichment at the spore poles suggests a specific role in assembly of the polar appendage.

YabG control of the mother cell line of gene expression

We found that the expression of cdeM and cotA is severely curtailed in the ΔyabG mutant (Fig 3). Expression of yabG begins soon after asymmetric division, under the control of σ^E and is maintained during later stages with the (weaker) contribution of σ^K (S6 Fig). In contrast, while cdeC expression is first detected after engulfment completion under the direct control of σ^K, at the onset of the main period of σ^K activity, expression of at least cdeM occurs later, when the forespore turns phase-bright (S14 Fig) [25]. The mechanism by which expression of cdeM and possibly other genes is delayed is presently unknown but it could involve YabG. A simple explanation for the effect of YabG on cotA and cdeM transcription is that the protease is involved in the removal of a factor that represses transcription of some σ^K-dependent genes. SpoIIID is an ancillary regulatory protein, produced under the control of σ^E, required for the production and activation of σ^K and importantly, also required together with σ^K to enhance expression of cdeC [35,89]. One possibility is that SpoIIID represses and thereby delays the expression of a late class of σ^K-controlled genes, which includes cdeM and cotA, and that YabG is required for its removal. The timing of this event would be determined by the rise in yabG expression when the forespore turns phase-bright (Fig 8A; see above), which is also when expression of at least cdeM is detected. In any event, YabG acts at the level of the cotA and cdeM promoters because expression of these genes from the cotE promoter bypasses the need for yabG for both their transcription and the association of CotA and CdeM with the spore surface (Fig 4). We cannot discard the possibility, however, that YabG is involved in the proteolytic activation of a putative activator (Fig 8A). The effect of YabG on cdeC expression could be indirect, in that the absence of YabG transcription of some late σ^K dependent genes (cdeM and cotA) is blocked while the expression of earlier σ^K targets (such as cdeC) is increased [25].

The control of the mother cell line of gene expression by YabG has not been reported for B. subtilis. In this organism, the mother cell line of gene expression is divided into several temporal and epistatic classes through the action of ancillary transcription factors that work together with σ^E (SpoIIID, GerR) and σ^K (GerE) [32,33,90–92]. Only SpoIIID is found in C. difficile. To what extent YabG influences gene expression in the mother cell and whether this role of YabG is unique to C. difficile remains to be studied.

Role of YabG in host colonization

At least in strain 630Δerm, spores of the yabG mutants show slow and less efficient germination as assessed by the drop in OD₆₀₀ of a spore suspension (S8 Fig). The plating efficiency of yabG spores in TA-BHI plates, however, is very close to that of WT spores, in agreement with earlier work [42]. It is therefore unclear whether inefficient germination contributes to the impaired colonization by the yabG mutant in the hamster assay. However, because of their increased permeability (S10 Fig), host-produced lysozyme may contribute to the germination of yabG spores in vivo. Full-length CotE accumulates in yabG spores (Fig 2A) but it is not known whether interdomain processing is a requirement for colonization and virulence [23]. The altered surface of cdeM spores likely explains the impaired colonization of mice [26] and the structural alterations of yabG spores may likewise affect the colonization ability of the mutant (Fig 8B)

The role of the YabG protease in coat/exosporium assembly, spore germination and host colonization raises the possibility that inhibitors of the enzyme may serve as chemotherapeutic agents to prevent proper morphogenesis of the C. difficile spore in vivo and transmission of the organism.

Ethics statement

All animal procedures were performed with prior approval from the Texas A&M Institutional Animal Care and Use Committee under the approved Animal Use Protocol number 2017–0102. Animals showing signs of disease were euthanized by CO₂ asphyxia followed by thoracotomy as a secondary means of death, in accordance with Panel on Euthanasia of the American Veterinary Medical Association. Texas A&M University’s approval of Animal Use Protocols is based upon the United States Government’s Principles for the Utilization and Care of Vertebrate Animals Used in Testing, Research and Training and complies with all applicable portions of the Animal Welfare Act, the Public Health Service Policy for the Humane Care and Use of Laboratory Animals, and all other federal, state, and local laws which impact the care and use of animals.

Strains and growth conditions

Bacterial strains and their relevant properties are listed in S1 Table. The Escherichia coli strain DH5α (Invitrogen) was used for molecular cloning and BL21(DE3) (Novagen) was used for the over-production of WT His_10-YabG and its variants and CspBA-Strep-tag by auto-induction [73]; HB101 (RP4) was used as the donor in C. difficile conjugation experiments [93]. Luria-Bertani medium was routinely used for growth and maintenance of E. coli with. ampicillin (100 μg/ml) or chloramphenicol (15 μg/ml) added if required. The C. difficile strains used in this study are congenic derivatives of the wild-type strain 630Δerm [94]. C. difficile strains were grown anaerobically (5% H₂, 15% CO₂, 80% N₂) at 37°C in brain heart infusion (BHI) medium (Difco) and in 70:30 plates for sporulation induction. When necessary, taurocholate (TA) (0.1% wt/vol), thiamphenicol (15 μg/ml) and/or cefoxitin (25 μg/ml) were added to the medium. For animal experiments, spores and vegetative cells were enumerated on taurocholate-cycloserine-cefoxitin-fructose agar (TCCFA) medium (Difco). All plasmids and oligonucleotide primers used are listed in S2 and S3 Tables, respectively. The inserts in all of the plasmids herein constructed were verified by DNA sequencing.

In-frame deletion of yabG using CRISPR-Cas9

A CRISPR-Cas9 system [95,96] was used to create an in-frame deletion of yabG (CD630_35690) in strain 630Δerm ΔpyrE, (S7A and S7B Fig). The forward primer to amplify the sgRNA, P5, used to direct the Cas9 nuclease to the yabG gene, was designed by adding the 20 nucleotides crRNA “SEED” region, identified with the Benchling CRISPR guide design tool, for the sgRNA [95,97]. P5 was used with P6 to amplify the sgRNA (159 bp). The homology arms (HA) coding for the yabG mutant allele were joined by PCR. The left homology arm (LHA) fragment (645 bp) was generated with P7 and P8 while the right homology arm (RHA) fragment (729 bp) was generated with P9 and P10 using as template chromosomal DNA of C. difficile 630Δerm; the two fragments were joined by overlapping PCR. The resulting fragment (1390 bp) was cleaved with AscI and AsisI while the sgRNA fragment was cleaved using SalI and AsiSI. Both fragments were cloned between the SalI and AscI sites of pMTL431521 [95] to yield pEM28. pEM28 was introduced into 630Δerm ΔpyrE by conjugation [98]. Transconjugants were selected as described [98]. Chromosomal DNA of the transconjugants was isolated as previously described [99]. PCR using primers P1 and P2 confirmed deletion of yabG in 630Δerm ΔpyrE (S7A and S7B Fig). The resulting strain is AHCD1150 (630Δerm ΔyabG ΔpyrE).

pyrE reversion and in trans-complementation

To restore the pyrE gene, pMTL-YN1 [94] was conjugated into strain 630Δerm ΔpyrE ΔyabG (AHCD1150). Transconjugants were plated onto minimal medium and analysed by PCR, using primers P3 and P4 (S7C and S7D Fig), to verify pyrE reversion. Finally, plasmid loss was tested by patch plating positive clones onto BHI supplemented with cefoxitin and thiamphenicol; this screen yielded strain AHCD1203 (630Δerm ΔyabG pyrE⁺). For complementation of the ΔyabG ΔpyrE strain in single copy at pyrE, yabG with its promoter region (a 1585 bp fragment) was PCR-amplified with P11 and P12 and the fragment inserted between the XhoI and BamHI sites of pMTLYN1C [94] to produce pEM39. In addition, yabG^C207A was PCR-amplified with P11 and P13 and chromosomal DNA from 630Δerm to obtain a fragment of 1192 bp. A second fragment of 419 bp was amplified using P14 and P12. The PCR fragments were fused by overlapping PCR using P11 and P12 and the resulting fragment cloned between the XhoI and BamHI sites of pMTLYN1C [94] to produce pEM41. pEM39 and pEM41 were conjugated into C. difficile AHCD1150 (ΔyabG) to introduce the yabG or yabG^C207A alleles at pyrE. Both strains were analysed by PCR to verify the insertion of the allele of interest at the pyrE locus using primers P3 and P4 (S7C and S7D Fig). Finally, colonies of yabG^C and ΔyabG^C207A were patch plated onto BHI with cefoxitin and thiamphenicol to test for loss of the plasmids. Plasmid-cured strains were named AHCD1204 (referred to as yabG^C) and AHCD1205 (yabG^C207A).

A yabG-SNAP^Cd transcriptional fusion

The promoter region of yabG was PCR-amplified from genomic DNA of 630Δerm using P15 and P16, to produce a 268 bp fragment. The SNAP^Cd gene [29,30] was PCR amplified from pFT47 [29] using P17 and P18. The two pieces were joined by PCR and the resulting 803 bp fragment cloned between the EcoRI and XhoI sites of pMTL84121 [100] to produce pEM7 which was conjugated into the 630Δerm, sigE::ermB (AHCD533) and sigK::ermB (AHCD535) strains [29] (S1 Table).

Placing cdeM and cotA under the control of the cotE promoter

P_cotE-cotA was constructed by PCR-amplifying the cotE promoter from the genomic DNA of C. difficile 630 using primers P49/P50 and P51/P52. The two fragments were joined by PCR to produce a product of 1246 bp that was inserted between the EcoRI and HindIII sites of pMTL84121 to yield pSR77 which was conjugated into C. difficile 630Δerm, ΔyabG and ΔyabG pyrE::yabG^C207A strains yielding AHCD1502, AHCD1503, and AHCD1504, respectively. P_cotE-cdeM was constructed by PCR-amplifying the cotE promoter from the genomic DNA of C. difficile 630 using primers P53/P54 and P54/P55. The two pieces were joined by PCR and inserted between the EcoRI and HindIII sites of pMTL84121 to give pCAF3. Plasmid pCAF3 was conjugated into 630Δerm (WT) producing strain AHCD 817.

Translational YabG-SNAP^Cd fusions

To construct a C-terminal SNAP^Cd-tag fusion to yabG^WT, the sequence containing the yabG promoter and coding regions without the stop codon was PCR-amplified using P15 and P30 and chromosomal DNA from 630Δerm and cloned into pFT58 [29] yielding pEM5 (P_yabGyabG^WT-snap^Cd). To construct a C-terminal SNAP^Cd-tag fusion to yabG^C207A, the yabG promoter region was PCR-amplified using P15 and P13. The yabG^C207A coding sequence was amplified using primers P30 and P14 The two fragments were joined by PCR and the resulting fragment was cleaved using EcoRI and BamHI and cloned into pFT58. This produced plasmid pEM40 (P_yabG-yabG^C207A- snap^Cd).

cspBA-strep-tag into pACYCDuet-1

The cspBA coding sequence was PCR-amplified using primers P31 and P32. The resulting fragment was inserted between the NcoI and NotI sites of pACYC-duet (Novagen) to create the cspBA-strep-tag-expression plasmid pEM23.

his₁₀-yabG in pET16b

The yabG coding sequence was PCR-amplified using primers P33 and P34. The resulting fragment was inserted between the BamHI and XhoI sites of pET16b (Novagen) to create pEM6. Mutations generating single Ala substitutions were introduced in yabG using primer pairs P14/P13, P37/P38, P39/P40, P41/P42, P43/P44 and pEM6 as the template; this created plasmids pEM12, pEM13, pEM21, pEM38, and pEM24, respectively.

Spore production and purification

Cultures (150 ml) in BHI medium were incubated at 37°C under anaerobic conditions for 7 days. Cells were collected by centrifugation (at 4800xg, for 10 min, 4°C) resuspended in cold water and stored for 48 hours at 4°C. The suspensions were then collected by centrifugation and the sediment resuspended in 1 ml of PBS-tween 20. Spores were purified on density gradients of Gastrografin (Bayer) [101]. The spore titer in the suspension was measured spectrophotometrically at an OD_580.

Spore fractionation and mass spectrometry

Spores were resuspended in 50 μl of decoating buffer (10% glycerol,4%SDS,10% β-mercaptoethanol, 1mM DTT, 250 Mm Tris Ph 6,8) to a final OD₅₈₀ of 4.0. The suspension was boiled for 5 minutes and the spores collected by centrifugation. The supernatant, corresponds to a cortex/coat/exosporium fraction. The spore sediment was washed twice with PBS with 0.1% Tween-20, and incubated with 50mM Tris-HCl pH 8 with 2mg/ml lysozyme for 2 h at 37°C to digest the spore cortex peptidoglycan and release core/cortex-associated proteins.

Whole cell extracts and immunobloting

Whole cell extracts were prepared from E. coli or sporulating cells of C. difficile as described before [29]. The whole cell extracts and the cortex/coat/exosporium and core/cortex fractions of spores were analysed by 15% SDS-PAGE and immunoblotting. Antibodies were used at the following dilutions: anti-YabG (1:1000), anti-CdeC (1:500), anti-CdeM (1:15000), anti-CotA (1:1000), anti-CspB, anti-CspC, and anti-SleC (1:3000), anti-GPR (1:10000), anti-SNAP-tag at 1:1000 and anti-His tag (at 1:1000). A rabbit secondary antibody conjugated to horseradish peroxidase (Sigma) was used at a dilution of 1:5000; an anti-mouse IgG (whole molecule)-peroxidase (Sigma) was used at a dilution of 1:2000 for the detection of the SNAP-tag. All the immunoblots were developed with the Super Signal Pico Plus Chemiluminescent Substrate (Thermo Scientific). For protein identification, protein bands were excised from Coomassie-stained gels, digested with trypsin and analysed by matrix-assisted laser desorption ionization (MALDI) mass spectrometry.

RNA extraction and quantitative RT-PCR analysis

Sporulating cells were collected from cultures of the 630Δerm strain and the ΔyabG mutant after 14 h and 20 h of growth on 70:30 agar plates. RNA was extracted from three independent cultures, using the RNeasy mini kit (Qiagen), according to the manufacturer’s instructions. cDNAs synthesis and real-time quantitative PCR were as previously described [102,103]. In each sample, the quantity of cDNAs for each gene was normalized to the quantity of cDNAs of the rpoC gene. The primers used for each marker are listed in S3 Table. The relative transcript changes were calculated using the 2^-ΔΔCt method as described.

Spore lysozyme resistance and lysozyme permeability assays

Lysozyme resistance was assessed based on the method described in [79]. Spores (1x10⁷) of the WT strain 630Δerm, yabG mutants or the complementation strain were resuspended in a volume of 0.5 ml PBS buffer and incubated for 30 min at 37°C with lysozyme (250 μg/ml) followed by heat treatment for 10 min at 80°C, serial dilution and plating on BHI agar plates containing 0.1% TA to determine surviving colony forming units (CFU)/ml. The number of phase-dark spores resulting from exposure to lysozyme (under the conditions above, but not heat treated) was also scored using phase contrast microscopy.

Spore Germination and DPA content

Density-gradient-purified spores were resuspended in BHI to a final OD₆₀₀ of 1 and heat activated for 10 min at 80°C. Taurocholic acid (TA) (Sigma-Aldrich) in BHI was added to a final concentration of 0.5%. Germination was followed under anaerobic conditions (5% H₂, 15% CO₂, 80% N₂) at 37°C, by following the decrease in the OD₆₀₀ of the spore, until no significant changes were detected (data is shown for the first 90 min). The DPA released was measured by incubating spores at 37°C in PBS supplemented with 0.5% TA for 1 hour as described [51]; the same amount of spores was incubated at 37°C in PBS without TA addition (negative control) or at 100°C with 0.5% TA (positive control). The spore suspension was centrifuged (at 6000 xg for 3 min at room temperature) and the released DPA estimated by measuring the OD₂₇₀ of the supernatant. The efficiency of spore germination was also estimated using a plate assay as described [42].

Microscopy

For SNAP labelling, samples were withdrawn from sporulating cultures at 14h and 20h growth into 70:30 media [53], incubated with the TMR-Star substrate (New England Biolabs), processed for phase contrast and fluorescence microscopy and imaged as described [29]. The intensity of the FM4-64 signal in the spore polar appendage was quantified using ImageJ (http://rsbweb.nih.gov/ij/) by drawing a 0.25 μm diameter circle in the appendage region. The data from three independent experiments was represented using SuperPlots [104]. For the characterization of spores of strains producing either YabG^WT- or YabG^C207A-SNAP^Cd in the WT or ΔyabG background, with TMR-Star as described above and with the membrane dye Mitotracker Green (MTG, 0.5 ng/mL; Invitrogen). Super-resolution Structured Illumination Microscopy (SR-SIM) as conducted as described previously [99].

Scanning (EM) and transmission electron microscopy (TEM)

Thin sectioning TEM of density gradient-purified C. difficile spores was as described previously [29]. For SEM, highly purified spores were fixed with 2.5% glutaraldehyde, 1% formaldehyde, 0.1M phosphate buffer during 30 min at room temperature. Samples were washed 3 times with 0.2 M CaCl₂ and 0.1% Tween 20, dehydrated at room temperature with increasing concentrations of ethanol washes up 100% and finally dehydrated in 100% acetone during 5 min. Samples were mounted on glass slides and covered with gold. Scanning electron microscopy was performed on a Hitachi SU-8010 field emission gun operated at 1.5kV.

AlphaFold2 modelling

AlphaFold2 [60] was used to predict the structure of C. difficile YabG. No template structures were used in the prediction, which was iterated for up to 48 recycles, followed by energy refinement with AMBER using default settings implemented in ColabFold [105] and using MMseqs2 for creating multiple sequence alignments [106]. The confidence of the modelling was assessed by the pLDDT metric and the predicted alignment error (PAE), i.e., the uncertainty about the interface (S2 Fig). Values of pLDDT > 90 are expected to have high accuracy. Five YabG models were generated; rank 1 model is represented in Fig 1C. The statistics associated with each of the five models are shown in S2 Fig.

Animal studies

All animal studies were performed with prior approval from the Texas A&M University Institutional Animal Care and Use Committee. Female Syrian golden hamsters, 80g - 120g, were housed individually in cages and had ad libitum access to food and water for the duration of the experiment. To induce susceptibility to C. difficile infection, hamsters were gavaged with 30 mg/kg clindamycin [84,85]. After 5 days, hamsters were gavaged with 1.000 spores of WT C. difficile (n = 7) or C. difficile ΔyabG (n = 8) and monitored for signs of disease (lethargy, poor fur coat and wet tail). Hamsters showing signs of disease were euthanized by CO₂ asphyxia followed by thoracotomy as a secondary means of death in accordance with Panel on Euthanasia of the American Veterinary Medical Association. Fecal samples were collected daily and weighed, suspended in 1 mL of sterile water and dissociated with a pipette. The samples were serially diluted and plated on to TCCFA agar medium to enumerate both vegetative C. difficile cells and spores. CFUs were tabulated and expressed as CFU/g feces. The data has been set to the median values and the limit of detection was 10³ CFU/g.