Complexity of Clostridial Trx systems

We analyzed the composition of Trx systems in a database containing 349 genomes representative of the bacterial diversity from Megrian et al. [38] using BLAST [39] with TrxA and TrxB sequences from Escherichia coli and C. difficile. Distinction between NTR and non-NTR was made by identification of the NAD(P)-binding site [11] (S1A Fig) in TrxB sequences. Results were then reported on a schematic tree of the prokaryote phylogeny [40] (Fig 1A). A high variability of the number of TrxAs is observed as previously described [9,41,42]. Conversely, most of the bacteria, including those for which Trx systems have been extensively studied such as E. coli, Mycobacterium tuberculosis and Helicobacter pylori, harbor a unique TrxB. Some phyla contain more than one TrxB, which is often correlated with the presence of a non-NTR. Among them, Firmicutes is the most remarkable one with up to four TrxBs in the same bacterium. Within this phylum, a diversity of the number of TrxBs is observed, with only one TrxB for the Bacilli and Tissierellales, and two to four copies for the Negativicutes and Clostridia with a few exceptions (Fig 1B). In the latter two groups, most of the bacteria have both types of TrxB: NTR and non-NTR, which are probably FFTR [11]. Despite the numerous studies on the role of Trx systems in bacterial physiology, their role in Clostridia and the implication of the presence of multiple systems have never been studied.

Download:

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

Fig 1. Composition of Trx systems across Bacteria.

Composition of the Trx systems among the bacterial phylogeny (A) or among Firmicutes (B). The schematic phylogeny is based on Witwinosky et al. [40] (A) and on Taib et al. [43] (B). TrxA and TrxB are indicated by circles and squares, respectively. NTR are indicated in green and non-NTR in orange. Dark border indicates conservation among the phyla (A) or within the family (B), forms without border indicate presence in some individuals of the phyla but not in all of them. Chimeric enzymes are proteins with fused TrxA and TrxB domains. Firmicutes are highlighted in yellow (A). C. difficile is highlighted in bold red.

https://doi.org/10.1371/journal.ppat.1012001.g001

We focused on C. difficile, a genetically manipulatable bacterium with a well-characterized lifecycle. The reference strain (630Δerm) of our model bacterium has three TrxAs (TrxA1/CD630_16900, TrxA2/CD630_30330, TrxA3/CD630_23550) and three TrxBs (TrxB1/CD630_16910, TrxB2/CD630_21170, TrxB3/CD630_23560). TrxB2 and TrxB3 harbor NAD(P)H binding sites with GxGxxA and HRRxxxR motifs [11], suggesting that these two proteins are NTRs (S1A Fig). By contrast, TrxB1 lacks these NAD(P)H binding motifs and clusters with other clostridial FFTRs, suggesting that this copy uses ferredoxin as cofactor (S1A–S1B Fig). Finally, TrxA1 and TrxA2 harbor a classical TrxA WCGPC motif [5], while TrxA3 presents an atypical GCEPC motif described in other Clostridia [44] (S1C Fig).

Regulation of C. difficile Trx systems

A first step in deciphering the complexity of Trx systems is the analysis of their genetic organization and the regulation of their expression. The trxA1 and trxB1 genes form a two-gene operon, while trxA3 and trxB3 are part of the grd operon encoding the glycine-reductase [45] (Figs 2A and S1D). By contrast, the trxA2 and trxB2 genes are monocistronic and located in two distinct loci on the chromosome. Analysis of the genome-wide transcription-start site (TSS) mapping data [46] shows the presence of σ^A-recognized promoters upstream of the TSSs of trxA2 and the grd operon containing trxA3B3. 5’RACE technique was performed to identify the TSSs of the trxA1B1 operon and the trxB2 gene. Both TSSs were preceded by consensus sequences corresponding to promoters recognized by σ^B, the sigma factor of the general stress-response [33] (S2A Fig). We then compared by qRT-PCR the expression of the trx genes in the wild-type (WT) 630Δerm strain and in the sigB::erm mutant [33] (Fig 2B). As expected, no significative differences were observed for the expression of the trxB3 and trxA2 genes in the two strains, while the expression of the trxB1 and trxB2 genes decreased in the sigB mutant compared to the WT strain, confirming the σ^B-dependent control of these genes.

Download:

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

Fig 2. Expression of C. difficile Trx systems.

(A) Genetic organization and transcriptional regulation of trx genes. The four trx loci are represented, with trxA genes in light blue and trxB genes in green. The two organizations of trxA2 (in trxB4 positive and negative strains) are represented with the additional trxB4 represented in pink. (B, C) Expression of trx genes was measured by qRT-PCR in (B) WT strain and sigB mutant after 4.5 h of growth in TY and in (C) WT strain after 24 h of growth in TY in anaerobiosis or at 1% O₂. Mean and Standard Deviation (SD) are shown. Experiments were performed in 6 biological replicates. One sample t-tests were used with comparison of the fold change to 1. (D) Pictures of C. difficile WT and sigB mutant strains carrying a transcriptional fusion P_trxA1/B1-FAST^CD. Cells were cultivated during 24 h in TY liquid medium in anaerobic or hypoxic conditions (1% O₂) and observed at high magnification (60X). Images are overlays of bacterial autofluorescence (AF) (green) and FAST (red). Scale bars represent 10 μm. (E) Average P_trxA1B1-FAST^CD fluorescence intensity of individualized bacteria from acquired images of panel D. Each group consists in the measure of the average P_trxA1B1-FAST^CD of 600 cells from two independent experiments. Kruskal-Wallis tests were performed followed by Dunn’s multiple comparison test. *: p-value<0.05, ** <0.01, *** <0.001 and **** <0.0001.

https://doi.org/10.1371/journal.ppat.1012001.g002

As C. difficile faces low O₂ tensions in the gut [47], and because of the role of Trx systems in oxidative stress response in other bacteria, we investigated the effect of O₂ on the expression of trx genes. Using qRT-PCR, we showed that exposure to 1% O₂ for 24 h led to a significant increase of expression of the trxB1 and trxA2 genes while the expression of trxB2 and trxB3 remained unchanged (Fig 2C). To confirm the induction of trxA1B1 expression, we used a transcriptional fusion between the promoter of this operon and the FAST^CD reporter gene [48,49]. The FAST system is O₂ independent and functions both in anaerobiosis and in the presence of O₂. This system has been successfully used in Clostridium acetobutylicum [49]. While expression of the P_trxA1B1-FAST^CD fusion was not detected in anaerobiosis, it was significatively induced in the presence of 1% of O₂ (Fig 2D and 2E). The absence of fluorescent signal in the presence of O₂ in the sigB::erm mutant confirmed the σ^B-dependent control of this operon. Finally, we detected in the WT strain a heterogeneity of expression of the P_trxA1B1-FAST^CD fusion, as observed previously for other σ^B targets [50].

In summary, the trx loci are differentially regulated (Fig 2A). The trxA1B1 operon is σ^B-regulated and is induced in presence of O₂, while trxB2 is not induced by O₂ despite its σ^B-dependent regulation. The trxA2 gene is also induced in presence of O₂ but is transcribed from a σ^A-dependent promoter, indicating that O₂ induction and σ^B-dependent control are not linked. Finally, the trxA3B3 genes are part of the grd operon and are likely transcribed by σ^A. A glycine riboswitch has been previously identified upstream of the grd operon [51], suggesting an induction by glycine, as demonstrated in a recent study [52].

Trx systems are not essential in C. difficile

To investigate the role of the three Trx systems in C. difficile physiology, we inactivated the different trx genes. We were able to obtain mutants for all trx genes except trxB3 despite several attempts, and all the corresponding multi-mutants were obtained. We also obtained complemented strains for the different mutants with plasmids carrying the corresponding genes under the control of their own promoters. We were able to obtain a triple trxA mutant indicating that Trx systems are not essential in C. difficile. In addition, no growth defects were observed for any of the single or multiple mutants in TY medium (S3A Fig). However, the survival of the trxA triple mutant in stationary phase was strongly impacted, with a defect of more than two-logs already visible at 24 h (S3B Fig). The survival of the ΔtrxA1/ΔtrxA2 double mutant was also affected, but to a lesser extent than the triple mutant. No survival defect was observed for the trxB mutants (S3C Fig). These results indicated that even if TrxAs are not necessary for C. difficile growth, they are important for its survival during stationary phase, with a more important role of TrxA1 and TrxA2 compared to TrxA3.

Two Trx systems favor the survival of C. difficile when exposed to O₂ and inflammation-related stresses

Even though C. difficile is a strict anaerobe and a gut resident, this pathogen is exposed to various low O₂ tensions during its infectious cycle [47]. C. difficile must also deal with immune cells [32] and inflammation-produced molecules [27]. All these stresses target principally thiols [53], and we demonstrated that some trx genes were induced by low O₂ tensions or controlled by σ^B. Thus, we investigated the role of Trx systems in stress tolerance by assessing the survival of all the trx mutants upon exposure to different molecules.

We first evaluated the survival in the presence of 1% O₂ to mimic the O₂-level at the mucus layer [47] and at 0.1% of O₂ which corresponds to O₂-tension in a dysbiotic gut lumen [47]. The trxB1::erm/ΔtrxB2 double mutant and the ΔtrxA1/ΔtrxA2/ΔtrxA3 triple mutant did not grow on plates when exposed to 1% O₂ in contrast to the WT strain, showing the importance of Trx systems in O₂ tolerance (Fig 3A and 3B). Furthermore, the ΔtrxA1/ΔtrxA2 double mutant displayed the same phenotype as the triple mutant at this O₂ tension, whereas the two other trxA double mutants had a survival rate comparable to the WT strain. These results suggest that TrxA1 and TrxA2 but not TrxA3 are important for survival in the presence of O₂. The O₂ tolerance of all single mutants was similar to the one of the WT strain, suggesting functional redundancy for TrxA1 and TrxA2 as well as TrxB1 and TrxB2. The phenotypic complementation of double and triple mutants by only one of the above-mentioned partners further confirms this conclusion (S4A Fig). The trxA3 gene did not complement the trxA triple mutant (S4A Fig), supporting the absence of involvement of TrxA3 in O₂ tolerance. The same results were obtained in the presence of 0.1% of O₂ (S4B Fig).

Download:

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

Fig 3. Trx systems are required for survival to oxidative and nitrosative stresses.

(A, B) Samples were serially diluted, plated in duplicate on TY Taurocholate (Tau) plates and incubated either in anaerobiosis or in hypoxia at 1% O₂ for 64 h. Survival was calculated by doing the ratio between CFUs in the last dilution with stress and CFUs in the last dilution without stress. Survival was then normalized as the ratio of the mutant vs the WT. (C, D) Samples were serially diluted and plated on TY and (C) TY + DEA NONOate 750 μM or (D) TY + HClO 0.1% and incubated for 24 h. (E) Samples were incubated in glycylglycine buffer in the presence or absence of H₂O₂ 250 μM for 30 min, serially diluted and plated on TY plates. Mean and SD of survival are shown. Experiments were performed in 5 biological replicates. For all assays, ordinary one-way ANOVA was performed followed by Dunett’s multiple comparison test. The comparisons are made with the WT strain. *: p-value<0.05, ** <0.01, *** <0.001 and **** <0.0001.

https://doi.org/10.1371/journal.ppat.1012001.g003

We then wanted to study the role of Trx systems in an inflammation context by evaluating the tolerance of trx mutants to molecules produced by immune cells. As observed for O₂ tolerance, the trxB1::erm/ΔtrxB2 double mutant, the ΔtrxA1/ΔtrxA2 double mutant and the trxA triple mutant were more sensitive to DEA NONOate, a NO donor (Fig 3C), to hypochlorite (HClO) (Fig 3D) and to hydrogen peroxide (H₂O₂) (Fig 3E), confirming the key role of these four proteins in stress response to inflammation-produced molecules. By contrast, the ΔtrxA1/ΔtrxA3 and ΔtrxA2/ΔtrxA3 mutants had a survival rate comparable to the WT strain (Fig 3C–3E), indicating a marginal role of TrxA3 in the survival to oxidative and nitrosative stress, as observed for O₂ tolerance. Assays on complemented strains (S4 Fig) validated these observations.

Altogether, these results point out that despite the existence of specific detoxication enzymes for O₂ and for oxidative and nitrosative stresses, Trx systems, and more precisely two TrxAs (TrxA1 and TrxA2) and two TrxBs (TrxB1 and TrxB2), are crucial for C. difficile survival when exposed to these stresses.

A supplementary TrxB is part of the stress-response arsenal in some strains of C. difficile

An in silico investigation for additional Trx partners in NCBI available C. difficile assembled genomes led to the identification, in some strains, of a fourth TrxB named TrxB4. In these strains, the trxB4 gene is located downstream of the trxA2 gene, which is monocistronic in the 630Δerm strain. The trxB4 gene when present forms an operon with trxA2. The trxB4 gene is widespread in clades 3 and 5 of C. difficile but is also found sporadically in some strains of other clades (Fig 4A). The maximum likelihood tree of the trxB4 gene follows the core-genome C. difficile phylogeny (S5A Fig), suggesting an ancestral origin of trxB4. We analyzed the genetic environment of trxA2 in C. difficile strains representative of the diversity of the species (Fig 4B). We found that a variable region was delimited by conserved genes, the trxA2 gene on one side and CD3026 encoding a hypothetical protein on the other side. In the 630Δerm, this variable region is composed of genes encoding a phosphotransferase system (CD3027 to CD3031) and a PLP-dependent aminotransferase (CD3032). These genes are absent in trxB4-containing strains, at the exception of the ancestral SA10-050 strain, which is used to root the tree (Fig 4A) [54]. In clade 5 strains carrying trxB4, the variable region is composed of an ABC transporter, a membrane protein, a two-component system and trxB4. These genes are also present in the SA10-050, confirming that all these genes, including trxB4, were ancestral and that several independent gene losses events probably occurred. Consistently, the trxB4-containing strains outside of clade 5 harbor scars of the two variable regions, illustrating at least two independent deletions (Fig 4B).

Download:

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

Fig 4. The fourth TrxB of C. difficile.

(A) Repartition of the trxB4 gene in a data set of 194 genomes. The core genome was used to reconstruct C. difficile phylogeny. The presence of trxB4 was assessed using BLASTN [55] and indicated by pink boxes in the outer layer. Final tree was produced with iTOL [56]. (B) Genetic environment of trxA2 genes. Sequence of trxA2 environments were analyzed using MicroScope platform [57]. (C-E) Complementation of the trxB1/trxB2 double mutant by the trxB1, the trxB2 or the trxB4 gene. (C) Samples were serially diluted, plated in duplicate on TY Tau plates and incubated either in anaerobiosis or in hypoxia at 1% O₂ for 64 h. (D, E) Samples were serially diluted and plated on TY and on (D) TY + HClO 0.1% or (E) TY + DEA NONOate 750 μM. Plates were incubated for 24 h. Experiments were performed in 5 biological replicates. Mean and SD are shown. For stress assays, one-way ANOVA was performed followed by Dunett’s multiple comparison test. *: p-value<0.05, ** <0.01, *** <0.001 and **** <0.0001.

https://doi.org/10.1371/journal.ppat.1012001.g004

As TrxB1, TrxB4 lacks the NAD(P)H binding motifs and clusters with clostridial FFTRs, strongly suggesting that TrxB4 is a FFTR (S1A–S1B Fig). Given the importance of TrxB1 and TrxA2 in stress-response, we investigated the role of TrxB4 in this adaptive process. We complemented the trxB1::erm/ΔtrxB2 double mutant with the trxB4 gene of the E1 strain [58] expressed under the control of its own promoter, as we show that the σ^A-recognized promoter located upstream trxA2 was conserved between the 630Δerm and E1 strains (S5B Fig). The expression of the trxB4 gene in the trxB1::erm/ΔtrxB2 mutant restored a level of O₂-survival comparable to the complementation obtained with the trxB1 and trxB2 genes (Fig 4C). The same observations were made with HClO and NO (Fig 4D and 4E), confirming the involvement of TrxB4 in stress-response.

Two Trx systems are required to cope with disulfide-bond formation induced by some bile salts

Bile acids are inextricably linked to the life cycle of C. difficile. Primary bile acids such as cholate (CHO), which are abundant in the CDI context, allow germination [21,23], whereas secondary bile acids such as deoxycholate (DOC) are known to prevent C. difficile colonization through growth inhibition [59]. C. difficile colonizes the gut at very low concentration of secondary bile acids [23], but faces higher levels during the microbiota recovery phase [60], leading to biofilm formation and/or clearance of the bacteria [61]. Hence, tolerance to bile acids is important for C. difficile transmission, persistence, and CDI recurrence.

To determine the implication of Trx systems in bile acids susceptibility, we exposed the trx mutants to DOC and CHO (Fig 5A and 5B). As observed for inflammation-related stresses, we found a decrease in survival of multi-mutants of trxA1, trxA2, trxB1 and trxB2 genes, suggesting their importance for C. difficile survival in presence of DOC, in agreement with a function in stress-response. For CHO, we observed a lesser effect compared to DOC, with only a significant sensitivity for the trxA triple mutant. The sensitivity of the complemented multi-mutants was in agreement with these results, with a phenotypic complementation only by trxA1 or trxA2 and trxB1 or trxB2 (S6A and S6B Fig). Interestingly, this phenotype was not seen when glycodeoxycholate (GlyDOC) was used (S6C Fig), suggesting a specific action of CHO and DOC.

Download:

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

Fig 5. Trx systems are required to tolerate disulfide-bond forming activity of bile acids.

Samples were serially diluted and plated on TY and on (A) TY + DOC 0.03%, (B) TY + CHO 0.4%, (C) TY + Triton X-100 0.01%, (D) TY + SDS 0.001% or (F) TY + diamide 0.2 mM. The plates were incubated for (F) 24 h or (A-D) 48 h. (E) WT and ΔtrxA1/ΔtrxA2/ΔtrxA3 mutant (trxA) were serially diluted and plated on TY and on TY containing either DOC 0.03%, DTT 0.1% or DOC 0.03%, and DTT 0.1%. The plates were incubated for 48 h. Survival was calculated by doing the ratio between CFUs in the last dilution with chemical and CFUs in the last dilution without chemical. Experiments were performed in 5 biological replicates. Mean and SD are shown. For all assays, ordinary one-way ANOVA was performed followed by Dunett’s multiple comparison test. *: p-value<0.05, ** <0.01, *** <0.001 and **** <0.0001.

https://doi.org/10.1371/journal.ppat.1012001.g005

To assess if this increased sensitivity of the mutants was due to the detergent activity of bile acids, we exposed the trx mutants to other detergents (Triton X-100 and SDS) (Fig 5C and 5D). We observed no sensitivity of the trx mutants at concentrations described to be critical for growth of a detergent-sensitive strain, the ΔbusAA mutant, used as a control (S6D and S6E Fig) [62]. The trx mutants are therefore not sensitive to detergents, suggesting that their bile acids-sensitivity is linked to another physico-chemical property. Bile acids have been described to form disulfide bonds [63], which could explain the observed phenotypes. To test this hypothesis, we exposed the trxA triple mutant to DOC in presence of dithiothreitol (DTT), a disulfide-bond reducing agent [64] (Fig 5E). The complete restoration of survival in presence of DTT confirmed that a disulfide bond-dependent mechanism is responsible for the increased susceptibility of trx mutants to DOC. This hypothesis is also supported by the fact that DOC-sensitive trx mutants were also more sensitive to diamide, a chemical triggering disulfide-bond formation [64] (Figs 5F and S6F). Conjointly, these results show the importance of Trx systems in the repair of disulfide-bonds induced by some bile acids, CHO and DOC.

The third Trx system is involved in sporulation

Sporulation is a key step of C. difficile life cycle. After gut colonization, the formation of spores is triggered by an undetermined signal. The newly-formed spores disseminate in the environment during diarrhea. As this metabolically inactive form is resistant to many stresses including air and acidic pH, it allows the transmission to new hosts. We therefore evaluated the contribution of Trx systems to sporulation. We first evaluated the sporulation capacity of the single mutants (Fig 6A). The ΔtrxA3 mutant showed a significantly lower rate of sporulation, suggesting a role of TrxA3 in this process. The sporulation rate of the ΔtrxA1/ΔtrxA3 and the ΔtrxA2/ΔtrxA3 mutants were even lower (Fig 6B), while the ΔtrxA1/ΔtrxA2 mutant had the same sporulation efficiency than the WT strain. These results suggest a major role for TrxA3 and a slight contribution of the two other TrxAs to sporulation in the absence of TrxA3.

Download:

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

Fig 6.

TrxA3 contributes to spore formation (A-D) Effect of trxA gene inactivation on sporulation. Sporulation rate of WT strain and (A) trxA single mutants, (B) trxA double mutants and (C) ΔgrdAB mutant was evaluated daily over 4 days by numeration of total cells and spores by serial dilution and plating on TY + Tau before (total CFUs) and after (spores) ethanol shock. (D) Sporulation rate of WT strain in TY or TY + 20 mM glycine was evaluated. Experiments were performed in 5 biological replicates. Mean of sporulation rate and Standard Error of the Mean (SEM) are shown. For sporulation assays, t-tests were performed comparing the sporulation rate of the condition and the sporulation rate of the WT in TY. *: p-value<0.05, ** <0.01.

https://doi.org/10.1371/journal.ppat.1012001.g006

The trxA3 gene is part of the glycine reductase (grd) operon (S1D Fig). As other proteolytic Clostridia, C. difficile is able to extract energy from Stickland fermentations [15]. The grd operon contains the grdABCDEX genes encoding the enzymes involved in glycine reduction, and a Trx system (trxA3/trxB3) required for the reduction of the selenoenzyme GrdA [65]. To determine if the sporulation defect was associated with the glycine-reductase activity, we studied the sporulation of a ΔgrdAB mutant (Fig 6C). We observed a sporulation defect, suggesting that GrdAB contributes to the efficiency of spore formation. As glycine is a known co-germinant of C. difficile spores [59], we hypothesized that the observed reduction of sporulation efficiency could be due to a reduced consumption of glycine. Addition of glycine in the media decreased the sporulation rate of the WT strain at 24 h (Fig 6D), but at a lower extent than mutations of grdAB or trxA3, suggesting the existence of another mechanism to explain the observed phenotypes. Altogether, these results show that the third Trx system of C. difficile, TrxA3 and likely TrxB3, plays a role in the control of sporulation likely through a GrdAB-dependent mechanism.

Regulation and role of the spore Trx system

The link between Trx systems and the spores is not restricted to the grd operon. Interestingly, TrxA1 and TrxB1 have been detected in the spore proteome of the 630Δerm strain [66]. We used a P_trxA1B1-trxA1’-FAST^CD translational fusion to confirm the presence of TrxA1 in the spores. We confirmed that this protein fusion was able to complement the ΔtrxA1/ΔtrxA2 mutant in a O₂-survival experiment (S7A Fig). TrxA1 was detected both in the vegetative cells and in the sporulating cells, mainly in the forespore (Fig 7A). Using a P_trxA1B1-FAST^CD transcriptional fusion (Fig 7B), and the membrane specific marker MTG, we showed that the trxA1B1 operon was expressed in the vegetative cell (yellow arrows), in the mother cell (pink arrows) and in the forespore (blue arrows). The expression of this fusion was detectable in all the compartments at the stage of asymmetric division. From the engulfment to the maturation of the spore, the expression was significantly increased in the mother cell compared to the vegetative cell, and higher in the forespore compared to the mother cell (Figs 7C and S7B). By contrast, the expression of the P_trxA1B1-FAST^CD fusion was not visible anymore in the phase bright forespore (Fig 7C). In a sigB::erm mutant (Fig 7B and 7C), the expression of the fusion was no longer detectable in the vegetative cells consistently with our previous observations (Fig 2D) but also in the mother cell, suggesting that the expression in this compartment was σ^B-dependent. Interestingly, the expression of this fusion in the sigB::erm mutant was still visible in the forespore, suggesting the existence of an alternative σ^B-independent transcription of trxA1B1 during sporulation.

Download:

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

Fig 7. Regulation of the trxA1B1 operon during sporulation.

(A) The location of TrxA1 during sporulation was evaluated using the translational P_trxA1B1-trxA1’-FAST^CD fusion. Bacteria were incubated in anaerobiosis for 48 h in sporulation medium and observed at high magnification (100X). Images are overlays of bacterial autofluorescence AF (green) and FAST (red). Scale bars represent 10 μm. White arrows indicate a forespore. (B) Expression of the trxA1/B1 operon was monitored using the transcriptional P_trxA1B1-FAST^CD fusion in the WT strain and in the sigB::erm mutants. Bacteria were cultured for 48 h in anaerobiosis in sporulation medium. Images are overlays of phase contrast, MTG (green) and FAST (red). Scale bars represent 10 μm. Yellow arrows indicate vegetative cells, blue arrows forespores and pink arrows mother cells. (C) Expression of the trxA1/B1 over the stages of sporulation. The different stages of sporulation of the WT strain are shown. Yellow arrows indicate vegetative cells, blue arrows forespores and pink arrows mother cells. The schematic bacteria indicate the sigma factors present in the compartments over sporulation. The sigB::erm mutant is shown at the stage of engulfment. The sigF::erm mutant blocked at the stage of asymmetric division forms disporic cells and the sigG::erm mutant is blocked at the stage of engulfment. Scale bars represent 3 μm. (D) Promoter identification through 5’RACE using RNA extracted from the sigB::erm mutant grown in sporulation medium. The TSS (+1) is indicated in red. Upstream this TSS, σ^B and σ^G boxes [46] are represented in orange and green, respectively.

https://doi.org/10.1371/journal.ppat.1012001.g007

Four sigma factors are involved in sporulation: σ^E and σ^K in the mother cell and σ^F and σ^G in the forespore [67]. Given the fact that the trxA1B1 operon was expressed in the forespore in the sigB::erm mutant, σ^F and σ^G are the best candidates to be the alternative sigma factors that transcribe this operon during sporulation. However, the corresponding promoter was not identified by our first 5’RACE performed with RNA extracted from exponentially growing cells (S2A Fig). We mapped again the TSS using RNA extracted from a sigB::erm mutant after 24 h of growth in sporulation medium. We were able to identify a new TSS and the consensus sequences of a promoter recognized by σ^G (Figs 7D and S7C and S7D). To confirm this control, we introduced the P_trxA1B1-FAST^CD transcriptional fusion in the sigF::erm and the sigG::erm mutants [67] (Fig 7B and 7C). As σ^F is necessary for sigG expression [68], a σ^F-transcribed gene is not expressed in a sigF::erm mutant but expressed in a sigG::erm mutant, while a σ^G-transcribed gene is expressed neither in a sigF::erm nor in a sigG::erm mutant. In a disporic sigF::erm mutant cell, the P_trxA1B1-FAST^CD fusion was expressed only in the mother cell, likely under the control of σ^B, but not in the forespores (Fig 7C). The same observation was made for the sigG::erm mutant blocked following engulfment completion. These results confirm the σ^G-dependent control of the expression of the trxA1B1 operon during sporulation in agreement with the second promoter identified.

To understand the function of this Trx system in the spore, we first performed a sporulation assay using a ΔtrxA1/B1 double mutant (Fig 8A). This mutant has a similar sporulation rate as the WT strain, except after 24 h where the percentage of spores was significantly higher. This quicker sporulation is not due to a difference in growth (S7E Fig). The TrxA1/B1 system might thus contribute to delay sporulation but is clearly not required for sporulation. We then performed a germination assay using spores of both the WT and the ΔtrxA1/B1 mutant (Fig 8B). The germination was followed in anaerobiosis and in air (Fig 8B–8C) through measurement of the decrease of OD_600nm of a spore suspension following addition of 1% taurocholate. The WT and ΔtrxA1/B1 mutant spores germinated similarly in anaerobiosis (Fig 8B), but the total OD_600nm reduction was significantly lower in the mutant in aerobiosis (Fig 8C), suggesting a lower efficiency of germination. We also evaluated the germination in aerobiosis after 20 and 180 min under the microscope (Fig 8D and 8E). After 20 min, all the spores of the WT strain were phase-dark, while a subpopulation of approximately 15% of spores of the ΔtrxA1/B1 mutant remained phase-bright. After 180 min, this subpopulation only represented 5% of spores, which was still significantly higher than the WT strain. In addition, the reduced decrease of OD_600nm observed in aerobiosis for the ΔtrxA1/B1 was still observed when DTT was added (S7F Fig). This result suggests that the reduced efficiency of germination is not due only to the disulfide bond reduction activity of Trx systems. Moreover, as a phenotype was only visible in aerobiosis, it seems that this germination defect is due to an increased sensitivity to oxidative stress. We also performed an outgrowth assay in anaerobiosis of the mutant, but we observed no difference with the WT strain (S7G Fig). However, as this assay requires anaerobiosis to observe growth, we are not able to conclude if the germination defect observed in aerobiosis would lead to an outgrowth defect in similar conditions.

To validate the hypothesis of a link with oxidative stress and knowing the impact of these TrxA1/B1 in oxidative stress tolerance (Fig 3), we tested the viability of the spores when exposed to a sporicidal molecule, HClO (Fig 8F). After 10 min of exposure at 0.1% HClO, we found that the survival of ΔtrxA1/B1 spores was more impacted than the one of the WT spores with a 33-fold decrease of spore survival. These results confirm the role of the TrxA1B1 system in the protection of the spore against oxidative stress and suggest a role during germination.

Download:

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

Fig 8. The role of TrxA1/TrxB1 in spore germination and hypochlorite resistance.

(A) Sporulation rate of the ΔtrxA1/trxB1 mutant was evaluated daily over 4 days as described in Fig 6. Mean of sporulation rate and SEM are shown. (B—E) Role of TrxA1/B1 in germination. ~10⁷ spores of WT strain and of ΔtrxA1/B1 mutant were exposed to 1% taurocholate to induce germination, either (B) in anaerobiosis or (C) in air. OD_600nm was monitored every 5 min to evaluate germination. Experiments were performed in 5 replicates with at least 2 independent spore suspensions. Mean and SD are shown. (D) ~10⁷ spores of WT strain and of ΔtrxA1/B1 mutant were exposed to 1% taurocholate to induce germination. After 0, 20 and 180 min, spores were washed and fixed in 4% PFA before observation under phase contrast microscopy at high magnification (60X). Scale bars represent 5 μm. Yellow arrows indicate bright phase spores. (E) Percentage of bright phase spores were quantified from acquired images of panel C from 900 cells from two independent experiments. Mean and SD are shown. (F) Role of TrxA1/B1 in resistance to HClO treatment. 10⁸ spores were exposed to 0.1% of HClO for 10 min, neutralized with 1% sodium thiosulfate, washed and diluted and spotted on TY Tau plates for numeration. Survival was estimated by calculating the ratio of CFUs between treated and non-treated spores. Experiments were performed in 5 biological replicates. Mean and SD are shown. For sporulation assays, t-tests were performed comparing the sporulation rate of the WT and the mutant. For germination assays, two-way ANOVA were performed. For comparison of the proportion of bright phase spores, multiple unpaired t tests were performed. For sporicidal assays, unpaired t-tests were performed. *: p-value<0.05, ** <0.01.

https://doi.org/10.1371/journal.ppat.1012001.g008

Phylogenetic analysis of bacterial Trx systems

A dataset of 349 diversity-representative bacteria was used as the database for all bacteria [38]. A dataset of 67 bacteria was used as the Firmicutes database. BLASTP [39] was used using TrxB sequence from E. coli (accession number P0A9P5—TRXB_ECO57) and TrxB1 sequence from C. difficile. Presence of the NAD(P)H binding motifs (GxGxxA and HRRxxxR motifs [11]) were investigated for detection of NTR, while non-NTR were defined by the absence of the motifs. Schematic phylogeny was made following the one in Witwinoski et al. [40] for all bacteria and following Taib et al. [43] for Firmicutes. The number of TrxAs and TrxBs was reported following BLASTP results.

Bacterial strains and culture media

C. difficile strains and plasmids used in this study are listed in S1 and S2 Tables. C. difficile strains were grown anaerobically (5% H₂, 5% CO₂, 90% N₂) in TY (Bacto tryptone 30 g.L^-1, yeast extract 20 g.L^-1, pH 7.4), in Brain Heart Infusion (BHI; Difco) or in sporulation medium [90]. For solid media, agar was added to a final concentration of 17 g.L^-1. When necessary, thiamphenicol (Tm, 15 μg.mL^-1), erythromycin (Erm, 2.5 μg.mL-¹) and cefoxitin (Cef, 25 μg.mL^-1) were added to C. difficile culture. E. coli strains were grown in LB broth. When indicated, ampicillin (Amp, 100 μg.mL^-1) and chloramphenicol (Cm, 15 μg.mL^-1) were added to the culture medium. When indicated, the spore germinant taurocholate (Tau) was added in plates at 0.05%.

Construction of C. difficile mutant strains

All primers used in this study are listed in S3 Table. The trxB1::erm mutant was obtained by using the ClosTron gene knockout system as previously described [91]. The PCR product generated by overlap extension that allows intron retargeting to trxB1 was cloned between the HindIII and BsrG1 sites of pMTL007 to obtain pDIA6190. The ΔtrxB2, ΔtrxA1, ΔtrxA2, ΔtrxA3, ΔtrxA1/B1 and ΔgrdAB knock-out mutants were obtained by using the allele chromosomic exchange method [92]. Briefly, PCR were performed to amplify 1 kb fragments located upstream and downstream of the targeted genes. Using the Gibson Assembly Master Mix (Biolabs), purified PCR fragments were cloned into the pMSR plasmid [92] linearized by inverse PCR and treated by DpnI. The sequence of the resulting plasmids was verified by sequencing. These plasmids (S1 Table) introduced in the HB101/RP4 E. coli strain were then transferred by conjugation into C. difficile strains. Transconjugants were selected on BHI plates supplemented with Tm and C. difficile selective supplement (SR0096, Oxoid). Single crossover integrants were identified by a faster growth on the selective plates. These clones were restreaked into new BHI plates containing Cef and Tm. They were then restreaked on BHI plate containing 200 ng.mL^-1 of aTc. This compound allowed the expression of the CD2517.1 toxin gene cloned under the control of the P_tet promoter and the selection of a second crossover event [92]. The resulting clones were checked by PCR for the expected deletion. Steps were repeated in each different mutant to generate C. difficile multi-mutants.

Complementation of the different mutants

For the construction of the complementation plasmids, different strategies were used (S3 Table). To complement the mutants with trxA1 or trxB1, a fragment containing their promoter region and the trxA1 plus trxB1 genes (IMV1183/IMV1151) was first cloned between the XhoI and BamHI sites of pMTL84121 giving pDIA7025. The plasmid pDIA7025 (pMTL84121-P-trxA1-trxB1) was then linearized by inverse PCR to delete either trxA1 (IMV1200/IMV1201) or trxB1 (IMV1331/CM13) to give pDIA7042 or pDIA7129, respectively. The trxA2 gene or the trxB2 gene with their promoter regions were amplified by PCR using CA24/CA25 (trxA2) or IMV1202/IMV1203 (trxB2) and cloned into pMTL84121 linearized by inverse PCR with the oligonucleotides CM13 and IMV993 using the Gibson Assembly method giving pDIA7122 and pDIA7050. A PCR fragment corresponding to the grd operon promoter region and the grdX-trxB3-trxA3 genes was amplified using IMV1235/CA43. The PCR product was cloned by the Gibson Assembly method into pMTL84121 to produce pDIA7162. The grdX and trxB3 genes were deleted from plasmid pDIA7162 by inverse PCR using IMV1300/IMV1389 giving pDIA7164. For the plasmid containing trxB4, a PCR fragment corresponding to the promoter region and the trxA2 and trxB4 genes was first obtained using CA26/CA27 and DNA from the E1 strain as template. The fragment was cloned into pMTL84121 by the Gibson Assembly method to give pDIA7113. Inverse PCR was then performed with CA75/CA76 to remove the trxA2 gene giving pDIA7156.

All the plasmids were verified by sequencing, introduced into HB101/RP4 E. coli strain and then transferred into C. difficile strains by conjugation. Transconjugants were selected on BHI plates supplemented with Tm and C. difficile selective supplement. The presence of the cloned gene was then verified by PCR.

RNA extraction, qRT-PCR and 5’RACE

Cultures of the WT strain were inoculated in TY at OD_600nm 0.05 and incubated for 24 h in anaerobiosis or with 1% O₂ (BugBox M from Baker Ruskinn). To ensure proper gas diffusion, cultures were made with 2 ml in 6-well plates. Cultures of the WT strain and the sigB::erm mutant were harvested after 4.5 h of culture in TY. Pellets were resuspended in the RNApro solution (MP Biomedicals) and RNA was extracted using the FastRNA Pro Blue Kit (MP Biomedicals). cDNAs synthesis and real-time quantitative PCR were performed as previously described [51,93]. In each sample, the quantity of cDNAs of a gene was normalized to the quantity of cDNAs of the gyrA gene. The relative change in gene expression was recorded as the ratio of normalized target concentrations (the threshold cycle ΔΔCt method) [94]. Experiment was performed in at least 6 biological replicates.

To determine the transcription initiation site of trxA1 and trxB2, the 5’ RACE (Rapid Amplification of cDNA Ends, Invitrogen kit) technique was used. Using SuperScript II reverse transcriptase in the buffer provided (20 mM Tris-HCl (pH 8.4), 50 mM KCl, MgCl2 2.5 mM, dNTP 0.4 mM, DTT 0.01 mM), a single-strand cDNA was synthesized from a site internal to the unknown 5’ end of the mRNA using an oligonucleotide specific of each gene. For the trxA1 gene, we used either mRNA extracted from the WT strain after 5 h of growth in TY or from the sigB mutant after 24 h of growth in sporulation medium. The matrix mRNA was eliminated by treatment with H and T1 RNAses. A polyC tail was then added to the 3’ end of the cDNA via the terminal deoxynucleotidyl transferase. A PCR was then performed using a primer complementary to the polyC tail (containing an arbitrary sequence in 5’) and another primer specific to each cDNA. A second amplification of the obtained PCR product is performed with oligonucleotides specific for the start of each gene and a primer hybridizing with the arbitrary sequence located upstream of the polyC (Abridged Anchor Primer APP). The PCR products were cloned into the plasmid pGEMT easy (Promega). For each gene, 4 to 8 plasmids corresponding to white colony-forming transformants were extracted and sequenced to determine the TSS.

Fusion with the FAST^CD reporter system

To develop a new fluorescent reporter system for use in C. difficile, we turned to the FAST-tag reporter, a 14-kDa monomeric protein, engineered from the photoactive yellow protein, that interacts rapidly and reversibly with 4-hydroxybenzylidene rhodamine derivatives [48,49]. We designed a synthetic FAST cassette with a codon usage optimized for C. difficile (Integrated DNA Technology), which we termed FAST^CD. This cassette was cloned into pFT47 [67] by replacement of the SNAP gene by the FAST^CD gene to give pDIA7190. This plasmid was then linearized by inverse PCR using IMV1103 and either IMV1440 (transcriptional fusions) or IMV1498 (translational fusion). We amplified by PCR a 210 bp DNA fragment corresponding to the promoter region upstream of the trxA1B1 genes using primers IMV1105/IMV1443 (S2 Table). Using the Gibson Assembly method, we inserted this promoter region into pDIA7190 to obtain pDIA7194 carrying a transcriptional P_trxA1B1-FAST^CD fusion. The promoter region and the complete sequence of trxA1 without its stop codon using IMV1105/IMV1499 was also amplified and cloned into pDIA7190 to obtain pDIA7236 carrying a translational P_trxA1B1-trxA1’-FAST^CD fusion. These plasmids were introduced into E. coli HB101 (RP4) and then transferred to C. difficile 630Δerm, sigB::erm, sigF::erm or sigG::erm strain by conjugation. Transconjugants were selected on BHI agar plates containing Tm and Cfx.

Fluorescence microscopy and image analysis

For FAST labelling, cells were harvested after 24 h of growth in anaerobiosis or at 1% of O₂ in TY or after 48 h in sporulation medium in anaerobiosis. Cells were washed once (PBS), fixed in 4% paraformaldehyde (PFA) in PBS for 20 min at room temperature and gently washed 4 times (PBS). Cells were then resuspended in 100 μl in PBS and the TF-coral substrate (Twinkle Factory) was added at a final concentration of 5 μM. The membrane dye Mitotracker Green (MTG) (0.5 μg/ml, ThermoFisher) was added before fixation and the suspension was incubated at room temperature for 15 minutes in the dark.

For phase-contrast and fluorescence microscopy, 5 μL of vegetative and/or sporulating cells were mounted on 1.7% agarose-coated glass slides to keep the cells immobilized and to obtain sharper images. The images were taken with exposure times 600 ms for FAST and FITC. Vegetative cells were observed using a Nikon Eclipse TI-E microscope 60x Objective and captured with a CoolSNAP HQ2 Camera. For the spores, cells were observed using a Delta Vision Elite microscope equipped with an Olympus 100x objective and captured with sCMOS camera. The images were analyzed using ImageJ. Acquisition parameters were similar for all samples of an experiment. Average of FAST fluorescence intensity consists of fluorescence intensity of at least 600 bacteria from two independent experiments from 3 different microscopic fields.

Growth curves, survival and sporulation assays

Overday cultures of C. difficile strains were inoculated by a 1:50 dilution in TY medium. After 5 h of growth, new bacterial suspensions were prepared at an OD_600nm of 0.05 in TY and 100 μl were used in 5 technical replicates for preparation of 96 wells plates that were sealed with a gas-impermeable adhesive film (MicroAmp Optical Adhesive Film, Applied Biosystems). OD_600nm was monitored every hour at 37°C for 24 h by using a GloMax Explorer plate-reader (Promega).

For sporulation assays, 5 ml bacterial suspensions were prepared from an overday culture at an OD_600nm of 0.05 in TY. At 0, 24, 48, 72 and 96 h, the number of total cells was determined by serial dilution and plating on TY Tau. The spore number was determined by replating the dilution after an ethanol shock (volume 1:1 of absolute ethanol) of 1h. Sporulation rate was estimated by calculating the ratio of spores to total cells over time. Total cells data from 0, 24 and 48 h were used for survival curves. Experiments were performed in 5 biological replicates.

O₂-tolerance assays

Overday cultures in TY Tau were used to prepare bacterial suspensions at an OD_600nm of 0.5. After serial dilution (non-diluted to 10⁻⁵), 5 μl of each dilution were spotted on TY Tau plates. Plates prepared in duplicate were incubated at 37°C either in anaerobia or in the presence of 0.1% or 1% of O₂ for 64 h. The last dilution allowing growth after incubation was recorded and the CFUs of this dilution were counted. Survival was calculated by doing the ratio between CFUs in the last dilution in hypoxia and CFUs in the last dilution in anaerobiosis. The results were then normalized by doing the ratio between the results obtained for a mutant with the results of the reference strain, 630Δerm or 630Δerm pMTL84121 for plasmid-carrying strains. Experiments were performed in 5 biological replicates.

Stress tolerance assays

NO, HClO, DOC, CHO, SDS, Triton X-100 and diamide stress assays were performed in TY by using the serial dilution protocol as indicated in the O₂-tolerance section. Dilutions of a suspension at an OD_600nm of 0.5 were plated on TY plate and in TY plate containing either 750 μM of DEA NONOate (Sigma-Aldrich), 0.1% of NaClO (Sigma-Aldrich), 0.03% (725 μM) of DOC (Sigma-Aldrich), 0.4% (9.3 mM) of CHO (Sigma-Aldrich), 0.1% (2.2 mM) of GlyDOC (Sigma-Aldrich), 0.01% of Triton X-100 (Merck), 0.003% of SDS (Euromedex) or 0.2 mM of diamide (Sigma-Aldrich). When indicated, DTT (Sigma-Aldrich) was added at 0.1%. For each plate, the last dilution allowing growth was recorded after incubation at 37°C for 24 h (NO, HClO, diamide) or 48 h (DOC, CHO, SDS, triton). The survival was evaluated by doing the ratio of the CFUs in presence of the stress on the CFUs on the control plate. For H₂O₂, overday cultures were used to prepare two bacterial suspensions per strain at an OD_600nm of 0.5 in glycyl-glycine buffer (glycylglycine 50 mM, glucose 0.2%, pH 8). H₂O₂ (Honeywell) was added in one of the suspensions to a final concentration of 250 μM. After 30 minutes, the suspensions were serially diluted and plated on TY plate and survival was calculated as mentioned above. Experiments were performed in 5 biological replicates.

Spore production, germination, outgrowth and sporicidal assays

Spore suspensions were prepared as previously described [95]. Briefly, 200 μl from overnight cultures of C. difficile strains were plated on SMC agar plates and were incubated at 37°C for 7 days. Spores scraped off with water were incubated for 7 days at 4°C. Cell fragments and spores were separated by centrifugation using a HistoDenz (Sigma-Aldrich) gradient [96]. Spores were enumerated by CFU calculation on TY Tau.

Spore germination was monitored by OD_600nm as previously described [95]. Briefly, a spore suspension was resuspended in BHI and exposed to 1% Tau, and the OD_600nm was measured every 5 min for 1 h, either in anaerobiosis or in air, with or without DTT at 0.1%. Experiments were performed in 5 independent replicates, with at least 2 independent spore preparations. For evaluation under the microscope, a spore suspension was resuspended in BHI and exposed to 1% Tau. After 0, 20 and 180 min, spores were washed and fixed in 4% PFA. Spores were then observed under phase contrast microscopy under 60x magnification using a Nikon Eclipse TI-E microscope 60x Objective and captured with a CoolSNAP HQ2 Camera. The images were analyzed using ImageJ. Percentage of bright phase spores were quantified from 900 cells from two independent experiments from three different microscopic fields.

Outgrowth was monitored by OD_600nm as previously described [97]. A spore suspension was resuspended in BHI and exposed to 1% Tau, and the OD_600nm was measured every 10 min for 1 h followed by every 30 min for 10 h. OD_600nm was normalized by the initial OD_600nm. Experiments were performed in 5 independent replicates, with at least 2 independent spore preparations.

The sporicidal assay was performed as previously described [98]. Spore suspensions were exposed for 10 min to 0.1% HClO or to water. 1 volume of 1% sodium thiosulfate was then added to neutralize HClO, and the spores were washed twice in water. After resuspension, spores were serial diluted and spotted on TY Tau for quantification. Survival corresponds to the ratio between treated spores on untreated spores. Experiments were performed in 5 independent replicates, with at least 2 independent spore preparations.

C. difficile core genome phylogenetic tree

Using Bio.Entrez package of Biopython [99] we retrieved all the accession numbers for C. difficile available in GenBank database [100] (retrieved on 2022-12-27), with the following query: ("Clostridioides difficile"[Organism] OR "Clostridium difficile"[Organism]) AND ("4000000"[SLEN]: "10000000"[SLEN])". 2251 accession numbers were fetched. For those of them that had a sequence associated and whose molecular type was "genomic DNA", we retrieved their sequences and metadata. We thus obtained a dataset of 212 C. difficile genomes. We added three custom genomes to this data set (E1, CD10165 and SA10050) [54,58], obtaining a data set of 215 sequences. Redundant strains were removed, leading to the final data set of 194 genomes. We annotated these genomes in our data set, extracted their core genome and aligned it with PanACoTA [101] (v1.3.1, the gene annotation was performed with Prodigal [102]). We then reconstructed a phylogenetic tree for the core genome with IQ-TREE 2 [103] (v2.2.2, with model selection [104]: -m TEST, partition [105] by 1st, 2nd and 3rd codon positions, and 1000 ultrafast bootstraps [106]). In the resulting tree, we collapsed the non-informative branches (of less than 1/2 mutations) with gotree [107]. The final figure was produced with iTOL [56].

Gene-specific phylogenetic trees

We detected the orthologs of the four trxB genes (trxB1, trxB2, trxB3, and trxB4) in the 194 genomes of our data set with BLASTN [55]. For each of the four genes we created a multiple sequence alignment with Clustal Omega [108] available in SeaView [109] (v5.0.5). Each alignment was manually curated. We reconstructed the phylogenetic trees for each gene with the same procedure as for the core tree. We then compared them to the core tree with ETE3 toolkit [110]: the trees were highly compatible, with 96–98% of their branches present in the core-genome tree. The final figure was produced with Dendroscope [111].

Protein alignment, distance tree and synteny analysis

TrxA and TrxB sequences from diverse bacteria were aligned with MAFFT software [112] using the Auto Strategy option of the software. Alignment was visualized on Jalview software [113]. For TrxBs, distance tree was obtained using neighbor joining method with a JTT substitution model. TrxB from Desulfovibrio vulgaris [114] was used as outgroup to root the tree. Final tree was produced with iTOL [56]. For synteny analysis, grd operons and trxA2 genetic context from Clostridia and C. difficile strains were analyzed using the MicroScope website [57].

Statistical analysis and data presentation

For survival assays, multiple unpaired t-tests were performed. For qPCRs, one sample t-tests were performed with comparison of the fold change to 1. For microscopy, Kruskal-Wallis tests were performed followed by Dunn’s multiple comparison test. For stress assays, one-way ANOVA were performed. When significant, multiple comparison was performed using Dunett’s multiple comparison test. For sporulation assays, unpaired t-tests were performed comparing the sporulation rate of two conditions. For germination and outgrowth assays, two-way ANOVA were performed. For comparison of the proportion of bright phase spores, multiple unpaired t tests were performed. For sporicidal assays, unpaired t-test were performed.

Bar plots, curves, violin plots and statistical analysis were performed with GraphPad Prism 10.0.1 (170). Figs 2A, 4B, S1D and S2A were produced with BioRender.com.

* indicates p-value<0.05, ** <0.01, *** <0.001 and **** <0.0001.