Identification of Clostridioides difficile mutants with increased daptomycin resistance

ABSTRACT

Daptomycin is a cyclic lipopeptide antibiotic used to treat infections caused by some Gram-positive bacteria. Daptomycin disrupts synthesis of the peptidoglycan (PG) cell wall by inserting into the cytoplasmic membrane and binding multiple forms of the undecaprenyl carrier lipid required for PG synthesis. Membrane insertion requires phosphatidylglycerol, so studies of daptomycin can provide insight into assembly and maintenance of the cytoplasmic membrane. Here, we studied the effects of daptomycin on Clostridioides difficile, the leading cause of healthcare-associated diarrhea. We observed that growth of C. difficile strain R20291 in the presence of sub-MIC levels of daptomycin resulted in a chaining phenotype, minicell formation, and lysis—phenotypes broadly consistent with perturbation of membranes and PG synthesis. We also selected for and characterized eight mutants with elevated daptomycin resistance. The mutations in these mutants were mapped to four genes: cdsA (cdr20291_2041), ftsH2 (cdr20291_3396), esrR (cdr20291_1187), and draS (cdr20291_2456). Of these four genes, only draS has been characterized previously. Follow-up studies indicate these mutations confer daptomycin resistance by two general mechanisms: reducing the amount of phosphatidylglycerol in the cytoplasmic membrane (cdsA) or altering the regulation of membrane processes (ftsH2, esrR, and draS). Thus, the mutants described here provide insights into phospholipid synthesis and identify signal transduction systems involved in cell envelope biogenesis and stress response in C. difficile.

IMPORTANCE

C. difficile is the leading cause of healthcare-associated diarrhea and is a threat to public health due to the risk of recurrent infections. Understanding biosynthesis of the atypical cell envelope of C. difficile may provide insight into novel drug targets to selectively inhibit C. difficile. Here, we identified mutations that increased daptomycin resistance and allowed us to better understand phospholipid synthesis, cell envelope biogenesis, and stress response in C. difficile.

INTRODUCTION

Clostridioides difficile is a Gram-positive anaerobic opportunistic pathogen and is the leading cause of healthcare-associated diarrhea. It is estimated that in the United States alone, C. difficile causes over 200,000 infections, about 12,000 deaths, and over one billion dollars in medical costs annually (1). The Centers for Disease Control and Prevention has classified C. difficile as an urgent threat to public health. C. difficile infections can range from mild diarrhea to pseudomembranous colitis and toxic megacolon (1, 2). These infections often occur after antibiotic treatment disrupts the normal gut flora, creating an environment for C. difficile to colonize (1, 3, 4). C. difficile infections usually respond well to treatment with antibiotics, but these drugs may further disrupt the normal gut flora, leading to recurrent infections that are difficult to eradicate (2). There is a need for treatments that target C. difficile selectively.

Daptomycin is a cyclic lipopeptide antibiotic that is used clinically to treat recalcitrant infections with Gram-positive bacteria such as methicillin-resistant Staphylococcus aureus and vancomycin-resistant Enterococcus faecium (5, 6). Although daptomycin’s exact mechanism of action remains to be elucidated, its antimicrobial activity requires both Ca²⁺ and phosphatidylglycerol (7). Previously, it was thought that Ca²⁺-daptomycin oligomers interacted with phosphatidylglycerol to insert into the membrane and form pore-like structures causing depolarization of the membrane and subsequently cell death (8, 9). However, more recent evidence demonstrates that in the presence of phosphatidylglycerol, daptomycin can complex with multiple derivatives of the undecaprenyl carrier lipid required for synthesis of peptidoglycan (PG) (7). It is thought that this disrupts PG biosynthesis; however, daptomycin also causes changes in membrane fluidity by rearrangement of fluid membrane microdomains (10).

In some bacterial species, it has been shown that daptomycin resistance is linked to lowered amounts of phosphatidylglycerol, presumably due to reduced binding opportunities for daptomycin. In Bacillus subtilis and S. aureus, mutations in a phosphatidylglycerol synthase, pgsA, decrease phosphatidylglycerol levels in the membrane and increase daptomycin resistance (11–13). Phosphatidylglycerol levels in the membrane can also be reduced by modification. MprF synthesizes lysyl-phosphatidylglycerol, a positively charged phospholipid, by adding L-lysine to phosphatidylglycerol and translocating lysyl-phosphatidylglycerol to the outer leaflet of the membrane (14, 15). In B. subtilis and S. aureus, mutants of mprF lacking lysyl-phosphatidylglycerol have decreased daptomycin resistance (11, 16). Conversely, overexpression of mprF in B. subtilis or gain of function mutations in mprF in S. aureus increases daptomycin resistance, presumably due to depletion of phosphatidylglycerol (11, 16). It has also been postulated that since lysyl-phosphatidylglycerol is positively charged, the increase in daptomycin resistance could be due to a decrease in the net negative charge of the membrane and consequent reduced affinity for the Ca²⁺-daptomycin complex (11).

The cell envelope is a common target for antibiotics, so understanding the biogenesis of the cell envelope in C. difficile may identify novel strategies to inhibit C. difficile selectively with minimum impact on commensals in the gut microbiome. One promising target is the PG wall, which in C. difficile is unusual in that it contains primarily 3-3 crosslinks rather than the 4-3 crosslinks that predominate in most bacteria (17). In addition, the C. difficile PG is highly deacetylated, with ~80%–90% of GlcNAc residues being deacetylated (17–21). Deacetylation of GlcNAc in C. difficile is the most critical factor in providing a high level of resistance to lysozyme (19, 20, 22, 23). Another noteworthy and potentially druggable feature of the C. difficile cell envelope is the S-layer, a proteinaceous, paracrystalline structure surrounding the outer surface of the cell wall. The S-layer is made up of 29 cell wall proteins, the most abundant being SlpA, which is secreted and proteolytically processed into high- and low-molecular-weight proteins (24, 25). The S-layer is essential for virulence, and an slpA null mutant has attenuated toxin production (26).

A third potential source of antibiotic targets in the C. difficile cell envelope is the cytoplasmic membrane, but comparatively little is known about this essential cell structure. In fact, only recently has membrane polar lipid composition been defined (27). The membrane of C. difficile is composed of ~30% phosphatidylglycerol, 16% cardiolipin, and 50% glycolipids (27). It is also noteworthy that C. difficile lacks phosphatidylethanolamine, phosphatidylserine, and lysyl-phosphatidylglycerol (27). Notably, glycolipids make up 50% of the C. difficile membrane, much higher than in model organisms like Escherichia coli and B. subtilis (28–30). There are four documented glycolipids in C. difficile, namely, monohexose-diradylglycerol (MHDRG; 14% of polar lipids), dihexose-diradylglycerol (DHDRG; 15% of polar lipids), trihexose-diradylglycerol (THDRG; 5% of polar lipids), and aminohexosyl-hexosyldiradylglycerol (HNHDRG; 16% of polar lipids) (27). While MHDRG, DHDRG, and THDRG have been identified in other bacteria (31, 32), HNHDRG has only been found in C. difficile (27). We recently identified the genes (hexSDFRK) required for synthesis of HNHDRG in C. difficile (33). These genes were also required for resistance to daptomycin and bacitracin (33). In addition, we isolated mutants of C. difficile with increased resistance to lysozyme that revealed a novel two-component regulatory system DraRS. We found that DraRS senses and responds to lipid-II binding antibiotics, including daptomycin, and that activation of DraRS leads to increased daptomycin resistance in part by increasing expression of hexSDF (34).

Surotomycin, a daptomycin derivative, is a cyclic lipopeptide antibiotic that was in development to treat C. difficile infections, but development has stagnated due to conflicting results in phase III clinical trials. One phase III clinical trial found that surotomycin performed as well but not better than vancomycin, while another trial found that surotomycin was not as efficacious as vancomycin (35–37). Previously, Adams et al. selected for surotomycin-resistant mutants in C. difficile 630 and E. faecium (38). They identified mutations in cardiolipin synthases in both C. difficile and E. faecium, clsB (cd630_3404; cdr20291_3226) and cls (HMPREF0351_11068), respectively. Mutations in the membrane protease ftsH2 (cd630_3559; cdr20291_3396) and envelope stress response gene esrR (cd630_1345; cdr20291_1187) were also associated with increased surotomycin resistance. However, none of these mutations were complemented or otherwise shown to be responsible for the increased surotomycin resistance.

Here, we describe that sub-MIC levels of daptomycin cause cell division defects including chaining and production of minicells. We also isolated eight independent mutants of C. difficile R20291 with increased daptomycin resistance. These strains harbored unique single nucleotide polymorphisms (SNPs) in four different genes: cdr20291_2041 (cdsA), cdr20291_3396 (ftsH2), cdr20291_1187 (esrR), and cdr20291_2456 (draS). We confirmed mutations in these genes were responsible for increased daptomycin resistance. We then used a subset of these mutants to dissect the synthesis of phospholipids and better understand assembly of the cell envelope of C. difficile.

RESULTS

Daptomycin effects on C. difficile viability and morphology

To determine the effect of sub-MIC concentrations of daptomycin on C. difficile growth and morphology, we subcultured overnight cultures of R20291 into sub-MIC levels of daptomycin and followed growth by measuring the OD₆₀₀. Wild-type (WT) R20291 has a daptomycin MIC of ~4 µg/mL. Treatment of R20291 with 0.5 or 1 µg/mL levels of daptomycin slowed the growth rate, while 0.25 µg/mL had no noticeable effect (Fig. 1A). At 5 hours post inoculation, we evaluated changes to cell morphology by phase-contrast microscopy as well as fluorescence microscopy to detect membranes (FM4-64) and DNA (Hoechst 33342). Daptomycin-treated cells showed a distinct chaining phenotype (Fig. 1C and E through G). The length of cells was measured regardless of their division or separation status, i.e., chained cells were treated as one unit. A minimum of 113 units were measured per condition. The average length increased by twofold to threefold, and some chains were more than 10-fold longer than any observed in the untreated control culture. We also found a striking number of minicells (around 3%–4%), which were devoid of DNA as evidenced by the lack of staining with Hoechst 33342. Finally, some cells became phase bright, and there were signs of lysis (Fig. 1C through G). Cell chaining, formation of minicells, and lysis are consistent with prior evidence that daptomycin interferes with cell division and PG synthesis (39).

Fig 1.

Effect of sub-MIC daptomycin on C. difficile cell morphology. (A) The C. difficile strain R20291 was subcultured 1:100 into TY with varied daptomycin concentrations, grown at 37°C, and imaged by microscopy after 5 hours. (B) The length of bacterial units increased with daptomycin concentrations from 6.2 ± 1.3 µm to 12.4 ± 10.9 at 0.5 µg/mL daptomycin and 16 ± 13 µm at 1 µg/mL daptomycin. A minimum of 113 units were measured. (C–F) The increased unit length was largely due to chaining. Other effects included the formation of minicells (red carets and insets), phase bright cells (black caret), and lysis (white caret). There were 3%–4% minicells at all concentrations of daptomycin (139–191 cells or chains counted). (F) Staining with Hoechst 33342 showed that minicells were generally devoid of DNA. (G) Stain FM4-64 was used to show membranes. Black size bar: 10 µm for C and D; white size bar 5 µm for E–G. Data shown are representative of two experiments.

Identification of daptomycin resistant C. difficile mutants

We were interested in identifying genes involved in daptomycin resistance in C. difficile. Using wild-type R20291, which has a daptomycin MIC of ~4 µg/mL, we subcultured eight independent cultures of R20291 into tryptone-yeast (TY) containing 4 µg/mL of daptomycin. We then subcultured these into TY, grew them overnight, and subcultured again into TY with increased daptomycin. We passaged these cells three to four times until they grew at >16 µg/mL of daptomycin. We then isolated individual colonies from each culture and determined their sensitivity to daptomycin (Table 1).

TABLE 1.

Unique SNPs in daptomycin-resistant mutants

Strain	Locus tag	Gene	Ref aa change	R20291 NC_013316 position	Reference nucleotide position change	Score	Variant coverage	Variant fraction	Daptomycin MIC (μg/mL)a	BioSample accession number	SRA Accession Number
R20291	WT								4
CDE3577	CDR20291_3396	ftsH2	Glu263Stop	4039248	787G>T	3123.95	105	1	27	SAMN39244212	SRR27410865
CDE3578	CDR20291_2456	draS	Ala689Val	2880703	2066C>T	2763.65	89	1	53	SAMN39244213	SRR27410864
CDE3579	CDR20291_1187	esrR	Arg49Ile	1418752	146G>T	2800.39	91	1	32	SAMN39244214	SRR27410863
CDE3581	CDR20291_3396	ftsH2	Glu61_Gly62 frameshift	4039853	181_182insT	3468.77	84	0.98	32	SAMN39244215	SRR27410862
CDE3832	CDR20291_2041	cdsA	Ile217Lys	2392913	650T>A	2910.27	97	1	8	SAMN39244217	SRR27410860
CDE3831	CDR20291_2041	cdsA	Asp100Tyr	2393265	298G>T	3298.17	111	1	16	SAMN39244216	SRR27410861
CDE3833	CDR20291_2041	cdsA	Pro13Leu	2393525	38C>T	3014.67	104	1	16	SAMN39244218	SRR27410859
CDE3834	CDR20291_2456	draS	Glu120Stop	2882411	358G>T	3216.12	108	1	>32	SAMN39244219	SRR27410858

^a Average MIC from three independent experiments.

We isolated genomic DNA and performed whole-genome sequencing using Illumina 150-bp paired-end reads to identify potential mutations associated with increased daptomycin resistance. We identified unique SNPs in all eight suppressors which affected four different genes: cdr20291_2041 (cdsA), cdr20291_3396 (ftsH2), cdr20291_1187 (esrR), and cdr20291_2456 (draS) (40–43). Three of the eight independent SNPs were in a homolog of cdsA (P13L, D100Y, and I217K), two were in ftsH2 (E263* and E61-frameshift), one was in esrR (R49I), and two were in draS (E120* and A689V) (Table 1). This work was performed using C. difficile R20291; however, the genes identified here are present in most if not all strains of C. difficile. We previously demonstrated that activation of the DraRS two-component regulatory system increased daptomycin resistance in C. difficile (34). Interestingly, mutations in ftsH2 and esrR had been found in a selection for surotomycin resistance; however, their importance for resistance was never confirmed (38). Here, we characterize the role of ftsH2, esrR, and cdsA mutations in resistance to daptomycin.

Partial loss-of-function mutations in cdsA increase daptomycin resistance

 We isolated three independent mutants with SNPs in cdr20291_2041, which is homologous to cdsA from B. subtilis (Table 1). CdsA is a putative cytidine diphosphate-diacylglycerol (CDP-diacylglycerol) synthase, and in other organisms mutations in cdsA increase daptomycin resistance (44). CdsA is an enzyme required for catalyzing the addition of cytidine triphosphate to diacylglycerol to produce CDP-diacylglycerol, an essential precursor of all phospholipid biosynthesis (Fig. 2C) (45).

Fig 2.

Decreased phospholipid synthesis increases daptomycin resistance. (A) Spontaneous cdsA mutants increase daptomycin MIC. (B) Complementation of cdsA^D100Y with P_xyl-cdsA restores daptomycin MIC to wild-type levels. Data were analyzed using one-way analysis of variance using Dunnett’s multiple-comparison test. ****P < 0.0001; ***P < 0.001; **P < 0.01; if not shown, P > 0.05. (C) The predicted phospholipid synthesis pathway in C. difficile. (D) Partial depletion of CdsA or PgsA increases daptomycin resistance. Cells of strain R20291 harboring CRISPRi plasmids were grown overnight in TY Thi₁₀. Samples were serially diluted, and 5 µL of each dilution was spotted onto TY Thi₁₀ + 0.1% xylose with or without 0.1 µg/mL of daptomycin. Plates were photographed after incubation overnight. (E) Phosphatidylglycerol and cardiolipin levels of cdsA^D100Y normalized to WT. Data shown are representative of at least three experiments.

We found that each of the cdsA mutants shows approximately twofold to fourfold increased daptomycin resistance compared to wild type (Fig. 2A; Table 1). Wild-type levels of daptomycin resistance were restored in the cdsA^D100Y mutant by a plasmid with a xylose-inducible copy of the wild-type gene (P_xyl-cdsA; Fig. 2B), demonstrating directly that cdsA^D100Y is responsible for daptomycin resistance and implying that it is a loss-of-function mutation. It is likely that the same conclusions can be drawn for the other cdsA mutations.

A transposon library screen by Dembek et al. identified cdsA as essential for in vitro growth in C. difficile (46). Thus, the mutations that we isolated in cdsA are likely to retain partial function. To test this notion, we used Clustered Regularly Interspaced Short Palindromic Repeats interference (CRISPRi) to knock down cdsA expression (47). We designed two independent sgRNAs to cdsA and cloned them into pIA33, which has a xylose-inducible dCas9 and allows for different degrees of gene silencing depending on the amount of xylose added to the growth medium (47). In the absence of xylose, strains containing sgRNAs targeting cdsA were able to grow similarly to the negative control strain with an sgRNA that does not target anywhere in the genome (Fig. S1A). However, when strains with CRISPRi sgRNAs targeting cdsA were plated on media with 1% xylose, there was an approximate 5-log loss of viability in a spot titer assay (Fig. S1B). This confirms that cdsA is essential for C. difficile. We scraped cells from the 0 dilution spot on the 1% xylose plate and examined them by microscopy using phase contrast and FM4-64 membrane stain (Fig. S1C through L). We found that cells depleted for CdsA were often misshapen, elongated, bent, or curvy (Fig. S1E through H) when compared to the negative control (Fig. S1C and D). To determine if decreasing CdsA levels would increase daptomycin resistance, we performed a spot titer assay on plates with 0.1 µg/mL of daptomycin and only 0.01% xylose to partially knockdown cdsA expression. Indeed, partial knockdown of cdsA expression resulted in a 1,000-fold increase in viability when compared to the negative sgRNA (Fig. 2D). This further confirms that decreased CdsA activity increases daptomycin resistance.

Decreased phosphatidylglycerol synthesis increases daptomycin resistance

The C. difficile genome does not encode homologs of genes known to be involved in the biosynthesis of phosphatidylserine and phosphatidylethanolamine, and these phospholipids are not detected in C. difficile (27). Thus, we predicted the cdsA mutations likely impact synthesis of phosphatidylglycerol and cardiolipin, which compose approximately 30% and 15% of the membrane lipids, respectively (27). We performed lipidomic analysis on the cdsA^D100Y mutant and found that both phosphatidylglycerol and cardiolipin were decreased compared to wild type (Fig. 2E). Thus, in C. difficile, partial loss-of-function of cdsA increases daptomycin resistance by decreasing the abundance of phosphatidylglycerol, as reported previously in other organisms (11–13).

This finding led us to ask whether changes to other genes involved in phosphatidylglycerol synthesis also altered daptomycin resistance. In C. difficile, pgsA, encoding for phosphatidylglycerol synthase, is an essential gene (46). We, therefore, used CRISPRi to study its role by partial knockdown of expression. First, we confirmed that pgsA is essential, as we detect a loss of viability when strains containing sgRNAs targeting pgsA were plated on 1% xylose to induce CRISPRi (Fig. S1B). Microscopy revealed that depletion of PgsA results in cells that are generally longer and more curved compared to the negative controls (Fig. S1B). Thus, loss of either CdsA or PgsA results in similar viability and cell morphology defects.

We then asked if partial knockdown of pgsA expression increases daptomycin resistance. When cells were plated on media with 0.01% xylose, knockdown of pgsA was sufficiently subtle that growth was not substantially affected compared to the negative control (Fig. 2D). When cells were plated on media with 0.01% xylose and low levels of daptomycin (0.1 µg/mL), we found that strains with sgRNAs targeting pgsA had improved viability compared to the negative control (Fig. 2D). This is consistent with a decrease in phosphatidylglycerol increasing daptomycin resistance.

Loss of cardiolipin decreases daptomycin resistance

Cardiolipin synthase (Cls) converts two molecules of phosphatidylglycerol to cardiolipin (Fig. 2C). Interestingly, a mutation in a putative Cls encoding gene cdr20291_3226 (clsB) was associated with increased surotomycin resistance in C. difficile. However, it was not confirmed if this mutation was responsible for the altered daptomycin resistance (38). Based on these observations, we sought to determine if cells lacking cardiolipin synthases are more sensitive to daptomycin due to increased levels of phosphatidylglycerol.

We used the known cardiolipin synthases ClsA and YwjE from B. subtilis to perform a BLAST search for homologs of Cls in C. difficile R20291. We identified two putative cls genes in C. difficile, cdr20291_0193 (clsA), and cdr20291_3226 (clsB) (Fig. S2), neither of which were identified as individually essential (46). Since cls genes are often redundant, we used CRISPR to construct ΔclsA, ΔclsB, and ΔclsA ΔclsB mutants. We isolated lipids from WT, ΔclsA, ΔclsB, and ΔclsA ΔclsB mutants as previously described (48). We then separated the lipids using thin-layer chromatography (TLC) and stained for phospholipids using molybdenum blue (48). We found that the ΔclsA ΔclsB double mutant lacked any detectable cardiolipin and showed a corresponding, slight increase in phosphatidylglycerol (Fig. 3A). We then performed lipidomic analysis on the ΔclsA, ΔclsB, and ΔclsA ΔclsB mutants and found that cardiolipin was absent in the ΔclsA ΔclsB double mutant and that phosphatidylglycerol levels slightly increased. Loss of clsA or clsB alone did not alter the cardiolipin levels (Fig. 3E and F). Cardiolipin production could be restored in the ΔclsA ΔclsB strain by expression of either clsA or clsB alone (Fig. 3B and E; Fig. S7). This suggests that both ClsA and ClsB are responsible for cardiolipin synthesis in C. difficile. We found that loss of ClsA and ClsB had no major effect on cell size or shape compared to wild type (Fig. S3).

Fig 3.

Mutants lacking cardiolipin synthesis are more sensitive to daptomycin. (A) Wild type, ΔclsA ΔclsB, ΔclsA, and ΔclsB lipid extracts and phosphatidylglycerol (PG) and cardiolipin (CL) lipid standards were separated by TLC and visualized with molybdenum blue. (B) Extracts from wild type with an empty vector pAP114 (EV), ΔclsA ΔclsB EV, ΔclsA ΔclsB P_xyl-clsA, and ΔclsA ΔclsB P_xyl-clsB were separated by TLC and visualized with molybdenum blue. (C) ΔclsA ΔclsB decreases daptomycin MIC twofold compared to wild type and ΔclsA and ΔclsB do not change daptomycin MIC. (D) Expressing P_xyl-clsA or P_xyl-clsB in ΔclsA ΔclsB restores daptomycin MIC to levels of wild type or higher (EV: empty vector). (E) Cardiolipin (CL) and (F) phosphatidylglycerol (PG) levels of ΔclsA ΔclsB, ΔclsA, and ΔclsB normalized to WT. Data were analyzed using one-way analysis of variance using Dunnett’s multiple-comparison test. ****P < 0.0001; ***P < 0.001; **P < 0.01; if not shown, P > 0.05. Data shown are representative of three experiments.

Next, we measured the MIC of daptomycin against the cls mutants. We found that deletion of either clsA or clsB alone had no effect on daptomycin resistance (Fig. 3C). However, the ΔclsA ΔclsB double mutant was twofold more sensitive to daptomycin (Fig. 3C). Expression of either clsA or clsB in the ΔclsA ΔclsB mutant increased daptomycin resistance to wild-type levels or above (Fig. 3D). This suggests that the decrease in daptomycin resistance is due to loss of cardiolipin, presumably because this leads to an increase in phosphatidylglycerol levels.

Previous work identified an SNP in clsB (ClsB^D79N) that was associated with increased surotomycin resistance (38). To evaluate if this mutation increased daptomycin resistance, we constructed a vector harboring P_xyl-clsB^D79N, introduced it into both WT and the ΔclsA ΔclsB strain, and determined the daptomycin MIC. We found that production of ClsB^D79N increased daptomycin resistance 4- and 16-fold in WT and the ΔclsA ΔclsB mutant, respectively (Fig. 3D). We then compared the effect of the ClsB^D79N mutant on phospholipid biosynthesis by performing TLC. However, we did not detect a dramatic shift in the phospholipid profile of cells producing ClsB^D79N (Fig. 3B). Our data confirm that ClsB^D79N is indeed responsible for the increased daptomycin MIC reported elsewhere (38). However, the mechanism of resistance does not appear to be due to an increase in phosphatidylglycerol levels and remains to be elucidated.

Loss-of-function mutations in esrR increase daptomycin resistance

We identified one strain with an SNP in cdr20291_1187, which we named esrR for envelope stress response regulator. The SNP that confers daptomycin resistance changes Arg 49 to Ile and confers a fourfold to eightfold increase in daptomycin resistance (Fig. 4B). To confirm that the SNP esrR^R49I was responsible for increased daptomycin resistance, we expressed wild-type esrR from a xylose-inducible promoter, which restored daptomycin sensitivity to parental levels (Fig. 4B). esrR is the first gene in a predicted three-gene operon which includes esrAB (Fig. 4A). EsrR belongs to the PadR family of transcriptional repressors (Pfam accession PF03551), most of which control genes involved in stress responses (49–52). Thus, we predicted that the EsrR^R49I mutation inactivates the repressor, which results in increased expression of the esrRAB operon. Indeed, reverse transcription-quantitative PCR (RT-qPCR) revealed increased levels of esrA and esrB mRNA compared to wild type, and repression was largely restored by a P_xyl-esrA⁺ plasmid (Fig. S4).

Fig 4.

Loss-of-function mutations in esrR increase daptomycin resistance. (A) Diagram of the esrRAB operon including esrR (cdr20291_1187), esrA (cdr20291_1188), and esrB (cdr20291_1189). (B) Spontaneous esrR mutant, esrR^R49I, increases daptomycin resistance compared to wild type. Complementation of esrR^R49I with P_xyl-esrR restored daptomycin MIC to wild-type levels (EV: empty vector, pAP114). Expression of P_xyl-esrAB in wild-type increases daptomycin MIC fourfold compared to wild type. (C) CRISPRi knockdowns using two independent sgRNAs to target esrRAB in esrR^R49I decreased the daptomycin MIC back to wild-type levels compared to esrR^R49I with a negative sgRNA control (Neg). Data were analyzed using one-way analysis of variance using Dunnett’s multiple-comparison test. ****P < 0.0001; ***P < 0.001; **P < 0.01; if not shown, P > 0.05. Data shown are from at least three experiments.

Since EsrR is an auto-repressor, we hypothesized that increased expression of esrAB was responsible for increased daptomycin resistance. CRISPRi is usually polar on downstream genes (47, 53). We, therefore, used CRISPRi to knock down expression of the whole esrRAB operon using two independent sgRNAs targeting esrR. We found that CRISPRi knockdown of esrRAB in the esrR^R49I mutant decreased daptomycin resistance to near wild-type levels (Fig. 4C). In contrast, knockdown of esrRAB in a wild-type background did not have a significant impact on daptomycin resistance (Fig. 4C). Conversely, expressing esrAB from a xylose-inducible promoter resulted in a fourfold increase in daptomycin resistance (Fig. 4A). In total, these findings are consistent with increased expression of esrAB being responsible for increased daptomycin resistance in the esrR^R49I mutant.

Expression of esrRAB is induced by cell envelope stress

We previously found that expression of esrRAB was induced in response to daptomycin (34). To confirm and extend this observation, we constructed a P_esr-rfp reporter plasmid and introduced it into wild-type R20291. We then grew cells containing the reporter to mid-log and incubated with sub-MIC levels of daptomycin, vancomycin, bacitracin, and ampicillin for 2 hours. The cells were fixed and exposed to oxygen overnight to allow chromophore maturation (54, 55). We then analyzed expression at the single-cell level using flow cytometry (55, 56). We found that incubation with daptomycin, vancomycin, and bacitracin at the highest concentrations led to approximately fivefold induction of P_esr-rfp, while ampicillin led to approximately twofold induction (Fig. S6). Taken together, these data suggest that esrRAB is induced by cell envelope stress and can provide resistance to daptomycin.

Loss of FtsH2 increases daptomycin resistance

Two independent mutants had SNPs in the gene cdr20291_3396 (ftsH2), E263* and E61-frameshift, both of which increased daptomycin resistance approximately eightfold compared to wild type (Fig. S5A). FtsH is a membrane-bound AAA+ protease that utilizes ATP to unfold and translocate proteins into a catalytic chamber for degradation (57). In E. coli and B. subtilis, FtsH has been associated with protein quality control and stress response (58, 59). C. difficile encodes two FtsH homologs, FtsH1 (cdr20291_1722) and FtsH2, while most other organisms only encode a single FtsH.

Because our ftsH2 SNPs involved a premature stop codon and an early frameshift, we presumed that loss of FtsH2 function conferred elevated daptomycin resistance. This was confirmed by introducing a P_xyl-ftsH2 plasmid into the ftsH2 E61-frameshift mutant. Expressing wild-type ftsH2 in the mutant strain restored wild-type levels of daptomycin resistance (Fig. S5B). We also used CRISPRi to knock down the expression of ftsH2 in wild type and observed that daptomycin resistance increased approximately fourfold to eightfold compared to a negative control sgRNA (Fig. S5C). To further confirm the role of FtsH2 in daptomycin resistance, we used CRISPR to create a deletion of ftsH2. The ΔftsH2 mutant also showed an increase in daptomycin resistance (Fig. 5A), which could be complemented by expressing wild-type ftsH2 under a xylose-inducible promoter and thereby restore daptomycin sensitivity to wild-type levels (Fig. 5A). However, a catalytic mutant, P_xyl-ftsH2^{H433A E434A}, was unable to complement the ΔftsH2 mutant (Fig. 5A). Taken together, these data clearly demonstrate that loss of FtsH2 increases daptomycin resistance and suggest that the proteolytic activity of FtsH2 is required for daptomycin susceptibility.

Fig 5.

Loss of the membrane-bound AAA+ protease FtsH2 increases daptomycin resistance. (A) Deletion of ftsH2 increases the daptomycin MIC. Deletion of ftsH2 can be complemented with P_xyl-ftsH2 to restore daptomycin MIC to wild-type levels but cannot be complemented by P_xyl-ftsH1 or P_xyl-ftsH2^{H433A E434A}. (B) CRISPRi knockdowns using two independent sgRNAs to target ftsH1 in wild type or an ΔftsH2 mutant did not further increase the daptomycin MIC. Data were analyzed using one-way analysis of variance using Dunnett’s multiple-comparison test. ****P < 0.0001; ***P < 0.001; **P < 0.01; if not shown, P > 0.05. Data shown are from at least three experiments.

FtsH1 does not have a role in daptomycin resistance

As noted, C. difficile encodes two FtsH homologs, FtsH1 and FtsH2, while most other organisms encode a single FtsH homolog. It is not known if the two homologs have redundant or distinct functions in C. difficile. To determine if loss of FtsH1 also increases daptomycin resistance, we used CRISPRi to knock down ftsH1 expression in wild type, but there was no effect on daptomycin resistance (Fig. 5B). To determine if the presence of FtsH2 could mask a role for FtsH1, we used CRISPRi to knock down the expression of ftsH1 in an ΔftsH2 mutant. We found that loss of both FtsH1 and FtsH2 has the same daptomycin resistance phenotype as loss of FtsH2 alone (Fig. 5B). Finally, we introduced a P_xyl-ftsH1 plasmid into our ΔftsH2 mutant and tested daptomycin resistance, with the result that ftsH1 is unable to restore daptomycin sensitivity to wild-type levels (Fig. 5A). Taken together, these data show that FtsH1 does not impact daptomycin resistance and that FtsH1 and FtsH2 likely have distinct functions and targets in the cell.

DISCUSSION

New treatment options for C. difficile infections are needed to overcome issues with relapse due to inadvertent killing of host commensal bacteria by frontline antibiotics like vancomycin and metronidazole (60–62). A better understanding of C. difficile cell envelope biogenesis might provide new targets and improved approaches to treating infections. Among the structures that make up C. difficile’s envelope, the cytoplasmic membrane is understudied compared to the PG wall and S-layer (17–26). Here, we investigated the response of C. difficile to daptomycin, which according to current models kills bacteria by inhibiting PG synthesis. Specifically, daptomycin forms a tripartite complex with phosphatidylglycerol and the undecaprenyl lipid that helps ferry PG building blocks across the cytoplasmic membrane (7). Because daptomycin’s mechanism of action involves phosphatidylglycerol, studies of daptomycin are expected to shed light on the cytoplasmic membrane.

Exposure of growing C. difficile to sub-MIC levels of daptomycin inhibited growth and induced multiple morphological defects, including cell chaining, minicell formation, and lysis. These defects, especially lysis, are broadly consistent with inhibition of PG synthesis through sequestration of the undecaprenyl carrier lipid. However, it is not obvious how interfering with the undecaprenyl cycle would promote minicell formation, which is a consequence of improper placement of the division septum. It is possible that insertion of daptomycin’s hydrophobic tail into the membrane and/or binding of daptomycin to phosphatidylglycerol perturbs membrane curvature or composition in ways that promote assembly of the division apparatus at inappropriate sites.

In a separate line of investigation, we isolated and confirmed eight independent daptomycin resistance mutants which turned out to harbor unique SNPs in four different genes cdr20291_2041 (cdsA), cdr20291_3396 (ftsH2), cdr20291_1187 (esrR), and cdr20291_2456 (draS). Subsequent characterization of these mutants revealed that resistance mechanisms fell into two categories: (i) altering the level of phosphatidylglycerol and (ii) altering transcriptional or proteolytic regulatory systems.

Mutations altering synthesis of phosphatidylglycerol impact daptomycin resistance

In other organisms, loss of phosphatidylglycerol increases daptomycin resistance because daptomycin requires phosphatidylglycerol to target the cell (7, 11–13). Consistent with this, three of our daptomycin-resistant mutants had partial loss-of-function mutations in cdsA, which encodes a CDP-diacylglycerol synthase. We extended this result by targeted CRISPRi knockdown of a second phosphatidylglycerol synthesis gene, pgsA, which encodes a phosphatidylglycerol phosphate synthase. As predicted, decreasing expression of either cdsA or pgsA with CRISPRi enhanced daptomycin resistance. Another prediction of the model is that increasing phosphatidylglycerol levels should decrease daptomycin resistance. Indeed, we found that preventing conversion of phosphatidylglycerol to cardiolipin by deleting two cardiolipin synthase genes, clsA and clsB, rendered C. difficile more sensitive to killing by daptomycin. Finally, both cdsA and pgsA were reported to be essential by Tn-seq in C. difficile (46). We confirmed essentiality by CRISPRi knockdown and further showed that depletion of these phospholipid synthesis enzymes results in elongated and misshapen cells (Fig. S1). These shape defects can be rationalized by the fact that the PG synthesis machinery is embedded in the cytoplasmic membrane. It will be interesting to examine whether manipulating phosphatidylglycerol levels causes mislocalization of penicillin-binding proteins or other PG synthesis proteins (63, 64).

Mutations altering signal transduction impact daptomycin resistance

A common mechanism for increased daptomycin resistance in other organisms is the activation of a signal transduction system like LiaRS in B. subtilis that is induced by and mediates resistance to daptomycin (11, 65–67). LiaRS homologs in E. faecalis and LiaRS orthologs VraRS in S. aureus also confer daptomycin resistance (68–73). Our selection for daptomycin resistance identified mutations in two different signaling systems. One was the recently described two-component regulatory system, DraRS (34). Although DraRS does not show any homology to LiaRS sensing domains, it nevertheless responds to lipid II-inhibiting antibiotics like daptomycin (34). Constitutive activation of DraRS increases daptomycin resistance, in part by increasing expression of hexSDF, which are required for synthesis of the glycolipid HNHDRG (34). To date, HNHDRG has only been found in C. difficile membranes and is normally present at about 16%. Increased levels of HNHDRG presumably decrease the binding of daptomycin to the membrane.

One of our daptomycin-resistant mutants harbored an SNP in a PadR family repressor that we named esrR (Fig. 4B). The SNP changes Arg 49 to Ile. Mutations in esrR were previously associated with increased surotomycin resistance (38). Our data suggest that EsrR is an auto-repressor and that esrR^R49I is a loss-of-function mutation that derepresses expression of esrRAB (Fig. S4). We found that increased expression of esrAB in wild-type C. difficile increases daptomycin resistance (Fig. 4B). Previous RNA-seq data showed expression of esrRAB was induced by daptomycin (34). We confirmed this observation using a P_esr-rfp reporter and also showed that the reporter responded to a variety of cell envelope stresses (Fig. S6).

EsrA contains a domain of unknown function, DUF1700, which overlaps with a predicted HAAS domain (HTH-associated α-helical signaling domain) that contains a GxP motif predicted to mediate protein-protein interactions. EsrB contains a Toastrack domain predominated by β-sheets (74). Streptococcus pneumoniae encodes PtvRAB which provide increased vancomycin resistance (52). These proteins are not homologous to EsrRAB, but they share the same domain architecture suggesting they could function in a similar manner. PtvR is a PadR repressor; PtvA has HAAS DUF1700 domains, while PtvB contains DUF4097 or Toastrack domain. In S. pneumoniae, expression of ptvRAB is induced by vancomycin in a PtvR-dependent manner (52). In B. subtilis, LiaF and LiaG also contain Toastrack domains and are both regulated by LiaRS, a two-component system induced by daptomycin and required for daptomycin resistance (11, 65–67, 74). LiaF is a negative regulator of LiaRS, and LiaG is predicted to be a membrane-anchored protein (67, 75). The role of LiaG is unknown, but data suggest that LiaG promotes recruitment of LiaIH to the membrane to confer daptomycin resistance by an unknown mechanism (75). Proteins containing Toastrack domains are also found in the LiaFSR system in E. faecalis that is associated with daptomycin resistance (72, 73). It is unknown how EsrAB confers daptomycin resistance in C. difficile, but one possibility is that EsrAB forms a complex to sequester daptomycin, recruiting other proteins to the membrane like the Lia system in B. subtilis (76). Another possibility is that EsrAB facilitates changes to the membrane as has been observed in E. faecalis (72, 77, 78).

Mutations in AAA+ protease, FtsH2, increase daptomycin resistance

We identified two independent strains with mutations in ftsH2 that increased daptomycin resistance, and we confirmed that loss of ftsH2 also increased daptomycin resistance (Fig. S5A; Fig. 5A). A previous study found a C. difficile strain with a mutation in ftsH2 has increased surotomycin resistance (38). FtsH is a membrane-bound AAA+ protease present in many organisms that is associated with protein quality control and stress response (58, 59). Of the two FtsH homologs in C. difficile, only FtsH2 is associated with daptomycin resistance. The proteolytic targets of FtsH2 are unknown, and where FtsH targets have been identified, they vary between organisms (58, 59). We hypothesize that in the absence of ftsH2, the accumulation of its proteolytic targets increases daptomycin resistance. It appears this is not specific to C. difficile as a recent study found loss of FtsH in E. faecalis increased daptomycin resistance and decreased metabolic activity (79).

MATERIALS AND METHODS

Bacterial strains, media, and growth conditions

Bacterial strains are listed in Table 2. The C. difficile strains used in this study are derivatives of R20291. C. difficile strains were grown on TY medium consisting of 3% tryptone, 2% yeast extract, and 2% agar (for solid medium). TY was supplemented as needed with thiamphenicol at 10 µg/mL (Thi₁₀). Conjugations were performed on solid brain-heart infusion (BHI) media (3.65% BHI, 2% agar) and plated on TY with thiamphenicol at 10 µg/mL, kanamycin at 50 µg/mL, and cefoxitin at 8 µg/mL. C. difficile strains were maintained at 37°C in an anaerobic chamber (Coy Laboratory Products) in an atmosphere of approximately 2%–2.5% H₂, 5% CO₂, and ~93% N₂. E. coli strains were grown in lysogeny broth medium (1% tryptone, 0.5% yeast extract, 0.5% NaCl, and 1.5% agar for solid medium) at 37°C with chloramphenicol at 10 µg/mL and ampicillin at 100 µg/mL as needed. Daptomycin (Fresenius Kabi LLC) was ordered from the University of Iowa hospital pharmacy. Stocks were prepared at 50 mg/mL in water, filter sterilized, and stored at −20°C.

TABLE 2.

Strains

Species and strain	Genotype and/or description	Source or reference
E. coli strains
OmniMAX-2 T1R	F′ [proAB + lacIq lacZΔM15 Tn10(Tetr) Δ(ccdAB)] mcrA Δ(mrr-hsdRMS-mcrBC) φ80(lacZ)ΔM15 Δ(lacZYA-argF)U169 endA1 recA1 supE44 thi-1 gyrA96 relA1 tonA panD	Invitrogen
HB101/pRK24	F− mcrB mrr hsdS20(rB− mB−) recA13 leuB6 ara-14 proA2 lacY1 galK2 xyl-5 mtl-1 rpsL20	(80)
MG1655	Wild Type
B. subtilis strains
BS49	Tn916 donor strain, Tetr	(81)
C. difficile strains
R20291	Wild-type strain from UK outbreak (ribotype 027)
CDE3577	R20291 ftsH2E263*
CDE3578	R20291 draSA689V
CDE3579	R20291 esrRR49I
CDE3581	R20291 ftsH2E61-Frameshift*
CDE3832	R20291 cdsAI217K
CDE3831	R20291 cdsAD100Y
CDE3833	R20291 cdsAP13L
CDE3834	R20291 draSE120*
CM106	R20291 cdsAD100Y pAP114
CM105	R20291 cdsAD100Y pCE936
AP447	R20291 pIA34	(47)
CDE3925	R20291 pCI5571
CDE3926	R20291 pCI5572
CM113	R20291 pHD09 pgsA KD g1
CM202	R20291 pHD10 pgsA KD g2
CM180	R20291 ΔclsA ΔclsB
CDE4532	R20291 ΔclsA
CM186	R20291 ΔclsB
AP441	R20291 pAP114	(47)
CM193	R20291 ΔclsA ΔclsB pAP114
CM194	R20291 ΔclsA ΔclsB pCE1135
CDE4650	R20291 ΔclsA ΔclsB pCE1136
CDE4725	R20291 ΔclsA ΔclsB pCE1156
AP419	R20291 esrRR49I pAP114
AP434	R20291 pCE861
AP433	R20291 esrRR49I pCE861
AP431	R20291 pCE867
AP448	R20291 esrRR49I pIA34
AP423	R20291 pCE858
AP437	R20291 esrRR49I pCE858
AP425	R20291 pCE859
AP424	R20291 esrRR49I pCE859
CDE4634	R20291 pCE1128 PesrR-rfp fusion
AP422	R20291 ftsH2E61-Frameshift* pAP114
AP421	R20291 ftsH2E61-Frameshift* pCE862
BZ201	R20291 ΔftsH2 pAP114
BZ202	R20291 ΔftsH2 pCE913
BZ203	R20291 ΔftsH2 pCE862
BZ605	R20291 ΔftsH2 pCE1147
BZ118	R20291 pBZ102
BZ119	R20291 pBZ103
BZ204	R20291 ΔftsH2 pIA34
BZ205	R20291 ΔftsH2 pBZ102
BZ206	R20291 ΔftsH2 pBZ103
AP432	R20291 pCE865
AP430	R20291 pCE866

Microscopy

Microscopy procedures were adapted from previous reports (54, 55). Cells were immobilized using thin agarose pads (1%). Phase-contrast micrographs were recorded on an Olympus BX60 microscope equipped with a 100× UPlanApo objective (numerical aperture, 1.35). Micrographs were captured with a Hamamatsu Orca Flash 4.0 V2 + complementary metal oxide semiconductor camera. Excitation light was generated with an X-Cite XYLIS LED light source. Membranes were stained with the lipophilic dye FM4-64 (Life Technologies) at 10 µg/mL. Cells were imaged immediately without washing. Red fluorescence was detected with the Chroma 49008 filter set (538 to 582 nm excitation filter, 587 nm dichroic mirror, and a 590 to 667 nm emission filter). DNA staining was achieved with 45 µg/mL Hoechst 33342 (Invitrogen). Cells were incubated at room temperature for 15 min and imaged without washing. Blue fluorescence was detected with the Olympus filter set U-MWU (330–385 nm excitation filter, 400 nm dichroic mirror, and a 420 nm barrier emission filter). The Olympus cellSens software was used to measure cell length.

Plasmid and bacterial strain construction

All plasmids are listed in Table 3. Plasmids were constructed using Gibson Assembly (New England Biolabs, Ipswich, MA). Regions of the plasmids constructed using PCR were verified by DNA sequencing. Oligonucleotide primers used in this work were synthesized by Integrated DNA Technologies (Coralville, IA) and are listed in Table S1. All plasmids were propagated using OmniMax-2 T1R as a cloning host and conjugated into C. difficile using HB101/pRK24 (80). CRISPR-Cas9 deletion plasmids were passaged through E. coli strain MG1655 before transformation into B. subtilis strain BS49 (81). CRISPR-Cas9 plasmids were built on the backbone of pJK02 (82) with some modification (20).

TABLE 3.

Plasmids

Plasmid	Relevant features	Reference	Primers used
pAP114	Pxyl::mCherryOpt cat	(47)
pIA33	Pxyl::dCas9-opt Pgdh::sgRNA-rfp cat	(47)
pIA34	Pxyl::dCas9-opt Pgdh::sgRNA-neg cat	(47)
pCE1135	Pxyl::clsA cat		6511–6512
pCE1136	Pxyl::clsB cat		6513–6514
pCE861	Pxyl::padR cat		5222–5223
pCE862	Pxyl::ftsH2 cat		5220–5221
pCE867	Pxyl::cdr20291_1188–1189 cat		5224–5225
pCE913	Pxyl::ftsH1 cat		5551–5552
pCE936	Pxyl::cdsA cat		5605–5606
pCE858	Pxyl::dCas9-opt Pgdh::sgRNA-padR-1 cat		5226
pCE859	Pxyl::dCas9-opt Pgdh::sgRNA-padR-2 cat		5227
pCE865	Pxyl::dCas9-opt Pgdh::sgRNA-ftsH2-1 cat		5218
pCE866	Pxyl::dCas9-opt Pgdh::sgRNA-ftsH2-2 cat		5219
pCI5571	Pxyl::dCas9-opt Pgdh::sgRNA-cdsA-1 cat		5571
pCI5572	Pxyl::dCas9-opt Pgdh::sgRNA-cdsA-2 cat		5572
pBZ102	Pxyl::dCas9-opt Pgdh::sgRNA-ftsH1-1 cat		5537
pBZ103	Pxyl::dCas9-opt Pgdh::sgRNA-ftsH1-2 cat		5538
pHD09	Pxyl::dCas9-opt Pgdh::sgRNA-pgsA-1 cat		4923
pHD10	Pxyl::dCas9-opt Pgdh::sgRNA-pgsA-2 cat		4924
pCE1147	Pxyl::ftsH2H433A E434A cat		6672–6673
pCE1156	Pxyl::clsBD79N cat		6754–6755
pCM111	Pxyl::Cas9-opt Pgdh::sgRNA-clsA cat		6151–6154, 5858
pCE1116	Pxyl::Cas9-opt Pgdh::sgRNA-clsB cat		6205–6208, 6217
pBZ108	Pxyl::Cas9-opt Pgdh::sgRNA-ftsH2-3 cat		5290–5293, 5546
pDSW1728	Ptet-mCherry cat	(54)
pCE1128	PesrR-rfp cat		6569–6570

For xylose-inducible overexpression constructs, genes of interest were amplified using PCR; the oligonucleotides are listed in Table S3 in the supplemental material. PCR amplicons were inserted into the plasmid pAP114 at the SacI and BamHI sites, as described previously (47). For CRISPRi constructs, two guides were created for each targeted gene of interest and amplified using PCR. The PCR amplicons were then inserted into the pIA33 backbone at MscI and NotI sites as previously described (47). Oligonucleotide sequences for the guides are listed in Table S1 in the supplemental material.

Whole genome sequencing and SNP analysis

Overnight cultures (5 mL) of the parent strain and daptomycin-resistant mutants were used for DNA extraction (NEB Monarch Genomic DNA Purification Kit). DNA quality was determined by gel electrophoresis. DNA sequencing was performed by SeqCenter (Pittsburgh, PA) using 150-bp paired end reads on Illumina NextSeq 2000. SNP analysis was performed using PATRIC or BV-BRC (42, 43). The sequences were aligned using BWA-mem (41), and the SNPs were identified using FreeBayes (40). We obtained a median base coverage of >120 for each strain. We focused on SNPs that had high SNP score, high variant coverage, and variant fraction (Table 1).

Antibiotic MIC determination

Overnight cultures of C. difficile were subcultured, grown to late log phase (OD₆₀₀ of 1.0), and then diluted into TY to 10⁶ CFU/mL. A series of antibiotic concentrations was prepared in a 96-well plate in 50-µL TY broth. Wells were inoculated with 50 µL of the dilute late-log-phase culture (0.5 × 10⁵ CFU/well) and grown at 37°C for 16 hours. For strains containing vectors with xylose-inducible elements, 1% xylose was included in the overnight culture, the subculture, and the media in the MIC plate. After the incubation period, the 96-well plates were spun down for 5 min at 5,000 × g, and the MIC was determined based on the presence of cell pellets. Data are reported as the average from three independent experiments.

Lipid extraction

Lipid extractions were performed as previously described (48). Overnight cultures of C. difficile were subcultured in 500 mL of TY and grown to log phase (OD₆₀₀ 0.6–0.7), and pellets were harvested at 3,000 × g. Cell pellets were washed in cold 20 mM MOPS pH 7.2, resuspended in 1 mL water, and transferred to a glass centrifuge tube, and 3.75 mL of chloroform-methanol (1:2, vol/vol) was added. The mixture was incubated at room temperature for 1–2 hours with the cap off and vortexed periodically. The mixture was then spun for 10 min at 3,000 × g, and the supernatant decanted into a clean, glass centrifuge tube. The remaining pellet was resuspended in 4.75-mL methanol-chloroform-water (2:1:0.8, vol/vol/vol). The mixture was then spun for 10 min at 3,000 × g, and the supernatant decanted to combine with the first supernatant. To the supernatants, 2.5 mL each of chloroform and water was added and spun for 10 min at 3,000 × g. The lower, chloroform phase was removed, transferred to a screw cap centrifuge tube, and let sit open overnight to evaporate in a 35°C heating block in a fume hood. The resulting lipid extract was dissolved in chloroform.

TLC and lipid staining

TLC and lipid staining were performed as previously described with modifications (48). Silica gel glass TLC plates were activated by heating for 30 min at 125°C. Once cooled, 0.2 mg of lipid extracts or 10 µg of lipid standards (Avanti Polar Lipids, Alabaster, AL) was spotted ~1 cm from the bottom of the plate. The plate was then placed into a TLC developing tank and run until the solvent reached ~0.5 cm from the top of the plate. The solvent conditions used were chloroform-hexane-methanol-glacial acetic acid (5:3:1:0.5, vol/vol/vol/vol). After drying, the plate was sprayed with molybdenum blue spray reagent (Sigma Aldrich, St. Louis, MO) with a glass atomizer to visualize phospholipids.

P_esr-rfp reporter assays

Cells harboring P_esr-rfp transcriptional fusions were subcultured into TY thi₁₀ and grown at 37°C. Antibiotics were added at mid-log, and cultures were incubated for an additional 2 hours. To fix cells, 500-µL culture was mixed with 120-µL fixation cocktail: 100 µL of a 16% (wt/vol) paraformaldehyde aqueous solution (Alfa Aesar, Ward Hill, MA) and 20 µL of 1 M NaPO₄ buffer (pH 7.4). The sample was vortexed, incubated in the anaerobic chamber at 37°C for 30  min and then on ice 30  min, and removed from the chamber. The fixed cells were washed twice with 1 mL of PBS, resuspended in 50 µL of PBS, and left in the dark for 18 hours to allow for chromophore maturation (54, 55).

Cells were analyzed at the Flow Cytometry Facility at the University of Iowa using the Becton Dickinson LSR II instrument with a 561-nm laser, a 610/20-nm-band-pass filter, and a 600 LP dichroic filter. Data were analyzed using BD FACSDiva software.

Lipidomics

Overnight cultures were grown in TY (supplemented with 1% xylose and Thi₁₀ when needed) in biological triplicate. Overnight cultures were then subcultured to an OD₆₀₀ of 0.05 into 500-mL TY (supplemented with 1% xylose and Thi₁₀ when needed). Subcultures were allowed to grow until an OD₆₀₀ of 0.6 to 0.7 was reached, at which point the cells were harvested and pelleted at 8,000 × g for 20 min. Biological replicates (3) were grown on different days.

Lipid extraction and liquid chromatography with tandem mass spectrometry (LC-MS/MS) were performed by Cayman Chemical Company. After thawing, cells were mixed with 5-mL methanol, transferred to 7-mL Precellys tubes containing 0.1-mm ceramic beads (Bertin Technologies; CK01 lysing kit), and homogenized with three cycles at 8,800 rpm for 30 s, with 60-s pauses between cycles. Then, 800 mL of the homogenized mixtures was transferred to 8-mL screw-cap glass tubes. A methyl tert-butyl ether (MTBE)-based liquid-liquid extraction protocol was used by first adding 1.2-mL methanol containing a mixture of deuterated internal standards covering several major lipid categories (fatty acids, glycerolipids, glycerophospholipids, sphingolipids, and sterols) and then 4-mL MTBE. The mixture was incubated on a tabletop shaker at 500 rpm at room temperature for 1 hour and then stored at 4°C for 60 hours to maximize lipid extraction. After bringing the samples to room temperature, phase separation was induced by adding 1-mL water to each sample. The samples were vortexed and then centrifuged at 2,000 × g for 15 min. The upper organic phase of each sample was carefully removed using a Pasteur pipette and transferred into a clean glass tube. The remaining aqueous phase was reextracted with 2 mL of the upper phase of MTBE/methanol/water at 10:3:2.5 (vol/vol/vol). After vortexing and centrifuging, the organic phase was collected and combined with the initial organic phase. The extracted lipids were dried overnight in a SpeedVac vacuum concentrator.

The dried lipid extracts were reconstituted in 200-mL n-butanol–methanol at 1:1 (vol/vol) and transferred into autosampler vials for analysis by LC-MS/MS. Aliquots of 5 mL were injected into an Ultimate 3000 ultraperformance liquid chromatography system connected to a Q Exactive Plus Orbitrap mass spectrometer (Thermo Scientific). An Accucore C30 2.6-mm, 150-by-2.1-mm HPLC column (Thermo Scientific) was used, using mobile phases A [acetonitrile/water/formic acid 60:40:0.1 (vol/vol/vol), containing 10 mM ammonium formate] and B [acetonitrile/isopropanol/formic acid 10:90:0.1 (vol/vol/vol), containing 10 mM ammonium formate]. Lipids were eluted at a constant flow rate of 300 mL/min using a gradient from 30% to 99% mobile phase B over 30 min. The column temperature was kept at a constant 40°C. Polarity switching was used throughout the gradient to acquire high-resolution MS data (resolution, 75,000) and data-dependent MS/MS data.

Data analysis was performed using Lipostar software (version 2; Molecular Discovery) for detection of features (peaks with unique m/z and retention time), noise and artifact reduction, alignment, normalization, and lipid identification. Automated lipid identification was performed by querying the Lipid Maps Structural Database (LMSD), modified by Cayman to include many additional lipids not present in the LMSD. To allow for comparison between strains, the summed peak areas of lipids with the same head groups were normalized to the wild-type values.

RNA extraction and RT-qPCR

Strains were grown overnight in TY Thi₁₀, subcultured 1:100 into TY Thi₁₀, and grown to an OD₆₀₀ of 0.4–0.6. Cultures were adjusted to 1% xylose and incubated for an additional 1 hour. The cultures were then mixed 1:1 with a solution of 50% ethanol and 50% acetone and immediately frozen at −80°C. RNA was extracted using an RNeasy extraction kit (Qiagen) according to the manufacturer’s instructions. An additional DNase treatment step was added using the TURBO DNA-free kit (Invitrogen). The list of primers used to quantitate cDNA levels of different samples is provided in Table S1 in the supplemental material for esrA (TEQ173-174), esrB (TEQ175-176), and rpoB (TEQ165-166). Experiments were performed in technical triplicate on three biologically independent replicates. Data were normalized to RNA levels of the C. difficile housekeeping gene rpoB (19, 23).

DATA AVAILABILITY

The whole genome sequencing data of daptomycin resistance mutants were deposited under the BioProject ID PRJNA1060763, and the corresponding BioSample and SRA accession numbers are available in Table 1.

SUPPLEMENTAL MATERIAL

The following material is available online at https://doi.org/10.1128/jb.00368-23.

Supplemental material. jb.00368-23-s0001.pdf.

Figures S1 to S7 and Table S1.

ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.