Cyclic Di-GMP and VpsR Induce the Expression of Type II Secretion in Vibrio cholerae

ABSTRACT

Vibrio cholerae is a human pathogen that alternates between growth in environmental reservoirs and infection of human hosts, causing severe diarrhea. The second messenger cyclic di-GMP (c-di-GMP) mediates this transition by controlling a wide range of functions, such as biofilms, virulence, and motility. Here, we report that c-di-GMP induces expression of the extracellular protein secretion (eps) gene cluster, which encodes the type II secretion system (T2SS) in V. cholerae. Analysis of the eps genes confirmed the presence of two promoters located upstream of epsC, the first gene in the operon, one of which is induced by c-di-GMP. This induction is directly mediated by the c-di-GMP-binding transcriptional activator VpsR. Increased expression of the eps operon did not impact secretion of extracellular toxin or biofilm formation but did increase expression of the pseudopilin protein EpsG on the cell surface.

IMPORTANCE Type II secretion systems (T2SSs) are the primary molecular machines by which Gram-negative bacteria secrete proteins and protein complexes that are folded and assembled in the periplasm. The substrates of T2SSs include extracellular factors, such as proteases and toxins. Here, we show that the widely conserved second messenger cyclic di-GMP (c-di-GMP) upregulates expression of the eps genes encoding the T2SS in the pathogen V. cholerae via the c-di-GMP-dependent transcription factor VpsR.

INTRODUCTION

Vibrio cholerae is a major bacterial pathogen responsible for the diarrheal disease cholera, causing an estimated 2.8 million infections each year resulting in approximately 91,000 deaths (1). V. cholerae is endemic to coastal waterways in tropical countries, where it persists in the environment as a biofilm on chitinous surfaces and periodically causes outbreaks in human populations. V. cholerae can rapidly spread and multiply under favorable environmental conditions, as seen in the recent 2010 Haiti outbreak (2). A fundamental property that allows V. cholerae to cause disease is its ability to transition from environmental reservoirs to human hosts. This transition is in part regulated by the second-messenger molecule cyclic di-GMP (c-di-GMP) (3–5).

c-di-GMP is a nearly ubiquitous second-messenger molecule in bacteria that controls a range of physiological functions, including biofilm formation, motility, and virulence factor expression (6). c-di-GMP is synthesized by diguanylate cyclases (DGCs) (7) and degraded by phosphodiesterases (PDEs) (8, 9). Together, these enzymes alter the concentration of c-di-GMP in the cell in response to environmental inputs. c-di-GMP has been shown to repress motility and virulence to promote a sessile, biofilm-associated lifestyle by stimulating the production of exopolysaccharide (EPS) matrix substances and adhesins while inhibiting flagellar activity or expression (3, 10–12). Levels of intracellular c-di-GMP have been proposed to be high in environmental reservoirs, inducing biofilm formation to promote survival, whereas levels are reduced in human hosts, allowing virulence factor gene expression (4). However, recent results from our laboratory suggest that in vivo levels of c-di-GMP may be dependent upon spatial localization within the intestine (13).

The type II secretion system (T2SS) of V. cholerae also contributes to environmental persistence and host disease (14). The T2SS is a multiprotein complex that spans the inner and outer membranes of many Gram-negative bacteria. Proteins destined for export via the T2SS are first translocated across the cytoplasmic membrane via the SEC or TAT pathway (15, 16), where they assemble in the periplasm before being exported as fully folded proteins into the extracellular milieu (17). In V. cholerae, the T2SS consists of 13 proteins, 12 of which are encoded by contiguous genes comprising the extracellular protein secretion (eps) gene cluster. Many T2SS-dependent proteins are degradative enzymes or toxins that contribute to bacterial pathogenesis; thus, the T2SS is considered an important molecular machine necessary for virulence (18). Within the host, V. cholerae causes diarrhea by T2SS-dependent secretion of cholera toxin. In addition to cholera toxin, V. cholerae exports, via the T2SS, other extracellular factors, including chitinases, proteases, DNase, and pilin, which aid in its ability to successfully occupy diverse ecological niches (14). The T2SS also secretes the three proteins RbmA, RbmC, and Bap1, which are necessary for robust, shear-resistant biofilm formation (10, 19–21). Other major bacterial pathogens, such as Escherichia coli and Pseudomonas aeruginosa, secrete virulence factors encoded by genes with considerable similarity to those of V. cholerae through the T2SS (17, 22–24).

Until recently, little was known about transcriptional regulation of the T2SS. A microarray of V. cholerae found a significant increase in epsC to epsN transcripts upon induction of a DGC (25). Concurrent with our experiments, another study found that two promoters were predicted to be upstream of epsC; the downstream promoter was predicted to be σ⁷⁰ dependent, while the upstream promoter was σ^E dependent (26). Moreover, two V. cholerae DGCs were shown to induce an epsC-lux transcriptional fusion in E. coli (26).

In this work, we demonstrate the molecular mechanism by which the T2SS eps gene cluster is induced by c-di-GMP.

RESULTS

Transcriptional control of the eps gene cluster of V. cholerae.

Previous reports have suggested that the eps gene cluster in V. cholerae is induced by c-di-GMP (26, 33). To determine the mechanism of this regulation, we constructed a transcriptional fusion of a 628-bp region upstream of the first gene in the T2SS operon, epsC (referred to as PepsC), to the lux operon on the plasmid pBBRlux (Fig. 1A). Significant expression from PepsC-lux was observed in V. cholerae, suggesting that this region of DNA contained the promoter(s) for the eps gene cluster.

FIG 1.

c-di-GMP induces epsC. (A) Putative eps operon. Open reading frames are shown in dark gray, the PepsC sequence used in initial promoter analyses is shown in light gray, and the numbers indicate DNA base pair numbers on the N16961 V. cholerae chromosome 1 as a reference sequence. (B) PepsC was transcriptionally fused to the lux operon. Luminescence production following overexpression of the DGC QrgB and the vector control was determined in the V. cholerae ΔvpsL mutant. The gray bars represent noninduced cultures; the black bars represent cultures induced by 100 μM IPTG. A promoterless pBBRlux plasmid was used (Control). Relative light units (R.L.U.) were calculated by dividing the raw luminescence by the optical density at 595 nm. The error bars indicate standard deviations of three biological replicates. (C) c-di-GMP induces transcription of epsC, epsG, and epsH as measured by qRT-PCR. R.Q., relative quantification. *, P < 0.05; ns, not significant.

To determine if the region is upregulated by c-di-GMP, we examined expression of this transcriptional fusion in a V. cholerae cell that expressed the Vibrio harveyi DGC QrgB from a plasmid. Importantly, we demonstrated that this is a moderately active DGC that generates intracellular concentrations of c-di-GMP analogous to those of the low-cell-density quorum-sensing state in V. cholerae (34). Thus, the concentration of c-di-GMP generated by QrgB is physiologically relevant. Our results showed that PespC-lux was significantly induced by c-di-GMP (3.7-fold), indicating that this region upstream of espC is regulated by changes in c-di-GMP (Fig. 1B).

In a previous genetic screen to identify promoters of V. cholerae that are induced by the second-messenger molecule c-di-GMP (30), we identified a promoter located upstream of the epsG gene in the putative epsC-epsN operon (see Fig. S1A in the supplemental material). As epsC in induced by c-di-GMP and our initial studies of the epsG-lux promoter showed modest induction, we explored whether this was also the case for the promoter upstream of epsG. However, unlike PepsC, further examination revealed that c-di-GMP did not consistently induce PepsG-lux (see Fig. S1B in the supplemental material).

The eps gene cluster appears to be structured as an operon, although this has not been formally demonstrated. We therefore wondered if induction of PepsC by c-di-GMP leads to increased expression of the genes downstream of the epsC promoter. To determine this, we used quantitative reverse transcriptase PCR (qRT-PCR) analysis of the epsC, epsG, and epsH genes to quantify relative RNA levels with and without qrgB induction. We examined epsG and espH, as these genes are located downstream of the novel epsG promoter that we identified. Indeed, relative transcript abundance of epsC, epsG, and epsH was induced 2- to 3-fold with increased levels of c-di-GMP upon induction of qrgB (Fig. 1C). These data suggest that c-di-GMP induces transcription of the eps gene cluster in V. cholerae via activation of the epsC promoter. This result is in agreement with a prior microarray study that showed the entire eps operon was induced by c-di-GMP (33).

Characterization of the epsC promoter.

To identify the upstream region of PepsC necessary for expression and c-di-GMP induction, we constructed 5′ promoter truncations of the PepsC DNA fragment that was analyzed (Fig. 1) and measured the expression of these transcriptional fusions at low and high c-di-GMP levels via QrgB expression in V. cholerae. All the fragments of PepsC-lux that contained sequence from −155 and longer (numbered relative to the epsC translational start site) exhibited promoter activity and were stimulated by increased c-di-GMP equivalently to our original PepsC-lux fusion (Fig. 2). Alternatively, truncating to −124 virtually abolished all expression of the promoter and induction by c-di-GMP. A promoterless pBBRlux plasmid served as the negative control.

FIG 2.

Functional analysis of the epsC promoter. (Left) The epsC promoter was analyzed by generating 5′ deletions. The numbers on the left indicate the 5′ ends of the constructs relative to the epsC open reading frame. The arrows indicate transcriptional start sites determined by 5′ RACE. The slashes indicate that the sequence is not to scale. (Right) Luminescence production from PepsC-lux following overexpression of qrgB was determined in the V. cholerae ΔvpsL mutant. The gray bars represent noninduced cultures; the black bars represent cultures induced by 100 μM IPTG. A promoterless pBBRlux plasmid was used as the negative [(-)] control. The dotted lines indicate the critical region for promoter activity. The error bars indicate standard deviations. *, P < 0.05.

To further understand the transcriptional control of epsC, we determined potential transcriptional start sites upstream of epsC using 5′ rapid amplification of cDNA ends (5′ RACE). Two putative transcriptional start sites were identified in PepsC located at bases −203 and −88 relative to the epsC translation start site (Fig. 2). These transcriptional start sites lie immediately downstream of the σ^E and σ⁷⁰ promoters predicted by Zielke et al. based on sequence analysis (26). In that study, the σ^E 5′ promoter was named P2, while the σ⁷⁰ 3′ promoter was named P1. For consistency, we refer to the epsC promoters using this nomenclature. Based on the result shown in Fig. 2, we conclude that the transcriptional start site located at −88, PepsC1, is the primary site for transcription initiation under the conditions we examined here and that c-di-GMP induction of PepsC1 requires a sequence encoded 36 to 67 bp upstream of the PepsC1 transcriptional start site.

Induction of PepsC1 requires the transcription factor VpsR.

To further understand the c-di-GMP induction of PepsC1, we sought to identify transcription factors necessary for this regulation. Three c-di-GMP-dependent transcription factors have been identified in V. cholerae: FlrA, VpsT, and VpsR (30, 35, 36). We measured expression of PepsC in mutants carrying deletions of the genes for these transcription factors and observed that PepsC was fully induced by c-di-GMP in an flrA or vpsT deletion mutant. However, we found that c-di-GMP-mediated increase in bioluminescence was greatly reduced in a ΔvpsR mutant, suggesting that VpsR is required for the c-di-GMP-mediated regulation of eps genes (Fig. 3).

FIG 3.

c-di-GMP induction of PepsC1 requires VpsR. Luminescence production from the PepsC fusion following induction of qrgB with 100 μM IPTG (black bars) and under noninducing conditions (gray bars) was determined in ΔvpsL, ΔvpsL ΔflrA, ΔvpsL ΔvpsT, and ΔvpsL ΔvpsR strains. The error bars indicate the standard deviations of three biological replicates.

Analysis of the epsC promoter sequence identified a 14-bp putative VpsR binding site, TTTAACGTTTGAGA (Fig. 4C), located from −138 to −124, 36 bp upstream of PepsC1. This binding site matches 11/14 bases in the recently described VpsR binding site consensus sequence (37). Promoter truncation analysis indicated the region encoding this putative binding site is essential for c-di-GMP induction of epsC (Fig. 2). To determine if VpsR can bind to the epsC upstream region, we amplified the DNA in the −228 and −124 epsC-lux fusions (see Fig. 2) and performed electrophoretic mobility shift assay (EMSA) analysis with purified VpsR. Increasing amounts of VpsR bound to and shifted the 228-bp fragment but did not shift the 124-bp DNA probe (Fig. 4A). Multiple bands of different sizes were observed, suggesting that VpsR can bind to the epsC promoters in different multimeric states. Nonspecific interaction of VpsR with all the DNA probes could be observed at the highest concentrations of VpsR.

FIG 4.

VpsR directly binds to the epsC promoter. (A) EMSA of PepsC containing the fragments indicated in Fig. 2A with purified VpsR at the following concentrations: lanes 1 and 6, no protein; lanes 2 and 7, 0.9 μM; lanes 3 and 8, 1.7 μM; lanes 4 and 9, 3.1 μM; and lanes 5 and 10, 4.0 μM. (B) EMSA of the 5′ truncations shown in Fig. 2A with or without 0.9 μM VpsR. (C) Map of PepsC indicating the σ^E-dependent promoter (P2) and the σ⁷⁰-dependent promoter (P1) with the putative VpsR binding site. The VpsR binding site sequence is shown compared to the consensus sequence from Zamorano-Sanchez et al. (37). Bases that do not match the canonical binding site are shaded.

We further confirmed the location of the VpsR binding site by generating probes from all the epsC promoter truncations shown in Fig. 2 and analyzed VpsR binding using EMSA. Consistent with our genetic results, VpsR bound to every fragment except the −124 fragment (Fig. 4B). This is the only fragment that is not regulated by c-di-GMP and does not encode the putative VpsR binding site. These results indicate that VpsR specifically binds to PepsC at a VpsR binding site located from −124 to −155 relative to the epsC translation start site. We also attempted EMSA in the presence of excess c-di-GMP; however, we did not observe any change in binding affinity or binding patterns (data not shown). This is consistent with previously published results showing that c-di-GMP does not impact VpsR binding to target promoters (30, 37).

Mutation of the VpsR binding site abolishes c-di-GMP induction of the eps gene cluster.

Our results thus far indicate that transcription of epsC is induced by c-di-GMP and that the promoter region of epsC specifically binds to VpsR. To demonstrate that VpsR is the transcription factor necessary for c-di-GMP induction of epsC, we mutated the VpsR binding site upstream of PepsC1 on the genome of V. cholerae (labeled RBM for VpsR binding site mutant [see Fig. S2 in the supplemental material]). To validate that these mutations disrupted VpsR binding to the epsC promoter and subsequent induction by c-di-GMP, we first constructed a transcriptional PepsC-RBM-lux fusion into the pBBRlux vector (labeled PepsC-RBM). PepsC-RBM had basal levels of expression lower than those of PepsC, and it was not responsive to c-di-GMP (Fig. 5A). These results are consistent with disruption of VpsR binding to the promoter. We also observed in an EMSA that VpsR binding to this mutant promoter fragment was abolished compared with that of an equivalent promoter fragment encoding the wild-type (WT) VpsR binding site (Fig. 5B). Because epsC has multiple promoters, we confirmed, using qRT-PCR, that expression of epsC in the RBM mutant was uninducible by c-di-GMP (Fig. 5C). In addition, these results showed that expression of epsC in RBM was only mildly reduced (2-fold or less) compared to WT V. cholerae (Fig. 5C). These experiments demonstrated that VpsR binding to the epsC promoter is necessary for c-di-GMP induction.

FIG 5.

Mutation of the putative VpsR binding sites disrupts c-di-GMP induction of espC. For all the graphs, high c-di-GMP levels (black bars) were generated by inducing the DGC QrgB with 100 μM IPTG, while normal levels of c-di-GMP (gray bars) were the uninduced control. (A) Transcriptional fusions of wild-type or RBM (i.e., VpsR binding site mutant) epsC to lux were measured under high- and low-c-di-GMP conditions. *, P < 0.05. (B) Binding of VpsR to the WT promoter (lanes 1 to 4) or the VpsR binding site promoter (lanes 5 to 8) with no protein (lanes 1 and 5) or at VpsR concentrations of 0.37 μM (lanes 2 and 6), 0.74 μM (lanes 3 and 7), or 1.5 μM (lanes 4 and 8). (C) Transcription of epsC as measured by qRT-PCR in the wild-type or RBM mutant strain. The error bars indicate standard deviations.

c-di-GMP induction of epsC does not impact secretion of toxin or biofilm formation.

We hypothesized that induction of the eps genes would enhance phenotypes that are dependent on secretion of proteins through the T2SS. Cholera toxin is the best-characterized substrate of the T2SS (22, 32). Because all our previous experiments were performed in Luria-Bertani (LB) medium, we wished to examine toxin secretion under these conditions. As cholera toxin is not highly expressed in LB medium in the V. cholerae El Tor strain that we utilized, the plasmid pWD615, which constitutively expresses the etxB gene, was used. etxB encodes the B subunit of enterotoxin as a proxy for cholera toxin (CtxB). CtxB and EtxB share 88% protein sequence similarity and are both secreted by the T2SS at over 90% efficiency. They have previously been used interchangeably in E. coli and V. cholerae (38).

EtxB secretion was examined by centrifuging the cells to separate toxin that was in the supernatant versus toxin that was cell associated. EtxB was detected using a standard GM1 enzyme-linked immunosorbent assay (ELISA) (32). We observed secretion of EtxB at all time points examined, but this secretion was unaffected by c-di-GMP levels in the cell (Fig. 6). These data show that increased levels of c-di-GMP do not impact secretion of extracellular toxin by the T2SS under the conditions tested here.

FIG 6.

c-di-GMP does not impact secretion of EtxB. Secretion of the E. coli toxin EtxB constitutively expressed from a plasmid was determined at high and low levels of c-di-GMP. The black bars represent cultures induced with 100 μM IPTG; the gray bars represent noninduced cultures. The cultures were grown for the indicated times following 1/1,000 dilution. The cells were removed by centrifugation and lysed. Both the supernatant and cell-associated EtxB levels were determined using a GM1 ELISA. The error bars represent the standard deviations (n = 3).

Since the T2SS, c-di-GMP, and VpsR are all implicated in biofilm formation, we hypothesized that induction of the eps genes by c-di-GMP would alter biofilm formation. Biofilm formation was measured using the BacTiterGlo assay, which quantifies viable cells in the biofilm. Because c-di-GMP and VpsR are well known to induce biofilm formation (3, 30, 39), we quantified biofilm formation in the WT strain compared to the RBM mutant at high versus low c-di-GMP levels. The only difference between these strains is that induction of eps genes does not occur in the RBM mutant. Importantly, the rest of the biofilm regulatory cascade remains intact. The RBM and WT strains exhibited equivalent biofilm formation in the presence of high or low c-di-GMP levels (Fig. 7), showing that under the conditions we tested, transcriptional induction of epsC by c-di-GMP does not impact biofilm formation.

FIG 7.

Disruption of VpsR binding at epsC does not impact biofilm formation. Shown is biofilm formation as measured by quantifying viable bacteria (see Materials and Methods) of the WT or RBM mutant at high (black bars) levels of c-di-GMP generated by inducing the QrgB DGC enzyme with 100 μM IPTG versus normal levels of c-di-GMP (gray bars), representing the uninduced samples. The error bars represent the standard deviations (n = 3).

Induction of the eps operon by c-di-GMP increases EpsG on the cell surface.

The gene epsG encodes the major pseudopilin of the T2SS in V. cholerae (22). It has been suggested that polymerization and contraction of EpsG are the driving forces for secretion of folded proteins from the periplasm across the outer membrane (16). As transcription of the epsG gene was increased by c-di-GMP, we used a polyclonal antibody that binds to EpsG to determine if protein expression was similarly increased. Based on previous observations that the T2SS of P. aeruginosa and Klebsiella oxytoca form extracellular pseudopili upon overexpression of the major pseudopilin genes (40–42), we hypothesized that increased EpsG could lead to formation of an extracellular pilus. To test this hypothesis, using the method of Nunn et al. (43), we sheared extracellular pili from intact WT or RBM V. cholerae strains expressing QrgB or its corresponding active-site mutant (QrgB*) using a syringe and needle and observed the amount of EpsG that was associated with the cellular fraction or sheared pilus fraction using Western blot analysis. Quantification of the periplasmic beta-lactamase, which was expressed from the QrgB or QrgB* expression plasmid in the cell periplasm, was performed pre- and postshearing to determine cell lysis. An increase of 2-fold in beta-lactamase activity was detected in the cell supernatant after shearing, indicating that only modest disruption of membrane integrity occurred during this process (see Fig. S3 in the supplemental material). In the WT or RBM strain, only a small amount of EpsG could be detected in the sheared fraction at low c-di-GMP concentrations (Fig. 8). However, upon induction of QrgB in the WT strain, a significant amount of EpsG was detected in the sheared fraction. Alternatively, extracellular EpsG was greatly reduced in the RBM mutant at high c-di-GMP levels compared with the WT strain. These results suggest that c-di-GMP induction of the eps gene cluster by VpsR results in EpsG extracellular pilus formation.

FIG 8.

c-di-GMP induces extracellular EpsG. Extracellular EpsG was detected by Western blotting for six strains. Lanes: 1, WT V. cholerae pEpsG overexpression; 2, RS01-V. cholerae ΔepsG; 3, WT V. cholerae-QrgB*; 4, WT V. cholerae-QrgB; 5, RBM-QrgB*; 6, RBM-QrgB. The experiment was repeated for four biological replicates, yielding the same results, and one representative blot is shown.

DISCUSSION

Our results indicate that the transcription of the T2SS in V. cholerae is induced in a VpsR-dependent manner by c-di-GMP, and we demonstrated that this increased transcription increases production of EpsG. Until very recently, expression of the T2SS in V. cholerae was thought to be constitutive. However, our results, as well as a recent article by Zielke et al., demonstrated transcriptional control of the eps operon (26). It has been shown in V. cholerae that T2SS mutants lead to cell envelope stress and that the cell responds by induction of the σ^E stress response (44). Recently it has also been shown that the more upstream promoter, PepsC2, is induced by σ^E (26). This regulation makes logical sense, as increased synthesis of the epsC-epsN operon may be a response to periplasmic stress. Zielke et al. suggested that transcriptional control of the eps operon is a mechanism to appropriately modulate total cellular secretion capacity, although the impact of transcriptional regulation on secretion itself was not measured in the study (26). We likewise hypothesized that increased expression of the eps genes mediated by c-di-GMP would enhance T2SS-dependent phenotypes. However, contrary to our hypothesis, secretion of extracellular toxin and biofilm formation were unchanged when the epsC-epsN operon was induced, suggesting that overall eps gene expression does not necessarily directly correlate with T2SS activity.

One phenotype that we did observe to be associated with c-di-GMP induction of epsC was increased extracellular EpsG. Shearing pili using expulsion through a syringe is a standard method to remove extracellular appendages (43), and this approach has been shown not to significantly disrupt the bacterial membrane (41), consistent with our findings (see Fig. S3 in the supplemental material). Our results suggest that c-di-GMP induction of the eps gene cluster through VpsR leads to the production of EpsG pili on the surface of the cell. The nature and function of these pili are interesting topics for further study, but our results suggest that c-di-GMP could alter the T2SS in V. cholerae to function more like type IV pili (42).

It is interesting that in V. cholerae the quorum-sensing master regulator HapR induces epsC-epsN expression (26). HapR also reduces the intracellular concentration of c-di-GMP (34). Therefore, these regulatory effects are conflicting, as hapR mutants have high c-di-GMP levels but lower epsC-epsN expression. These observations suggest that the regulation of epsC-epsN by HapR must be through a c-di-GMP-independent mechanism that is epistatic to c-di-GMP induction of PepsC1.

The T2SS is necessary for proper biofilm development in V. cholerae, in part through secretion of three extracellular proteins (21), and c-di-GMP is similarly essential for biofilm formation (3, 6, 34). VpsR binds c-di-GMP and upregulates the vpsT and the vps gene clusters in a c-di-GMP-dependent manner (30, 39, 45). Thus, VpsR appears to be a molecular link connecting the intracellular c-di-GMP concentration to both Vibrio polysaccharide biosynthesis and type II secretion. The only other gene shown to be directly regulated by VpsR outside the biofilm gene cascade is aphA, encoding a transcription factor which positively regulates virulence factor expression (46). Therefore, this study identifies epsC as an additional direct target of VpsR and supports the idea that the VpsR–c-di-GMP regulon extends beyond the control of biofilm formation (47).

The T2SS is conserved in many bacterial species and is important in pathogenesis (17, 24). Additionally, more than 75% of bacteria contain DGC and PDE domains that are involved in c-di-GMP synthesis and degradation (48). A functional link between the T2SS and biofilms has been established in V. cholerae, as well as other gammaproteobacteria, such as E. coli (49). We therefore predict that c-di-GMP regulation of the T2SS might be widespread in other important bacterial pathogens.

MATERIALS AND METHODS

Bacterial strains, culture conditions, and DNA manipulation.

All experiments utilized V. cholerae El Tor biotype strain C6706str2 or mutant derivatives (Table 1). Plasmids were introduced into V. cholerae through biparental mating with E. coli S17 λpir as the donor and verified by antibiotic selection and culturing on thiosulfate-citrate-bile salts-sucrose agar plates (Difco). Unless otherwise stated, bacteria were incubated at 35°C with shaking at 220 rpm in LB medium (Accumedia). Agar plates were made with 15 g/liter agar (Accumedia). Antibiotics were used at the following concentrations: kanamycin, 100 μg/ml; chloramphenicol, 10 μg/ml; and polymyxin B, 10 U/ml. Protein expression vectors were induced with 100 μM isopropyl-β-d-thiogalactoside (IPTG) unless otherwise stated. All compounds were purchased from Sigma. V. cholerae strain RS01 (ΔepsG) was generated using the pKAS32 vector, as described previously (27). Relevant plasmids and primers are shown in Table 2.

TABLE 1.

Bacterial strains

Strain	Description	Reference
V. cholerae
C6706str2	Wild type	50
CW2034	ΔvpsL	34
WN310	ΔvpsR ΔvpsL	30
JP1195	ΔvpsT ΔvpsL	30
pDS54	ΔvpsL ΔflrA	35
RS01	ΔepsG	This study
E. coli S17	λpir	51

TABLE 2.

Plasmids and primers

Plasmid or primer	Description	Primer sequence(s)a (5′→3′)	Reference
Plasmids
pAEKlv8	−628 epsC promoter cloned into pBBRlux	F, ATAACTAGTCATAAGGAATAATCCGGC; R, ATACGGATCCAAATTTCCACGTTATTCC	This study
pAEK6	−451 epsC promoter cloned into pBBRlux	F, AGACACTAGTGCGTTGGTCTGAGATC; R, ATACGGATCCAAATTTCCACGTTATTCC	This study
pAEKsv1	−228 epsC promoter cloned into pBBRlux	F, ATAACTAGTGCCACATTGCCTCTCTAAGC; R, ATACGGATCCAAATTTCCACGTTATTCC	This study
pAEK5	−153 epsC promoter cloned into pBBRlux	F, AGACACTAGTCAAGCAAGTCGAC; R, ATACGGATCCAAATTTCCACGTTATTCC	This study
pAEKepsc2	−124 epsC promoter cloned into pBBRlux	F, AGACACTAGTCACTTCGCTCCAC; R, ATACGGATCCAAATTTCCACGTTATTCC	This study
pCMW75	qrgB expression vector		34
pCMW98	qrgB (GG→AA) active-site mutant expression vector		34
pBBRlux	Promoterless reporter backbone and vector control expressing Lux operon		52
pWD615	etxB expression vector		31
pEVS141	Vector control for expression vectors		53
Primers to make RDM mutant and RS01
USF		ATACACTAGTGATCACTCGCCAATTGGCG	This study
USR		TAAGGAGGATATTCATATGAGTTACTCCACACTATGTCG
DSR		ATGTGTTGACTGACCGAGCG
DSF		GAAGCAGCTCCAGCCTACACGCTTGGCTAATTAGCGGTAAC
5′-RACE primers
GSP		GCGCATGCTCTACCGCCCAAT	This study
For nested amplification		CTGGCCCTCCCCAAGCGACAA, GTCATTCAATATTGGCAGGT

^aF, forward; R, reverse.

PCR was performed using standard methods with Invitrogen HiFi Taq polymerase. Transcriptional fusions of test promoters were cloned into the SpeI and BamHI restriction sites of the pBBRlux plasmid using restriction endonucleases (Fermentas or New England BioLabs) and ligated with T4 DNA ligase (New England BioLabs). For gene expression studies, luminescence was measured in opaque, white 96-well microtiter plates (Corning) after 6 h of growth following 1/1,000 dilution on either a SpectraMax M5 plate reader (Molecular Devices) or an Envision multimode plate reader (2104-0020; Perkin Elmer). Statistical significance for all gene expression studies, including the examination of transcriptional reporter fusions and qRT-PCR, was determined using an unpaired Student t test.

Identification of transcriptional start sites.

RNA was prepared from CW2034 containing the plasmids pAEKlv8 and p6f12. CW2034 is a ΔvpsL mutant of C6706str2 that does not flocculate at high c-di-GMP concentrations and was thus used for experiments such as RNA extraction and gene expression. Cultures were lysed using TRIzol reagent according to the manufacturer's instructions (Invitrogen) from 2 ml of cells at an optical density at 600 nm (OD₆₀₀) of 0.5. The transcriptional start sites were determined by 5′ RACE (Invitrogen) according to the manufacturer's instructions. Two rounds of nested amplification using the primers listed in Table 2 were used with a 0.1% dilution of the original PCR mixture.

qRT-PCR.

Overnight cultures of CW2034 with the plasmids pCMW75 and pEVS141 were diluted 1:1,000 in 3 ml LB medium, induced with 1 mM IPTG, and grown to an OD₆₀₀ of 1.0. The cultures were harvested and pelleted by centrifugation, and the supernatant was discarded. The cells were lysed, and RNA was extracted with the RNeasy RNA extraction kit (Qiagen) according to the manufacturer's instructions. The RNA was reverse transcribed to cDNA with the GoScript reverse transcription system (Promega). Quantitative PCR was carried out with TaqMan reagents (Invitrogen). The 16S rRNA gene was used as a reference gene, and differences were determined using ΔΔC_T analysis.

Detection of EpsG on cell surfaces and measurement of cell lysis.

V. cholerae strains were grown on solid LB agar plates containing the appropriate antibiotics at 100 μg/ml and 500 μM IPTG at 35°C overnight. Bacteria were scraped off the agar and resuspended in PBS (136 mM NaCl, 2 mM KCl, 4 mM Na₂HPO₄, 1 mM KH₂PO₄, pH 7.4) by pipetting. Bacterial suspensions were normalized to an OD₆₀₀ of 5.8 in 6 ml of PBS, and a 300-μl preshearing aliquot was reserved for the nitrocefin assay. The remaining normalized bulk bacterial suspensions were passed through a 26-gauge hypodermic needle three times with a syringe to shear extracellular appendages. A 300-μl aliquot of each sheared suspension was reserved for nitrocefin assay, while the remaining sheared bulk suspensions were centrifuged twice at 10,000 × g in a Sorvall RC-513 centrifuge (DuPont Industries) for 5 min at 4°C to remove cells and large insoluble aggregates. The supernatants were then centrifuged at 150,000 × g in an L7-65 ultracentrifuge (Beckman) for 30 min at 4°C to pellet the sheared material, such as extracellular pili. The bulk supernatant was decanted, and the pelleted materials were resuspended in the residual supernatant. Total protein concentrations were determined by Bradford assay, and the samples were mixed with 2× loading buffer (120 mM Tris, pH 6.8, 2% SDS, 2% β-mercaptoethanol, 20% glycerol, 0.02% bromophenol blue), followed by 10 min of boiling; 0.5 μg of total protein from each sample was loaded into a 4 to 20% Mini-Protean TGX precast protein gel (Bio-Rad) and run by standard electrophoresis techniques. The resulting gel was analyzed by standard Western blotting techniques (28) using a polyclonal antibody to EpsG-His₆, as described previously (29), and a secondary antibody (horseradish peroxidase [HRP]-conjugated goat anti-rabbit IgG; ThermoFisher). Antibody detection was done with Pierce ECL Plus Western blotting substrate (ThermoFisher). The blot was visualized using an Amersham Imager 600 (GE Healthcare Life Sciences).

Cell fragility was determined by measuring extracellular beta-lactamase activity. After cells were scraped from the plates in the cell-shearing assay and resuspended in phosphate-buffered saline (PBS), 300 μl was removed to quantify beta-lactamase before cell shearing, and the same volume was removed after passing the cells through a 26-gauge hypodermic needle. Each of these samples was kept on ice, and the cells were pelleted by centrifugation. Two hundred microliters of supernatant was removed from each sample, which represented the beta-lactamase that had leaked from the cells. The cell pellet and remaining supernatant were resuspended in 100 μl of 2× spheroplast medium (100 mM Tris, pH 8.0, with 4,000 U/ml of polymyxin B sulfate) and incubated at 37°C for 30 min to disrupt the outer membrane to release beta-lactamase. The resulting spheroplasts were pelleted by centrifugation, and the supernatant was removed for the assay. Beta-lactamase was assayed using the chromogenic substrate nitrocefin (Cayman Chemical, Ann Arbor, MI). The substrate was prepared by adding 20 μl of 0.5 mg/ml nitrocefin in 0.1 M phosphate buffer, pH 7.0, 80 μl of H₂O, and 50 μl 0.2 M phosphate buffer, pH 7.0, to each well of a black, opaque 96-well plate. A 50-μl sample was added to the substrate, and activity was measured by reading the OD₄₈₂ every 30 s for 6 min. The final optical densities of the reaction mixtures are shown in Fig. S3 in the supplemental material.

EMSA.

VpsR purification was carried out as previously described (30). DNA probes were generated by PCR with the FAM (6-carboxyfluorescein)-labeled primers CMW234 and CMW235, which flank the SpeI and BamHI restriction sites of pBBRlux, amplifying the inserted DNA in the vector using the appropriate transcriptional fusions as template DNA. The DNA was purified using Promega SV gel and PCR cleanup kits. Ten-nanomolar probes were incubated with 50 μg/ml poly(dI-dC) at 30°C for 30 min with VpsR in a 20-μl total reaction volume balanced with VpsR buffer (20 mM sodium phosphate, 250 mM NaCl, 20% glycerol, and 0.05% β-mercaptoethanol). Following incubation, 2 μl of 80% glycerol was added to each reaction mixture, and appropriate volumes of the reaction mixture were loaded onto 5% polyacrylamide-Tris-acetic acid-EDTA gels. Electrophoresis was carried out for 1 h at 95 V and visualized on a Typhoon FLA 9000 scanner (GE Healthcare Life Sciences).

GM1 ELISA.

V. cholerae containing the EtxB-expressing plasmids pWD615 (31) and pCMW75 were grown overnight, diluted 1:100 in LB medium, and grown with shaking. One milliliter of cells was harvested in late exponential phase and pelleted by centrifugation. The supernatant and cells were separated and kept on ice, and the cells were resuspended in 1 ml of PBS and then lysed by sonication. Enterotoxin B was quantified by GM1 ELISA as previously reported (32).

Biofilm assay.

V. cholerae WT biofilms were grown on a minimum biofilm eradication concentration (MBEC) plate, where biofilms form on pegs attached to the lid. Overnight cultures in LB medium were diluted 1:500, and then 160 μl was incubated for 8 h at 35°C with no shaking. The MBEC lid with attached biofilm was then transferred to a new microtiter plate containing 160 μl of PBS for 5 min to remove planktonic bacteria. Subsequently, the lid was transferred to an opaque, black 96-well plate (PerkinElmer) containing 160 μl of 25% BacTiter-Glo (Promega) for 5 min before measuring the luminescence on a SpectraMax M5 plate reader (Molecular Devices).