Bile Salts Promote ToxR Regulon Activation during Growth under Virulence-Inducing Conditions

ABSTRACT

Cholera is an epidemic disease caused by the Gram-negative bacterium Vibrio cholerae. V. cholerae is found in aquatic ecosystems and infects people through the consumption of V. cholerae-contaminated food or water. Following ingestion, V. cholerae responds to host cues to activate the expression of critical virulence genes that are under the control of a hierarchical regulatory system called the ToxR regulon. The ToxR regulon is tightly regulated and is expressed in vitro only under special growth conditions referred to as AKI conditions. AKI conditions have been instrumental in elucidating V. cholerae virulence regulation, but the chemical cues within AKI medium that activate virulence gene expression are unknown. In this study, we fractionated AKI medium on a reverse-phase chromatography column (RPCC) and showed that the virulence-activating molecules were retained on the RPCC column and recovered in the eluate. Liquid chromatography–high-resolution mass spectrometry (LC-HRMS) analysis of the eluate revealed the presence of a known ToxR regulon activator, taurocholate, and other bile salts. The RPCC eluate activated the ToxR regulon when added to noninducing medium and promoted TcpP dimerization in a two-hybrid system, consistent with taurocholate being responsible for the virulence-inducing activity of AKI medium. Additional experiments using purified bile salts showed that the ToxR regulon was preferentially activated in response to primary bile acids. The results of this study shed light on the chemical cues involved in V. cholerae virulence activation and suggested that V. cholerae virulence genes are modulated in response to regionally specific bile acid species in the intestine.

INTRODUCTION

Vibrio cholerae is a Gram-negative bacterium that causes the epidemic disease cholera (1). V. cholerae is native to aquatic environments, from which people acquire cholera from ingesting V. cholerae-contaminated food or water. Following ingestion, V. cholerae colonizes the small intestine and produces a potent enterotoxin that is responsible for the production of a secretory diarrhea that can rapidly lead to dehydration and death. Cholera is an ancient disease, with accounts of cholera-like disease dating back to antiquity, and was responsible for seven global pandemics since the 1800s. We are amid the seventh cholera pandemic, with cholera continuing to emerge in areas where access to clean water is limited. The first six cholera pandemics were thought to have been caused by classical biotype strains of V. cholerae. The seventh pandemic coincided with the emergence of a new biotype, the El Tor biotype, which has since displaced classical strains in causing epidemic cholera.

Intestinal colonization by V. cholerae is dependent on the expression of virulence genes, which are under the control of a hierarchical regulatory system known as the ToxR regulon, which activates virulence gene expression in response to environmental cues in the gastrointestinal tract (2). The ToxR regulon consists of five primary regulatory proteins: AphA, AphB, TcpP, ToxR, and ToxT (3). The induction of the regulon is initiated by AphA and AphB binding to the tcpPH promoter to activate its transcription (4, 5). TcpP then binds, along with ToxR, to the toxT promoter to activate toxT transcription. ToxT is the most downstream regulator in the ToxR regulon and directly activates the expression of the operons that encode the two primary V. cholerae virulence factors, the toxin-coregulated pilus (TCP) and cholera toxin (CT) (3). TCP is a type IV pilus that is required for colonization of the small intestine, and CT is the enterotoxin that is responsible for the secretory diarrhea that is the hallmark of the disease cholera.

ToxR and TcpP are one-component transcriptional regulators that play a central role in V. cholerae virulence gene expression (6, 7). Both proteins contain a periplasmic sensing domain (PPD) that is attached to a cytoplasmic DNA-binding domain by a single transmembrane-spanning domain (6, 7). The toxR and tcpP genes are in separate operons, and both genes are positioned upstream of a gene encoding another membrane protein (toxS and tcpH, respectively), which are required for the full regulatory activity of both proteins (8, 9). The PPDs of both ToxR and TcpP are thought to sense and respond to environmental cues to regulate virulence gene expression. Several PPD agonists have been defined for ToxR, including bile salts (10, 11), indole (12), and the cyclic dipeptide cyclo(Phe-Pro) (13, 14). Studies suggest that these compounds destabilize the ToxR PPD to promote ToxS binding and increased ToxR activity (11). TcpP activity has been shown to be indirectly modulated by disulfide stress caused by the primary bile salt taurocholate (TCA) (15). Taurocholate has been shown to inhibit DsbA reductase activity, which promotes the formation of intermolecular disulfide bonds between two TcpP monomers to generate stable homodimers, which then bind to the toxT promoter to activate its expression (15, 16). TcpH contributes to TcpP activity by protecting it from proteolytic degradation (9). In addition to ToxR and TcpP, the remaining ToxR regulon components have also been shown to respond to environmental cues. These include aphA, which is regulated by the quorum-sensing systems and LeuO (13, 17); aphB, whose transcription is regulated by OmpR (18) and whose activity is modulated by oxygen and low pH (19, 20); and ToxT, which has been shown to be regulated posttranscriptionally in response to bicarbonate and fatty acids (21, 22).

In El Tor V. cholerae strains, the ToxR regulon is not expressed during growth under standard laboratory conditions, which was a barrier for in vitro studies on V. cholerae virulence. In 1986, Iwanaga et al. developed artificial in vitro culture conditions that supported ToxR regulon activation and, thus, high-level production of virulence factors in El Tor vibrios (23). These virulence-inducing conditions were referred to as “AKI conditions” and require the growth of V. cholerae statically at 37°C for 4 h in 10 ml of AKI medium (1.5% Bacto peptone, 0.4% yeast extract, 0.5% NaCl, 0.3% NaHCO₃) in a 150-mm by 15-mm test tube before the culture is transferred to a 100-ml Erlenmeyer flask and incubated with constant shaking at 37°C overnight. A subsequent study showed that NaHCO₃ could be omitted from the medium (24). AKI conditions have since been widely adopted by the cholera research field for the study of V. cholerae virulence and have been shown to reproduce many virulence phenotypes that have been observed during mammalian infection.

AKI conditions have been instrumental in advancing our understanding of the mechanisms involved in V. cholerae virulence. However, the molecular basis for AKI medium-dependent activation of the ToxR regulon is not understood. In particular, the chemical cues in AKI medium that are responsible for ToxR regulon activation are unknown. Recent studies from our laboratory documented that passing AKI medium through a C₁₈ chromatography column resulted in attenuated CT and TCP production (18). As the C₁₈ column-extracted AKI medium did not affect V. cholerae growth, this observation suggested that the virulence-inducing molecule(s) that was present in AKI medium was retained on the C₁₈ chromatography column. In this report, we investigated this observation to characterize the virulence-activating compound(s) in AKI medium. We found that AKI medium contained bile acids, including taurocholate, and that the presence of the bile acid-containing fraction was critical for the virulence-inducing properties of AKI medium. We document that these activating molecules facilitated TcpP dimerization and activated the expression of the bile-regulated vexCD promoter, suggesting that the activating molecules were likely bile acids. We further show that the ToxR regulon was activated in vitro in response to primary bile acids, including taurocholate, glycocholate (GCA), and glycochenodeoxycholate (GCDCA).

RESULTS

AKI medium contains virulence-activating compounds.

In a recent study, we documented that passing AKI medium through a C₁₈ chromatography column attenuated its ability to support virulence factor production in O1 El Tor V. cholerae (18). To confirm this, we quantified CT production in V. cholerae O1 El Tor strain JB58 following growth under AKI conditions in AKI broth and AKI broth that had been passed through a C₁₈ chromatography column (C₁₈-extracted AKI medium). The results confirmed our previous study and showed that CT production in V. cholerae was suppressed during growth in the C₁₈-extracted AKI medium (Fig. 1A). This suggested that passing the AKI medium over the C₁₈ column had removed the virulence-activating molecule(s) that was present in the AKI medium. The decrease in CT production was not a result of growth differences as JB58 exhibited similar growth kinetics in the two media (data not shown). Based on this result, we hypothesized that the virulence-activating compound(s) in AKI medium was being retained on the C₁₈ chromatography column. To test this, we eluted the retained molecules from the C₁₈ column with 100% methanol. The resulting eluate was then concentrated by evaporation and resuspended in an equivalent volume of C₁₈-extracted AKI medium. The resulting medium was then used to assess CT production in V. cholerae during growth under AKI conditions. The results confirmed that adding back the C₁₈ column eluate to the extracted AKI medium restored CT production to a level that was equivalent to that in the unextracted medium (Fig. 1A). From this, we concluded that that the virulence-activating molecule(s) in AKI medium was present in the C₁₈ column eluate.

FIG 1.

AKI medium contains molecules that modulate virulence factor production. (A) Cholera toxin (CT) production following growth of V. cholerae strain JB58 overnight under virulence-inducing conditions in AKI medium, AKI medium extracted on a C₁₈ chromatography column, or C₁₈-extracted AKI medium reconstituted with the C₁₈ chromatography column AKI medium eluate. (B) CT production following growth of V. cholerae strain JB58 overnight in AKI broth or LB under AKI growth conditions or shaken growth conditions with the addition of the C₁₈ chromatography column AKI medium eluate or the C₁₈ chromatography column LB medium eluate. The results are the means and standard deviations (SD) from three independent experiments. Statistical relevance compared to the WT cultured in AKI medium under AKI conditions was calculated using one-way analysis of variance (ANOVA) with Dunnett’s post hoc test. *, P ≤ 0.05; ns, not significant. ND, not detected.

We next tested if the AKI medium C₁₈ column eluate supported virulence factor production during growth under non-virulence-inducing conditions by adding the C₁₈ column AKI medium eluate to lysogeny broth (LB). We then quantified CT production by V. cholerae JB58 under AKI conditions (i.e., 4 h of static growth followed by aerated growth on a shaker) and non-AKI conditions (i.e., constant shaking). The results showed that there was no detectable CT produced when JB58 was cultured in LB under AKI conditions (Fig. 1B). In contrast, the addition of the AKI medium C₁₈ column eluate to the LB medium resulted in high-level CT production. The C₁₈ column-dependent induction of CT production was dependent on the growth of the culture under AKI conditions as indicated by the drastic decrease in CT production when the C₁₈- column eluate-supplemented LB culture was incubated under non-AKI growth conditions (Fig. 1B). This provided additional confirmation that virulence-inducing compounds in AKI broth were present in the C₁₈ column eluate. Furthermore, these results suggested that the initial 4-h static growth period used during growth under AKI conditions was critical for virulence activation, consistent with previous studies showing that AphB activity was activated via a thiol switch during incubation under reduced-oxygen conditions (19).

The above-described results suggested that the AKI medium C₁₈ column eluate contained molecules that were critical for virulence activation in V. cholerae. As a control, we also extracted LB medium on a C₁₈ chromatography column and tested if the LB medium C₁₈ column eluate supported CT production when added to LB. The results showed that the LB medium C₁₈ column eluate did not support CT production (Fig. 1B), which suggested that the virulence-activating compounds were present only in AKI medium.

Virulence-activating compounds in AKI medium induce toxT transcription.

The production of V. cholerae virulence factors is positively regulated by the ToxR regulon (3). As the C₁₈ column AKI medium eluate contained virulence-inducing molecules, we sought to determine at which point in the ToxR regulon the activating molecules were functioning. We therefore cultured wild-type (WT) strain JB58 bearing plasmid-based transcriptional reporters for the primary ToxR regulon genes under AKI conditions in AKI medium or C₁₈-extracted AKI medium. We then quantified reporter expression at 3.5 h for the toxR, aphA, aphB, tcpP, and toxT genes and 5 h for the genes encoding CT and TCP production (i.e., ctxA and tcpA). The results showed that toxR, aphB, and tcpP expression did not change during growth in C₁₈-extracted AKI medium, relative to the unextracted AKI medium (Fig. 2A, C, D), whereas aphA expression was increased in C₁₈-extracted AKI medium (Fig. 2B). In contrast, the expression of toxT was significantly reduced following growth in the C₁₈-extracted AKI medium relative to the unextracted AKI medium (Fig. 2E). Consistent with this, the expression levels of the two genes downstream of ToxT in the ToxR regulon, tcpA (Fig. 2F) and ctxA (Fig. 2G), were also significantly reduced during growth in the C₁₈-extracted AKI medium relative to unextracted AKI medium. This is consistent with the data in Fig. 1 showing that the virulence factor-inducing molecule(s) was retained in the C₁₈ column AKI medium eluate.

FIG 2.

Effect of C₁₈-extracted AKI medium on ToxR regulon expression. WT V. cholerae JB58 harboring transcriptional reporters for the indicated ToxR regulon genes was cultured under AKI conditions in either AKI medium or C₁₈-extracted AKI medium for 3.5 h (toxR, aphA, aphB, tcpP, toxT, ompR, and leuO) or 5 h (ctxA and tcpA) when gene expression was quantified as Miller units (MU) or relative light units per optical density. Data represent the means ± SD from three independent experiments performed in triplicate. Statistical significance was determined using Student’s t test. *, P ≤ 0.05. RLU, relative light units.

OmpR and LeuO were previously shown to repress ToxR regulon expression, with OmpR repressing aphB transcription (18) and LeuO directly repressing aphA transcription (25). We therefore tested if growth in C₁₈-extracted AKI medium affected ompR or leuO expression. The results showed that there was no difference in ompR expression between the two media (Fig. 2H), while leuO expression was significantly decreased in C₁₈-extracted AKI medium (Fig. 2I). This indicated that neither ompR nor leuO contributed to virulence attenuation in the C₁₈-extracted AKI medium. This conclusion was further supported by the observation that aphA expression was increased while aphB expression was unchanged during growth in the C₁₈-extracted AKI medium (Fig. 2B). From these results, we conclude that the virulence-activating molecule(s) in AKI medium functions by inducing toxT transcription.

Induction of toxT expression by the C₁₈ column AKI medium eluate does not require the periplasmic domain of ToxR.

The expression of toxT is under the control of ToxR and TcpP, which together bind at the toxT promoter to activate its expression (7, 26). ToxR is a one-component regulator that contains a PPD connected to a cytoplasmic DNA-binding domain by a transmembrane-spanning domain (6). ToxR activity is modulated by environmental cues (e.g., bile salts and cyclic dipeptides), which appear to interact with its PPD to effect transcriptional changes at target genes, such as leuO (11, 13, 25). However, the ToxR PPD is dispensable for the ToxR-dependent activation of toxT transcription (13, 25, 27). We therefore tested whether toxT activation via the bioactive molecule(s) in AKI medium was dependent on the ToxR PPD. To do this, we compared toxT-lacZ expression in WT strain JB58 and an isogenic strain expressing a toxR allele lacking its PPD (ToxR^ΔPPD) during growth in LB or LB supplemented with the AKI medium C₁₈ column eluate. The ToxR^ΔPPD allele used in these studies was previously shown to be membrane localized and to support the expression of both virulence genes and porin genes (25, 27). The results showed that the addition of the AKI medium C₁₈ column eluate to LB medium activated toxT expression to equivalent levels in both strains (Fig. 3), indicating that the ToxR PPD was dispensable for toxT activation by the bioactive molecule(s) that is present in AKI medium and suggesting that the inducing signal likely does not function by altering ToxR activity at the toxT promoter.

FIG 3.

Activation of toxT transcription by the AKI C₁₈ column extract eluate is independent of the ToxR periplasmic domain. V. cholerae WT strain JB58 and an isogenic strain expressing toxR lacking its periplasmic sensing domain (toxR^ΔPPD) bearing a toxT-lacZ transcriptional reporter were cultured under AKI conditions in LB or LB supplemented with the C₁₈ column AKI medium eluate for 4 h when toxT expression was determined as Miller units (MU). Statistical significance was determined using Student’s t test. *, P ≤ 0.05 (relative to the LB control).

The C₁₈ column AKI medium eluate promotes TcpP dimerization.

Activation of toxT transcription is mediated by ToxR and TcpP, which together bind at the toxT promoter to activate its expression. As the above-described data indicated that the C₁₈ column AKI medium eluate activated toxT expression in a ToxR PPD-independent manner, we hypothesized that the AKI medium eluate may have affected TcpP activity. TcpP activity is promoted by the formation of homodimers that are stabilized by intermolecular disulfide bonds between the PPDs of two adjacent TcpP monomers by a mechanism that is promoted by taurocholate (15). We therefore tested if passing AKI medium through a C₁₈ chromatography column affected TcpP dimerization. We used a previously established bacterial adenylate cyclase two-hybrid system to assay TcpP dimerization in Escherichia coli (15, 28). In this system, full-length tcpP was translationally fused to the T25 and T18 domains of the Bordetella pertussis adenylate cyclase (cyaA). The resulting plasmids were introduced into an E. coli cyaA mutant, and the cultures were grown in the presence of the inducer isopropyl-β-d-1-thiogalactopyranoside (IPTG) before β-galactosidase activity, which is used as a quantitative readout for protein dimerization, was assayed. As a control, we first tested the effect of taurocholate addition on TcpP dimerization during growth in LB (Fig. 4A). The results showed a taurocholate concentration-dependent increase in β-galactosidase activity, which indicated that taurocholate stimulated TcpP homodimerization and confirmed previous reports (15, 16) (Fig. 4A). We next tested the effect of AKI medium on TcpP dimerization (Fig. 4B). The results showed that the basal level of β-galactosidase activity was elevated in AKI medium relative to LB medium (Fig. 4A and B). As the growth kinetics of the tested E. coli cyaA strain were similar in both media (not shown), this suggests that AKI medium contains a compound(s) that promotes TcpP dimerization. The growth of the reporter strain in C₁₈-extracted AKI medium resulted in decreased β-galactosidase activity, suggesting that the compound(s) responsible for TcpP dimerization was retained on the C₁₈ column. Reconstituting the C₁₈-extracted AKI medium with the C₁₈ column AKI medium eluate restored TcpP dimerization to a level that was equivalent to that in unextracted AKI medium, confirming that the active compound(s) was retained on the C₁₈ column (Fig. 4B). Based on these results, we conclude that the bioactive molecule(s) that is responsible for the activation of the ToxR regulon in AKI medium likely functions by promoting TcpP dimerization.

FIG 4.

The C₁₈ column AKI medium eluate promotes TcpP dimerization. Plasmids bearing tcpP fused to the T18 and T25 domains of B. pertussis adenylate cyclase were introduced into an E. coli cyaA mutant. The resulting strain was cultured in LB medium with 0.5 mM IPTG at 30°C for 8 h when β-galactosidase activity was measured and reported as Miller units (MU). Statistical relevance compared to LB was calculated using one-way ANOVA with Dunnett’s post hoc test. (A) The addition of taurocholate to LB medium promotes TcpP homodimerization, with maximum activity occurring at 400 μM taurocholate. (B) TcpP homodimerization in AKI medium, C₁₈-extracted AKI medium, and C₁₈-extracted AKI medium reconstituted with the C₁₈ column AKI medium eluate. Statistical relevance compared to AKI medium was calculated using one-way ANOVA with Dunnett’s post hoc test. *, P ≤ 0.05; **, P ≤ 0.005.

The C₁₈ column AKI medium eluate activates vexCD expression.

As previous studies showed that TcpP dimerization is mediated by taurocholate, the above-described two-hybrid results led us to hypothesize that AKI medium contained bile salts. To test this, we used the vexCD operon promoter as a surrogate reporter for the presence of bile salts in AKI medium. The vexCD operon encodes a bile acid-specific efflux system. Studies have previously shown that this system is specifically activated in response to exogenous bile salts but not other detergent-like compounds (29, 30). We cultured WT strain JB58 bearing a vexCD-lacZ reporter in AKI medium, C₁₈-extracted AKI medium, LB, and LB containing the C₁₈ column AKI medium eluate for 3 h when we quantified vexCD-lacZ expression. The results showed a relatively high vexCD expression level in AKI medium and an ∼4-fold reduction in vexCD expression in the C₁₈-extracted AKI medium (Fig. 5). The decreased vexCD expression in the C₁₈-extracted AKI medium suggested that AKI medium likely contained bile salts and that the bile salts were retained on the C₁₈ chromatography column when AKI medium was passed through the column. The basal level of vexCD expression in LB medium was reduced relative to that in AKI medium, suggesting a lower concentration of bile salts in LB medium. The addition of the C₁₈ column AKI medium eluate to the LB medium resulted in a large increase in vexCD expression (Fig. 5), suggesting that the vexCD-inducing molecule(s) was enriched in the C₁₈ chromatography column AKI medium eluate. Together, these data suggest that AKI medium likely contains bile salts and that the bile salts were retained on the C₁₈ chromatography column and enriched in the AKI medium C₁₈ column eluate.

FIG 5.

The AKI C₁₈ column extract eluate induces the expression of the vexCD RND efflux system. WT V. cholerae strain JB58 harboring a vexCD-lacZ transcriptional reporter was incubated in AKI broth, C₁₈-extracted AKI broth, LB, or LB supplemented with the C₁₈ column AKI medium eluate. The cultures were incubated at 37°C with shaking for 3 h when aliquots were collected and used to quantify vexCD-lacZ expression as Miller units (MU). The results represent the means plus SD from three independent experiments performed in triplicate. Statistical relevance compared to growth in AKI broth was calculated using one-way ANOVA with Dunnett’s post hoc test. *, P ≤ 0.05.

AKI medium contains multiple bile salts.

The finding that vexCD expression was increased upon the addition of the C₁₈ column AKI medium eluate to LB medium suggested that AKI medium may contain bile salts. To test this, samples of AKI medium, C₁₈-extracted AKI medium, and the C₁₈ chromatography column AKI medium eluate were submitted to the University of Pittsburgh Health Sciences Metabolomics and Lipidomics Core (www.metabolomics.pitt.edu) for qualitative bile acid analysis by liquid chromatography–high-resolution mass spectrometry (LC-HRMS). The results confirmed the presence of multiple bile acid species in AKI medium, including taurocholate, a known TcpP activator (Table 1). The bile salt species were effectively removed from AKI medium following passage of the AKI medium over the C₁₈ chromatography column and were recovered in the C₁₈ chromatography column eluate (Table 1). Note that the reported bile salt analyte/control ratios cannot be used for quantitative comparisons between the samples as the results are influenced by chemical complexity and spectral overlap in the three test samples. The identification of taurocholate in AKI medium and its enrichment in the AKI medium C₁₈ column eluate further validated the above-described data showing that AKI medium and the C₁₈ column AKI medium eluate promoted TcpP dimerization (Fig. 4). From these results, we conclude that AKI medium contains bile salts, which are likely responsible for the AKI medium-dependent induction of virulence factor production in O1 El Tor V. cholerae.

TABLE 1.

LC-HRMS analysis of AKI medium, C₁₈-extracted AKI medium, and the AKI medium C₁₈ column eluate^a

Bile salt	Ratio (analyte/control) (SD)
	AKI medium	C18-extracted AKI medium	AKI medium C18 column eluate
CA	45.922 (0.747)	ND	49.751 (4.224)
CDCA	0.516 (0.025)	ND	1.332 (0.153)
DCA	4.130 (0.061)	ND	4.925 (0.363)
GCA	53.100 (0.980)	ND	57.575 (3.538)
GCDCA	2.598 (0.084)	ND	2.881 (0.173)
GDCA	10.547 (0.157)	ND	12.124 (0.799)
GLCA	0.047 (0.003)	0.0008 (0.0004)	0.076 (0.006)
GUDCA	0.028 (0.002)	ND	0.032 (0.002)
TCA	58.806 (1.503)	0.0050 (0.0040)	61.0378 (5.051)
TDCA	6.221 (0.106)	0.0020 (0.0003)	6.955 (0.472)
TLCA	0.121 (0.005)	0.0050 (0.0001)	0.179 (0.012)
TUDCA	0.357 (0.007)	ND	0.410 (0.036)

^aAbbreviations: CA, cholate; CDCA, chenodeoxycholate; DCA, deoxycholate; GCA, glycocholate; GCDCA, glycochenodeoxycholate; GDCA, glycodeoxycholate; GLCA, glycolithocholate; GUDCA, glycoursodeoxycholate; TCA, taurocholate; TDCA, taurodeoxycholate; TDCDA, taurochenodeoxycholate; TLCA, taurolithocholate; TUDCA, tauroursodeoxycholate; ND, not detected.

Bile salts activate virulence factor production in El Tor V. cholerae.

The observation that the C₁₈ column AKI medium eluate contained virulence-activating molecules and was enriched for bile salts led us to test whether specific bile salts present in AKI medium could activate virulence gene expression. To this end, we cultured V. cholerae JB58 bearing a ctxA-lux reporter under AKI conditions in LB medium supplemented with 400 μM or 800 μM of the following commercially available primary and secondary bile salts: cholate (CA), chenodeoxycholate (CDCA), deoxycholate (DCA), glycochenodeoxycholate (GCDCA), glycocholate (GCA), glycodeoxycholate (GDCA), taurocholate (TCA), and taurodeoxycholate (TDCA) (Fig. 6A). AKI medium (inducing medium) and LB medium (noninducing medium) were included as positive and negative controls, respectively. The inoculated cultures were incubated statically for 4 h when reporter expression was recorded and normalized by the optical density of the culture. The results revealed that all of the bile salts, except DCA, induced ctxA expression to levels exceeding those of the LB control (Fig. 6B). The results showed that GCDCA treatment exhibited the greatest induction of ctxA expression, followed by TCA, GCA, and TDCA. CA showed a concentration-dependent increase in ctxA expression at 400 and 800 μM. Cholate at 800 μM activated ctxA to a level that was similar to that observed in AKI medium (Fig. 6). DCA was previously reported to poorly induce tcpA expression, which was consistent with our results (15, 31). In contrast to the other bile salts, we noted that the addition of DCA to the culture medium resulted in a significant growth defect (not shown).

FIG 6.

Effect of bile salts on V. cholerae virulence gene expression. (A) Schematic of the pathway for bile acid production and processing in humans. Primary bile acids are produced in the liver and secreted into the intestine. Thereafter, the primary bile acids are modified by the host microbiota to generate secondary bile acids. Bile acids promoting CT production are in red, and bile acids that did not promote CT production are in blue. Chenodeoxycholate (CDCA) is in black and was not tested. (B) V. cholerae harboring a ctxA-luciferase fusion transcriptional reporter plasmid was cultured under AKI conditions in AKI broth or LB supplemented with 800 μM of the indicated bile salts for 4 h when reporter expression was measured and normalized by the culture optical density. The results represent the means plus SD from three technical replicates. Abbreviations: CA, cholate; DCA, deoxycholate; GCDCA, glycochenodeoxycholate; GCA, glycocholate; GDCA, glycodeoxycholate; TCA, taurocholate; TDCA, taurodeoxycholate; ng, no growth.

Effect of bile salts on CT and TCP production.

To assess whether the bile salt-specific induction of ctxA described above correlated with CT and TCP production, we repeated the above-described experiments with WT strain JB58 and quantified CT and TcpA production following growth overnight under AKI conditions in the bile salt-supplemented LB medium. The results showed that TCA was the most potent bile salt in promoting both CT and TcpA production (Fig. 7). TCA appeared to exhibit concentration-dependent induction of CT production, with the highest level of CT production at 800 μM (Fig. 7A). Similar results were observed for TcpA production, where treatment with 800 μM TCA activated TcpA production to a level that was similar to that of AKI medium-grown cultures (Fig. 7B). The results for GCA mirrored the TCA results, but the level of CT and TcpA production was lower than that observed with TCA. GDCA also activated CT and TcpA production, with CT levels reaching ∼20 to 25% of the level produced in AKI medium. Significantly, while GCDCA activated ctxA expression to a level that was higher than that observed in AKI medium (Fig. 6B), this did not translate into a high level of CT or TcpA production. A similar phenomenon was observed with CA, which induced ctxA to levels similar to those in AKI medium at 800 μM (Fig. 6B) but induced CT production to <10% of the AKI medium level. The remaining bile salts (e.g., DCA, CA, GDCA, and TDCA) poorly stimulated CT and TcpA production, with levels being less than 10% of the AKI level. It was noteworthy that among the tested bile salts, primary bile salts appeared to best support CT and TcpA production, whereas secondary bile salts poorly activated CT production (Fig. 6A and Fig. 7). From these results, we conclude that conjugated primary bile salts are the most efficient activators of the ToxR regulon, with taurocholate being the most potent activator, as previously reported.

FIG 7.

Effect of bile acids on CT and TCP production. (A) Relative CT production by V. cholerae strain JB58 following growth under AKI conditions in AKI medium or LB medium supplemented with 0, 400, or 800 μM of the indicated bile salts. The results represent the means plus SD from at least three independent experiments. (B) TcpA Western blotting of V. cholerae strains cultured in 400 μM (top) or 800 μM (bottom) of the indicated bile salts. The TcpA band is marked by the arrow on the left. Abbreviations: CA, cholate; CDCA, chenodeoxycholate; DCA, deoxycholate; GCDCA, glycochenodeoxycholate; GCA, glycocholate; GDCA, glycodeoxycholate; TCA, taurocholate; TDCA, taurodeoxycholate.

Western immunoblot assays of V. cholerae cultured under AKI conditions in LB supplemented with 400 or 800 μM bile salts produced corresponding results, with TCA resulting in TcpA production to levels similar to those observed in the AKI medium control (Fig. 7B). GCA produced detectable TcpA but at reduced levels relative to those of TCA and those in the AKI medium culture. The remainder of the tested bile salts did not induce levels of TcpA production that were obviously different from those of the LB medium control.

DISCUSSION

V. cholerae colonization of the small intestine is dependent on its ability to respond to environmental cues in the gastrointestinal tract. While the characterization of these signals in vivo is complicated, the development of artificial virulence gene-inducing conditions using AKI medium and AKI conditions has facilitated the discovery of many factors involved in virulence regulation in O1 El Tor V. cholerae. Furthermore, AKI growth conditions have been shown to replicate virulence-associated phenotypes observed during mammalian infection. However, the chemical cues that facilitate the AKI medium-dependent activation of the ToxR regulon are poorly understood. In this study, we documented that AKI medium contains bile salts that are critical for ToxR regulon activation and virulence factor production. Furthermore, our data suggest that the ToxR virulence regulon is preferentially activated in response to primary bile acids, suggesting the possibility that individual bile salt species may function as regional cues to modulate gene expression during V. cholerae pathogenesis.

We present multiple lines of evidence to suggest that the presence of bile salts in AKI medium is responsible for its ability to support ToxR regulon activation. The virulence-inducing activity of AKI medium correlated with the presence of bile acids, as confirmed by mass spectrometry showing that bile acids, including taurocholate, were present in AKI medium, depleted from the AKI medium following passage over a C₁₈ chromatography column, and recovered in the C₁₈ column eluate. Furthermore, the observation that the active molecules in AKI medium functioned independently of the ToxR PPD and promoted TcpP dimerization suggested that it was likely that the presence of taurocholate in AKI medium was responsible for ToxR regulon activation. The observation that taurocholate addition to LB medium largely reproduces the inducing activity and virulence factor production capability of AKI medium supports this notion.

The active form of TcpP, which binds to the toxT promoter, is thought to be a homodimer. Taurocholate contributes to virulence activation by indirectly promoting interprotein disulfide bond formation between cysteines that are present in the PPDs of two adjacent TcpP monomers to generate stable homodimers (Fig. 8) (15). The mechanism behind taurocholate-dependent activation was linked to taurocholate-specific effects on the oxidation state of DsbA (16). When active, DsbA inhibits TcpP dimerization by catalyzing the formation of intramolecular disulfide bonds between two TcpP PPD cysteine residues. However, the presence of taurocholate results in DsbA accumulating in an reduced form that cannot act on newly formed TcpP to generate intramolecular disulfide bonds (15, 16). This frees the two cysteine residues to form intermolecular disulfide bonds to promote the formation of active TcpP homodimers. While we demonstrate that AKI medium contains taurocholate, we note that it is possible that multiple bile salts, or potentially other medium components that coextract with bile salts on the C₁₈ column, could also contribute to virulence activation in AKI medium (see below).

FIG 8.

Model of the function of bile salts in the activation of the ToxR regulon. (A) During the growth of V. cholerae under noninducing conditions (e.g., LB medium), DsbA functions to oxidize the two cysteine residues in the periplasmic domain (PPD) of TcpP, resulting in the formation of intramolecular disulfide-bonded TcpP, which cannot form stable dimers that are required for the activation of toxT transcription. (B) The presence of taurocholate in AKI medium (and presumably in vivo) inhibits DsbA reductase activity, allowing TcpP to dimerize and form intermolecular disulfide bonds that stabilize the homodimer to facilitate TcpP binding to the toxT promoter in conjunction with ToxR to activate toxT expression. ToxT then directly activates the expression of the genes required to produce cholera toxin (CT) and the toxin-coregulated pilus (TCP).

Our results documented that LB medium did not support ToxR regulon activation. Given that LB medium differs from AKI medium in its nitrogen source (Bacto tryptone versus Bacto peptone), we inferred that Bacto peptone is the source of the virulence-inducing molecules in AKI medium. While Bacto tryptone is derived from a pancreatic digest of casein, Bacto peptone is produced from an enzymatic digest of animal protein. The presence of bile acids in Bacto peptone suggests that the animal tissues used to produce Bacto peptone either contain or are contaminated with bile acids. Without further analysis, we cannot exclude the possibility that Bacto tryptone also contains bile acids, but if it does contain bile acids, our results indicate that they are present at a lower concentration than that found in Bacto peptone. The broad spectrum of primary and secondary bile acids identified in the mass spectrometry analysis suggests that the tissues used to produce Bacto peptone may have contained intestine, liver, and/or gallbladder tissue.

Bile acids are amphipathic molecules that play an important role in digestion by solubilizing ingested fats (reviewed in reference 32). The primary bile salts (e.g., chenodeoxycholate, cholate, and their taurine and glycine conjugates) are produced in the liver, stored in the gallbladder, and secreted into the proximal small intestine at high-millimolar concentrations (Fig. 7A). Bile acids have been shown to affect numerous V. cholerae phenotypes that have been linked to bile salt-dependent effects on ToxR and TcpP activity. For example, bile acids appear to directly interact with the ToxR PPD to induce conformational changes that affect its activity at target promoters (10, 11). Bile acids have also been shown to affect the redox state of ToxR and to inhibit ToxR proteolysis under starvation conditions (33, 34). Other metabolites, including the cyclic dipeptide cyclo(Phe-Pro) and indole, have also been shown to activate the ToxR regulon through a ToxR PPD-dependent mechanism (10, 12, 13, 35). A recent study further suggests that ToxR can respond to cell metabolites to modulate adaptive responses (25). The present studies showed that the ToxR PPD was dispensable for virulence gene activation, suggesting the possibility that ToxR may play a passive role in virulence activation. Such a function is supported by studies showing that ToxR functions in virulence activation by recruiting TcpP to the toxT promoter (36). This leads to the intriguing question of whether TcpP activation occurs after recruitment to the toxT promoter or if ToxR recruits already activated TcpP to the toxT promoter.

The reporter studies presented here revealed that multiple bile salts increased ctxA transcription when added to LB medium (Fig. 6B), but conjugated primary bile salts appeared to be the most potent ctxA activators, whereas secondary bile salts poorly activated ctxA expression. This phenomenon was also reflected in CT and TcpA production, where only primary bile salts supported virulence factor production when added to LB medium (Fig. 7). These results were similar to those of previous studies (15, 31) and suggest that V. cholerae has evolved to discriminate between bile acid species. This could have broad implications for V. cholerae pathogenesis as bile acid metabolism by the host microbiota results in regional differences in bile acid species in the gastrointestinal tract (32, 37). After primary bile acids are released into the proximal small intestine, they are transformed by the intestinal flora during transit through the intestine to generate secondary bile acids. These transformation reactions include deconjugation to form unconjugated bile acids and dihydroxylation to generate deoxycholate. Thus, the composition of bile acid species varies from the proximal to the distal intestine (32, 37). This suggests that specific bile acid species can be used by V. cholerae as regional cues to modulate adaptive responses. For example, the conjugated primary bile acids predominate in the upper intestinal tract and therefore likely serve as an environmental cue to activate virulence gene expression via TcpP following host entry. Likewise, a previous study showed that conjugated primary bile acids also activate the V. cholerae type VI secretion system (T6SS) (38). As the T6SS is thought to provide a competitive benefit to V. cholerae for intestinal colonization, bile acid sensing in the proximal intestine likely promotes colonization by multiple mechanisms. Interestingly, the latter study also found that the secondary bile acid deoxycholate, which predominates in the distal intestine, repressed the T6SS. Whether secondary bile salts have similar effects on virulence factor production in vivo has not been demonstrated. Deoxycholate has been shown to activate CT production in classical biotype strains through a ToxT-independent mechanism (39), but the addition of deoxycholate to AKI medium does not suppress CT or TcpA production O1 El Tor strain N16961 (J. E. Bina, unpublished results).

There was a disconnect between ctxA transcription and CT production following GCDCA and GCA treatment. While the addition of taurocholate to LB medium strongly activated ctxA expression and resulted in high-level CT and TcpA production, the addition of GCDCA to LB medium strongly activated ctxA expression but did not result in a correspondingly high level of CT or TcpA production. GCA activated ctxA transcription to AKI medium levels when present at 800 μM in LB medium but also did not support high-level CT or TcpA production. The discrepancy between ctxA expression and CT/TcpA production during GCDCA and GCA treatment suggests the possibility that individual bile salts may have additional effects on virulence gene expression downstream of TcpP activation. It remains unclear what these effects may be, but one possibility is that specific bile salts may influence the activity of secondary effectors on virulence factor production such as fatty acids (21, 40, 41), bicarbonate, and/or cyclic di-GMP (42–45).

MATERIALS AND METHODS

Bacterial strains, culture conditions, and plasmids.

Bacterial strains and plasmids used in this study are listed in Table 2. V. cholerae N16961 strain JB58 (ΔlacZ Sm^r) was used as the wild type (WT) in all experiments. Escherichia coli strain EC100λpir (Epicentre, Madison, WI, USA) was used as a host for plasmid amplification. All E. coli and V. cholerae strains were routinely grown in lysogeny broth (LB) (46) or on LB agar at 37°C, supplemented with antibiotics when needed. The induction of the ToxR regulon was accomplished by culturing V. cholerae strains under AKI conditions as follows. LB cultures of the test strains grown overnight were individually diluted (10⁻³) into 10 ml of AKI broth (15 g Bacto peptone, 4 g Difco yeast extract, and 5 g of NaCl per liter [pH 7.4]) in 150- by 15-mm glass test tubes (47). The inoculated test tubes were then incubated statically for 4 h at 37°C or until the optical density at 600 nm (OD₆₀₀) reached ≥0.08 before the cultures were transferred to 125-ml Erlenmeyer flasks and incubated with shaking at 37°C overnight for CT quantification. Antibiotics were used at the following concentrations when necessary: carbenicillin (Cb) at 100 μg/ml and kanamycin at 50 μg/ml. All the bile acid chemicals were purchased from Sigma-Aldrich (St. Louis, MO).

TABLE 2.

Strains and plasmids

Strain or plasmid	Description	Source or reference
Strains
E. coli
EC100λpir	F− mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80dlacZΔM15 ΔlacX74 recA1 endA1 araD139 Δ(ara leu)7697 galU galK λ− rpsL (Smr) nupG pir+	Epicentre
BTH101	F− cya-99 araD139 galE15 galK16 rpsL1 (Strepr) hsdR2 mcrA1 mcrB1	28
V. cholerae
JB58	O1 El Tor strain N16961; ΔlacZ Smr	49
SS4	JB58 toxRΔPPD	25
Plasmids
pXB192	toxT promoter fused to lacZ in pTL61T; Ampr	48
pXB193	ctxAB promoter fused to lacZ in pTL61T; Ampr	14
pXB194	tcpA promoter fused to lacZ in pTL61T; Ampr	14
pXB201	toxR promoter fused to lacZ in pTL61T; Ampr	14
pXB202	aphA promoter fused to lacZ in pTL61T; Ampr	14
pXB203	aphB promoter fused to lacZ in pTL61T; Ampr	14
pXB231	vexCD promoter fused to lacZ in pTL61T; Ampr	50
pXB266	leuO promoter fused to lacZ in pTL61T; Ampr	51
pTB18	Promoter probe vector for constructing transcriptional fusions to the lux operon; Ampr	35
pTB20	tcpP promoter fused to the lux operon in pTB18; Ampr	35
pTB17	toxT promoter fused to the lux operon in pTB18; Ampr	35
pTB23	ctxA promoter fused to the lux operon in pTB18; Ampr	35
pKD9	ompR promoter fused to lacZ in pTL61T; Ampr	18
pTL61T	Reporter plasmid for transcriptional fusions to lacZ	52
pUT18C-tcpP	N terminus of TcpP fused to the T18 fragment of adenylate cyclase in pUT18C; Ampr	15
pKT25-tcpP	C terminus of TcpP fused to the T25 fragment of adenylate cyclase in pKT25; Kanr	15
pUT18C	T18 fragment of adenylate cyclase under the control of the lac promoter; Ampr	28
pKT25	T25 fragment of adenylate cyclase under the control of the lac promoter; Kanr	28

Preparation of C₁₈-extracted medium.

Sep-Pak C₁₈ chromatography cartridges (Waters) were preconditioned with 10 ml of 100% methanol followed by 10 ml of sterile double-distilled water (ddH₂O) before 50 ml of growth medium was passed through the cartridge, and the flowthrough was collected, filter sterilized, and used as C₁₈-extracted medium. Thereafter, compounds retained on the C₁₈ columns were eluted with 10 ml of 100% methanol. The eluates were then concentrated by evaporation, resuspended in 50 ml of the indicated medium, and filter sterilized prior to use.

Transcriptional reporter assays.

V. cholerae cells harboring the indicated reporters were cultured under AKI conditions or in LB. At the indicated times, aliquots were collected in triplicate and assayed for reporter expression. β-Galactosidase activity was quantified as previously described (10) and is reported as Miller units (MU). Luminescence production was determined using a Berthold-Sirius (Bad Wildbad, Germany) luminometer and is reported as photons per second per OD₆₀₀ unit. All the transcriptional reporter experiments were performed independently at least three times.

CT and TCP production.

CT and TCP production was assessed by a GM1 enzyme-linked immunosorbent assay and TcpA Western blotting as previously described (48).

Analysis of bile salts in AKI medium, C₁₈-extracted AKI medium, and the C₁₈ column eluent.

Sep-Pak C₁₈ cartridges (Waters) were preconditioned with 10 ml of 100% methanol followed by 10 ml of sterile ddH₂O before 50 ml of AKI broth was passed through the cartridge, and the flowthrough was collected and used as C₁₈-extracted AKI medium. Molecules that were retained on the C₁₈ column following the passage of AKI medium were eluted from the column with 10 ml (5× concentrated) of 100% methanol and used as the C₁₈ eluent. Three independently generated replicates of AKI medium, C₁₈ chromatography column-extracted AKI medium, and the C₁₈ chromatography column AKI medium eluate were submitted to the University of Pittsburgh Health Sciences Metabolomics and Lipidomics Core (www.metabolomics.pitt.edu) for qualitative bile acid analysis by liquid chromatography–high-resolution mass spectrometry (LC-HRMS), as follows.

(i) Sample preparation.

Polar metabolite pool extraction was performed by the addition of deuterated (D₃)-creatinine, (D₃)-alanine, (D₄)-taurine, and (D₃)-lactate (Sigma-Aldrich) to the sample media as an internal standard for a final concentration of 10 μM. After 3 min of vortexing, the supernatant was cleared of protein by centrifugation at 16,000 × g. Three microliters of the cleared supernatant was subjected to online LC-MS analysis. Bile acid extraction was performed by adding 450 μl acetonitrile–3% HCl to 50 μl of the cleared supernatant, spiked with 5 μg/ml each of (D₄)-taurochenomuricholic acid (TcMCA), (D₅)-β-muricholic acid (bMCA), (D₄)-tauroursodeoxycholic acid (TUDCA), (D₄)-taurocholic acid (TCA), (D₄)-glycocholic acid (GCA), (D₄)-cholic acid (CA), (D₄)-ursodeoxycholic acid (UDCA), (D₄)-taurochenodeoxycholic acid (TCDCA), (D₄)-taurodeoxycholic acid (TDCA), (D₄)-glycoursodeoxycholic acid (GUDCA), (D₄)-chenodeoxycholic acid (CDCA), (D₄)-taurolithocholic acid (TLCA), (D₄)-glycolithocholic acid (GLCA), (D₄)-deoxycholic acid (DCA), (D₄)-glycochenodeoxycholic acid (GCDCA), (D₄)-glycodeoxycholic acid (GDCA), and (D₄)-lithocholic acid (LCA). After vortexing for 3 min, 300 μl was transferred to a clean tube and dried to completion under nitrogen gas before resuspension in 50 μl of a 1:1 solution of methanol-water. Three microliters was subjected to online LC-MS analysis.

(ii) Bile acid LC-HRMS method.

Analyses were again performed by online semitargeted LC-HRMS. Briefly, concentrated bile acid extracts were injected via a Thermo Vanquish ultrahigh-performance liquid chromatography (UHPLC) system and separated over a reversed-phase Phenomenex Luna C₁₈(2) column (2.1 by 100 mm, 5-μm particle size) maintained at 55°C. For the 20-min LC gradient, the mobile phase consisted of the following: solvent A (water–5 mM ammonium acetate–0.012% formic acid) and solvent B (100% methanol). The gradient was as follows: 0 to 0.4 min with 52% solvent B, an increase to 100% solvent B over 13.1 min, a hold at 100% solvent B for 2 min, and reequilibration at 52% solvent B for 5 min. The Thermo IDX tribrid mass spectrometer was operated in negative-ion mode, scanning in ddMS² mode (2 microscans) from m/z 100 to 800 at a resolution of 120,000 with automatic gain control (AGC) targets of 2e5 for full scans and 2e4 for MS² scans using high-energy collisional dissociation (HCD) fragmentation at stepped collision energies of 15, 35, and 50. The source ionization setting was 2.6 kV for the spray voltage. Source gas parameters were 45 for sheath gas, 12 for auxiliary gas at 320°C, and 3 for sweep gas. Calibration was performed prior to analysis using Pierce FlexMix ion calibration solutions (Thermo Fisher Scientific). Integrated peak areas were then extracted manually using Quan Browser (Xcalibur version 2.7; Thermo Fisher) and normalized to the internal standards.

Bacterial two-hybrid system to assess TcpP dimerization.

The TcpP two-hybrid system was generously provided by Jun Zhu (University of Pennsylvania) and used as previously described (15). Briefly, E. coli BTH101 cells (lacking cyaA) bearing pUT18C and pKT25 containing the TcpP periplasmic and membrane-spanning domains fused separately to the T18 and T25 domains of the B. pertussis adenylate cyclase gene were subcultured in the indicated media containing 0.5 mM isopropyl-β-d-1-thiogalactopyranoside. The cultures were then incubated at 30°C without shaking until they reached an OD₆₀₀ of ∼0.3, at which time they were assayed for β-galactosidase activity.