Host Cell Contact Induces Expression of Virulence Factors and VieA, a Cyclic di-GMP Phosphodiesterase, in Vibrio cholerae

Abstract

Vibrio cholerae, a noninvasive bacterium, colonizes the intestinal epithelium and secretes cholera toxin (CT), a potent enterotoxin that causes the severe fluid loss characteristic of the disease cholera. In this study, we demonstrate that adherence of V. cholerae to the intestinal epithelial cell line INT 407 strongly induces the expression of the major virulence genes ctxAB and tcpA and the virulence regulatory gene toxT. No induction of toxR and tcpP, which encode transcriptional activators of toxT, was observed in adhered bacteria, and the adherence-dependent upregulation of toxT expression was independent of ToxR and TcpP. A sharp increase in the expression of the vieA gene, which encodes a cyclic di-GMP (c-di-GMP) phosphodiesterase, was observed in INT 407-adhered V. cholerae immediately after infection. Induction of toxT, ctxAB, and tcpA in INT 407-adhered vieA mutant strain O395 ΔvieA was consistently lower than in the parent strain, although no effect was observed in unadhered bacteria, suggesting that VieA has a role in the upregulation of toxT expression specifically in host cell-adhered V. cholerae. Furthermore, though VieA has both a DNA binding helix-turn-helix domain and an EAL domain conferring c-di-GMP phosphodiesterase activity, the c-di-GMP phosphodiesterase activity of VieA is necessary and sufficient for the upregulation of toxT expression.

INTRODUCTION

Vibrio cholerae, a Gram-negative, noninvasive bacterium, is the causative agent of the diarrheal disease cholera (1). Cholera continues to be a major public health concern, especially in developing countries, in many of which it exists in endemic form and frequently assumes epidemic proportions. For successful infection of its human host, V. cholerae must colonize the intestine and produce several virulence factors, including cholera toxin (CT) and a toxin-coregulated pilus (TCP) that greatly enhances colonization of the intestinal epithelium by the bacterium (2, 3). The hierarchical expression of several regulatory proteins that make up the ToxR regulon controls the expression of a subset of virulence factors, including CT and TCP (4, 5). At the top of the hierarchy are ToxR and TcpP, which synergistically promote the transcription of the toxT gene (6, 7). ToxT, in turn, activates the expression of several virulence genes, including ctxAB and tcpA, which code for CT and TcpA, the major subunit of TCP, respectively (8). Under laboratory conditions, the ToxR regulon is maximally expressed in cells grown at 30°C in Luria broth (LB), pH 6.6, and an osmolarity equivalent to 66 mM NaCl. Paradoxically, physicochemical parameters thought to characterize the intestinal lumen, namely, a temperature of 37°C, an alkaline pH, and high osmolarity, repress the expression of the ToxR regulon in vitro (9, 10). Furthermore, bile, a major constituent of the small intestine, represses the expression of virulence factors in V. cholerae (11, 12, 13). These observations suggest that as-yet-unidentified signals encountered by V. cholerae in the intestine may overcome the repression of the ToxR regulon imposed by bile, an alkaline pH, and a temperature of 37°C.

Contact of bacteria with eukaryotic cells has been recognized as a signal capable of triggering the expression of specific bacterial genes. Interaction with host cells induces the expression of the yop genes in Yersinia pseudotuberculosis (14). In Escherichia coli, the interaction of the P pili with their host cell receptors has been shown to induce the transcription of the bar gene, which is essential for the response to iron limitation (15). Host cell contact-dependent transcriptional upregulation of components of the type IV pili has been reported in Neisseria meningitidis (16) and Actinobacillus pleuropneumoniae (17). In this report, we demonstrate that the major virulence genes of V. cholerae and the vieA gene, which encodes a cyclic di-GMP (c-di-GMP) phosphodiesterase (PDE), are strongly induced in bacteria associated with epithelial cell lines. The PDE activity of VieA was necessary for the host cell contact-dependent upregulation of virulence genes of V. cholerae.

MATERIALS AND METHODS

Bacterial strains, plasmids, oligodeoxynucleotides, and growth conditions.

The bacterial strains and plasmids used in this study are listed in Table 1. All of the DNA oligonucleotides used are listed in Table S1 in the supplemental material. Bacterial strains were grown with aeration in LB to the mid-exponential phase, washed with phosphate-buffered saline (PBS), suspended in Dulbecco's modified Eagle's medium (DMEM), and used to infect a confluent monolayer of INT 407 or another cell line as mentioned in Results. For estimation of virulence gene expression under in vitro virulence-inducing conditions, V. cholerae was grown in LB (pH 6.6) at 30°C to the logarithmic phase (optical density [OD] of approximately 0.5) for estimation of toxT expression or to the late logarithmic phase (OD of approximately 1.2) for ctxAB expression. For growth in LB, antibiotics were used where required at the following concentrations, streptomycin, 100 μg/ml; ampicillin, 50 μg/ml; kanamycin, 30 μg/ml; chloramphenicol, 3 μg/ml. In all clones constructed with pBAD vectors (Table 1), 0.2% arabinose was used to induce gene expression.

Table 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant characteristic(s)	Source or reference
V. cholerae strains
O395	O1, classical; wild type, Smr	Laboratory collection
O395 toxT::lacZ	Derivative of O395N1	18
O395 ctx::lacZ	Derivative of O395N1	Laboratory collection
JJM43	O395 ΔtoxR	3
O395 ΔtcpP	O395 ΔtcpP	19
O395 ΔvieA	Derivative of O395; Smr	This study
E. coli strains
DH5α λpir	F− (lacZYA-argF)U169 recA1 endA1 hsdR17 supE44 thi-1 gyrA96 relA1 λ::pir	Laboratory collection
SM10λpir	thi recA thr leu tonA lacY supE RP4-2-Tc::Mu λ::pir	Laboratory collection
Plasmids
pCVD442	oriR6K mobRP4 sacB, Apr	20
pVM7	pBR327carrying toxR gene	6
ptcpP	pACYC184 carrying V. cholerae tcpP gene	21
pvieA	pBAD 24 carrying V. cholerae vieA gene	This study
pAT1568	pBAD33::VieA-EAL-His6, Cmr	22
pAT1615	pBAD33::VieAEAL(E170A)-His6, Cmr	22
pAT1385	pCVD442::ΔvieA, Apr	23

Cell culture.

The human embryonic intestinal cell line INT 407 (obtained from NCCS, Pune, India) was grown in low-glucose DMEM with 0.37% (wt/vol) sodium bicarbonate, 10% (vol/vol) heat-inactivated newborn calf serum, and 1% (vol/vol) penicillin-streptomycin (5,000 U/ml penicillin G sodium salt and 5,000 μg/ml streptomycin sulfate) at 37°C in a 5% (vol/vol) CO₂ atmosphere. Hep-2, HeLa, and HT-29 cells were grown under identical conditions, except that high-glucose DMEM was used. Differentiation of HT-29 cells was performed as described previously (24), and alkaline phosphatase production was used as a marker of cell differentiation (25). All tissue culture reagents were obtained from Gibco (Invitrogen Corporation). For bacterial adherence assays, cells were grown in 25-cm² flasks.

Adherence assays.

INT 407 cells were grown to confluence in DMEM containing 10% newborn calf serum and washed thrice with PBS, and DMEM without serum was added. V. cholerae cultures grown to the mid-exponential phase in LB were suspended in DMEM as described above, added to the INT407 monolayer at a multiplicity of infection (MOI) of 50, and incubated for 45 min at 37°C in a 5% CO₂ atmosphere. Nonadhered bacteria in the culture supernatant were removed, the infected cell line was washed four times with PBS, and DMEM containing 10% serum was added (referred to as 0 h). After incubation for different periods of time, adhered bacteria were released by lysing the INT 407 cells with 1% Triton X-100 in PBS, followed by vigorous pipetting. The bacterial CFU count was determined by serial dilution and plating. Treatment with 1% Triton X-100 had no effect on the viability of V. cholerae. Assays of the adherence of V. cholerae to other cell lines were performed similarly. Parallel experiments were performed with V. cholerae under identical conditions but without a cell line. These cells were also washed, incubated in DMEM containing 10% serum, and finally treated with 1% Triton X-100. All experiments were performed in duplicate and repeated at least thrice. The data are expressed as means ± standard deviations. The statistical significance of differences in the data was analyzed with the Student t test, and P values of <0.05 were considered significant.

RNA isolation and RT-PCR.

RNA was extracted from unadhered and adhered V. cholerae with guanidium isothiocyanate as described previously (26). The RNA was treated with RNase-free DNase 1 (1 U/μg, amplification grade; Invitrogen) in the presence of an RNase inhibitor (RNasin; Invitrogen). For real-time quantitative reverse transcription (qRT)-PCR, One-Step SYBR green PCR Master Mix (TaKaRa) was used in accordance with the manufacturer's instructions. qRT-PCR was performed in triplicate (unless otherwise stated) for the genes of interest and the 16S rRNA gene in an iCycler IQ5 real-time PCR detection system (Bio-Rad). A dissociation curve was generated at the end of each cycle to verify that a single product was amplified with software provided with the system. The change in SYBR green fluorescence was monitored by the system software, and the threshold cycle (C_T) above the background for each reaction was calculated. From C_T values, the relative levels of expression of the genes of interest were calculated by the 2^−ΔΔCT method (27), with the 16S rRNA gene as an internal reference. The statistical significance of the differences observed was calculated with the two-sample t test. A P value of <0.05 was considered significant. Primers specific for ctxAB, tcpA, toxT, toxR, tcpP, and the 16S rRNA gene were used as described previously (28). The sequences of the primers used are given in Table S1 in the supplemental material.

Construction of a V. cholerae vieA deletion mutant.

An in-frame deletion of vieA was constructed in V. cholerae strain O395 with plasmid pAT1385 as described previously (23). The gene deletion was confirmed by PCR of genomic DNA with primers vieA3 and vieA4 (see Table S1 in the supplemental material) flanking the deleted region and sequencing. Phenotypic confirmation of the mutations was obtained by motility assay (see Fig. S1 in the supplemental material).

RESULTS

Virulence gene expression is induced in V. cholerae adhered to the intestinal epithelial cell line INT 407.

The expression of the major virulence genes ctxAB and tcpA was examined in V. cholerae following adherence to INT 407 cells and compared to that in unadhered bacteria grown under identical conditions without a cell line or isolated from the supernatant of an infected INT 407 monolayer. The V. cholerae O395 bacteria used to infect the INT 407 cell line were grown under nonpermissive conditions (LB, pH 8.6, 37°C) (6); hence, very little ctxAB or tcpA expression was detected in the bacteria immediately after infection (Fig. 1). The expression decreased further when the bacteria were incubated in DMEM without a cell line and in bacteria isolated from the supernatant of infected INT 407 monolayers. However, the expression of both ctxAB and tcpA increased progressively in INT 407-adhered bacteria and the expression of both of the genes was significantly higher in the adhered bacteria than in the unadhered control. After 5 h, quantitation by real-time RT-PCR indicated an approximately 12-fold increase in ctxAB expression and a 10-fold increase in tcpA expression in the adhered V. cholerae bacteria compared to the unadhered bacteria at 5 h (Fig. 1). Parenthetically, no difference in the expression of ctxAB and tcpA was observed in unadhered bacteria grown in DMEM without a cell line and those isolated from the supernatant of infected INT 407 monolayers (data not shown). Hence, in subsequent experiments, unadhered bacteria grown in DMEM without a cell line were used as unadhered controls.

Fig 1.

Virulence gene expression in V. cholerae adhered to INT 407 cells. RNA was extracted from INT 407-adhered V. cholerae at different times after adherence for estimation of ctxAB and tcpA expression by qRT-PCR. As a control experiment (labeled unadhered), ctxAB and tcpA gene expression was estimated in V. cholerae incubated under identical conditions without a cell line. The expression of each gene was normalized to that of the internal reference, the 16S rRNA gene. Relative expression levels were calculated as the expression of a gene divided by that in unadhered V. cholerae at the beginning of the experiment (0 h), which was arbitrarily defined as 1. Data presented are averages of five independent experiments, and error bars represent standard deviations. The significance of differences in the expression of the ctxAB or tcpA genes in unadhered and adhered V. cholerae cells was determined by Student's t test and is indicated as follows: **, P < 0.01; *, P < 0.05.

Strain O395 carrying the lacZ gene transcriptionally fused to ctxAB (O395 ctxAB::lacZ, Table 1) was also used to study the expression of virulence genes following adherence to INT 407 cells. The adherence of the strain to INT 407 cells was similar to that of parent strain O395 (see Table S2 in the supplemental material). β-Galactosidase activity was estimated in adhered bacteria and compared to that in unadhered bacteria. The results obtained indicated that ctxAB expression was upregulated about 9.5-fold upon adherence (Table 2), similar to that obtained by RT-PCR (Fig. 1). The expression of the housekeeping genes encoding the rRNA and the RNA polymerase β subunit (rpoB) and recA in adhered bacteria was comparable to that in unadhered bacteria (data not shown), suggesting that host cell adherence did not affect V. cholerae transcription in general.

Table 2.

Upregulation of ctxAB and toxT expression in INT 407-adhered V. cholerae

V. cholerae strain and time (h) postadherence	β-Galactosidase activitya (Miller units/108 CFU)		Fold changeb	LB (pH 6.6) at 30°C
	Unadhered	Adhered		Culture OD	β-Galactosidase activitya (Miller units/108 CFU)
O395 ctxAB::lacZ
3	780 ± 75	2,021 ± 145	2.7 ± 0.5	1.2	1,800 ± 100
5	407 ± 50	3,854 ± 533	9.3 ± 0.57
7	370 ± 19	1,025 ± 160	2.7 ± 0.43
O395 toxT::lacZ
3	490 ± 25	4,412 ± 335	9 ± 0.7	0.5	2,200 ± 500
5	785 ± 79	4,302 ± 200	5.2 ± 0.5
7	1,153 ± 157	3,177 ± 612	2.7 ± 0.23

^aResults of three independent experiments are presented as means ± standard deviations.

^bFold change indicates expression in adhered bacteria compared to that in unadhered bacteria.

Expression of the virulence regulatory gene toxT in INT 407-associated V. cholerae.

Since ToxT is the direct transcriptional activator of ctxAB and tcpA (8), both of which are upregulated in INT 407-adhered V. cholerae, toxT expression in adhered and unadhered V. cholerae at different times after adherence to INT 407 was estimated next (Fig. 2). RT-PCR analysis indicated that at 5 h postadherence, the expression of toxT in the adhered bacteria was about 7-fold higher than that in unadhered bacteria; thereafter, toxT expression declined (Fig. 2B).

Fig 2.

Effect of adherence on toxT expression. (A) Schematic indicating the positions of primers (P1/P2 and P3/P4), the toxT promoter (P_x), and the tcpA-tcpF promoter (P_A). The expression of toxT at different times after the adherence of V. cholerae O395 to INT 407 and growth in LB (pH 6.6, 30°C) was estimated by qRT-PCR with primers P1 and P2 (B) or P3 and P4 (C). toxT expression in unadhered V. cholerae at the beginning of the experiment (0 h) was arbitrarily defined as 1. Results of three independent experiments are presented as means ± standard deviations. The significance of differences in the expression of the toxT gene in unadhered and adhered V. cholerae cells was determined by Student's t test and is indicated as follows: ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Strain O395 toxT::lacZ was also used to infect INT 407 cells, and the amount of β-galactosidase produced by INT 407-associated bacteria was monitored at different times after adherence and compared to that produced by unadhered bacteria (Table 2). An increase in β-galactosidase activity was observed in INT 407-adhered bacteria, and the pattern of toxT induction in strain O395 toxT::lacZ adhered to INT 407 was similar to that of strain O395.

The in vitro conditions known to promote the maximum expression of virulence genes in V. cholerae classical biotype strain O395 are growth in LB (pH 6.6) at 30°C and are referred to as virulence-inducing conditions (6). Under these conditions, maximum toxT expression occurs in the logarithmic phase (OD of approximately 0.5), whereas ctxAB expression is maximum in the early stationary phase (OD of approximately 1.2). RT-PCR assays indicated that the amount of toxT produced by INT 407-adhered V. cholerae O395 was comparable to that produced by the bacteria during in vitro growth under virulence-inducing conditions (Fig. 2B). However, the toxT expression in INT 407-adhered strain O395 toxT::lacZ was about 2-fold higher than that measured under in vitro virulence-inducing conditions (Table 2). The reason for this discrepancy was not clear. It was noted that ctxAB expression in INT 407-adhered strain O395 toxT::lacZ was also about 2-fold higher than under in vitro virulence-inducing conditions (Table 2).

The toxT gene is located downstream of the tcpA-tcpF gene cluster. It has been demonstrated that the expression of toxT might occur from a promoter (P_X, Fig. 2A) that is immediately upstream of toxT and the ToxT protein thus produced activates ToxT-dependent transcription from the upstream tcpA promoter (P_A, Fig. 2A). This might read through a proposed terminator between the tcpF and toxT genes to result in continued toxT transcription (19). Whether toxT expression in adhered V. cholerae cells depends on the positive autoinduction of toxT from the tcpA promoter was examined. Since transcription initiation at the tcpA promoter and readthrough to toxT produce a transcript 105 bp longer than that initiated at the toxT promoter (19), RT-PCR was performed with a primer pair spanning a region from upstream of the toxT transcription initiation site to within the toxT coding sequence (toxT long-transcript-specific primers P3 and P4, Fig. 2A). Significant upregulation of the longer toxT transcript was observed in INT 407-adhered V. cholerae, although in unadhered bacteria, the amount of the long toxT transcript actually decreased up to 5 h, after which a small increase was observed (Fig. 2C). These results suggested that while toxT expression occurred primarily from the proximal promoter (P_X, Fig. 2A) in unadhered bacteria, a significant part of toxT transcription was initiated at the distal tcpA promoter (P_A, Fig. 2A) in INT 407-adhered V. cholerae.

Upregulation of toxT expression requires contact between V. cholerae and host cells.

Whether any component(s) secreted by INT 407 into the medium was responsible for the induction of virulence genes in V. cholerae was next examined. For this purpose, INT 407 cells were grown to confluence in complete DMEM, washed, suspended in fresh complete DMEM, and incubated for 5 h. The medium covering the cell line was collected, and V. cholerae cells grown to the logarithmic phase were suspended in the “conditioned” DMEM or in “fresh” DMEM and incubated in a 5% CO₂ atmosphere for 3 to 5 h. RNA was extracted from the bacteria and used to estimate the expression of the ctxAB, tcpA, and toxT genes by qRT-PCR. No significant difference in the expression of the virulence genes was observed in V. cholerae grown in “fresh” or “conditioned” DMEM (see Fig. S2 in the supplemental material). We also considered the possibility that adherence of V. cholerae to INT 407 cells might induce the secretion by the INT 407 cells of some component that is responsible for the induction of virulence genes in V. cholerae. However, no induction of toxT was observed in V. cholerae suspended in conditioned medium prepared from V. cholerae-infected INT 407 cells (see Fig. S2 in the supplemental material). From these results, it is reasonable to conclude that the induction of toxT expression in INT 407-associated V. cholerae is not due to a medium component or any component secreted by INT 407 cells and that induction requires the direct contact of bacteria with INT 407 cells.

Role of the upstream regulators ToxR and TcpP.

The expression of toxT requires the transmembrane DNA binding proteins TcpP and ToxR, which act synergistically to activate the expression of the toxT gene (4). Since toxT expression was higher in INT 407-associated V. cholerae, the expression of toxR and tcpP in adhered bacteria was examined. No significant difference in the expression of the toxR and tcpP genes in adhered and unadhered bacteria was observed (data not shown).

To investigate further if ToxR and TcpP have a role in the upregulation of toxT in INT 407-associated V. cholerae, toxT expression was examined in toxR mutant strain JJM43 (2) and tcpP mutant strain O395 ΔtcpP (19) following adherence to INT 407. In the unadhered state, toxT expression was about 36- and 48-fold lower in JJM43 and O395 ΔtcpP, respectively, than in the wild-type parent strain but was still detectable by qRT-PCR (Fig. 3). toxT expression was significantly greater in the mutant strains adhered to INT 407 cells than in unadhered mutant strains (Fig. 3). Complementation of strains JJM43 and O395 ΔtcpP with plasmids pVM7 (6) and ptcpP (21), carrying the cloned toxR and tcpP genes, respectively, restored toxT expression in both control and adhered bacteria (Fig. 3). These results suggested that although there is a requirement for ToxR and TcpP for the optimum expression of toxT under all conditions, the host cell adherence-dependent upregulation of toxT expression occurred independently of ToxR and TcpP.

Fig 3.

toxT expression in toxR and tcpP mutant strains. The expression of toxT in unadhered and INT 407-adhered strains O395, ΔtoxR (JJM43), ΔtoxR (JJM43)/ptoxR(pVM7), O395 ΔtcpP, and O395 ΔtcpP/ptcpP was examined at 5 h after adherence. toxT expression is expressed relative to that in unadhered strain O395, which was arbitrarily defined as 1. Data represent the means of three experiments, and error bars indicate the standard deviations. The significance of differences in the expression of the toxT gene in unadhered and adhered V. cholerae cells was determined by Student's t test and is indicated as follows: ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Induction of vieA in INT 407-adhered V. cholerae.

The VieSAB three-component system was identified as a regulator of virulence genes in V. cholerae, and it has been demonstrated that the VieA response regulator is required in the V. cholerae classical biotype for maximal expression of the ctxAB and toxT genes in M9 medium containing specific amino acids (29). The effect was due to reduction of the intracellular c-di-GMP concentration by the c-di-GMP PDE activity of VieA (22, 30). To examine whether VieA has a role in the upregulation of toxT expression in INT 407-adhered V. cholerae, vieA expression was first assayed in these bacteria. A sharp increase in the expression of vieA was observed immediately after adherence when vieA expression was about 10-fold higher in adhered bacteria than in unadhered bacteria (Fig. 4A). After 1 h, however, vieA expression in the adhered bacteria decreased although it continued to be about 2- to 5-fold higher in adhered than in unadhered bacteria examined up to 7 h (Fig. 4A).

Fig 4.

Role of VieA in virulence gene expression. (A) V. cholerae O395 was added to INT 407 monolayers at an MOI of 50 and incubated for 45 min, after which unadhered bacteria were removed from the culture supernatant and DMEM containing 10% serum was added to the infected cell line (referred to as 0 h). The expression of vieA in unadhered and INT 407-adhered V. cholerae O395 was estimated by qRT-PCR at different times after adherence. vieA expression in unadhered V. cholerae at 0 h was arbitrarily defined as 1. Results of three independent experiments are presented as means ± standard deviations. The significance of differences in the expression of the vieA gene in unadhered and adhered V. cholerae cells was determined by Student's t test and is indicated as follows: **, P < 0.01; *, P < 0.05. (B) The expression of toxT, ctxAB, and tcpA in INT 407-adhered strains was examined. The expression of the genes in unadhered O395 was arbitrarily defined as 1. Results of three independent experiments are presented as means ± standard deviations. The significance of differences in the expression of the toxT, ctxAB, or tcpA gene in unadhered and adhered V. cholerae cells was determined by Student's t test and is indicated as follows: ***, P < 0.001; **, P < 0.01; *, P < 0.05.

To investigate if VieA has a role in the induction of toxT and other virulence genes in INT 407-adhered V. cholerae, a O395 ΔvieA strain was constructed (Table 1). The expression of toxT, ctxAB, and tcpA was examined in the mutant strain following adherence to INT 407 cells and also in unadhered bacteria. Although the expression of these genes in unadhered bacteria of the wild-type and ΔvieA strains was similar, INT 407-dependent upregulation in adhered bacteria was consistently lower in the O395 ΔvieA mutant than in the parent strain (Fig. 4B). The overexpression of vieA from plasmid pVieA resulted in a significant increase in toxT, ctxAB, and tcpA expression in adhered bacteria, although little effect was observed in unadhered bacteria (Fig. 4B). Taken together, these results suggest that VieA has a role in the upregulation of toxT, ctxAB, and tcpA expression specifically in host cell-adhered V. cholerae.

The VieA protein contains a DNA binding helix-turn-helix (HTH) domain and an EAL domain conferring c-di-GMP PDE activity (30). To examine which activity of VieA is necessary for the regulation of toxT expression, plasmid pAT1568 carrying a fragment of the vieA gene (amino acids 110 to 386) that includes the EAL domain but not the HTH domain (Table 1), was introduced into wild-type O395 and the ΔvieA mutant strain. The expression of the VieA EAL domain increased toxT expression in INT 407-associated bacteria to a level comparable to that produced by the full-length vieA gene (Fig. 4B). However, the expression of the VieA EAL domain containing an E170A point mutation (pAT1615, Table 1) that destroys c-di-GMP PDE activity did not result in toxT upregulation (Fig. 4B). These results strongly suggest that toxT expression in INT 407-associated V. cholerae is upregulated by the reduction of c-di-GMP levels.

toxT expression in V. cholerae adhered to other cell lines.

It has been documented in early studies that V. cholerae adheres to many different cell lines. The effect of adherence to a variety of cell lines on toxT expression was examined next. V. cholerae O395 toxT::lacZ was allowed to adhere to Hep-2 and HeLa cells and the polarized cell line HT-29. HT-29 cells were differentiated under conditions that induce brush border formation (24), and both undifferentiated and differentiated, polarized HT-29 cells were used for adherence assays. V. cholerae adhered to all of the cell lines with efficiency comparable to that observed for INT 407 (data not shown). Furthermore, adherence to all of the cell lines resulted in the upregulation of toxT expression and maximum upregulation was observed in V. cholerae adhered to polarized HT-29 cells (Fig. 5). Interestingly, very little toxT expression was observed in V. cholerae adhered to a plastic surface (Fig. 5). These results indicate that adherence to eukaryotic cells, and not mere abiotic surface contact, is responsible for the induction of virulence genes in V. cholerae.

Fig 5.

toxT expression in V. cholerae adhered to different cell lines. V. cholerae strain O395 toxT::lacZ was used to infect the INT 407, Hep-2, HeLa, undifferentiated HT-29, differentiated HT-29 cell lines and an abiotic (plastic) surface under identical conditions. At 5 h postadherence, β-galactosidase activity was assayed in unadhered and adhered bacteria and expressed in Miller units per 10⁸ bacteria. Results of three independent experiments are presented as means ± standard deviations.

DISCUSSION

The expression of the ToxR virulence regulon of V. cholerae is strongly influenced by a number of environmental factors that exert their effects at different levels of a regulatory cascade. Temperature and pH affect the expression of the tcpP gene at the top of the hierarchy, while bile affects the cascade at a later stage by modulating the activity of ToxT (4, 5). In this study, we examined the expression of the ToxR regulon in V. cholerae following adherence to the intestinal epithelial cell line INT 407 and found an additional level of complexity to govern the expression of virulence factors. The expression of toxT was strongly induced in INT 407 cell-associated V. cholerae (Fig. 2). The induction required direct contact of the bacteria with INT 407 cells and was not due to a component secreted by the cell line.

What could be a possible mechanism for the increased toxT expression observed in adhered bacteria? Since numerous studies have established that the expression of toxT is controlled primarily by two synergistically acting activators, ToxR and TcpP, the role of these activators in the upregulation of toxT expression in INT 407-associated V. cholerae was investigated first. With toxR and tcpP mutant strains, it was demonstrated that host cell contact-dependent toxT upregulation per se was independent of ToxR and TcpP (Fig. 3). Next, the role of c-di-GMP was considered, in view of the fact that it has been elegantly demonstrated that alteration of the c-di-GMP concentration affects the expression of virulence genes and a low level of c-di-GMP promotes toxT expression (22). A remarkable increase in the expression of vieA, which encodes a c-di-GMP PDE, was noted immediately after the adherence of V. cholerae to INT 407 cells (Fig. 4A). Furthermore, although toxT expression in unadhered bacteria was similar in the wild-type and vieA mutant strains (data not shown), toxT induction in host cell-adhered vieA mutant strains was significantly lower than in the wild-type strain (Fig. 4B), suggesting that VieA has a role in the host cell contact dependent-upregulation of toxT. The VieA protein contains three conserved domains, an N-terminal phosphoreceiver domain, an EAL domain that has c-di-GMP PDE activity, and a C-terminal DNA binding HTH domain (30). We have demonstrated that the expression of the EAL domain is necessary and sufficient to upregulate toxT expression in adhered bacteria (Fig. 4B). From these results, it may be reasonable to conclude that VieA upregulates toxT expression in adhered bacteria by reducing the intracellular c-di-GMP concentration. V. cholerae genome contains about 30 genes with potential c-di-GMP PDE activity (31), and it is likely that other PDEs also contribute to a reduction of the c-di-GMP levels in adhered bacteria; this may account for the fact that toxT induction, though reduced, was still observed in adhered vieA mutant strains (Fig. 4B).

It has been demonstrated that toxT transcription may occur from two different promoters, one directly upstream of the toxT open reading frame and another further upstream that is activated by ToxT itself (19). Although the exact mechanism by which c-di-GMP levels control toxT expression is not known, it is attractive to hypothesize that the burst of vieA expression that occurs immediately after adherence might be sufficient to activate toxT expression from the proximal promoter. The ToxT protein thus produced promotes a positive feedback loop that activates its own transcription from the distal promoter (“long” toxT transcript), resulting in sustained toxT expression. This hypothesis is supported by the observation that although little “long” toxT transcript was detected initially, the amount increases with time in adhered bacteria but not in unadhered bacteria (Fig. 2C). The hypothesis can also explain why vieA expression diminishes after an initial burst in adhered bacteria (Fig. 4A).

Although it is now becoming increasingly clear that c-di-GMP is a ubiquitous second messenger that might link a variety of environmental signals to specific adaptive responses of bacteria, only a few signals that regulate c-di-GMP levels have been identified (32). These include oxygen (33), blue light (34, 35), nutrient starvation (36), antibiotics (37, 38) bile salts (39), mucin (40, 41), etc. The results described in this report indicate that host cell contact may be a signal that regulates c-di-GMP levels through activation of c-di-GMP PDEs in order to upregulate the expression of virulence factors.