Quorum Sensing Regulatory Cascades Control Vibrio fluvialis Pathogenesis

Abstract

Quorum sensing (QS) is a process by which individual bacteria are able to communicate with one another, thereby enabling the population as a whole to coordinate gene regulation and subsequent phenotypic outcomes. Communication is accomplished through production and detection of small molecules in the extracellular milieu. In many bacteria, particularly Vibrio species, multiple QS systems result in multiple signals, as well as cross talk between systems. In this study, we identify two QS systems in the halophilic enteric pathogen Vibrio fluvialis: one acyl-homoserine lactone (AHL) based and one CAI-1/AI-2 based. We show that a LuxI homolog, VfqI, primarily produces 3-oxo-C10-HSL, which is sensed by a LuxR homolog, VfqR. VfqR-AHL is required to activate vfqI expression and autorepress vfqR expression. In addition, we have shown that similar to that in V. cholerae and V. harveyi, V. fluvialis produces CAI-1 and AI-2 signal molecules to activate the expression of a V. cholerae HapR homolog through LuxO. Although VfqR-AHL does not regulate hapR expression, HapR can repress vfqR transcription. Furthermore, we found that QS in V. fluvialis positively regulates production of two potential virulence factors, an extracellular protease and hemolysin. QS also affects cytotoxic activity against epithelial tissue cultures. These data suggest that V. fluvialis integrates QS regulatory pathways to play important physiological roles in pathogenesis.

INTRODUCTION

Bacteria often exchange chemical signals to help them monitor their population densities through a phenomenon referred to as quorum sensing (QS) (1). The genus Vibrio includes more than 30 species, many of which are associated with human diseases, and have described QS systems for both interbacterial and intrabacterial communication (2). Among them, the acyl-homoserine lactone (AHL) system in Vibrio fischeri is well characterized and is used as a model system for many AHL-producing Gram-negative bacteria (3). The AHL signal molecule is produced by the AHL synthase LuxI and recognized by the LuxR receptor, leading to altered gene expression of downstream genes. In Vibrio cholerae, the causative agent of cholera (4), the major QS signal molecules are 3-hydroxytridecan-4-one (cholerae autoinducer-1 [CAI-1]) and AI-2 (5). Changes in these autoinducer levels correspond to repression or derepression of the major QS regulator, HapR. At a low cell density and signal concentration, a phosphorelay system is active, resulting in the phosphorylation of the terminal acceptor LuxO, a DNA-binding response regulator protein. Phospho-LuxO, together with a sigma factor σ⁵⁴, activates transcription of the genes encoding a set of small RNAs that, in conjunction with RNA chaperone, Hfq, bind to and destabilize the transcript of HapR, the master QS regulator (6). Alternatively, at high cell density, QS molecules interact with their cognate sensors, leading to the dephosphorylation of LuxO. Consequently, LuxO is inactivated and HapR is expressed (7, 8; see also two review articles [9, 10] for additional details). V. cholerae uses QS to control phenotypic factors critical to environmental survival, pathogenesis, and transmission (11–16). In addition to the AHL system, V. fischeri possesses a QS system similar to that of V. cholerae autoinducers, and these two regulate one another by positive feedback (17). Both AHL and CAI-1/AI-2 QS systems are present in other Vibrio species, including V. harveyi and V. anguillarum, although these two QS systems do not seem to interact. Lastly, V. vulnificus QS system is similar to that of V. cholerae (2).

Vibrio fluvialis is an emerging human pathogen of increasing public health significance. This is largely due to its food-borne nature ability to cause mild to moderate dehydration, vomiting, fever, abdominal pain, and diarrhea (18, 19). It was first isolated in 1977 from a patient suffering from severe diarrhea (18) and can be frequently detected in both human diarrheal stool samples and aquatic environments (20, 21). It has been reported that V. fluvialis elicits intestinal fluid when fed to suckling mice (22) and produces an array of virulence factors, including proteases, cytolysins, and toxins that elongate Chinese hamster ovary (CHO) cells (23, 24). However, regulation of the virulence factors is not well understood in V. fluvialis. Our recently completed draft sequence of a V. fluvialis clinical isolate (unpublished data) may shed some light onto the matter. Sequence analysis indicates that the V. fluvialis genome encodes two QS systems. In the present study, we confirmed the role of annotated QS genes in autoinducer production and regulation. We found that in V. fluvialis the CAI-1/AI-2 QS system negatively regulates the AHL QS system, unlike the positive regulation described for some other Vibrio species. We further demonstrated that QS systems in V. fluvialis play important roles in regulating several potential virulence factors.

MATERIALS AND METHODS

Bacterial strains, plasmids, and standard culture conditions.

V. fluvialis strain XJ85003 was isolated from a human with diarrhea in Xinjiang Province of China in 1985. The strain has been deposited in the Culture Collection of the Chinese Centers for Disease Control. A spontaneous streptomycin-resistant mutant of XJ85003 was used as the wild-type strain in all experiments in the present study. Unless noted otherwise, V. fluvialis strains were grown with aeration in brain heart infusion (BHI) broth at 37°C, and antibiotics were used in the final concentrations: streptomycin (100 μg/ml), ampicillin (100 μg/ml), kanamycin (50 μg/ml), chloramphenicol (5 μg/ml), and tetracycline (2 μg/ml). In-frame deletions of vfqI, vfqR, hapR, cqsA, luxS, luxO, and vfh were constructed by cloning the flanking regions of these genes into the suicide vector pWM91 containing a sacB counterselectable marker (25). The resulting plasmids were introduced into V. fluvialis by conjugation, and deletion mutants were selected for double homologous recombination events. Transcriptional fusion reporters were constructed by cloning promoter sequences of the genes of study (∼0.5-kb sequences upstream of the start codon) into pBBR-lux, which contains a promoterless luxCDABE reporter (26). Plasmids containing either P_BAD-vfqR or P_BAD-hapR were constructed by cloning vfqR (PCR amplified by using the primers 5′- CCGAATTCATGCAGAAAATTCTCCGTC -3′ and 5′- CGTCGACTCAGACGTAAGGATTAATG -3′) or hapR (using the primers 5′- CCGAATTCATGGACGCATCTATAGAG -3′ and 5′- GGCTGCAGTTAGTGATCGCGTTTATA -3′) coding sequences into pBAD24 digested with EcoRI/SalI or EcoRI/PstI, respectively (27). Then, 0.01 to 0.1% arabinose was added to the medium to induce P_BAD promoters. Information on all of the plasmids and oligonucleotides used here are available upon request.

Autoinducer production assays and AHL identification.

The levels of CAI-1 and AI-2 in culture supernatants were measured as described previously (8, 28), with minor modifications. Briefly, for CAI-1 production, 2.5% (vol/vol) culture supernatant was added to a 1:10 dilution of an overnight culture of MM920 (grown and diluted in Luria-Bertani [LB] medium) and incubated at 30°C for 3 h with agitation. The MM920 strain does not produce CAI-1 (ΔcqsA) and cannot sense AI-2 (ΔluxQ) (8). For AI-2 production, 10% (vol/vol) supernatant was added to a 1:5,000 dilution of an overnight culture of BB170 (grown and diluted in autoinducer bioassay medium) and incubated at 30°C for 6 h with agitation. The BB170 strain can only sense AI-2 (ΔluxN), so it detects exogenous AI-2 at a low cell density (29). In both cases, luminescence was measured with a Bio-Tek Synergy HT spectrophotometer. To detect AHL production and use thin-layer chromatography to examine AHL contents, the Agrobacterium tumefaciens KYC55(pJZ372)(pJZ384)(pJZ410) bioassay strain (30) was used according to a previously published protocol (31).

Purification and identification of AHLs produced by V. fluvialis were based on the protocol described previously (32, 33) with minor modifications. Briefly, a 1-liter cell-free culture fluid of V. fluvialis wild type was grown in LB medium to late log phase and extracted with ethyl acetate. The extract was dried and resuspended in 1 ml of ethyl acetate. Electrospray ionization-tandem mass spectrometry (ESI-MS/MS; Quattro Micro, Waters Micromass Co.) values were recorded. For all samples, the precursor ions of the mass 102 ion (P) were identified. The fragmentation spectra of all those precursor ions whose mass corresponded to those of AHLs were recorded. The AHL structure was then deduced according to the following formula:

$$\begin{array}{c}n=\frac{m/z-1-171}{28}{C}_{n\times 2+4}-\text{HSL};\\ n=\frac{m/z-1-187}{28}3-\text{hydroxyl}-C_{n\times 2+4}-\text{HSL};\\ n=\frac{m/z-1-185}{28}3-\text{oxo}-C_{n\times 2+4}-\text{HSL}\end{array}$$ (1)

Gene expression analysis.

The strains containing transcriptional lux fusion plasmids were grown at 37°C with appropriate antibiotics. Cultures were withdrawn at the time points indicated, and luminescence was measured using a Bio-Tek Synergy HT spectrophotometer and normalized against the optical density at 600 nm (OD₆₀₀). To determine whether V. fluvialis QS regulates the heterologous V. harveyi luxCDABE operon, the pBB1 cosmid, carrying the V. harveyi luxCDABE operon, was introduced into V. fluvialis strains by conjugation, and light production curves as a function of the cell densities were generated at 30°C as described previously (34).

Quantitative reverse transcription-PCR (qRT-PCR) was also used to examine gene expression in V. fluvialis. RNA was isolated from various V. fluvialis cultures by using an RNeasy minikit (Qiagen), and DNA was removed by using the DNA-free kit (Ambion). The cDNA was synthesized from 1 μg of RNA by reverse transcription with Superscript II reverse transcriptase (Invitrogen) according to the manufacturer's instructions. Controls lacking reverse transcriptase were included. qRT-PCR experiments were performed by CFX96 (Bio-Rad), using SYBR Premix Ex Taq (TaKaRa). All primer pairs amplified the target gene with efficiencies of 98 to 102% (data not shown). For each sample, the mean cycle threshold of the test transcript was normalized to that of recA.

Protease and hemolysin assay.

For protease production assays, V. fluvialis strains were incubated on 3% (vol/vol) skim milk plate overnight at 37°C (12). For hemolysin assays, wild-type and QS mutant strains were grown to mid-log phase at 37°C. Portions (5 μl) of concentrated cultures were then spotted onto Columbia blood plates and incubated for 48 h at 30°C.

Cytotoxicity assays.

To test the effects of V. fluvialis on host cells, lactate dehydrogenase (LDH) cytotoxicity assays were used. Epithelial HT29 cell lines were cultured in Dulbecco modified Eagle medium (DMEM), supplemented with 10% fetal bovine serum, 2.5 mM l-glutamine, and 100 μg of streptomycin/ml at 37°C with 5% CO₂. A total of 2 × 10⁵ cells were seeded in each well of 24-well plates and cultured overnight. The tissue culture medium was removed, and the cells were washed with DMEM three times before treatment. V. fluvialis strains were grown on three independent BHI agar plates overnight at 37°C, and bacterial cells then were resuspended in DMEM. Approximately 10⁷ bacterial cells were then added to the each tissue culture well and incubated for 6 h at 37°C with 5% CO₂. The LDH assays were performed using a CytoTox 96 nonradioactive cytotoxicity assay kit (Promega). The percent cytotoxicity was calculated by normalizing the LDH released when treated with 2% Triton X-100.

Nucleotide sequence accession numbers.

The DNA sequences of V. fluvialis QS genes reported in the present study have been deposited in the GenBank database under accession numbers KC153048 (vfqI), KC153049 (vfqR), KC153050 (cqsA), KC153051 (luxS), KC153052 (cqsS), KC153053 (luxP), KC153054 (luxQ), KC153055 (hapR), KC153056 (luxO), and KC153057 (luxU).

RESULTS AND DISCUSSION

V. fluvialis produces at least three types of QS molecules.

It has been reported that many Vibrio species produce multiple QS molecules (autoinducers) (35). To date, no study on V. fluvialis QS has been reported. We therefore first examined autoinducer production in V. fluvialis cell-free culture supernatants collected from different time points of growth. We used a V. cholerae bioassay strain to detect CAI-1 activity (8), a V. harveyi strain to detect AI-2 (29), and an A. tumefaciens strain to detect AHLs (30). As shown in Fig. 1A, V. fluvialis produced all three QS molecules in a cell-density-dependent manner: autoinducer concentrations were low in the culture supernatants collected from early time points, and the accumulation of autoinducers reached maximal levels in late-log-phase to early-stationary-phase cultures. CAI-1 and AI-2 produced by different Vibrio species are typically structurally identical, whereas AHLs produced by different bacteria differ in the length and modifications of the acyl side chain (9). We thus further determined the structures of AHLs produced by V. fluvialis by ESI-MS/MS (32, 33).Three AHL molecules were present in V. fluvialis culture supernatants (Fig. 1B and C): a major component, 3-oxo-C10-HSL, and two minor components C10-HSL and 3-oxo-C12-HSL. Two AHLs were also detected by using TLC analysis visualized by a bioassay reporter (30) (Fig. 1C, inset). These data suggest that in V. fluvialis at least three different types of QS molecules are produced: CAI-1, AI-2, and AHLs.

Fig 1.

Autoinducer production in V. fluvialis. (A) Accumulation of CAI-1, AI-2, and AHLs during V. fluvialis growth. Wild-type V. fluvialis was grown in BHI medium at 37°C. At the time points indicated, cell-free supernatants were assayed for the activities of CAI-1 (circles), AI-2 (squares), and AHLs (triangles) using the bioassay strains described in Materials and Methods and reported as the fold induction after being normalized against the BHI medium background. The data are means and standard deviations of three independent experiments. (B and C) Structural determination of AHLs produced by V. fluvialis. Cell-free supernatants from late-log-phase wild-type cultures were extracted and subjected to ESI-MS/MS. The retention times for all precursor ions whose masses corresponded to those of AHLs were recorded (B), and the mass spectrometry of the major peak is shown (C). The AHL structure was deduced as described in Materials and Methods (32, 33). The inset shows the results of a TLC analysis of V. fluvialis AHLs. C₁₈ reversed-phase TLC plates (Whatman) were used to separate AHLs and were detected by using overlaid agar medium containing cultures of an AHL bioassay strain and X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) (31). Synthetic AHL standards were purchased from Sigma-Aldrich.

VfqI/VfqR, the LuxI/LuxR homologs, are required for AHL production.

The LuxI/LuxR system of V. fischeri is considered as the QS model in Gram-negative bacteria (9), and homologs of luxI and luxR have been identified in a large number of bacterial genomes. Analysis of the V. fluvialis genome draft sequence identified LuxI and LuxR homologs, sharing 50 and 38% identity at the amino acid level, respectively. We therefore annotated them as VfqI (V. fluvialis QS I) and VfqR (V. fluvialis QS R). To determine whether VfqI and VfqR are functional in V. fluvialis, we first examined AHL production in vfqI and vfqR in-frame deletion mutants. Deletion of vfqI virtually abolished AHL production (Fig. 2A), indicating that VfqI is likely the AHL synthase responsible for AHL production in V. fluvialis. The vfqR mutant also produced a significantly reduced amount of AHLs (Fig. 2A), suggesting that VfqR is required to activate vfqI. The expression of vfqI and vfqR in trans in vfqI and vfqR mutants, respectively, restored AHL production (Fig. 2A, gray bars). To confirm that VfqR activates vfqI transcription, we examined vfqI expression using a transcriptional luminescent reporter in wild-type and vfqI and vfqR mutant strains. In this system, the expression of vfqI in vfqI mutants was significantly reduced, but the expression was rescued by the addition of crude AHLs extracted from wild-type culture supernatants (Fig. 2B). The vfqI expression was likewise low in vfqR mutants; however, supplementing crude AHLs in vfqR cultures could not restore vfqI expression. These data indicate that both VfqR and VfqI are required for vfqI expression and thus AHL production. It is likely that AHLs produced by VfqI serve as the ligand for VfqR, thus modulating its activity, perhaps in a similar manner to other LuxR-family proteins (36). Of note, the expression of vfqI was decreased 3-fold in vfqI and vfqR mutants compared to the wild type, but AHL production was reduced >100-fold. It is possible that this discrepancy results from a high basal level expression of the vfqI reporter plasmid. Interestingly, when we examined vfqR expression, we found that whereas vfqR expression was increased in both vfqI and vfqR mutants, vfqR expression was reduced by exogenous crude AHL only in vfqI mutants and not in vfqR mutants (Fig. 2C). These data suggest that in this AHL-based QS system, VfqR-AHL represses vfqR expression via a negative-feedback loop.

Fig 2.

Roles of VfqI and VfqR in AHL production and regulation. (A) AHL production. The wild type (WT; pBAD24) and vfqI (pBAD24), vfqI (P_BAD-vfqI), vfqR (pBAD24), and vfqR (P_BAD-vfqR) mutants of V. fluvialis were grown in BHI medium in the presence of 0.01% arabinose until late log phase. Cell-free spent media were subjected to AHL bioassay (31), and the β-galactosidase activity (traI-lacZ) was reported in Miller Unit (44). The data are the means and standard deviations for three independent experiments. (B and C) Expression of vfqI (B) and vfqR (C) mutants. The wild type and vfqI and vfqR mutants containing vfqI-luxCDABE (B) or vfqR-luxCDABE (C) plasmids were grown in BHI medium until mid-log phase. When indicated, 3% concentrated (100×) culture supernatants from the wild type were included in the medium. Luminescence was then measured and is reported as light units/OD₆₀₀. The data are means and standard deviations for three independent experiments.

CAI-1 and AI-2 produced by CqsA and LuxS regulate the expression of hapR.

CAI-1 and AI-2 QS autoinducer signals are present in V. cholerae and many other Vibrio species to regulate various physiological functions. In V. fluvialis genome, we identified the homologs of V. cholerae CqsA (CAI-1 synthase), LuxS (AI-2 synthase), LuxO (σ⁵⁴-dependent regulator), and HapR (QS master regulator, a TetR-family transcriptional regulator similar to LuxR of V. harveyi [37]). To test whether these gene homologs are functional in V. fluvialis, we examined CAI-1 and AI-2 production in various in-frame deletion mutants. As expected, deletion in cqsA and luxS abolished CAI-1 and AI-2 production, respectively. CAI-1 and AI-2 production were not affected by deletion of LuxO or HapR (Fig. 3A). We examined the relationship between hapR expression and CAI-1/AI-2, LuxO and HapR by measuring hapR-lux reporter expression in various mutants during growth. We found that hapR expression was reduced in cqsA-luxS double mutants (Fig. 3B), but not in cqsA or luxS single mutants (data not shown), suggesting that either CAI-1 or AI-2 is sufficient to activate hapR expression. The expression of hapR was higher in the luxO mutant, indicating that LuxO has a negative regulatory effect on hapR expression, similar to the autoinducer systems of V. cholerae and V. harveyi. Subsequent work will clarify the involvement, if any, of small RNA regulation of hapR expression in V. fluvialis, as seen in V. cholerae. Furthermore, compared to the wild type, hapR expression was elevated in hapR mutants at higher bacterial densities (Fig. 3B), a finding indicative of the involvement of an autorepression mechanism in the regulation of hapR expression. This regulation is likely to be direct, since hapR-lux expression in E. coli was significantly reduced when HapR was overexpressed (Fig. 3C).

Fig 3.

The effects of CAI-1/AI-2 components on autoinducer production and hapR expression. A. CAI-1 and AI-2 production. Wild type (WT) and mutant strains were grown in BHI medium until late log phase. CAI-1 (left panel) and AI-2 (right panel) production were measured as described in Materials and Methods, and the results are reported as the percentage of that produced in wild type. The data are means and standard deviations for three independent experiments. (B) hapR expression. Wild type and mutants containing hapR-luxCDABE plasmids were grown in BHI medium, the luminescence was measured, and the CFU were counted during growth. (C) Expression of hapR-luxCDABE in E. coli with a vector control or with a P_BAD-hapR in the presence or absence of 0.01% arabinose. Luminescence was then measured and is reported as light units/OD₆₀₀. The data are means and standard deviations for three independent experiments.

HapR negatively regulates AHL-based QS system by repressing vfqR expression.

To test whether there is cross talk between the AHL and CAI-1/AI-2 QS systems, we examined whether mutations in AHL systems affect hapR expression and vice versa. We found that the deletion of either vfqR or vfqI did not affect hapR-lux reporter expression at the early log, mid-log, or stationary phase of growth (data not shown), suggesting that the AHL QS system does not regulate the CAI-1/AI-2 system. On the other hand, deletion of hapR greatly affected the AHL QS system in mid-log-phase (Fig. 4) and stationary-phase (data not shown) growth: in hapR mutants, AHL production was significantly increased relative to the wild type (Fig. 4A). This is likely due to increased vfqI and vfqR expression in hapR mutants (Fig. 4A). In hapR mutants, repressor function was partially complemented when hapR was expressed under the control of a P_BAD promoter (Fig. 4A, black bars). To test the regulatory effect of HapR on vfqR, we measured vfqR-lux reporter expression in an E. coli in the presence of inducible P_BAD-hapR or vector control. As seen in Fig. 4B, vfqR was highly expressed in the absence of HapR or arabinose, whereas arabinose-induced HapR strongly repressed vfqR-lux. These results suggest that HapR may directly repress vfqR expression. It is also possible that HapR acts as a repressor through a shared component present in E. coli.

Fig 4.

HapR regulation of AHL-VfqR. (A) Effect of hapR deletion on AHL production and vfqI and vfqR expression. The hapR mutants (white bars) or hapR mutants complemented hapR in trans (black bars) containing either vfqI-luxCDABE or vfqR-luxCDABE plasmids were grown in BHI medium until mid-log phase. Then, 0.1% arabinose was added to induce P_BAD-hapR expression. AHL production and vfqI and vfqR expression were assayed and normalized against that of the wild type. The data are means and standard deviations for three independent experiments. (B) E. coli DH5α harboring a vfqR-luxCDABE plasmid and either pBAD24 (vector control) or P_BAD-hapR were grown in LB medium in the absence or in the presence of 0.01% arabinose until mid-log phase. Luminescence was then measured and is reported as light units/OD₆₀₀. The data are means and standard deviations for three independent experiments.

QS regulates potential V. fluvialis virulence factors.

To investigate targets regulated by these two V. fluvialis QS systems, we first introduced a plasmid harboring the V. harveyi luxCDABE operon into the wild type and QS mutants of V. fluvialis. It has been shown in both V. harveyi and V. cholerae that QS is involved in regulation of this operon (8, 12). Lux expression displayed a characteristic U-shape curve (Fig. 5A, filled dots), indicating that luxCDABE is regulated by QS in V. fluvialis. Moreover, as expected, Lux expression was abolished in hapR mutants, but not in vfqR and vfqI mutants (Fig. 5A), demonstrating that HapR is able to recognize and activate V. harveyi QS-regulated promoters, which share similarity with those in V. cholerae. Of note, the cell-density-dependent Lux induction in V. fluvialis is not as dramatic as in that of V. harveyi and V. cholerae (8). It is possible that the V. fluvialis HapR differs in certain residues so that it does not recognize the lux promoter as well as LuxR and V. cholerae HapR.

Fig 5.

QS-regulated target genes in V. fluvialis. (A) pBB1. Wild-type (WT) and mutant strains containing a plasmid harboring a V. harveyi luxCDABE operon were grown in BHI medium at 30°C. At the OD₆₀₀ indicated, the luminescence was then measured and is reported as light units/OD₆₀₀. The data are means and standard deviations for three independent experiments. (B) Extracellular protease production. Wild-type and mutant strains were streaked onto skim milk plates and incubated overnight at 37°C. (C) Expression of vfh and hemolysin production. The wild type and mutants were grown on BHI medium plates overnight, and the cells were collected. RNA was purified and subjected to qRT-PCR to measure the vfh expression. The data were normalized against recA mRNA level in each total RNA sample. The data are means and standard deviations for three independent experiments. For the lower part of the panel, wild-type and QS mutant strains were grown to mid-log phase at 37°C. Then, 5 μl of concentrated cultures was spotted onto Columbia blood plates, followed by incubation for 48 h at 30°C. (D) Cytotoxicity assays. Overnight cultures of wild-type and mutant strains were added to HT29 tissue culture cells at a multiplicity of infection of ∼100, followed by incubation for 6 h at 37°C with 5% CO₂. The lactate dehydrogenase (LDH) assays were performed using a CytoTox 96 nonradioactive cytotoxicity assay kit (Promega), and the results are presented as the percentage of LDH released from 2% Triton X-100 treatments. The data are means and standard deviations for three independent experiments. ***, P < 0.005; **, P < 0.05; *, P < 0.5; ns, no significance (Student t test).

It has been reported that extracellular proteases in V. cholerae may serve as detachases that are important for the release of bacteria from the intestinal surface during colonization (12, 38) and are important in a Caenorhabditis elegans model (39). To study QS-related extracellular protease production, wild-type and vfqR, vfqI, and hapR mutant strains were grown on milk agar plates. Only the hapR mutant was protease deficient on this selective medium (Fig. 5B), suggesting that protease production is regulated by CAI-1/AI-2 QS system in V. fluvialis. Analysis of genomic draft sequences revealed one metalloprotease gene annotated in V. fluvialis (data not shown). Future experiments will investigate the role of HapR in regulating this gene, as well as the role this protease plays in pathogenesis.

The extracellular hemolysin Vfh has been purified and characterized in V. fluvialis in previous work (40). We identified the gene encoding Vfh in V. fluvialis draft genome sequences, and deletion of the vfh gene abolished hemolysin production in V. fluvialis (Fig. 5C). There are additional two genes predicted to encode hemolysin-like proteins; however, deletion of these two genes did not affect the hemolytic activity of V. fluvialis (data not shown). In V. cholerae, QS negatively regulates hemolysin gene expression (41). To test whether QS controls hemolysin expression in V. fluvialis, we examined vfh expression in the wild type and QS mutants using qRT-PCR (Fig. 5C, top panel) and hemolysin production using Columbia blood agar plate assays (Fig. 5C, lower panel). We found that vfh expression and hemolytic activity were significantly reduced compared to the wild type in vfqI and vfqR mutants but were higher in hapR mutants. Since HapR negatively regulates vfqR expression (Fig. 4), these results suggest that vfh expression is activated by the AHL-VfqR QS system and repressed by the CAI-1/AI-2-HapR QS system. To further test whether QS affects cytotoxic activity of V. fluvialis on host cells, we applied the wild type and QS mutants to HT29 tissue cultures and performed LDH assays after incubation. We found that the cytotoxicity was decreased in both vfqI and vfqR mutants but not in hapR mutants (Fig. 5D). Cytotoxicity was strikingly reduced in vfh mutants, suggesting that hemolysin, or possibly another QS-regulated virulence factor, plays a major role in V. fluvialis cytotoxic activity. Interestingly, in vfh-vfqR double mutants, virtually no cytotoxic activity was detected (Fig. 5D, last column). These results suggest that in addition to activating vfh expression, the VfqR-AHL QS system also regulates the production of other extracellular factors contributing to cytotoxic activity.

In the present study, we discovered that there are at least two functional QS systems in pathogenic V. fluvialis: the CAI-1/AI-2-HapR system and the AHL-VfqR system. We identified QS regulatory components, including signal synthases, transcriptional regulators, and QS-regulated gene targets. These QS systems cross talk and employ feedback regulation mechanisms. For example, HapR represses vfqR expression (Fig. 4), and HapR and VfqR also negatively regulate their own expression (Fig. 2C and 3B). To our knowledge, V. fluvialis represents the only human pathogen studied to date that possesses both AHL and CAI-1/AI-2 QS systems. A model summarizing these interactions is shown in Fig. 6. Many bacteria have multiple QS regulatory systems that comprise an integrated hierarchical regulatory network (42). In the well-characterized V. fischeri and P. aeruginosa QS systems, one QS system positively regulates another QS system (17, 43). The physiological significance of the negative regulatory cascades in V. fluvialis QS systems is not yet clear; however, we hypothesize that it permits fine-tuning of gene regulation, allowing V. fluvialis to better adapt to surrounding environments, particularly during infection. For example, V. fluvialis causes cholera-like symptoms, but the illness is not as severe as cholera (19, 21). We speculate that V. fluvialis use the negative regulatory mechanisms to control virulence gene expression so that hosts survive the infection. This may maximally benefit propagation of V. fluvialis. Further in vivo study is required to better understand the relationship between V. fluvialis pathogenesis and QS regulation.

Fig 6.

Proposed working model for QS regulation in V. fluvialis. CAI-1/AI-2 signals activate the expression of HapR, which activates protease production and represses both hapR and vfqR expression. VfqI is an AHL synthase and synthesizes 3-O-C10-HSL, which may bind to VfqR as its cognate ligands. VfqR-AHL complex activates expression of vfqI, vfh, and other unknown genes. It also autorepresses its own expression.