|Home | About | Journals | Submit | Contact Us | Français|
Vibrio parahaemolyticus is an inhabitant of estuarine and marine environments that causes seafood-borne gastroenteritis worldwide. Recently, a type 3 secretion system (T3SS2) able to secrete and translocate virulence factors into the eukaryotic cell has been identified in a pathogenicity island (VP-PAI) located on the smaller chromosome. These virulence-related genes have previously been detected only in clinical strains. Classical virulence genes for this species (tdh, trh) are rarely detected in environmental strains, which are usually considered to lack virulence potential. However, during screening of a collection of environmental V. parahaemolyticus isolates obtained in the North Adriatic Sea in Italy, a number of marine strains carrying virulence-related genes, including genes involved in the T3SS2, were detected. In this study, we investigated the pathogenic potential of these marine V. parahaemolyticus strains by studying their adherence ability, their cytotoxicity, their effect on zonula occludin protein 1 (ZO-1) of the tight junctions, and their effect on transepithelial resistance (TER) in infected Caco-2 cells. By performing a reverse transcription-PCR, we also tested the expression of the T3SS2 genes vopT and vopB2, encoding an effector and a translocon protein, respectively. Our results indicate that, similarly to clinical strains, marine V. parahaemolyticus strains carrying vopT and vopB2 and that other genes included in the VP-PAI are capable of adhering to human cells and of causing cytoskeletal disruption and loss of membrane integrity in infected cells. On the basis of data presented here, environmental V. parahaemolyticus strains should be included in coastal water surveillance plans, as they may represent a risk for human health.
Vibrio parahaemolyticus is an halophilic inhabitant of estuarine and marine environments and a leading cause of seafood-borne gastroenteritis worldwide, frequently due to ingestion of uncooked shellfish.
The pathogenicity of V. parahaemolyticus classically has been correlated with production of haemolytic toxins TDH and TRH, the first toxin being responsible for the Kanagawa phenomenon (6, 8, 14, 17, 27). TDH causes a number of cytotoxic effects, including erythrocyte lysis, disruption of the microtubule cytoskeleton, ion influx into cultured cells, cell rounding, and disruption of epithelial barrier function (4, 7, 22). Less is known about the targets of TRH, although studies with the purified protein have shown that the toxin induces lysis of erythrocytes and fluid accumulation in the rabbit ileal loop model (6, 23). Kanagawa-positive V. parahaemolyticus strains carrying the tdh gene show a very high capability of adhering to human intestinal cells (5) and compromising the integrity of the epithelial barrier. The loss of membrane integrity, which can be monitored by measuring the transepithelial resistance (TER), may contribute to the diarrhea associated with V. parahaemolyticus infections, similar to the effects caused by other enterovirulent bacteria.
Some studies examining the contribution of TDH and TRH to V. parahaemolyticus virulence have focused on effects induced by the two individual purified toxins that cause drastic disruption of the epithelial barrier function (4, 16, 22, 23). However, Lynch et al. (11) reported that destruction of the epithelial barrier can occur independently of toxin production, as shown from studies conducted on strains lacking the tdh and trh genes. In addition, some strains lacking classical virulence factors caused a decrease in TER values in eukaryotic cells, accompanied by an increase in paracellular permeability associated with dramatic disruption of the actin cytoskeleton. In the same report (11), it was demonstrated that purified TDH protein was not able to alter the epithelial barrier. Lynch and collaborators concluded that previous reports indicating a role for TDH in cytoskeleton disruption resulted from use of elevated toxin concentrations, seemingly higher than those encountered in the bacterial culture supernatant (11). A role for TDH and/or TRH was not excluded, but it was suggested that these toxins may exert their effects at later stages of infection. These data, together with other data obtained in recent years (12, 19, 28), indicated that TDH or TRH is not the only bacterial factor determining V. parahaemolyticus virulence and that other factors might contribute to the pathogenic potential directed to human hosts. A number of these factors have been subsequently detected and well characterized (10, 12, 20, 28).
Genome sequencing of the V. parahaemolyticus clinical strain RIMD2210633 revealed that this bacterial species possesses two sets of type 3 secretion system (T3SS) genes: T3SS1, located on chromosome 1, and T3SS2, located on chromosome 2 (12). T3SS1 genes display high similarity with those from Yersinia T3SS, while T3SS2 genes are only partially similar to other bacterial secretion systems reported to date (20). Following this discovery, a number of studies have been conducted to elucidate the role of these genes in the pathogenicity process (1, 9, 10, 20). Using in vitro human cell lines and ileal loop tests, it was shown, using clinical strains, that T3SS1 is responsible for the cytotoxicity of V. parahaemolyticus for eukaryotic cells (20) while T3SS2 is involved in cytotoxicity and enterotoxicity (9, 20).
A good model for in vitro studies on the intestinal barrier is represented by the human intestinal Caco-2 cell line, which has been used to test the cytotoxicity caused by V. parahaemolyticus. Although the cytotoxicity of this bacterial species to human cells has been clearly proven (22), the virulence factors involved in this phenomenon have not been fully described. A V. parahaemolyticus mutant strain with both copies of tdh deleted, which completely abolished the hemolytic activity, still showed fluid accumulation in the rabbit intestine, suggesting that an unknown virulence factor(s) may be involved in the toxicity caused by V. parahaemolyticus (21).
Until recently, no data were available on secretion system genes in environmental V. parahaemolyticus and those strains were assumed to lack the T3SS2 system (12). For this reason, the recently reported finding that some environmental strains can carry a T3SS2 system (3) represents an interesting discovery regarding the potential virulence of marine bacterial strains. In the current study, we investigated the pathogenic potential of environmental V. parahaemolyticus strains isolated from the Venice Lagoon area in the Northern Adriatic Sea (Italy) by analyzing their ability to adhere to eukaryotic cells, the damage they might cause to the cell structures, and their cytotoxic effect on these cells.
A collection of 59 environmental V. parahaemolyticus strains isolated from the North Adriatic Sea (Italy) was analyzed in this study concerning the pathogenicity potential.
The entire collection was screened for adherence and cytotoxicity (Table (Table1).1). A subset of six strains (see Table Table11 and below), representing the different classes of environmental strains and carrying specific virulence-related genes, were selected and used to further evaluate the pathogenic potential of the marine strains (tests on transepithelial resistance and cell structural damage). Two of these strains possessed the tdh and orf8 genes that have been considered markers of the pandemic O3:K6 V. parahaemolyticus strain (15). These two strains also contained a sequence variation in the toxRS gene known as toxRS/new that is also considered to be a marker for pandemic strains (18). The six strains subjected to the more extensive characterization were as follows.
(i) Two strains (VPeVEpan and VPeVEpan2) carrying pandemic markers tdh and orf8 (13, 15) and the toxRS/new gene (18) and isolated from water and plankton, respectively, at the site Caleri in the Venetian Lagoon (2).
(ii) Two strains (VPe23 and VPe28, both tdh negative and trh negative) representing strains containing T3SS2 genes such as vopB2 and vscC2, encoding translocation proteins (10, 26), vopT and vopP, encoding effector proteins (9, 26), and other genes located on the pathogenicity island (VP-PAI), such as vopC, encoding a gene homologous to the Escherichia coli cytotoxic necrotizing factor, and VPA1376, encoding a protein homologous to Vibrio cholerae VPI protein (12).
(iii) Two strains (VPe93 and VPe122, negative for tdh, trh, and T3SS2 genes) representative of strains not carrying the considered virulence-related genes.
Bacteria were grown in tryptic soy broth (TSB) (Difco Laboratories, Detroit, MI) supplemented with 1% NaCl at 37°C, unless otherwise indicated. Special growth conditions for the different types of experiments are described in the following protocols. Cell growth was monitored with an LKB spectrophotometer at a 640-nm wavelength (optical density of 640 nm [OD640]).
Experiments were conducted with bacterial strains grown in two different media, Luria-Bertani (LB) broth supplemented with 3% NaCl and Dulbecco's modified Eagle medium (DMEM) supplemented with 5% fetal bovine serum. For RNA extraction and DNase treatment, the RNase minikit (Qiagen) was used. A quantity of extracted RNA, ranging between 2.5 and 5 μg, was reverse transcribed using Superscript III First-Strand synthesis system (Invitrogen, according to the manufacturer's instructions). For each reverse transcription (RT) reaction, an RT-negative (RT−) control including all components except the SuperScriptIII RT enzyme was prepared. The first-strand cDNA obtained was amplified directly by PCR, using the Easy-A high-fidelity PCR cloning enzyme (5 U/μl).
Bacterial cells were grown in brain heart infusion broth containing 0.5% mannitol and 3% NaCl (BHIM3) and incubated at 37°C without shaking until an optical density of ca. 0.8 to 0.9 (A620) was obtained. For these experiments, HeLa cells were grown in monolayer in Dulbecco's modified Eagle medium supplemented with 10% fetal bovine serum and 1% mannose (DMEM/F12) and used at passage level 22 to 25. After the HeLa cells were washed with phosphate-buffered saline (PBS), bacterial cells were added to the medium at a multiplicity of infection (MOI) of 10:1, mixed, and incubated at 37°C for two different periods of time, 30 and 60 min. After incubation, cells were washed three times with PBS, fixed in 70% methanol for 10 min, and Giemsa stained for 30 min. Cells were washed three times with PBS and dried. Finally, the coverslips were removed from the wells, assembled on the microscope slides, and examined microscopically to enumerate adherent bacteria.
For each bacterial strain tested, 100 HeLa cells were observed, on average, at the microscope and the number of bacteria adhering to each of the eukaryotic cells was counted. At least two separate experiments were conducted in duplicate, and three different examiners read the slides. After the averaged number of adhered bacteria/cell was calculated for each one of the strains, the adherence index was determined by referring to one of the four defined ranges: 11 to 15 bacteria/cell, 7 to 10 bacteria/cell, 3 to 6 bacteria/cell, and 0 to 2 bacteria/cell.
For cytotoxicity assays, three separate cultures of bacterial cells were grown in TSB-1% NaCl, pH 8.2, at 37°C overnight with shaking. Caco-2 cells were seeded at 2 × 104 cells/well in 96-well plates and cultured for 48 h to confluence. Cells were used between passages 18 and 24. Three different Caco-2 cell monolayers were cocultured with PBS-washed bacteria at a multiplicity of infection (MOI) of 10:1 for 5 h. The release of lactate dehydrogenase (LDH) into the medium was quantified using a Cytotox96 nonradioactive cytotoxicity kit (Promega), following the manufacturer's instructions. The results represent the means of six independent determinations, with bars showing one standard deviation.
Because different types of target cells release different amounts of LDH, a preliminary experiment was performed comparing LDH released from Caco-2 cells and other cell lines (T84 intestinal cells, HCT-8 cells, HeLa cell line), using different concentrations of human cells to determine the most appropriate conditions for in vitro study.
Transwell (12-well polycarbonate transwell filter support, 0.4-μm pore size, and 12-mm diameter; Corning Inc., NY) filters were treated with sterile type I collagen (10 μl of collagen, 0.5 mg/ml on each coverslip) and sterilized under UV light overnight before Caco-2 cells were added to the apical compartment of each well. Caco-2 cells were seeded at an initial density of 8 × 104 cells/cm2 at passage level 18 to 24 in DMEM supplemented with 10% fetal bovine serum, 100 U ml−1 penicillin, 100 mg ml−1 streptomycin, 1× nonessential amino acids (NEAA), and 1 mM sodium pyruvate. The bacterial cells were added gently drop by drop (final volume, 500 μl), and 1.5 ml of the same medium was added to the basolateral compartment. The cells were incubated at 37°C in a 5% CO2 atmosphere until polarized monolayers were formed, approximately 21 days, and the medium was changed every second day. After 10 days, the TER was tested using an EVOM-G ohmmeter (World Precision Instruments).
The TER was successively tested after 21 days in the first experiment, 22 days in the second experiment, and 25 days in the third experiment. Only wells displaying baseline resistance readings greater than 350 Ω/cm2 were used for experiments. Before the experiments were started, the instrument was prepared so that the resistance range was approximately 2,000 Ω in order to have readings of 100 Ω, 200 Ω, etc. In every experiment a blank (Transwell without cells) was used and measured hourly. A volume of 2 μl of V. parahaemolyticus overnight culture was added to the upper chamber of each well, and TER measurements were obtained hourly. Also, the TER of two transwells containing uninfected cells was measured hourly. For this experiment, strains were grown overnight in LB broth supplemented with 3% NaCl at 37°C with shaking.
Caco-2 cells were seeded at an initial density of 8 × 104 cells/cm2 onto type I collagen-coated (0.5 mg/ml) coverslips and incubated at 37°C in 24 wells in the presence of 5% CO2 for 5 days in high-glucose medium (DMEM supplemented with 5% fetal bovine serum, 1× NEAA, and 1 mM sodium pyruvate). Type I collagen-coated coverslips were previously sterilized by overnight incubation under UV. The medium was removed from the wells, and each well was gently washed with sterile phosphate-buffered saline (PBS) three times and filled with 500 μl of medium. Cells were infected with 2 μl of an overnight culture of bacteria grown in LB broth supplemented with 3% NaCl at 37°C with shaking. Cells were incubated at 37°C for two different periods of time, 3 h and 6 h. Following infection, coverslips were washed three times with PBS for 5 min and fixed in 2.5% paraformaldehyde for 10 min at 37°C, after which the cells were rinsed with PBS supplemented with 0.5% Triton X-100 three times before being blocked for 30 min with 1% normal goat serum in 500 μl PBS/well, supplemented with 0.1% Triton X-100. The cells were incubated for 45 min with rabbit anti-zonula occludin protein 1 (ZO-1) (N-term) (diluted 1:80; Zymed, San Francisco, CA) in PBS supplemented with 1% normal goat serum and 0.2% bovine serum albumin. The cells were washed three times with PBS and labeled with Alexa-Fluor 568 goat anti-rabbit antibody (diluted 1:80; Molecular Probes) in PBS supplemented with 1% normal goat serum and 0.2% bovine serum albumin for an incubation period of 30 min. Cells were washed three times with PBS and labeled to detect actin protein with Alexa-Fluor 488 phalloidin (diluted 1:200; Molecular Probes) in PBS for an incubation period of 20 min. The coverslips were washed with PBS three times for 5 min, dried, coated with ProLong antifade kit, and dried overnight covered by foil. Finally, samples were visualized using an inverted microscope (Nikon Eclipse TE 2000-E).
The results were statistically analyzed with analysis of variance (ANOVA) at the P < 0.05 significance level with the SPSS software. Data corresponding to adherence and cytotoxicity assays were also compared using the Student's t test with a significance level of P < 0.01.
In a recent study, we demonstrated the presence of T3SS2 and other virulence-related genes in a number of environmental V. parahaemolyticus strains isolated in the North Adriatic Sea (Italy) in the area of the Venetian Lagoon (3). In addition to T3SS2 genes (vopB2, vscC2, vopT, and vopP) and genes vopC and VPA1376 encoding proteins homologous to E. coli and V. cholerae virulence factors, the virulence genes included in the environmental strain screening were trh and ure encoding TRH and urease, respectively, and the virulence-related open reading frame (ORF) gene VP0394, carried on a 22-kb pathogenicity island-like element located on chromosome I and whose predicted protein product has homology to a DNA methyltransferase protein (28). Since the presence of a gene in a bacterial genome does not ensure that this gene is functional, it is necessary to demonstrate gene transcription by detecting the corresponding mRNA using reverse transcriptase PCR or other methods.
In a previous report (3), the transcription of the genes vopC and vcsC2 was tested. In the present study, we focused on the expression of two other T3SS2 genes, vopT and vopB2, which encode proteins that function as effector and translocon, respectively, as recently described (10). These two genes have been detected in the two marine pandemic strains and 12 environmental strains carrying only the secretion system. The transcription of these two genes was evaluated in a V. parahaemolyticus strain carrying the genetic pandemic markers tdh, orf8, and toxRS/new and some T3SS2 genes (VPeVEpan), in the strain VPe23 carrying only the T3SS2 genes, and in the reference pandemic strain QM97097 (see Table Table1).1). The gene gyrB was included in the study as a constitutively expressed housekeeping gene.
As demonstrated previously (3) for genes vopC and vcsC2, the results obtained indicate that vopT and vopB2 are transcribed in the three V. parahaemolyticus strains, under our experimental conditions, as shown by RT-PCR analysis and resulting in production of amplified corresponding cDNA (Fig. (Fig.11).
An initial step in the pathogenic process for many bacteria is adherence to the host cell before delivering toxins or causing damage to the cell structure. Earlier studies reported the capability of Kanagawa phenomenon (KP)-positive strains to adhere to intestinal cells (5). Therefore, we investigated whether environmental strains from our study were capable of adhering to human cells.
Preliminary experiments were performed to identify the optimal infection period that allowed bacterial adherence to human cells. An infection time of 60 min was found to be optimal, since a longer period of incubation resulted in nonspecific attachment to the coverslips by most of the strains, making it difficult to evaluate the specific capability to adhere to human cells.
Evaluation of adherence capability of the strains was based on an arbitrary designed adhesion index, considering that maximum adherence (very good adherence, 11 to 15 bacteria/cell) corresponded to the capability to adhere to eukaryotic cells of the reference strain QM97097 (KP-positive clinical strain). Other adhesion ranges were defined as good adhesion (7 to 10 bacteria/cell), medium adhesion capability (3 to 6 bacteria/cell), and poor or no adhesion capability (0 to 2 bacteria/cell).
Of the 59 environmental strains tested for adherence to HeLa cells, 13% showed very strong adherence, 30% good adherence, 48% medium adherence, and 8% poor or no adherence. Among strains showing very strong or good adherence, 54% carried virulence-related genes, while among strains showing medium or poor/no adherence, only 37% carried virulence-related genes (Table (Table11).
Results for strains representative of the entire screened collection are shown in Fig. Fig.2.2. As seen in this figure, the strain not carrying virulence-related genes (VP393) showed poor adherence to HeLa cells, compared to the strain carrying the T3SS2 genes (VPe23), which showed good capability to adhere to human cells. This effect was more evident in the VPeVEpan strain, which showed an adherence capability similar to that of the reference clinical strain QM97097. Adherence phenotypes shown by the other strains are presented in Table Table1.1. It should be noted, however, that some strains lacking virulence genes showed good adherence (Table (Table1),1), indicating that adhesion is a necessary but not sufficient condition for bacteria to succeed in the infection process. The results indicate that there is not a strict correlation between presence or absence of the genes tested in this study and adherence.
The cytotoxicity analysis was performed by measuring release of lactate dehydrogenase (LDH) by Caco-2 cells infected with the different environmental strains (LDH is released upon cell lysis).
In order to use a consistent MOI (multiplicity of infection) index of 10 bacteria:1 cell in each experiment, standard growth curves were prepared.
Results shown in Fig. Fig.33 indicate that, compared to the reference strain, the highest cytotoxic activity was detected in the strain VPeVEpan containing tdh and T3SS2 genes, a lower but still relevant activity in strains carrying T3SS2 but not tdh (VPe23), and very low cytotoxic activity in strains not carrying virulence-associated genes (VPe93). The results represent the means of six independent determinations, with bars showing one standard deviation. As shown in Table Table1,1, the same experiment conducted with strains belonging to the different groups gave similar results, without statistically significant differences (Student's t test), compared with the corresponding representative strains presented in the figure.
The effects of infection by environmental V. parahaemolyticus strains on epithelial barrier integrity were examined by measuring TER (transepithelial resistance) across a Caco-2 cell monolayer.
As shown in Fig. Fig.4,4, at time zero (T0), cells infected with any of the strains tested gave epithelial resistance values of ca. 400 Ω. During the first hour of infection (T1), the values decreased dramatically (70 to 80 Ω) in all cases except uninfected cells (negative control) and cells infected with strains lacking virulence-related genes, for which the values slightly decreased (only by a few ohms). The TER values of uninfected cells remained fairly constant, showing a slight decrease of about 100 Ω over 7 h (Fig. (Fig.4),4), and cells infected with strains lacking virulence-related genes decreased slightly more, about 150 Ω over 7 h. Strains VPeVEpan and Vpe23 (both containing T3SS2) showed a gradual decrease after the first hour, which became more marked in the VpeVEpan strain by the third hour of infection (T3) and then remained roughly constant during the rest of the experiment. The KP-positive reference strain QM97097 induced a gradual decrease in resistance value until T6 (about 4 h of infection), and then resistance dropped dramatically for the remaining 3 h.
Alterations in TER are often accompanied by changes in localization of the tight junction proteins. The effect of infecting Caco-2 cells with environmental V. parahaemolyticus strains carrying virulence-related genes on localization of ZO-1 (peripheral protein) and filamentous actin was examined using immunofluorescence microscopy.
In uninfected cells, the protein ZO-1 and filamentous actin were located at the periphery of the apical side of the cell monolayer; the location of both proteins revealed the integrity of the cells showing intact and relaxed borders (Fig. (Fig.55 A). The same localization of ZO-1 and filamentous actin was observed in cells infected with strains lacking virulence-related genes (Fig. (Fig.5B).5B). No disruption of the cell monolayer was observed, confirming the absence of cytotoxicity in cells infected with these strains. The only detectable effect on the cells was in border shape, which assumed a curly profile. In contrast, when cells were infected with the KP-positive pandemic strain used as a positive reference strain, the cell monolayer was disrupted and presented a discontinuity in the peripheral borders of the cells, and a complete rearrangement of the cytoskeleton was seen (Fig. (Fig.5D).5D). This result was expected, since the capability of strains producing both TDH and TRH toxins to disrupt the cell barrier is well established.
Infection of cell monolayers with environmental strains carrying T3SS2 genes (VPe23 and VPe28) resulted in a clear disruption of ZO-1 and filamentous actin (Fig. (Fig.5C).5C). The disruption was even stronger when the infection was induced by the VPeVEpan strain carrying both T3SS2 and tdh, compared to that induced by strains carrying only T3SS2; the cytoskeleton completely lost integrity, and drastic rearrangement of actin was observed (Fig. 5E and F). Moreover, in experiments performed with strains carrying virulence-related genes, most of the cells detached from the coverslip, indicating cytotoxicity caused by the infecting bacterial cells and the capability of these bacteria to produce virulence proteins.
The pathogenicity of V. parahaemolyticus classically has been correlated with production of hemolytic toxins TDH and TRH, the first toxin being responsible for the Kanagawa phenomenon.
A number of data obtained in recent years (12, 19, 28) indicated that TDH or TRH is not the only bacterial factor determining V. parahaemolyticus virulence and that other factors might contribute to the pathogenic potential directed to human hosts. Several of these factors have been successively detected and well characterized (10, 12, 20, 28), as is the role of secretion systems in mediating the delivery of bacterial virulence proteins to eukaryotic cells. In V. parahaemolyticus it has been demonstrated that at least two sets of T3SS genes exist: T3SS1, located on chromosome 1 and responsible for the cytotoxicity of the bacterial species for eukaryotic cells (20), and T3SS2, shown to be involved in cytotoxicity and enterotoxicity (9, 10, 20). Other genes have been proposed as being associated with the V. parahaemolyticus pathogenic potential, e.g., urease and methyltransferase (3).
To evaluate if a direct correlation exists between the presence of virulence genes, such as tdh, trh, the T3SS2 genes, and other virulence-associated genes, and the potential capability of environmental strains to cause infection in humans, the effective pathogenic potential of some marine V. parahaemolyticus strains isolated from the Italian coastline was analyzed by studying the in vitro interaction with human cells, including adherence capability, cytotoxicity, and the ability to induce structural alterations in infected cells.
An initial step in this study, the transcription of T3SS2, was confirmed via RT-PCR, showing that two T3SS2 genes, vopT, encoding a homolog of the Pseudomonas exoenzyme T and contributing to cytotoxicity, and vopB2, encoding a translocator involved in the contact-dependent activity of pore formation in infected cells, are expressed under standard growth conditions. Transcription of other T3SS2 genes carried by strains from the same environmental V. parahaemolyticus collection was previously reported (3).
Our data on the adherence phenotype of more than 50 environmental strains indicate that there is not an unequivocal direct correlation between the presence of tdh, trh, T3SS2, or other virulence genes and the ability of the environmental strains to adhere to human cells. Hence, adherence, demonstrated under conditions employed in this study, does not seem to be a pathogenic factor specific for V. parahaemolyticus isolates carrying virulence genes. This result is consistent with other data in the literature (25) and indicates that adherence cannot be used as a discriminatory element to distinguish pathogenic and nonpathogenic V. parahaemolyticus strains.
To investigate further the pathogenic potential of environmental V. parahaemolyticus strains, the cytotoxicity of strains carrying and not carrying T3SS2 genes was evaluated. The cytotoxic effect of the environmental V. parahaemolyticus strains seems to be associated more with the TDH protein, as shown by results obtained with strains VPeVEpan and VPeVEpan2 (a high level of cytotoxicity, similar to that caused by the reference pandemic strain QM97097), in comparison to the strains carrying only the T3SS2 genes and demonstrating a cytotoxicity half that of the reference strain. Although a previous study on tdh-deprived mutants excluded a role for TDH in cytotoxicity (21), other reports (22, 24) have provided support for the cytotoxic effect of TDH on human cells. Our results are in accordance with the latter and, moreover, suggest an additive effect of T3SS2 and TDH on the cytotoxic damage caused by the environmental strain VPeVEpan.
On the other hand, the cytotoxicity of strains not carrying virulence genes was detectable, measured as release of LDH, although at a much lower level (25%). The possibility that other genes not analyzed in this study are in some way involved in the toxicity to eukaryotic cells cannot be excluded.
In the experiments performed using Caco-2 cell monolayers, two representative strains among those carrying T3SS2 (VPe23 and VPe28) induced a decrease in TER values which was associated with a profound reorganization of the cytoskeleton: infection led to the redistribution and aggregation of actin and ZO-1 within Caco-2 cells in accordance with previous reports (11) for strains lacking the tdh and trh genes. Because the ability to cause disruption of tight junctions and a consequent decrease of TER values is very similar for strains carrying only T3SS2 and those with the tdh gene, it appears that these effects are mainly associated with the secretion system, as reported for strains lacking TDH prior to the identification of T3SS2 as a virulence factor (11). The effects on the cytoskeleton and tight junctions could be assumed to comprise a phenotype distinguishing environmental V. parahaemolyticus with pathogenic potential from nonpathogenic V. parahaemolyticus.
On the basis of data obtained from the experiments described here employing representative V. parahaemolyticus strains, it can be concluded that the environmental strains analyzed, including those lacking the classical pathogenic factors tdh and trh but carrying other virulence-associated genes such as those belonging to T3SS2, showed at least some of the virulence characteristics typical of clinical strains: they adhere efficiently to human cells and once in contact with human intestinal cells cause disruption of the membrane tight junctions, compromising the intestinal barrier. We hypothesize a role for this secretion system in the pathogenicity of the marine strains, but further studies are needed to provide direct evidence to support this hypothesis. The results reported in this paper support the fact that environmental V. parahaemolyticus strains might constitute a public health concern and a risk to human health and should be taken into consideration in water quality surveillance plans.
The V. parahaemolyticus environmental strains analyzed in this study were isolated within the framework of the international research project VibrioSea, cofunded by the Centre National d'Etudes Spatiales (CNES), the Institut Pasteur, France, and the Universities of Verona and Genoa, Italy. This project has been developed in the context of a doctorate fellowship, Cooperint2007, funded by the University of Verona, Italy, and assigned to G. Caburlotto. This work was supported by NIH grant R01 AI 19716 to J. B. Kaper and partially by grant NBCH2070002 to R. R. Colwell received from the Department of Homeland Security.
We thank Patricia B. Lodato and Jane Michalski (Department of Microbiology and Immunology, University of Maryland School of Medicine, Baltimore, Maryland) for their suggestions and technical support in the part concerning the expression of virulence-related genes. We thank Alessio Fasano (University of Maryland School of Medicine, Baltimore, Maryland) for the technical support in immunofluorescence microscopy.
Editor: S. M. Payne
Published ahead of print on 17 May 2010.