This study was undertaken to examine the diversity in biochemical activity, serotype, and the presence of known virulence markers (the tdh, trh, and T3SS genes) among 144 V. parahaemolyticus isolates. To our knowledge, no other study has characterized such a diverse panel of V. parahaemolyticus isolates at this level of detail. The isolates examined in this study included clinical strains from across North America isolated from July 2006 to November 2007 and oyster isolates obtained from market oysters collected across the United States in 2007. While the focus of this study was to identify differences between clinical and oyster isolates in order to provide insights into strain virulence, it should be noted that there are likely pathogenic strains among the group of oyster isolates. All potentially pathogenic (based on the presence of tdh and/or trh) oyster isolates were included in the study, as well as a proportion of oyster isolates negative for both tdh and trh similar to the proportion of clinical isolates available with the same hemolysin genotype. Our sample selection criteria, including preferential selection of tdh-negative, trh-negative strains from oysters, introduces a possibility of sample selection bias in the statistical comparisons, but the relative proportions of strains negative for both tdh and trh in the two groups of isolates, clinical and oyster, were similar.
As in previous reports, the API 20E test was unreliable for the identification of the V. parahaemolyticus
), but all strains were identified by the presence of the V. parahaemolyticus
gene. The only novel trait observed in this set of isolates was the ability to produce β-galactosidase (ONPG test), but this was not a differential feature between clinical and oyster isolates. Historically, the majority (>90%) of V. parahaemolyticus
strains have been reported to be negative for ONPG (3
), but 46% and 57% of the clinical and oyster isolates, respectively, were positive. This may be an evolution of V. parahaemolyticus
that was previously undocumented due to the decrease in the use of biochemical examination of isolates because of its lack of reliability for species identification. Previously, urease production has been linked to the presence of the trh
). In this study, 95 of 96 (99%) isolates that produced urease also harbored the trh
gene. However, three trh
-positive strains were negative for urease production. Since this study did not test for the presence of the ure
gene, it is possible that these strains simply did not express the gene for urease production under the experimental conditions used.
The most prevalent serotype seen in this study was O1:KUT, one of the serogroups associated with pandemic V. parahaemolyticus
). Eleven clinical and 7 oyster isolates belonged to serotype O1:KUT. However, it has been reported that serotype O1:KUT alone is not a reliable indicator of a pandemic lineage isolate (7
). The remaining pandemic serotypes, O3:K6, O4:K68, and O1:K25, were infrequently identified among the clinical isolates but were not identified in any of the oyster isolates. Taken together, these data indicate a likely low prevalence of these pandemic strains in U.S. market oysters. The second most prevalent serotype was O4:K12, the most common serotype in human illnesses from Washington State (12
). All 11 of the O4:K12 isolates were from patient stool specimens. Although only three were from the state of Washington, the majority of others were from states that may have received seafood products from Washington. Thirteen additional serotypes were found only in clinical isolates, and nine serotypes were found only in oyster isolates. As such, serotype was the trait least shared by clinical and oyster isolates. Future work on virulence models is planned to determine whether certain serotypes, or groups of serotypes, can be predictive of pathogenic potential.
The hemolysin genes, tdh
, have generally been considered reliable indicators of strain virulence (2
). Additionally, it has been suggested that strains containing T3SS2 should also be considered more virulent than others (5
). More than 90% of clinical V. parahaemolyticus
strains with tdh
and T3SS2 were isolated from stool specimens, supporting the hypothesis that these hemolysin genes and the presence of T3SS2 are predictive of food-borne illness risk. In contrast, 27% of clinical V. parahaemolyticus
strains in this study were negative for tdh
, and T3SS2, so they would, by convention, be considered avirulent. This group of strains was diverse with respect to the isolation source, with one (5%) from blood, five (24%) from wounds, seven (33%) from stool specimens, and eight (38%) reported as “other.” This could indicate that these strains are more opportunistic or have decreased virulence potential, since “other” may refer to infections resulting from recreational water activities as opposed to seafood consumption. This information does raise some concerns about the reliability of the tdh
, and T3SS2 genes as predictors of overall strain virulence. However, a definitive association of each isolate with a patient with food-borne illness is lacking, making it difficult to determine whether different sets of virulence factors may play a role in wound infections caused by V. parahaemolyticus
versus food-borne illnesses.
Market oyster isolates in this study were preferentially selected for harboring the tdh
gene. This precludes any conclusions about the distribution of strains negative for both tdh
, but all strains of other genotypes (positive for both tdh
positive and trh
negative and trh
positive) isolated during the previous study (9
) were included. All of the isolates positive for tdh
(whether positive or negative for trh
) were from oysters harvested from Gulf and Mid-Atlantic state estuaries from late spring to early winter. This is in agreement with a previous study of Gulf oysters and water, where pathogenic V. parahaemolyticus
strains were isolated more frequently when water temperatures were elevated (18
); however, there is a report describing an inverse correlation between water temperature and the isolation of tdh
-positive V. parahaemolyticus
strains from oysters (11
A majority of tdh
-positive isolates were obtained from Canada, Maine, and Washington during the early summer, indicating a potential preferential distribution of these strains in northern areas. This is reflected in the previous report that trh
-positive V. parahaemolyticus
strains constitute a higher proportion of the total V. parahaemolyticus
population in the Mid-Atlantic than in other areas during the summer (9
). It is also interesting that the remaining tdh
-positive V. parahaemolyticus
strains were isolated from Texas oysters harvested from January through May, since this is one of the few areas of the Gulf Coast where the prevalence of trh
-positive strains had not been well documented.
In this study, 10 of the V. parahaemolyticus
isolates were missing at least one of the four T3SS1 genes tested (even when amplified by simplex PCR), in contrast to previous reports that T3SS1 is present in all strains of V. parahaemolyticus
). VP1686 and/or VP1694 (vscF
) were the genes missing from these isolates. The functions of these genes are not yet defined, and VP1686 is a “putative” open reading frame (ORF), indicating that the loss or absence of these genes may not impair the functionality of T3SS1. Because some weak amplification was observed by simplex PCR, it is possible that these strains do possess the VP1686 and VP1694 genes but have a divergent sequence that cannot be amplified efficiently with the current primer set.
Noriea et al. suggested that VPA1362 (vopB2
) of T3SS2α may be a more reliable predictor of virulence than tdh
, based on the absence of vopB2
in environmental isolates that contained other T3SS2α genes (31
). As in that study, the vopB2
gene specific to T3SS2α was amplified in all nine of the tdh
-negative clinical isolates. Additionally, neither of the two oyster tdh
-negative isolates amplified VPA1362 (vopB2
) by multiplex PCR. Weak amplification was observed from the two oyster isolates by simplex PCR, suggesting a potential sequence variation in the vopB2
gene that might be utilized as a more specific indicator of strain virulence than tdh
As reported previously (31
), all clinical and oyster isolates that were tdh
negative and trh
positive amplified all four genes specific for T3SS2β. Whereas a previous report demonstrated the absence of T3SS2α and T3SS2β in isolates positive for both tdh
), all clinical isolates and all but one oyster isolate with tdh
amplified all four T3SS2β genes. The one remaining oyster isolate positive for both tdh
amplified three of the four genes, with no amplification of the vopC
effector protein. This is the first report of the presence of T3SS2β in isolates positive for both tdh
. This difference may be attributed to diversity in the strain panel and the use of simplex PCR to examine all isolates, whereas only the multiplex PCR was utilized in previous reports.
This study demonstrated more-reliable amplification of the T3SS genes by simplex PCR than by the previously published multiplex PCR assays. The anomalies observed in the PCR amplifications of these genes bring to light many questions about the sequence diversity of these systems. The primer sets utilized in the current study were designed based on the limited number of available sequences, which were predominately from clinical isolates. It is likely that a higher diversity of these systems will be revealed as the number of isolates examined increases. The sequence divergence of the T3SSs from the strains studied, particularly the T3SSα vopB2 gene, is a subject of investigations currently under way in our laboratory.
In summary, this study characterized a relatively diverse group of V. parahaemolyticus clinical and oyster isolates for biochemical differences and the distribution of pathogenicity factors. Surprisingly, more than one-quarter of the clinical isolates were negative for both tdh and trh and did not possess T3SS2. These results indicate that the virulence of V. parahaemolyticus is more complex than historically believed: the tdh and/or tdh and T3SS2 genes are not necessarily predictive of pathogenic potential. Serotype was a distinguishing feature of the clinical isolates; 17 of the serotypes were found only in clinical isolates. However, the variety of serotypes may be too wide for use as a predictor of virulence. Overall, this study exposes a higher level of complexity in the virulence potential of V. parahaemolyticus and brings to light concerns about the reliability of the long-standing virulence markers of the species. More-discriminatory analyses of these and additional isolates are under way in our laboratory with the goal of elucidating more-reliable predictors of V. parahaemolyticus virulence.