INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

Pathogenic Vibrio parahaemolyticus strains cause gastroenteritis in humans after they consume contaminated foods, most often raw or partially cooked fish and shellfish (1, 7, 15). The organism is widely disseminated in estuarine environments throughout the world (9) and has been detected in oysters on Canada's Pacific coast (12) and in the environment of western North America as far north as Alaska (30). Outbreaks of V. parahaemolyticus food poisoning are most common in Japan and southeast Asia (1, 22), though they occur occasionally in other parts of the world. The number of cases detected in Canada to this date is extremely low (14). Prior to 1997, only sporadic cases of locally acquired gastroenteritis caused by urease-positive, Kanagawa hemolysin-negative V. parahaemolyticus were detected on Canada's Pacific coast (11, 13). A prospective study in this region demonstrated that V. parahaemolyticus infection was locally acquired in only 10 patients during a 3-year period, 1984 to 1987 (11), and that the isolates were urease-positive, Kanagawa hemolysin-negative strains (11). A recent outbreak of gastroenteritis in July and August 1997 appears to have been caused by ingestion of uncooked or undercooked oysters contaminated with V. parahaemolyticus (6). V. parahaemolyticus was found in oysters from several harvesting areas in the region, though only in low numbers. In addition, no deficiencies in oyster processing or distribution that could account for the outbreak were identified.

Subspecies typing of V. parahaemolyticus may be useful for tracking the organisms implicated in the recent Canadian west coast outbreak as well as for gaining insight into the ecology of the organisms in Canadian waters. A number of molecular methods for typing V. parahaemolyticus have been described. Restriction fragment length polymorphism (RFLP) analysis of virulence or virulence-associated genes has been a valuable tool for typing several bacterial species, including Escherichia coli O157:H7 (26) and Vibrio cholerae (3, 4, 32, 34). In V. parahaemolyticus, the thermostable direct hemolysin (TDH) and the TDH-related hemolysin, encoded by the tdh and trh genes, respectively, are important virulence factors involved in the production of gastroenteritis (20, 27). However, it appears that due to conservation of nucleotide sequences flanking the tdh and trh genes in V. parahaemolyticus, RFLP analysis of these genes lacks discriminatory power (21). Use of the arbitrarily primed PCR method gave higher resolution in that study. Wong et al. (33) have recently described a method for pulsed-field gel electrophoresis (PFGE) capable of separating 130 strains into 39 patterns constituting 14 PFGE types. Analysis of 16S and 23S ribosomal gene RFLPs and enterobacterial repetitive intergenic consensus sequence (ERIC) PCR have previously proven useful for subtyping V. cholerae (2, 24, 25). The genes for the polar flagellum of V. parahaemolyticus have been cloned and sequenced and are arranged at two separate loci within the genome of this organism (17, 28). Since flagellar gene RFLPs have been useful for characterization of other organisms, most notably Campylobacter jejuni (18, 19, 22), RFLPs of the V. parahaemolyticus polar flagellum were investigated to determine if they would be similarly useful for typing of this organism.

We therefore compared four methods for subtyping strains from a recent V. parahaemolyticus outbreak on Canada's Pacific coast: PFGE, ribotyping, ERIC PCR, and RFLP analysis of the polar flagellum gene locus (Fla locus RFLP analysis). ERIC PCR appeared to have the best discriminatory power, followed closely by ribotyping and PFGE. Fla locus RFLP analysis was not as discriminatory as the other methods, but it may provide an alternate rapid method for molecular typing.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Strains. Bacterial isolates from patients (38 isolates) and environmental sources (16 isolates), collected during the 1997 outbreak by hospital microbiology staff and laboratory personnel at the B.C. Centre for Disease Control, were submitted to the Laboratory Centre for Disease Control (LCDC) in Ottawa for further characterization. Six other human V. parahaemolyticus isolates from sporadic cases of gastroenteritis were submitted by provincial laboratories to the LCDC laboratories; a total of 60 isolates were characterized. Bacteria were maintained on slants of Institut Pasteur maintenance medium (10 g of Difco peptone per liter, 5 g of Difco beef extract per liter, 3 g of NaCl per liter, 2 g of Na₂HPO₄ · 12H₂O per liter, 8 g of Difco granulated agarose per liter [pH 7.4]) at room temperature in the dark for long-term storage. Subcultures frozen at 80°C in brain heart infusion broth (Difco) containing 15% glycerol were used to prepare working cultures.

DNA extraction. Bacteria grown overnight on nutrient agar plates were suspended in TNE buffer (10 mM Tris, 20 mM NaCl, 50 mM EDTA [pH 7.5]). Sodium dodecyl sulfate and proteinase K were added to yield final concentrations of 0.5% and 100 µg/ml, respectively, and the bacterial suspensions were incubated at 55°C for 30 min. DNA was extracted with phenol-chloroform-isoamyl alcohol (25:24:1), precipitated with isopropanol, washed with 70% ethanol, and resuspended in TE buffer (10 mM Tris-HCl, 1 mM EDTA [pH 8]).

ERIC PCR. The ERIC 1R primer (5'-ATG TAA GCT CCT GGG GAT TCA C-3') (31) used in this investigation was obtained from the DNA Core Facility, Bureau of Microbiology, LCDC. The reaction mixture (25 µl per reaction) consisted of a 200 µM concentration of each deoxynucleoside triphosphate, 1× PCR buffer II (Perkin-Elmer, Foster City, Calif.), 1.5 mM MgCl₂, 2 µM ERIC 1R primer, 1.25 U of AmpliTaq (Perkin-Elmer), and 100 ng of template DNA. PCR was performed in a Perkin-Elmer 9600 thermal cycler. Following an initial denaturation at 94°C for 5 min, genomic DNA was amplified through 35 cycles of 94°C for 10 min, 52°C for 1 min, and 72°C for 1 min. PCR was completed with a final extension at 72°C for 10 min. Products were separated by electrophoresis on 1.5% agarose gels containing 0.5× TBE (10× TBE is 0.89 M Tris base, 0.89 M boric acid, and 0.02 M disodium EDTA [pH 8.4]; Boehringer Mannheim, Laval, Quebec, Canada), followed by ethidium bromide staining and photography with a UV transilluminator.

Long PCR and Fla locus RFLP analysis. PCR primers designed to amplify a 6.6-kb fragment comprising most of the 8.3-kb locus containing V. parahaemolyticus polar flagellar genes flaB, flaA, flaG, flaH, flaI, flaJ, and flaK, (17, 26) were obtained from the DNA Core Facility. The forward primer, 5'-TAC CTA AAC AAC GCA AAC TCA GCA CA-3', lies within flaB, the first gene in the operon. The reverse primer, 5'-TGC GGT ATG ACG AAT AGT GAA TG-3', is contained in flaK, the last gene in the operon. Long PCR utilized reagents from the GeneAmp XL PCR kit (Perkin-Elmer) in 50-µl reaction mixtures containing 1× XL buffer II, a 200 µM concentration of each deoxynucleoside triphosphate, 30 pmol of each primer, 1.2 mM magnesium acetate, 3 U of recombinant Tth DNA polymerase XL, and 40 ng of DNA template. Amplification was accomplished in a Perkin-Elmer 2400 thermal cycler according to the following cycle parameters: initial denaturation at 94°C for 1 min; 16 cycles of 94°C for 15 s, 60°C for 30 s, and 68°C for 7 min; 17 cycles of 94°C for 15 s, 60°C for 30 s, and 68°C for 7 min 15 s; and final extension at 72°C for 10 min. After 6 µl of each PCR product was digested with a restriction enzyme in a total volume of 30 µl for 3 h at 37°C, 15 µl of the final digest was analyzed on 2% agarose gels, stained, and visualized as summarized above. Restriction enzymes tested were chosen on the basis of the published flagellar locus gene sequence and included DdeI, RsaI, TaqI, and Sau3AI; double digestion by PstI and SspI was also performed. RsaI performed best in these initial experiments and was used for all subsequent analyses.

Ribotyping. Genomic DNA was restricted for 6 h with BglI, and fragments were separated on a 0.8% agarose gel. A supercoiled DNA ladder (Gibco BRL, Burlington, Ontario, Canada) linearized with PvuII was used as a size standard. After electrophoresis, gels were stained with ethidium bromide and observed under UV light to ensure that the DNA was completely digested. After denaturation and neutralization, DNA in the gel was capillary blotted onto a Hybond-N+ membrane (Amersham Life Science Inc., Oakville, Ontario, Canada) and cross-linked to the membrane with a Stratalinker 2400 instrument (Stratagene; PDI Bioscience, Aurora, Ontario, Canada). All subsequent procedures were done according to the protocol for hybridization in tubes supplied with the enhanced chemiluminescence (ECL) direct nucleic acid labeling and detection kit (Amersham Life Science Inc.). Briefly, membranes were prehybridized in ECL hybridization buffer for 1 h at 42°C. The rrnB rRNA operon of E. coli was cut from plasmid pKK3535 (15), purified from the gels, and used as a probe for the rRNA genes (16). A supercoiled DNA ladder (Gibco BRL) was used as the probe for the size standard. Each probe (200 ng) was denatured by boiling 5 min, labeled directly with horseradish peroxidase according to the manufacturer's directions, added immediately to the hybridization buffer, and allowed to hybridize at 42°C overnight. Blots were then washed once in 5× SSC (20× SSC is 3.0 M NaCl plus 0.3 M sodium citrate, pH 7.0; Boehringer Mannheim), three times at 55°C in primary wash buffer without urea (0.1× SSC containing 0.4% sodium dodecyl sulfate), and twice with 2× SSC, and they were developed with ECL detection reagent followed by exposure to Hyperfilm-MP photographic film (Amersham Life Science Inc.). No cross-reactivity of the supercoiled DNA ladder probe with V. parahaemolyticus DNA was seen on blots labeled only with this probe.

PFGE. Isolates were grown overnight at 37°C with shaking in brain heart infusion broth with or without 2% NaCl added. Bacteria were collected by centrifugation, washed with 500 µl of 75 mM NaCl-25 mM EDTA (pH 7.9), centrifuged again, and suspended in 200 µl of 10 mM Tris containing 20 mM NaCl and 50 mM EDTA (pH 7.9). A suspension equivalent to McFarland standard 8 was prepared by adding 200 µl of 2% low-melting-point agarose (Boehringer Mannheim) solution in TNE buffer. This suspension was transferred to disposable plug molds (Bio-Rad Laboratories Ltd., Richmond, Calif.) and cooled to 4°C. Plugs were suspended in 1 ml of 250 mM EDTA (pH 8) containing 2 mg of proteinase K per ml and incubated at 55°C for 30 min. After this time, 1 ml of 1.0% N-lauroylsarcosine (wt/vol) was added, and incubation was continued overnight at 55°C. Agarose plugs were washed three times with TE buffer (pH 7.5) for 30 min per wash and then were washed five times with distilled water for a minimum of 20 min per wash. After equilibration with restriction enzyme digestion buffer A (Boehringer Mannheim) for 45 min, agarose-embedded DNA was digested with 80 U of restriction enzyme overnight at 30°C. Initial experiments on selected isolates indicated that NotI and ApaI gave similar results; therefore, ApaI was used in all analyses. Digestion was stopped by the addition of 100 µl of loading dye (30% glycerol, 0.25% bromophenol blue, and 0.25% xylene cyanol FF in distilled water), and plugs were melted for 15 min at 69°C. Forty microliters of sample was added to a 1% agarose gel made by using 0.5× TBE. Electrophoresis was performed for 24 h at 6 V/cm with a 1- to 18-s linear ramp time in a Bio-Rad CHEF-DR II electrophoresis system. Gels were cooled to 14°C throughout the run. After staining with ethidium bromide, banding patterns were visualized with a UV transilluminator and photographed.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Patterns from the four typing methods were compared visually. PFGE, ribotyping, and ERIC PCR differentiated strains into 15, 14, and 13 similar types, respectively, while Fla locus RFLP analysis generated five discrete types (Table 1). In addition, PFGE or ERIC PCR patterns differing by only one or two bands were designated subtypes, resulting in 19 and 20 different patterns, respectively. When the results from all four methods were combined, at least 35 patterns were obtained from the 60 isolates. The largest group, 14 strains, had identical patterns generated by all four methods and contained both outbreak strains and an environmental strain (T97-67), indicating that bacteria causing clinical cases of disease were likely derived from environmental sources.

All isolates from the 1997 V. parahaemolyticus outbreak on Canada's Pacific coast segregated into four ribotype pattern groups (A, B, C, and F), which also contained four strains isolated from the environment and three from sporadic cases of disease occurring in previous years (Table 1). The remaining 12 environmental isolates were distributed in eight different ribotypes, while each of the 2 remaining isolates from sporadic cases had a unique ribotype pattern (Fig. 1). These data indicate that only a subset of the bacteria present in the environment was represented in the outbreak. Two isolates (T97-63 and -64) were considered untypeable by this method after repeated failures to obtain sufficient quantities of restriction endonuclease-digested DNA for analysis; only smears were seen on gels, suggesting that rapid nonspecific degradation of DNA took place. Degradation of DNA was occasionally seen with some samples (Fig. 1, lane 5), but subsequent reisolation of DNA from these strains had good yields and clear ribotype patterns (data not shown).

View larger version (74K): [in this window] [in a new window]   FIG. 1.   Representative ribotype patterns of BglI-digested DNA. Pattern designations are given in parentheses. Lane 1, T97-31 (A); lane 2, T97-32 (B); lane 3, T97-33 (C); lane 4, T97-54 (D); lane 5, T97-57 (degraded DNA from isolate exhibiting pattern E in subsequent analysis); lane 6, T97-45 (F); lane 7, T97-58 (G); lane 8, T97-60 (H); lane 9, T97-52 (I); lane 10, T97-61 (J); lane 11, T97-65 (K); lane 12, T97-66 (L); lane 13, T97-472 (M); lane 14, T97-474 (N); lane M, PvuII-digested supercoiled DNA, with sizes of selected standard bands shown at the right.

Analysis of PFGE patterns revealed that the outbreak strains could be divided into four types, A, B, C, and D. Strains in the A group could be further divided into A1, A2, and A3 subtypes, differing by one to three bands (Table 1; Fig. 2). Within the outbreak strains, the PFGE pattern generally coincided with the ribotype pattern, so that strains with ribotype pattern A had PFGE type A1, A2, or A3, strains with ribotype pattern B had PFGE type B, strains with ribotype pattern C had PFGE type C, and strains with ribotype pattern F had PFGE type E (Table 1). The only exception was strain T97-37, which had ribotype pattern B but PFGE type C. PFGE patterns B and C differed by more than 10 bands (Fig. 2), suggesting that isolate T97-37 was not closely related to other isolates having the same ribotype, ERIC PCR, and Fla locus RFLP analysis patterns. Environmental isolates showed greater variability, including eight different types among nine strains plus six untypeable strains. With the exception of strain T97-67, discussed above, none of the environmental strains had the same PFGE types as outbreak strains. Similarly, strains isolated from sporadic cases exhibited different PFGE types than either outbreak or environmental strains.

View larger version (136K): [in this window] [in a new window]   FIG. 2.   Representative PFGE patterns for V. parahaemolyticus chromosomal DNA digested with ApaI. Pattern designations are given in parentheses. Lane M,  DNA concatemer size standard, with sizes of selected bands given at the left; lane 1, T97-31 (A1); lane 2, T97-34 (A2); lane 3, T97-450 (B); lane 4, T97-42 (C); lane 5, T97-45 (D); lane 6, T97-52 (E); lane 7, T97-61 (I); lane 8, T97-65 (K).

ERIC PCR differentiated the outbreak strains into six types, A, B, C, D, H, and I (Table 1; Fig. 3). Dominant bands in the ERIC PCR patterns were characteristic of types, while differences in fainter bands allowed differentiation of subtypes. Subtypes A1 and A2 differed by one band, subtypes B2 and B3 each had a one-band difference compared to subtype B1, and subtypes H2 and H3 each had a one-band difference compared to subtype H1. A total of 10 subtypes were seen among outbreak isolates (Table 1; Fig. 3). Environmental strains were grouped into seven different types, three subtypes of which (A1, B1, and D) were also found in outbreak strains. Unique ERIC PCR patterns were found in all isolates from sporadic cases prior to 1997. Groups of outbreak isolates that were apparently identical by ribotype and PFGE type were differentiated by ERIC PCR (Table 1). In addition, while ribotyping found environmental isolates T97-53, -55, -62, and -67 to be identical to 18 outbreak strains producing ribotype pattern B, ERIC PCR identified 16 of these outbreak isolates as having identity solely with strain T97-67. ERIC PCR also detected identity between environmental isolates T97-63 and -64 and the outbreak strain T97-461. In total, 18 of 38 outbreak isolates were found to be identical to environmental isolates when either ERIC PCR or ribotyping was used for comparison. Patterns generated by ERIC PCR were reproducible in at least three different assays, though the intensities of some of the fainter bands occasionally varied somewhat.

View larger version (83K): [in this window] [in a new window]   FIG. 3.   ERIC 1R PCR profiles from selected V. parahaemolyticus isolates. Pattern designations are in parentheses. Lane M, 100-bp ladder; lane 1, T97-31 (A1); lane 2, T97-32 (B1); lane 3, T97-33 (C1); lane 4, T97-56 (D); lane 5, T97-65 (G); lane 6, T97-450 (I); lane 7, T97-474 (L); lane 8, negative control.

The polar flagellum locus containing the flaA and flaB flagellin genes, genes for hook-associated proteins (flaG and flaH), and other genes encoding proteins comprising the flagellar apparatus (flaK, flaJ, and flaK) was amplified by PCR and the resulting product was digested with RsaI, which exhibited the greatest number of polymorphisms and an optimal complexity of banding patterns (Fig. 4). Only five different patterns were obtained by Fla locus RFLP analysis. Type B was common to organisms from the 1997 British Columbia outbreak that also had ribotype B, organisms from the environment, and isolates from sporadic cases of V. parahaemolyticus gastroenteritis prior to 1997 (Table 1). Fla locus RFLP analysis and ribotyping identified the same 18 outbreak isolates as being identical to environmental isolates. The second most common Fla locus RFLP pattern, type C, was associated only with outbreak strains having ribotypes C and F, while type A was associated only with outbreak strains that had ribotype A and with two isolates from sporadic cases before 1997. Fla locus RFLP types D and E were each associated with only a single environmental isolate.

View larger version (97K): [in this window] [in a new window]   FIG. 4.   RFLP patterns obtained by RsaI digestion of the product of PCR of the V. parahaemolyticus Fla locus. Pattern designations are given in parentheses. Lanes M, 100-bp ladder; lane 1, T97-31 (A); lane 2, T97-32 (B); lane 3, T97-33 (C); lane 4, T97-58 (D); lane 5, T97-56 (E).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The molecular typing methods used here were helpful for differentiating isolates obtained from patients during the outbreak. There was a strong correspondence among ribotypes, PFGE types, and Fla locus RFLP types for the 38 outbreak isolates when compared with a background of heterogeneous environmental strains, suggesting that the outbreak isolates were derived from a limited number of initial sources. The correspondence of ERIC PCR types with types identified by the other three methods was not as strong. One environmental isolate, however, was found to be identical to several outbreak isolates when all four typing methods were considered, confirming the environmental origin of the outbreak strains.

ERIC PCR typing and ribotyping identified 17 outbreak isolates as being identical to one environmental strain, whereas PFGE almost completely segregated outbreak strains from environmental and sporadic isolates. This finding implies that these methods measure genetic events that occur independently, and it makes interpretation of the relationships between strains more complex. It is clear that no single method is sufficient to unambiguously examine the genetic relatedness among the strains. At the same time, there appears to be a tremendous genetic heterogeneity in the endemic environmental V. parahaemolyticus population that is reflected in the large number of types associated with human disease that were seen here.

V. parahaemolyticus is endemic to Canada's Pacific Northwest region (12), though it cannot be isolated from oysters during periods when water temperatures are cool. It seems likely that, as noted for other geographical areas (9), this organism may reside undetected in environmental reservoirs until conditions are appropriate for its association with and growth in oysters or other shellfish. While ribotyping and Fla locus RFLP analysis indicated that at least some of the types found in outbreak and environmental isolates in 1997 were also present in isolates from sporadic cases prior to 1997, no identity from year to year could be found by the other two methods. Genetic changes resulting in variant ribotypes or Fla locus RFLP patterns may occur more slowly than changes producing different patterns in PFGE or ERIC PCR analyses, though the biological processes that could generate such differences are unknown. Karaolis et al. (10) have calculated a rate of one rRNA restriction site change every 6 years in V. cholerae; in this and other bacterial species, the stability of ribotype patterns has allowed the development of standard ribotyping schemes (5, 24). On the other hand, genetic changes that alter PFGE patterns can occur during the course of an outbreak (29). Ribotyping may therefore be the most useful method for examining similarity in isolates that are temporally or geographically separated.

Simpson's index of diversity has been adapted for use as a method for deriving a numerical index of the discriminatory ability of single or combined typing systems (8). The numbers derived are presented as the discrimination index. For each typing method, the discrimination index represents the percentage of occasions that two strains sampled randomly from a population fall into different types.

In this investigation, ribotyping exhibited a discrimination index slightly lower than that of PFGE or ERIC PCR but much higher than that of Fla locus RFLP analysis (Table 2). Despite the fact that ribotyping is slower and somewhat more labor-intensive than the other three methods, the lower rate of change of ribotype patterns, as discussed above, may be useful for evaluating potential links between environmental strains and isolates from infected patients. In this investigation, ribotyping was used in the initial grouping of strains (Table 1). The resulting groups appeared to have a logical coherence overall when other typing methods were included, suggesting that ribotyping detected an underlying similarity in the biology of these bacteria. Ribotyping may also be useful for investigating changes in populations of environmental isolates over longer periods of time. Isolates with ribotypes B and C were isolated from patients in the years before the 1997 outbreak; these types were also found among outbreak strains, suggesting that long-term carriage in the environment may occur.

ERIC PCR has the best discriminatory value (Table 2), is rapid, and is relatively easy to perform, making it the method of choice when a single method is to be used for typing V. parahaemolyticus. However, the potentially rapid change in ERIC PCR types could also obscure relationships between strains that otherwise might be evident if other methods were used. ERIC typing in conjunction with ribotyping (Table 2) has the best discrimination index and typing ability of all the methods. This combination may be used to elucidate most of the relevant genetic and epidemiological relationships among V. parahaemolyticus strains.

While PFGE had a high diversity index, evaluation of PFGE results was complicated by the relatively large number of strains (14 of 60 [23%]) that were untypeable as a result of DNA degradation, despite precautions to limit the activity of endogenous proteases and DNases. This difficulty was not encountered in previous work (33), though it may limit the usefulness of this method in the future. Comparison of our PFGE data with those of Wong et al. (33) indicates that the two methods have similar complexities in the size ranges separated. Though PFGE appeared to be an excellent technique for differentiation of outbreak, environmental, and sporadic strains in this study, the high proportion of untypeable isolates may limit the utility of this method in the future.

Fla locus RFLP analysis had the lowest discriminatory power of any technique used in this study, though it was useful for confirmation of types determined by other methods. The fact that fewer patterns were generated by Fla locus RFLP analysis than by the other three techniques and the conservation of these patterns from year to year suggest that the fla genes within this locus are relatively conserved. The choice of more variable genetic loci for RFLP analysis may provide more detailed information about the genetic relationships among isolates.

In previous investigations of V. parahaemolyticus in the Pacific Northwest of North America, locally acquired infections were found only when the organism could be detected in the local environment (11). V. parahaemolyticus strains have been recovered from the environment in this area when water temperatures were greater than 14°C and salinities were less than 13%. The outbreak in 1997 occurred in a year of exceptionally warm ocean surface temperatures and heavy rainfall associated with El Niño conditions conducive to the growth of V. parahaemolyticus. Climatic changes associated with global warming could make these conditions more common in the future, increasing the risk of outbreaks associated with V. parahaemolyticus in this area. Ribotyping and ERIC PCR can be used to further characterize V. parahaemolyticus populations in the environment. Investigations on whether the bacteria responsible for British Columbia outbreaks are endemic to the Canadian west coast or imported from other areas when environmental conditions are favorable may aid risk analysis and prevention.

Counts of V. parahaemolyticus from the oysters obtained from local harvesting areas during the 1997 British Columbia outbreak were below the levels considered necessary to cause human illness (6). Though no deficiencies in oyster processing or distribution were identified in initial investigations, improper handling may have occurred at some point before consumption. In Taiwan, outbreaks were most often associated with commercially prepared school seafood lunches (23), suggesting that mishandling of seafood ingredients may be a common factor. The molecular typing methods described here could be used to aid trace-back investigations aimed at determining the source of improperly handled foods and to create opportunities for intervention or remediation. They should be useful for the epidemiological investigation, intervention, and control of human disease resulting from infection with V. parahaemolyticus.