Predator-Prey Interactions between Halobacteriovorax and Pathogenic Vibrio parahaemolyticus Strains: Geographical Considerations and Influence of Vibrio Hemolysins

ABSTRACT Halobacteriovorax is a genus of naturally occurring marine predatory bacteria that attack, replicate within, and lyse vibrios and other bacteria. This study evaluated the specificity of four Halobacteriovorax strains against important sequence types (STs) of clinically relevant Vibrio parahaemolyticus, including pandemic strains ST3 and ST36. The Halobacteriovorax bacteria were previously isolated from seawater from the Mid-Atlantic, Gulf of Mexico, and Hawaiian coasts of the United States. Specificity screening was performed using a double agar plaque assay technique on 23 well-characterized and genomically sequenced V. parahaemolyticus strains isolated from infected individuals from widely varying geographic locations within the United States. With few exceptions, results showed that Halobacteriovorax bacteria were excellent predators of the V. parahaemolyticus strains regardless of the origins of the predator or prey. Sequence types and serotypes of V. parahaemolyticus did not influence host specificity, nor did the presence or absence of genes for the thermostable direct hemolysin (TDH) or the TDH-related hemolysin, although faint (cloudy) plaques were present when one or both hemolysins were absent in three of the Vibrio strains. Plaque sizes varied depending on both the Halobacteriovorax and Vibrio strains evaluated, suggesting differences in Halobacteriovorax replication and/or growth rates. The very broad infectivity of Halobacteriovorax toward pathogenic strains of V. parahaemolyticus makes Halobacteriovorax a strong candidate for use in commercial processing applications to enhance the safety of seafoods. IMPORTANCE Vibrio parahaemolyticus is a formidable obstacle to seafood safety. Strains pathogenic to humans are numerous and difficult to control, especially within molluscan shellfish. The pandemic spread of ST3 and ST36 has caused considerable concern, but many other STs are also problematic. The present study demonstrates broad predatory activity of Halobacteriovorax strains obtained along U.S. coastal waters from the Mid-Atlantic, Gulf Coast, and Hawaii toward strains of pathogenic V. parahaemolyticus. This broad activity against clinically relevant V. parahaemolyticus strains suggests a role for Halobacteriovorax in mediating pathogenic V. parahaemolyticus levels in seafoods and their environment as well as the potential application of these predators in the development of new disinfection technologies to reduce pathogenic vibrios in molluscan shellfish and other seafoods.

salt requirements and tolerances (reviewed in reference 1). The life cycle of Halobacteriovorax involves an attack-phase predator that seeks out, attacks, and digests a hole in the cell wall of susceptible prey bacteria through which it enters the intraperiplasmic space, forming a predator-prey structure called a bdelloplast. The prey dies, and the Halobacteriovorax, using host nutrients, elongates until the nutrients are depleted, at which time it septates and divides into progeny cells. The progeny are released into the extracellular milieu after lysis of the former host's cell wall to initiate another cycle of infection (2,3).
Previous specificity studies indicate that Vibrio parahaemolyticus is a common prey for Halobacteriovorax (1,4,5). This, in conjunction with the ubiquitous distribution of Halobacteriovorax in salt water (1,6), suggests that Halobacteriovorax may be one of nature's tools to limit the proliferation of V. parahaemolyticus populations in the marine environment (1,7). However, many of the susceptibility studies on the predation of V. parahaemolyticus by Halobacteriovorax have used environmental Vibrio isolates, not clinically isolated V. parahaemolyticus strains. There are many strains of V. parahaemolyticus that cause shellfishassociated illnesses, particularly from the consumption of raw or undercooked oysters, as well as from other raw seafoods. Unfortunately, there is only limited information available on the susceptibility of specific sequence types (STs) and serotypes of human-pathogenic V. parahaemolyticus to predation by Halobacteriovorax.
Vibrio parahaemolyticus is widespread within the marine environment throughout temperate and tropical regions of the world. It is considered a common cause of seafoodrelated bacterial illnesses. An estimate of the number of reported and underreported foodborne cases of V. parahaemolyticus acquired in the United States annually ranges between 34,664 and 58,027 (8). In a study involving clinical isolates of V. parahaemolyticus from North America, Jones et al. (9) identified numerous strains from stools and wounds of infected individuals, as well as apparently nonclinical isolates derived directly from oysters. They also determined whether the strains contained potential virulence factors, including the thermostable direct hemolysin gene (tdh) and the thermostable direct hemolysin-related hemolysin gene (trh) (10). Those results were provided in more detail for these same strains in a recent paper by Miller et al. (11), who also reported genomic sequences for all the V. parahaemolyticus isolates.
This study evaluates four strains of Halobacteriovorax obtained from the Gulf of Mexico, Hawaii, and tributaries of the Delaware Bay (12,13) to determine their ability to prey upon and kill 23 well-characterized and clinically relevant V. parahaemolyticus isolates from the Jones et al. study (9). The vibrios represent six STs, including multiple ST3 and ST36 strains obtained from around the United States. The ST3 and ST36 strains are pandemic strains that include serotypes O3:K6 and O4:K12, respectively (14)(15)(16)(17)(18). Two ST631 strains are also included in the evaluation. ST631 is known to be endemic in northern states (19,20). We also evaluated whether Halobacteriovorax strains preferentially infect V. parahaemolyticus strains that are common to the same geographical area or if they are broadly infectious toward V. parahaemolyticus isolates from other, more distant habitats. Additionally, we evaluated the infectivity of each Halobacteriovorax isolate toward V. parahaemolyticus strains in the presence and absence of genes for virulence-associated hemolysins to determine if the hemolysins might protect the vibrios from predation.

RESULTS AND DISCUSSION
Halobacteriovorax isolates and BLAST searches. Four Halobacteriovorax isolates were originally obtained from seawater from sites along the Delaware Bay (Mid-Atlantic), the Gulf Coast of Alabama, and Hawaii (13) (11), were used as potential prey for the four Halobacteriovorax strains. The vibrios were kindly provided by the U.S. Food and Drug Administration (FDA). Genomic sequences of each of these isolates were previously determined and are given in the work of Miller et al. (11). Sequence types and serotypes of these strains and the state where each illness was reported are shown in Fig. 1.
Specificity studies and geographic implications. The ability of Halobacteriovorax to predate upon different clinical strains of V. parahaemolyticus obtained from different geographic areas was determined by plaque assay. Results show that most Halobacteriovorax strains are predatory toward the vibrios (Fig. 1). Very broad specificity was observed for the Halobacteriovorax strains; all four strains predated upon 21 or 22 of the 23 V. parahaemolyticus strains plus the V. parahaemolyticus strain RIMD 2210633 (ST3, O3:K6 pandemic strain that was isolated in Japan) (Fig. 1). The RIMD strain was used for our original isolation of the Halobacteriovorax strains from seawater (5). Only two of the V. parahaemolyticus strains (CDC strain K5582, an ST631 isolate from Georgia, and K4859, an untypeable strain from Hawaii) were resistant to predation by three of the four Halobacteriovorax isolates (Fig. 1). The K5582 isolate from Georgia was infected only by Halobacteriovorax G3, which was isolated from seawater from the neighboring state of Alabama, while the K4859 strain from Hawaii was infected only by the Hawaiian strain of Halobacteriovorax (H4). Three other Vibrio strains from Georgia and another three from Hawaii did not show any predator-prey preferences based on the locations from which the Halobacteriovorax and vibrios were isolated (Fig. 1).
Results showed generally broad specificity of Halobacteriovorax isolates from different regions of the United States against strains of V. parahaemolyticus from widely varying  Miller et al. (11). c Abbreviations: "Unt" indicates that the sequence type is untypeable using the typing method of González-Escalona et al. (21); "Kuk" signifies that the K serotype is untypeable. d Host Vp is the V. parahaemolyticus strain that was used to originally isolate the four Halobacteriovorax strains used in this study.
Since Halobacteriovorax bacteria are not known to replicate outside a host bacterium, and the above-described assays used the original host bacterium (V. parahaemolyticus RIMD 2210633) to initially isolate and subsequently propagate the predators, it was necessary to remove the host cells before performing the specificity testing in other V. parahaemolyticus strains. This was accomplished using an enrichment, filtration, dilution method as previously published (5). The resulting Halobacteriovorax stocks were tested to ensure successful removal of the vibrios in the enrichment culture. Controls showed no Vibrio contamination in any of the filtered and diluted Halobacteriovorax stocks used in our host specificity studies.
Prey specificity as affected by sequence type and serotype. Predator-prey interactions were determined with six different sequence types and 12 serotypes of V. parahaemolyticus. Results for five isolates of ST3 and 11 isolates of ST36 demonstrate predator-prey interactions (Fig. 1). The ST3 isolates comprised serotypes O3:K6 (3 each), O4:K68 (1 each), and O1:Kuk, where Kuk stands for K untypeable (1 each). Halobacteriovorax preyed upon the five ST3 strains and an additional ST3 strain used as the original host (V. parahaemolyticus RIMD 2210633). The ST36 isolates consisted of serotypes O4:K12 (7 each), O4:K13 (1 each), O4:Kuk (2 each), and O4:K63 (1 each). Neither the ST nor the serotype appeared to affect the predation of Halobacteriovorax on any of these V. parahaemolyticus strains. The Halobacteriovorax bacteria were also quite capable of infecting other STs and serotypes of clinically important V. parahaemolyticus isolates (Fig. 1).
Plaque sizes. Specific plaque diameters were not measured here because sizes varied depending somewhat on the density of the plaques. Many plaque assay plates contained plaques that were too numerous to count, where most of the plaques could not be individually discriminated. Plaques were first visible usually 2 to 3 days postplating, or slightly longer for faint plaques, and slowly increased in size over the 5-to 7-day incubation period. Where individual sizes were observable, plaques ranged from pinpoint (;1 mm in diameter) to large (;1 cm in diameter). A breakdown of relative plaque sizes (pinpoint, small, medium, and large) is shown by the colored and lettered blocks in Fig. 1. Relative plaque sizes give some clues about the growth and replication rates of individual Halobacteriovorax strains on any given V. parahaemolyticus host strain under the prevailing conditions. As evident from Fig. 1, Halobacteriovorax G3 produced small, medium, and large plaques depending on the Vibrio strain. In contrast, Halobacteriovorax S11 produced only pinpoint or small plaques regardless of the Vibrio strain, except in the original host (RIMD 2210633), where plaque sizes were small to medium. Most of the pinpoint-sized plaques were found in the ST3, O3:K6 vibrios, while Halobacteriovorax strains OS1, G3, and H4 produced small and/or mediumsized plaques in the majority of ST36 strains, with occasional large plaques as well (Fig. 1). Based on the larger plaques on the ST36 strains than on the ST3 strains, it appears that Halobacteriovorax more efficiently attacked or replicated in ST36 than in ST3 strains. This suggests that some Vibrio sequence types may contain resistance factors against certain predators.
Large plaques were most often produced by Halobacteriovorax G3 and were about twice as likely to form with G3 as with OS1 or H4. Large plaques were also produced by OS1, G3, and H4 in the original host V. parahaemolyticus RIMD 2210633. No large or medium plaques were produced by S11 in the Vibrio strains, except in the original host Vibrio, which had small and medium plaques. Some predators, like S11, may invade and/or reproduce within the vibrios more slowly, perhaps due to suboptimal growth conditions for a particular isolate. Previously, we showed that the salinity of the plaque assay medium can affect plaque sizes, with smaller plaques formed when the salinity of the medium is higher than the salinity of the seawater from which the Halobacteriovorax bacteria were originally obtained (5). In the present study, plaque assay medium contained approximately 3% NaCl (30 ppt) and the salinities of the seawater from which the S11 and OS1 strains were originally obtained were 23.1 ppt and 25.4 ppt, respectively. This 2.3-ppt difference in salinity may be the reason the OS1 plaques were occasionally larger than the S11 plaques. Both strains came from neighboring locations along the Delaware Bay (12).
Clear versus faint (cloudy) plaques. Plaques were clear unless designated faint (or cloudy) as indicated in Fig. 1 by green boxes labeled "F." The cloudy plaques likely represent incomplete killing of the host vibrios, perhaps due to the presence or development of Vibrio resistance factors by some of the V. parahaemolyticus strains. Faint plaques remained cloudy from the first visualization of the plaques (usually 3 to 4 days) until the end of the incubation period (7 days). The faint plaques were associated with three V. parahaemolyticus strains (K5528, K5435, and K5282) that were isolated from Georgia, Washington State, and Hawaii, respectively, following challenge with Halobacteriovorax isolates obtained from the Mid-Atlantic and Hawaii (OS1, S11, and H4). The Gulf strain (G3) did not produce faint plaques in any of the Vibrio strains. Faint plaques were present in one of the five strains of ST3 but none of the 11 strains of ST36. Faint plaques were present only in the absence of one or both V. parahaemolyticus hemolysins. More information about the effect of the lack of specific hemolysins on Halobacteriovorax infectivity is discussed below.
Influence of Vibrio hemolysins on Halobacteriovorax predation. Vibrio parahaemolyticus contains hemolysins that are potential virulence factors in humans. They include the thermostable direct hemolysin (TDH) and the TDH-related hemolysin (TRH), encoded by the tdh and trh genes, respectively. Results of host specificity studies were compared with the presence or absence of tdh and/or trh (Fig. 1). Halobacteriovorax successfully attacked most of the vibrios regardless of hemolysin presence or absence. Fourteen of the Vibrio strains were both tdh and trh positive, of which 13 strains were readily attacked by all four predators. The exception was CDC strain K5582, an ST631 strain, which was susceptible to only Halobacteriovorax G3. Three of the V. parahaemolyticus strains were both tdh and trh negative and were predated upon by all Halobacteriovorax strains (Fig. 1). Only five of the V. parahaemolyticus strains were tdh negative, four of which were strains isolated in Hawaii. Only one of the tdh-negative strains (CDC K4859 from Hawaii) was resilient to predation by three of the four predators. The other four vibrios were susceptible to attack by all four Halobacteriovorax strains. Five of the vibrios were tdh positive and trh negative, including the host Vibrio RIMD 2210633, and were all predated upon by the four Halobacteriovorax strains. Four of these were ST3, which is commonly observed to be tdh positive and trh negative (21,22). Conversely, only two of the vibrios were tdh negative and trh positive, of which one was attacked by all four predators but the other was attacked only by H4. Overall, it appears that these two hemolysin genes, either separately or together, do not prevent plaque formation on V. parahaemolyticus. However, faint plaques were produced only when either or both the tdh and/or trh genes were absent, specifically for Vibrio strains K5528 (tdh positive/trh negative), K5435 (tdh negative/trh positive), and K5282 (tdh negative/trh negative) from Georgia, Washington State, and Hawaii, respectively (Fig. 1). The same results were consistently obtained upon repeated assays. Therefore, it appears that the absence of these Vibrio hemolysins contributes to survival of a portion of the vibrios, the likely source of the cloudiness in three of the Vibrio strains. For other Vibrio hosts, the presence or absence of one or both hemolysins appears to have no effects on Halobacteriovorax infectivity, since plaques were clear, indicating total or nearly total elimination of the vibrios. Further study of possible resistance factors in some strains of V. parahaemolyticus and virulence factors in Halobacteriovorax strains is warranted.
Considerations of host specificity by Halobacteriovorax isolates. Previously, we assessed the predatory ability of Halobacteriovorax strains OS1, OR7, S11, and G3 to prey upon five strains of V. parahaemolyticus, two strains of Vibrio vulnificus, and a single strain of Vibrio alginolyticus (5). The Halobacteriovorax strains were predatory toward all five V. parahaemolyticus strains tested at that time but were not predatory toward the other Vibrio species, thus demonstrating that these Halobacteriovorax strains have a strong affinity for V. parahaemolyticus. That was likely due, in part, to the fact that our Halobacteriovorax strains were originally isolated on V. parahaemolyticus strain RIMD 2210633. If we had originally isolated Halobacteriovorax strains specific for some other genus or species of bacterium, then they may not have shown an affinity toward V. parahaemolyticus strains, or their affinities may have been significantly reduced. Thus, the initial organism upon which Halobacteriovorax strains are isolated will likely bias the results toward selection of genus-or species-specific Halobacteriovorax strains. Clearly, other bacterial strains can serve as prey to Halobacteriovorax (reviewed in reference 1).
Halobacteriovorax bacteria as potential biocontrol agents. The broad specificity of these Halobacteriovorax isolates toward V. parahaemolyticus from diverse and geographically distinct regions of the United States is welcome news for the development of biocontrol measures against pathogenic strains of V. parahaemolyticus. These findings justify our further efforts to formulate a cocktail (mixture) of Halobacteriovorax strains for treating Vibriocontaminated shellfish and other seafoods with the goals to reduce V. parahaemolyticus contamination and render the products safer to consume. Such treatments could have broad applications in reducing V. parahaemolyticus in market oysters, clams, and mussels using depuration, a commercial process employed internationally to reduce potential pathogens in contaminated shellfish (reviewed in reference 23). A simple pretreatment of the shellfish prior to depuration could be effective in reducing V. parahaemolyticus levels. Likewise, a simple dip of fish products in a solution containing Halobacteriovorax might safely kill vibrios on the surface of fish fillets without the need for either disinfecting chemicals or heat treatment. Based on our testing here, a cocktail containing Halobacteriovorax strains G3 and H4 should effectively target all the V. parahaemolyticus strains used in this study. Previous pilot studies have successfully demonstrated the efficacy of predatory bacteria in reducing bacterial pathogens in experimental infections in laboratory rats (24), plants (25), and seafoods, including fish (26,27) and shellfish (28). Further research on the use of Halobacteriovorax and other predatory bacteria as potential biocontrol agents in food processing, disease prevention and treatment, and farming and aquaculture operations is warranted.
Summary. The specificities of four strains of Halobacteriovorax from the Mid-Atlantic, Gulf Coast, and Hawaii were determined on 23 clinical strains of V. parahaemolyticus obtained from regions around the United States. Broad predatory activity was observed with 21 or 22 of the 23 strains serving as prey. The remaining two vibrios appeared resistant to three of the four Halobacteriovorax strains, and reasons for this resistance are uncertain. Halobacteriovorax species obtained from three distinctly different geographic locations were readily capable of preying upon V. parahaemolyticus isolated from both similar and remote habitats. All five ST3 and 11 ST36 vibrio strains were susceptible to attack by the four Halobacteriovorax strains, providing some hope that these pandemic strains may be moderated, at least partially, by naturally occurring Halobacteriovorax in the marine environment. Plaques were generally larger in the ST36 than the ST3 strains, suggesting more rapid predator entry into its prey and/or faster replication and growth within ST36 strains. The Gulf Coast strain of Halobacteriovorax (G3) commonly produced the largest plaques. This is good news, since V. parahaemolyticus levels nationally are highest in the warmer waters of the Gulf, where the more rapidly replicating G3 strain is present. Finally, we saw no evidence that the presence or absence of the well-known V. parahaemolyticus hemolysin genes tdh and trh has any effects on Halobacteriovorax-induced plaque formation, although the lack of one or both genes may have led to the formation of faint plaques, suggesting only partial death of three of the Vibrio strains. Although the abovedescribed studies cover only a thin slice of the many tasks needed to define Halobacteriovorax interactions with V. parahaemolyticus and other pathogenic species, our work provides strong evidence that pathogenic strains of V. parahaemolyticus are likely to be readily susceptible to Halobacteriovorax predation within the marine environment.

MATERIALS AND METHODS
Sources of bacterial isolates. Halobacteriovorax strains were obtained from seawater from the coast of Alabama in the Gulf of Mexico (strain G3), riverine sites along the U.S. Mid-Atlantic coast of the Delaware Bay (strains OS1 and S11), and Keyhole Point near Kailua-Kona, HI (strain H4) (12,13). They were isolated on V. parahaemolyticus strain RIMD 2210633, which was previously sequenced by Makino et al. (29) and for which chromosome sequences were entered into GenBank under accession numbers BA000031 and BA000032. This Vibrio is a pandemic strain isolated from Japan in 1996 and previously shown to be an ST3, serotype O3:K6 strain that is tdh positive and trh negative (29). Twenty-three clinical strains of V. parahaemolyticus were kindly provided by the U.S. Food and Drug Administration (FDA) under a material transfer agreement (MTA). These strains were previously described by Jones et al. (9) and Miller et al. (11) and are listed in Fig. 1. The isolates caused illnesses among individuals in a variety of U.S. states (Fig. 1). The V. parahaemolyticus and Halobacteriovorax isolates were maintained at 280°C as 30% glycerol stocks.
Halobacteriovorax enrichments. The Halobacteriovorax strains were enriched in their original host V. parahaemolyticus RIMD 2210633 in previously autoclaved and 0.22-mm-filtered, 30-ppt natural seawater for 24 to 48 h. After incubation, each enrichment was filtered through a 0.45-mm Acrodisc syringe filter (Pall Corp., Timonium, MD) and then diluted 1:1,000 in sterile 30-ppt natural seawater to dilute out any remaining bacterial host cells to extinction in accordance with our previously published enrichment, filtration, dilution technique (5). Although filtration is usually sufficient to separate non-Vibrio host cells from the smaller Halobacteriovorax, filtration alone was shown to be ineffective in removing Vibrio minicells, which can pass through 0.45-mm filters (5). The 1:1,000 dilution of Halobacteriovorax was tested for residual vibrios by pour plate technique by plating 100 mL of each enrichment in LB agar containing 3% NaCl. This served as a control to ensure the Halobacteriovorax strains were not contaminated with their original host Vibrio prior to host specificity testing of the predator against new strains of V. parahaemolyticus. Any carryover of the original host Vibrio to new Vibrio strains has the potential to give false-positive results in host specificity assays.
PCR and 16S rRNA gene sequencing of Halobacteriovorax isolates. To ensure the four predators selected for analysis were all different strains of Halobacteriovorax, PCR primers were designed to amplify different Halobacteriovorax strains, but not residual host Vibrio DNA that could be present in the filtered enrichments, either from Vibrio minicells or from DNA from lysed vibrios, or both. Using the three named species of Halobacteriovorax in GenBank, ( Predator-prey specificity assays. In preparation for specificity assays using a double agar plaque assay procedure, the vibrios were subcultured by inoculation onto Difco thiosulfate citrate bile salts sucrose agar (TCBS; Becton, Dickinson and Company, Sparks, MD) plates and incubated overnight at 26°C. A colony was then picked and enriched in 10 mL of Difco Luria-Bertani (LB) broth (Becton, Dickinson and Company) containing 3% NaCl. Cultures were incubated at 37°C at 200 rpm to an optical density at 600 nm (OD 600 ) of ;0.2. While the vibrios were incubating, bottom and top agars were prepared for plaque assays using autoclaved and 0.22-mm-filtered, 30-ppt natural seawater from the Delaware Bay. Bottom agar consisted of Bacto agar (15 g/L) and 0.1% polypeptone peptone medium (Pp20), both from Becton, Dickinson and Company, while the top agar consisted of Bacto agar (7.5 g/L) and 0.1% Pp20. After autoclaving, both agars were cooled to 50°C in a water bath. To prepare bottom agar plates, 20 mL was pipetted into sterile petri dishes. For top agar, 7.5 mL of autoclaved top agar was pipetted into sterile tubes. All tubes were maintained in the water bath at 50°C. After the bottom agar solidified, specificity testing on the Vibrio cultures was initiated when they reached an OD 600 of ;0.2. At this time, 6.5 mL of sterile seawater, 100 mL of the filtered and 1:1,000-diluted Halobacteriovorax enrichments, and 1 mL of the V. parahaemolyticus culture were combined with the 7.5 mL of top agar, mixed by gentle inversion to prevent bubble formation, and poured over the bottom agar layer. After solidification, plates were inverted, incubated at 26°C, and monitored daily for the formation of plaques for 5 to 7 days, after which plaque presence or absence and appearance were recorded along with relative plaque sizes. Plates that contained pinpoint, small, or faint plaques were incubated for the full 7 days.
Specificity comparisons. Comparisons were drawn to evaluate (i) the ability of predators to invade and kill different V. parahaemolyticus strains, (ii) possible geographic preferences for predators from one area to infect vibrios from the same or other, more distant areas, (iii) the relative sizes and appearances of the plaques produced by different predatory strains on different Vibrio strains, (iv) predator-prey interactions based on Vibrio sequence types, (v) predator-prey interactions based on Vibrio serotypes, and (vi) predator-prey interactions as affected by the presence or absence of Vibrio tdh and/or trh hemolysin genes.
Data availability. Halobacteriovorax strains used in this study are available through a material transfer agreement (MTA) with the U.S. Department of Agriculture (USDA), while Vibrio strains used are available through an MTA with the U.S. Food and Drug Administration (FDA). Partial 16S rRNA sequences for the four Halobacteriovorax strains used in this study are provided in Table S1 in the supplemental material.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, DOCX file, 0.01 MB.

ACKNOWLEDGMENTS
This work was supported by USDA, ARS in-house funds under CRIS project number 8072-42000-090-000D.
Vibrio parahaemolyticus strains were generously provided by the U.S. Food and Drug Administration.
The use of trade names or commercial products in this publication is solely for the purpose of providing specific information and does not imply recommendation or endorsement by the U.S. Department of Agriculture (USDA) or the U.S. Food and Drug Administration (FDA).