Zebrafish as a model for Vibrio parahaemolyticus

Microbiology (2013), 159, 2605–2615

DOI 10.1099/mic.0.067637-0

Zebrafish as a model for Vibrio parahaemolyticus virulence Rohinee N. Paranjpye, Mark S. Myers, Evan C. Yount3 and Jessica L. Thompson4 Correspondence Rohinee N. Paranjpye

Northwest Fisheries Science Center, National Marine Fisheries Service, National Oceanic and Atmospheric Administration, 2725 Montlake Boulevard E, Seattle, WA 98112, USA

[email protected]

Received 11 March 2013 Accepted 19 September 2013

Vibrio parahaemolyticus is a Gram-negative, naturally occurring marine bacterium. Subpopulations of strains belonging to this species cause an acute self-limiting gastroenteritis in humans and, less commonly, wound infections. In vivo models to differentiate avirulent and virulent strains and evaluate the pathogenic potential of strains of this species have been largely focused on the presence of known virulence factors such as the thermostable direct haemolysin (TDH), the TDH-related haemolysin (TRH) or the contributions of the type 3 secretion systems. However, virulence is likely to be multifactorial, and additional, yet to be identified factors probably contribute to virulence in this bacterium. In this study, we investigated an adult zebrafish model to assess the overall virulence of V. parahaemolyticus strains. The model could detect differences in the virulence potential of strains when animals were challenged intraperitoneally, based on survival time. Differences in survival were noted irrespective of the source of isolation of the strain (environmental or clinical) and regardless of the presence or absence of the known virulence factors TDH and TRH, suggesting the influence of additional virulence factors. The model was also effective in comparing differences in virulence between the wild-type V. parahaemolyticus strain RIMD2210633 and isogenic pilin mutants DpilA and DmshA, a double mutant DpilA : DmshA, as well as a putative chitin-binding protein mutant, DgbpA.

INTRODUCTION Vibrio parahaemolyticus is a naturally occurring Gramnegative pathogen that is responsible for the majority of bacterial infections due to consumption of seafood (Mead et al., 1999; Newton et al., 2012). Infection is usually acquired by consumption of raw or undercooked shellfish, particularly oysters. The bacterium can cause a severe but self-limiting watery diarrhoea often with abdominal cramping, nausea, vomiting, fever and chills. These symptoms usually occur within 24 h of ingestion, with illness lasting approximately 3–4 days. Severe disease is rare and occurs more commonly in persons with weakened immune systems. Wound infections can occur when an open wound is exposed to seawater (CDC, 2005; Johnson et al., 1984). Reports of such infections have been increasing in recent years (Weis et al., 2011). 3Present address: LifeLong Medical Care, Oakland, CA, USA. 4Present address: Alaska Fisheries Science Center, National Marine Fisheries Service, National Oceanic and Atmospheric Administration, Seattle, WA, USA. Abbreviations: TDH, thermostable direct haemolysin; TRH, TDH-related haemolysin; T3SS, type 3 secretion system; AIC, Akaike Information Criterion; REP-PCR, repetitive extragenic palindromic sequence-based PCR.

067637

Printed in Great Britain

The thermostable direct haemolysin (TDH) and the TDHrelated haemolysin (TRH) have been regarded as the major virulence factors of this pathogen. TDH exhibits enterotoxic activity in vivo as well as cytotoxic effects on specific cell lines. TRH has a more dominant role in cytotoxicity (Honda & Iida, 1993; Nishibuchi et al., 1992; Park et al., 2004). The rabbit or rat ileal loop model has been traditionally used to detect fluid accumulation due to the enterotoxic activity of TDH in vivo (Ljungh & Wadstrom, 1982; Miyamoto et al., 1980; Takeda, 1982). Sequencing of the genome of V. parahaemolyticus RIMD2210633, a tdh+ strain, identified the presence of two type 3 secretion systems (T3SSs), T3SS1 and T3SS2 (Makino et al., 2003). Studies conducted in vitro have shown that T3SS1 induces cytotoxicity in several cell lines while T3SS2 can induce cytoxicity in Caco-2 cells and enterotoxicity in vivo (Kodama et al., 2008; Park et al., 2004). In a recent extensive study undertaken to clarify the role of TDH and the two T3SSs, T3SS2 rather than TDH was suggested to be the major contributor to enterotoxicity; however, both T3SS1 and TDH contributed to lethality in a mouse model, suggesting that TDH and/or other proteins(s) that are part of T3SS1 are also likely to contribute to virulence (Hiyoshi et al., 2010). More recently, two different animal models were used to assess the differential contributions of T3SS1 and T3SS2 to V. parahaemolyticus virulence (Pin˜eyro et al., 2605

R. N. Paranjpye and others

2010). Using neonatal pigs inoculated orogastrically, the authors demonstrated that T3SS2, but not T3SS1, was required for gastrointestinal disease (acute diarrhoea and vomiting). In the same study, the authors also showed that T3SS1 was necessary to cause mortality in mice when inoculated via the intrapulmonary route, raising the possibility that an unrecognized T3SS1-dependent exotoxigenic factor is involved in lethality (Pin˜eyro et al., 2010). In another study, infant rabbits challenged orogastrically were used to assess the ability of V. parahaemolyticus to damage host intestines (Ritchie et al., 2012). In most cases, these animal models require specialized facilities and have been primarily useful to examine the contribution of specific toxins or effectors in the pathogenesis of V. parahaemolyticus. Zebrafish (Danio rerio) have been successfully used as a model to study the virulence mechanisms utilized by a number of Gram-negative as well as Gram-positive pathogens of fish and humans (Brenot et al., 2004; Davis et al., 2002; Goldberg et al., 2008; Menudier et al., 1996; Moyer & Hunnicutt, 2007; Neely et al., 2002; O’Toole et al., 2004; Phelps et al., 2009; Pressley et al., 2005; Prouty et al., 2003; Rawls et al., 2007; Rojo et al., 2007). The presence of an innate and adaptive immune system and its genetic tractability, as well as the ease of breeding and laboratory handling make zebrafish an ideal host model for analysing bacterial infections (Stemple & Driever, 1996; Sullivan & Kim, 2008; Zon, 1999). The species V. parahaemolyticus includes strains that are genetically very diverse, with multiple factors involved in its pathogenicity. In this study we used the zebrafish model to determine the overall differences in virulence of V. parahaemolyticus strains from clinical and environmental sources. The model was also used to assess differences in virulence between a wild-type V. parahaemolyticus strain and isogenic mutants of three surface-expressed proteins that could potentially have a role in virulence: two type IV pilins, pilA and mshA, and a surface-expressed colonization factor gbpA that links environmental survival and human infection in Vibrio cholerae (Kirn et al., 2005). V. parahaemolyticus encodes two type IV pilins, PilA (or ChiRP) and MshA, homologous to corresponding genes in Vibrio vulnificus and V. cholerae (Chiavelli et al., 2001; Fullner & Mekalanos, 1999; Paranjpye & Strom, 2005; Watnick et al., 1999). In a previous study, we have shown that PilA in V. vulnificus contributes to virulence in an iron-overloaded mouse model, and deletion of PilA resulted in a 1-log increase in the LD50 as compared with the wild-type V. vulnificus strain (Paranjpye & Strom, 2005). In V. cholerae, the MshA pilus contributes to adherence to plankton and to formation of biofilms and thus is important in environmental survival; however, this pilus does not appear to have a role in virulence (Thelin & Taylor, 1996; Watnick et al., 1999). A third protein, GbpA, in V. cholerae has a dual role in attachment to plankton and human epithelial cells (Kirn et al., 2005). In the present 2606

study, we assessed the virulence of PilA, MshA and GbpA in V. parahaemolyticus using the zebrafish model.

METHODS Bacterial strains and growth conditions. Bacterial strains used in

this study are listed in Table 1. V. parahaemolyticus strains were grown overnight (18 h) in trypticase soy broth (TSB) with a final NaCl concentration of 2 %, at 25 uC, while Escherichia coli cells were grown at 37 uC, both with shaking at 150 r.p.m. All cultures were adjusted with TSB to an OD600 of 1.0. Cultures were washed twice with PBS before use as an inoculum. Serial 10-fold dilutions of cultures were plated on tryticase soy agar (TSA; 2 % NaCl) to confirm the concentration of the inoculum. When heat-killed V. parahaemolyticus RIMD2210633 was used as a control, cultures were heattreated at 60 uC for 2 h followed by two washes with PBS. Heat-killed cultures were plated on TSA and confirmed to be non-viable. V. parahaemolyticus strains with deletion mutations in pilA, mshA and gbpA were constructed by the strategy for tagged in-frame mutations using crossover PCR as outlined (Link et al., 1997; Weis et al., 2011) and described by Frischkorn et al. (2013). In constructing the mutations, 366 bp of pilA, 444 bp of mshA and 1410 bp of gbpA were deleted leaving only 18 bp of the 59 end and 36 bp of the 39 end of each gene. The double mutant, V. parahaemolyticus DpilA : DmshA, was generated by conjugating the plasmid with DmshA into V. parahaemolyticus DpilA. V. parahaemolyticus DgbpA was constructed by the same technique but obtained from S. Moseley (University of Washington, Seattle, WA). Complemented mutant strains, V. parahaemolyticus DpilA(+pilA), V. parahaemolyticus DmshA(+mshA) and V. parahaemolyticus DgbpA(+gbpA) were constructed by conjugating the respective plasmids with the cloned genes (pilA, mshA, gbpA) into the respective mutant strains (Frischkorn et al., 2013; Table 1). Complementation of the mutant strains was verified by measuring the ability of the strains to form biofilms on abiotic surfaces and by adherence to chitin beads (Frischkorn et al., 2013) using published protocols (O’Toole et al., 1999). Zebrafish care and infection. Zebrafish (WT AB) were raised at

our facility at the Northwest Fisheries Science Center and were between 5 and 6 months old. In some experiments used for comparison zebrafish (3–3.5 cm in length and 0.4–0.5 g) were obtained from a commercial supplier and maintained in separate tanks. Care and feeding of zebrafish followed established protocols (http://zfin.org/zf_info/zfbook/zfbk.html). Intraperitoneal challenge. For challenge experiments, zebrafish

were injected intraperitoneally midway between the pectoral fin and the anus with 10 ml of the inoculum using a repeater dispenser (Hamilton #83700) and a 33 gauge needle (Hamilton 1750LTSN, 33/ 0.3759/PT4) following previously published protocols (Lefebvre et al., 2009). Zebrafish were housed in glass aquaria (76 litres) in deionized water supplemented with aquarium salts (Instant Ocean; Aquarium Systems) at a concentration of 60 mg l–1, and transferred to 19 litre aquaria with the same water prior to inoculation. Water temperature was maintained at 25 uC. Each tank was equipped with an air filter. Zebrafish were not fed on the day of inoculation. Tris-buffered tricaine at a concentration of 320 mg ml21 was used to kill the fish on completion of the experiment. All aquaria and water were disinfected with 20 % sodium hypochlorite after each experiment. Filters from air pumps were sterilized by autoclaving. Virulence evaluation based on survival time and statistical analysis. In experiments where strains were compared for differences

in virulence based on survival time, ten zebrafish were inoculated with each strain at a concentration of 106 c.f.u. per 10 ml of the inoculum Microbiology 159

Zebrafish model for Vibrio parahaemolyticus virulence

Table 1. Bacterial strains used in study Strain

Relevant characteristics

V. parahaemolyticus RIMD2210633 RIMD DpilA RIMD DmshA RIMD DpilA : DmshA RIMD DgbpA

Wild-type Wild-type Wild-type Wild-type Wild-type

(clinical isolate) with disrupted pilA with disrupted mshA with disrupted pilA and mshA with disrupted gbpA

RIMD DpilA(+pilA) RIMD DmshA(+mshA) RIMD DgbpA(+gbpA)

RIMD DpilA complemented with pilA RIMD DmshA complemented with mshA RIMD DgbpA complemented with gbpA

and observed every 30 min for the first 10 h and terminated after 5 days. Heat-killed V. parahaemolyticus RIMD2210633 were also included in the analysis. Experiments were repeated at least once. Control fish injected with 10 ml PBS were included with each experiment. Survival fractions (median survival times) were determined using the product limit (Kaplan–Meier) method and the survival curves compared using the logrank test (GraphPad Prism version 4). A P-value .0.05 indicated that the survival curves were not significantly different. The inoculum level of 106 c.f.u. per 10 ml, slightly above the LD50 of the wild-type strain RIMD2210633, allowed for precise monitoring and observations. When differences between the wild-type strain and isogenic mutant strains were compared, groups of five zebrafish were injected intraperitoneally with serial 10-fold dilutions of the inoculum, ranging from 103 to 108 c.f.u. of the V. parahaemolyticus strains being compared. Fish were observed for 5 days after challenge. Experiments were repeated three times. The LD50 was calculated using the method of Reed & Muench (1938). Experiments were also analysed using logistic regression where replicate, log dose and strain are all included as fixed effects (Venables & Ripley, 2002). For each experiment, we tested for differences between X and Y and X and Z; to assess the suitability of this model we used the Akaike Information Criterion (AIC) to compare it with simpler models without strain, and replicate, and more complex models with interactions between replicate and log dose, and strain and log dose. If the simpler or more complex models provided a better fit (lower AIC), we accounted for this when comparing the strains. Histology Zebrafish were examined grossly for any externally visible

abnormalities and then killed with MS 222 just before they succumbed to infection. Fish were then carefully opened with a midline incision of the abdomen by fine dissecting scissors to expose the internal organs to fixative, and the entire fish was preserved in Dietrich’s fixative (Fournie et al., 2000) at approximately a 1 : 20 (v/v) tissue to fixative ratio. To ensure uniform and complete fixation, tissues were fixed for 3 days on a rotor, or other agitating device. After fixation for 3 days, the samples were removed from fixative, rinsed in two or three changes of water, and placed in 70 % ethanol. The fish were then bisected along the sagittal plane using a fresh razor blade, making the cut parallel to and on the left side of the spinal cord, before they were processed. The bisected fish were then loaded into cassettes for processing, using a Shandon Hypercenter XP. The tissue processing protocol followed that specified by the Zebrafish International Resource Center, University of Oregon, Eugene, OR (http://zebrafish.org/zirc/health/diseaseManual.php). Processed tissues were infiltrated with a 50 : 50 ratio of Fisher Paraplast PLUS tissue embedding medium and Surgipath Formula ‘R’ infiltrating and embedding paraffin. Tissues were cut in the sagittal plane in step sections at 5–7 mm thicknesses with a high-profile disposable blade. Sections of gills, eyes, brain, spinal cord and all internal organs were http://mic.sgmjournals.org

Reference/source Makino et al. (2003) Frischkorn et al. (2013) Frischkorn et al. (2013) Frischkorn et al. (2013) S. Moseley (University of Washington, Seattle, WA) Frischkorn et al. (2013) Frischkorn et al. (2013) Frischkorn et al. (2013)

taken to ensure proper and complete histopathological evaluation. Sections were routinely stained by haematoxylin and eosin, using the protocols as described in the document Diseases of Zebrafish in Research Facilities provided by the Zebrafish International Research Center, as above. Sections were examined by light microscopy on a Nikon Optiphot microscope, and selected sections were photographed by digital photomicroscopy on a Nikon Eclipse E600 microscope, using a SPOT camera and SPOT software, by an experienced fish histopathologist (M. S. M.). All major organs were examined for pathological changes, with a focus on alterations in the erythrocytes.

RESULTS V. parahaemolyticus can establish an infection in zebrafish In initial experiments, zebrafish (n510) were injected intraperitoneally with serial 10-fold dilutions of the wildtype strain V. parahaemolyticus RIMD2210633 at doses from 16103 to 16108 c.f.u. The abdominal cavity and the area surrounding the site of injection showed haemorrhaging about 4 h after inoculation, and mortalities were observed in fish challenged with the higher doses beginning 7 h after inoculation. No haemorrhaging was observed in the zebrafish injected with PBS alone. Similarly, no pathological signs or mortality were observed in zebrafish inoculated with heat-killed V. parahaemolyticus or with a non-pathogenic strain of E. coli (DH5a). The survival of zebrafish was dose dependent and LD50 values were similar using zebrafish of the wild-type AB strain (LD5055.466105) and those from a commercial source (LD5055.736105), suggesting that V. parahaemolyticus can infect zebrafish in a reproducible and dose-dependent manner. Histopathology of V. parahaemolyticus intraperitoneal infection As V. parahaemolyticus are not natural pathogens of fish it was necessary to determine the pathology of infection in this model. Zebrafish were injected intraperitoneally with the wild-type V. parahaemolyticus RIMD2210633 strain at a concentration of 16108 c.f.u. per 10 ml and killed at intervals of 0, 3, 6, 9 and 10 h post-inoculation. Controls were injected with 10 ml PBS. A higher inoculum was used 2607


for histological analysis than the LD50 of the wild-type V. parahaemolyticus strain to ensure that the progression of infection could be recorded within 1 day and samples for post-inoculation histological analyses collected prior to degradation of tissues due to death. The two midsagittal halves of each fish were sectioned for histological analysis and tissue sections of 3–7 zebrafish were examined at each time interval. Upon dissection, severe bleeding was noted in the abdominal cavity with disintegration of the organs. Histological examination of tissue sections revealed significant damage to the erythrocytes in the vasculature, spleen, choroid gland in the eye and to the skeletal muscle fibres in the retroperitoneal muscle (Table 2). Within 3 h after infection, over half the fish displayed alterations of the cell membrane, rounding of the nuclei and crenation of the erythrocytes, and significant coagulative necrosis of the muscle fibres in the retroperitoneal muscle. At 3 h, the spleen was affected in over a third of the fish, with increased plasma proteins and haematin pigment present in the ellipsoids; the tissue sections of the chorion around the eye indicated rounding and pyknosis of erythrocyte nuclei in 40 % of the fish examined (Table 2). At 6 h postchallenge, the area around the abdominal cavity grossly displayed considerable haemorrhaging and ascites in some cases. By 6 h, 100 % of the fish showed obvious lysis of erythrocytes with naked erythrocytic nuclei in the general vasculature, which continued for the duration of the experiment. Significant coagulative necrosis of the retroperitoneal muscle was also evident in all the fish examined at 6 h. At this time point, increased plasma proteins in ellipsoids and haematin pigment in the splenic pulp was noted in two-thirds of the samples examined and in all samples 8 h after challenge (Table 2). In the choroid gland

at 6 h, rounding and pyknosis of red blood cell nuclei in the vasculature had increased to 60 % (Table 2). Additionally, erythrocyte lysis was observed in the haemopoietic kidney in 25 % of the fish at 10 h postinjection. In summary, histological changes in fish injected intraperitoneally with V. parahaemolyticus included degeneration and eventual lysis of erythrocytes, increased vascular plasma proteins and coagulation necrosis of skeletal muscle in the retroperitoneal region. These conditions were more severe and more prevalent as the post-exposure period increased from 3 to 10 h postinjection (Table 2). None of these alterations to the erythrocytes in the organs examined or coagulation necrosis in the retroperitoneal musculature was present in the ten control fish examined. We also compared the pathology of infection in moribund zebrafish injected intraperitoneally with V. parahemolyticus strains with or without the haemolysin genes (tdh and trh) (Table 3). The data in this table suggest that the prevalence of early, less severe effects on erythrocytes (rounding of nuclei, cell membrane alterations, crenation) was not affected by the presence or absence of the TDH or TRH haemolysins in V. parahemolyticus, with nearly all injected fish showing these changes to the erythrocytes, regardless of genotype. The only statistically significant difference (by Fisher’s exact test) was the prevalence of the less severe changes (membrane lysis, nuclei with marginated chromatin, or condensed, fragmented, pyknotic or naked nuclei) in erythrocytes in the general vasculature of fish injected with tdh2/trh+ isolates (P50.025). In fish injected with the tdh2/trh+ genotype, there was a general trend for reduced prevalence of increased plasma proteins

Table 2. Frequency of occurrence (number of fish affected) of histopathological features and lesions in adult zebrafish at various intervals following intraperitoneal injection of V. parahaemolyticus RIMD2210633 (1¾108 c.f.u. per 10 ml) None of the features/lesions was detected in the five control zebrafish examined.

ND,

Not done. RBC, red blood cell. Time post-inoculation

Vasculature erythrocyte (RBC) changes Rounding of RBC nuclei, cell membrane alterations, crenation Lysis of RBC, naked RBC nuclei Increased plasma proteins Spleen Increased plasma proteins, haematin pigment in ellipsoids Congestion Calcified cell debris Choroid in eye Rounding of RBC nuclei, pyknotic RBC nuclei Obvious RBC lysis Kidney Obvious lysis of RBCs in haemopoietic tissue Retroperitoneal muscle Coagulative necrosis

2608

3h

6h

8h

9h

10 h

3/5 1/5 4/5

5/5 5/5 4/5

7/7 6/7 7/7

3/3 3/3 2/3

4/4 4/4 4/4

1/3 0/3 0

2/3 1/3 0

5/5 0/5 0

ND

ND

2/2 0/2 1/2

2/5 0/5

3/5 0/5

6/7 1/7

3/3 2/3

3/4 3/4

0/5

0/5

0/7

0/3

1/4

3/5

5/5

6/7

3/3

3/4

ND

Microbiology 159


Table 3. Comparison (frequency of occurrence or number of fish affected) of histopathological features and lesions in moribund adult zebrafish following intraperitoneal injection of V. parahaemolyticus strains with or without the haemolysins (tdh or trh) None of the features/lesions was detected in the control zebrafish examined. V. parahaemolyticus strains tdh+/trh+ Vasculature, erythrocyte (RBC) changes; systemic Rounding of RBC nuclei, cell membrane alterations, crenation Lysis of RBCs, nuclei with marinated chromatin, or condensed, fragmented, pyknotic, naked RBC nuclei Increased plasma proteins Spleen Increased plasma proteins, haematin pigment in ellipsoids Congestion Choroid in eye Rounding, condensation, pyknosis of RBC nuclei Obvious RBC lysis Kidney Obvious RBC lysis in haemopoietic tissue Retroperitoneal muscle Coagulative necrosis

tdh”/trh+

tdh”/trh”

4/6 5/6

10/10 2/10*

4/4 3/4

5/6

5/10

3/4

4/4 4/4

2/4 3/4

0/2 0/2

4/6 2/6

7/10 0/10

4/4 0/4

2/6

0/10

0/4

6/6

6/10

4/4

*Significantly lower prevalence than in the tdh+/trh+ strain, by Fisher’s exact test, P50.034.

in the general vasculature, increased plasma proteins, congestion and haematin pigment in ellipsoids of the spleen, and obvious erythrocyte lysis in the chorion and blood vessels of the kidney, as compared with the tdh+/ trh+ genotype. Similarly, in fish injected with V. parahemolyticus lacking both haemolysins (tdh2/trh2) the tissues in the spleen did not appear to be affected. There was also reduced prevalence of obvious lysis of erythrocytes in the chorion and blood vessels of the kidney, as compared with the tdh+/trh+ genotype. However, there was no discernible pattern of alteration in the prevalence of coagulative necrosis of the retroperitoneal musculature among the three V. parahaemolyticus strains, with nearly all fish affected with this lesion. Comparison of the virulence potential of genotypically diverse V. parahaemolyticus strains In a previous study we compared the genetic diversity of 149 clinical and environmental isolates from the Pacific north-west and other geographical regions (Paranjpye et al., 2012). The study examined the variation in 20 genes, which included the virulence-associated genes (tdh, trh, urease), genes associated with T3SS1 and T3SS2, and the pandemic virulence markers (orf8, toxRSnew). All strains encoded T3SS1 and T3SS2 with the sequences of several genes representative of either T3SS2a or TTSS2b (Makino et al., 2003; Okada et al., 2009). Results from a second study conducted in our laboratory show that these isolates represent different sequence types based on MLST and repetitive extragenic palindromic sequence-based PCR (REP-PCR) profiles (Turner et al., 2013). http://mic.sgmjournals.org

In this study, we evaluated the pathogenic potential of 23 V. parahaemolyticus clinical and environmental isolates that were part of the two previous studies (Table 4). These isolates represented different genetic groups, MLST sequence types and REP-PCR groups based on our previous work (Paranjpye et al., 2012; Turner et al., 2013). Groups of ten zebrafish were injected intraperitoneally with equal concentrations of each strain (106 c.f.u. per 10 ml) and their survival was compared as described in Methods. Survival curves for each strain tested were reproducible and in replicate experiments the survival curves were not significantly different when compared for each strain (P.0.05) (Table 4, Fig. 1). Survival curves of two representative strains of each genotype (tdh±/trh±) are presented in Fig. 1. The average median survival times for six of the nine clinical isolates that were tdh+ and trh+ (all from the Pacific north-west) were between 6.25 and 10.25 h. The other three clinical isolates (also from the Pacific north-west) were less virulent when tested in our model and had average median survival times between 22 and 43.5 h (Table 4). The difference in virulence of these isolates may reflect the immune status of the host or possibly the route of infection (wound versus gastroenteritis). The survival times for the three environmental isolates (one from oyster and two from water) that were also tdh+/trh+ were significantly lower (6.125–7.5 h) and comparable to the six clinical isolates, suggesting that these environmental isolates could potentially be pathogenic under the right conditions. Of the six tdh+/trh2 isolates tested, five were of environmental origin and showed distinct differences in virulence with average median 2609


Table 4. Comparison of survival times for zebrafish challenged with clinical and environmental strains of V. parahaemolyticus Ten zebrafish were challenged in each trial. Heat-killed V. parahaemolyticus RIMD2210633 were used as controls.

ND,

Not done.

Strain

Source

tdh

trh

toxRS new

orf8

Mean survival times (±SEM, h)

P-value*

No. of trials

REP-PCR group

MLST (ST)D

12315 EN9701173 EN9901310 846d SRPC10290d EN2910 260 50 9710290 3644 EN9701072 EN9701121 551d 930 863 605 747 RIMD2210633d 48256 96Q 901128d 805d 970107

Clinical Clinical Clinical Oyster Clinical Clinical Water Water Clinical Clinical Clinical Clinical Water Net tow Net tow Net tow Water Clinical Clinical Clinical Clinical Water Oyster

+ + + + + + + + + + + + + + + + + + 2 2 2 2 2

+ + + + + + + + + + + + 2 2 2 2 2 2 + + 2 2 2

2 2 2 2 2 2 + 2 2 2 2 2 + + + + + + 2 + 2 2 +

2 2 2 2 2 2 2 2 2 2 2 2 + + + + + + 2 2 2 2 2

7.36±2.30 7.25±0.50 6.25±0.35 7.50±1.09 6.75±1.34 22.0±0 7.25±1.06 6.13±1.24 7.63±0.18 22.0±0 10.30±6.72 43.50±6.36 22.0±0 22.0±0 22.0±0 31.0±10.61 28.50±9.19 6.15±1.11 23.83±3.18 9.63±4.42 15.0±9.90 5.75±1.06 13.25±1.77

0.2119 0.2427 0.1417 0.7530 0.8824 0.8537 0.8751 0.563 0.8866 0.2406 0.0718 0.6512 0.3833 0.887 0.6017 0.4661 0.1704 0.5217 0.6862 0.9378 0.7866 0.3738 0.3691

2 2 2 3 8 2 2 2 2 2 2 2 2 2 2 2 2 5 3 2 2 2 2

1 1 1 1 1 1 4 4 8 8 8 13 2 2 2 2 2 2 14 31 11 25 19

36 36 36 36 36 36 34 34 43 43 43 50 3 3 3 3 ND

3 ND ND

135 ND

131

*P.0.05 indicates that survival times between trials were not significantly different. DSequence type. dSurvival curves shown in Fig. 1.

survival times between 22 and 31 h as compared with 6.15 h for the clinical isolate, RIMD2210633, which belonged to the same MLST sequence type and REP-PCR group. Two clinical isolates, 96Q and 48256, which were tdh2/trh+ differed in virulence based on survival times, again verifying that the model can potentially differentiate the degree of virulence between strains encoding the same haemolysins. Our analysis also included one clinical (strain 901128) and two environmental isolates (805 and 970107) that were tdh2/trh2 which had survival times between 5 and 22 h, suggesting the contribution of additional virulence factors besides the known haemolysins. No mortality was observed in any of the zebrafish injected with PBS or heat-killed V. parahaemolyticus RIMD2210633. Comparison of the survival curves of strains within the REP-PCR and MLST groups suggests differences based on virulence unrelated to source of isolation (Table 4). For example, one of the six V. parahaemolyticus strains (EN2910, a clinical isolate) that were part of the same REP-PCR group (group 1) and MLST sequence type (ST 36) had a higher median survival time than other strains in the same group. Similarly, three environmental strains from REP-PCR group 2 and MLST ST3 had higher median 2610

survival times as compared with the clinical strain RIMD2210633 (also ST3), again pointing to other factors that might contribute to differences in virulence. Similar differences were also noted in the other REP-PCR and MLST groups in the study (Table 4). These observations suggest that the zebrafish model can distinguish potential differences in the virulence of strains in addition to the variation based on known haemolysins, source of isolation or genetic differentiation methods such as REP-PCR and MLST. Detection of V. parahaemolyticus mutants with attenuated virulence To evaluate the ability of the zebrafish model to assess the contribution of putative virulence genes we compared survival of the wild-type V. parahaemolyticus strain with that of the pilin mutants, pilA and mshA, the double mutant pilA : mshA (Frischkorn et al., 2013) and a strain with a mutation in gbpA, a secreted attachment factor (Kirn et al., 2005). In separate experiments performed in triplicate, we first compared the virulence of the wild-type V. parahaemolyticus RIMD2210633 with V. parahaemolyticus DpilA and the complemented mutant V. parahaemolyticus DpilA(+pilA) strain. In a second set of experiments, we Microbiology 159


(a) Percentage survival

100

trial 1 trial 2 trial 3

75

75 50

25

25 0 0

10

20

30

40

50

trial 1 trial 2

(c) 100 Percentage survival

100

50

0

0

10

20

30

40

(d) 100

75

75

50

50

25

25

50

trial 1 trial 2 trial 3 trial 4 trial 5

0

0 0

10

20

30

40

(e) 100 Percentage survival

trial 1 trial 2 trial 3 trial 4 trial 5 trial 6 trial 7 trial 8

(b)

trial 1 trial 2

75

0

50

20

30

40

(f) 100

50

25

25

50

trial 1 trial 2

75

50

0

10

0 0

10 20 30 40 Time post-inoculation (h)

50

0

10 20 30 40 Time post-inoculation (h)

50

Fig. 1. Survival curves (Kaplan–Meir) of zebrafish following intraperitoneal challenges of V. parahaemolyticus strains (a) 846 (oyster isolate), (b) SPRC 10290 (clinical isolate), (c) 551 (water isolate), (d) RIMD2210633 (clinical isolate), (e) 805 (water isolate) and (f) 901128 (clinical isolate) all at a dose of 106 c.f.u. per 10 ml.

compared the wild-type strain, the DmshA and the double mutant DpilA : DmshA; and lastly we compared the wild-type strain with the DgbpA strain. Comparison of virulence as measured by LD50 is summarized in Table 5. Experiments were also analysed by logistic regression as described in Methods. In the first set of experiments, the wild-type strain was significantly different from DpilA (P50.001) but not different from the complemented pilA mutant (P50.365). When the wild-type strain was compared with the DmshA and the double mutant DpilA : DmshA, virulence of the wild-type strain was significantly different from that of the DpilA : DmshA strain (P50.00441) but not from that of the DmshA strain (P50.415). Finally, virulence of the DgbpA strain was not significantly different from that of the wild-type strain (P50.600). For both experiments, the model on which inference was based (fixed effects for replicate log http://mic.sgmjournals.org

dose and strain) had lowest AIC when compared with the simpler and more complex models. Histopathological observations of zebrafish inoculated with mutant strains Sagittal tissue sections of halved zebrafish were similarly examined in this histological study, and the observations (noted as the percentage of fish presenting the histopathological features or lesions) are summarized in Table 6. The prevalence of lesions was compared between fish injected with the wild-type strain at 8 h post-injection (when behaviour consistent with morbidity was first apparent) and fish injected with the mutant strains when morbidity first became apparent. Lesions in the general vasculature observed in fish injected with the wild-type strain – such as 2611


wild-type strain. The only significant (P,0.05) difference was reduced prevalence of necrosis of the retroperitoneal muscle and accumulation of plasma proteins in tissues of fish challenged with the mshA mutant strain as compared with the wild-type strain (Table 6). In the absence of both pilA and mshA, 50 % of the erythrocytes in the general vasculature and choroid in the eye showed significant differences as compared with fish injected with the wild-type strain. Both samples (n52) showed significant congestion of the spleen and there was a significant reduction in the prevalence of coagulative necrosis in the retroperitoneal musculature, as compared with fish injected with the wildtype strain (Table 6). In fish challenged with the DgbpA mutant there was also a reduction in changes in the vasculature, spleen and choroid in the eye as compared with those injected with the wild-type strain (Table 6). These differences were less pronounced than those observed in fish challenged with the pilA or pilA : mshA mutants. In summary, histological changes in the vasculature, spleen, choroid in the eye and kidney were significantly reduced in fish challenged with the pilA mutant as compared with those challenged with the wild-type strain.

Table 5. Comparison of virulence (LD50) of the wild-type V. parahaemolyticus RIMD2210633 strain with the type IV pilin and gbpA mutants (triplicate experiments) Strains compared WT and DpilA DpilA and DpilA(+pilA) WT and DpilA(+pilA) WT and DmshA WT and DpilA : DmshA WT and DgbpA

Mean (±SEM) difference in LD50 (log10) 0.974±0.169* 0.690±0.219* 0.335±0.220 0.227±0.133 0.780±0.089* 0.292±0.003

*Significantly different when compared by logistic regression (P,0.005).

rounding of erythrocyte nuclei with cell membrane crenation and other alterations (Fig. 2a), erythrocyte lysis and nuclear condensation, fragmentation, pyknosis or naked erythrocyte nuclei (Fig. 2a), increased plasma proteins and haematin pigment in the splenic ellipsoids, and obvious rounding of erythrocytes in the choroid gland (Fig. 2c) – were not observed in fish injected with the pilA mutant strain (Fig. 2b, d); in fact, the prevalence of all these lesions was significantly reduced in the pilA mutant strain as compared with that in fish injected with the wild-type strain (Table 6). The prevalence of most abnormalities in erythrocytes in the vasculature, spleen and choroid in the eye, in fish injected with the mshA mutant strain was less pronounced as compared with those injected with the

DISCUSSION Our study describes an easy-to-use animal model that is useful in assessing the virulence of V. parahaemolyticus isolates irrespective of the presence of specific known haemolysins, such as tdh and trh, or secretion systems (T3SS), thus allowing a useful screen for further studies.

Table 6. Comparison (frequency of occurrence or number of fish affected) of histopathological features and lesions in moribund adult zebrafish following intraperitoneal injection of wild-type V. parahaemolyticus or the type IV pilin (DpilA), DmshA, DpilA : DmshA and gbpA (DgbpA) mutant strains Zebrafish were inoculated with 106 c.f.u. per 10 ml. RBC, red blood cell. V. parahaemolyticus strain

Vasculature, erythrocyte (RBC) changes; systemic Rounding of RBC nuclei, cell membrane alterations, crenation Lysis of RBC, nuclei with marginated chromatin, or condensed, fragmented, pyknotic, naked RBC nuclei Increased plasma proteins Spleen Increased plasma proteins, haematin pigment in ellipsoids Congestion Choroid in eye Rounding, condensation, pyknosis of RBC nuclei Obvious RBC lysis Kidney Obvious RBC lysis in haemopoietic tissue Retroperitoneal muscle Coagulative necrosis

WT

DpilA

DmshA

DpilA : DmshA

DgbpA

7/7 6/7

0/3* 0/3*

4/5 4/5

1/2 0/2

1/4 1/4

7/7

0/3*

2/5*

0/2*

1/4

5/5 0/5

0/2* 0/2

2/3 2/3

0/1 1/1

0/2 0/2

6/7 1/7

0/3* 0/3

3/5 0/5

1/2 0/2

1/4 1/4

0/5

0/3

0/5

0/2

0/4

7/7

2/3

2/4*

1/2

4/4

*Significantly lower prevalence than in the wild-type strain, by Fisher’s exact test, P,0.05. 2612

Microbiology 159


(a)

(b)

(c)

(d)

Fig. 2. Histopathological lesions in moribund zebrafish after infection with wild-type (WT, tdh+/trh”) or various mutant strains of V. parahaemolyticus. (a, c) Abnormalities seen in fish injected with the WT strain. (a) Rounding of red blood cell nuclei (RBC) and obvious RBC membrane lysis (arrows) in the vasculature (bulbus arteriosis of heart) as compared with the oval/elliptical shape of the nuclei in intact RBCs present in the bulbus arteriosis of fish injected with the DpilA mutant strain (b). (c) Rounding, RBC membrane crenation and lysis of RBCs (arrows) in the choroid of the eye of a fish injected with wild-type bacteria, as compared with the oval/elliptical shape of the nuclei in normal, intact RBCs with obviously intact plasma membranes present in the choroid of fish injected with the DpilA mutant strain (d). Visible in the lower portion of this micrograph are several layers of the zebrafish retina. All photomicrographs shown are from tissues sections stained with haematoxylin and eosin. Bars, 50 mm.

The model could also effectively distinguish differences in isogenic mutants of type IV pilins, PilA and MshA, and wild-type strains. Zebrafish are easy to breed, reach sexual maturity by 3 months and have well-developed innate and adaptive immune systems. Moreover, zebrafish colonies can be set up relatively easily in any laboratory, making it a convenient animal model for preliminary analysis of V. parahaemolyticus virulence. The species V. parahaemolyticus is very diverse and complex and shows evidence of significant strain-specific http://mic.sgmjournals.org

differences, in addition to the presence or absence of the known virulence determinants tdh, trh, T3SS2 and the pathogenicity islands such as VPA1, that may contribute to the virulence of these bacteria (Hurley et al., 2006; Izutsu et al., 2008; Kodama et al., 2008; Makino et al., 2003; Pin˜eyro et al., 2010). For example, in a study by Park et al. (2004), deletion of both copies of the known haemolysin tdh in the clinical strain V. parahaemolyticus RIMD2210633 abolished haemolytic activity but did not affect cytotoxicity to HeLa cells. Furthermore, the mutant strain still showed partial fluid accumulation in rabbit ileal loops, suggesting that additional unknown virulence factors may be involved in the pathogenicity of the strain (Park et al., 2004). Several in vivo or in vitro models have been used to assess the distinct effects of pathogenicity such as cytotoxicity and enterotoxicity (Ljungh & Wadstrom, 1982; Miyamoto et al., 1980; Takeda, 1982; Nishibuchi et al., 1992; Park et al., 2004). However, comparison of the cumulative virulence of V. parahaemolyticus strains has been difficult to assess. Our study demonstrates that V. parahaemolyticus can infect zebrafish in a dose-dependent manner. The bacteria were lethal to zebrafish inoculated intraperitoneally with an LD50 of 5.76105 for the tdh+/trh2 clinical strain, RIMD2210633. The model was effective in evaluating differences in the pathogenic potential of V. parahaemolyticus strains irrespective of the source of isolation (environmental or clinical). We observed differences in isolates based on survival time regardless of the presence or absence of the known haemolysins, TDH or TRH, suggesting that this model could also be useful to evaluate the potential presence and/or contribution of additional (as yet unknown) virulence factors in strains encoding presently known virulence factors, e.g. TDH, TRH, T3SS2 and VPA1. Differences were also noted in isolates that grouped together using the typing methods, MLST and REP-PCR, suggesting that the model might provide a more refined screen for differentiating V. parahaemolyticus isolates of varying virulence. Early systematic changes in erythrocytes were observed in all isolates regardless of the presence or absence of the haemolysins TDH and TRH, suggesting the possible contribution of additional as yet unknown factors that affect erythrocytes as all isolates encode both T3SS1 and T3SS2. More pronounced changes were observed in the vasculature of the spleen and kidney in fish injected with tdh+/trh+ isolates, a finding that probably indicates the cumulative contribution of these haemolysins in virulence. The reproducible differences noted in the survival of the wild-type strain compared with the isogenic pilin mutants DpilA and DmshA and the DgbpA strain suggest that the model is useful to compare the virulence of wild-type strains and isogenic mutants. Our study suggests that, similar to V. vulnificus pilA, V. parahaemolyticus pilA contributes to virulence in this model (Paranjpye & Strom, 2005). In addition, similar to the V. cholerae MSHA pilin, the V. parahaemolyticus MSHA pilin does not appear to contribute to virulence, and is likely to have a role in 2613


environmental persistence of the bacterium (Chiavelli et al., 2001; Tacket et al., 1998; Thelin & Taylor, 1996; Watnick et al., 1999). In a recently published study (Frischkorn et al. 2013), the ability of DgbpA to form biofilms was not significantly different from that of the wild-type strain. In this study, the virulence of the V. parahaemolyticus gbpA mutant was not significantly different from that of the wild-type strain. This suggests that the role of this putative chitin-binding protein of V. parahaemolyticus in colonization of chitinous surfaces in the environment and in potential attachment to GlcNAc residues may be different from what has been observed for the homologous protein in V. cholerae (Kirn et al., 2005). Alterations in the erythrocytes were not noted in fish challenged with the DpilA strain as compared with those challenged with the wild-type or DmshA strain, suggesting that pilA could potentially be involved in adherence to erythrocytes. In an earlier study, the homologous pilA mutant in V. vulnificus was defective in adherence to epithelial cells (Paranjpye & Strom, 2005). On the other hand, histological changes in tissue sections of fish challenged with the mshA mutant strain were not significantly different from those of the wild-type, suggesting that the MSHA pilin is probably not involved in adherence in vivo. Virulence of the DmshA strain and the wild-type strain was also not significantly different, confirming that similar to the homologous mshA pilin in V. cholerae, V. parahaemolyticus mshA probably does not have a role in virulence in our model (Tacket et al., 1998; Thelin & Taylor, 1996). Notable histopathological differences were, however, observed in tissue sections of fish challenged with the gbpA mutant. The precise role of this putative surface-expressed factor in pathogenesis will need further study. In summary, the zebrafish infection model for V. parahaemolyticus provides a low-cost, reproducible, easy-to-use method to screen strains of this bacterium with varying virulence potential regardless of the source of isolation. The model also provides a means to evaluate the potential contribution of other, as yet unknown virulence factors. In our study, the model was effective in differentiating isogenic mutants from wild-type strains based on LD50, and therefore could be useful to evaluate the virulence potential of individual genes.

REFERENCES Brenot, A., King, K. Y., Janowiak, B., Griffith, O. & Caparon, M. G. (2004). Contribution of glutathione peroxidase to the virulence of

Streptococcus pyogenes. Infect Immun 72, 408–413. Centers for Disease Control and Prevention (CDC) (2005). Vibrio

illnesses after Hurricane Katrina – multiple states, August-September 2005. MMWR Morb Mortal Wkly Rep 54, 928–931. Chiavelli, D. A., Marsh, J. W. & Taylor, R. K. (2001). The mannose-

sensitive hemagglutinin of Vibrio cholerae promotes adherence to zooplankton. Appl Environ Microbiol 67, 3220–3225. Davis, J. M., Clay, H., Lewis, J. L., Ghori, N., Herbomel, P. & Ramakrishnan, L. (2002). Real-time visualization of mycobacterium-

macrophage interactions leading to initiation of granuloma formation in zebrafish embryos. Immunity 17, 693–702. Fournie, J., Krol, R. & Hawkins, W. (2000). Fixation of fish tissues. In

The Laboratory Fish, The Handbook of Experimental Animals, pp. 569– 578. Edited by G. Ostrander. New York: Academic Press. Frischkorn, K. R., Stojanovski, A. & Paranjpye, R. (2013). Vibrio

parahaemolyticus type IV pili mediate interactions with diatomderived chitin and point to an unexplored mechanism of environmental persistence. Environ Microbiol 15, 1416–1427. Fullner, K. J. & Mekalanos, J. J. (1999). Genetic characterization of a

new type IV-A pilus gene cluster found in both classical and El Tor biotypes of Vibrio cholerae. Infect Immun 67, 1393–1404. Goldberg, J. B., Hancock, R. E., Parales, R. E., Loper, J. & Cornelis, P. (2008). Pseudomonas 2007. J Bacteriol 190, 2649–2662. Hiyoshi, H., Kodama, T., Iida, T. & Honda, T. (2010). Contribution of

Vibrio parahaemolyticus virulence factors to cytotoxicity, enterotoxicity, and lethality in mice. Infect Immun 78, 1772–1780. Honda, T. & Iida, T. (1993). The pathogenicity of Vibrio parahaemo-

lyticus and the role of the thermostable direct haemolysin and related haemolysins. Rev Med Microbiol 4, 106–113. Hurley, C. C., Quirke, A., Reen, F. J. & Boyd, E. F. (2006). Four

genomic islands that mark post-1995 pandemic Vibrio parahaemolyticus isolates. BMC Genomics 7, 104–123. Izutsu, K., Kurokawa, K., Tashiro, K., Kuhara, S., Hayashi, T., Honda, T. & Iida, T. (2008). Comparative genomic analysis using microarray

demonstrates a strong correlation between the presence of the 80-kilobase pathogenicity island and pathogenicity in Kanagawa phenomenonpositive Vibrio parahaemolyticus strains. Infect Immun 76, 1016– 1023. Johnson, D. E., Weinberg, L., Ciarkowski, J., West, P. & Colwell, R. R. (1984). Wound infection caused by Kanagawa-negative Vibrio

parahaemolyticus. J Clin Microbiol 20, 811–812. Kirn, T. J., Jude, B. A. & Taylor, R. K. (2005). A colonization factor

links Vibrio cholerae environmental survival and human infection. Nature 438, 863–866. Kodama, T., Hiyoshi, H., Gotoh, K., Akeda, Y., Matsuda, S., Park, K. S., Cantarelli, V. V., Iida, T. & Honda, T. (2008). Identification of

ACKNOWLEDGEMENTS We thank Martin Liermann for help with statistical analysis, Kathi Lefebvre for guidance in zebrafish inoculations and Tiffany Linbo for raising and care of the zebrafish. We also thank Steve Moseley for critical review of the manuscript. This work was supported by the NOAA West Coast Center for Oceans and Human Health (WCCOHH) as part of the NOAA Oceans and Human Health Initiative. The WCCOHH is part of the National Marine Fisheries Service’s Northwest Fisheries Science Center, Seattle, WA, USA. 2614

two translocon proteins of Vibrio parahaemolyticus type III secretion system 2. Infect Immun 76, 4282–4289. Lefebvre, K. A., Tilton, S. C., Bammler, T. K., Beyer, R. P., Srinouanprachan, S., Stapleton, P. L., Farin, F. M. & Gallagher, E. P. (2009). Gene expression profiles in zebrafish brain after acute

exposure to domoic acid at symptomatic and asymptomatic doses. Toxicol Sci 107, 65–77. Link, A. J., Phillips, D. & Church, G. M. (1997). Methods for generat-

ing precise deletions and insertions in the genome of wild-type Escherichia coli: application to open reading frame characterization. J Bacteriol 179, 6228–6237. Microbiology 159

Zebrafish model for Vibrio parahaemolyticus virulence Ljungh, A. & Wadstrom, T. (1982). Toxins of Vibrio parahaemolyticus

and Aeromonas hydrophila. Toxin Reviews 1, 257–307. Makino, K., Oshima, K., Kurokawa, K., Yokoyama, K., Uda, T., Tagomori, K., Iijima, Y., Najima, M., Nakano, M. & other authors (2003). Genome sequence of Vibrio parahaemolyticus: a pathogenic

mechanism distinct from that of V cholerae. Lancet 361, 743–749. Mead, P. S., Slutsker, L., Dietz, V., McCaig, L. F., Bresee, J. S., Shapiro, C., Griffin, P. M. & Tauxe, R. V. (1999). Food-related illness

Pin˜eyro, P., Zhou, X., Orfe, L. H., Friel, P. J., Lahmers, K. & Call, D. R. (2010). Development of two animal models to study the function of

Vibrio parahaemolyticus type III secretion systems. Infect Immun 78, 4551–4559. Pressley, M. E., Phelan, P. E., III, Witten, P. E., Mellon, M. T. & Kim, C. H. (2005). Pathogenesis and inflammatory response to Edwardsiella

tarda infection in the zebrafish. Dev Comp Immunol 29, 501–513.

and death in the United States. Emerg Infect Dis 5, 607–625.

Prouty, M. G., Correa, N. E., Barker, L. P., Jagadeeswaran, P. & Klose, K. E. (2003). Zebrafish-Mycobacterium marinum model for myco-

Menudier, A., Rougier, F. P. & Bosgiraud, C. (1996). Comparative

bacterial pathogenesis. FEMS Microbiol Lett 225, 177–182.

virulence between different strains of Listeria in zebrafish (Brachydanio rerio) and mice. Pathol Biol (Paris) 44, 783–789.

Rawls, J. F., Mahowald, M. A., Goodman, A. L., Trent, C. M. & Gordon, J. I. (2007). In vivo imaging and genetic analysis link bacterial motility

Miyamoto, Y., Obara, Y., Nikkawa, T., Yamai, S., Kato, T., Yamada, Y. & Ohashi, M. (1980). Simplified purification and biophysicochemical

and symbiosis in the zebrafish gut. Proc Natl Acad Sci U S A 104, 7622–7627.

characteristics of Kanagawa phenomenon-associated hemolysin of Vibrio parahaemolyticus. Infect Immun 28, 567–576.

Reed, L. J. & Muench, H. (1938). A simple method of estimating fifty

Moyer, T. R. & Hunnicutt, D. W. (2007). Susceptibility of zebra fish

Ritchie, J. M., Rui, H., Zhou, X., Iida, T., Kodoma, T., Ito, S., Davis, B. M., Bronson, R. T. & Waldor, M. K. (2012). Inflammation and

Danio rerio to infection by Flavobacterium columnare and F. johnsoniae. Dis Aquat Organ 76, 39–44.

percent endpoints. Am J Hyg 27, 493–497.

Neely, M. N., Pfeifer, J. D. & Caparon, M. (2002). Streptococcus-zebrafish

disintegration of intestinal villi in an experimental model for Vibrio parahaemolyticus-induced diarrhea. PLoS Pathog 8, e1002593.

model of bacterial pathogenesis. Infect Immun 70, 3904–3914.

Rojo, I., de Ila´rduya, O. M., Estonba, A. & Pardo, M. A. (2007). Innate

Newton, A., Kendall, M., Vugia, D. J., Henao, O. L. & Mahon, B. E. (2012). Increasing rates of vibriosis in the United States, 1996-

immune gene expression in individual zebrafish after Listonella anguillarum inoculation. Fish Shellfish Immunol 23, 1285–1293.

2010: review of surveillance data from 2 systems. Clin Infect Dis 54 (Suppl. 5), S391–S395.

Stemple, D. L. & Driever, W. (1996). Zebrafish: tools for investigating

Nishibuchi, M., Fasano, A., Russell, R. G. & Kaper, J. B. (1992).

Sullivan, C. & Kim, C. H. (2008). Zebrafish as a model for infectious

Enterotoxigenicity of Vibrio parahaemolyticus with and without genes encoding thermostable direct hemolysin. Infect Immun 60, 3539– 3545.

Tacket, C. O., Taylor, R. K., Losonsky, G., Lim, Y., Nataro, J. P., Kaper, J. B. & Levine, M. M. (1998). Investigation of the roles of toxin-coregulated

O’Toole, G. A., Pratt, L. A., Watnick, P. I., Newman, D. K., Weaver, V. B. & Kolter, R. (1999). Genetic approaches to study of biofilms. Methods

pili and mannose-sensitive hemagglutinin pili in the pathogenesis of Vibrio cholerae O139 infection. Infect Immun 66, 692–695.

Enzymol 310, 91–109.

Takeda, Y. (1982). Thermostable direct hemolysin of Vibrio

O’Toole, R., Von Hofsten, J., Rosqvist, R., Olsson, P. E. & Wolf-Watz, H. (2004). Visualisation of zebrafish infection by GFP-labelled Vibrio

Thelin, K. H. & Taylor, R. K. (1996). Toxin-coregulated pilus, but not

anguillarum. Microb Pathog 37, 41–46. Okada, N., Iida, T., Park, K.-S., Goto, N., Yasunaga, T., Hiyoshi, H., Matsuda, S., Kodama, T. & Honda, T. (2009). Identification and

characterization of a novel type III secretion system in trh-positive Vibrio parahaemolyticus strain TH3996 reveal genetic lineage and diversity of pathogenic machinery beyond the species level. Infect Immun 77, 904–913. Paranjpye, R. N. & Strom, M. S. (2005). A Vibrio vulnificus type IV

pilin contributes to biofilm formation, adherence to epithelial cells, and virulence. Infect Immun 73, 1411–1422. Paranjpye, R., Hamel, O. S., Stojanovski, A. & Liermann, M. (2012).

cellular differentiation. Curr Opin Cell Biol 8, 858–864. disease and immune function. Fish Shellfish Immunol 25, 341–350.

parahaemolyticus. Pharmacol Ther 19, 123–146. mannose-sensitive hemagglutinin, is required for colonization by Vibrio cholerae O1 El Tor biotype and O139 strains. Infect Immun 64, 2853–2856. Turner, J. W., Paranjpye, R. N., Landis, E. D., Biryukov, S. V., Gonza´lez-Escalona, N., Nilsson, W. B. & Strom, M. S. (2013).

Population structure of clinical and environmental Vibrio parahaemolyticus from the Pacific Northwest Coast of the United States. PLoS ONE 8, e55726. Venables, W. N. & Ripley, B. D. (editors) (2002). Modern Applied

Statistics with S. (Statistics and Computing), 4th edn. New York: Springer.

Genetic diversity of clinical and environmental Vibrio parahaemolyticus strains from the Pacific Northwest. Appl Environ Microbiol 78, 8631–8638.

Watnick, P. I., Fullner, K. J. & Kolter, R. (1999). A role for the

Park, K.-S., Ono, T., Rokuda, M., Jang, M.-H., Iida, T. & Honda, T. (2004). Cytotoxicity and enterotoxicity of the thermostable direct

Weis, K. E., Hammond, R. M., Hutchinson, R. & Blackmore, C. G. M. (2011). Vibrio illness in Florida, 1998-2007. Epidemiol Infect 139,

hemolysin-deletion mutants of Vibrio parahaemolyticus. Microbiol Immunol 48, 313–318.

mannose-sensitive hemagglutinin in biofilm formation by Vibrio cholerae El Tor. J Bacteriol 181, 3606–3609.

591–598. Zon, L. I. (1999). Zebrafish: a new model for human disease. Genome

Phelps, H. A., Runft, D. L. & Neely, M. N. (2009). Adult zebrafish

Res 9, 99–100.

model of streptococcal infection. Curr Protoc Microbiol, Chapter 9, Unit 9D.1.

Edited by: H. Hilbi

http://mic.sgmjournals.org

2615