Characterization of clinical Vibrio parahaemolyticus strains in Zhoushan, China, from 2013 to 2014

Vibrio parahaemolyticus is recognized as major cause of foodborne illness of global public health concern. This study collected 107 strains of V. parahaemolyticus during active surveillance of diarrheal diseases in hospitals in Zhoushan during 2013 to 2014 and investigated their serotypes, virulence genes (tdh, trh, and orf8), antimicrobial resistance, and genotypes. The dominant serotypes of the 107 clinical strains were O3:K6, O4:K8, and O4:KUT with 87.9% and 3.7% of the strains carrying the virulence genes tdh and trh, respectively. Molecular typing by pulsed-field gel electrophoresis indicated divergence among the clinical strains. Most isolates were sensitive to the common antimicrobial agents used against the Vibrio species except ampicillin. We conclude that continuous surveillance of V. parahaemolyticus in diarrhea patients is a public health priority and is useful for conducting risk assessment of foodborne illnesses caused by V. parahaemolyticus.


Introduction
Vibrio parahaemolyticus is a gram-negative halophilic human pathogen that naturally occurs in marine or estuarine environments, which is frequently isolated from a variety of seafood, such as shrimp, oyster, and fish [1][2][3]. Infections caused by V. parahaemolyticus can cause acute human gastroenteritis with major symptoms of headache, abdominal pain, and diarrhea, and in some cases, wound infection and septicemia [4][5][6]. This foodborne pathogen is considered the leading cause of seafood-derived illness in many countries around the world, including the United States, Thailand, Malaysia, Japan, Korea, and China [6][7][8][9]. In China, a total of 322 gastroenteritis outbreaks involving 9,041 illnesses and 3,948 hospitalizations due to V. parahaemolyticus infection were reported from 2003 to 2008 [10].
Zhoushan Fishery is the largest fishery in China. The output of aquatic products in Zhoushan is around 1.3 million tons annually, meanwhile, the offshore ocean fishing output account formore than one tenth of the total output in China [11]. Many citizens of Zhoushan in China regularly consume raw or undercooked seafood. Therefore, V. parahaemolyticus is a common regional pathogen, which causes more than 70% of the food-poisoning cases in Zhoushan [12]. Despite the high risk of V. parahaemolyticus, information regarding the prevalence or molecular epidemiology of this pathogen is not readily available.
The main objective of the present study was to investigate V. parahaemolyticus strains collected from stool specimens of outpatients with diarrheal disease in Zhoushan hospitals and to analyze their serotypes, virulence genes, genotypic traits, and antimicrobial resistance. The investigation will provide vital information to government agencies for performing risk assessments of V. parahaemolyticus infections.

Bacterial isolates
A total of 107 clinical V. parahaemolyticus isolates were used in the present study. They were collected from the stool specimens of 107 sporadic diarrhea outpatients from the active surveillance hospital: Zhoushan hospital. All isolates were collected from 2013 (41 isolates) to 2014 (66 isolates). The 107 isolates were stored at -80˚C in trypticase soy broth (TSB) containing 20% glycerol for further analysis.

Analysis of virulence genes and Kanagawa phenomenon
Vibrio parahaemolyticus can be detected by a species-specific marker (tlh) with the pathogenic strains carrying an additional tdh and/or trh gene. In addition, a distinctive eighth open reading frame of the filamentous phage f237, orf8, has been suggested as an useful genetic marker for identification of those pathogenic strains [13][14][15].
Determination of tlh tdh, and trh genes as well as the marker (orf8) for the V. parahaemolyticus isolates was performed using multiplex polymerase chain reaction (PCR) according to the methods of Ward and Bej [16] ( Table 1). The pathogenic group was defined as tdh+ and/or trh+; all other isolates were assigned to the non-pathogenic group [15,17].
Analysis of the Kanagawa phenomenon produced by the V. parahaemolyticus isolates was conducted using Wagatsuma blood agar (OXOID, Cambridge, UK) according to China National Food Safety Standard methods, GB 4789.7-2013. Each V. parahaemolyticus isolate was cultured on a plate of tryptic soy agar (OXOID, Cambridge, UK) containing 3% NaCl for Table 1. Oligonucleotide primers used for polymerase chain reactions [16].

Target genes Primers (Probes) Sequence (5'-3') Amplicon Size (bp)
tlh 18 h and then spotted on a Wagatsuma agar plate. The Wagatsuma agar plates were incubated for 18-24 h at 37˚C and examined for Kanagawa phenomenon within 24 h.

Serotyping
Serological analysis of lipopolysaccharide (O) and capsular (K) antigens of the V. parahaemolyticus isolates was identified by performing agglutination tests using a commercial V. parahaemolyticus antiserum test kit (Denka Seiken, Tokyo, Japan) according to the manufacturer's instructions.

Pulsed-field gel electrophoresis (PFGE)
The PFGE analysis of the V. parahaemolyticus isolates was performed according to the Pulse-Net protocol (http://www.pulsenetinternational.org). Briefly, using NotI enzyme to restrict the chromosomal DNA. The restriction fragments were resolved with 1% seakem gold agarose gel in 0.5% Tris-boric acid-EDTA buffer using a CHEF Mapper XA system (Bio-Rad Laboratories, Richmond, Calif., USA). The XbaI digested DNA from Salmonella enterica serotype Braenderup strain H9812 was used as a molecular size marker. The PFGE patterns were analyzed using BioNumerics software (Applied Maths, Belgium). Clustering was performed using the unweighted pair group method and the Dice correlation coefficient with a position tolerance of 1.5%. Clusters were defined on the basis of a 90% similarity cutoff [18].

Antimicrobial susceptibility testing
The antimicrobial susceptibility of the 107 clinical V. parahaemolyticus isolates was tested using the broth microdilution method.    (Table 2).

PFGE and cluster analysis
Analysis of the genotypes of the 107 clinical V. parahaemolyticus isolates found that 4 of the isolates were non-typable due to DNA degradation during endonuclease digestion. PFGE analysis of the remaining 103 isolates yielded 23 distinguishable patterns at a 90% similarity threshold (Fig 2). Among them, 35 isolates were clustered in pattern 9, which contained isolates belonging to two different serotypes, O4:K8 and O4:KUT. Furthermore, the isolates, which were O3:K6 serotype were not all included in the same group, such as VP012 and VP029 were clustered in pattern 14 and the other O3:K6 isolates were clustered in pattern 10, 11, 12 and 13.

Discussion
V. parahaemolyticus has emerged as an important public health concern worldwide as a pathogen that causes gastroenteritis related to the consumption of various seafood, including crab, shrimp, lobster, fish, and oysters [19]. In the present study, 107 clinical V. parahaemolyticus strains were collected from 2013 to 2014 in Zhoushan, we identified and analyzed the virulence genes, serotypes, molecular typing by PFGE, and antimicrobial resistance profiles of these clinical strains.
As shown by epidemiological investigation, thermostable direct hemolysin (TDH) is recognized as one of the major pathogenic factors in V. parahaemolyticus, and is present in 92% to 95% clinical isolates [19,20], and it contributes to the formation the "Kanagawa phenomenon" (KP) [19], which has been regarded as an important indicator for the identification of pathogenic and non-pathogenic strains [19]. In the present study, among 94 tdh+ clinical strains, 15 strains did not show KP, moreover, we found one clinical strain (1/107) that tested positive for KP, though it carried neither tdh nor trh. Therefore, the KP is not a reliable indicator for identifying pathogenic and non-pathogenic V. parahaemolyticus strains [19].
As shown in Table 2, 87.9% of the strains carried tdh, and only 4 strains carried trh, that is, 91.5% of the strains carried tdh and/or trh gene. This is similar to a previous study reporting 92.5% of the clinical isolates of V. parahaemolyticus collected in Shanghai as carrying the tdh and/or trh gene [20]. Furthermore, 11.2% of clinical strains from the diarrheal patients  contained neither the tdh nor the trh gene, which is similar to a previous study reporting 11.8% of clinical isolates of V. parahaemolyticus collected in Jiangsu province as carrying neither the tdh gene nor the trh gene [15]. It has been reported that there were outbreaks caused by strains lacking of tdh and/or trh [21]. Although other pathogenic factors have been recognized, such as type III secretion system 1 and 2 and biofilm formation, the understanding of pathogenicityis still incomplete [21]. Meanwhile, 28 strains contained orf8. The orf8 is suggested to be a marker for detecting recent serotype O3:K6 and other pandemic group strains [16], which were considered as a good target for detection of pathogenic pandemic V. parahaemolyticus O3:K6 serotype strains. As shown in the current study, the hemolysin genes (tdh and trh) and orf8 were tested, furthermore, more pandemic markers, such as toxRS/new and type III secretion system could be analyzed to reveal comprehensive pathogenic and pandemic characteristics of clinical V. parahaemolyticus strains in Zhoushan. Since V. parahaemolyticus was discovered 60 years ago [22], it has been considered as a major cause of foodborne illnesses all over the world with serotype O3:K6 being firstly isolated in 1996 in India and subsequently worldwide [15,23]. In this study, 31 (29%) of 107 clinical V. parahaemolyticus isolates collected over 2 years from 2013 to 2014 were identified as O3:K6, which is lower than the previous study reporting 54%, 62% and 66% of the clinical isolates of V. parahaemolyticus collected in Guangdong, Shanghai and Jiangsu respectively [15,20,24]. These results demonstrate that V. parahaemolyticus strains with different serotypes can share common molecular types, therefore, the serotyping was of limited use because there was no match between serotype and PFGE cluster. PFGE has been used as the "gold standard" method to assess genetic diversity and clonal relatedness among V. parahaemolyticus isolates from different sources [18,20,25].
Antibiotics can be used to treat V. parahaemolyticus infection, and they should be based on the antimicrobial susceptibilities of V. parahaemolyticus. In the present study, most isolates were sensitive to the antimicrobial agents recommended by CLSI for Vibrio species (not V. cholerae) except ampicillin (85.8%). Similar findings have been reported by Chen et al (87.1%) [17]. As shown in Table 3, 100% of the clinical strains in the present study were susceptible to Meropenem, Imipenem, Cefoxitin, Ceftazidime, Trimethoprim/sulfamethoxazole, Chloramphenicol, Amoxicillin/clavulanic acid, Ciprofloxacin, Levofloxacin and Tetracycline.
To the best of our knowledge, this is the first study to investigate virulence factors, serotypes, genotypes, and antimicrobial susceptibility of the clinical strains of V. parahaemolyticus collected in Zhoushan over 2 years through a continuous surveillance of V. parahaemolyticus isolated from patients during warm months. Continuous monitoring of V. parahaemolyticus strains in Zhoushan is necessary to generate information useful for the control of foodborne illnesses caused by V. parahaemolyticus.

Conclusions
This study analyzed characteristics of V. parahaemolyticus strains in diarrheal patients in Zhoushan. Different serotypes were detected in the clinical isolates. Although the clinical V. parahaemolyticus  strains isolated in Zhoushan are not resistant to most common antimicrobial agents, it will be necessary to keep a close attention to the emergence of antimicrobial resistant strains and strengthen the management of antimicrobials. Therefore, this information is useful for the control and treatment of foodborne illnesses caused by V. parahaemolyticus in Zhoushan.