Prevalence and virulence of Vibrio parahaemolyticus isolated from clinical and environmental samples in Huzhou, China Authors: Peng Zhang, Xiaofang Wu, Lei Ji, Wei Yan, Liping Chen, Zhonghao Lu, Deshun Xu, Yunfeng Zha, Dafang Xu, Fenfen Dong Published: 2024 ABSTRACT Background Vibrio parahaemolyticus has emerged as the leading cause of seafood-associated infections worldwide. Previous studies have shown that V. parahaemolyticus can be detected in both environmental and clinical samples. However, the molecular characteristics of V. parahaemolyticus isolated from these sources remain unknown. ResultsThis study examined 128 strains of V. parahaemolyticus isolated from clinical and environmental samples collected between 2020 and 2023 in Huzhou, China. We identified 73 serotypes; O10:K4, O3:K6, and O4:KUT were the dominant serotypes among clinical isolates. We examined the proliferation and motility of major epidemic strains from environmental and clinical samples. Genetic diversity and evolution were assessed by average nucleotide identity (ANI), phylogenetic tree construction, and multilocus sequence typing (MLST). Furthermore, we identified 13 novel sequence types (STs) among the environmental isolates, indicating that V. parahaemolyticus strains are widely distributed and evolve rapidly in the environment in Huzhou, China. We found 206 virulence genes among these isolates, indicating that environmental isolates possess numerous virulence genes. Additionally, we detected 4 strains carrying the tdh or trh gene, which may increase their pathogenicity. The prediction results of antibiotic resistance genes shown that environmental isolates may carry up to 104 resistance genes, compared to 30 in clinical isolates.ConclusionsWe observed that the environmental serotypes of V. parahaemolyticus exhibit greater diversity compared to clinical isolates, which are predominantly concentrated in three major serotypes. Furthermore, a considerable genetic distance was found between most clinical and environmental isolates. Notably, some clinical isolates show a closer genetic proximity to environmental isolates. Additionally, the distribution of virulence genes, specifically T3SS and tdh, significantly differs among isolates from these two distinct sources. The prediction results for antibiotic resistance genes suggest that environmental isolates may harbor a broader spectrum of resistance genes. The findings of this study provide new insights into the phylogenetic relationships between V. parahaemolyticus strains from clinical and environmental sources, and they enhance the MLST database. Background The gram-negative bacterium Vibrio parahaemolyticus has attracted significant attention due to its association with seafood-related food poisoning outbreaks and is the leading cause of seafood-associated infections worldwide [1]. This bacterium naturally occurs in marine and estuarine environments, as well as seafood (e.g., fish and shellfish) [2]. Infection occurs through the consumption of raw or undercooked seafood [3]. V. parahaemolyticus causes three major clinical syndromes: gastroenteritis, wound infections, and septicemia [4]. The number of human infections associated with V. parahaemolyticus has reportedly expanded worldwide [5]. Recently, V. parahaemolyticus has been identified as a major foodborne pathogen in food poisoning incidents in China, raising public health concerns. V. parahaemolyticus is classified by serotyping determined by the combination of somatic (O) and capsular (K) antigens. Serotype O3:K6 was first detected in India [6]. Since then, several serovariants of the pandemic O3:K6 strains have been identified, such as O4:K68, O1:K25, O1:26, and O1:KUT, which are also recognized as pandemic clones [7]. Classically, the pathogenicity of V. parahaemolyticus is related to the production of an exotoxin, thermostable direct hemolysin (TDH) and/or TDH-related hemolysin (TRH) [8–10]. Several subtyping methods, such as multilocus sequence typing (MLST), have been developed to investigate the genetic diversity of outbreaks caused by V. parahaemolyticus and other foodborne pathogens [11]. As sequencing costs decrease, whole-genome sequencing is becoming increasingly common for bacterial genome analysis. To assess the homogeneity and heterogeneity between clinical and environmental V. parahaemolyticus isolates from Huzhou City, China, we used the conventional serotyping method to identify the main serotypes of V. parahaemolyticus in environmental and clinical samples. We also examined the proliferation and motility of isolates from different sources and serotypes. We further assessed the genetic diversity among V. parahaemolyticus strains by whole-genome sequence analysis. Additionally, we analyzed the V. parahaemolyticus isolates by average nucleotide identity (ANI) and phylogenetic tree construction to characterize the differences in the genome and evolution of each isolate. Additional information regarding multilocus sequence typing (MLST) and the distribution of virulence genes and antibiotic resistance genes among clinical and environmental V. parahaemolyticus strains would enrich the MLST database and epidemiological archive; it would also support the development of methods for efficient risk assessment concerning this common foodborne pathogen. Methods Sample collection and V. parahaemolyticus serotyping V. parahaemolyticus was isolated from clinical (stool) and environmental samples collected at food markets during a foodborne disease surveillance program from 2020 to 2023 in several regions of Huzhou, China. The isolation, identification, and serotyping of V. parahaemolyticus were performed as previously described [12]. We selected 128 V. parahaemolyticus isolates (69 from stool samples and 59 from environmental samples) for sequencing and further analysis. V. parahaemolyticus in the environment was mainly isolated from fish, shrimp, and water. The detailed information of these isolates was shown in Table 1. Table 1Information of each isolate in this studyisolateSourceSerotypePCR result trh tlh S2668Large yellow croakerO3:K6-+S3842Macrobrachium rosenbergiiO1:K25-+S3845Macrobrachium shrimp storage waterO5:K17-+S3846breamO3:KUT-+S3850ShrimpO1:KUT-+S3762shrimpO11:KUT-+S3234River shrimp waterO10:KUT-+S3213ShrimpO5:KUT-+S3216ShrimpO5:K20-+S3217SnailO5:K20-+S3763River bass storage waterO1:K9-+S3433Freshwater fish storage waterO1:KUT-+S3738ShrimpO4:K34-+S3740SnailO4:K32-+S3431SnailO2:K3-+S3610River shrimp storage waterO1:KUT-+S3609ShrimpO2:KUT-+S3843SticklebackO2:K3-+S3849ShrimpO10:K66-+S3761ShrimpO2:K9-+S3215ShrimpO1:K68-+S3230SnailO2:K59-+S2671Crucian carp storage waterO3:K6-+S3611SnailO2:KUT-+S3608ShrimpO1:KUT-+S3607SnailO8:K39-+S3434ShrimpO4:KUT++S3432River bassO5:KUT-+S3169CrucianO2:K33-+S3606Sediments at the bottom of fish pondsO1:K36-+S3323ShrimpO1:KUT-+S3322ShrimpO3:K29-+S3214Large yellow croakerO3:K33-+S3758BassO3:K1-+S3844HairtailO10:K66-+S3318CrucianO5:K30-+S3231CrucianO10:KUT-+S3232Crucian carp storage waterO2:KUT-+S3737CrucianO5:KUT-+S3847Freshwater fish storage waterO6:KUT-+S3320CrucianO3:K28-+S3319SticklebackO2:K28-+S3168BassO2:K28-+S3171Macrobrachium rosenbergiiO1:KUT-+S3170Macrobrachium rosenbergiiO10:KUT-+S3172Macrobrachium rosenbergiiO1:KUT-+S3848ShrimpO1:KUT-+S3212CrucianO1:K38-+S3233ShrimpO2:KUT-+S3235CrucianO3:KUT-+S3317BassO3:K28-+S3736Crucian carp sedimentO1:K46-+S3739ShrimpO4:KUT-+S3735CrucianO3:KUT-+S3760SnailO3:K37-+S3759Sediments at the bottom of fish pondsO3:K46-+S3840CrucianO3:KUT-+S3841Bighead fishO2:K3-+S3851ShrimpO3:K37-+S1424PatientO10:K4-+S1598PatientO10:K4-+S1474PatientO10:K4-+S2641PatientO3:K6-+S2543PatientO3:K6-+S2622PatientO3:K6-+S2661PatientO4:KUT-+S1504PatientO10:K4-+S1596PatientO3:K6-+S1475PatientO10:K4-+S0433PatientO10:K4-+S1517PatientO10:K4-+S0180PatientO4:KUT-+S0187PatientO4:KUT-+S0469PatientO10:K4-+S0134PatientO3:K6-+S0183PatientO4:KUT-+S0261PatientO3:K6-+S0179PatientO4:KUT-+S0181PatientO4:KUT-+S0184PatientO4:KUT-+S0185PatientO4:KUT-+S0189PatientO4:KUT-+S0341PatientO10:K4-+S0182PatientO4:KUT-+S2619PatientO10:K4-+S2592PatientO10:K4-+S0240PatientO3:K6-+S0338PatientO3:K6-+S2542PatientO10:K4-+S2491PatientO10:K4-+S2393PatientO4:KUT-+S2642PatientO4:KUT-+S2395PatientO4:KUT-+S2605PatientO10:K4-+S2590PatientO10:K4-+S2640PatientO3:K6-+S2398PatientO4:KUT-+S2394PatientO4:KUT-+S2722PatientO10:K4-+S2591PatientO10:K4-+S2588PatientO10:K4-+S1473PatientO10:K4-+S1472PatientO10:K4-+S1589PatientO3:K6-+S0131PatientO3:K6-+S0124PatientO4:KUT-+S0453PatientO10:K4-+S0466PatientO10:K4-+S2724PatientO10:K4-+S2589PatientO10:K4-+S1523PatientO10:K4-+S2535PatientO4:KUT-+S1484PatientO10:K4-+S2544PatientO4:KUT-+S2660PatientO10:K4-+S0468PatientO10:K4-+S0463PatientO10:K4-+S0364PatientO10:K4-+S0360PatientO10:K4-+S2623PatientO3:K6-+S2646PatientO4:KUT-+S2659PatientO10:K4-+S0470PatientO10:K4-+S2584PatientO10:K4-+S2585PatientO10:K4-+S0467PatientO10:K4-+S0391PatientO10:K4-+S2399PatientO4:KUT-+ Table 1Information of each isolate in this studyisolateSourceSerotypePCR result trh tlh S2668Large yellow croakerO3:K6-+S3842Macrobrachium rosenbergiiO1:K25-+S3845Macrobrachium shrimp storage waterO5:K17-+S3846breamO3:KUT-+S3850ShrimpO1:KUT-+S3762shrimpO11:KUT-+S3234River shrimp waterO10:KUT-+S3213ShrimpO5:KUT-+S3216ShrimpO5:K20-+S3217SnailO5:K20-+S3763River bass storage waterO1:K9-+S3433Freshwater fish storage waterO1:KUT-+S3738ShrimpO4:K34-+S3740SnailO4:K32-+S3431SnailO2:K3-+S3610River shrimp storage waterO1:KUT-+S3609ShrimpO2:KUT-+S3843SticklebackO2:K3-+S3849ShrimpO10:K66-+S3761ShrimpO2:K9-+S3215ShrimpO1:K68-+S3230SnailO2:K59-+S2671Crucian carp storage waterO3:K6-+S3611SnailO2:KUT-+S3608ShrimpO1:KUT-+S3607SnailO8:K39-+S3434ShrimpO4:KUT++S3432River bassO5:KUT-+S3169CrucianO2:K33-+S3606Sediments at the bottom of fish pondsO1:K36-+S3323ShrimpO1:KUT-+S3322ShrimpO3:K29-+S3214Large yellow croakerO3:K33-+S3758BassO3:K1-+S3844HairtailO10:K66-+S3318CrucianO5:K30-+S3231CrucianO10:KUT-+S3232Crucian carp storage waterO2:KUT-+S3737CrucianO5:KUT-+S3847Freshwater fish storage waterO6:KUT-+S3320CrucianO3:K28-+S3319SticklebackO2:K28-+S3168BassO2:K28-+S3171Macrobrachium rosenbergiiO1:KUT-+S3170Macrobrachium rosenbergiiO10:KUT-+S3172Macrobrachium rosenbergiiO1:KUT-+S3848ShrimpO1:KUT-+S3212CrucianO1:K38-+S3233ShrimpO2:KUT-+S3235CrucianO3:KUT-+S3317BassO3:K28-+S3736Crucian carp sedimentO1:K46-+S3739ShrimpO4:KUT-+S3735CrucianO3:KUT-+S3760SnailO3:K37-+S3759Sediments at the bottom of fish pondsO3:K46-+S3840CrucianO3:KUT-+S3841Bighead fishO2:K3-+S3851ShrimpO3:K37-+S1424PatientO10:K4-+S1598PatientO10:K4-+S1474PatientO10:K4-+S2641PatientO3:K6-+S2543PatientO3:K6-+S2622PatientO3:K6-+S2661PatientO4:KUT-+S1504PatientO10:K4-+S1596PatientO3:K6-+S1475PatientO10:K4-+S0433PatientO10:K4-+S1517PatientO10:K4-+S0180PatientO4:KUT-+S0187PatientO4:KUT-+S0469PatientO10:K4-+S0134PatientO3:K6-+S0183PatientO4:KUT-+S0261PatientO3:K6-+S0179PatientO4:KUT-+S0181PatientO4:KUT-+S0184PatientO4:KUT-+S0185PatientO4:KUT-+S0189PatientO4:KUT-+S0341PatientO10:K4-+S0182PatientO4:KUT-+S2619PatientO10:K4-+S2592PatientO10:K4-+S0240PatientO3:K6-+S0338PatientO3:K6-+S2542PatientO10:K4-+S2491PatientO10:K4-+S2393PatientO4:KUT-+S2642PatientO4:KUT-+S2395PatientO4:KUT-+S2605PatientO10:K4-+S2590PatientO10:K4-+S2640PatientO3:K6-+S2398PatientO4:KUT-+S2394PatientO4:KUT-+S2722PatientO10:K4-+S2591PatientO10:K4-+S2588PatientO10:K4-+S1473PatientO10:K4-+S1472PatientO10:K4-+S1589PatientO3:K6-+S0131PatientO3:K6-+S0124PatientO4:KUT-+S0453PatientO10:K4-+S0466PatientO10:K4-+S2724PatientO10:K4-+S2589PatientO10:K4-+S1523PatientO10:K4-+S2535PatientO4:KUT-+S1484PatientO10:K4-+S2544PatientO4:KUT-+S2660PatientO10:K4-+S0468PatientO10:K4-+S0463PatientO10:K4-+S0364PatientO10:K4-+S0360PatientO10:K4-+S2623PatientO3:K6-+S2646PatientO4:KUT-+S2659PatientO10:K4-+S0470PatientO10:K4-+S2584PatientO10:K4-+S2585PatientO10:K4-+S0467PatientO10:K4-+S0391PatientO10:K4-+S2399PatientO4:KUT-+ Information of each isolate in this study Motility test Motility testing was performed using a modified version of an existing method [13]. Briefly, a single colony was cultured to exponential phase in fresh Tryptic Soy Broth (TSB) supplemented with 3% NaCl. Two-microliter aliquots of this bacterial suspension were added to the center of swimming plates (LB broth, 3% NaCl, 0.2% agar) and swarming plates (LB broth, 3% NaCl, 1.5% agar). The plates were incubated at 36 °C, 30 °C, or 25 °C for 8 h or 48 h, respectively; the colony diameters were recorded. All experiments were performed in triplicate, and the average values were reported. Growth assays Single colonies were selected, inoculated into TSB supplemented with 3% NaCl, grown to exponential phase at 36 °C, and diluted with fresh TSB medium supplemented with 3% NaCl to an optical density at 600 nm (OD600) of 0.05. The suspensions were placed on a shaker at 36 °C for 8 h, and the OD600 was measured with a spectrophotometer every hour. Three replicates were performed for each strain. Statistical analysis was conducted by two-way analysis of variance (ANOVA). Whole-genome sequencing In total, 128 isolates were selected from both environmental and clinical sources for genome sequencing. DNA extraction and library construction were performed as previously described [12]. FASTP was used to filter the data for each sample and to remove adapters and low-quality data [14]. Bacterial genomes with quality score Q30 > 90% were used for further analysis. High-quality paired-end reads were assembled using SPAdes v3.13.2 to obtain the assembled genome sequence [15]. Seven housekeeping genes—dnaE, gyrB, recA, dtdS, pntA, pyrC, and tnaA—were used for characterization of V. parahaemolyticus. A V. parahaemolyticus MLST map of the 128 genomes was obtained by uploading whole-genome sequences to PubMLST (https://pubmlst.org/). Phylogenetic analysis A phylogenetic tree was constructed based on 128 single nucleotide polymorphisms (SNPs) among all of the isolates. Variant calling against the V. parahaemolyticus RIMD 2,210,633 (GCF_000196095.1) genome was performed using the Snippy pipeline (https://github.com/tseemann/snippy). Core SNP alignment was produced with snippy-core v.4.6.0. Recombinant regions were filtered out from the alignment using Gubbins v.3.0.0 [16]. SNPs were extracted from alignments using SNP-sites [17]. Core SNP alignment to infer a phylogenetic tree was performed using IQ-TREE with the maximum likelihood method using the TVMe + ASC + R5 nucleotide substitution model with 1000 bootstraps replicates [18–20]. The 53 clinical isolates were constructed using the same method as the 128 isolates; however, the nucleotide substitution model employed for analysis was K3P + ASC + R2. The tree was visualized with FigTree v.1.4.4, and different serotypes were highlighted in distinct colors. Virulence factor and antibiotic resistance genes prediction Virulence factors were predicted from the assembled genomes of each isolate using the Virulence Factor Database (VFDB) VFanalyzer tool (http://www.mgc.ac.cn/cgi-bin/VFs/v5/main.cgi?func=VFanalyzer). The annotation of antibiotic resistance genes was performed using the Comprehensive Antibiotic Resistance Database (CARD) with BLASTP (parameters: --evalue 1e-5 --query-cover 90 --subject-cover 90 --range-cover 90 --id 60). Analysis of trh gene by PCR A fresh colony ring was ground in 200 µL of sterilized water. The bacterial suspension was heated at 100 °C for 10 min and then centrifuged at 10,000 rpm for 5 min. The supernatant containing strain DNA was collected and stored at − 80 °C until use. Following the manufacturer’s instructions, the virulence genes of V. parahaemolyticus isolates were analyzed by fluorescent PCR using a commercial V. parahaemolyticus Triple Nucleic Acid Testing Kit (Biogen Biological Co., Ltd., Shenzhen, China). Average nucleotide identity The ANI was calculated using fastANI v. 1.34 [21]. All strain pairs were tested using the “Many to Many” method in fastANI and the “-matrix” option. The heat map was generated using pheatmap in R. Results Serotyping Seventy-three different serotypes were identified for the environmental and clinical isolates. The environmental isolates had 62 different serotypes, while the number of clinical isolates was 22 different serotypes. Among the environmental isolates, O1:KUT and O2:KUT were the two most common serotypes; O10:K4, O3:K6, and O4:KUT were the three most common serotypes among the clinical isolates (Fig. 1). However, no serotypes were widely present in either the environmental or clinical isolates (Fig. 1). These findings indicated that most V. parahaemolyticus strains isolated from the environment were not pathogenic and differed from clinical isolates of V. parahaemolyticus strains. Fig. 1Serotypes of 128 V. parahaemolyticus strains isolated from clinical and environmental samples Serotypes of 128 V. parahaemolyticus strains isolated from clinical and environmental samples Growth curve and motility Next, we examined differences in cell growth and motility among strains with different serotypes and from different sources. We selected four epidemic strains—S3850 (serotype O1:KUT), S3611 (serotype O2:KUT), S2640 (serotype O3:K6), and S2660 (serotype O10:K4)—from both environmental and clinical isolates for growth curve and motility testing. First, we monitored the OD600 of these four isolates over time with a spectrometer. The O3:K6 isolate showed faster cell growth during the exponential phase compared with the other three strains, but differences among the four isolates were not statistically significant (Fig. 2A). These results suggested that differences in serotype do not affect cell proliferation. V. parahaemolyticus produces both a polar flagellum for movement through liquid medium (swimming) and a lateral flagellar system for movement over solid surfaces (swarming) [22, 23]. Therefore, we next characterized motility by measuring the crawling diameter of the isolates. The O3:K6 and O10:K4 isolates exhibited stronger swarming ability than the other two strains at all three temperatures examined (25 °C, 30 °C, and 36 °C) (Fig. 2B). The O10:K4 isolate showed stronger swimming ability than the other three strains at 36 °C and 30 °C. The O2:KUT isolate showed weaker swimming and swarming abilities than the other three isolates at 25 °C. Therefore, we concluded that serotype was not associated with cell proliferation, but isolates with O10:K4 and O3:K6 serotypes had slightly stronger motility relative to environmental isolates. Fig. 2A Cell growth and (B) mobility of four different serotypes of V. parahaemolyticus strains isolated from clinical and environmental samples. (A) Plot of cell density (OD600) against time of O10:K4 clinical isolate, O3:K6 clinical isolate, O2:KUT environmental isolate, and O1:KUT environmental isolate. We performed three independent experiments (error bars represent SD). (B) Average diameters of swarming and swimming of 4 isolates. We performed three independent experiments. We analyzed data for significance using Student’s t-test A Cell growth and (B) mobility of four different serotypes of V. parahaemolyticus strains isolated from clinical and environmental samples. (A) Plot of cell density (OD600) against time of O10:K4 clinical isolate, O3:K6 clinical isolate, O2:KUT environmental isolate, and O1:KUT environmental isolate. We performed three independent experiments (error bars represent SD). (B) Average diameters of swarming and swimming of 4 isolates. We performed three independent experiments. We analyzed data for significance using Student’s t-test ANI analysis We next examined whether there were significant genomic differences between environmental and clinical isolates using the ANI, a well-documented and robust method for comparing genomes and assessing relationships among species [24]. In total, 128 environmental and clinical isolates were selected for genome sequencing and further analysis. Pairwise comparisons showed a minimum ANI of ~ 98.1% for all strains, indicating relatively small differences among them (Fig. 3). We noted that the differences among clinical isolates were significantly smaller than those among environmental isolates. Based on the serotype analysis of clinical isolates, the difference between O10:K4 and O3:K6 was significantly smaller than between O10:K4, O3:K6, and O4:KUT. Additionally, we found that the clinical isolates with the O4:KUT serotype could be divided into two clusters, and the difference between these two clusters was greater than the difference between the O10:K4 and O3:K6 isolates (Fig. 3). Fig. 3ANI percentage analysis of all isolates. ANI analysis was visualized using R ANI percentage analysis of all isolates. ANI analysis was visualized using R Phylogenetic analysis of V. parahaemolyticus isolates We explored the genetic diversity among the isolates by determining their phylogenetic relationships. For this purpose, we constructed a phylogenetic tree based on the core SNPs of 128 isolates (Fig. 4A). All environmental isolates were distributed in clade 1, and almost all clinical isolates were distributed in clade 2, with the exception of two groups of O4:KUT. We found that O4:KUT had five sequence types (STs): ST2516, ST3594, ST2980, ST332, and ST3. We discovered that certain strains lacked STs, potentially because mutations in their housekeeping genes resulted in sequencing results that do not align with any known alleles in current databases. The genetic distances between the O4:KUT isolates of ST2516 and ST332 and the environmental isolates were much shorter than the genetic distances between the O4:KUT isolates of ST2516 and ST332 and the clinical isolates (Fig. 4A). Given that most clinical isolates within clade 2 form a tight cluster, we performed a separate phylogenetic analysis on these 53 isolates. This analysis revealed that O10:K4 isolates are relatively concentrated, whereas O3:K6 isolates exhibit a more dispersed distribution (Fig. 4B). Although some O4:KUT strains were ST3, the genetic distances between these O4:KUT isolates were relatively large (Fig. 4B). This phylogenomic-based analysis also supported the ANI results. The STs of the environmental isolates were diverse and showed a wide distribution, indicating significant genomic differences among these isolates. Additionally, the clinical isolates with serotypes O10:K4 and O3:K6 had distant genetic relationships with the environmental isolates, suggesting that strains detected in the environment were not responsible for human illness (Fig. 4). Fig. 4Maximum-likelihood phylogenetic trees based on core SNP distances, and bootstrap values are shown as percentages of 1000 replicants. A A phylogenetic tree depicting the genetic relationships among 128 isolates. Strain IDs in red represent environmental isolates, and those in green represent clinical isolates. ST, serotype and isolation time of each isolate are shown on the right. Orange indicates newly identified STs. B The phylogenetic tree presents detailed information for the 53 isolates within clade 2, as shown in Fig. 4A Maximum-likelihood phylogenetic trees based on core SNP distances, and bootstrap values are shown as percentages of 1000 replicants. A A phylogenetic tree depicting the genetic relationships among 128 isolates. Strain IDs in red represent environmental isolates, and those in green represent clinical isolates. ST, serotype and isolation time of each isolate are shown on the right. Orange indicates newly identified STs. B The phylogenetic tree presents detailed information for the 53 isolates within clade 2, as shown in Fig. 4A Distribution of virulence genes in V. parahaemolyticus isolates Next, we investigated the differences in virulence genes between environmental and clinical isolates. We detected 206 virulence genes in these isolates, with each isolate carrying at least 140 virulence genes. The environmental isolates carried an average of 146 virulence genes, whereas the clinical isolates carried an average of 154 virulence genes (Fig. 5A). These virulence genes mainly participate in inheritance, antiphagocytosis, chemotherapy resistance, mobility, iron uptake, quorum sensing, and the secretory system. We investigated the reasons for the differences in pathogenicity between clinical and environmental isolates and found that 11 virulence genes—mshB, mshC, pilA, vopA, vopC, vopL, vopT, vcrD2, vscC2, vscN2, and tdh—were mainly distributed in the clinical isolates (Fig. 5B). These genes were primarily involved in adherence, chemotaxis and motility, the secretion system, and toxins. TRH is considered a major virulence factor as well [25]. We used PCR to assess the presence of the trh gene and found that only isolates S3434 was positive for it (Table 1). This finding suggests that the trh gene is not the cause of disease in the clinical isolates. T6SS genes are also known to contribute towards the pathogenicity of V. parahaemolyticus. We retrieved all T6SS1 and T6SS2 sequences from NCBI and utilized BLAST to analyze the distribution of these gene clusters among our isolate sequences. Our analysis revealed that nearly all isolates possess all the genes of the T6SS2 cluster. Furthermore, clinical isolates predominantly contain most of the T6SS1 genes. In contrast, some environmental isolates were found to possess most of the T6SS1 genes, while others lacked any T6SS1 genes (Fig. 5C). We concluded that these 11 genes may be responsible for the pathogenicity of clinical isolates and the distribution of the T6SS genes is experiencing changes. Fig. 5Heat map of virulence gene distribution in 128 V. parahaemolyticus strains. A All predicted virulence genes are shown at the bottom. The strain names are shown on the right. B Virulence genes with significant differences in distribution are shown on the left. The strain names are shown at the bottom. Red and blue indicate the presence and absence of genes, respectively. C Heat map depicting the distribution of T6SS genes across 128 V. parahaemolyticus isolates. Red and blue indicate the presence and absence of genes, respectively Heat map of virulence gene distribution in 128 V. parahaemolyticus strains. A All predicted virulence genes are shown at the bottom. The strain names are shown on the right. B Virulence genes with significant differences in distribution are shown on the left. The strain names are shown at the bottom. Red and blue indicate the presence and absence of genes, respectively. C Heat map depicting the distribution of T6SS genes across 128 V. parahaemolyticus isolates. Red and blue indicate the presence and absence of genes, respectively Detection of antibiotic resistance genes A total of 112 antibiotic resistance genes were annotated in 128 isolates using CARD. Almost all the isolates carried 30 identical resistance genes, including acrB, CARB-17, CARB-18, CARB-19, CARB-20, CARB-21, CARB-22, CARB-24, CARB-25, CARB-26, CARB-27, CARB-28, CARB-29, CARB-30, CARB-33, CARB-34, CARB-35, CARB-36, CARB-38, CARB-40, CARB-41, CARB-43, CARB-44, CARB-46, CARB-47, CARB-48, crp, msbA, txr, and ugd. Aside from these 30 antibiotic resistance genes, no other such genes were detected in the clinical isolates. In contrast, environmental isolates showed greater variation. The S3216 isolate stood out with the highest number of antibiotic resistance genes, totaling 104. The sul2 gene was primarily found in isolates S3213, S3216, S3233, S3317, and S3319. Specific resistance genes such as aph(3’)-Ib, aph(6)-Id, tet(A), tet(C), and tet(G) were exclusively distributed in S3317 and S3319. The qnrC gene was found only in S3432 and S3434, while the fosC2 gene was unique to S3843 (Fig. 6). Therefore, we concluded that almost all clinical and environmental isolates among these 128 strains shared the same 30 resistance genes, but there were significant differences among some of the environmental isolates. Fig. 6Heat map showing the distribution of antibiotic resistance genes in 128 V. parahaemolyticus strains. Red and blue indicate the presence and absence of genes, respectively Heat map showing the distribution of antibiotic resistance genes in 128 V. parahaemolyticus strains. Red and blue indicate the presence and absence of genes, respectively Discussion We analyzed the serotypes and virulence genes of 128 V. parahaemolyticus isolates from clinical and environmental samples in Huzhou, China, from 2020 to 2023. The serotype distribution of clinical isolates in Huzhou City from 2020 to 2023 was similar to that reported previously in Jiangsu Province [26]. O3:K6 was the dominant serotype responsible for gastroenteritis outbreaks in Jiangsu Province [26]. However, there were no dominant serotypes other than O3:K6 in Jiangsu Province from 2018 to 2021 [26]; the present study demonstrated the emergence of O10:K4 and O4:KUT as dominant serotypes among clinical isolates from 2020 to 2023 (Fig. 1), consistent with a previous report [27]. Our results were consistent with previous research showing that the frequency of O4:KUT strains has increased significantly since 2013 [28]. Moreover, O10:K4 was first discovered in 2020 [29] and replaced O3:K6 as the dominant serotype in 2021 [30]. Additionally, the dominant serotypes of the environmental isolates were consistent with previous studies, such that O1 and O2 were dominant [31]. V. parahaemolyticus frequently undergoes recombination, and recombination in the vicinity of gene clusters encoding O and K antigens contributes to serotype transformation [32]. Previous reports indicated that the growth rate of V. parahaemolyticus is closely related to temperature, pH, and NaCl concentration [33]. Serotype O3:K6 has shown more rapid growth than O1:KUT and O4:K8 in the presence of 1% and 3% NaCl at 15 °C, leading to pandemic outbreaks of O3:K6 infection [33]. In the present study, the bacterial growth rate was independent of serotype and isolation source at 30 °C, regardless of O3:K6 or O10:K4 serotype (Fig. 2A). Future research with numerous strains of O3:K6, O10:K4, and other serotypes is necessary to clarify whether this is a serotype-specific growth characteristic. Flagellar motility in Vibrio spp. is associated with several cellular processes, such as movement, colonization, adhesion, biofilm formation, and virulence [34, 35]. Isolates with serotypes O3:K6 and O10:K4 had stronger swimming and swarming capacities relative to the environmental isolates (Fig. 2B), which also explains why clinical isolates are more pathogenic than environmental isolates. In this study, we identified 49 STs among 128 V. parahaemolyticus isolates, 13 of which were novel (Fig. 4). The diversity of serotypes and STs among environmental isolates of V. parahaemolyticus is attributable to frequent recombination events in the bacteria rather than mutation [11]. The phylogenetic tree and MLST results showed that O3:K6, O10:K4, and some O4:KUT isolates belonged to ST3 (Fig. 4), consistent with a previous study [29]. Our findings suggest close genetic relationships among O10:K4, O4:KUT, and O3:K6. Furthermore, ST3 was the dominant ST among the 69 clinical strains, and all ST3 strains (n = 53) belonged to the pandemic O3:K6 serotype, consistent with previous reports from China and other countries [11, 36]. ST3 is the only sequence type with an international distribution [37]. ST3 is largely clinical and representative of pandemic clones in Asia and America [38–40]. A survey in China showed that approximately 74.3% and 94.0% of V. parahaemolyticus infections in patients from Beijing [41] and Zhejiang [42] respectively, were caused by ST3, whereas 85.4% of the O3:K6 serotype isolates from Shenzhen belonged to ST3 [43]. Taken together, these observations indicate that ST3 plays an important role in human infection. These findings showed that V. parahaemolyticus strains in the Huzhou area exhibited greater polymorphisms within environmental strains than among clinical strains. Notably, five O4:KUT isolates were distributed in different clades in the phylogenetic tree, and these O4:KUT isolates belonged to different STs: ST3, ST332, ST3594, ST2980, and ST2516 (Fig. 4). Reports indicate that the detection rate of the O4:KUT serotype exceeded that of O3:K6 and became the highest in 2013 in Zhejiang Province [44]. In subsequent years, the rate of O4:KUT remained high, including in Huzhou City [27]. Two main types of O4:KUT have been reported: O4:KUT-recAin and O4:KUT2 [28, 44]. O4:KUT-recAin has a large fragment insertion in the housekeeping gene recA at nucleotide position 497; it also has the ability to cause outbreaks [28]. The MLST analysis revealed that none of the O4:KUT isolates harbored large insertions within the recA gene, suggesting the absence of O4:KUT-recAin strains in the current study. O4:KUT2 was identified as ST332, which may be identical to our O4:KUT ST332. These observations also indicated that the K antigen constantly changes to produce multiple types of O4:KUT. It will be essential to monitor whether strain virulence changes with alterations in the K antigen. Among the isolates examined in the present study, we detected 206 virulence genes involved in inheritance, antiphagocytosis, chemotherapy resistance, mobility, iron uptake, quorum sensing, and the secretory system (Fig. 5A). Eleven genes showed significant differences in distribution between clinical and environmental isolates (Fig. 5B); these genes are involved in inheritance, chemotherapy resistance, motility, the secretory system, and toxins. Generally, V. parahaemolyticus strains capable of producing TDH and/or TRH belong to a pathogenic population of this species. However, this population constitutes a very small portion of the total (0–6%) [45–47], indicating that pathogenic strains are relatively rare. In this study, we detected tdh in all clinical isolates, but four strains carrying this gene were also present among the environmental isolates. We detected the trh gene only in S3434 using PCR, while whole genome sequencing failed to detect this gene. We validated the PCR results by conducting alternative PCR and sequencing the amplified products. The inconsistency between these two detection methods may be due to the higher detection sensitivity of qPCR, which allows it to identify genes present in lower copy numbers. Therefore, the pathogenicity of these four isolates requires further confirmation. In addition to hemolysins, the pathogenicity of V. parahaemolyticus strains is closely associated with the presence of molecular syringes, which allow the injection of effectors [48]. The type III secretion system (T3SS) can assist with the direct injection of virulent proteins or effectors into host cells [49]. V. parahaemolyticus has two different T3SS systems, designated as T3SS1 and T3SS2. T3SS1 is present in almost all clinical and environmental isolates [50], consistent with our findings; it can release VopQ, VopS, VPA0450, and VopR (VP1638) into host cells to promote cell death. T3SS2 is another secretion system (mainly found in clinically isolated strains [51]) that can release the effectors VopC, VopT, VopZ, VopA/P, VopV, VopL, and VPA1380 into host cells, thereby inhibiting host cell growth and promoting cell death. Most of the 11 genes identified in the present study are T3SS2 secretion effectors, consistent with previous findings [51]. T6SS is also a secretory device capable of secreting effectors into host mammalian cells as well as target bacterial cells. It was reported that V. parahaemolyticus isolates contain a T6SS on chromosome 2, named T6SS2, whereas pathogenic isolates encode another T6SS on chromosome 1, named T6SS1 [52, 53]. T6SS1 is encoded by a main gene cluster (vp1386-vp1420) [54]. The T6SS2 cluster contains 22 genes (locus tags vpa1025 to vpa1046). Utilizing VFanalyzer for the prediction of virulence genes did not show much information regarding the distribution of T6SS genes. Consequently, we employed the BLAST method to further analyze the distribution of these genes. We found that almost all isolates, including environmental and clinical isolates, carry all the genes of T6SS2 which was consistent with previous research [52, 53]. In contrast, the T6SS1 system is predominantly found in clinical isolates. A subset of environmental isolates has been found to harbor a significant number of T6SS1 genes, underscoring the need for further investigation into their potential virulence. Notably, we detected significant differences in the mannose-sensitive hemagglutinin type IV (MSHA) pilus genes mshA and mshC between the environmental and clinical isolates; other genes, such as mshD, mshE, and mshF, showed no differences according to source. MSHA pili are required for V. parahaemolyticus biofilm formation and persistence [55]; they play critical roles in intestinal colonization by the pathogen [56, 57]. Our results suggest that mshA and mshC play important roles in the pathogenicity of V. parahaemolyticus. Clinical and environmental isolates from the 128 strains almost all carried 30 identical resistance genes, including acrB, various CARBs, crp, msbA, txr, and ugd (Fig. 6). Compared to previous studies, this study found that the V. parahaemolyticus isolates carried several new resistance genes that were widely prevalent, such as acrB, crp, msbA, txr, and ugd. The protein encoded by acrB can transport a range of antibiotics, such as chloramphenicol, erythromycin, and tetracycline, playing a crucial role in bacterial multidrug resistance [58]. CARBs, or carbenicillin-hydrolyzing β-lactamases, are widely distributed among distant bacteria, primarily through mobile genetic elements [59–61]. In this study, we found that lots of CARBs were present in all isolates, a finding that aligns with previous research [62]. CRP regulatory factors are known to regulate multidrug resistance in Escherichia coli by inhibiting the expression of MdtEF multidrug efflux pumps [63]. The absence of CRP significantly increases the resistance of E. coli strains to ampicillin, azithromycin, erythromycin, and crystal violet [63]. MsbA is an essential inner membrane protein in E. coli and a member of the ABC transporter superfamily [64]. Studies have shown that tet35 and txr are essential for tetracycline resistance in heterologous E. coli hosts. E. coli carrying tet35 and txr can act as energy-dependent tetracycline efflux pumps, albeit with lower efficiency compared to tetA [65]. The gene ugd encodes UDP glucose dehydrogenase, which is responsible for synthesizing UDP glucuronic acid, a compound associated with intrinsic resistance to cationic antimicrobial peptides [66]. The high prevalence of these antibiotic resistance genes detected in V. parahaemolyticus warrants attention. In addition to the 30 common resistance genes, environmental isolates have been identified as carrying multiple other resistance genes. These include sul2, various forms of dfrA, NDM, aac(3), aadA, tet, fosC2 and qnrC genes. Sul2 and tetA are representative resistance genes for sulfonamides and tetracyclines, respectively, and are known for their widespread distribution in the environment [67]. Notably, sul2 has been found in highly artificially polluted environments, indicating its association with anthropogenic pollution [68], suggesting that sul2 in V. parahaemolyticus may be obtained from the environment. The dfrA genes detected in S3216, encodes a dihydrofolate reductase protein [69]. Resistance to trimethoprim is primarily mediated through the acquisition of mobile dfrA genes, which are predominantly found in the Enterobacteriales [70]. Multiple variants of NDM were also detected in the S3216 isolate. The NDM β-lactamase has become one of the most widespread carbapenemases globally due to its rapid evolution and transmission through multidrug-resistant (MDR) plasmids [71, 72]. The bla NDM gene is primarily found in the Enterobacteriaceae family [73]. The aac(3) enzyme is another type of resistance gene detected in the S3216 and S3217 isolates. The AAC(3)-II subclass confers resistance to gentamicin, netilmicin, tobramycin, sisomicin, 2’-N-ethylnetimycin, 6’-N-ethylnetimycin, and dibekacin and is widely present in Enterobacteriaceae and other Gram-negative clinical isolates [74, 75]. In addition, aac(3)-II alleles are typically found on mobile genetic elements, such as plasmids and transposons, which facilitate horizontal transfer between different pathogens [76–78]. The aadA gene variants are another type of resistance gene detected in the S3216 isolate. According to reports, the aadA gene encodes an adenyltransferase enzyme that confers resistance to streptomycin and spectinomycin and is present in Gram-negative bacteria [79, 80] and Staphylococcus aureus [81]. In addition, we detected the qnrC gene in the S3432 and S3434 isolates. qnrC, like qnrA, qnrB, and qnrS, is a plasmid-borne qnr gene that provides low-level quinolone resistance [82]. fosC2 is also a resistance gene located on plasmids, mediating resistance to fosfomycin, and is primarily found in the Enterobacteriaceae family [83].We speculate that the resistance genes identified in the environmental isolates may have been transmitted via plasmids. Conclusions In summary, we found no differences in proliferation between environmental and clinical isolates of V. parahaemolyticus; however, the clinical isolates showed significantly stronger motility than the environmental isolates. We demonstrated that the V. parahaemolyticus population in environmental samples is genetically diverse. Among the 59 isolates, we identified 46 different sequence types (STs), 13 of which were novel. We obtained information regarding genetic diversity among isolates from the MLST database. Moreover, we detected 4 environmental isolates carrying tdh or trh, suggesting that the new threat to public health posed by these environmental strains should not be overlooked. The findings of this study revealed three groups of O4:KUT that pose a threat to public health, and K antigen evolution should receive greater attention. The detection of virulence genes in the present study has enhanced the overall understanding of the pathogenesis of V. parahaemolyticus. Predictions of antibiotic resistance genes indicate that V. parahaemolyticus isolates from environmental sources may harbor a greater number of resistance genes. Furthermore, the evolution and dissemination of these resistance genes could potentially broaden the resistance spectrum of V. parahaemolyticus isolates originating from clinical sources.