Virulence, Resistance, and Genomic Fingerprint Traits of Vibrio cholerae Isolated from 12 Species of Aquatic Products in Shanghai, China

Vibrio cholerae is a waterborne bacterium and can cause epidemic cholera disease worldwide. Continuous monitoring of V. cholerae contamination in aquatic products is imperative for assuring food safety. In this study, we determined virulence, antimicrobial susceptibility, heavy metal tolerance, and genomic fingerprints of 370 V. cholerae isolates recovered from 12 species of commonly consumed aquatic products collected from July to September of 2018 in Shanghai, China. Among the species, Leiocassis longirostris, Ictalurus punetaus, Ophiocephalus argus Cantor, and Pelteobagrus fulvidraco were for the first time detected for V. cholerae. Toxin genes ctxAB, tcpA, ace, and zot were absent from all the V. cholerae isolates. However, high occurrence of virulence-associated genes was detected, such as hapA (82.7%), hlyA (81.4%), rtxCABD (81.4%, 24.3%, 80.3%, and 80.8%, respectively), and tlh (80.5%). Approximately 62.2% of the 370 V. cholerae isolates exhibited resistance to streptomycin, followed by ampicillin (60.3%), rifampicin (53.8%), trimethoprim (38.4%), and sulfamethoxazole-trimethoprim (37.0%). Moreover, ∼57.6% of the isolates showed multidrug resistant phenotypes with 57 resistance profiles, which was significantly different among the 12 species (multiple antimicrobial resistance index, p < 0.001). Meanwhile, high incidence of tolerance to heavy metals Hg2+ (69.5%), Ni2+ (32.4%), and Cd2+ (30.8%) was observed among the isolates. The enterobacterial repetitive intergenic consensus-polymerase chain reaction (ERIC-PCR)-based fingerprinting profiles classified the 370 V. cholerae isolates into 239 different ERIC-genotypes, which demonstrated diverse genomic variation among the isolates. Overall, the results in this study meet the increasing need of food safety risk assessment of aquatic products.


Introduction
V ibrio cholerae is a Gram-negative bacterium that is autochthonously inhabited in aquatic environments worldwide. The bacterium can cause cholera, a severe diarrheal disease that is typically transmitted via contaminated water and person-to-person contact. 1,2 It is estimated that V. cholera caused roughly 2.9 million cases of cholera and 95,000 deaths annually worldwide between 2008 and 2012. 3 In recent years, cholera remained endemic in developing countries, and the most recent outbreak was reported in Mozambique in Aquatic animals such as fish, shellfish, and crustaceans are important reservoirs and vectors of V. cholerae. 4 For instance, V. cholerae was isolated from *30 fish species belonging to 9 different orders within the Actinopterygii class. 4 Recently, pathogenic V. cholerae (serotype: O1) was found in Rastrineobola argentea in Lake Victoria, Tanzania in 2017. 5 Thus, continuous monitoring of aquatic products and identification of risk factors of V. cholerae are imperative for assuring food safety.
Antimicrobial agents are widely applied in aquaculture industry for effective control of pathogenic bacteria; nevertheless, the indiscriminate use of antibiotics leads to the emergence of antibiotic-resistant bacteria, particularly in developing countries. 19 Resistant V. cholerae isolates that originated from clinical and environmental sources have been reported. 18,[20][21][22][23][24] For example, during an outbreak of cholera in Guinea-Bissau in 1996-1997, casefatality rates increased from 1% to 5.3% after the emergence of multidrug-resistant (MDR) V. cholerae. 20 Chomvarin et al. reported that V. cholerae O1 strains (n = 35) isolated from clinical samples and natural surface water sources in Thailand were resistant to trimethoprim (TM)/sulfamethoxazole and/or tetracycline (TET)/or ampicillin (AMP). 21 Recently, Verma et al. reported that 99% of V. cholerae strains (n = 438) from stool samples of diarrheal patients in India were resistant to more than two antibiotics. 24 Lepuschitz et al. investigated V. cholerae isolates (n = 82) recovered from an Austrian lake, and they found that the majority of the isolates showed resistance to sulfonamides (n = 80), and some to streptomycin (STR; n = 32) and AMP (n = 17). 23 Xu et al. reported that *65.3% of V. cholerae isolates (n = 400) from four species of commonly consumed fish in Shanghai, China were resistant to STR, followed by AMP (44.5%) and rifampicin (RIF; 24.0%). 18 Heavy metals used in mining, fossil fuel combustion, and agricultural practices can enter the food chain, such as cadmium (Cd), chromium (Cr), mercury (Hg), and Pb. 25 For instance, Li and Xie examined heavy metal contents in eleven species of fish collected from the Three Gorges Reservoir in China, and they found that the concentrations of Hg in carnivorous fish were higher than those in omnivorous fish. 26 Vu et al. analyzed heavy metal concentrations of five species of fish in the Houjing River in Taiwan, and they found that heavy metal concentrations of arsenic (As), Cd, Cr, copper (Cu), nickel (Ni), and Pb were higher than officially permissible limits. 27 Recently, Xu et al. also reported that *49.3% of the V. cholerae isolates (n = 400) were tolerant to Hg 2+ , followed by Zn 2+ (30.3%), and Pb 2+ (12.0%). 18 The People's Republic of China is the world's largest producer, consumer, and exporter of aquatic products. For instance, Scophthalmus maximus (known as turbot) is one of the most important economic marine flatfish species in Europe, which was introduced to China in the 1990s, and its production reached 640,000 tons in 2013 in China. 28 Among crustaceans, the largest output was yielded from Penaeus vannamei with 1,672,287 tons in 2017 in China. 29 In addition, Ostrea gigas thunberg is the world's widely cultured shellfish, and its production reached 4,879,422 tons in 2017 in China. 29 In our prior studies, we surveyed prevalence traits of V. cholerae strains isolated from four species of fish, three species of shrimps, and water samples from the Yangtze River Estuary in 2011-2017. 18,30,31 In this study, we determined virulence, resistance, and genetic diversity of 370 V. cholerae isolates recovered from 12 species of aquatic products, including nine species of fish, one species of crustacean, and two species of shellfish (see the Materials and Methods section). To our knowledge, four species of fish thereof were for the first time monitored for V. cholerae, including Ictalurus punetaus, Leiocassis longirostris, Ophiocephalus argus Cantor, and Pelteobagrus fulvidraco. This study meets the increasing need for decreasing and controlling pathogenic V. cholerae persistence in aquatic products.
Identification of V. cholerae isolates V. cholerae was identified by biochemistry and molecular biology methods as described in our previous study. 18 Briefly, properly diluted samples were spread on to selective thiosulfate citrate bile salts sucrose (TCBS, pH 8.5, 3.0% NaCl; Beijing Land Bridge Technology Co., Ltd., Beijing, China) agar plates, and they were incubated at 37°C for 16-18 hr. Yellow colonies were considered to be presumptive V. cholerae, and they were then inoculated into Double-Arginine Hydrolase Test Medium and Esculin Medium (Muwei Biotechnology Co., Ltd., Shanghai, China) for the arginine dihydrolase test and esculin hydrolysis test, respectively. V. cholerae isolates showing negative reactions in both two biochemical tests were further identified by polymerase chain reaction (PCR) assays. V. cholerae GIM 1.449 (Guangdong Culture Collection Center, Guangzhou, China) was used as a positive control strain.
V. cholerae-specific lolB and bacterial 16S ribosomal RNA (rRNA) genes were detected by PCR assays as previously described. 18,32 Genomic DNA was extracted by using a thermal lysis method. 18 A 20 mL of PCR reaction mixture contained 8 mL of DNase/RNase-free deionized water (Tiangen Biotech Co., Ltd., Beijing, China), 10 mL of 2 · Taq Master Mix (Novoprotein Technology Co., Ltd., Shanghai, China), 0.5 mL of each primer (5 mM) (Table 1), and 1 mL of genomic DNA. PCR reactions were performed by using Mastercycler Ò pro PCR thermal cycler (Eppendorf, Hamburg, Germany) under the conditions as previously described. 18 Amplicons were analyzed by agarose gel electrophoresis, visualized, and recorded by using a UVP EC3 Imaging system (UVP, LLC, Upland, CA) as previously described. 18 The PCR products were validated by DNA sequencing at Shanghai Sangon Biological Engineering Technology and Services Co., Ltd. (Shanghai, China). Sequence analysis was performed by using Basic Local Alignment Search Tool (BLAST) software against the GenBank database.

Detection of virulence and virulence-associated genes
The major virulence genes (ctxAB and tcpA) (Kumar et al. ) 33 and virulence-associated genes (ace, zot, rtxABCD, hapA, hlyA, tlh, mshA, and pilA). 6,11,18,34 were detected by PCR assay using the primer pairs listed in Table 1. The PCR reaction conditions varied depending on the melting temperature (Tm) values of the primer pairs. All the oligonucleotide primers used in this study were synthesized by Sangon. The genomic DNA of V. cholerae ATCC39315 (N16961) was used as a positive control.
Antibiotic susceptibility and heavy metal tolerance assays V. cholerae isolates were measured for in vitro susceptibility to 10 antimicrobial agents, and tolerance to 8 heavy metals according to the methods described in our previous studies. 18,35 The antimicrobial agents (Oxoid) included 10 mg AMP, 30 mg chloramphenicol (CHL), 10 mg STR, 10 mg gentamicin (CN), 30 mg kanamycin (KAN), 5 mg RIF, 100 mg spectinomycin (SPT), 30 mg TET, 5 mg TM, and 25 mg SXT (sulfamethoxazole [23.75 mg]-trimethoprim [1.25 mg]). The heavy metals included CdCl 2 , CrCl 3 , CuCl 2 , PbCl 2 , HgCl 2, NiCl 2 , MnCl 2 , and ZnCl 2 (analytical reagents; Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) with concentrations tested in the range from 3.125 to 3,200 mg/mL. Escherichia coli strains ATCC25922 and K12 (Institute of Industrial Microbiology, Shanghai, China) were used as quality control strains. 18,35 Enterobacterial repetitive intergenic consensus-PCR assay The enterobacterial repetitive intergenic consensus-PCR (ERIC-PCR) was performed according to a method previously described. 18 The ERIC-PCR reactions were performed under the following conditions: initial denaturation at 95°C for 8 min, followed by 32 cycles of 94°C for 30 sec, 52°C for zot-1129R AACCCCGTTTCACTTCTACCCA 1 min, and 65°C for 8 min, and a final extension at 65°C for 16 min. The ERIC-PCR products (6 mL/sample) were analyzed by electrophoresis at 120 V for about 30 min on a 1.0% agarose gel. Amplified DNA fragments were visualized and recorded as described earlier. DNA banding patterns generated by the ERIC-PCR were clustered by the unweighted pair group method with arithmetic mean (UPGMA) by using the Dice coefficient. 18 A dendrogram was constructed by using BioNumeric software v.7.6. 6 Simpson's index of diversity was calculated as previously described. 36 Statistical analysis Data analysis was performed by using SPSS software version 17.0 (SPSS, Inc., Chicago, IL). The multiple antimicrobial resistance index (MARI) of isolates was defined as previously described. 37 The MARI is often used to determine the antibiotic resistance-associated health risk. Statistically significant differences between aquatic products and MARI of resistant isolates were determined by one-way analysis of variance followed by appropriate post hoc text (Turkey). 18

Prevalence of V. cholerae isolates in 12 species of aquatic products
In this study, V. cholerae strains were isolated from 92 samples of aquatic products, including 9 species of fish, 1 species of crustacean, and 2 species of shellfish. Approximately 3,376 yellow single colonies were randomly selected from the selective TCBS agar plates for further identification. Approximately 30.5% (1,028/3,376) of the yellow colonies were detected negative for the arginine dihydrolase activity test and esculin hydrolysis test, and they were positive for the V. cholerae-specific gene lolB. The results were confirmed by DNA sequencing of the lolB and 16S rRNA genes. Approximately 95.6% (983/1,028) of the V. cholerae isolates were recovered from the fish, 3.2% (33/1,028) from the crustacean, and 1.2% (12/1,028) from the shellfish samples.
A pure culture of randomly selected 370 V. cholerae isolates was further analyzed and reported in this study, including 328 isolates from the fish (A. nobilis, C. auratus, C. idellus, O. argus Cantor, I. punetaus, L. longirostris, P. fulvidraco, P. pekinensis and S. maximus), 12 isolates from the shrimp (P. vannamei), and 30 isolates from the shellfish (O. gigas thunberg and P. magellanicus) samples.
Virulence and virulence-associated gene profiles in the V. cholerae isolates Virulence and virulence-associated gene profiles of the 370 V. cholerae isolates were obtained (Fig. 1). All the V. cholerae isolates were detected negative for the toxin genes ctxAB and tcpA. Moreover, none of the isolates carried the virulence-associated genes ace, zot, and pilA. In contrast, the hapA (82.7%), hlyA (81.4%), rtxBCD (80.3%, 81.4%, and 80.8%, respectively), and tlh (80.5%) genes were prevalent in the 370 V. cholerae isolates, *24.3% of which also carried the rtxA gene. A low percentage of the mshA gene (6.0%) was observed among the isolates.
As shown in Fig. 1, the V. cholerae isolates in the three types of aquatic products had different virulence genotypes. The mshA gene was only present in V. cholerae isolates that originated from the fish samples; whereas the hapA, hlyA, rtxCABD, and tlh genes were detected positive in the isolates from the fish, crustacean, and shellfish with different percentages. Among the V. cholerae isolates in fish, the hapA gene accounted for the highest proportion (83.5%), followed by rtxC (82.0%) and hlyA (82.0%). The occurrence of the hapA, hlyA, rtxBCD, and tlh genes was observed to be similar in the shrimp isolates (66.7%). All the V. cholerae isolates in shellfish were detected positive for the hapA (100%), hlyA (100%), rtxBCD (100%), and tlh (100%) genes, most of which also carried the rtxA gene (91.7%). In contrast, a lower incidence of the rtxA gene was observed in the fish (19.9%) and shrimp (43.3%) isolates.
Resistance to STR was the most prevalent (25.0-100%) among the V. cholerae isolates that originated from the 12 species of aquatic products, followed by to RIF (12.5-100%). Moreover, approximately 7.4-100% of the isolates from all the species showed resistance to AMP, except the P. magellanicus. In contrast, the isolates from all the species were sensitive to CHL, except the C. auratus. Also, all the isolates from P. pekinensis were sensitive to CHL, TET, SXT, and TM. However, the occurrence of resistant isolates to the later 2 drugs was higher in the other 11 species of aquatic products (SXT: 8.3-85.7%; TM: 8.3-92.9%) (Supplementary Table S1).
Different heavy metal tolerance profiles were also observed among the 12 species of aquatic products (Supplementary Table S2). The isolates from C. auratus, I. punetaus, O. argus Cantor, and S. maximus were tolerant to the maximum number of heavy metals (five of eight heavy metals), followed by C. idellus and L. longirostris (four of eight heavy metals), as well as A. nobilis, P. pekinensis, O.  gigas thunberg and P. magellanicus (three of eight heavy metals). Moreover, the highest percentage of tolerance to Ni 2+ was observed in the isolates from C. idellus (90.0%) and L. longirostris (90.0%), when compared with the isolates from the other species (0-36.3%). In addition, the unique resistance to Mn 2+ was detected in the isolates from C. idellus (3.3%), and to Pb 2+ it was observed from C. auratus (3.3%) and S. maximus (1.6%).

Genomic fingerprints of the V. cholerae isolates
Genetic diversity of the 370 V. cholerae isolates recovered from the 12 species of aquatic products was evaluated by the ERIC-PCR assay. The obtained fingerprinting profiles comprised various sizes of DNA bands, consistent with a previous report. 18 On the basis of the fingerprinting profiles, all the isolates were classified into 239 different ERICgenotypes, and 167 isolates (69.9%) thereof were assigned as singletons. Approximately 94.0% (n = 157) and 6.0% (n = 10) of these singletons were recovered from the fish and shellfish, respectively, whereas none was recovered from the crustacean samples. The UPGMA algorithm grouped all the 239 ERIC genotypes into 18 distinct clusters at a 73.0% similarity cutoff level (Figure not shown). Most isolates had a similarity coefficient of 50.0-100%, and the Simpson's diversity index was 0.8150. These results demonstrated high genetic diversity of the 370 V. cholerae isolates recovered from the 12 species of aquatic products.

Comparison of the MDR and heavy metal tolerance
Our data also revealed that among the 370 V. cholerae isolates, *57.6% (n = 213) exhibited MDR phenotypes with 57 resistance profiles. Given that high incidence of heavy metal tolerant isolates was also observed, therefore, we further investigated the relationship between the MDR and heavy metal tolerance of the isolates. The 213 MDR isolates were classified into 11 distinct clusters (I to XI) with 148 ERIC-genotypes (Fig. 5). Approximately half of the MDR isolates were grouped into Cluster II (51.6%, n = 110) with 56 ERIC-genotypes. Cluster IX was the second largest cluster (16.4%, n = 35); it consisted of 35 MDR isolates with 29 ERIC-genotypes, followed by the Cluster III, Cluster VI, and Cluster VII containing 7.5% (n = 16), 7.0% (n = 15), and 6.1% (n = 13) of the MDR isolates, respectively. The remaining isolates (11.3%, n = 24) were classified into Clusters I, IV, V, VIII, X, and XI with percentages in the range from 3.3% to 0.5%.
Among the 213 MDR isolates, *40.4% (n = 86) were tolerant to 1 heavy metal, followed by 32.9% (n = 70), 14.1% (n = 30), and 3.8% (n = 8) to 2, 3, and 4 heavy metals, respectively. Various resistance profiles were observed in different phylogenetic clusters. For instance, the Hg/AMP/ RIF/STR/SXT/TM resistance profile was predominant in the largest Cluster II In addition, the MDR isolates with identical ERICgenotypes had similar resistance profiles to antimicrobial agents, but different tolerance profiles to heavy metals. For instance, two isolates, V. c-L. longirostris 0805-18 and V. c-P. pekinensis 0807-08, recovered from L. longirostris and P. pekinensis, respectively, shared an identical ERICgenotype Vc00008, and antibiotic resistance profile (AMP/RIF/STR), but former isolate resistant to Hg 2+ and Ni + , and the latter only resistant to Hg 2+ . Overall, these data demonstrated considerable genetic diversity of the 213 MDR V. cholerae isolates with various heavy metal tolerance phenotypes.

Discussion
Outbreaks and prevalence of foodborne diseases not only are a major burden on global health care systems but also result in a huge negative impact on economic growth and social stability. 38 V. cholerae is ubiquitous in aquatic environments worldwide. Continuous monitoring of V. cholerae contamination in aquatic products is crucial for assuming food safety. In this study, 370 V. cholerae strains were isolated and identified from the 12 species of aquatic products. Among these species, I. punetaus, L. longirostris, O. argus Cantor, and P. fulvidraco were for the first time analyzed for V. cholerae. The I. punetaus (known as channel catfish) is native to North America and its production reached 382,306 tons in 2017 in China. 29 L. longirostris (known as longsnout catfish) is a rare and valuable freshwater economic fish in China, O. argus Cantor (known as snakehead fish) is also a popular freshwater fish and is regarded as a good nourishing tonic by Chinese people, 39 and P. fulvidraco (known as yellow catfish) is a small fish that lives in shoals. The total output of these 3 species reached 227,454, 483,141, and 480,032 tons in 2017 in China. 29 In this study, all the 370 V. cholerae isolates were absent for the toxic ctxAB and tcpA genes, consistent with our prior studies, 18,30,31 and some other literature. 22,[40][41][42][43] The ace, zot genes within the CTX element of pathogenic V. cholerae were also absent from all the isolates tested in this study. A similar observation was also reported based on non-O1/ O139 V. cholerae strains of environmental origins. 11,22,40,43 Recently, Olivares et al. reported that non-O1/O139 V. cholerae, isolated from an 81-year-old woman with diarrheal disease, carried the rtxA gene. 44 Within the rtxCABD operon, the rtxA gene encodes an RTX toxin, whereas the rtxBD and rtxC genes encode two secretion proteins and a toxin activator, respectively. 45 In this study, the intact rtxCABD operon was observed in about 24.3% of the 370 V. cholerae isolates, which was remarkably lower than those of rtxB + /rtxC + /rtxD + genes (80.3%, 81.4%, and 80.8%, respectively). The lower occurrence of the rtxA gene (24.3%) was also reported by Schirmeister et al. (27.8%), 11 but it was different from that (83.0%) by Xu et al. 18 It will be interesting to investigate the possible mechanism whereby the rtx gene cluster is truncated in V. cholerae in future research.
The HlyA is a pore-forming toxin that causes ion leakage and, ultimately, eukaryotic cell lysis. 46 Previous research has indicated that the HlyA and TLH prolonged V. cholerae colonization and pathogenesis to the epithelial cell. 47 In this study, the hlyA and tlh genes were present in *81.4% and 80.5% of the V. cholerae isolates, respectively, consistent with the previous report. 18 Ceccarelli et al. also reported that the hlyA gene was detected positive in 83% of V. cholerae isolates (n = 395) from oyster, sediment, and water samples collected from 2009 to 2012 in Maryland, United States. 14 In addition, the hapA gene encoding a Zn-dependent hemagglutinin protease facilitates V. cholerae detachment from the intestinal mucosa. 48 In this study, the hapA gene was present in *82.7% of the V. cholerae isolates, consistent with previous reports, such as 95.0% of the isolates (n = 400) from fish 18 and 98% of the isolates (n = 794) from clinical and environmental samples. 49 Meena et al. also reported that 46.5% of V. cholerae strains isolated from water samples in India carried the hapA gene. 50 The MSHA is very important for surface colonization of V. cholerae. 51 In this study, the mshA gene was detected positive in about 6.0% of the V. cholerae isolates, which was higher than that (0.8%) reported by Xu et al., 18 but lower than that (98.8%) by Rahman et al. 52 The PilA, belonging to the putative Type IV pilus of V. cholerae, is a virulence factor for Vibrio vulnificus. 53 In this study, the pilA gene was absent from all the 370 V. cholerae isolates tested, but this gene was present in 0.8% of the V. cholerae isolates in a previous report. 18 Taken together, in this study, the ctxAB, tcpA, ace, zot, and pilA genes were absent from all the 370 V. cholerae isolates recovered from the 12 species of aquatic products; however, the other virulenceassociated genes were prevalent (79.2-82.7%), such as rtxBCD, hlyA, hapA, and tlh.
It has been reported that China is the largest antibiotic producer and importer in the world. 54 The inappropriate use of antimicrobial agents has resulted in high detection frequencies in surface water and sediment samples of lakes and rivers in China, for example, the Taihu Lake, 55 the Yangtze River, 56 and the Yellow River Delta. 57 In this study, *60% of the V. cholerae isolates from the 12 species of aquatic products were resistant to STR and AMP. Xu et al. also reported that 44.5% and 65.3% of the V. cholerae isolates, recovered from four species of fish collected in 2017 in Shanghai, China, were resistant to AMP and STR, respectively. 18 Moreover, in this study, *57.6% of the isolates exhibited MDR phenotypes with 57 resistance profiles, which were significantly different among the 12 species (MARI, p < 0.001). Notably, 20.3% of the V. cholerae isolates recovered from the fish samples showed MARI values >0.4, indicating that their sources posed a high risk of antimicrobial contamination.
Heavy metals are cytotoxic to humans at low concentrations. 58 In this study, we observed that *69.5%, 32.4%, and 30.8% of the V. cholerae isolates recovered from the 12 species of aquatic products were tolerant to Hg 2+ , Ni 2+ , and Cd 2+ , respectively, when compared with the quality control strain. Tolerant V. cholerae isolates to the three heavy metals were more than those (Hg 2+ , 49.3%; Ni 2+ , 0.3%; and Cd 2+ , 4.8%) in the previous report. 18 Wang et al. investigated the circumstances of a rice-fish-farming system, and they found that Pb, Cd, Hg, As, and Cr constituted the most severe toxic heavy metal pollution in south China. 59 In this study, only two isolates (V. c-S. maximus 0706-87, V. c-C. auratus 0709-17) showed tolerance to Pb 2+ , but none to Cr 3+ . In addition, 41.6% of the 370 V. cholerae isolates were tolerant to 2 or more heavy metals, of which notably 9 V. cholerae isolates were tolerant to 4 heavy metals. Different heavy metal tolerance profiles were also observed among the V. cholerae isolates from the 12 species of aquatic products. These results suggested that various levels of heavy metal pollution may exist in aquatic farming environments.
The ERIC sequences are intergenic repetitive units in E. coli and other members of the Enterobacteriaceae, such as V. cholerae. 60 The ERIC-PCR has been applied in clonal diversity and genotyping of V. cholerae isolates from clinical and environmental origins. 6,18,61,62 For instance, Xu et al. reported 328 ERIC-genotypes among 400 V. cholerae isolates derived from 4 species of fish. 18 In this study, the 370 V. cholerae isolates recovered from the 12 species of aquatic products were classified into 239 ERIC-genotypes, 69.9% of which were assigned as singletons. These data, coupled with previous studies, demonstrated considerable genetic diversity of V. cholerae in aquatic products.
In conclusion, in this study, 370 V. cholerae isolates were isolated from the 12 species of commonly consumed aquatic products, which were collected from the two large aquatic product markets in Shanghai, China, from July to September of 2018. Among these species, V. cholerae was for the first time detected in the species I. punetaus, L. longirostris, O. argus Cantor, and P. fulvidraco. The CT genes ctxAB and tcpA, as well as toxin-associated genes ace, zot, and pilA were absent from all the 370 V. cholerae isolates, indicating no contamination of epidemic V. cholerae strains in the aquatic product samples tested. Nevertheless, high detection frequencies of the other virulence-associated genes were observed among the isolates, including the hapA (82.7%), hlyA (81.4%), rtxCABD (81.4%, 24.3%, 80.3%, and 80.8%, respectively), and tlh (80.5%).
Meanwhile, *62.2% of the 370 V. cholerae isolates were resistant to STR, followed by AMP (60.3%), RIF (53.8%), TM (38.4%), and SXT (37.0%). Approximately 57.6% of the isolates exhibited MDR phenotypes with 57 resistance profiles, which was significantly different among the 12 species (MARI, p < 0.001). High incidence of tolerance to heavy metals Hg 2+ (69.5%), Ni 2+ (32.4%), and Cd 2+ (30.8%) was observed among the isolates. Approximately 17.4% of the MDR isolates were tolerant to three or four heavy metals, implying a high risk of antibiotic and heavy metal contamination in the aquaculture environment for the aquatic animals tested. The obtained ERIC-PCR fingerprinting profiles classified all the 370 V. cholerae isolates into 239 different ERIC-genotypes, and 69.9% thereof were assigned as singletons, which demonstrated considerable genomic variation among the isolates.
In future research, it will be interesting to further investigate coevolution mechanisms of the antibiotic resistance and heavy metal tolerance of V. cholerae isolates in aquatic products. Overall, this study provided data in support of food safety risk assessment of aquatic products, and aquatic animal health management in aquaculture industry.

Disclosure Statement
No competing financial interests exist.

Funding Information
This study was supported by the grants from the Science and Technology Commission of Shanghai Municipality (grant No. 17050502200) and the National Natural Science Foundation of China (grant No. 31671946).