First Report of Potentially Pathogenic Klebsiella pneumoniae from Serotype K2 in Mollusk Tegillarca granosa and Genetic Diversity of Klebsiella pneumoniae in 14 Species of Edible Aquatic Animals

Klebsiella pneumoniae can cause serious pneumonitis in humans. The bacterium is also the common causative agent of hospital-acquired multidrug-resistant (MDR) infections. Here we for the first time reported the genetic diversity of K. pneumoniae strains in 14 species of edible aquatic animals sampled in the summer of 2018 and 2019 in Shanghai, China. Virulence-related genes were present in the K. pneumoniae strains (n = 94), including the entB (98.9%), mrkD (85.1%), fimH (50.0%), and ybtA (14.9%) strains. Resistance to sulfamethoxazole-trimethoprim was the most prevalent (52.1%), followed by chloramphenicol (31.9%), and tetracycline (27.7%), among the strains, wherein 34.0% had MDR phenotypes. Meanwhile, most strains were tolerant to heavy metals Cu2+ (96.8%), Cr3+ (96.8%), Zn2+ (91.5%), Pb2+ (89.4%), and Hg2+ (81.9%). Remarkably, a higher abundance of the bacterium was found in bottom-dwelling aquatic animals, among which mollusk Tegillarca granosa contained K. pneumoniae 8-2-5-4 isolate from serotype K2 (ST-2026). Genome features of the potentially pathogenic isolate were characterized. The enterobacterial repetitive intergenic consensus polymerase chain reaction (ERIC-PCR)–based genome fingerprinting classified the 94 K. pneumoniae strains into 76 ERIC genotypes with 63 singletons, demonstrating considerable genetic diversity in the strains. The findings of this study fill the gap in the risk assessment of K. pneumoniae in edible aquatic animals.


Introduction
Klebsiella pneumoniae can cause life-threatening pneumonitis in humans [1,2]. The bacterium is also the common causative agent of nosocomial infection diseases, including septicemia, meningitis, osteomyelitis, thrombophlebitis, and urinary tract infections (UTIs) [3,4]. K. pneumoniae was initially isolated from the lungs of dead pneumonia patients in 1882 [5]. The bacterium is found growing in the environment, specifically in areas such as soil, water, and plant matter [6]. K. pneumoniae strains are classified into at least 79 serotypes, of which serotypes K1, K2, K5, K20, K54, and K57 are closely related to bacterial pathogenesis [7][8][9]. Studies have also indicated that K1 and K2 are high-risk clones with high virulence and multidrug resistance [10]. The rising incidence of high-virulent and multidrug-resistant (MDR) K. pneumoniae is one of the major clinical and public health issues worldwide [11].
The new finding of K. pneumoniae present in commonly edible aquatic animals poses tremendous danger to food safety and human health. Person-to-person or person-to-item interaction during food preparation may transmit K. pneumoniae [31]. Temperatures over 35 • C have been shown to reduce K. pneumoniae growth and metabolic activity. At 60 • C, there is a significant drop in growth; however, K. pneumoniae still exhibits some metabolic activity (i.e., not completely inactivated) [32]. Food handlers should thus be aware of their involvement in the spread (or control) of the pathogenic bacterium.
China is the world's largest producer, exporter, and consumer of aquatic products and has contributed approximately 60% (64,636,700 tons) to the global output in 2021 (National Bureau of Statistics, http://www.stats.gov.cn/ (accessed on 1 August 2022)). Continuous monitoring and identification of risk factors in the aquatic products are crucial for food safety and human health. Thus, the major objectives of this study were to survey the prevalence, virulence, and resistance of K. pneumoniae in commonly the consumed 41 species of aquatic animals collected in Shanghai, China. Of these, 39 species have not ever been detected having K. pneumoniae. Genetic diversity and coresistance between antibiotics and heavy metals of 94 K. pneumoniae strains recovered in the survey were also investigated. The findings of this study will fill gaps in the risk assessment of K. pneumoniae in edible aquatic animals.

Sample Collection
The 41 species of aquatic animals included mollusks (n = 21), fish (n = 17), and crustaceans (n = 3) (Table S1). Samples (n = 108) were collected in sterile plastic bags (Nanjing Maojie Microbial Technology Co., Ltd., Nanjing, China) from the Shanghai Jiangyang Aquatic Market and Shanghai Luchao Port Aquatic Market in July, August, and September of 2018 and 2019 in Shanghai, China [22]. The samples were stored in iceboxes (700 × 440 × 390 mm) and immediately transported to the laboratory at Shanghai Ocean University (Shanghai, China) for analysis within 24 h at 4 • C [22].

Isolation and Identification of K. pneumoniae
K. pneumoniae was isolated and identified in accordance with the National Standards of the People's Republic of China (SN/T 1962(SN/T -2007. The samples were pretreated as described in our recent research [33]. Briefly, an aliquot (25 g) of each sample was mixed with 225 mL of 1× phosphate-buffered saline (PBS, pH 7.4-7.6, Shanghai Sangon Biological Engineering Technology and Services Co., Ltd., Shanghai, China), homogenized, serial diluted, and spread onto the MacConkey Inositol Adonitol Carbenicillin (MIAC, Beijing Land Bridge Technology Co., Ltd., Beijing, China) agar plates [34]. The red colonies grown on the MIAC agar plates at 37 • C were randomly picked out for further identification. K. pneumoniae ATCC13883 was used as the control strain.
The matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF/MS) platform (Microflex LT/SH, MALDI Biotyper, Bruker, Germany) was used to identify presumptive K. pneumoniae colonies; the platform was run under the same conditions and parameters in our recent studies [34,35]. Mass spectra were collected and analyzed by using the Microflex LT Mass Spectrometer (Bruker 422 Daltonics, Bremen, Germany) [34,35].
Biochemical tests, including capsular staining, Gram's staining, and dynamic tests (Qingdao Haibo Bio Co., Ltd., Qingdao, China), were also used to identify K. pneumoniae strains, in accordance with the manufacturer's instructions and the Chinese Government Standard (SN/T 1962(SN/T -2007. K. pneumoniae strains were also confirmed using the polymerase chain reaction (PCR) assay to amplify the 16S rRNA gene with the universal bacterial primers (Table S2) and DNA sequencing analysis [33,34].

Identification of Virulence Genes
Virulence genes (aerobactin, alls, entB, fimH, iroN, magA, mrkD, rmpA, traT, wcaG, and ybtA) were detected using the PCR assay [15,[36][37][38][39][40][41]. All oligonucleotide primers (Table S2) used in this study were synthesized by the Sangon (Shanghai, China). The genomic DNA of strains was prepared using the TaKaRa MiniBEST Bacterial Genomic DNA Extraction Kit version 3.0 (TaKaRa Biomedical Technology Co., Ltd., Beijing, China). The 20 µL PCR reaction mixture contained 8 µL of DNase/RNase-free deionized water (Tiangen Biotech Co., Ltd., Beijing, China), 10 µL of 2× Taq Master Mix (Novoprotein Technology Co., Ltd., Shanghai, China), 0.5 µL of each primer (5 µM), and 1 µL of DNA template. PCR amplification was performed as described in our previous research [33,34,42], but with different annealing temperatures and elongation times, based on the melting temperatures of the primer pairs and the predicted sizes of the PCR products (Table S2). The obtained PCR products were analyzed as previously described [33,34,42]. DNA sequencing was carried out by Suzhou Jinweizhi Biotechnology Co., Ltd., Suzhou, China. The basic local alignment search tool (BLAST) of National Center for Biotechnology Information (NCBI) was used for sequence analysis (https://www.ncbi.nlm.nih.gov/ (accessed on 19 August 2021)). The virulence gene sequences of K. pneumoniae in the NCBI sequence databases were compared in the BLAST analysis. The standard strain K. pneumoniae ATCC13883 (Guangdong Culture Collection Center, Guangzhou, China) was used as a positive control strain in this study.

Enterobacterial Repetitive Intergenic Consensus-PCR (ERIC-PCR) Assay
The ERIC-PCR was used for genotyping of K. pneumoniae strains as described in previous studies (Table S2). The 20 µL reaction mixture contained 6 µL of DNase/RNasefree deionized water (Tiangen Biotech Co., Ltd., Beijing, China), 10 µL of 2 × Taq Master Mix (Novoprotein Technology, Shanghai, China), 1 µL of each primer (5 µM), and 2 µL of a DNA template. ERIC-PCR reactions were performed using the primers ERIC1R and ERIC2 (Table S2) at 95 • C for 30 s, 52 • C for 1 min, and at 65 • C for 8 min for 32 reaction cycles [34]. Amplified DNA fragments were observed, recorded, and analyzed; the unweighted pair group with arithmetic averages (UPGMA) method was used for cluster analysis; and the Simpson index was calculated to assess the diversity of K. pneumoniae strains [33,34,42,43].
Functional assignments were inferred using the standalone BLAST software [48]. The virulence factor database (http://www.mgc.ac.cn/VFs (accessed on 2 April 2022)), the comprehensive antibiotic resistance database (CARD, http://arpcard.Mcmaster.ca, Version 1.1.3) [49], and the BacMet database (http://bacmet.biomedicine.gu.se/ (accessed on 21 September 2022)) were utilized to identify virulence, antibiotic and heavy metal resistance-related genes, respectively. Mobile genetic elements (MGEs) were described and identified according to the methods described in our recent research [48]. The software was run with the default settings.

Statistical Analysis
The SPSS version 17.0 software (SPSS Inc., Chicago, IL, USA) was used for statistical analysis. The multiple antibiotic resistance index (MARI) values were calculated according to the formula described previously [50]. In this study, all tests were performed in triplicate.

K. pneumoniae in the 41 Species of Aquatic Animals
K. pneumoniae in the 41 species of aquatic animals was isolated, of which 39 species have not ever been detected for the bacterium, except Mytilus edulis and Ostrea gigas thunberg (Table S1). Approximately 567 red colonies were picked from the MIAC agar plates, and 16.6% (n = 94) of the colonies were identified as K. pneumoniae by the MALDI-TOF/MS analysis. The 94 K. pneumoniae strains also tested positive for the capsular staining and negative for the Gram's staining and dynamic testing. The sequencing of the bacterial 16S rRNA gene validated the results as well (Table S3).
The 94 K. pneumoniae strains were isolated from the aquatic animals originating from different regions in China, including Shanghai City and the Fujian, Zhejiang, Jiangsu, Shandong, and Hunan provinces. The highest percentage of K. pneumoniae were found in the samples from Shandong Province ( (Table S4).

Virulence-Related Genes in the K. pneumoniae Strains by PCR Assay
All the 94 K. pneumoniae strains tested negative for the virulence-related genes aerobactin, allS, iroN, rmpA, traT, and wcaG using the PCR assay. However, higher incidences of the entB (98.9%) and mrkD (85.1%) genes were observed, followed by the fimH (50.0%) and ybtA (14.9%) genes among the K. pneumoniae strains. The amplified products of the entB, mrkD, ybtA, and fimH genes were confirmed by DNA sequencing (Table S3).
On the basis of the limited numbers of K. pneumoniae strains obtained in this study, distinct antibiotic resistance profiles were also found among the K. pneumoniae strains in the 14 species of aquatic animals ( Table 4). For example, all the K. pneumoniae strains (n = 6) in M. anguillicaudatus were resistant to CHL, CIP, KAN, NOR, TET, and SXT. The strains from S. strictus (n = 3), M. veneriformis (n = 2), and B. areolate (n = 1) were all resistant to SXT. Conversely, the resistance to GEN was present only in a few strains from C. cahayensis

Heavy Metal Tolerance and MDR of the K. pneumoniae Strains
To get insights into the coresistance mechanism between the MDR and heavy metal tolerance, we further investigated the 32 MDR K. pneumoniae strains by using a phylogenetic analysis. As shown in Figure 3, the MDR strains were grouped into six clusters (Clus-

Genome Features of the K. pneumoniae 8-2-5-4 Isolate from Serotype K2
The 94 K. pneumoniae strains were also detected for the virulent serotypes (K1, K2, K5, K20, K54, and K57) by using the PCR assay, and the results revealed that the K. pneumoniae 8-2-5-4 isolate recovered from T. granosa tested positive for the serotype K2 gene, whereas the others tested negative for the virulent serotypes tested. Therefore, the genome sequence of this isolate was determined by using the Illumina Hiseq 10 technique. The obtained genome size of K. pneumoniae 8-2-5-4 (serotype K2) was 5,432,647 bp with 57.32% GC contents (Figure 4). Approximately 107,851 clean single reads were obtained and assembled into 64 scaffolds with a sequencing depth of 271.87-fold (on average). Notably, approximately 5042 protein-coding genes were annotated, of which 1230 encoded unknown proteins. MGEs were also identified, including GIs (n = 11), prophage gene clusters (n = 1), and ISs (n = 2) ( Table 5). The draft genome sequence of K. pneumoniae 8-2-5-4 isolate from T. granosa was deposited in the GenBank database under accession number JALJQQ000000000. Additionally, the MLST analysis revealed that K. pneumoniae 8-2-5-4 strain from serotype K2 belonged to the ST-2026. Approximately 89 genes involved in bacterial pathogenesis were identified in the K. pneumoniae 8-2-5-4 genome, which were related to adherence, antiphagocytosis, invasion, serum resistance, the iron uptake system, regulation, and the secretion system (Table S6). These results highlighted the potential virulence of K. pneumoniae 8-2-5-4. Circles from inside to outside were described in our recent research [48]. The reference genome was K. pneumoniae subsp. pneumoniae HS11286 (GenBank accession no. CP003200.1) (the third circle), and the draft genome sequence of K. pneumoniae 8-2-5-4 (the fourth circle) was determined in this study (GenBank accession no. JALJQQ000000000). Circles from inside to outside were described in our recent research [48]. The reference genome was K. pneumoniae subsp. pneumoniae HS11286 (GenBank accession no. CP003200.1) (the third circle), and the draft genome sequence of K. pneumoniae 8-2-5-4 (the fourth circle) was determined in this study (GenBank accession no. JALJQQ000000000).
Approximately 89 genes involved in bacterial pathogenesis were identified in the K. pneumoniae 8-2-5-4 genome, which were related to adherence, antiphagocytosis, invasion, serum resistance, the iron uptake system, regulation, and the secretion system (Table S6). These results highlighted the potential virulence of K. pneumoniae 8-2-5-4. K. pneumoniae 8-2-5-4 showed resistance to TET. Unexpectedly, approximately 36 resistance-related genes were identified in the K. pneumoniae 8-2-5-4 genome (Table S7), of which most genes were involved in the antibiotic efflux, antibiotic inactivation, reduced permeability to antibiotics, and antibiotic target alteration. For example, 15 genes were related to the resistance to cephalosporin and fluoroquinolone, and 12 to TET, consistent with the resistance phenotype of this bacterium.
These results indicated the potential risk in of consuming T. granosa contaminated with K. pneumoniae 8-2-5-4.

Discussion
K. pneumoniae is the second leading cause of human bloodstream infections caused by Gram-negative bacteria [1,51]. Klebsiella species have also been identified as the third leading cause of hospital-acquired infections (HAI) in the United States [52]. Nevertheless, to the best of our knowledge, the literature on K. pneumoniae in aquatic environments and aquatic products is rare so far. Barati et al. [53] reported 55 K. pneumoniae strains isolated from aquatic environment samples collected in Perak, Malaysia [53]. Effendi et al. [54] isolated 3 K. pneumoniae strains from fish samples collected in East Java [54]. Recently, Håkonsholm et al. [27] reported 78 K. pneumoniae strains isolated from marine bivalve mollusks collected along the Norwegian coast [27]. In this study, we surveyed K. pneumoniae in the 41 species of aquatic animals collected in the summer of 2018 and 2019 in Shanghai, China. Of these, 39 species had no history of carrying K. pneumoniae. Our results revealed that K. pneumoniae was present in the nine species of mollusks, three species of fish, and two species of crustaceans. The results of this study also revealed the high abundance of K. pneumoniae in benthic animals, such as C. cahayensis, M. antiquata, and P. clarkii. These findings highlighted the urgent need for policy and research on the risk assessment of the edible aquatic animals for the K. pneumoniae contamination.
None of the 94 K. pneumoniae strains identified in this study carried the virulencerelated genes aerobactin, allS, iroN, rmpA, traT, or wcaG. However, higher percentages of the entB (98.9%) and mrkD (85.1%) genes were detected, followed by the fimH (50.0%) and ybtA (14.9%) genes among the K. pneumoniae strains, which harbored nine virulence-related gene profiles. Moreover, the K. pneumoniae strains in the 14 species of aquatic animals had different virulence gene profiles. These results demonstrated a variable risk of potentially pathogenic K. pneumoniae strains in the 14 species of edible aquatic animals.
The ERIC-PCR-based fingerprinting method has been used for the relatedness or differentiation analysis of pathogenic bacteria, including K. pneumoniae [55]. For instance, Zhang et al. [41] reported 60 ERIC genotypes among 61 K. pneumoniae strains from 1200 retail foods sampled from May 2013 to April 2014 in China [41]. The ERIC-PCR conditions have been well established in our research group and applied for the genotyping of Vibro parahaemolyticu, Vibrio cholerae, and Klebsiella oxytoca strains isolated from various aquatic products [33,34,42]. In this study, the ERIC-PCR-based fingerprinting of the 94 K. pneumoniae strains revealed 76 ERIC genotypes with 63 singletons. These results demonstrated the considerable genetic diversity of the K. pneumoniae strains in aquatic animals.
Antimicrobial agents are widely applied in aquaculture to prevent animal diseases caused by pathogenic microorganisms [56]. However, the misuse or overuse of antibiotics comes with serious negative effects, such as the development of antibiotic resistance [57]. The emergence of MDR pathogens has become a global challenge and major burden for the treatment of infectious diseases [58]. K. pneumoniae is a common MDR pathogen for HAI associated with high morbidity and mortality because of the limited treatment options [59]. For example, Zhang et al. [60] reported that 7.5% of K. pneumoniae strains (n = 12) had MDR phenotypes, which were isolated from clinical samples in Wenzhou, China [60]. Fatima et al. [61] found that all K. pneumoniae strains (n = 21) were resistant to a minimum of six and a maximum of 14 antibiotics tested, which were isolated from urine (n = 15) and sputum (n = 6) samples from different patients with UTIs and respiratory tract infections (RTIs) at the Boland Medical Center in Quetta, Pakistan [61]. In this study, approximately 34.0% of the K. pneumoniae strains displayed MDR phenotypes. Notably, the highest rate (69.2%) of MDR strains was found in the fish. Moreover, the MDR phenotypes were predominant among K. pneumoniae strains in M. anguillicaudatus (100%), all of which were resistant to six antimicrobial agents. These results indicated higher antimicrobial exposure levels or contaminated sources in these aquatic animals.
The results of this study also revealed that the SXT resistance was predominant (52.1%) among the K. pneumoniae strains, one-third of which were also resistant to CHL (31.9%) and TET (27.7%). SXT belonged to the fluoroquinolone antibiotics and was frequently detected in surface water environments [62]. A high occurrence (28%) of SXT-resistant K. pneumoniae strains was also found in untreated hospital waste in Bangladesh [58]. Consistent with the results in this study, the higher occurrences of resistance to SXT, CHL and TET were also detected in K. oxytoca strains (n = 125) isolated from the 41 species of aquatic animals in our previous research [34]. In this study, certain percentages of intermediate susceptibility to CIP (18.1%) and KAN (13.8%) may suggest a potential resistance trend of these drugs. Additionally, our data also revealed different antibiotic resistance profiles among the K. pneumoniae strains in different types and species of aquatic animals. For example, the resistance rates of CHL, CIP, KAN, NOR, TET, and SXT were higher among the strains in M. anguillicaudatus, suggesting serious exposure of the aquatic animal to these antibiotic drugs. These results also highlighted the risk of the transmission of antibiotic resistance in K. pneumoniae to human populations via the food chain.
In recent decades, rapid industrialization, urbanization, and agricultural modernization may have led to the increased environment pollution by releasing a substantial amount of hazardous heavy metals [63]. For example, Zhao et al. [64] collected surface sediment samples in 14 typical intertidal areas across China, from July to September in 2015. They found that 11 areas were exposed to the moderate ecological risk of heavy metals in the sediment [64]. In this study, the 94 K. pneumoniae strains had different heavy metaltolerance profiles. For instance, the tolerance to Cu 2+ and Cr 3+ were the most prevalent (96.8%, and 96.8%), followed by Zn 2+ (91.5%), Pb 2+ (89.4%), and Hg 2+ (81.9%). Consistent with these results, high percentages of the tolerance to Cu 2+ (84.8%), Pb 2+ (80.8%), Cr 3+ (66.4%), Zn 2+ (66.4%), and Hg 2+ (49.6%) were also observed among the K. oxytoca strains, in our previous study [34]. Copper is an essential trace element needed for the human body, but it is toxic if it exceeds the specified limit [65]. Chromium can cause skin allergies and increase the risk of lung cancer [66]. Direct or indirect exposure to Pb 2+ leads to various health risks, such as cancer, stroke, memory problems, and neurological problems [67]. Metallic, inorganic mercury is a potent neurotoxicant [68]. In this study, the data showed that most K. pneumoniae strains were tolerant to Cr 3+ (100% to 92.3%), Cu 2+ (100% to 95.0%), Hg 2+ (92.3% to 76.2%), Pb 2+ (100% to 85.0%), and Zn 2+ (95.2% to 90.0%). These results suggested that the heavy metal pollution likely occurred in the aquaculture environment of the aquatic animals.
A comparison of the genomic fingerprinting profiles of the 32 MDR K. pneumoniae strains further revealed a close relatedness between the MDR and heavy metal tolerance. For example, all the MDR K. pneumoniae strains were tolerant to two or more heavy metals. Notably, eight K. pneumoniae strains were resistant to six antibiotics and also tolerant to four to six heavy metals. Frequent and persistent heavy metal pollution had a profound impact on the composition and activity of the microbial community, likely cross-selected for antibiotic resistance, and vice versa [69].
The serotypes K1, K2, K5, K54, and K57 of K. pneumoniae strains are closely associated with invasive diseases of the bacterium [9]. K1/K2 strains are generally more virulent than the others [64]. As a causative agent of pyogenic liver abscess, K2 is also frequently reported in community-acquired pneumonia [70]. In this study, our results revealed that the K. pneumoniae 8-2-5-4 strain isolated from T. granosa belonged to the serotype K2, whereas no virulent serotypes (K1, K2, K5, K54, and K57) were found in the other 93 K. pneumoniae strains. Mucoviscosity-associated gene A (magA gene) encodes a polysaccharide polymerase enzyme that is specific to serotype K1 [71,72]. In this study, the magA was absent from all the K. pneumoniae strains.
A draft genome sequence (5,432,731 bp) of K. pneumoniae 8-2-5-4 (serotype K2) was determined using the Illumina Hiseq 10 technique. This strain was identified as K. pneumoniae ST-2026. Sequences of the ST-2026 from four K. pneumoniae strains, which were isolated from chicken, from Asia (unknown detailed source), were first deposited in the public database by Zhang Lifeng, National Institute for Communicable Disease Control and Prevention, Chinese Center for Disease Control and Prevention, China, on 21 September 2015 (https://cge.food.dtu.dk/services/MLST/ (accessed on 23 November 2022)). These suggested the possible transmission of K. pneumoniae ST-2026 strains between chicken and the aquatic animal T. granosa. Remarkably, approximately 89 virulence-related genes were identified in the K. pneumoniae 8-2-5-4 genome, which related mainly to adherence, antiphagocytosis, invasion, the iron uptake system, regulation, the secretion system, and serum resistance. Of these, the highest proportion (25.8%) of the virulence-related genes functioned in adhesion. Studies have shown that adherence is essential for the colonization, biofilm formation, and epithelial cell invasion of pathogenic K. pneumoniae [73]. In this study, comparative genomic analysis also revealed 36 antibiotic resistance-related genes in the K. pneumoniae 8-2-5-4 genome. For example, the TolC gene, which encoded an important mediator of antibiotic resistance in K. pneumoniae for the outer membrane efflux channel [29], was involved in the resistance to various drug classes. The ramA gene encoded an intrinsic regulator in K. pneumoniae, which played an important role in the bacterial response to antibiotic exposure [74]. The TEM-116 and SHV-38 genes encoded extended-spectrum β-lactamases (ESBLs) that can hydrolyze penicillins, broad-spectrum cephalosporins, and monobactams [30]. The diversity of resistance genes, gene variance, and selective pressure from the environment may result in the difference between resistance phenotypes and resistance genotypes. Meanwhile, approximately 70 heavy metal-tolerance genes were identified in the K. pneumoniae 8-2-5-4 genome. For example, 16 genes were involved in Cu tolerance, 13 genes in Zn tolerance, and 19 genes in the tolerance to multiple heavy metals (e.g., Cr 3+ , Co 2+ , Cd 2+ , Hg 2+ , Ni 2+ , and Mn 2+ ), consistent with the Cr 3+ /Cu 2+ /Hg 2+ /Zn 2+ phenotype of this bacterium. These results provided the genome-level evidence for the potential virulence and coresistance of K. pneumoniae 8-2-5-4 isolate from T. granosa. The pathogenesis of the bacterium should be further investigated in the future research by using cell and animal-mode techniques.

Conclusions
In this study, we for the first time investigated the prevalence, virulence, and resistance of K. pneumoniae in 41 commonly consumed species of aquatic animals collected in the summers of 2018 and 2019 in Shanghai, China. In these 39 species, the bacterium has never been detected. K. pneumoniae was found in the 14 species of aquatic animals, including A. woodiana, B. areolata, Blotchy rock cod, C. auratus, C. cahayensis, E. sinensis, H. rubra, M. anguillicaudatus, M. antiquata, M. veneriformis, N. cumingi Crosse, P. clarkii, S. strictus, and T. granosa. Except P. clarkii, the other 13 species had no history of carrying K. pneumoniae. The high abundance of K. pneumoniae was found in benthic animals, such as C. cahayensis, M. antiquata, and P. clarkii.
Overall, the results of this study demonstrated the considerable genetic diversity and tight coresistance relatedness of the K. pneumoniae strains from the 14 species of aquatic animals, and they fill gaps in the risk assessment of K. pneumoniae in edible aquatic animals.
Supplementary Materials: The following supporting information can be downloaded at https: //www.mdpi.com/article/10.3390/foods11244058/s1: Table S1-The 41 species of aquatic animals and the isolated K. pneumoniae strains; Table S2-The oligonucleotide primers used in this study [75,76]; Table S3-The sequenced 16S rRNA and virulence-related genes in the representative K. pneumoniae strains; Table S4-The distribution of the K. pneumoniae strains in the aquatic animal samples originating from different regions and fish markets in China; Table S5-Heavy metal tolerance of the 94 K. pneumoniae strains evaluated in this study; Table S6-The predicted virulence-related genes in the K. pneumoniae 8-2-5-4 genome; Table S7-The predicted antibiotic resistance-related genes in the K. pneumoniae 8-2-5-4 genome; Table S8-The prediction heavy metal resistance-related genes in the K. pneumoniae 8-2-5-4 genome.