Genomic Analysis of Aeromonas veronii C198, a Novel Mcr-3.41-Harboring Isolate from a Patient with Septicemia in Thailand

The resistance of Gram-negative bacteria to colistin, mediated by plasmid-borne mcr genes, is an emerging public health concern. The complete genome sequence (4.55 Mb) of a clinical isolate of Aeromonas veronii biovar veronii obtained from a patient with septicemia was determined using short-read and long-read platforms. This isolate (C198) was found to harbor a novel mcr-3 gene, designated mcr-3.41. Isolate C198 revealed adjacent mcr-3.41 and mcr-3-like genes. It contained one chromosome and two plasmids, both of which encoded a RepB replication protein. Other antimicrobial resistance genes, including blacphA3, blaOXA-12, tetA, rsmA, and adeF, were also present. Isolate C198 was resistant to amoxicillin–clavulanate, ampicillin–sulbactam and tetracycline, and showed intermediate resistance to trimethoprim–sulfamethoxazole. The isolate was susceptible to piperacillin–tazobactam, carbapenem, third-generation cephalosporins, fluoroquinolones, chloramphenicol, and aminoglycosides. Putative virulence genes in the C198 genome encoded type II, III, and VI secretion systems; type IV Aeromonas pili; and type I fimbria, flagella, hemagglutinin, aerolysin, and hemolysins. Multilocus sequence typing revealed a novel sequence type (ST), ST720 for C198. Phylogenetic analysis of the single nucleotide polymorphisms in C198 demonstrated that the strain was closely related to A. veronii 17ISAe. The present study provides insights into the genomic characteristics of human A. veronii isolates.


Introduction
Aeromonas veronii is a Gram-negative bacterium found in a variety of environmental niches, including water, seafood, meat, and vegetables, and occasionally in the feces of healthy individuals [1,2]. It can cause several diseases in humans, including wound infections, pneumonia, hemolytic uremic

General Characteristic of A. veronii C198
In 2016, a Gram-negative bacterium was isolated from a blood specimen of a 68-year-old male with septicemia who was admitted to a tertiary hospital in southern Thailand. He was treated with ceftriaxone, and he eventually recovered. Based on the results of a matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) analysis, the isolate, which was named strain C198, was determined to be A. veronii. Analysis of its genome sequence using average nucleotide identity (ANI) and Kraken2 assigned the isolate to the biovar veronii, out of the two A. veronii biovars (veronii and sorbria). The ANI value between strain C198 and the type strain of A. veronii CECT 4257 T was 96.15%, above the species cut-off level of 95%.
A vast majority (95.4%) of the Aeromonas spp. infections in humans are caused by only four species: A. caviae (37.26%), A. dhakensis (23.49%), A. veronii (21.54%), and A. hydrophila (13.07%) [1]. Although the global impact of Aeromonas in human infections is unknown, a study in California revealed that the annual incidence of Aeromonas infections was 10.5 cases per million people [1]. The estimated incidence of Aeromonas bacteremia in France was 0.66 cases per million people [9]. Another study showed that the incidence of bacteremia in Taiwan was 76 cases per one million people [10]. In Thailand, Aeromonas spp. have been clinically isolated from human cases of bacteremia and peritonitis [11][12][13].

Genome Features
The whole-genome sequence of A. veronii C198 is composed of 4,575,001 bp (N 50 = 4,550,752 bp) with 58.6% GC content. The final genomic assembly had three contigs, one circular chromosome (4,550,752 bp), and two circular plasmids (13,923 and 10,326 bp). The genome was predicted to contain 4223 genes, including 4065 coding sequences (CDS), 11 5S rRNA genes, 10 16S rRNA genes, 10 23S rRNA genes, 123 tRNA genes, and 4 ncRNA genes. There were no clustered regularly interspaced short palindromic repeats. Tekedar et al. (2019) reported a comparison of 41 publicly available A. veronii genomes from various sources. Their genome size ranged from 4.28 to 4.95 Mb, and their G + C ratios ranged from 58.1 to 58.9%. Of these 41 genomes, only strain AVNIH1 had a plasmid [2].
The two plasmids of the A. veronii C198 isolate are 13,923 and 10,326 bp in length, encoding 15 and 12 proteins, respectively. The larger plasmid contains a tetA gene that was predicted to encode the tetracycline efflux MFS transporter, Tet (A). Both plasmids carried genes encoding AAA family ATPases, EamA family transporters, relaxases, Tn3-like element TnAs1 family transposases, integrase domain-containing proteins, type II toxin-antitoxin system RelE/ParE family toxins, cysteine hydrolases, ribbon-helix-helix CopG family proteins, GGDEF domain-containing proteins, sel1 repeat family proteins, and transcriptional regulators. The gene encoding the plasmid replication protein RepB was also present in both plasmids. PlasmidFinder and PLACNETw did not identify the Inc and MOB replicon types, respectively.
The single nucleotide polymorphism (SNP) phylogeny demonstrated that C198 was closely related to a fish isolate (17ISAe) in Korea ( Figure 2). There are four highly conserved genetic subgroups for this: (1) USA dairy cattle isolates and an isolate from Pamvotida Lake, Greece; (2) strain ML09-123 (USA) and strain TH0426 (catfish isolate) from China; (3) human isolates (strains CECT4257 T , CCM4359, and AER397) from the USA and a sediment isolate from China (B565); and (4) surface water isolates from the USA and Germany [2]. As shown in Figure [2]. Interestingly, human A. veronii isolates are distributed throughout the phylogenetic tree, which means they are genetically diverse. protein RepB was also present in both plasmids. PlasmidFinder and PLACNETw did not identify the Inc and MOB replicon types, respectively.
The single nucleotide polymorphism (SNP) phylogeny demonstrated that C198 was closely related to a fish isolate (17ISAe) in Korea ( Figure 2). There are four highly conserved genetic subgroups for this: (1) USA dairy cattle isolates and an isolate from Pamvotida Lake, Greece; (2) strain ML09-123 (USA) and strain TH0426 (catfish isolate) from China; (3) human isolates (strains CECT4257 T , CCM4359, and AER397) from the USA and a sediment isolate from China (B565); and (4) surface water isolates from the USA and Germany [2]. As shown in Figure 2 [2]. Interestingly, human A. veronii isolates are distributed throughout the phylogenetic tree, which means they are genetically diverse.

Antimicrobial Susceptibility and Resistance Genes
Antimicrobial susceptibility testing revealed that A. veronii C198 was resistant to amoxicillinclavulanate, ampicillin-sulbactam, and tetracycline, but showed intermediate resistance to trimethoprim-sulfamethoxazole (Table 1). It was susceptible to piperacillin-tazobactam, carbapenem, third-generation cephalosporins, fluoroquinolones, chloramphenicol, and aminoglycosides (Table 1). Some studies have reported that A. veronii is usually resistant to amoxicillin-clavulanate and is susceptible to carbapenem, third-and fourth-generation cephalosporins, aminoglycosides, monobactam, and fluoroquinolones [1,8,14,15]. The minimal inhibitory concentration (MIC) of colistin for our isolate was 2 µg/mL, which could not be interpreted as susceptible, intermediate, or resistant by either the Clinical Laboratory Standards Institute (CLSI) (M45) or EUCAST guidelines due to the lack of a clinical breakpoint for colistin in Aeromonas spp. However, this MIC was similar to those of A. veronii strains 172, Z2-7, and w55, which were also 2 µg/mL [8,14,15].   Whole-genome sequencing revealed the presence of mcr-3, bla cphA3 , bla OXA-12 , tetA, rsmA, and adeF genes, which are likely responsible for the observed resistance to various antimicrobials (Tables 1  and 2). Most of these resistance genes are located on the chromosome, except for tetA, which is located on the larger (13,923 bp) plasmid. The presence of bla cphA3 and bla OXA-12 may contribute to the observed resistance to ampicillin-sulbactam and amoxicillin-clavulanate. However, these genes do not confer resistance to piperacillin-tazobactam, third-generation cephalosporins, or carbapenem, because our isolate was susceptible to these antimicrobials. The tetA gene likely contributed to tetracycline resistance, and mcr-3.3 may be responsible for the observed MIC of colistin (2 µg/mL).
Comparison of the antimicrobial resistance genes in our isolate to those in other A. veronii isolates revealed the presence of mcr-3 in the genomes of isolates from China, Korea, and the USA ( Table 2). The pattern of antimicrobial resistance genes in C198 was similar to those in strains 126-14 and HX3, which were isolated from humans and alligators, respectively, in China; only the tetA gene was different. Almost all isolates contained bla cphA , bla OXA-12 , adeF, rsmA, and EF-Tu mutant (R234F). This suggests that these genes are conserved in A. veronii. The current study revealed that mcr-3 genes were present in some, but not all, A. veronii genomes ( Table 2), indicating that A. veronii does not inherently carry the mcr-3 gene. However, A. veronii may be a reservoir for the dissemination of mcr-3 genes to other bacteria, as previously reported [14,16].
The genetic organization surrounding mcr-3 in the C198 isolate was examined. As shown in Figure 3, the mcr-3 and mcr-3-like genes were adjacent to each other. Similar findings have been reported for two adjacent mcr-3 genes in other Aeromonas spp., including A. veronii isolates [8,14]. Among the previously evaluated genes, only mcr-3.3 conferred colistin resistance, whereas mcr-3-like genes did not affect the MIC of colistin [14]. Both the mcr-3 and mcr-3-like genes of the C198 isolate were located on the chromosome and were flanked by diacylglycerol kinase (dgk) and ISAeca5 family transposase genes downstream, and an EamA family transporter and the IS3 family transposase genes upstream (Figure 3). This is similar to the gene arrangement observed in isolate HX3. A common gene found in all mcr-3-harboring A. veronii isolates is dgk, which is located downstream of the mcr-3-like gene in C198 (Figure 3).   S1). All of these data indicated that the MCR-3.41 was more closely related to MCR-3.25 than to MCR-3.3. The phylogenetic tree of all mcr-3 protein sequences is shown in Figure 4. Our novel mcr-3.41 was clustered with mcr-3.25 according to the alignment analysis result as described above. An analysis of the mcr-3-like genes among strains C198, 172, and HX3 demonstrated 100% identity among C198, 126-14, and HX3, and 99.81% identity between C198 and 172 (Supplementary File S1).

Virulence Factors
The virulence genes detected in the C198 isolate included genes encoding type II, III, and VI secretion systems, tap type IV pili, type I fimbria, flagella, hemagglutinin, aerolysin, and hemolysins (Table 1). A. veronii possesses a type II secretion system that exports hydrolytic enzymes, hemolysin, and aerolysin [1,2]. Type VI systems are known to inject protein effectors, such as G repeat proteins (VgrG) and hemolysin-co-regulated protein (Hcp), directly into the cytosol of the target cell [1]. However, another study reported that seven of nine human A. veronii isolates lack a type III system, indicating that specific system may not be crucial for its virulence in mammals [2].
Hemolysins, including aerolysin, are capable of forming pores in the cell membrane, leading to the osmotic lysis of target cells [1]. In addition, aerolysins may alter the permeability of blood cells and other eukaryotic cells, resulting in cell lysis [17]. Tap type IV pili, type I fimbria, flagella, and hemagglutinin are involved in host cell adhesion [1,17].
The pathogenesis of A. veronii infection is complex and is an area of active investigation. A recent study identified several virulence factors that were positively correlated with the pathogenicity of A. veronii [18]. These virulence genes contribute significantly to the development of infections, and A. veronii isolates carrying more virulence genes were more virulent in mice [17,18].

Bacterial Strain
A laboratory-based surveillance program for the detection of emerging species of antimicrobialresistant bacteria in 11 hospitals (Sakon Nakhon, Nakhon Phanom, Surin, Udonthani, Mukdahan, Bueng Kan, Chumporn, Surat Thani, Tak, Phayao, and Bangkok) within representative provinces in Thailand has been ongoing since 2016. The A. veronii strain described in this study was isolated from the blood of a patient with septicemia admitted to a tertiary hospital in southern Thailand.
A sequencing library was generated using the NEBNext Ultra II DNA Library Prep Kit for Illumina (New England Biolabs, Ipswich, MA, USA) according to the manufacturer's recommendations. We applied Fastp v0.19.5 [20] for quality filtering of Illumina reads. Adapters were trimmed using Skewer v0.2.2 [21]. Quality checking of the reads was performed using FastQC v0.11.8 (https: //www.bioinformatics.babraham.ac.uk/ projects/fastqc/). Hybrid assemblies of ONT and Illumina data were generated using Unicycler v0.4.8 [22], and the quality of the genome sequences was checked using QUAST v5.0.2 [23]. The circular DNA structures of the bacterial chromosome and This is a free program. No company owns. We used it via website, so, I added the reference is enough.plasmids were computationally produced using Unicycler software. Genome sequences were submitted to the NCBI Prokaryotic Genome Annotation Pipeline v4.12 for validation. Default parameters were used for all software, unless otherwise specified.
Multilocus sequence typing (MLST) analysis to determine the STs of Aeromonas spp. was performed using PubMLST (https://pubmlst.org/aeromonas/). A Google BURST analysis of the STs was performed using PHYLOViZ 2.0 [31]. Genomic sequences were compared using a reference genome-based SNP strategy using REALPHY [32]. A phylogenetic tree was constructed from the REALPHY data using MEGA-X with the neighbor-joining method (500 bootstrap replicates) by applying the Tamura three-parameter model [33]. The reference genomes of 40 isolates of A. veronii (isolated from humans (n = 16), animals (n = 18), and external environments (n = 6) were downloaded from GenBank for the REALPHY analysis (Table 2). In the case of phylogenetic analysis of all mcr-3 proteins, the tree was constructed using MEGA-X via the neighbor-joining method with 500 bootstrap replicates by applying the Poisson model [33]. The mcr-3 protein sequences used in this study are shown in Supplementary File S2.

Accession Numbers
The complete sequence was submitted to GenBank under the BioProject accession number: PRJNA525849, BioSample accession number: SAMN15587301, and accession number: JACEGL000000000.

Ethics
The Human Research Ethics Committee Office of Osaka University reviewed this protocol and approved this study. The ethics approval number was 14468-5. This study was conducted according to the principles of the Declaration of Helsinki.

Conclusions
This study revealed the clinical isolate A. veronii obtained from a patient with septicemia contained adjacent mcr-3 and mcr-3-like genes. In addition to these, other antimicrobial resistance genes and virulence genes were also present. Phylogenetic analysis demonstrated that the strain was closely related to the A. veronii strain obtained from a discus fish. Combining ONT and Illumina sequencing data provided insights into the genomic characteristics of human A. veronii isolates.