Comparative Genomic Analysis of Virulent Vibrio (Listonella) anguillarum Serotypes Revealed Genetic Diversity and Genomic Signatures in the O-Antigen Biosynthesis Gene Cluster

Vibrio anguillarum is the most frequent pathogen affecting fish worldwide. The only known virulent strains of V. anguillarum are serotypes O1, O2, and O3. Genetic differences between the serotypes that could shed insight on the evolution and serotype differences of this marine pathogen are unknown. Here, we fully sequenced and characterized a strain of V. anguillarum O1 (J382) isolated from winter steelhead trout (Oncorhynchus mykiss irideus) in British Columbia, Canada. Koch’s postulates using the O1 strain were replicated in naïve lumpfish (Cyclopterus lumpus) and compared to O2. Phenotypic and genotypic comparisons were conducted for serotypes O1, O2, and O3, using biochemical tests and bioinformatic tools, respectively. The genome of V. anguillarum O1 (J382) contains two chromosomes (3.13 Mb and 1.03 Mb) and two typical pJM1-like plasmids (65,573 and 76,959 bp). Furthermore, V. anguillarum O1 (J382) displayed resistance to colistin sulphate, which differs from serotype O2 and could be attributed to the presence of the ugd gene. Comparative genomic analysis, among the serotypes, showed that intra-species evolution is driven by insertion sequences, bacteriophages, and a different repertoire of putative ncRNAs. Genetic heterogeneity in the O-antigen biosynthesis gene cluster is characterized by the absence or the presence of unique genes, which could result in differences in the immune evasion mechanisms employed by the respective serotypes. This study contributes to understanding the genetic differences among V. anguillarum serovars and their evolution.


Introduction
Reports of vibriosis in freshwater, brackish water, and marine aquatic animals caused by the Gram-negative bacterium, commonly known as Vibrio anguillarum (Listonella anguillarum), have been recorded as early as 1893 and continue today [1,2]. This pathogen is important because it causes high morbidity and mortality in aquaculture, resulting in (BioSample: SAMN25088161). The strain was identified as serotype O1, using serotyping and a species-specific PCR, using the rpoS gene [25]. The bacterium was revived from cryopreservation and routinely grown on trypticase soy agar (TSA; Difco, Franklin Lakes, NJ, USA), supplemented with 2% NaCl and 1.5% bacto agar (Difco, Franklin Lakes, NJ, USA), and incubated for 48 h at 15 • C [5]. The routine cultures produced following the standard procedures, as described above, were maintained in 3 mL of trypticase soy broth (TSB; Difco, Franklin Lakes, NJ, USA), and supplemented with 2% NaCl at 15 • C for 18-24 h until the mid-logarithmic phase was reached (optical density (OD 600 ) ≈ 0.7 (~4.1 × 10 8 CFU/mL)). When required, TSB was supplemented with 100 µM 2,2-dipyridyl (iron-limited condition) and 100 µM of FeCl 3 (iron-rich condition). Luria-Bertani broth (10 g tryptone, 10 g NaCl; 5 g yeast extract; double distilled H 2 O up to 1 L) was modified to 0%, 0.5%, or 2% NaCl to investigate the NaCl requirements. For investigating the production of siderophores, chrome azurol S (CAS) plates were used [25]. The bacterial hemolytic activity was evaluated using TSA plates supplemented with 5% sheep blood.

Physiological Characterization of V. anguillarum J382
The bacterial growth was determined under different temperatures, NaCl concentrations, and sugar requirements. For optimum temperature requirement analysis, the microorganism was grown on 2% NaCl TSA at 4, 15, 28, and 37 • C for 48 h. To evaluate the NaCl requirements, the microorganism was grown in 3 mL of LB broth supplemented with 0%, 0.5%, 1.5%, and 2% NaCl and incubated at 15 • C for 24 h. The oxidation/fermentation (OF) of semi-solid media (5.0 g NaCl; 0.3 g di-potassium phosphate; 2.0 g peptone; 0.03 g bromothymol blue; 3.0 g agar; 10 g glucose; double distilled H 2 O up to 1 L) supplemented with 1% of individual sugars (glucose, arabinose, glycogen, maltose, lactose) was used to determine the sugar utilization. The hemolytic activity and motility were determined by standard methods [3].

Siderophore Detection
The secretion of siderophores was tested using CAS plates by previously described assays [5] with minor modification. Briefly, V. anguillarum J382 was first cultured in 3 mL 2% NaCl TSB, as described above, at 15 • C for 24 h until the mid-log phase was reached. Subsequently, 30 µL of the inoculum was grown at 15 • C for 24 h, with shaking in 2% NaCl TSB (bacterial control) and the media supplemented with either 100 µM of FeCl 3 (iron-rich condition), or 100 µM 2,2-dipyridyl (iron-limited condition). After incubation, the cells were pelleted at 4200× g at 15 • C for 10 min, washed and resuspended in 1 mL of PBS. The siderophore secretion was analyzed by inoculating 10 µL of the resuspended cells onto the CAS agar plates and incubating for 48-72 h at 15 • C. The bacterial colonies were visually checked for the presence of a yellow-orange halo around them, as an indication of the secretion of siderophores.

Biochemical, Enzymatic, and Physiological Characterization
The biochemical and enzymatic profiles were performed using API 20 NE, API 20 E, and API ZYM strips, according to the manufacturer's instructions (bioMerieux, Marcyl'Étoile, France). After incubating the strips with V. anguillarum J382 at 15 • C for 48 h, the results were analyzed using apiweb software (bioMerieux). Tests for catalase and oxidase methods were carried out using standard methods [3].

Bacterial Infection and Tissue Sampling
Infection assays of V. anguillarum J382 were carried out in lumpfish of 50-70 g at the federal certified AQ3 biocontainment zone in the Cold-Ocean Deep-Sea Research Facility (CDRF). Three groups of experimental fish were set up in 3 tanks (n = 45 per tank) of 500 L and were acclimatized for 2 weeks under optimal conditions (8-10 • C, 95-110% air saturation, ambient photoperiod, and UV-treated, filtered, flow-through seawater). The V. anguillarum J382 was culture, as described (Section 2.1), until the mid-logarithmic phase (optical density (OD 600 ) ≈ 0.7 (~4.1 × 10 8 CFU/mL)). The cells were harvested by centrifugation at 4200× g for 10 min at room temperature and washed three times with phosphate-buffered saline (PBS; 136 mM NaCl, 2.7 mM KCl, 10.1 mM Na 2 HPO 4 , 1.5 mM KH 2 PO 4 (pH 7.2)). After washing, the cells were diluted to~1.0 × 10 6 CFU/mL and 1.0 × 10 7 CFU/mL for bath infection. Quantification of the bacterial was carried out using the plate count method. The lumpfish in tanks 1 and 2 were bath-infected for 30 min with freshly prepared 10 6 CFU/mL and 10 7 CFU/mL bacterial inoculum of V. anguillarum J382, respectively, while the lumpfish in tank 3 were used as the non-infected control. Mortality was monitored for 15 d and a one-way ANOVA statistical analysis was carried out using GraphPad Prism 9.0.0 (121) software. At 5 days post-infection (dpi), 6 fish were sampled from each tank, euthanized with 400 mg/L of MS222, and the liver, head kidney, and spleen were aseptically removed and placed in sterile homogenized bags (Nasco, Whirl-Pak ® , Fort Atkinson, WI, USA). The weight of each organ was taken, and the samples were homogenized and adjusted to a final volume of 1 mL (weight: volume) using sterile PBS. The tissue mixtures were serially diluted in sterile PBS (1:10) and plated onto 2% NaCl TSA at 15 • C for 48 h. Bacterial identification was conducted using the API system, according to the manufacturer's instructions, and bacterial count per gram to tissue was determined as below [28]: Dilution Factor × 10 Tissue weight (grams) 2.5. Genomic DNA Preparation and Whole Genome Sequencing V. anguillarum J382 was grown until the mid-log phase, as described above. The bacterial cells were then pelleted by centrifugation at 4200× g for 10 min at room temperature and washed with sterile PBS 3 times. A commercial DNA extraction kit (Wizard High-Molecular-Weight DNA Extraction Kit; Promega, Madison, WI, USA) was used to extract the genomic DNA, according to the manufacturer's protocol. The purity and integrity of the extracted DNA was analyzed by spectrophotometry (Genova Nano Spectrophotometer, Jenway; Staffordshire, UK) and on 0.8% agarose gel using gel electrophoresis, respectively. High-quality DNA samples were sent for library preparation and whole genome sequencing using 2 platforms, namely, PacBio SMRT cell and Illumina MiSeq PE250, at Genome Quebec (Montreal, QC, Canada). The whole genome sequence was submitted to the National Center for Biotechnology Information (NCBI) database.

Data Preprocessing, Genome Assembly, Mapping and Annotation
The PacBio reads were assembled at Genome Quebec using the Hierarchical Genome Assembly Process (HGAP) assembler v.4 [29] with 30X coverage cutoff [28,30]. Confirmation of the assembly was carried using the CLC Genomics Workbench (CGWB) v.20.0.4 (Qiagen, Hilden, Germany). The Illumina raw reads were trimmed, and low-quality reads were filtered out using the CGWB and the reads quality was analyzed by FastQC v12 [31]. The PacBio long reads were assembled into circular contigs using the CGWB. The PacBio assembled genome was polished using the trimmed high-quality Illumina reads and the CGWB de novo tool in the genome finishing module tools with default settings. The plasmids were assembled using the Illumina reads that did not align with the V. anguillarum J382 chromosomes, following the above-mentioned assembly protocol. The assembled chromosomes and plasmids were annotated using the Rapid Annotation Subsystem Technology pipeline (RAST) [32] and comprehensive genome analysis implemented in PATRIC v3.6.12 [33]. The assembled chromosomes and plasmids were re-annotated by the NCBI prokaryotic genome annotation pipeline (PGAP). The obtained 2 chromosomes and plasmids were further mapped using the web server CGView Server BETA with Proksee annotation (https://sites.ualberta.ca/~stothard/ (accessed on 13 March 2023)).

Plasmid Analysis
The number of plasmids in V. anguillarum J382 was determined using alkaline extraction, described elsewhere [34]. The extracted plasmids were detected on 0.5% agarose gel at 40 V overnight. The gels were stained with ethidium bromide and viewed using the Invitrogen iBright CL1500 Imaging System (Thermo Fisher Scientific, Waltham, MA, USA).

Comparative Genomic Analysis, Synteny and Phylogeny Analysis
The CGWB was used to compare the genomes listed in Table 1. Chromosomes 1 and 2 were analyzed separately. Whole genome multiple alignment was performed using the whole genome analysis tool with default settings (minimum initial seed length = 15 (allow mismatches in seeds), minimum alignment block length = 100 (rearrange contigs)). The average nucleotide identity comparison was conducted using the default parameters: minimum similarity fraction = 0.8, minimum length fraction = 0.8. Comparative heat maps were constructed following the default parameters of the distance measure of 1 Pearson correlation and complete linkage criteria. For synteny studies, whole genome dot plots were constructed between V. anguillarum J382 and representative genomes in serotypes O1 (V. anguillarum 87-9-116), O2 (V. anguillarum J360), and O3 (V. anguillarum CNEVA NB11008), using the parameters: minimum initial seed length = 15 and allow mismatches in the seeds (default parameters). The separate whole genome multiple alignment files for chromosomes 1 and 2 created using the CLC Bio Workbench were exported as FASTA files, used to construct phylogenetic trees in MEGA 11 [35] using the neighbor-joining statistical method [36] with 1000 replicate bootstraps. The strain P. damselae subsp. Piscicida 91-197 was used as the out group to root the tree.

Determination of O-Antigen Biosynthesis Genes
The O-antigen biosynthesis genes, their corresponding pathways, and the sugars they encode were determined using bioinformatic tools. The KEGG platform [29] was used to annotate and identify the O-antigen biosynthesis genes in V. anguillarum J382 and the representative genomes in the serotypes O1 (V. anguillarum 87-9-116), O2 (V. anguillarum J360), and O3 (V. anguillarum CNEVA NB11008). Firstly, whole genome amino acid sequences in FASTA format were downloaded from either RAST for V. anguillarum J382 or from the NCBI for the remaining genomes, and KO identifiers/K numbers (assignment of orthologous genes based on sequence similarity) were assigned to the genes. Two platforms implemented in KEGG were used for the K number assignment, namely BlastKOALA [29] that uses BLASTP sequence similarity search (parameters: taxonomy group of the genome, Bacteria and KEGG GENES database search file, genus_prokaryotes) and GhostKOALA [37], which performs automatic KO assignment by GHOSTX sequence similarity search (parameters: KEGG GENES database search file, genus_prokaryotes). The resulting files were uploaded into KEGG Mapper, Reconstruct [38] for O-antigen biosynthesis pathways identification in the genomes, including the resulting oligosaccharides. The predicted O-antigen oligosaccharides in each serotype were manually presented as a Venn diagram. Comparative analysis and visualization of the O-antigen biosynthesis genes, among the three serotypes, was performed using ggplot2 [39] implemented in R statistical software (v4.1.2; R Core Team 2021). The LPS profiles were evaluated by SDS-PAGE and visualized by silver staining [40][41][42].

Statistical Analysis
Statistical analysis of the mortality between the bath challenged groups and the control groups was conducted using a one-way ANOVA in GraphPad Prism 9.0.0 (121).

Phenotypic Characterization of V. anguillarum J382
The phenotypic characterization of V. anguillarum J382 shown in Table 2 and Figure S1, revealed that this strain exhibited the characteristics of marine Vibrio spp. [50,51]. These included halophilic metabolism; positive reactions for indole, catalase, and oxidase; the fermentation of D-glucose, mannitol, sorbitol, and arabinose; the reduction of nitrate to nitrite; and the hydrolysis of L-arginine. V. anguillarum J382 was sensitive to the vibriostatic agent 0129, and the antibiotics chloramphenicol, oxytetracycline, tetracycline, sulfamethoxazole, and oxalinic acid, and resistant to ampicillin, similar to other members in the genus [50]. Interestingly, V. anguillarum J382 was resistant to colistin sulphate. The siderophore-dependent iron acquisition was also analyzed. V. anguillarum J382 was able to synthesize siderophores under iron-limited conditions (2% NaCl TSB supplemented with 100 µM 2,2-dipyridyl), whilst sideorophore synthesis was not observed under iron-rich conditions (2% NaCl TSB supplemented with 100 µM FeCl 3 ). The hemolysin activity of V. anguillarum J382 on sheep blood agar was temperature dependent. Hemolysis was observed after incubation at 28 • C, while there was no hemolysis at 15 • C. Arginine 38 (sensitive) Phenylacetic acid -+: growth or positive reaction; +++: substantial growth; -: negative growth or reaction.

Clinical Signs and Fish Mortality
The lumpfish bath-infected with V. anguillarum J382 exhibited clinical signs that are consistent with vibriosis, including skin discoloration, hemorrhage on the fins and fish body, lesions around the mouth, reddened anus, bloody eyes, and exophthalmia (Figure 1a-g). Mortality in the two doses evaluated, started at 5 days post-infection (dpi) and all the fish were dead by 7 and 9 dpi after the high (10 7 CFU/mL) and lower dose (10 6 CFU/dosemL), respectively ( Figure 1h). There was a significant difference in mortality between the infected groups and the control group (p = 0.0153).

Clinical Signs and Fish Mortality
The lumpfish bath-infected with V. anguillarum J382 exhibited clinical signs that a consistent with vibriosis, including skin discoloration, hemorrhage on the fins and fis body, lesions around the mouth, reddened anus, bloody eyes, and exophthalmia (Figu 1a-g). Mortality in the two doses evaluated, started at 5 days post-infection (dpi) and a the fish were dead by 7 and 9 dpi after the high (10 7 CFU/mL) and lower dose (1 CFU/dosemL), respectively ( Figure 1h). There was a significant difference in mortali between the infected groups and the control group (p = 0.0153).  (e) reddening of the anus; (f) exophthalmia; and (g) hemorrhage in the eye. The scale bar was added using ImageJ software; (h) daily survival rate of V. anguillarum J382 infected lumpfish. Colonization of V. anguillarum J382 in lumpfish organs at 5 dpi using (i) 10 6 CFU/mL and (j) 10 7 CFU/mL.

Tissue Colonization of V. anguillarum J382 Infected Lumpfish
Tissues from all the organs (heart, spleen, head kidney, and liver) isolated from the lumpfish infected with 10 6 CFU/mL or 10 7 CFU/mL doses of V. anguillarum J382 were heavily colonized with bacteria. The bacterial loads ranged from >10 7 CFU/g tissue in the heart to >10 10 CFU/g tissue in the other tissue organs (Figure 1i,j). The highest bacterial load (up to 10 12 CFU/g tissue) was recorded in the spleen for both the 10 6 CFU/mL and 10 7 CFU/mL doses. The bacteria isolated from the colonized fish organs were characterized using the API system ( Figure S2) and the phenotypic characterization confirmed the identity of the recovered bacteria as V. anguillarum.

Genome Sequencing and Characterization
The complete genome of V. anguillarum J382 consists of two chromosomes and two plasmids (accession numbers CP091185 for chromosome 1, CP091186 for chromosome 2, CP091187 for plasmid pVA_O1_1, and CP091188 for plasmid pVA_O1_2). The sizes of the two chromosomes are 3,133,133 bp and 1,038,699 bp (Table S1), with GC content of 44.64% and 44.04%. Chromosome 1 has 300 subsystems, 2914 protein-coding sequences (CDS), 95 tRNA, and 28 rRNA sequences (Figure 2a). On the other hand, chromosome 2 has 58 subsystems, 1007 CDS, and 4 tRNA sequences (Figure 2b). Table S2 shows the genes associated with pathogenesis and environmental adaption that were identified in chromosomes 1 and 2. Chromosome 1 contains most of the genes related to virulence, whose functions are related to iron acquisition and metabolism, motility, and chemotaxis, siderophore biosynthesis and transport, type II and IV secretion systems and toxins, and superantigens. Both chromosomes contain the genes related to cell wall and capsule formation, invasion and intracellular resistance, resistance to toxic compounds, antibiotic resistance, and hemolysins.

Tissue Colonization of V. anguillarum J382 Infected Lumpfish
Tissues from all the organs (heart, spleen, head kidney, and liver) isolated from the lumpfish infected with 10 6 CFU/mL or 10 7 CFU/mL doses of V. anguillarum J382 were heavily colonized with bacteria. The bacterial loads ranged from >10 7 CFU/g tissue in the heart to >10 10 CFU/g tissue in the other tissue organs (Figure 1i,j). The highest bacterial load (up to 10 12 CFU/g tissue) was recorded in the spleen for both the 10 6 CFU/mL and 10 7 CFU/mL doses. The bacteria isolated from the colonized fish organs were characterized using the API system ( Figure S2) and the phenotypic characterization confirmed the identity of the recovered bacteria as V. anguillarum.

Genome Sequencing and Characterization
The complete genome of V. anguillarum J382 consists of two chromosomes and two plasmids (accession numbers CP091185 for chromosome 1, CP091186 for chromosome 2, CP091187 for plasmid pVA_O1_1, and CP091188 for plasmid pVA_O1_2). The sizes of the two chromosomes are 3,133,133 bp and 1,038,699 bp (Table S1), with GC content of 44.64% and 44.04%. Chromosome 1 has 300 subsystems, 2914 protein-coding sequences (CDS), 95 tRNA, and 28 rRNA sequences (Figure 2a). On the other hand, chromosome 2 has 58 subsystems, 1007 CDS, and 4 tRNA sequences (Figure 2b). Table S2 shows the genes associated with pathogenesis and environmental adaption that were identified in chromosomes 1 and 2. Chromosome 1 contains most of the genes related to virulence, whose functions are related to iron acquisition and metabolism, motility, and chemotaxis, siderophore biosynthesis and transport, type II and IV secretion systems and toxins, and superantigens. Both chromosomes contain the genes related to cell wall and capsule formation, invasion and intracellular resistance, resistance to toxic compounds, antibiotic resistance, and hemolysins. Two similar plasmids (pVA_O1_1 and pVA_O1_2) were identified in V. anguillarum J382, as shown in Figure 3a,b. Summarized details of the two plasmids are shown in Table  S1. The plasmid pVA_O1_1 is 65,573 bp in length, whilst the plasmid pVA_O1_2 is 76,959 bp in length, with similar GC content of 42.59% and 42.6%, respectively. Both plasmids have two subsystems, and 69 CDS for plasmid pVA_O1_1 and 106 CDS for plasmid pVA_O1_2. The genes contained in both plasmids are the same except that plasmid pVA_O1_2 contains multiple copies of genes, such as 4′-phosphopantetheinyl transferase (entD), ferric vibriobactin, enterobactin transport system, ATP-binding protein (viuG), The maps were drawn using CGView BETA Server and the annotation was carried out using Prokka software implemented in CGView.
Two similar plasmids (pVA_O1_1 and pVA_O1_2) were identified in V. anguillarum J382, as shown in Figure 3a,b. Summarized details of the two plasmids are shown in Table S1. The plasmid pVA_O1_1 is 65,573 bp in length, whilst the plasmid pVA_O1_2 is 76,959 bp in length, with similar GC content of 42.59% and 42.6%, respectively. Both plasmids have two subsystems, and 69 CDS for plasmid pVA_O1_1 and 106 CDS for plasmid pVA_O1_2. The genes contained in both plasmids are the same except that plasmid pVA_O1_2 contains multiple copies of genes, such as 4 -phosphopantetheinyl transferase (entD), ferric vibriobactin, enterobactin transport system, ATP-binding protein (viuG), isochorismatase (entB), isochorismate synthase (entC), 2,3-dihydroxybenzoate-AMP ligase (entE), histidine decarboxylase (hdc), and genes that encode for the efflux ABC transporter, permease/ATP-binding protein, transport ATP-binding protein CydCD, and non-ribosomal peptide synthetase modules, siderophore biosynthesis (Supplementary File S2). The genes associated with virulence and adaptation identified in both plasmids are related to iron acquisition and metabolism, siderophore biosynthesis and transport, and toxins and superantigens (Table S2). Verification of the presence of the plasmids in V. anguillarum J382 revealed a single DNA band around 23 Mda (Figure 3c), which could be due to the small size difference between the plasmids. A blastn search in the NCBI database revealed that both plasmids, pVA_O1_1 and pVA_O1_2, are 100% and 99.98% identical to plasmid pJM1, respectively (Table S3).
isochorismatase (entB), isochorismate synthase (entC), 2,3-dihydroxybenzoate-AMP ligase (entE), histidine decarboxylase (hdc), and genes that encode for the efflux ABC transporter, permease/ATP-binding protein, transport ATP-binding protein CydCD, and nonribosomal peptide synthetase modules, siderophore biosynthesis (Supplementary File S2). The genes associated with virulence and adaptation identified in both plasmids are related to iron acquisition and metabolism, siderophore biosynthesis and transport, and toxins and superantigens (Table S2). Verification of the presence of the plasmids in V. anguillarum J382 revealed a single DNA band around 23 Mda (Figure 3c), which could be due to the small size difference between the plasmids. A blastn search in the NCBI database revealed that both plasmids, pVA_O1_1 and pVA_O1_2, are 100% and 99.98% identical to plasmid pJM1, respectively (Table S3).   The average nucleotide identities between the V. anguillarum J382 chromosomes 1 and 2, and chromosomes from other V. anguillarum strains, revealed high ANI values (99-100%), especially with serotype O1 strains. The ANI values and GC% among the V. anguillarum strains were conserved regardless of the fish host and spatial distribution (Table 1). Both chromosomes 1 and 2 of V. anguillarum J382 are very similar to chromosomes 1 and 2 from the serotype O1 strains V. anguillarum 775, V. anguillarum M3, V. anguillarum 87-9-116, V. anguillarum 425, and V. anguillarum ATCC-68554 (Figure 4a,b).

Average Nucleotide Identities (ANI) and Phylogeny
The average nucleotide identities between the V. anguillarum J382 chromosomes 1 and 2, and chromosomes from other V. anguillarum strains, revealed high ANI values (99-100%), especially with serotype O1 strains. The ANI values and GC% among the V. anguillarum strains were conserved regardless of the fish host and spatial distribution (Table 1). Both chromosomes 1 and 2 of V. anguillarum J382 are very similar to chromosomes 1 and 2 from the serotype O1 strains V. anguillarum 775, V. anguillarum M3, V. anguillarum 87-9-116, V. anguillarum 425, and V. anguillarum ATCC-68554 (Figure 4a,b). The phylogeny was inferred using the neighbor-joining method, the Kimura method was used for computing the evolutionary distance, and 1000 bootstrap replicates were used. P. damselae subsp. piscicida was used as the outgroup.
The phylogenetic relationships provided insights on the closest neighbors of the V. anguillarum J382 chromosomes 1 and 2. The neighbor-joining method revealed that The phylogeny was inferred using the neighbor-joining method, the Kimura method was used for computing the evolutionary distance, and 1000 bootstrap replicates were used. P. damselae subsp. piscicida was used as the outgroup.
The phylogenetic relationships provided insights on the closest neighbors of the V. anguillarum J382 chromosomes 1 and 2. The neighbor-joining method revealed that chromosome 1 from whole genomes of V. anguillarum are clustered into four groups (Figure 4c). Cluster I consists of serotype O1, cluster II consists of serotype O2, cluster III consists of serotype O3, and cluster IV consists of a strain of unknown serotype. Phylogenetic analysis of V. anguillarum J382 chromosome 2 revealed five clusters (Figure 4d). Clusters I and III consisted of chromosomes 2 from serotype O1, whilst clusters II, IV, and V consisted of serotypes O2, O3, and unknown/other, respectively. Surprisingly, both chromosomes 1 and 2 of V. anguillarum VIB43, a strain that was earlier reported to be serotype O1, clustered together with serotype O2 strains, V. anguillarum J360 and VIB12. Both phylogenetic trees revealed that in the Vibrio genus, V. anguillarum is closely related to V. cholerae.

Synteny Analysis
Comparisons between the chromosomes of V. anguillarum J382 with those of conspecific isolates were performed to study synteny. Dot plot visualization of the serotype O1 chromosome 1 (V. anguillarum J382 vs. V. anguillarum 87-9-116) showed high similarity (Figure 5a), although there was genomic rearrangement of the locally collinear blocks ( Figure S4a). The visualization of V. anguillarum J382 chromosome 1 against V. anguillarum J360 chromosome 1 (serotype O2) (Figure 5b), and V. anguillarum J382 chromosome 1 against V. anguillarum CNEVA NB 11008 chromosome 1 (serotype O3) (Figure 5c), revealed genomic gaps, inversions, and orthologs. This was supported by the locally collinear blocks arrangements that also showed genomic rearrangements ( Figure S4b,c). Similar results were observed when synteny was studied for V. anguillarum J382 chromosome 2 (Figures 5d-f and S4D-F). chromosome 1 from whole genomes of V. anguillarum are clustered into four groups (Figure 4c). Cluster I consists of serotype O1, cluster II consists of serotype O2, cluster III consists of serotype O3, and cluster IV consists of a strain of unknown serotype. Phylogenetic analysis of V. anguillarum J382 chromosome 2 revealed five clusters ( Figure  4d). Clusters I and III consisted of chromosomes 2 from serotype O1, whilst clusters II, IV, and V consisted of serotypes O2, O3, and unknown/other, respectively. Surprisingly, both chromosomes 1 and 2 of V. anguillarum VIB43, a strain that was earlier reported to be serotype O1, clustered together with serotype O2 strains, V. anguillarum J360 and VIB12. Both phylogenetic trees revealed that in the Vibrio genus, V. anguillarum is closely related to V. cholerae.
Only one oligosaccharide is common to O1, O2, and O3 serotypes. One oligosaccharide is also shared between O1 and O2, or between O1 and O3, and two oligosaccharides are shared between O2 and O3 (Figure 7a).
Lipopolysaccharide extracted from isolates, V. anguillarum J382 (O1) and V. anguillarum J360 (O2) displayed distinct LPS profiles between the two serotypes ( Figure  7b). Several LPS bands were missing in serotype O2 in the O-antigen protein region. Also, there were differences in the core/lipid A protein band in the band sizes, as well as the number of bands.

Characteristics of ncRNAs in V. anguillarum Serotypes O1 and O2
A total of 122 ncRNA were annotated in V. anguillarum J382, of which 89 ncRNAs and 33 ncRNAs were in chromosome 1 and chromosome 2, respectively (Supplementary file 2). There were multiple copies of the ncRNAs CsrB_RF00018, CsrC_RF00084, Thr_leader_RF00506, Leu_leader_RF00512, MicX_RF01808, RsaJ_RF01822, and The genes galE, manA, manB, wecB, wecC, glmU, wbp, and the oligosaccharide UDP-Nacetyl-α-D-mannosaminuronate, are common among the three serotypes (Figure 7a and Table S4). Interestingly, the wecA gene is only found in serotype O1, and wbpO is unique to both serotypes O2 and O3 (Table S4). In the three serotypes studied, two genes (galE and manB) are found in chromosome 2 and the rest of the studied genes in chromosome 1 (Table S4).
Only one oligosaccharide is common to O1, O2, and O3 serotypes. One oligosaccharide is also shared between O1 and O2, or between O1 and O3, and two oligosaccharides are shared between O2 and O3 (Figure 7a).
Lipopolysaccharide extracted from isolates, V. anguillarum J382 (O1) and V. anguillarum J360 (O2) displayed distinct LPS profiles between the two serotypes ( Figure 7b). Several LPS bands were missing in serotype O2 in the O-antigen protein region. Also, there were differences in the core/lipid A protein band in the band sizes, as well as the number of bands.
The comparison of annotated ncRNAs between serotype O1 and O2 yielded 47 common ncRNAs, while 38 were predicted only in serotype O1 (J382) and 14 only in serotype O2 ( Figure S5). The comparison revealed that there are ncRNAs predicted only in one serotype and absent in another (Figure 8). The ncRNAs that were predicted only in serotype O1 are linked to functions, such as pathogenesis, RNA modification, and the bacterial signal recognition particle RNA (Figure 8a). Both serotypes have ncRNAs that originate from plasmids and viruses, but are not commonly shared (Figure 8a,b). Commonly, ncRNAs linked to functions, such as Hfq binding, protein synthesis, metabolism, virulence, iron regulation, stress response, quorum sensing, riboswitches, ribozymes, and transport, are found in both genomes of the two serotypes (Figure 8c).
The comparison of annotated ncRNAs between serotype O1 and O2 yielded 47 common ncRNAs, while 38 were predicted only in serotype O1 (J382) and 14 only in serotype O2 ( Figure S5). The comparison revealed that there are ncRNAs predicted only in one serotype and absent in another (Figure 8). The ncRNAs that were predicted only in serotype O1 are linked to functions, such as pathogenesis, RNA modification, and the bacterial signal recognition particle RNA (Figure 8a). Both serotypes have ncRNAs that originate from plasmids and viruses, but are not commonly shared (Figure 8a,b). Commonly, ncRNAs linked to functions, such as Hfq binding, protein synthesis, metabolism, virulence, iron regulation, stress response, quorum sensing, riboswitches, ribozymes, and transport, are found in both genomes of the two serotypes (Figure 8c).

Discussion
Infections with V. anguillarum in fish cause significant losses in the marine environment and aquaculture, although vaccines have reduced the risk of vibriosis in many of these cultured species [2]. The repertoire of aquatic animals that are susceptible to V. anguillarum includes important species, such as rainbow trout, salmon, turbot, sea bream, sea bass, ayu, cod, and eel [2]; cleaner fish such as lumpfish and wrasse [52]; ornamental fish [53]; and cultured bivalves [54]. V. anguillarum infections also have socio-economic impacts on the aquaculture industry, whereby the mortalities of aquatic animal species can reach up to 100%. For instance, annual economic losses due to V. anguillarum have been estimated to be between USD 18-30 million in Japan [54]. Herein, we characterized a strain of V. anguillarum serotype O1, isolated from winter steelhead trout in 1999. Although V. anguillarum is commonly reported in rainbow trout, this strain was isolated from the first reported vibriosis in winter steelhead [22]. The virulence of strain J382 was confirmed in lumpfish, where there was 100% mortality at 9 dpi in lumpfish ( Figure 1).
The V. anguillarum serotype O1 J382 showed a typical V. anguillarum phenotype and genome. It was earlier reported that most V. anguillarum serovar O1 exhibited resistance to colistin, but were sensitive to ampicillin and cephalothin, whilst the opposite was observed for most serovar 02 strains [27]. The V. anguillarum O1 (J382) strain was particularly resistant to colistin sulphate, which was contrary to the V. anguillarum serotype O2 (J360) isolated from lumpfish [5]. Previously, it has been suggested that some intrinsic antimicrobial resistance in the V. anguillarum serovar is dependent and the genes responsible are chromosomally encoded [27]. This was supported by the genomic analysis of V. anguillarum J382, where all the antibiotic-resistant genes were located on either chromosome 1 or chromosome 2 (Table S2).
Phenotypically, members of V. anguillarum are eurythermal having an optimal growth temperature of between 30-34 • C, with limited growth below 5 • C, but grow rapidly between 25-30 • C [55,56]. V. anguillarum have a halophilic metabolism and thrive at NaCl concentrations between 1% and 2%, allowing for survival in seawater for over 50 months. Some strains have a plasmid-mediated iron-sequestering system [2,55]. The characteristics exhibited by V. anguillarum J382 were in accordance with those mentioned above. Our study also verified that the virulence of V. anguillarum is dually modulated by temperature and iron availability. For instance, hemolysin activity was noted at 28 • C on sheep blood agar, whilst there was limited activity at 15 • C, and the iron-sequestering ability of V. anguillarum O1 was demonstrated under iron-limited conditions ( Figure S1).
Previous transcriptomic analysis of the V. anguillarum serotype O2 revealed that under iron starvation some virulence factors like LPS biosynthesis, heme receptor HuvA, siderophore piscibactin, MARTX toxins, and T6SS2 were upregulated at 15 • C, while hemolysin Vah1, T6SS1, ferrous iron transport, and the vanchrobactin siderophore system were upregulated at 25 • C [56]. Chemotaxis and flagellum-related genes were downregulated under iron starvation and at 15 • C. This iron and temperature dependent behavior of V. anguillarum needs to be considered when producing vaccines, as the vaccine might lack some important antigens thereby affecting the protection [56].
The genotypic characterization of V. anguillarum J382 revealed the presence of two chromosomes and two pJM1-like plasmids. The presence of two chromosomes is a general characteristic of Vibrio spp. that evolved as a survival strategy, allowing the rapid adaptation of the pathogen to different environments, such as seawater and different hosts [2]. The smaller chromosome 2 is thought to have been a megaplasmid that was taken up by an ancestral Vibrio spp., since its parA gene tends to cluster with homologues from plasmids and not from chromosomes [57], contains an integron region found in plasmids, contains only four tRNA genes, and no rRNA gene is present (Table S1). Chromosome 2 could have acquired genes from other species, as well as from chromosome 1, since most of its genes are identical copies [57]. On the other hand, the two pJM1-like plasmids found in V. anguillarum J382 are characteristic of serotype O1, which is absent in serotype O2 and encode an iron-sequestering system that is siderophore-dependent (siderophore anguibactin) [2].
The iron-sequestering system can efficiently dislocate host protein-bound iron making it available for uptake by the bacteria [54]. The iron-sequestering role is shared between the pJM1-like plasmids and chromosome 1 (Table S2) in serotype O1 strains. Another iron-uptake system found in O1 and O2 V. anguillarum serovars, known as the siderophore vanchrobactin, is independent of the pJM1 plasmid [54].
The average nucleotide identity (ANI), phylogeny, and synteny of V. anguillarum J382, with close and distant related species, was carried out. The V. anguillarum J382 chromosomes consistently showed similarity to the other serotypes O1 genomes, whilst the identity decreased when compared to serotype O2, O3, and an unspecified serotype strain. Whole genome ANI is a bioinformatic approach corresponding to DNA-DNA hybridization, which is a simple and useful method for the purpose of describing genetic relatedness that can effectively delineate bacterial species based on lineage-specific genes [57]. Evidence of evolution within the V. anguillarum species, especially among serotypes O1, O2, and O3, was displayed by the presence of genomic rearrangements, deletions, and inversions, as well as the presence of orthologs. However, the synteny of locally colinear blocks, among the three serotypes, revealed the conservation of functionally related genes. Colinear syntenic blocks are defined as conserved groups of genes across taxa that are consistently encoded in the same locality for gene expression regulation in prokaryotes [58,59]. The whole chromosome phylogenetic analysis of V. anguillarum confirmed earlier reports whereby the three serotypes were distinctly separated [5].
The genes encoding for O-antigen biosynthesis arranged in a gene cluster are generally located on the chromosomes and are separated into three groups, namely, nucleotide sugar biosynthesis, sugar transfer, and O-antigen processing [15]. The predicted V. anguillarum O-antigen gene clusters displayed variations among serotypes O1, O2, and O3. Evidence of inter-species lateral transfer of the O-antigen gene clusters exists and is partly attributed to O-antigen diversity in Gram-negative bacteria [15]. Common in most bacteria, sugars or sugar derivatives that comprise the O-antigen components are transferred from nucleotide sugar precursors [15]. In the present study, nucleotide sugar precursors uridine 5 -diphosphate-glucose (UDP-Glc), uridine diphosphate galactose (UDP-Gal), and uridine diphosphate N-acetylglucosamine (UDP-GlcNAc) that are common among Gram-negative bacteria [60] were also predicted in the three V. anguillarum serotypes. On the other hand, the UDP-sugar precursor synthesizing gene, ugd and wecA gene [15] were only identified in serotype O1.
The genomic characterization of V. anguillarum serotype O2 supported earlier findings of the LPS analysis that was conducted using nuclear magnetic resonance techniques and partial hydrolysis [16]. In this study, V. anguillarum J360 used as a serotype O2 representative strain, was shown to include only UDP-sugar biosynthesis pathways for the production of N-acetyl-α-D-mannosaminuronate, 2,4-diacetamido-2,4,6-trideoxy-u-glucose (N,N'-diacetylbacillosamine), 2-acetamido-3-amino-2,3-dideoxy-D-glucuronamide, and 2,3-diacetamido-2,3-dideoxy-d-mannuronic acid. These polysaccharides were earlier identified in serotype O2, together with an N-formyl-L-alanyl group [16]. For serotype O3, one dTDP-sugar biosynthesis and three UDP-sugar biosynthesis pathways were identified by bioinformatic analysis of the representative whole genome (CNEVA NB 11008). The wbpO/tviB gene unique to serotypes O2 and O3 is a key enzyme in the B-band LPS production of P. aeruginosa O6 [67]. This finding may suggest the presence of the LPS B-band in the two serotypes used. However, further LPS characterization experiments with these serotypes are required to elucidate this genetic finding. Furthermore, the absence of the rfbH and rfbI required for the Wzm/Wzt export system pathway suggests that both serotypes utilize alternative pathways for O-antigen translocation and polymerization. We, however, did not recognize any genes homologous to the genes encoding for Wzx and Wzy utilized by the Wzx/Wzy pathway, or the genes encoding WbbE and WbbF utilized by the synthase pathway in the O-antigen gene clusters [68]. The responsible genes might be located elsewhere on the chromosome as was found in Salmonella enterica groups D1 and B, where the wzy gene was found outside the O-antigen gene cluster [69].
Bacterial non-coding RNAs (ncRNA) are known to carry out various functions that range from genetic information processing, physiological adaptation, and metabolism [70,71]. The discovery of bacterial ncRNAs has primarily been conducted using computational searching using genomic and metagenomic sequence data [69]. Bacterial ncRNAs were first found within intergenic regions of E. coli, with high conservation among closely related species [72]. Herein, we annotated ncRNAs in the serotype O1 V. anguillarum J382 genome using computational tools. To check for the conservancy of the ncRNAs, a comparison was carried out between V. anguillarum J382 and V. anguillarum J360, a serotype O2 strain. A total of 122 ncRNAs were predicted in V. anguillarum J382, whose functions were linked to physiological processes, such as virulence, quorum sensing, and stress responses.
The comparison of ncRNAs in both V. anguillarum serotypes revealed that several of the common ncRNAs were linked to pathogenesis of the bacteria to the host. These include several Hfq binding ncRNAs, such as Qrr_RF00378. The Hfq chaperone in conjunction with qrr sRNAs in V. cholerae and related sRNAs in V. harveyi, was shown to destabilize the mRNA encoding the regulators LuxR and HapR in V. harveyi and V. cholerae, respectively, thereby suppressing quorum sensing [73]. Other ncRNAs involved in pathogenesis include the VrrA (VrrA_RF01456), which in V. cholerae, influence intestinal colonization and adhesion by regulating the expression of TcpA and OmpA, respectively, as well as and outer membrane vesicle (OMV) formation [74]. The ncRNAs like CsrB/CsrC (CsrB_RF00018/CsrC_RF00084), Spot 42 (Spot_42_RF00021), and RyhB (RyhB_RF00057) are also involved in the regulation of virulence factors, namely, expression, cytotoxicity, and iron metabolism, respectively, of Vibrio species [75]. Despite the limited attention given to bacterial microRNAs (miRNAs), they have been proven to exist in bacterial species, such as Streptococcus mutans, E. coli, and Thalassospira spp. [76], and in this study, miRNAs were also identified in both serotypes. Future research is required to elucidate the role of the miRNAs in bacteria.

Conclusions
V. anguillarum O1 J382, isolated from a natural infection in winter steelhead trout in the Pacific coast, was subjected to phenotypic and genomic characterization. V. anguillarum J382 isolate was highly virulent in Atlantic lumpfish, suggesting a broad host fish range. Also, J383 was colistin sulphate resistant, like other V. anguillarum O1 strains. Comparative genomics of the O-antigen gene clusters of serotypes O1, O2, and O3 revealed genetic variation in genes related to nucleotide sugar translocation mechanisms, O-antigen polymerization, and composition. Moreover, gene rearrangements between the serotypes were observed. The ANI, phylogeny, and synteny revealed the closest relatives of V. anguillarum J382, as well as genetic diversity among the different serotypes showing intra-species evolution. Some of the O-antigen biosynthesis pathways identified were earlier identified in distantly related Gram-negative bacteria, such as P. aeruginosa, indicating the inter-species lateral transfer of the O-antigen gene clusters. The ncRNAs were also annotated and they provided insight into the various putative functions they might carry out in V. anguillarum under physiological adaptation. This study contributes essential pathogenesis and comparative genomics data on virulent V. anguillarum serotypes, critical for further fundamental studies (e.g., gene function) and applications in fish health (e.g., vaccines).

Supplementary Materials:
The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/microorganisms11030792/s1, Figure S1: Phenotypic characterization tests conducted on V. anguillarum J382 isolated from winter Sseelhead trout in British Columbia, Canada. Figure S2: Characterization using API system of bacteria isolated from lumpfish organs (spleen, heart, liver, head kidney) infected with V. anguillarum J382 at 5dpi. Figure S3(a): Putative O-antigen biosynthesis pathways associated with V. anguillarum serotype O1 and the corresponding oligosaccharides (blue color). Figure S3 Figure S5: Comparison of ncRNAs in V. anguillarum serotypes O1 (J382) and O2 (J360). The ncRNAs in V. anguillarum J382 were predicted using the software StructRNAfinder and BEDTools: intersect v2.3, while J360 ncRNAs were retrieved from the NCBI. The ncRNA comparison was performed using jvenn software. Table S1: Annotation summary of V. anguillarum J382 using Rapid Annotation Subsystem Technology (RAST). Table S2: Genes associated with pathogenesis and environmental adaption in V. anguillarum J382 predicted by RAST and Patric 2.0. Table S3. Plasmids identical to V. anguillarum J382 plasmids pVA_O1_1 and pVA_O1_2, which were found after a blastn search in the NCBI database. Table S4