A Global Survey of Hypervirulent Aeromonas hydrophila (vAh) Identified vAh Strains in the Lower Mekong River Basin and Diverse Opportunistic Pathogens from Farmed Fish and Other Environmental Sources

This global survey of vAh brought together scientists that study fish disease to evaluate the evolution, geographical distribution, phylogeny, and hosts of vAh and other Aeromonas sp. isolates. In addition to vAh isolates from China and the United States, four new vAh isolates were isolated from the lower Mekong River basin in Cambodia and Vietnam, indicating the significant threat of vAh to modern aquaculture and the need for improved biosecurity to prevent vAh spread. ABSTRACT Hypervirulent Aeromonas hydrophila (vAh) has emerged as the etiologic agent of epidemic outbreaks of motile Aeromonas septicemia (MAS) in high-density aquaculture of farmed carp in China and catfish in the United States, which has caused millions of tons of lost fish. We conducted a global survey to better understand the evolution, geographical distribution, and phylogeny of vAh. Aeromonas isolates were isolated from fish that showed clinical symptoms of MAS, and pure cultures were screened for the ability to utilize myo-inositol as the sole carbon source. A total of 113 myo-inositol-utilizing bacterial strains were included in this study, including additional strains obtained from previously published culture collections. Based on a gyrB phylogeny, this collection included 66 A. hydrophila isolates, 48 of which were vAh. This collection also included five new vAh isolates from diseased Pangas catfish (Pangasius pangasius) and striped catfish (Pangasianodon hypophthalmus) obtained in Cambodia and Vietnam, respectively. Genome sequences were generated from representative vAh and non-vAh isolates to evaluate the potential for lateral genetic transfer of the myo-inositol catabolism pathway. Phylogenetic analyses of each of the nine genes required for myo-inositol utilization revealed the close affiliation of vAh strains regardless of geographic origin and suggested lateral genetic transfer of this catabolic pathway from an Enterobacter species. Prediction of virulence factors was conducted to determine differences between vAh and non-vAh strains in terms of virulence and secretion systems. Core genome phylogenetic analyses on vAh isolates and Aeromonas spp. disease isolates (55 in total) were conducted to evaluate the evolutionary relationships among vAh and other Aeromonas sp. isolates, which supported the clonal nature of vAh isolates. IMPORTANCE This global survey of vAh brought together scientists that study fish disease to evaluate the evolution, geographical distribution, phylogeny, and hosts of vAh and other Aeromonas sp. isolates. In addition to vAh isolates from China and the United States, four new vAh isolates were isolated from the lower Mekong River basin in Cambodia and Vietnam, indicating the significant threat of vAh to modern aquaculture and the need for improved biosecurity to prevent vAh spread.

to assess vAh dissemination among various farmed fish species in different regions of the world.
Despite the evidence that vAh strains are clonal and have recently spread from Asia to the United States, there are some genetic differences among vAh strains. In particular, while vAh isolates from carp species in China typically have a complete type VI secretion system (T6SS) (12,22), most vAh isolates from channel catfish in the United States, and especially from western Alabama, lack a complete T6SS and only carry hcp1, tssH, and vgrG1 (23). The carp vAh isolate NJ-35, which has a complete T6SS, has been found to express a phospholipase that contributes to biofilm formation and virulence in zebrafish (Danio rerio) (24). While lacking many T6SS-associated genes, the presence of hcp1 and vgrG1 have been found to contribute to vAh ML09-119 virulence (23), but the degree to which the T6SS plays a role in fish host specificity and virulence has yet to be defined. The evolution of vAh strains as they infect and replicate in different fish species is of significant interest. Our lack of knowledge regarding fish host range and geographic distribution also prompted us to conduct a vAh global survey. As a group of fish disease experts from around the world, we primarily sampled freshwater fish with disease symptoms characteristic of MAS and obtained pure bacterial cultures that were evaluated for growth on myo-inositol, a phenotype that has been consistent in vAh strains isolated from China and the United States. A phylogenetic analysis of myo-inositol-utilizing strains using gyrB sequences was conducted to further characterize disease isolates. Finally, for representative vAh and non-vAh strains, we conducted comparative genome analyses to provide further information on the phylogeny and predicted virulence factors of vAh strains. This study is a first step toward a better understanding of vAh worldwide distribution, uniting fish disease researchers in a network that can help track the distribution of vAh and developing methods to protect farmed fish against this emerging pathogen.

RESULTS AND DISCUSSION
Identification of myo-inositol-utilizing Aeromonas sp. strains. This global vAh survey relied upon an extensive network of microbiologists willing to participate in screening fish disease isolates and cryopreserved collections for the presence of myo-inositol utilizing A. hydrophila strains. There have been no previous reports of A. hydrophila strains with the ability to use myo-inositol as the sole carbon source other than vAh strains (i.e., ST251). Therefore, by screening bacterial isolates for growth on myo-inositol in a minimal medium, our goal was to rapidly and cost-effectively identify putative vAh strains from diverse locales. From this extensive survey, 43 myo-inositol-utilizing Aeromonas sp. strains were isolated from Pabda (Ompok pabda) from Bangladesh, Pangas catfish from Cambodia, lake water from Finland, Koi (Cyprinus rubrofuscus) from France, basa fish (Pangasius bocourti) from Malaysia, rainbow trout (Oncorhynchus mykiss) from Mexico, crab (Brachyura spp.) from Norway, trout (Oncorhynchus spp.) and human feces from Spain, and striped catfish from Vietnam (Table 1). Typical vAh strains cultured on tryptic soy agar (TSA) produce smooth, rounded, opaque colonies that have a light yellow color with a 2-to 3-mm diameter range after 24 h of incubation (25). The strains that showed a colony morphology consistent with vAh and evident growth on myo-inositol (i.e., increase in the optical density at 600 nm [OD 600 ] of .0.4 over 48 h) were further validated by molecular phylogenetic analyses (26,27).
The 43 myo-inositol-utilizing Aeromonas strains collected worldwide were subjected to vAh-specific and/or gyrB-targeted PCR using the primer sets listed in Table 2. A phylogenetic analysis was conducted using gyrB sequences from these isolates in addition to Aeromonas sp. type strains and previously described vAh strains from China and the United States (Fig. 1). The phylogeny of these strains revealed a great diversity of Aeromonas spp. that were obtained in this survey, including A. bestiarum, A. bivalvium, A. caviae, A. dhakensis, A. finlandensis, A. media, A. salmonicida, A. sobria, and A. veronii. Interestingly, some of these Aeromonas spp. had not been previously shown to have the ability to use myoinositol as a carbon source, including A. bestiarum, A. bivalvium, A. caviae, A. dhakensis, A. media, and A. veronii (26,(28)(29)(30). While these Aeromonas species were not the target of this survey, this adds to our knowledge of the use of myo-inositol among diverse Aeromonas species. Additionally, it suggests that this ability may contribute to the persistence of these bacteria in aquatic habitats and the virulence of these opportunistic pathogens in diverse warm-water fish species. Based on the gyrB phylogeny, A. hydrophila strains isolated from Spain, Mexico, Finland, Cambodia, and Vietnam grouped together and formed well-supported clades. Furthermore, the gyrB phylogeny indicated that all previously described vAh strains (i.e., ST251) grouped together within a monophyletic clade with bootstrap support, clearly distinct from other myo- inositol utilizing Aeromonas sp. strains (Fig. 1). Interestingly, the vAh clade included strains from Cambodia (CPF2-S1) and Vietnam (DT-TKT-2020-677, DT-TKT-2020-680, DT-TKT-2020-681, and DT-TTD-2020-734). The Cambodian vAh strain CPF2-S1 was one of five myo-inositol utilizing bacterial isolates that were positive for vAh-specific PCR and isolated from Pangas catfish in the Mekong River basin. The four vAh isolates from Vietnam were all obtained from diseased striped catfish in the Mekong River delta, and three of them (DT-TKT-2020-677, DT-TKT-2020-680, DT-TKT-2020-681) are closely related to a recently reported vAh strain, DT-TTD-2020-734, obtained from striped catfish in the Mekong River delta (31). These newly described vAh strains indicate that additional fish species are susceptible to vAh and that the Mekong River basin is an active region of vAh disease transmission. Inositol catabolism phylogeny. The evolutionary history of the inositol catabolism pathway among myo-inositol-utilizing Aeromonas spp. was inferred based on the amino acid sequences of IolA, IolC, IolD, IolE, IolG, InoE, InoF, and InoL, which were obtained from representative vAh and other Aeromonas sp. genomes (Fig. S1A to H). The variability in inositol gene content among these strains precluded a concatenated phylogenetic analysis. Among vAh strains, there were no differences observed in the evolutionary history of gene products required for myo-inositol transport (InoE, InoF, and InoL) or catabolism (IolA, IolC, IolD, IolE, and IolG), with all vAh strains present in the same monophyletic clade and having strong bootstrap support (Fig. S1A to H). This is consistent with the observation that vAh strains are clonal, including the newly isolated vAh strains from Cambodia (CPF2-S1) and Vietnam (e.g. DT-TTD-2020-734). Another consistent observation was that the inositol-related gene products from vAh strains share a close relationship with orthologous sequences from Enterobacter cloacae, which has been hypothesized to be the origin of the inositol catabolism pathway present in vAh strains (27). In contrast, the sequences obtained from other Aeromonas species were distantly related to those from vAh strains and E. cloacae, including A. dhakensis 1P11S3, A. dhakensis P3I3, A. dhakensis P3L3, A. media R100, A. sobria ESV-355, and A. sobria ESV-393. The evolutionary history of inositol utilization among Aeromonas sp. therefore appears to be complex, with horizontal gene transfer of inositol transport and catabolism postulated to play an important role. This survey revealed a large diversity of other Aeromonas species that can utilize myo-inositol. Future research should explore the role of myo-inositol utilization in the persistence and virulence of opportunistic Aeromonas sp. pathogens.
The role of myo-inositol utilization in vAh persistence and virulence should also be further explored. Channel catfish have been shown to synthesize myo-inositol in brain, kidney, and liver tissues, and soy-based fish feed containing a high concentration of phytic acid (inositol hexaphosphate) (32,33). The inositol derived from fish tissues and dietary sources may provide both a carbon source and an environmental signal that induces expression of vAh virulence factors. The transcriptional regulator IolR is responsible for the regulation of iol genes as well as other virulence factors in bacterial pathogens, such as Salmonella enterica (34,35). IolR has also been found to regulate autoaggregation and biofilm formation in the vAh strain NJ-35 (36). Furthermore, the presence of myo-inositol that accumulates in sediment from fish feed may help vAh to persist within the environment.  Average nucleotide identity (ANI). The pairwise ANI comparisons for 63 Aeromonas sp. genomes, including representative vAh strains from China, Cambodia, the United States, and Vietnam showed high ANI values (.99%) for all vAh strains (Fig. 2), which was consistent with the previous core genome-based phylogeny indicating the clonality of all known vAh strains (12). In contrast, only a few non-vAh A. hydrophila strains showed high ANI values compared with vAh strains, and most ANI values ranged from 96% to 97%. The exceptions to this were strains that had been putatively indicated as A. media, A. sobria, and A. veronii based on phylogenetic analyses, all of which had discrepancies between the species affiliation indicated by ANI values and their species affiliation indicated in GenBank as previously described (37). Based on these ANI data and a core genome-based phylogeny, the phylogenetic affiliations of several strains were revised (see below and Table 1). This survey also included diverse Aeromonas sp. isolated from diseased fish and other environments, as revealed by the A. hydrophila-Aeromonas sp. pairwise ANI comparisons that ranged from 67% to 93%.
Aeromonas core genome phylogenetic analysis. The phylogenetic relationships among the representative vAh and diverse Aeromonas sp. strains included in this survey were inferred based on a set of core genome sequences totaling 3.8 Mb (Fig. 3). A subset of vAh strains was included in the core genome phylogeny due to some of the strains lacking high-quality genome sequences (e.g., Vietnamese vAh strains DT-TKT-2020-681 DT-TKT-2020-677, DT-TKT-2020-680). Consistent with the gyrB phylogeny, the Aeromonas core genome phylogeny indicated that all vAh strains, including the newly identified strains from Cambodia and Vietnam, form a monophyletic clade with strong bootstrap support that is distinct from other A. hydrophila or other Aeromonas sp. strains (Fig. 3). While the clonal vAh clade showed little variation among its members for the core genome phylogeny, there was significant intraspecies genetic variability observed among the other Aeromonas sp. Strains, including within A. hydrophila, A. dhakensis, A. media, and A. sobria. Based on this core genome phylogeny (and ANI values), there were many bacterial isolates described as A. hydrophila that were affiliated with A. dhakensis, A. media, or A. sobria, and these revised phylogenetic affiliations have been indicated (Table 1). In this analysis, the exclusion of small fragments was set to 10 kbp because these fragments were found to be flanked by highly repetitive sequences, which were previously demonstrated to contribute less to the production of core genomes. This removal was chosen as a blanket approach to increase computational efficiency and decrease the noise generated from repetitive sequences, as this study is solely based on sequence-based comparisons. However, with the growing body of knowledge that shows repetitive regions as significant in regulation, future studies should focus on these noncoding regions. Virulence factors encoded in Aeromonas sp. genomes. Representative vAh and non-vAh genomes were evaluated for their encoded potential to secrete virulence factors (Fig. 4). In agreement with previous studies, vAh strains were universally found to encode complete type 2 secretion systems, which have been found to be essential to the virulence of a vAh strain isolated from a channel catfish in the United States (38). In contrast, type 3 secretion systems were only identified in non-vAh strains. Interestingly, the type 6 secretion systems (T6SS) were complete only in a subset of vAh strains as has been previously described (23). Most of the vAh isolates from China, with the one exception of strain GYK1, were predicted to possess the complete T6SS, which has been shown to contribute to biofilm formation and virulence in fish (24). The two new vAh isolates from Cambodia and Vietnam (CPF2-S1, DT-TTD-2020-734) possessed the entire T6SS, which further demonstrates their close relationship to vAh strains isolated from carp in China. In contrast, many of the vAh strains isolated from channel catfish in the United States lacked a complete T6SS, with the notable exception of S14-452 and other strains isolated from the Mississippi delta (23). VAh core genome phylogenetic analysis. The phylogenetic relationships among the representative vAh strains included in this survey were inferred based on a set of core genome sequences present in all sequenced vAh strains (Fig. 5). The vAh core genome phylogeny  Aeromonas genomes were annotated using RAST and submitted to MacSyFinder for secretion system analysis. Maximum independent E value and minimal profile coverage were set as the default, while the maximum E value was set as 1.0. Virulence factors include type 1 secretion system (T1SS), type 2 secretion system (T2SS), type 4 pili (T4P), tight adherence system (TAD), type 3 secretion system (T3SS), flagellum, a phylogenetic subtype of type 6 secretion system (T6SSi), and type 9 secretion system (T9SS).

Global Survey of Hypervirulent Aeromonas hydrophila
Microbiology Spectrum indicated that the newly identified strains obtained from diseased fish in Cambodia and Vietnam form a monophyletic clade with strong bootstrap support with vAh strains isolated from crucian carp (Carassius carassius) and mandarin fish (Siniperca chuatsi) in China. Moreover, these vAh isolates from Cambodia and Vietnam share a close relationship, indicating that they originated from a common ancestor. In contrast, two other strains isolated from carp in China, ZC1 and JBN2301, form a well-supported clade with vAh strains isolated from catfish in the United States (4). The successful isolation of vAh from farmed Pangas catfish in Cambodia and from farmed striped catfish in Vietnam broadens the knowledge of the geographical distribution of vAh and the fish species in which this emerging pathogen can cause disease. Due to the rapid growth of the live fish trade in Asia and beyond, this pathogen could be transmitted to more countries and infect more fish species without sufficient biosafety (39). This calls for future development of rapid and inexpensive diagnostic assays to identify vAh strains and aid in biosecurity precautions to prevent further dissemination of this virulent pathogen.

MATERIALS AND METHODS
Bacterial isolates. Fish that demonstrated the typical symptoms of MAS, especially with external hemorrhaging and in farms experiencing high fish mortality, were collected for diagnosis and autopsy at the local institution. Aeromonas sp. isolates were recovered from diseased fish from aquaculture ponds in Bangladesh, Cambodia, Finland, France, Ghana, Malaysia, Mexico, Norway, Pakistan, Spain, Thailand, and Vietnam ( Table 1). The fish species sampled were tilapia (Oreochromis niloticus), striped catfish, pabda, Pangas catfish, basa catfish, rainbow trout, carp (Cyprinidae spp.), and perch (Perca spp.), while in some cases isolates were obtained from other environmental samples such as lake water, crab, seafood, and human feces (Table 1). Organs with the highest concentration of vAh, including liver, spleen, and kidney, were used to inoculate tryptic soy agar (TSA) plates (Beckton Dickinson, New Jersey, USA) or other bacteriologic growth medium appropriate for A. hydrophila cultivation, and these cultures were incubated at 30°C for 24 to 48 h. The vAh strain ML09-119 served as a control for comparison.
Single colonies that showed A. hydrophila morphology were cultured on TSA (30°C, 24 h) to obtain isolated colonies. Three colonies of each strain were cultured separately in 2 mL of M9 broth medium supplemented with 5.5 mM myo-inositol as previously described (27). The vAh strain ML09-119 and the non-vAh strain AL06-06 served as positive and negative controls, respectively. Cultures were grown at 30°C for 48 h to record their growth as measured by the optical density at 600 nm (OD 600 ). The utilization of myo-inositol of an unknown isolate was monitored by turbidity and CFU counts (as previously described). An increase in turbidity (change in OD 600 of .0.4) was observed for myo-inositol-utilizing strains over 48 h. Pure cultures of myo-inositolutilizing strains were subsequently identified as vAh by phylogenetic analysis of gyrB sequences following previously described methods (12,40). Validated vAh strains were cryopreserved in tryptic soy broth (TSB) containing 20% glycerol at 280°C.
Phylogenetic analysis based on gyrB from myo-inositol-utilizing strains. Genomic DNA of the myo-inositol-utilizing isolates was isolated using the E.Z.N.A. bacterial DNA isolation kit according to the manufacturer's protocol (Omega Bio-Tek, Norcross, GA, USA). Bacterial DNA was quantified with a NanoDrop instrument (Thermo Fisher Scientific, Waltham, MA, USA) and used as a template for PCR amplify gyrB gene sequences using Aeromonas genus-level primer sets (Table 2) (41). To avoid the potential off-target priming and increase PCR specificity (42), touchdown PCR was conducted to generate gyrB products and performed on a Mastercycler Nexus thermo cycler (Eppendorf, Hamburg, Germany) with 50 ng of genomic DNA (gDNA) isolated from each strain, 25 mL of EconoTaq Plus green 2X master mix (Lucigen Corp., Middleton, WI, USA), and 0.5 mL of 20 mM reverse and forward primers. The thermal cycling parameters were 94°C for 3 min, 10 cycles of 94°C for 30 sec, 68°C for 30 sec (21°C per cycle), and 72°C for 1 min, and then 25 cycles of 94°C for 30 sec, 58°C for 30 sec, and 72°C for 1 min and a final extension at 72°C for 5 min.
The gyrB gene amplicons were Sanger sequenced as described previously (27,40) and assembled into consensus sequences using CLC Genomics Workbench (Qiagen, Inc., Aarhus, Denmark). The gyrB sequence reads were trimmed for quality, assembled into consensus sequences, and aligned with an existing gyrB sequence database obtained from Chinese and U.S. vAh and non-vAh strains (12), using ClustalW in MEGA X (43). The gyrB sequence database included sequences varying from 422 to 1,068 bp and included Aeromonas sp. type strains to confirm species affiliations. Phylogenetic relationships of the inositol-utilizing Aeromonas sp. strains and appropriate type strains were determined by the construction of a phylogenetic tree using MEGA X (43). In total, 100 strains were included in the tree, including 5 new vAh strains and 38 new non-vAh strains. Among the 66 A. hydrophila strains, some were removed due to poor sequence quality and/or the lack of an available viable culture from which to recover a better-quality gyrB sequence. The evolutionary history of the strains in the gyrB database was inferred using the maximum likelihood (ML) method (44). The ML analysis was conducted with 1,000 iterations for bootstrap support, with bootstrap values shown on each branch of the gyrB tree as a circle proportional to bootstrap support. The gyrB tree was annotated and visualized using iTOL v6 (45). The gyrB tree was rooted using the A. sobria type strain.
Genome sequencing based on the gyrB phylogeny. Representative isolates of different inositolutilizing Aeromonas lineages were selected for Illumina sequencing based on the results of the gyrB phylogenetic tree. The sequenced strains were selected to represent vAh and inositol-utilizing Aeromonas from multiple geographical locations, including Spain, Mexico, Cambodia, Vietnam, Bangladesh, and Malaysia. The fragment libraries were constructed using a Nextera XT DNA library prep kit (Illumina, San Diego, CA, USA) based on the manufacturer's protocol, followed by paired-end sequencing conducted on an Illumina MiSeq platform (46). Sequence reads were imported into CLC Genomics Workbench, which was used to trim sequences for quality, followed by de novo assembly using default settings. Draft genome contig sequences were generated for strains CPF2-S1 (Cambodia), 14  Inositol catabolism phylogeny. A genomic database was generated that included draft genome sequences of strains isolated in Mexico, Spain, Cambodia, Bangladesh, Malaysia, and Vietnam that were supplemented with genome sequences of vAh strains from China and the United States, as well as from Enterobacter spp. which are predicted to be the origin of the inositol catabolism pathway present in vAh strains (22,47). A total of 15 open reading frames (ORFs) were identified in the myo-inositol catabolism pathway in vAh strains, from which we selected proteins shown to be required for inositol utilization (IolA, IolC, IolD, IolE, and IolG) for phylogenetic analysis (48). The proteins predicted to be involved in myo-inositol transport, InoE, InoF, and InoL, were also included in the phylogenetic analyses (49). Given the unique evolutionary histories associated with these proteins involved in inositol catabolism, a separate phylogenetic analysis was conducted for each of these amino acid sequences. An ML tree with 1,000 iterations for bootstrap support was conducted on MEGA X as described above and visualized using ITOL v6.
Average nucleotide identity. To evaluate the overall genetic similarity between vAh, non-vAh, and myo-inositol-utilizing Aeromonas spp., the ANI values of the 63 Aeromonas genomes were compared using JSpeciesWS (50) and visualized with Daniel's XL Toolbox v7.3.4 (51). According to the criteria for taxonomic affiliation of new genomes, an ANI value of .95% indicated that two strains belong to the same species (52,53).
Aeromonas core genome analysis. Both noncoding and coding sequences of vAh and non-vAh strains, including A. sobria, A. media, A. dhakensis, and A. salmonicida strains, were used for a core genome phylogenetic analysis. In general, small fragments do not influence the overall quality of core genomes due to these small genomic regions being flanked by highly repetitive sequences. Removing small fragments helped to improve overall accuracy by decreasing the noise generated from repetitive sequences; therefore, any contigs less than 10 kb were filtered from Aeromonas genomes by limiting them from mapping to multiple regions. FASTA files of the filtered sequences were submitted to Mugsy v1.2.3, a multiple whole genome alignment tool, using default parameters (54). The alignment was processed by GBLOCK v0.91b for the identification of highly conserved regions across all Aeromonas spp. strains as previously described (55). The parameters used in GBLOCK for retention were dependent upon the input alignment as previously described (12). Briefly, a maximum of 8 contiguous regions, a minimum of 30 sequences for conserved regions, and 51 sequences for flanked positions were used for the gapped positions within a block. A ML tree based on the final alignment was generated using MEGAX with default parameters for the 60 Aeromonas spp. strains, including 28 vAh strains. The ML tree was further visualized using iTOL v6.
VAh core genome phylogenetic analysis. To assess the phylogeny of vAh strains isolated from Cambodia, China, the US, and Vietnam, a vAh core genome was created using both coding and noncoding sequences of representative vAh. Contigs less than 10 kb in size were removed to increase computational efficiency, and filtered data were submitted as FASTA files to the multiple whole-genome alignment tool Mugsy v1.2.3 (54), under default parameters. The resulting alignment was subsequently processed with GBLOCK v0.91b (55) in order to identify regions of high conservation across all isolates. Parameters for retention by GBLOCK are dictated by the input alignment and were the following: a minimum of 31 and 51 sequences for conserved and flanked positions, respectively, a maximum of 8 contiguous, but nonconserved positions, a minimum block length of 10, and one-half of the sequences allowed to possess gapped positions within a block. From the final alignment, a maximum likelihood (ML) phylogeny for the vAh isolates was inferred using RAxML v8.2.8 (56) under the general time reversible model of evolution with estimated proportions of invariable sites and rate variation among sites (i.e., GTR 1 I 1 G) and 1,000 bootstraps to determine branch supports, as described previously (22). Trees were visualized using iTOL v6.
Virulence factor prediction. Virulence factor prediction and identification of secretion systems followed previously described methods (23,57). Briefly, the secretion systems of Aeromonas strains were identified with the program MacSyFinder. The data set option was set as "unordered" to evaluate the draft genome of each strain. The minimal profile coverage was set to 0.5, the maximum E value was set to 1.0, and the maximum independent E value was 0.001. Secretion systems of vAh, non-vAh, and myo-inositolutilizing Aeromonas strains were identified and indicated with mandatory and accessory genes, and the corresponding copy numbers were determined.

SUPPLEMENTAL MATERIAL
Supplemental material is available online only. SUPPLEMENTAL FILE 1, PDF file, 2.1 MB.