Free access
Spotlight Selection
Research Article
24 June 2014

Functional Genomic Characterization of Virulence Factors from Necrotizing Fasciitis-Causing Strains of Aeromonas hydrophila

This article has a companion.


The genomes of 10 Aeromonas isolates identified and designated Aeromonas hydrophila WI, Riv3, and NF1 to NF4; A. dhakensis SSU; A. jandaei Riv2; and A. caviae NM22 and NM33 were sequenced and annotated. Isolates NF1 to NF4 were from a patient with necrotizing fasciitis (NF). Two environmental isolates (Riv2 and -3) were from the river water from which the NF patient acquired the infection. While isolates NF2 to NF4 were clonal, NF1 was genetically distinct. Outside the conserved core genomes of these 10 isolates, several unique genomic features were identified. The most virulent strains possessed one of the following four virulence factors or a combination of them: cytotoxic enterotoxin, exotoxin A, and type 3 and 6 secretion system effectors AexU and Hcp. In a septicemic-mouse model, SSU, NF1, and Riv2 were the most virulent, while NF2 was moderately virulent. These data correlated with high motility and biofilm formation by the former three isolates. Conversely, in a mouse model of intramuscular infection, NF2 was much more virulent than NF1. Isolates NF2, SSU, and Riv2 disseminated in high numbers from the muscular tissue to the visceral organs of mice, while NF1 reached the liver and spleen in relatively lower numbers on the basis of colony counting and tracking of bioluminescent strains in real time by in vivo imaging. Histopathologically, degeneration of myofibers with significant infiltration of polymorphonuclear cells due to the highly virulent strains was noted. Functional genomic analysis provided data that allowed us to correlate the highly infectious nature of Aeromonas pathotypes belonging to several different species with virulence signatures and their potential ability to cause NF.


A seminal paper regarding Aeromonas species as a cause of human infection was first published in 1968 (1). Since then, water-associated infections with aeromonads have been on the rise (2, 3). Further, the high resistance of Aeromonas species to both water chlorination, especially in biofilms, and multiple antibiotics (4) squarely resulted in the placement of this organism on the EPA's Contaminant Candidate List 2 (5) and its categorization as an emerging human pathogen. Indeed, Aeromonas species were isolated as pure cultures from 22% of the wounds of infected patients during the 2004 tsunami in southern Thailand (6). In addition to causing severe wound infections, this pathogen is associated with hemolytic-uremic syndrome and necrotizing fasciitis (NF) (711). The two earliest human cases of NF with Aeromonas were published in 1997 (12, 13), while the latest reported case of NF resulted from trauma and freshwater contact following a motorcycle accident (14).
Over the past 2 decades, several virulence factors of Aeromonas have been extensively characterized. For example, an aerolysin-related cytotoxic enterotoxin (Act) is a type 2 secretion system (T2SS)-secreted exotoxin with the ability to cause both diarrhea and severe tissue damage in the host (1519). The toxin leads to apoptosis or necrosis of host cells, depending upon the dose (1618, 20, 21). A functional T3SS and two potent and diverged T3SS effectors, namely, AexT and AexU, have been identified in aeromonads (22), with GTPase-activating protein (GAP) and ADP-ribosyltransferase (ADP-RT) activities, leading to host cell death. Likewise, two T6SS effectors, hemolysin-coregulated protein (Hcp) and the valine-glycine repeat G (VgrG) family of proteins (VgrG1,-2, and -3) have been reported and characterized (23, 24). The biological effects of translocated VgrG proteins are associated with their specific C-terminal extensions. In Aeromonas hydrophila SSU (recently reclassified as Aeromonas dhakensis sp. nov. comb. nov. (25), we have shown that VgrG1 carries a C-terminal vegetative insecticidal protein 2 (VIP-2) domain with ADP-RT activity, and its translocation into host cells leads to the ADP-ribosylation of actin, resulting in cytotoxic effects (24). Likewise, the T6SS-dependent translocation of Hcp was first shown by our group in A. dhakensis SSU-infected human colonic epithelial cells, which led to caspase 3 activation and subsequent apoptosis of host cells (23). In addition to its ability to be translocated into host cells, the secreted form of Hcp inhibited bacterial phagocytosis (26).
Aeromonads have various quorum-sensing (QS) systems that could modulate bacterial virulence genes (27, 28). Likewise, the ability of bacteria to swim and swarm contributes to the overall virulence of aeromonads (29, 30). It is unclear whether any of these virulence factors/mechanisms of Aeromonas species contribute to the pathogenesis of NF or whether NF is a host-mediated disease. During NF caused by group A Streptococcus (GAS), superantigens produced by the bacteria interact with major histocompatibility complex II of the antigen-presenting cells, as well as the β-chain of the T-cell receptor and the receptor of the CD28 costimulatory molecule (31). However, such superantigens have not yet been identified in aeromonads. Clearly, individuals who are somehow predisposed to NF genetically, physiologically, or immunologically are at risk (32). With this variation in host susceptibility, exposure to a large number of A. hydrophila bacteria of a particular pathotype in a unique environment could likely lead to a rapid, fulminant, and necrotic infection, requiring rapid surgical and chemotherapeutic intervention. To provide some answers, we performed genome-wide sequencing of NF-causing strains of A. hydrophila and compared them with other clinical and environmental isolates of aeromonads to identify any unique or pathotype-specific genes that could be implicated in NF. In addition, we developed an NF mouse model of infection and monitored the dissemination of bacteria in this model in real time by in vivo imaging.


Bacterial strains.

The sources of the Aeromonas strains and plasmids used in this study are listed in Table 1. Two A. hydrophila isolates successively cultured from a patient with NF were designated NF1 and NF2. These NF strains were recovered from the patient during and immediately after amputation surgery (culture of stump tissue) and preserved in Trypticase soy broth with 30% glycerol at −80°C. Following extended cultivation, two additional colony types, NF3 and NF4, were isolated from the same sheep blood agar (SBA) plate as NF2. These two colonies appeared much smaller than the dominant isolate, NF2.
TABLE 1 Strains and plasmids used in this study
Strain or plasmidRelevant characteristic(s)Source or reference
A. dhakensis CDCa
    SSU-RifrRifr strain of A. dhakensis SSULaboratory stock
    SSU-Rifr-lucStrain with Tn7-luciferase operonThis study
    SSU ΔahyRI mutantahyRI gene deletion mutant of A. dhakensis SSU; Smr Spr RifrLaboratory stock
    SSU ΔluxS mutantluxS gene deletion mutant of A. dhakensis SSU; Kmr RifrLaboratory stock
    SSU ΔqseB mutantqseB gene deletion mutant of A. dhakensis SSU; Kmr RifrLaboratory stock
A. hydrophila  
    ATCC 7966TCanned milk isolateATCCb,c
    ATCC 7966T-Rifr-lucRifr strain of A. hydrophila ATCC 7966T and Tn7-luciferase operonThis study
    NF1Initial isolate from patient wound siteThis studyc
    NF1-Rifr-lucStrain with spontaneous rifampin resistance and Tn7-luciferase operon 
    NF2-NF4Morphologically distinct isolates obtained successively from NF1 wound site following amputation and surgical debridementThis studyc
    NF2-Rifr-lucRifr strain of A. hydrophila NF2 with Tn7-luciferase operonThis study
    WITracheal aspirateThis studyc
    Riv3River water isolateThis studyc
A. jandaei  
    Riv2River water isolateThis studyc
    Riv2-Rifr-lucRifr strain of A. jandaei Riv2 with Tn7-luciferase operonThis study
A. caviae  
    NM22Stool sample isolate33c
    NM33Well water isolate33c
P. aeruginosa  
    PA103ExoA-positive strainLaboratory stock
    ΔexoA mutantexoA gene deletion mutant of P. aeruginosa PA103Laboratory stock
Plasmid pUTmini-Tn5::luxKm2KmrLaboratory stock
Origin, Philippines.
ATCC, American Type Culture Collection.
Origin, United States.
Since major pathogenic aeromonads form beta-hemolytic colonies relatively similar in colony size, morphology, and color on SBA plates, it is highly conceivable that the original culture from the amputated leg was a “mixed culture” of similar-looking Aeromonas strains, namely, A. hydrophila NF1 and NF2. No other Gram-positive or Gram-negative organisms were isolated from the cultures. During a second culture of tissue taken later from the amputation stump, a single colony of the persistent strain, NF2, was chosen for identification. Two environmental strains (A. jandaei Riv2 and A. hydrophila Riv3) were isolated from the river water from which the above-described patient acquired the infection. Two isolates that belonged to the A. caviae-media group (A. caviae NM22 and NM33) had similar DNA signatures and virulence factor-encoding genes and were isolated, respectively, from well water (NM33) and from a child with diarrhea (NM22) who consumed that contaminated well water (33). These isolates were confirmed in this study to be strains of A. caviae on the basis of phylogenetic analysis of the rpoD gene (data not shown). A. hydrophila strain WI, resistant to multiple antibiotics (ampicillin-sulbactam, amoxicillin-clavulanic acid, aztreonam, cefazolin, cefuroxime, ciprofloxacin, ertapenem, gentamicin, imipenem, and levofloxacin), was from the tracheal aspirate of a patient with acute adult respiratory and kidney failure. A. dhakensis SSU was obtained from the Centers for Disease Control and Prevention (CDC) and was originally obtained from a patient with diarrhea during a cholera-like outbreak in the Philippines.

Biochemical identification of Aeromonas isolates.

All of the bacterial strains were identified by conventional methods, specifically, Aerokey II (34) and additional biochemical tests (35). River water samples were collected and transported in Whirl-Pak bags (Nasco, Fort Atkinson, WI), and 10-ml aliquots were passed through 47-mm paper filters (pore size, 0.45 μm). The filters were placed on ampicillin-dextrin agar with vancomycin to select for aeromonads (this medium detects all Aeromonas species except [potentially] A. trota) (36).

Genome sequencing and annotation.

Genome sequencing of A. hydrophila NF1 and NF2 and A. dhakensis SSU was performed at the Emory Genome Center with GS-FLX (454 Life Sequencing, Branford, CT). The number of reads was 519,429 for strain NF1, 418,086 for NF2, and 261,221 for strain SSU. The estimated average coverage of the NF1, NF2, and SSU genomes was 23, 19-, and 11-fold, respectively. Genomes were assembled with Newbler software (37). The draft genome of strain NF1 was 4,809,530 bp on 170 contigs, that of NF2 was 4,789,684 bp on 154 contigs, and that of strain SSU was 4,879,288 bp on 281 contigs.
Genome sequencing of A. hydrophila Riv3 and WI, A. caviae NM22 and NM33, and A. jandaei Riv2 was performed at the University of Texas Medical Branch (UTMB) Molecular Genomics Core Facility with a HiSeq 1000 (Illumina, San Diego, CA) by using the 2 × 50-cycle paired-end protocol. Trimmed reads were assembled into contigs by using AbYSS (38). The draft genome of strain WI was 5,172,023 bp on 112 contigs, that of Riv2 was 4,478,089 bp on 43 contigs, and that of strain Riv3 was 4,861,432 bp on 58 contigs. The draft genomes of strains NM22 and NM33 contained 4,419,058 bp on 250 contigs and 4,470,246 bp on 162 contigs, respectively.
Genome sequencing of A. hydrophila NF3 and NF4 was performed on a MiSeq (Illumina, San Diego, CA) by using paired-end version 2 chemistry. The number of reads was 2,169,752 for strain NF3 and 6,762,904 for strain NF4. Paired-end FASTQ data sets were trimmed and assembled into contigs by using CLC Genomics Workbench, version 6.0.5 (CLC bio, Aarhus, Denmark). A draft genome of strain NF3 was 4,757,900 bp on 197 contigs, and that of NF4 was 4,779,984 bp on 156 contigs. The estimated average coverage of NF3 and NF4 genomes was 42- and 50-fold, respectively. Genomic contigs were annotated by using the RAST annotation server (39) to identify RNAs and protein-coding genes.

Comparative genomic analyses.

Comparative genomic analysis was performed to identify shared and dispensable genetic traits among the Aeromonas strains that were sequenced. In addition to the 10 genomes sequenced during this study, the closed genomes of A. hydrophila ATCC 7966T (GenBank accession no. CP000462.1), from a tin of milk with a fishy odor (40), and A. salmonicida A449 (GenBank BioProject PRJNA58631), from a brown trout (41), as well the draft genomes of A. hydrophila E1 and E2 (SRA063950), from a patient with a wound infection (42), A. aquariorum AAK1 (GenBank accession no. BAFL00000000.1), recently reclassified as A. dhakensis sp. nov. comb. nov. (25) and from the blood of a patient with septicemia and NF (43), A. veronii B565 (GenBank accession no. CP002607.1), from aquaculture pond sediment (44), and A. caviae Ae398 (GenBank Whole-Genome Shotgun project no. CACP00000000) from a child with profuse diarrhea (45), were used.
To facilitate comparisons, the genomes of these other seven aeromonads were also annotated by using the RAST annotation server. Genome-to-genome comparisons were performed primarily with a SEED viewer (46), which uses bidirectional protein-protein BLAST (blastp) sequence comparison of translated open reading frames. For all draft genomes, genes at the end of a contig or interrupted by contig gaps were analyzed by using bidirectional BLASTN analysis against all other genomes. Genomic regions (GRs), defined as regions present in one (unique) or more genomes and missing from at least one other genome (dispensable), were identified as previously reported (42) for the genomes of A. hydrophila ATCC 7966T, WI, Riv3, NF1 to NF4, E1, and E2; A. dhakensis SSU; and A. jandaei Riv2 (47, 48). Average nucleotide identity (ANI) by BLAST was computed with Jspecies (49). Evolutionary analyses were conducted in MEGA5 (50). The act (cytotoxic enterotoxin) gene locus was compared among different Aeromonas strains by using the Artemis Comparison Tool (51).

Core cluster alignment and phylogenetic inference.

The complete predicted proteome from the 16 Aeromonas genomes was searched against itself by using BLASTP with an E value cutoff of 1e−05. The best blast scores were converted into a normalized similarity matrix with the OrthoMCL (52) algorithm, which uses an additional step of Markov Clustering algorithm (MCL) to improve the sensitivity and specificity of the orthologous sequences identified. Core genes were identified as the protein-coding gene clusters that were shared by all of the 16 Aeromonas isolates used in this study. Multiple-sequence alignments of the core protein-coding genes were generated with the program MUSCLE (53) by using default settings. These core alignments were filtered for uninformative characters by GBLOCKS (54) by using default settings. Whole-genome phylogeny was determined with a concatenation of aligned individual core protein sequences, followed by neighbor joining (NJ) with NEIGHBOR in the PHYLIP package (55, 56). Distances were identified with the PROTDIST program within PHYLIP. The support of the data for each of the internal node of the phylogeny was estimated by using 100 bootstraps.

In vitro characterization of Aeromonas strains for various virulence traits.

On the basis of the literature and our own earlier studies, biofilm formation, swimming and swarming motility, protease production, and toxin/effector secretion via different secretion systems are correlated with the virulence of aeromonads in animal models. We have refined several of these in vitro assays to correlate gene expression with the functionality of the target gene to obtain reliable and quantitative data.

Crystal violet (CV) biofilm assay.

To measure solid-surface-associated biofilms, a modified biofilm ring assay (57, 58) was used. Biofilm formation was then quantified (59), and the results were normalized to 1 × 109 CFU to account for any minor differences in the growth rates of the various bacterial strains used. The CFU count was determined by serial dilution of the samples, followed by plating (33, 58).

Motility assay.

Luria-Bertani (LB) medium with 0.35% Difco Bacto agar (Difco Laboratories, Detroit, MI) was used to characterize swimming motility, while Difco nutrient broth with 0.5% Eiken agar (Eiken Chemical Co., Ltd., Tokyo, Japan) was used to measure the swarming motility of aeromonads (30, 42, 58). These methods provided better quantification of the motility data.

Measurement of protease activity.

Protease activity was measured in culture filtrates of Aeromonas isolates grown overnight (60). The hide powder azure substrate was used to measure protease activity (42). Protease activity was calculated per milliliter of culture filtrate per 108 CFU as determined by serial dilution of the samples and colony counting.

Measurement of hemolytic activity.

To measure hemolytic activity associated with Act, the culture filtrates from various Aeromonas isolates were first treated with trypsin (final concentration, 0.05%) at 37°C for 1 h and then subjected to a hemolytic activity assay with rabbit erythrocytes (61). Hemolytic activity titers were calculated as the absorbance at 540 nm of the hemoglobin release multiplied by the dilution factor of the culture filtrates. Units of hemolytic activity were reported per milliliter of cell filtrate/1 × 108 CFU (58). Although Act is active against erythrocytes of other animal and human species, rabbit red blood cells are more sensitive to the action of Act and hence were used in this study.
For the neutralization assay, culture filtrates of the strains studied were mixed with either preimmune (control) or hyperimmune rabbit serum (laboratory stock, 1:10 dilution) containing antibodies to Act (42, 62) and incubated at 37°C for 1 h before use.

Lactone production.

N-Acyl-homoserine lactone (AHL) production was detected by cross-streaking various Aeromonas strains against the biosensor Chromobacterium violaceum CV026 on LB agar plates (27, 42). Violet pigment production from C. violaceum CV026 by AHLs produced by Aeromonas isolates was scored on the basis of color intensity after overnight incubation of the plates at 37°C. The result was recorded as no lactone production (−) or a low (+), moderate (++), or high (+++) level of AHLs. An isogenic ahyRI mutant of A. dhakensis SSU (Table 1) was used as a negative control because of its inability to produce AHLs (62).

Western blot analysis.

Western blot analysis was used to detect the production and secretion of Hcp (a T6SS effector), AexU (a T3SS effector), and exotoxin A (ExoA) by various Aeromonas isolates. Briefly, the supernatants and cell pellets from cultures grown overnight were separated and the pellets were directly lysed in SDS-PAGE loading buffer. The proteins in the supernatants were first precipitated with 10% (vol/vol) trichloroacetic acid and then dissolved in the loading buffer before being subjected to SDS-PAGE. For Western blot analysis, specific antibodies to Hcp and AexU (available in the laboratory) (22, 26, 63) and to ExoA of Pseudomonas aeruginosa (LSBio, Inc., Seattle, WA) were used. At the amino acid level, ExoA of A. hydrophila is 65% homologous to that of P. aeruginosa (64). The bacterial strains were grown overnight in the LB medium at 37°C. The yield of ExoA produced by bacteria was influenced by the concentration of iron in the culture medium (65, 66).

Generation of bioluminescent Aeromonas strains.

The reporter strains were generated following triparental conjugation of rifampin-resistant (Rifr) Aeromonas isolates with Escherichia coli SM10 λpir carrying the pTNS2 plasmid and SM10 harboring the pUC18-mini-Tn7T-Km2-lux plasmid (67). The minitransposon system contains a lux luminescence operon with the native promoter and a kanamycin resistance (Kmr) cassette for transposon selection. This system allows site-specific transposition downstream of the glmS gene, which encodes a conserved glucosamine-6-phosphate synthetase, with the helper plasmid pTNS2 providing the transposase complex (67). The mutants were screened for bioluminescence by using an ImageQuant LAS4000 bioluminescence and fluorescence imaging workstation (GE Healthcare Sciences, Pittsburgh, PA). The reporter strains that were oxidase positive and emitted bioluminescence signals were selected for further characterization. Insertion sites were confirmed by PCR with primers PTn7R:5′-CACAGCATAACTGGACTGATTTC-3′ and GlmSFwd: 5′-GCCAGTATCCCATTGCCATG-3′, which, respectively, corresponded to the 5′ end of the Tn7 minitransposon and the 3′ end of the glmS gene in Aeromonas, followed by DNA sequencing. Rifr Aeromonas strains were generated by spontaneous mutation in response to antibiotic selection (200 μg/ml) to aid in the selection of transposon mutants.

Animal experiments. (i) Septicemic-mouse model of infection.

In a first set of experiments, groups (n = 6 to 23) of healthy female Swiss Webster mice (Taconic Farms) were infected via the intraperitoneal (i.p.) route with various aeromonads. The animals were infected with a dose of 5 × 107 CFU, and deaths were recorded for 14 days postinfection (p.i.). Before each study, a pilot experiment was performed with three different doses of SSU (8 × 106, 2 × 107, and 5 × 107 CFU). The dose chosen was that which was 100% lethal. For infection studies, Aeromonas cultures grown overnight were centrifuged and the pellets were washed three times in sterile phosphate-buffered saline (PBS) before being suspended in 1/10 of the original culture volume of PBS. Subsequently, each culture was titrated and inocula were prepared such that a 50-μl volume contained the intended infectious dose of the organism.

(ii) Mouse model of i.m. infection to mimic NF.

Mice (n = 5/group) were anesthetized with isoflurane, 50 μl of various Aeromonas cultures was injected intramuscularly (i.m.) into one of the legs at doses of 5 × 106 to 5 × 108 CFU, and deaths were recorded for 14 days p.i. The animals were also observed for the development of possible necrotic lesions around the injection site. The animals used for the time point studies were euthanized by carbon dioxide narcosis, followed by cervical dislocation.

Evaluation of bacterial dissemination after i.m. infection.

Animals (n = 5) infected as described above were euthanized at 24 or 48 h p.i. From each animal, the spleen and liver were collected aseptically and transferred to disposable tissue grinders (Fisher Scientific, Pittsburgh, PA) containing 1 ml of PBS for the spleen and 2 ml for the liver. These tissues were homogenized and serially diluted, and aliquots were subjected to bacterial counting after incubation on LB agar plates at 37°C for 24 h.

Gross pathological examination of lesions and processing of muscle tissues for histopathological analysis.

At the time of collection of internal organs for CFU determination, as well as from those mice that survived for longer times, the lesions associated with NF were grossly examined. In addition, leg muscle tissues adjacent to the injection site were collected and fixed in 10% buffered formalin. After 48 h of fixation, the tissues were processed and sectioned at 5 μm before staining with hematoxylin and eosin (H&E). The tissue sections were evaluated by light microscopy in a blinded fashion.

IVIBB in mice.

As we sought to characterize the i.m. model of mouse infection to mimic NF, we used in vivo imaging of bioluminescent bacteria (IVIBB) to evaluate colonization and dissemination of bacteria to peripheral organs. Consequently, strains SSU, ATCC 7966T, NF1, and Riv2 harboring the Tn7 minitransposon system with a lux operon were used.
Animals (n = 5) infected by the i.m. route with various bioluminescent Aeromonas strains were lightly anesthetized with isoflurane (24- and 48-h time points) at the indicated doses and subjected to in vivo imaging on an IVIS 200 bioluminescence and fluorescence whole-body imaging workstation (Caliper Life Sciences, Alameda, CA). At the same time points, mice were sacrificed after imaging and their livers, spleens, and leg muscles were subjected to CFU determination.
Finally, quantitation of radiance by regions of interest (ROIs), corresponding to the anatomical regions of liver, spleen, and leg muscle, was accomplished by using Living Image software (Caliper Life Sciences). ROIs of the same shape and area were replicated across images measuring total flux (ρ/s) to determine radiance as we previously described (68).

Statistical analysis.

All of the experiments were performed in triplicate, and statistical significance was analyzed by one-way analysis of variance (ANOVA). All of the animal data were subjected to Kaplan-Meier survival estimates, and a P value of ≤0.05 was considered significant.

Accession numbers.

The whole-genome shotgun projects and their associated short-read archive (SRA) files for the genomes described in this study have been deposited at NCBI under the following Biosample numbers: A. hydrophila WI, SAMN02597475; A. hydrophila NF1, SAMN02597476; A. hydrophila NF2, SAMN02597477; A. hydrophila NF3, SAMN02597478; A. hydrophila NF4, SAMN02597479; A. hydrophila Riv3, SAMN02597480; A. dhakensis SSU, SAMN02597481; A. jandaei Riv2, SAMN02597482; A. caviae NM22, SAMN02597483; and A. caviae NM33, SAMN02597484.


Genome-wide comparisons of various Aeromonas species.

Four A. hydrophila isolates, NF1 to NF4, two environmental river water isolates, Riv2 and Riv3, tracheal aspirate isolate WI of A. hydrophila, and diarrheal isolate SSU of A. dhakensis were sequenced, assembled, and annotated. Likewise, A. caviae NM22 (stool) and NM33 (water) isolates were sequenced and used for comparison analysis along with sequences from two recently described A. hydrophila strains, E1 and E2, from a case of human wound infection (42, 69). Additional comparisons were made with a number of previously published Aeromonas genomes (33, 40, 44, 45, 70).

ANI comparisons.

Pairwise ANI comparisons of the NF-causing strains revealed that NF2 to NF4 were very similar to each other, compared to the general level of ANI between isolates the same species (Table 2). However, NF3 and NF4 appeared to be more similar to each other than to NF2 (Table 2), and in general, the values were below what we would expect for “true” clonal relationships, i.e., identical strains. The ANI values are sensitive to differences in the sequencing and assembly technologies used, i.e., Roche 454 GS Junior versus Illumina MiSeq, as well as differences in coverage and the number of assembled contigs (which probably reflects differences in the number of repeat elements between strains). For example, the genome of strain NF3 was composed of 197 contigs, while those of NF2 and NF4 contained 154 and 156 contigs, respectively. Differences in gene content, measured by pairwise core genome size (Table 2), also suggested that the genomes of NF2 to NF4 were highly similar. Indeed, the core genome of these three strains (NF2 to NF4) contained 3,964 coding sequences (CDSs) at the 100% identity level, while the total core genome contains 4,230 CDSs. Visual inspection confirmed that many genes below 100% identity were the result of single-nucleotide polymorphisms (SNPs) altering the reading frame and truncating or lengthening the translated product, which is a common error found in pyrosequencing.
TABLE 2 ANIs and shared gene contents in pairwise comparisons of Aeromonas genomes
Strain (total no. of CDSs)ANI (%) or no. of CDSs shareda with strain:
1. A. hydrophila ATCC 7966T (4,279) 96.7896.6596.9696.9296.9696.7897.1296.8692.9986.4884.7685.586.4585.6485.66
2. A. hydrophila WI (4,660)3,986 96.3196.8496.8196.8696.6396.9596.7192.7886.3484.6485.2986.2985.5585.57
3. A. hydrophila Riv3 (4,358)3,9843,975 96.9296.996.9296.9296.9996.9792.9586.4284.6485.4286.3885.6185.63
4. A. hydrophila NF2 (4,366)3,9353,9463,923 99.9599.9896.9297.0296.8192.8386.3684.7985.3986.5285.6685.69
5. A. hydrophila NF3 (4,373)3,9033,9233,9004,253 99.9696.8696.9896.7692.7886.2484.7585.4386.4385.6785.67
6. A. hydrophila NF4 (4,375)3,9383,9533,9334,2854,296 96.9197.0296.8192.7886.3184.7485.3386.4185.685.6
7. A. hydrophila NF1 (4,361)3,9123,9513,9234,0213,9844,013 96.8396.7792.8586.3284.6785.4586.3385.5985.6
8. A. hydrophila E2 (4,241)3,8973,8673,9013,8353,8043,8313,807 96.8892.9886.3584.8385.3686.3985.6785.69
9. A. hydrophila E1 (4,373)3,8953,8923,9173,9553,9223,9503,9793,814 92.886.3284.7185.3486.4285.6285.64
10. A. dhakensis SSU (4,527)3,7593,8093,7653,8353,8203,8423,8503,6813,816 85.8884.5485.1486.4785.6585.68
11. A. salmonicida A449 (4,570)3,5643,6063,5613,6263,6023,6303,6393,4753,5983,605 82.9784.1784.1983.5583.62
12. A. jandaei Riv2 (4,107)3,4713,4723,4553,5183,5003,5243,5403,3923,5003,4753,366 88.7683.5382.7882.81
13. A. veronii B565 (4,045)3,4023,4263,3733,3783,3633,3863,4203,3083,3523,4073,2903,433 83.7482.882.82
14. A. caviae Ae398 (4,043)3,3393,4323,3613,3343,3193,3373,3693,2753,3243,3403,2853,1243,192 97.1797.18
15. A. caviae NM33 (4,172)3,3183,3623,3333,3103,2913,3163,3313,2623,3343,2653,2663,1103,1193,600 99.94
16. A. caviae NM22 (4,134)3,3003,3493,3163,2983,2833,3053,3103,2503,3183,2513,2493,0933,1033,5883,966 
ANIs (upper right) and shared gene contents (lower left) revealed by pairwise comparisons are shown. Values in bold indicate that the strains are of the same species on the basis of a threshold of 95.0%.
In order to correct for this problem, we built a robust tree by using only the 2,541 conserved full-length protein sequences (Fig. 1). This tree suggested that while isolates NF2 to NF4 were closely related, NF2 was a significantly more distant relative. In 696,126 positions in the concatenated protein alignment, there were 63 amino acid substitutions between NF3 and NF4, whereas there were 297 and 320 substitutions between NF2 and NF3 and between NF2 and NF4, respectively. The high numbers of substitutions may be, to some extent, an artifact of the use of different sequencing technologies (454 for NF2 and MiSeq for NF3 and NF4). However, this level of variation suggests that the three strains had a last ancestor in common well before they infected the patient or coinhabited the environmental niche directly before infection. For comparison, a recent simulation of within-host variation of the 2.9-Mbp Staphylococcus aureus genome showed that a diversity of four or five SNPs per genome should be expected from long-term clonal expansion with a constant population size of 5,000 (71).
FIG 1 NJ tree of strains built by using 2,514 conserved full-length predicted proteins concatenated together. The distance was calculated with PROTDIST (55, 56). The scale bar shows 0.001 substitution per site. There was strong bootstrap support for the tree, with all but three branches having 100/100 replicate congruence. The three branches with >50% but <100% support were all internal branches.
Strain NF1 was cultured from the initial surgical site and was distinctly different from the NF2 to NF4 group, supported by pairwise ANI values, gene content comparisons (Table 2), and the core protein phylogeny (Fig. 1). Interestingly, all four isolates (NF1 to NF4) were of the same molecular serotype on the basis of comparisons of the gene content and similarity of CDSs located between the waaL and rmlA genes. Further, the O-antigen region of A. dhakensis SSU was identical in terms of gene content to NF1 to NF4 but divergent in nucleotide identity at a level comparable to that of the overall genome nucleotide identity between these two groups of organisms (see Fig. S1 in the supplemental material).
Pairwise ANI comparisons of all of the A. hydrophila genomes examined in this study revealed a heterogeneous collection of isolates of this species (Table 2). This is not surprising, given that some strains were from environmental sources, while clinical isolates were from different patients, sites, and sources. However, it is interesting that clinical isolates from different patients and disease syndromes were just as divergent from environmental isolates as clinical isolates were, according to ANI. Genomic comparison of gene contents (Table 2), on the other hand, revealed that the genomes of clinical strains had more genes in common than did a clinical isolate and an environmental isolate (see below).

Core and pangenome.

In terms of a core genome, among seven A. hydrophila strains (NF2 was used to represent NF2 to NF4); we estimated a core genome of 3,617 CDSs. In terms of core genome stability, on average, 110 CDSs were subtracted from the core genome upon the addition of each A. hydrophila genome (see Fig. S2 in the supplemental material). Variation from this mean was high because of a large number of dispensable genes, as well as the geotemporal relatedness of strains from the same patient, such as NF1 to NF2 and E1 to E2. As expected, the core genome size decreased sharply when a genome from a different species was included in the analysis (see Fig. S2). To identify genetic features related to or indicative of virulent pathotypes, we identified dispensable and unique GRs from 11 sequenced genomes (9 of A. hydrophila, 1 of A. dhakensis, and 1 of A. jandaei) and determined the distribution of these features (Table 3; see Table S1 in the supplemental material).
TABLE 3 Distribution of GRs of A. hydrophila and A. dhakensis strains among Aeromonas species genomes
Virulence trait(s)Presence in or absence from strain:
ATCC 7966TWIRiv3NF2NF3NF4NF1E2E1SSUA449Riv2B565Ae398NM33NM22
Surface appendages and features                
    CFA/I (α C/U) fimbriae, AHA_0060++++++++++++
    CFA/I (α C/U) fimbriae, AHA_1021+++++++
    P (π C/U) fimbriae, AHA_0521+++++++++++++
    Phage-encoded P (π C/U) fimbriae+
    Tap type IV pilus++++++++++++++++
    MSHA BFP type IV pilus, mshaH acd mshaIIJK3LM_DOPQ++++++++++++++
    TAD Flp pilus+++
    Polar flagellum++++++++++++++++
    flaABG-fliDS cluster, AHA_1698–AHA_1702++++++++++++++
    Lateral flagella++++++++
Putative virulence factors/toxins                
    Cytotoxic enterotoxin/hemolysin, act+++++++++++
    Cytotonic enterotoxin, ast+++++++++
    Pore-forming cytolysin/hemolysin, hlyA+++++++++++
    TPS hemolysin+
    Cytotonic enterotoxin/lipase, alt++++++++++++++++
    Phospholipase/lecithinase/hemolysin (GCAT)++++++++++++++++
    Extracellular protease precursor, AHA_2712–AHA_2714++++++++++++++
    ExoA homolog+++++
    syp exopolysaccharide+++
    Group II capsule+++
    RTX toxin, AHA_1359, transporter cluster++++
    FHA family, RTX toxin+
    Ferrichrome iron uptake cluster, AHA_1951–AHA_1954++++++++++++++
    Ferric hydroxamate uptake (fhu) operon+++++++++++++++
    Pyoverdine siderophore operon, ASA_4368–ASA_4378++
    Ferric siderophore-TonB cluster, AHA_3433–AHA_3439+++++++++++++++
T3SSs and T6SSs                
    hcp (T6SS)++++++++++++
    yopH and chaperone (T3SS)++++++
    aexTU (T3SS)++++++++
    yopT and chaperone (T3SS)+
Resistance, including antimicrobials                
    ABC-type multidrug transport, AHA_0484–AHA_0486+++++++++++
    MATE pump, AHA_3203–AHA_3204+++++++++++++++
    DMT efflux, AHA_3820–AHA_3821++++++++++++
    Multidrug resistance protein B, AHA_4116–AHA_4117+++++++++++
    Acriflavin RND transporter, AHA_1320–AHA_1323+++
    NodT family RND efflux pump+++
    RND efflux pair, AHA_2959–AHA_2960+++++++++++++
    Macrolide efflux pump, macAB, AHA_0758–AHA_0761++++++++++++++
    Tripartite MRS+
    Polymyxin B resistance (arn)++++++++++++
    cepS, cephalosporinase+++++++++++++++
    ampS(H), class D beta-lactamase++++++++++++++++
    imiS(H), class 3 metallo-β-lactamase+++++++++++++
    LPS modification, cephalosporin hydroxylase, AHA_4152–AHA_4169++
    LPS modification, aminoglycoside 3′-N-acetyltransferase, AHA_4170–AHA_4182++++++
    Aminoglycoside 6′-N-acetyltransferase++++++++++
    Mercury, arsenic, heavy metal resistance tnp++++
    Heavy metal resistance tnp, TEM-1 beta-lactamase+
    Aminoglycoside 3′-PTS+
    Chloramphenicol acetyltransferase, mercury resistance, OXA-1 beta-lactamase transposon+
    Spectinomycin 9-O-adenylyltransferase, sulfonamide, OXA beta-lactamase+
    Macrolide 2′-PTS+
    Integrative conjugative element (46 kb), organic solvents++
    Quaternary ammonium compound, ethanolamine, propanediol+++++++
    Organic hydroperoxide resistance, AHA_3609–AHA_3611++++++++++++++
    Accessory arsenic resistance operon+++

Surface appendages.

Like those of other Gram-negative bacteria, Aeromonas genomes harbored genes for an assortment of fimbriae and pili. All of the genomes harbored the Tap type IV pilus (TFP) locus, whose pilin and subunit gene clusters were under considerably stronger evolutionary pressure than those from other Tap TFP gene clusters or the overall genomes (see Fig. S3 in the supplemental material). All of the genomes also contained genes for the mannose-sensitive hemagglutinin (MSHA) TFP locus (Table 3). The genome of A. salmonicida A449 was found to be missing a seven-gene internal cluster, mshNEGF2BAC. The genomes of both A. caviae NM22 and NM33 revealed a deletion of the majority of the msh locus genes because of the insertion of a mobile element, resulting in a truncated gene arrangement of mshH-acd-(mobile element)-mshDOPQ (Table 3).
The tight adherence (TAD)/fimbrial low-molecular-weight protein (Flp) pilus was present in A. hydrophila ATCC 7966T, as well as in A. salmonicida A449 and A. veronii B565. Three fimbriae were widely distributed among Aeromonas genomes, two α (class 5) alternate chaperone-usher fimbriae and one π (P) chaperone-usher fimbria. Additionally, the genome of A. hydrophila Riv3 contained a second π fimbria that was encoded on a prophage element (Table 3).
All Aeromonas genomes contained five genetic loci encoding a single polar, unsheathed flagellum, with one exception. The genomes of A. caviae NM22 and NM33 did not contain the flaABG fliDS locus, which encodes flagellins (flaABG), the cap protein (fliD), and the essential fliS gene product (Table 3). Accordingly, these two strains were nonmotile (data not shown). A 35-kb lateral flagellar gene cluster was present in the genomes of NF1 to NF4, E1, SSU, A449, and A. jandaei Riv2, all of which were highly pathogenic species or strains (Table 3). The gene content of this locus was highly conserved among Aeromonas species, and the gene cluster had the same chromosomal location in all of the strains; downstream from the gene encoding rRNA small-subunit methyltransferase I (RsmI, AHA_3894). Phylogenetic reconstruction of the evolutionary history of this locus argues for vertical transmission (see Fig. S4 in the supplemental material).


Several other virulence factor-encoding genes were identified in the genomes analyzed and included lipases, hemolysins, siderophore clusters, elastase, cytotonic/cytotoxic enterotoxins, T3SS, and the capsule polysaccharide-encoding loci (Table 3). The distribution of three of these factors, act (cytotoxic enterotoxin), exoA (ExoA), and a T3SS, correlated with virulent pathotypes. The act gene was present in ATCC 7966T, WI, NF1 to NF4, E1, SSU, A449, Riv2, and B565 but absent from all three of the A. caviae genomes and Riv3 (Table 3).
Closer inspection of the act genetic locus revealed variability in terms of gene content (Fig. 2A). For example, in the genomes of NF2 to NF4, a gene encoding a putative lipoprotein and a group 1 glutathionylspermidine synthase (EC was inserted downstream of the act gene. Additionally, the act gene was truncated in A. hydrophila NF2 to NF4 (567 bp), as opposed to other A. hydrophila and A. dhakensis act gene sequences, which were 1,482 bp in length (Fig. 2A). A putative arylsulfate sulfotransferase-encoding gene was inserted in this locus in the genomes of A. hydrophila Riv3, WI, and E2, of which only A. hydrophila WI possessed the act gene (Fig. 2A). Additionally, the act locus of Riv2 was approximately 600 bp longer than that of strain ATCC 7966T because of the insertion of a 516-bp MerR family transcriptional regulator-encoding gene between the act and thiC genes (data not shown). These alterations in the act gene locus might affect the expression of the gene in various Aeromonas isolates. The sizes of the act genes from other species of Aeromonas were 1,479 bp for A. salmonicida and 1,464 bp for A. jandaei Riv2 and A. veronii B565. Despite these findings, phylogenetic analysis of the act locus supported the hypothesis that this gene was vertically transmitted in Aeromonas genomes (Fig. 2B and C).
FIG 2 (A) Genetic organization of the act gene locus from Aeromonas species. Gene locus schematics are drawn to relative scale. Genes are depicted as follows. Gray arrows represent conserved flanking genes, i.e., that for the hypothetical protein FIG 00361276 (5′ or upstream) and thiC (3′ or downstream). Diagonally striped arrows represent the act gene. The white arrow represents hypothetical protein-encoding genes (nonhomologous). The vertically striped arrows represent putative arylsulfate sulfotransferase-encoding genes. The horizontally striped arrow represents the gene for a putative glutathionylspermidine synthase, group 1 (EC The regions between the gene schematics show homology, with darker red representing higher homology and lighter red indicating regions of lower homology, while colorless regions are not homologous and indicate insertions or deletions. The gene locus of strain ATCC 7966T is also representative of SSU, NF1, and E1, while Riv3 is also representative of E2. (B) Phylogenetic relationship of the act gene locus, including conserved flanking genes, i.e., that for hypothetical protein FIG 00361276 (5′ or upstream) and thiC (3′ or downstream), act, and other gene insertions, in Aeromonas. (C) Phylogenetic relationship of the act gene. The evolutionary reconstructions were inferred by using the NJ method. The bootstrap consensus tree was inferred from 1,000 replicates. Evolutionary distance is shown as the number of nucleotide differences.
Five genomes (A. hydrophila NF2 to NF4 and E1 and A. dhakensis SSU) contained a gene encoding an ExoA homologue, which is a NAD-dependent ADP-RT commonly associated with pathogenic Pseudomonas species (Table 3). The toxin leads to ADP-ribosylation of host elongation factor 2, resulting in the shutdown of protein synthesis and cell death (72). Importantly, this feature is a distinguishing trait among the three NF isolates, namely, NF2, NF3, and NF4. All of these isolates had this gene (100% identity among the isolates); however, this gene was absent from the genome of NF1. The exoA gene was also present in the genome of a highly virulent E1 strain isolated from a case of wound infection and a diarrheal A. dhakensis isolate, SSU (Table 3).

T3SSs, T6SSs, and their effectors.

The genomes of eight Aeromonas strains harbored a T3SS and included A. hydrophila NF1 to NF4 and E1, A. dhakensis SSU, A. salmonicida A449, and A. jandaei Riv2 (Table 3). The T3SS was inserted at the same locus, a serine tRNA (TGA) (at nucleotide positions 2983703 to 87 of A. hydrophila ATCC 7966T) (40), in all of the genomes of A. hydrophila and A. dhakensis. The T3SS is located on plasmid pAsa5 in A. salmonicida A449 and is flanked by transposases, indicating horizontal acquisition, while the T3SS was inserted between genes encoding a short-chain-specific acyl coenzyme A dehydrogenase (EC involved in isoleucine degradation and a cystathionine beta-lyase (EC in the genome of A. jandaei Riv2.
In terms of gene content, the 26.7-kb T3SSs (hypothetical ascUTSRQPO ascN aopN acr12 ascXY ascV acrRGVH aopBD axsCBAD ascCDEFGHIJKL) of the A. hydrophila strains analyzed in this study were identical (Table 3). The T3SS of A. dhakensis SSU was slightly larger, approximately 27.5 kb, because of the insertion of an aopX effector protein-encoding gene downstream of ascU (which is transcribed in the opposite direction). This gene was also present in the T3SS of A. salmonicida A449. The T3SS of A. jandaei Riv2 was considerably larger, at 30 kb, having accessory genes at both ends of the gene cluster. There was a 1,182-bp zinc metalloprotease-encoding gene located downstream of ascU and an exoU effector protein homologue and chaperone located downstream of the ascL gene.
Three putative T3SS effector protein gene clusters were identified outside the T3SS locus (Table 3). The aexTU effector-chaperone pair was inserted between a tryptophan-specific-permease-encoding gene and a phenylalanyl-tRNA synthetase beta subunit-encoding gene (AHA_0123 and AHA_0124, respectively). This effector protein was the most widely distributed of the three. Also present in this three-gene cluster was a homologue of the gene for Txp40, a 40-kDa insecticidal toxin. Interestingly, this gene cluster was present in the genome of A. hydrophila Riv3, even though the main T3SS gene locus was absent.
The yopH sycH effector-chaperone pair (Table 3) was located between a hexose phosphate transport protein (UhpT)-encoding gene and a phosphotransferase system (PTS) fructose-specific IIABC component-encoding gene. Its distribution was more restricted; it was present in the five strains of A. hydrophila with a T3SS locus and A. salmonicida A449 (Table 3). The genome of A. jandaei Riv2 also contained a yopT sycT effector-chaperone gene pair (Table 3) inserted between a siroheme synthase-encoding gene and a magnesium transporter-encoding gene (homologues of AHA_4121 and AHA_4122, respectively). Additionally, the genome of A. salmonicida A449 contained a yopO effector homologue and a putative chaperone-encoding gene pair.
In contrast to the T3SS, a gene cluster encoding a T6SS was more widely distributed; it was present in 12 of the 16 Aeromonas genomes that were compared (Table 3). These included the genomes of all nine A. hydrophila isolates, A. dhakensis SSU, A. salmonicida A449, and A. jandaei Riv2. All of the genomes that harbored the cluster had an 18-CDS region, impB to vgrG, in common. The 5′ region of the T6SS cluster was more variable, with most of the genomes containing accessory genes in comparison to ATCC 7966T, WI, and Riv3. Among the A. hydrophila and A. dhakensis genomes, the cluster was located between two tRNA genes, tRNA-Ser-GGA (AHA_1825) and tRNA-His-GTG (AHA_1851). The T6SS cluster was also adjacent to tRNA-His-GTG in the genome of A449, but there was a rearrangement in the 5′ region adjacent to the cluster. On the contrary, the cluster was inserted at a different chromosomal location in the genome of A. jandaei Riv2; however, its exact location could not be deduced because of a repetitive DNA (vgrG) sequence present at both ends of the gene cluster, which led to its assembly into a distinct contig.
In the genome of A. hydrophila ATCC 7966T, a single hcp gene and two vgrG genes were present in this main cluster. A smaller accessory T6SS-related gene cluster, at a separate chromosomal locus, contained a second hcp gene and one additional vgrG gene. This number of T6SS effector protein-encoding genes was apparently conserved among other Aeromonas genomes; however, because of the high homology among paralogues of each type of effector-encoding gene, the exact gene content could not always be concluded. The genomes of A. veronii B565 and A. caviae Ae398, NM22, and NM33 did not contain an effector-encoding gene, hcp or vgrG.

Resistance to antimicrobials.

The cepS(H) cephalosporinase gene was absent only from the genome of A. veronii B565, and the imiS(H) gene was missing only from the three A. caviae genomes (NM22, NM33, and Ae398) (Table 3). Two different aminoglycoside-modifying enzymes were found in some genomes of A. hydrophila, namely, ATCC 7966T, Riv3, E1, and E2, while several Aeromonas genomes contained at least one of these. The genome of A. hydrophila strain WI also possessed a number of modifying enzymes that contributed to antibiotic resistance (Table 3). Among these were genes that imparted resistance to aminoglycosides, chloramphenicol, streptomycin, sulfonamides, macrolides, tetracycline, ampicillin, and oxacillin. At least one of the oxacillin β-lactamases appeared to be an extended-spectrum β-lactamase (Table 3). Most of these genes were flanked by mobile-element genes.
In addition, several clusters of genes that impart resistance to heavy metals, quaternary ammonium compounds, organic hydroperoxide, and organic solvents were also identified, as well as genes for several putative multidrug efflux systems, including members of the RND, ABC, MATE, and MFS superfamilies (Table 3).

Other dispensable GRs.

In addition to the genomic features associated with virulence or survival in the clinical setting, there were a number of other dispensable GRs in A. hydrophila and A. dhakensis genomes (see Table S1 in the supplemental material). A number of these features were known operons involved in the catabolism or transport of specific substrates, such as N-acetylmuramic acid, N-acetylgalactosamine, thiamine, malate, arabinose, dl-lactate, l-cystine, taurine, sucrose, tungstate, N-ribosylnicotinamide, tetrathionate, phosphonate, and l-fucose (see Table S1). However, many of the dispensable features contained mostly hypothetical-protein-encoding genes, with perhaps one or two otherwise annotated genes. In general, the distribution of many of these dispensable genomic features correlated with species delineation. For example, many GRs were not present in the three A. caviae strains. Additionally, each genome contained a number of unique and, in rare cases, homologous prophage and prophage-like elements (data not shown). Of interest was the presence of a small integron, 8 to 13 kb, in the genomes of the four NF A. hydrophila strains, E1, and A. jandaei Riv2, of the zona occludens toxin prophage type (see Table S1).

In vitro evaluation of virulence characteristics.

We focused our studies on NF-causing strains (NF1 and NF2) of Aeromonas, as the mechanisms associated with the pathogenesis of NF are poorly understood. Among the environmental isolates, we studied two that were from river water (Riv2 and Riv3) through which the NF patient acquired the infection. These strains were compared to an environmental isolate (ATCC 7966T) of A. hydrophila (as a negative control), the genome of which we first sequenced and annotated (40), and that of highly virulent A. dhakensis SSU, whose sequence was annotated during this study. Our purpose was to discern if there were any pathogenic features unique to these strains, especially those associated with NF. A. hydrophila WI was chosen because of its resistance to multiple antibiotics.

Biofilm formation.

As shown in Fig. 3, the NF1 and SSU isolates formed comparable and significant biofilms. However, all of the other isolates produced statistically significantly lesser biofilms, which were minimal with the ATCC 7966T and Riv3 isolates.
FIG 3 Biofilm formation by Aeromonas strains was quantified on polystyrene plastic after 24 h of incubation at 37°C by CV staining. Three independent experiments were performed, and the arithmetic means ± standard deviations were plotted. An asterisk indicates a P value of ≤0.0001 as determined by one-way ANOVA. OD, optical density.

Swimming and swarming motility.

All of the isolates tested exhibited swimming motility (Fig. 4), albeit to various degrees. Among these isolates, NF1 and Riv2 showed a hypermotility phenotype compared to SSU. Both ATCC 7966T and WI were minimally motile. To more accurately measure the migration zone sizes of hypermotile strains (NF1 and Riv2), we used larger (100-mm instead of 60-mm) swimming agar plates. Our data indicated that while the swimming motility of SSU remained unaltered, one could clearly visualize the hypermotility phenotype of the NF1 and Riv2 isolates because of the larger surface area available to them for swimming (data not shown).
FIG 4 Swimming motility assay of various Aeromonas strains. We used 60-mm petri plates to determine swimming motility. Three independent experiments were performed, and the arithmetic means ± standard deviations were plotted. An asterisk indicates a P value of ≤0.0001 as determined by one-way ANOVA.
As noted in Fig. 5, NF1 and Riv2 swarmed to a higher degree than SSU, a difference that became statistically significant when 100-mm agar plates were used (data not shown). While ATCC 7966T, NF2, and Riv3 had moderate levels of swarming motility, WI did not swarm.
FIG 5 Swarming motility assay of various Aeromonas strains. We used 60-mm petri plates to determine swarming motility. Three independent experiments were performed, and the arithmetic means ± standard deviations were plotted. An asterisk indicates a P value of ≤0.0001 as determined by one-way ANOVA.

Protease and hemolytic activities.

As shown in Fig. S5 in the supplemental material, all of the Aeromonas strains studied produced a protease(s) to various degrees; however, no statistically significant difference between them and SSU was found. In terms of hemolytic activity associated with Act, we noted that the hemoglobin release from rabbit erythrocytes caused by NF2, WI, Riv2, Riv3, and ATCC 7966T was at a level much lower than that of SSU and NF1 (see Table S2 in the supplemental material). To demonstrate that most this hemolytic activity was associated with Act, we neutralized the toxin by using anti-Act polyclonal antibodies. Indeed, 75 to 98% of the Act-associated hemolytic activity of the bacterial strains tested was neutralized by these antibodies (see Table S2).

AHL production.

As shown in Fig. S6A in the supplemental material, all of the Aeromonas isolates (clinical versus water) produced similar levels of AHLs within 24 h, except for ATCC 7966T, as judged by the production of purple pigment by C. violaceum CV026 (73) (see Fig. S6B). We used an AHL-negative mutant of SSU (ΔahyRI) (62) as a negative control in the assay.

Production and secretion of Hcp and AexU.

Hcp secretion is T6SS dependent and a reliable indicator of the presence of a functional T6SS in bacteria (74). On the basis of Western blot analysis, Hcp production and secretion were noted in A. jandaei Riv2, as well as in A. dhakensis SSU (Fig. 6A). In contrast to ATCC 7966T, which did not synthesize Hcp, isolates NF1, NF2, Riv3, and WI synthesized Hcp but were unable to secrete it into the medium (Fig. 6A). As another positive control, we used the production and secretion of Hcp by Yersinia pestis, which also suggested the functionality of the T6SS in the plague bacterium.
FIG 6 Production and secretion of Hcp, AexU, and ExoA by various strains of Aeromonas and P. aeruginosa as determined by Western blotting. While only SSU and Riv2 secreted Hcp into the supernatant, the Hcp protein was identified in the cell pellets of all of the other strains studied, except for A. hydrophila ATCC 7966T (A). Isolates SSU, NF1, and NF2 produced AexU (B), while ExoA was secreted by isolates SSU and NF2, as well as by P. aeruginosa PA103 (C). Y. pestis (YP) served as a control. The molecular sizes of Hcp, AexU, and ExoA are indicated.
The AexU toxin is a T3SS effector identified in SSU, which possesses GAP and ADP-RT activities (63, 75). Both isolates NF1 and NF2 produced AexU, as was also noted in SSU (Fig. 6B). None of the other strains tested produced AexU.

Existence of ExoA homolog.

On the basis of genome-wide sequencing, a homolog of P. aeruginosa ExoA was found in SSU and NF2 isolates. Comparison of the DNA sequence of the exoA gene of P. aeruginosa with that of Aeromonas species revealed 70% homology. At the amino acid level, the sequence identity ranged from 64 to 65%. A sequence identity of 97 to 98% was noted for isolates SSU, NF2, and E1. To demonstrate that the exoA gene in the Aeromonas strains tested was expressed, we performed Western blot analysis with antibodies to P. aeruginosa ExoA. As noted in Fig. 6C, a secreted version of ExoA was seen in SSU and NF2 isolates. As a positive control, we used P. aeruginosa PA103, which also demonstrated a correct-size ExoA band of 66 kDa (76).
By analyzing the ExoA protein sequences at the NCBI conserved-domain database, a Pfam protein sequence search (, and SMART analysis (, it was revealed that the ExoA sequence from aeromonads possessed three domains similar to that found in P. aeruginosa. The ExoA sequences of SSU, NF2, and E1 had 639 amino acid residues. Interestingly also, the amino acid sequences of ExoA homologs in more recently sequenced A. hydrophila genomes, according to the NCBI database, were shorter than that of ExoA of SSU. Prediction of signal peptides in Archaea with the hidden Markov model revealed that the first 25 amino acid residues represent the leader peptide in SSU, NF2, and E1, similar to that seen in ExoA of P. aeruginosa PA103 (76).

Virulence associated with clinical and environmental aeromonads in a mouse model of infection. (i) Septicemic-mouse model.

To test the virulence of aeromonads, animals were injected with various A. hydrophila strains (SSU, ATCC 7966T, NF1, NF2, WI, and Riv3) or A. jandaei Riv2 via the i.p. route at a dose of 5 × 107 CFU. As shown in Fig. 7, 100% of the animals infected with A. dhakensis SSU and 90% of the mice challenged with NF1 or Riv2 died within 2 days p.i. The mortality rates were 35 and 45%, respectively, when the mice were infected with NF2 and Riv3 at the same infectious dose. Both strains ATCC 7966T and WI of A. hydrophila were the least virulent in a septicemic-mouse model, allowing the survival of 90 to 100% of the animals.
FIG 7 Virulence of various Aeromonas strains in a septicemic-mouse model of infection. Swiss-Webster mice (n = 6 to 23) were infected with a dose of 5 × 107 CFU of the strains shown by the i.p. route. The animals were monitored for death for 14 days. The data were statistically analyzed by using the Kaplan-Meier survival estimate, and the actual P values are presented. The number of animals used for each strain is shown in parentheses.

(ii) Developing NF by Aeromonas species in a mouse model.

Previously, the subcutaneous injection route was used on the footpads of humanized CD46 transgenic mice to induce Streptococcus pyogenes-associated NF (77). However, Aeromonas-related NF has always been linked to infections acquired through nonintact skin and exposure of wounds or skin lacerations to contaminated soil or water (9, 78, 79). To mimic this infection route, we challenged mice i.m. with strains SSU, NF1, NF2, and Riv2 or a mixture of NF1 and NF2 and used strain ATCC 7966T as a negative control. As depicted in Fig. 8A, A. jandaei Riv2 was highly virulent, resulting in 100% mortality in mice within 24 h p.i. at a dose of 5 × 108 CFU. Isolates SSU and NF2 also resulted in 100% mortality but at 48 h p.i.
FIG 8 Percent survival of mice after infection with various Aeromonas isolates by the i.m. route. (A) Animals were infected with isolate SSU, NF1, NF2, Riv2, or ATCC 7966T at 5 × 108 CFU/mouse and monitored for death for 14 days. (B) Animals were infected with NF1 and NF2 at the lower doses indicated or with a mixture of these two isolates at 1 × 108 CFU each per mouse. Isolates SSU, NF1, NF2, and Riv2 were statistically significantly more virulent than avirulent A. hydrophila strain ATCC 7966T (A). Likewise, NF2 at both the doses was statistically significantly more virulent than NF1 (B). The data were statistically analyzed by using Kaplan-Meier survival estimates, and the actual P values are presented.
Interesting were our findings that while NF2 was still 100% lethal to mice at a lower dose of 5 × 107 CFU by the i.m. route (data not shown), 65% of the animals survived when injected by the i.p. route with NF2 (Fig. 7). On the other hand, strain NF1, which was relatively more virulent than strain NF2 in a septicemic-mouse model of infection (Fig. 7), with 10% versus 65% survival, exhibited greatly reduced virulence (100% survival) when injected via the i.m. route at the same dose of 5 × 107 CFU (data not shown). At a higher dose of 5 × 108 CFU, animals infected with NF1 also succumbed to infection by day 6 (Fig. 8A). As expected, none of the animals infected with strain ATCC 7966T died.
As shown in Fig. 8B, while NF2 was still highly virulent (100% mortality) at doses of 1 × 108 to 2 × 108 CFU when given by the i.m. route, NF1 resulted in death in only 40% of the mice infected with a dose of 2 × 108 CFU. Since the NF1 and NF2 isolates were obtained at different times (NF1 from the initial wound site and NF2 after amputation) from the same patient, we determined whether mixing these two cultures in equal numbers could augment their virulence in a mouse model. Our data indicated that the virulence of NF2 predominated in the mixed culture, as 80% of the animals died by day 2 (Fig. 8B).

Dissemination of Aeromonas species after i.m. injection in a mouse model.

To evaluate the dissemination pattern of various Aeromonas isolates, infected mice (n = 5) were sacrificed at 24 or 48 h p.i. and their spleens and livers were subjected to CFU determination. In agreement with the animal survival data, greater bacterial loads were noted in the livers and spleens of mice in those groups that were infected with SSU (1 × 108 CFU), Riv2 (1 × 108 CFU), and NF2 (2 × 108 CFU) (Fig. 9A to C). The number of bacteria in these organs increased over the course of infection from 24 to 48 h. Because of the highly lethal nature of the Riv2 isolate, some of the animals died between 24 to 48 h after infection (Fig. 9B). Strain NF1 disseminated to the liver and spleen in lower numbers, even at a slightly higher infective dose (5 × 108 CFU) (Fig. 9D), than strains SSU, NF2, and Riv2 (Fig. 9A to C), with strain ATCC 7966T disseminating minimally to the livers and spleens of mice (1/5) over the course of 48 h of infection (Fig. 9E).
FIG 9 Bacterial dissemination patterns for various Aeromonas isolates in a mouse model of NF. Mice (n = 5) were infected by the i.m. route with the dose of isolate SSU (A), Riv2 (B), NF2 (C), NF1 (D), or ATCC 7966T (E) indicated. At 24 or 48 h p.i., the spleen and liver were aseptically removed from each mouse, homogenized, and subjected to bacterial colony counting. M1 to M5 represent individual mice, and some groups have fewer animals because of deaths p.i.

Gross and histopathological evaluations of tissues from Aeromonas species-associated NF.

We initially focused on two isolates, namely, SSU and NF1. At 24 h p.i. with isolate SSU (5 × 108 CFU), muscle tissues around the injection site revealed severe edematous and hemorrhagic swelling, unlike those of animals injected with PBS or lipopolysaccharide (20 μg) (Fig. 10A). These inflammatory changes were more progressive in nature at 48 h p.i. (Fig. 10B1 and B2). In addition, the leg tissues of mice infected with the NF1 isolate (2 × 108 CFU) were liquefied and showed gangrenous necrosis after 7 days, as shown in Fig. 10C.
FIG 10
FIG 10 Gross pathological examination of leg muscle tissues after infection with A. dhakensis SSU and A. hydrophila NF1. At 24 (A) or 48 (B1 and B2) h p.i., leg tissues from animals infected with isolate SSU at 5 × 108 CFU/mouse via the i.m. route were examined macroscopically. Note that the B2 animal expired at the time of examination. The other leg of the animal was given either PBS or LPS. Also, at 7 days p.i. (C), the animals that survived infection with strain NF1 at 2 × 108 CFU/mouse were sacrificed and the leg tissues were grossly observed for lesions. Only representative animals are shown.
Histopathological examination of skeletal muscle tissues from thigh regions adjacent to the injection site revealed various levels of multifocal and diffuse necrotic changes in striated muscle bundles, accompanied by edematous exudation and infiltration of polymorphonuclear cells (PMNs), specifically, neutrophils (Fig. 11A). Specifically, tissues collected from animals infected with strains ATCC 7966T and NF1 showed more PMN infiltration and edematic fluid but limited multifocal necrotic changes (Fig. 11A). On the other hand, tissues from animals infected with strains SSU and Riv2 exhibited extensive and diffuse necrotic changes with moderate levels of edema and infiltration of PMNs (Fig. 11A). Since strain NF2 was highly lethal to mice by the i.m. route, we reduced the infectious dose to 5 × 106 to 5 × 107 CFU. These tissues also revealed various levels of multifocal and local, as well as diffuse, necrotic changes in the skeletal muscle. In addition, infiltration of PMNs (specifically, neutrophils) and various severities of edematic fluid accumulation were noticed. As depicted in Fig. 11B, necrotic and degenerated myofibers could be seen (arrow) with significant inflammatory and edematous responses (asterisk) in animals infected with NF2.
FIG 11
FIG 11 Histopathological examination of mouse leg muscle tissues after infection by the i.m. route with various Aeromonas isolates. (A) Leg muscle tissues collected at 48 h p.i. from mice injected with strain ATCC 7966T, NF1, SSU, or Riv2 (at a dose of 5 × 108 CFU/mouse for ATCC 7966T and NF1 or 1 × 108 for SSU and Riv2) or with PBS (negative control) were processed for histopathological analysis and observed under a microscope after staining with H&E. A star symbols denote infiltration of PMNs, while squares, rectangles, and long or short arrows show necrotic lesions in skeletal muscle tissues with loss of characteristic striations. Triangles indicate edematous swelling. Each of these images represents a group of five samples. (B) Animals were infected with highly virulent isolate NF2 at 5 × 106 or 1 × 107 CFU/mouse. An arrow indicates skeletal muscle necrosis, and an asterisk indicates infiltration of PMNs. An image representative of five mice is shown. Magnifications (for panels A and B), 100× (top rows) and 400× (bottom rows).

Dissemination of bioluminescent bacteria in a mouse model of NF.

We infected mice (n = 10) by the i.m. route in the right hind leg with a challenge dose of 5 × 108 CFU of each of the above-mentioned strains, except Riv2. Since Riv2 exhibited higher virulence than the other strains tested, two lower infectious doses (5 × 107 and 1 × 108 CFU) were used. After infection, we measured bioluminescence in the legs, liver, and spleen. In addition, we determined bacterial counts in the livers and spleens of five animals per time point at 24 and 48 h p.i.
At 24 h p.i. for no strain SSU bioluminescence was observed beyond the baseline level, while two animals were positive for bacterial colonies in the liver and one was positive for bacterial colonies in the spleen (Fig. 12A). It should be noted that for animals to show bioluminescence, the number of organisms should be ∼105 CFU (68). At 48 h p.i., three mice were positive for bioluminescence in the leg at the injection site and two had luminescence extending into the peritoneal cavity. The bioluminescence observed also corresponded to the bacterial counts in the liver and spleen (Fig. 12A). Both of the animals that had strong luminescence in the peritoneal cavity showed bacterial counts in the range of 106 to 107 CFU in the liver and spleen. For each of the other strains (NF1, ATCC 7966T, and Riv2), strong bioluminescence was observed in the infected leg at 24 h p.i. (Fig. 12B to E), with the exception of one mouse in the ATCC 7966T group (Fig. 12C).
FIG 12
FIG 12 Bioluminescent imaging and corresponding bacterial loads of harvested livers and spleens infected with luminescent Aeromonas strains SSU (A), NF1 (B), ATCC 7966T (C) Riv2 (5 × 107 CFU) (D), and Riv2 (1 × 108 CFU) (E). Infected mice (n = 5) were imaged at 24 and 48 h p.i. and used for colony counting. For isolates SSU, NF1, and ATCC 7966T, the challenge dose was 5 × 108 CFU/mouse by the i.m. route. The CFU are reported for whole organs derived from colony plate counts.
At 48 h p.i., a decrease in luminescence was observed in the legs across all of the mice infected with strain ATCC 7966T, with a majority of the animals (4/5) having dissemination into the liver but remaining below 105 CFU. No bacteria were detected in the spleen (Fig. 12C). With strain NF1, luminescence was observed in all of the mice at both the 24- and 48-h time points, with dissemination of bacteria to the spleen and liver; however, luminescence was absent from the peritoneal cavity (Fig. 12B). In mice infected with strain Riv2 at a dose of 5 × 107 CFU, luminescence was observed at the injection site in the hind leg, as well as extending into the peritoneal cavity of one mouse, corresponding to viable plate counts of ∼107 CFU in the liver and spleen at 48 h (Fig. 12D). In mice infected with 1 × 108 CFU of Riv2, strong luminescence was observed at the injection site, as well as extending into the peritoneal cavities of two mice at 48 h, representing disseminating infection even at the lower dose (Fig. 12E). At both doses of Riv2, one mouse died by 24 h and two additional animals died at 48 h and were not included for imaging.


In this study, genetically diverse isolates of Aeromonas (belonging to A. hydrophila, A. jandaei, and A. caviae) were sequenced and analyzed to identify virulence factors associated with their distinct pathotypes. Our purpose was to discern if there were any pathogenic features unique to these strains, especially those associated with NF. NF is part of a family of diseases referred to as necrotizing soft tissue infections (NSTIs). Even milder NSTIs often require several surgical debridement (80). There are three major classifications of NSTIs, type I, involving mixed infections of anaerobes and aerobic bacteria (polymicrobial), type II, involving S. pyogenes, and type III, gas gangrene or clostridial myonecrosis, with all three requiring intensive care and rapid treatment. A variant of type I is salt water NF, in which an apparently minor skin wound is contaminated with salt water containing a Vibrio species. Even with rapid surgical intervention (81, 82), case fatalities can be as high as 50% (83) and leave surviving patients with lifelong disabilities and disfigurement (81, 82, 84).
Whole-genome sequencing of isolates NF1 to NF4 from a patient revealed that NF1 was genetically distinct and that strains NF2 to NF4 were not true clones in the sense that they likely did not derive from a last common ancestor either during the course of infection or in the environment directly prior to infection. Comparative genomic analysis, coupled with functional assays, of strains NF1 and NF2 and other clinical and environmental Aeromonas strains allowed us to investigate pathovar-specific virulence factors or markers.
Although isolates Riv2 and Riv3 were cultured from the same water source through which the patient with NF contracted the infection, they were genetically different (e.g., Riv3) from isolates NF1 to NF4 of A. hydrophila. We believe that since these water samples were collected at a later time subsequent to the infection of the patient, water microbiota changed because of fluctuations in the environmental conditions (85, 86). Nevertheless, our studies reinforce the idea that both benign and highly virulent strains can be present at all times in fresh and brackish waters, where aeromonads flourish. In fact, the advent of Aeromonas species as a cause of NF leads us to suggest that this condition is another variant of type 1, which can be termed “estuarine fasciitis.”
In the same vein, recent studies of Skwor et al. (85) recovered tetracycline- and ciprofloxacin-resistant Aeromonas strains from Lake Erie in Pennsylvania. Importantly, the prevalence of aerolysin and serine protease-encoding genes and their corresponding hemolytic and proteolytic/cytotoxic activities was also high in antibiotic-resistant strains of some Aeromonas species (e.g., A. hydrophila and A. veronii), suggesting that they contribute to bacterial virulence (85). Since the antibiotics of choice for the treatment of aeromonads are sulfamethoxazole-trimethoprim and ciprofloxacin (87), the presence of antibiotic-resistant Aeromonas strains in natural water becomes even more significant in the management of serious infections such as NF.
Aeromonads readily form biofilms; however, it is unclear whether there are any variations in biofilm formation among clonally diverse clinical and environmental isolates. The dynamics of biofilm formation depend on the environmental niches where the microbes hide during an infection and dictate the pathogenic nature of a given organism by modulating virulence gene expression (88). In this vein, we described the dynamics of mature, three-dimensional biofilm formation by A. dhakensis strain SSU (89). Indeed, highly virulent isolates NF1 and Riv2 formed biofilms comparable to that of SSU, and these likewise were better swimmers and swarmers, which overall contributed to their virulence (Fig. 3 to 5).
The AI-1-based QS system consists of the ahyRI genes, with ahyR encoding a transcription activator and the ahyI gene synthesizing an autoinducer AHL. The role of this system in biofilm regulation has been well established (27, 62, 90). However, in the present study, no direct correlation between AHL production and biofilm formation in the strains studied was noted (Fig. 3; see Fig. S6 in the supplemental material). It is plausible that in some Aeromonas strains, the regulation of biofilm formation is very tightly regulated, as dictated by specific environmental and nutritional niches found only under in vivo conditions and needing further exploration. Indeed, the expression of the locus that codes for lateral flagella and allows bacteria to swarm is linked to biofilm formation (91). Further, overlapping sets of regulators govern swarming motility and biofilm formation in a reciprocal fashion, depending upon the environmental cues that challenge microbes during an infection (92, 93).
In P. aeruginosa, it was noted that swarming isolates had diminished biofilm formation capabilities and that swarming motility occurred in the absence of either twitching or swimming motility. Further, the best swarmers secreted increased levels of proteases (94). On the contrary, the best E. coli biofilm formers displayed the greatest swimming motility (95). The above-mentioned scenario with E. coli seemed to fit with the SSU, NF1, and Riv2 isolates, as they exhibited greater swarming and swimming motilities (Fig. 4 and 5), formed better biofilms (Fig. 3), and were highly virulent in mouse models of infection (Fig. 7 and 8). However, on the basis of this study, no correlation seemed to exist between swarming motility (Fig. 5) and protease production (see Fig. S5 in the supplemental material).
The high hemolytic activity associated with Act, as well as the presence of the T3SS effector AexU also correlated with increased virulence of SSU and NF1 in mouse models of infection (Fig. 6 to 8; see Table S2 in the supplemental material). These results were corroborated by our earlier findings indicating that deletion of the act and aexU genes from A. dhakensis SSU significantly reduced their respective virulence in the septicemic-mouse model of infection (61, 63, 75).
Interesting, however, were our findings that although Riv2 and NF2 were highly virulent in one or both in vivo models of bacterial infection (Fig. 7 and 8), they produced low levels of Act (see Table S2 in the supplemental material). For NF2, it is likely related to the truncated nature of the act gene, while in the Riv2 isolate, the toxin gene might be regulated differently from that of isolates SSU and NF1. These isolates were still highly virulent in an NF mouse model (Fig. 8), despite low Act-associated hemolytic activity. This could be explained by the presence of a functional ExoA in NF2 and secretion of T6SS-associated Hcp or differences in T3SS effector proteins in Riv2, which could compensate for low Act-associated hemolytic activity (Fig. 6). The reverse seemed to be true for the NF1 isolate, as it had high hemolytic activity and produced AexU but did not contain the exoA gene or secrete Hcp. In this regard, isolate SSU had all of the weapons (e.g., the presence of Act, ExoA, AexU, and Hcp) (Fig. 6) necessary to inflict serious damage on the host.
The reduced hemolytic activity of Act in the WI isolate could be the result of an insertion of a putative arylsulfate-encoding gene upstream of the act gene that might have affected the expression of the gene (Fig. 2A). Finally, it is not clear why Riv3 had Act toxin-neutralizable hemolytic activity (see Table S2 in the supplemental material), as the genome of this isolate did not harbor the act gene. One plausible explanation is that the hemolytic activity seen in this isolate is contributed by another hemolysin for which we might have antibodies in our Act antitoxin preparation. Another possibility is that there might be a neutralizable cross-reactive epitope(s) among different hemolysins. Since the WI isolate did not secrete Hcp or produce AexU and ExoA, it was the least virulent in a mouse model of infection. The same logic applies to strain ATCC 7966T (Fig. 6 and 7).
Our earlier study indicated the presence of an exoA gene in the genome of wound isolate E1 (42); however, the functionality of the gene was not confirmed. Indeed, ExoA was secreted by the SSU, NF2, and PA103 isolates (Fig. 6), as well as by E1 (data not shown). While secretion of ExoA was demonstrated in the SSU and NF2 isolates when they were grown in LB medium at 37°C, ExoA was secreted from the E1 isolate at a low iron concentration of 0.05 μg/ml at 32°C (65). The low level of ExoA secretion detected in PA103 was probably related to the complex nutritional conditions needed for its synthesis (96).
In a mouse model of NF, the pathodymanics of lesion development by isolates SSU, NF1, NF2, and Riv2 (Fig. 11) closely mimicked the time course events reported for human NF. When NSTIs were first characterized, an overabundance of neutrophils at the infection site was recognized as a hallmark of the disease (97). In the mouse model of NF, we also noted abundant neutrophils at the infection site (Fig. 11). Therefore, it is plausible that neutrophils might have a role in the pathogenesis of NSTIs and in the extensive tissue damage that is characteristic of the disease.
The mechanism and significance of enhanced neutrophil recruitment observed in NSTIs are unknown. However, increases in the level of resistin, a proinflammatory peptide derived from neutrophils, have been correlated with disease severity in both NSTIs and septic shock (98). It is known that degranulation events in neutrophils lead to the release of azurophilic and specific granules into the extracellular space, exposing the surrounding tissue to toxic enzymes and antimicrobial peptides (99).
Strain NF2 was less lethal to mice when given by the i.p. route but was highly virulent when injected by the i.m. route (Fig. 7 and 8). These data raised the possibility that the microenvironment prevailing at the i.m. injection site would favor efficient adaptation of the bacteria to the host immune system and thus later led to systemic infection because of the presence of unique virulence factors. On the other hand, A. hydrophia strain ATCC 7966T, which lacks a T3SS and many other virulence factors (e.g., Hcp, ExoA) (40), remained avirulent in mice although the site of i.m. injection showed infiltration of PMNs (Fig. 11).
Substantiating the above-mentioned paradigm, the animal mortality pattern caused by isolates SSU, NF1, NF2, and Riv2 (Fig. 8) closely matched their dissemination profiles (Fig. 9) and correlated with the presence of key virulence factors. As alluded to above, a hallmark characteristic of NF is the dissemination of bacterial infection along fascial planes of the limbs, followed by, in a setting of severe infection, sepsis (81). To study this phenomenon in more detail, we used various bioluminescent reporter Aeromonas strains. In most of the strains tested, we observed that infection was limited to the hind leg, the site of bacterial injection, although dissemination and colonization of that region were extensive, with further dissemination of some strains (e.g., SSU, Riv2) to other organs occurring by 48 h. Unfortunately, since NF2 was not highly virulent in a septicemic-mouse model of infection, we did not construct this reporter strain. However, significant dissemination of this isolate to other organs was revealed by plate counting (Fig. 9C).
A. hydrophila ATCC 7966T had no lethality in this mouse model, and the infection was confined to the injection site, with little or no dissemination, even within the region of infection (Fig. 12C). In both strains SSU and Riv2, dissemination above the threshold for imaging was noted, which was confirmed by viable plate counts (Fig. 12A, B, D, and E).
In contrast, mice infected with NF1 showed more uniform dissemination, albeit at lower levels at 24 and 48 h (Fig. 12B), but this strain exhibited larger necrotic lesions in the infected legs of mice. Despite being isolated directly from the initial wound site from a patient with NF, the NF1 isolate seemed to show a lower propensity to disseminate and grow rapidly outside the fascial tissue of the infected leg (Fig. 12B). In fact, infection seemed to be confined to the leg, and only after an aggressive necrotic lesion was formed did the infection spread significantly (data not shown). One week p.i., several of the mice infected with NF1 developed lesions that encompassed the entire leg (Fig. 10C) before advancing across the peritoneum. Histologically, strain NF1 also seemed to localize at the injection site and thus to favor the infiltration of PMNs, eventually leading to localized necrotic changes rather than causing systemic dissemination (Fig. 11).
On the basis of the genomic and functional data presented in this paper, the role of virulence factors such as Act, ExoA, and T3SS and T6SS and their effectors and the functionality of the lateral flagella in causing NF are evident. However, one should be cautious in data interpretation, as the unique environment cues the aeromonads are exposed to during NF may trigger bacterial genes differently to contribute to the severity of the disease. While we did not identify a single, highly unique virulence factor in the NF-associated strains, nevertheless, it was quite evident that these strains were extremely virulent and were representative of many clinical strains described in our previous studies (33, 42).
Our findings are consistent with previous suggestions that there are particular virulent subsets of pathotypes within the A. hydrophila group. Recent findings obtained with the most prevalent NF-causing organism, GAS, suggested that NF is the by-product of superantigen genes responsible for toxins that are capable of bypassing normal intracellular antigen processing and presentation by major histocompatibility complex class II (100). In the absence of evidence of these or similar genes in Aeromonas, we propose that Aeromonas-related NF is the result of either a triad or quartet of pathovar-specific virulence factors that result in NF primarily in susceptible individuals. In sum, it appears that secretion/translocation of multiple toxins via different secretion systems along with other virulence mechanisms as summarized in Table 4 could be responsible for causing NF by aeromonads and needs further investigation.
TABLE 4 Virulence characteristics of various Aeromonas strains
StrainBiofilm formationMotilityHcp (T6SS)AexU (T3SS)ExoACytotoxic enterotoxin (Act)aVirulence in mice
A. dhakensis SSUHighHighHigh++++HighHighHigh
A. hydrophila ATCC 7966TLowModerateModerateModerateLowLow
A. hydrophila NF1HighHypermotileHypermotile++HighHighHigh
A. hydrophila NF2ModerateModerateModerate+++LowModerateHigh
A. hydrophila WIModerateLowLow+NDbModerateLowND
A. jandaei Riv2ModerateHypermotileHypermotile++LowHighHigh
A. hydrophila Riv3LowModerateModerate+NDModerateModerateND
Hemolytic activity associated with Act.
ND, not determined.
While this hypothesis remains to be proven, future studies on Aeromonas NF cases will provide information regarding the molecular and genetic features of these organisms. Studies such as those performed with GAS should be instrumental in revealing the intricacies of Aeromonas NF in vulnerable hosts.


The financial support provided to A.K.C. through Leon Bromberg and Robert E. Shope and John S. Dunn Distinguished Chair in Global Health endowments, UTMB, is gratefully acknowledged.
We thank the Genomics Core (specifically, Tom Wood and Steven Widen), UTMB, for sequencing several of the aeromonad genomes.

Supplemental Material

File (zam999105475so1.pdf)
ASM does not own the copyrights to Supplemental Material that may be linked to, or accessed through, an article. The authors have granted ASM a non-exclusive, world-wide license to publish the Supplemental Material files. Please contact the corresponding author directly for reuse.


Von Graevenitz A and Mensch AH. 1968. The genus Aeromonas in human bacteriology—report of 30 cases and review of the literature. N. Engl. J. Med. 278:245–249.
Presley SM, Rainwater TR, Austin GP, Platt SG, Zak JC, Cobb GP, Marsland EJ, Tian K, Zhang B, Anderson TA, Cox SB, Abel MT, Leftwich BD, Huddleston JR, Jeter RM, and Kendall RJ. 2006. Assessment of pathogens and toxicants in New Orleans, LA following Hurricane Katrina. Environ. Sci. Technol. 40:468–474.
Chen PL, Wu CJ, Chen CS, Tsai PJ, Tang HJ, and Ko WC. 16 November 2013. A comparative study of clinical Aeromonas dhakensis and Aeromonas hydrophila isolates in southern Taiwan: A. dhakensis is more predominant and virulent. Clin. Microbiol. Infect..
Palú AP, Gomes LM, Miguel MA, Balassiano IT, Queiroz ML, Freitas-Almeida AC, and de Oliveira SS. 2006. Antimicrobial resistance in food and clinical Aeromonas isolates. Food Microbiol. 23:504–509.
Chauret C, Volk C, Creason R, Jarosh J, Robinson J, and Warnes C. 2001. Detection of Aeromonas hydrophila in a drinking-water distribution system: a field and pilot study. Can. J. Microbiol. 47:782–786.
Hiransuthikul N, Tantisiriwat W, Lertutsahakul K, Vibhagool A, and Boonma P. 2005. Skin and soft-tissue infections among tsunami survivors in southern Thailand. Clin. Infect. Dis. 41:e93–96.
Figueras MJ, Aldea MJ, Fernandez N, Aspiroz C, Alperi A, and Guarro J. 2007. Aeromonas hemolytic-uremic syndrome. A case and a review of the literature. Diagn. Microbiol. Infect. Dis. 58:231–234.
Figueras MJ. 2005. Clinical relevance of Aeromonas sM503. Rev. Med. Microbiol. 16:145–153.
Monaghan SF, Anjaria D, Mohr A, and Livingston DH. 2008. Necrotizing fasciitis and sepsis caused by Aeromonas hydrophila after crush injury of the lower extremity. Surg. Infect. 9:459–467.
Tsai YH, Huang KC, Huang TJ, and Hsu RW. 2009. Case reports: fatal necrotizing fasciitis caused by Aeromonas sobria in two diabetic patients. Clin. Orthop. Relat. Res. 467:846–849.
Cui H, Hao S, and Arous E. 2007. A distinct cause of necrotizing fasciitis: Aeromonas veronii biovar sobria. Surg. Infect. 8:523–528.
Furusu A, Yoshizuka N, Abe K, Sasaki O, Miyazaki K, Miyazaki M, Hirakata Y, Ozono Y, Harada T, and Kohno S. 1997. Aeromonas hydrophila necrotizing fasciitis and gas gangrene in a diabetic patient on haemodialysis. Nephrol. Dial. Transplant. 12:1730–1734.
Joseph SW, Daily OP, Hunt WS, Seidler RJ, Allen DA, and Colwell RR. 1979. Aeromonas primary wound infection of a diver in polluted waters. J. Clin. Microbiol. 10:46–49.
Sever R, Lee Goldstein A, Steinberg E, and Soffer D. 2013. Trauma with a touch of fresh water: necrotizing fasciitis caused by Aeromonas hydrophilia after a motorcycle accident. Am. Surg. 79:E326–E328.
Xu XJ, Ferguson MR, Popov VL, Houston CW, Peterson JW, and Chopra AK. 1998. Role of a cytotoxic enterotoxin in Aeromonas-mediated infections: development of transposon and isogenic mutants. Infect. Immun. 66:3501–3509.
Ferguson MR, Xu XJ, Houston CW, Peterson JW, Coppenhaver DH, Popov VL, and Chopra AK. 1997. Hyperproduction, purification, and mechanism of action of the cytotoxic enterotoxin produced by Aeromonas hydrophila. Infect. Immun. 65:4299–4308.
Galindo CL, Fadl AA, Sha J, Pillai L, Gutierrez C Jr, and Chopra AK. 2005. Microarray and proteomics analyses of human intestinal epithelial cells treated with the Aeromonas hydrophila cytotoxic enterotoxin. Infect. Immun. 73:2628–2643.
Galindo CL, Gutierrez C Jr, and Chopra AK. 2006. Potential involvement of galectin-3 and SNAP23 in Aeromonas hydrophila cytotoxic enterotoxin-induced host cell apoptosis. Microb. Pathog. 40:56–68.
Ribardo DA, Kuhl KR, Boldogh I, Peterson JW, Houston CW, and Chopra AK. 2002. Early cell signaling by the cytotoxic enterotoxin of Aeromonas hydrophila in macrophages. Microb. Pathog. 32:149–163.
Galindo CL, Fadl AA, Sha J, Gutierrez C Jr, Popov VL, Boldogh I, Aggarwal BB, and Chopra AK. 2004. Aeromonas hydrophila cytotoxic enterotoxin activates mitogen-activated protein kinases and induces apoptosis in murine macrophages and human intestinal epithelial cells. J. Biol. Chem. 279:37597–37612.
Rosenzweig JA and Chopra AK. 2013. The exoribonuclease polynucleotide phosphorylase influences the virulence and stress responses of yersiniae and many other pathogens. Front. Cell. Infect. Microbiol. 3:81.
Sha J, Wang SF, Suarez G, Sierra JC, Fadl AA, Erova TE, Foltz SM, Khajanchi BK, Silver A, Graf J, Schein CH, and Chopra AK. 2007. Further characterization of a type III secretion system (T3SS) and of a new effector protein from a clinical isolate of Aeromonas hydrophila—part I. Microb. Pathog. 43:127–146.
Suarez G, Sierra JC, Sha J, Wang S, Erova TE, Fadl AA, Foltz SM, Horneman AJ, and Chopra AK. 2008. Molecular characterization of a functional type VI secretion system from a clinical isolate of Aeromonas hydrophila. Microb. Pathog. 44:344–361.
Suarez G, Sierra JC, Erova TE, Sha J, Horneman AJ, and Chopra AK. 2010. A type VI secretion system effector protein, VgrG1, from Aeromonas hydrophila that induces host cell toxicity by ADP ribosylation of actin. J. Bacteriol. 192:155–168.
Beaz-Hidalgo R, Martinez-Murcia A, and Figueras MJ. 2013. Reclassification of Aeromonas hydrophila subsp. dhakensis Huys et al. 2002 and Aeromonas aquariorum Martinez-Murcia et al. 2008. as Aeromonas dhakensis sp. nov. comb nov. and emendation of the species Aeromonas hydrophila. Syst. Appl. Microbiol. 36:171–176.
Suarez G, Sierra JC, Kirtley ML, and Chopra AK. 2010. Role of Hcp, a type 6 secretion system effector, of Aeromonas hydrophila in modulating activation of host immune cells. Microbiology 156:3678–3688.
Swift S, Karlyshev AV, Fish L, Durant EL, Winson MK, Chhabra SR, Williams P, Macintyre S, and Stewart GS. 1997. Quorum sensing in Aeromonas hydrophila and Aeromonas salmonicida: identification of the LuxRI homologs AhyRI and AsaRI and their cognate N-acylhomoserine lactone signal molecules. J. Bacteriol. 179:5271–5281.
Kozlova EV, Khajanchi BK, Sha J, and Chopra AK. 2011. Quorum sensing and c-di-GMP-dependent alterations in gene transcripts and virulence-associated phenotypes in a clinical isolate of Aeromonas hydrophila. Microb. Pathog. 50:213–223.
Kirov SM, Castrisios M, and Shaw JG. 2004. Aeromonas flagella (polar and lateral) are enterocyte adhesins that contribute to biofilm formation on surfaces. Infect. Immun. 72:1939–1945.
Kirov SM, Tassell BC, Semmler AB, O'Donovan LA, Rabaan AA, and Shaw JG. 2002. Lateral flagella and swarming motility in Aeromonas species. J. Bacteriol. 184:547–555.
Ramachandran G, Tulapurkar ME, Harris KM, Arad G, Shirvan A, Shemesh R, Detolla LJ, Benazzi C, Opal SM, Kaempfer R, and Cross AS. 2013. A peptide antagonist of CD28 signaling attenuates toxic shock and necrotizing soft-tissue infection induced by Streptococcus pyogenes. J. Infect. Dis. 207:1869–1877.
Baker C and Antonovics J. 2012. Evolutionary determinants of genetic variation in susceptibility to infectious diseases in humans. PLoS One 7:e29089.
Khajanchi BK, Fadl AA, Borchardt MA, Berg RL, Horneman AJ, Stemper ME, Joseph SW, Moyer NP, Sha J, and Chopra AK. 2010. Distribution of virulence factors and molecular fingerprinting of Aeromonas species isolates from water and clinical samples: suggestive evidence of water-to-human transmission. Appl. Environ. Microbiol. 76:2313–2325.
Carnahan AM, Behram S, and Joseph SW. 1991. Aerokey II: a flexible key for identifying clinical Aeromonas species. J. Clin. Microbiol. 29:2843–2849.
Abbott SL, Cheung WK, and Janda JM. 2003. The genus Aeromonas: biochemical characteristics, atypical reactions, and phenotypic identification schemes. J. Clin. Microbiol. 41:2348–2357.
Havelaar AH, During M, and Versteegh JF. 1987. Ampicillin-dextrin agar medium for the enumeration of Aeromonas species in water by membrane filtration. J. Appl. Bacteriol. 62:279–287.
Margulies M, Egholm M, Altman WE, Attiya S, Bader JS, Bemben LA, Berka J, Braverman MS, Chen YJ, Chen Z, Dewell SB, Du L, Fierro JM, Gomes XV, Godwin BC, He W, Helgesen S, Ho CH, Irzyk GP, Jando SC, Alenquer ML, Jarvie TP, Jirage KB, Kim JB, Knight JR, Lanza JR, Leamon JH, Lefkowitz SM, Lei M, Li J, Lohman KL, Lu H, Makhijani VB, McDade KE, McKenna MP, Myers EW, Nickerson E, Nobile JR, Plant R, Puc BP, Ronan MT, Roth GT, Sarkis GJ, Simons JF, Simpson JW, Srinivasan M, Tartaro KR, Tomasz A, Vogt KA, Volkmer GA, et al. 2005. Genome sequencing in microfabricated high-density picolitre reactors. Nature 437:376–380.
Simpson JT, Wong K, Jackman SD, Schein JE, Jones SJ, and Birol I. 2009. ABySS: a parallel assembler for short read sequence data. Genome Res. 19:1117–1123.
Aziz RK, Bartels D, Best AA, DeJongh M, Disz T, Edwards RA, Formsma K, Gerdes S, Glass EM, Kubal M, Meyer F, Olsen GJ, Olson R, Osterman AL, Overbeek RA, McNeil LK, Paarmann D, Paczian T, Parrello B, Pusch GD, Reich C, Stevens R, Vassieva O, Vonstein V, Wilke A, and Zagnitko O. 2008. The RAST server: rapid annotations using subsystems technology. BMC Genomics 9:75.
Seshadri R, Joseph SW, Chopra AK, Sha J, Shaw J, Graf J, Haft D, Wu M, Ren Q, Rosovitz MJ, Madupu R, Tallon L, Kim M, Jin S, Vuong H, Stine OC, Ali A, Horneman AJ, and Heidelberg JF. 2006. Genome sequence of Aeromonas hydrophila ATCC 7966T: jack of all trades. J. Bacteriol. 188:8272–8282.
Michel C. 1979. Furunculosis of salmonids: vaccination attempts in rainbow trout (Salmo gairdneri) by formalin-killed germs. Ann. Rech. Vet. 10:33–40.
Grim CJ, Kozlova EV, Sha J, Fitts EC, van Lier CJ, Kirtley ML, Joseph SJ, Read TD, Burd EM, Tall BD, Joseph SW, Horneman AJ, Chopra AK, and Shak JR. 2013. Characterization of Aeromonas hydrophila wound pathotypes by comparative genomic and functional analyses of virulence genes. mBio 4:e00064-13.
Wu CJ, Wang HC, Chen CS, Shu HY, Kao AW, Chen PL, and Ko WC. 2012. Genome sequence of a novel human pathogen, Aeromonas aquariorum. J. Bacteriol. 194:4114–4115.
Li Y, Liu Y, Zhou Z, Huang H, Ren Y, Zhang Y, Li G, and Wang L. 2011. Complete genome sequence of Aeromonas veronii strain B565. J. Bacteriol. 193:3389–3390.
Beatson SA, das Gracas de Luna M, Bachmann NL, Alikhan NF, Hanks KR, Sullivan MJ, Wee BA, Freitas-Almeida AC, Dos Santos PA, de Melo JT, Squire DJ, Cunningham AF, Fitzgerald JR, and Henderson IR. 2011. Genome sequence of the emerging pathogen Aeromonas caviae. J. Bacteriol. 193:1286–1287.
Overbeek R, Begley T, Butler RM, Choudhuri JV, Chuang HY, Cohoon M, de Crecy-Lagard V, Diaz N, Disz T, Edwards R, Fonstein M, Frank ED, Gerdes S, Glass EM, Goesmann A, Hanson A, Iwata-Reuyl D, Jensen R, Jamshidi N, Krause L, Kubal M, Larsen N, Linke B, McHardy AC, Meyer F, Neuweger H, Olsen G, Olson R, Osterman A, Portnoy V, Pusch GD, Rodionov DA, Ruckert C, Steiner J, Stevens R, Thiele I, Vassieva O, Ye Y, Zagnitko O, and Vonstein V. 2005. The subsystems approach to genome annotation and its use in the project to annotate 1000 genomes. Nucleic Acids Res. 33:5691–5702.
Chun J, Grim CJ, Hasan NA, Lee JH, Choi SY, Haley BJ, Taviani E, Jeon YS, Kim DW, Brettin TS, Bruce DC, Challacombe JF, Detter JC, Han CS, Munk AC, Chertkov O, Meincke L, Saunders E, Walters RA, Huq A, Nair GB, and Colwell RR. 2009. Comparative genomics reveals mechanism for short-term and long-term clonal transitions in pandemic Vibrio cholerae. Proc. Natl. Acad. Sci. U. S. A. 106:15442–15447.
Grim CJ, Kotewicz ML, Power KA, Gopinath G, Franco AA, Jarvis KG, Yan QQ, Jackson SA, Sathyamoorthy V, Hu L, Pagotto F, Iversen C, Lehner A, Stephan R, Fanning S, and Tall BD. 2013. Pan-genome analysis of the emerging foodborne pathogen Cronobacter spp. suggests a species-level bidirectional divergence driven by niche adaptation. BMC Genomics 14:366.
Richter M and Rossello-Mora R. 2009. Shifting the genomic gold standard for the prokaryotic species definition. Proc. Natl. Acad. Sci. U. S. A. 106:19126–19131.
Tamura K, Peterson D, Peterson N, Stecher G, Nei M, and Kumar S. 2011. MEGA5: molecular evolutionary genetics analysis using maximum likelihood, evolutionary distance, and maximum parsimony methods. Mol. Biol. Evol. 28:2731–2739.
Carver TJ, Rutherford KM, Berriman M, Rajandream MA, Barrell BG, and Parkhill J. 2005. ACT: the Artemis comparison tool. Bioinformatics 21:3422–3423.
Li L, Stoeckert CJ Jr, and Roos DS. 2003. OrthoMCL: identification of ortholog groups for eukaryotic genomes. Genome Res. 13:2178–2189.
Edgar RC. 2004. MUSCLE: a multiple sequence alignment method with reduced time and space complexity. BMC Bioinformatics 5:113.
Talavera G and Castresana J. 2007. Improvement of phylogenies after removing divergent and ambiguously aligned blocks from protein sequence alignments. Syst. Biol. 56:564–577.
Felsenstein J. 1988. Phylogenies from molecular sequences: inference and reliability. Annu. Rev. Genet. 22:521–565.
Felsenstein J. 1989. Mathematics vs. evolution: mathematical evolutionary theory. Science 246:941–942.
O'Toole GA and Kolter R. 1998. Initiation of biofilm formation in Pseudomonas fluorescens WCS365 proceeds via multiple, convergent signalling pathways: a genetic analysis. Mol. Microbiol. 28:449–461.
Khajanchi BK, Kozlova EV, Sha J, Popov VL, and Chopra AK. 2012. The two-component QseBC signalling system regulates in vitro and in vivo virulence of Aeromonas hydrophila. Microbiology 158:259–271.
Morohoshi T, Shiono T, Takidouchi K, Kato M, Kato N, Kato J, and Ikeda T. 2007. Inhibition of quorum sensing in Serratia marcescens AS-1 by synthetic analogs of N-acylhomoserine lactone. Appl. Environ. Microbiol. 73:6339–6344.
Erova TE, Pillai L, Fadl AA, Sha J, Wang S, Galindo CL, and Chopra AK. 2006. DNA adenine methyltransferase influences the virulence of Aeromonas hydrophila. Infect. Immun. 74:410–424.
Sha J, Kozlova EV, and Chopra AK. 2002. Role of various enterotoxins in Aeromonas hydrophila-induced gastroenteritis: generation of enterotoxin gene-deficient mutants and evaluation of their enterotoxic activity. Infect. Immun. 70:1924–1935.
Khajanchi BK, Sha J, Kozlova EV, Erova TE, Suarez G, Sierra JC, Popov VL, Horneman AJ, and Chopra AK. 2009. N-Acylhomoserine lactones involved in quorum sensing control the type VI secretion system, biofilm formation, protease production, and in vivo virulence in a clinical isolate of Aeromonas hydrophila. Microbiology 155:3518–3531.
Sierra JC, Suarez G, Sha J, Baze WB, Foltz SM, and Chopra AK. 2010. Unraveling the mechanism of action of a new type III secretion system effector AexU from Aeromonas hydrophila. Microb. Pathog. 49:122–134.
Lory S, Strom MS, and Johnson K. 1988. Expression and secretion of the cloned Pseudomonas aeruginosa exotoxin A by Escherichia coli. J. Bacteriol. 170:714–719.
Liu PV. 1973. Exotoxins of Pseudomonas aeruginosa. I. Factors that influence the production of exotoxin A. J. Infect. Dis. 128:506–513.
Bjorn MJ, Iglewski BH, Ives SK, Sadoff JC, and Vasil ML. 1978. Effect of iron on yields of exotoxin A in cultures of Pseudomonas aeruginosa PA-103. Infect. Immun. 19:785–791.
Choi KH, Gaynor JB, White KG, Lopez C, Bosio CM, Karkhoff-Schweizer RR, and Schweizer HP. 2005. A Tn7-based broad-range bacterial cloning and expression system. Nat. Methods 2:443–448.
Sha J, Rosenzweig JA, Kirtley ML, van Lier CJ, Fitts EC, Kozlova EV, Erova TE, Tiner BL, and Chopra AK. 2013. A non-invasive in vivo imaging system to study dissemination of bioluminescent Yersinia pestis CO92 in a mouse model of pneumonic plague. Microb. Pathog. 55:39–50.
Shak JR, Whitaker JA, Ribner BS, and Burd EM. 2011. Aminoglycoside-resistant Aeromonas hydrophila as part of a polymicrobial infection following a traumatic fall into freshwater. J. Clin. Microbiol. 49:1169–1170.
Reith ME, Singh RK, Curtis B, Boyd JM, Bouevitch A, Kimball J, Munholland J, Murphy C, Sarty D, Williams J, Nash JH, Johnson SC, and Brown LL. 2008. The genome of Aeromonas salmonicida subsp. salmonicida A449: insights into the evolution of a fish pathogen. BMC Genomics 9:427.
Worby CJ, Lipsitch M, and Hanage WP. 2014. Within-host bacterial diversity hinders accurate reconstruction of transmission networks from genomic distance data. PLoS Comput. Biol. 10:e1003549.
Iglewski BH, Liu PV, and Kabat D. 1977. Mechanism of action of Pseudomonas aeruginosa exotoxin A: adenosine diphosphate-ribosylation of mammalian elongation factor 2 in vitro and in vivo. Infect. Immun. 15:138–144.
McClean KH, Winson MK, Fish L, Taylor A, Chhabra SR, Camara M, Daykin M, Lamb JH, Swift S, Bycroft BW, Stewart GS, and Williams P. 1997. Quorum sensing and Chromobacterium violaceum: exploitation of violacein production and inhibition for the detection of N-acylhomoserine lactones. Microbiology 143(Pt 12):3703–3711.
Pukatzki S, McAuley SB, and Miyata ST. 2009. The type VI secretion system: translocation of effectors and effector-domains. Curr. Opin. Microbiol. 12:11–17.
Sierra JC, Suarez G, Sha J, Foltz SM, Popov VL, Galindo CL, Garner HR, and Chopra AK. 2007. Biological characterization of a new type III secretion system effector from a clinical isolate of Aeromonas hydrophila—part II. Microb. Pathog. 43:147–160.
Gray GL, Smith DH, Baldridge JS, Harkins RN, Vasil ML, Chen EY, and Heyneker HL. 1984. Cloning, nucleotide sequence, and expression in Escherichia coli of the exotoxin A structural gene of Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. U. S. A. 81:2645–2649.
Matsui H, Sekiya Y, Nakamura M, Murayama SY, Yoshida H, Takahashi T, Imanishi K, Tsuchimoto K, Uchiyama T, Sunakawa K, and Ubukata K. 2009. CD46 transgenic mouse model of necrotizing fasciitis caused by Streptococcus pyogenes infection. Infect. Immun. 77:4806–4814.
Borger van der Burg BL, Bronkhorst MW, and Pahlplatz PV. 2006. Aeromonas hydrophila necrotizing fasciitis. A case report. J. Bone Joint Surg. Am. 88:1357–1360.
Chern CH, How CK, and Huang LJ. 2006. Images in emergency medicine. Necrotizing fasciitis caused by Aeromonas hydrophila. Ann. Emerg. Med. 48:216–225.
Psoinos CM, Flahive JM, Shaw JJ, Li Y, Ng SC, Tseng JF, and Santry HP. 2013. Contemporary trends in necrotizing soft-tissue infections in the United States. Surgery 153:819–827.
Salcido RS. 2007. Necrotizing fasciitis: reviewing the causes and treatment strategies. Adv. Skin Wound Care 20:288–293, quiz 294–295.
Keung EZ, Liu X, Nuzhad A, Adams C, Ashley SW, and Askari R. 2013. Immunocompromised status in patients with necrotizing soft-tissue infection. JAMA Surg. 148:419–426.
Al Shukry S and Ommen J. 2013. Necrotizing fasciitis—report of ten cases and review of recent literature. J. Med. Life 6:189–194.
Johansson L, Thulin P, Low DE, and Norrby-Teglund A. 2010. Getting under the skin: the immunopathogenesis of Streptococcus pyogenes deep tissue infections. Clin. Infect. Dis. 51:58–65.
Skwor T, Shinko J, Augustyniak A, Gee C, and Andraso G. 2014. Aeromonas hydrophila and Aeromonas veronii predominate among potentially pathogenic ciprofloxacin- and tetracycline-resistant Aeromonas isolates from Lake Erie. Appl. Environ. Microbiol. 80:841–848.
Ibekwe AM and Lyon SR. 2008. Microbiological evaluation of fecal bacterial composition from surface water through aquifer sand material. J. Water Health 6:411–421.
Horneman AJ and Morris JG. 2012. Aeromonas. In Rose BD (ed), UpToDate: clinical reference library for physicians. Wolters Kluwer Health, Wellesley, MA.
Moormeier DE, Endres JL, Mann EE, Sadykov MR, Horswill AR, Rice KC, Fey PD, and Bayles KW. 2013. Use of microfluidic technology to analyze gene expression during Staphylococcus aureus biofilm formation reveals distinct physiological niches. Appl. Environ. Microbiol. 79:3413–3424.
Kozlova EV, Popov VL, Sha J, Foltz SM, Erova TE, Agar SL, Horneman AJ, and Chopra AK. 2008. Mutation in the S-ribosylhomocysteinase (luxS) gene involved in quorum sensing affects biofilm formation and virulence in a clinical isolate of Aeromonas hydrophila. Microb. Pathog. 45:343–354.
Lynch MJ, Swift S, Kirke DF, Keevil CW, Dodd CE, and Williams P. 2002. The regulation of biofilm development by quorum sensing in Aeromonas hydrophila. Environ. Microbiol. 4:18–28.
Canals R, Altarriba M, Vilches S, Horsburgh G, Shaw JG, Tomas JM, and Merino S. 2006. Analysis of the lateral flagellar gene system of Aeromonas hydrophila AH-3. J. Bacteriol. 188:852–862.
Caiazza NC, Merritt JH, Brothers KM, and O'Toole GA. 2007. Inverse regulation of biofilm formation and swarming motility by Pseudomonas aeruginosa PA14. J. Bacteriol. 189:3603–3612.
Shrout JD, Chopp DL, Just CL, Hentzer M, Givskov M, and Parsek MR. 2006. The impact of quorum sensing and swarming motility on Pseudomonas aeruginosa biofilm formation is nutritionally conditional. Mol. Microbiol. 62:1264–1277.
Murray TS, Ledizet M, and Kazmierczak BI. 2010. Swarming motility, secretion of type 3 effectors and biofilm formation phenotypes exhibited within a large cohort of Pseudomonas aeruginosa clinical isolates. J. Med. Microbiol. 59:511–520.
Wood TK, Gonzalez Barrios AF, Herzberg M, and Lee J. 2006. Motility influences biofilm architecture in Escherichia coli. Appl. Microbiol. Biotechnol. 72:361–367.
Daddaoua A, Fillet S, Fernandez M, Udaondo Z, Krell T, and Ramos JL. 2012. Genes for carbon metabolism and the ToxA virulence factor in Pseudomonas aeruginosa are regulated through molecular interactions of PtxR and PtxS. PLoS One 7:e39390.
Stamenkovic I and Lew PD. 1984. Early recognition of potentially fatal necrotizing fasciitis. The use of frozen-section biopsy. N. Engl. J. Med. 310:1689–1693.
Johansson L, Linner A, Sunden-Cullberg J, Haggar A, Herwald H, Lore K, Treutiger CJ, and Norrby-Teglund A. 2009. Neutrophil-derived hyperresistinemia in severe acute streptococcal infections. J. Immunol. 183:4047–4054.
Lacy P. 2006. Mechanisms of degranulation in neutrophils. Allergy Asthma Clin. Immunol. 2:98–108.
Lintges M, van der Linden M, Hilgers RD, Arlt S, Al-Lahham A, Reinert RR, Plucken S, and Rink L. 2010. Superantigen genes are more important than the emm type for the invasiveness of group A Streptococcus infection. J. Infect. Dis. 202:20–28.

Information & Contributors


Published In

cover image Applied and Environmental Microbiology
Applied and Environmental Microbiology
Volume 80Number 1415 July 2014
Pages: 4162 - 4183
Editor: H. Nojiri
PubMed: 24795370


Received: 10 February 2014
Accepted: 26 April 2014
Published online: 24 June 2014


Request permissions for this article.



Christopher J. Grim
Food and Drug Administration, Laurel, Maryland, USA
Elena V. Kozlova
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Duraisamy Ponnusamy
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Eric C. Fitts
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Jian Sha
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Michelle L. Kirtley
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Christina J. van Lier
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Bethany L. Tiner
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Tatiana E. Erova
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA
Sandeep J. Joseph
Department of Medicine, Division of Infectious Diseases, Emory University School of Medicine, Atlanta, Georgia, USA
Timothy D. Read
Department of Medicine, Division of Infectious Diseases, Emory University School of Medicine, Atlanta, Georgia, USA
Joshua R. Shak
Department of Medicine, Division of Infectious Diseases, Emory University School of Medicine, Atlanta, Georgia, USA
Sam W. Joseph
Department of Cell Biology and Molecular Genetics, University of Maryland, College Park, Maryland, USA
Maryland Institute of Applied Environmental Health, School of Public Health, University of Maryland, College Park, Maryland, USA
Ed Singletary
Doctors Hospital, Augusta, Georgia, USA
Tracy Felland
Mercy Hospital and Trauma Center, Janesville, Wisconsin, USA
Wallace B. Baze
Department of Veterinary Sciences, M. D. Anderson Cancer Center, Bastrop, Texas, USA
Amy J. Horneman
Pathology and Laboratory Medical Services, VA Maryland Health Care System, Baltimore, Maryland, USA
Ashok K. Chopra
Department of Microbiology and Immunology, University of Texas Medical Branch, Galveston, Texas, USA


H. Nojiri


Address correspondence to Ashok K. Chopra, [email protected].
C.J.G., E.V.K., D.P., E.C.F., and J.S. contributed equally to this report.

Metrics & Citations



  • For recently published articles, the TOTAL download count will appear as zero until a new month starts.
  • There is a 3- to 4-day delay in article usage, so article usage will not appear immediately after publication.
  • Citation counts come from the Crossref Cited by service.


If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. For an editable text file, please select Medlars format which will download as a .txt file. Simply select your manager software from the list below and click Download.

View Options

Figures and Media






Share the article link

Share with email

Email a colleague

Share on social media

American Society for Microbiology ("ASM") is committed to maintaining your confidence and trust with respect to the information we collect from you on websites owned and operated by ASM ("ASM Web Sites") and other sources. This Privacy Policy sets forth the information we collect about you, how we use this information and the choices you have about how we use such information.
FIND OUT MORE about the privacy policy