The ubiquitous bacterium Aeromonas hydrophila is a freshwater, facultatively anaerobic, chemoorganoheterotroph (Garrity et al., 2006) and the etiologic agent of disease in amphibians, birds, fishes, mammals, and reptiles, with the most common forms of disease being gastroenteritis, septicemia, and necrotizing fasciitis (Cipriano et al., 1984; Figueras et al., 2007; Monaghan et al., 2008; Janda and Abbott, 2010). Virulence in A. hydrophila is multifactorial, with disease resulting from the production and/or secretion of virulence factors, such as adhesins, cytotoxins, hemolysins, lipases, and proteases as well as the capacity to form biofilms, use specific metabolic pathways, and mediate virulence factor expression through quorum sensing (Allan and Stevenson, 1981; Cahill, 1990; Thornley et al., 1997; Beaz-Hidalgo and Figueras, 2013). The majority of experimental studies on identifying virulence determinants in Aeromonas spp. have been performed in the strain A. hydrophila SSU, which was later recognized to be affiliated with A. dhakensis on the basis of ANI and phylogeny comparisons (Grim et al., 2014). Adding confusion to this complexity, the literature on A. hydrophila is riddled with conflicting reports on the molecular determinants of virulence attributed to this species because of changes in classification and problems stemming from misidentification (Colston et al., 2014; Beaz-Hidalgo et al., 2015). The purpose of this review article is to provide an updated view on what is known about virulence factors in the aftermath of reclassification of A. hydrophila SSU.

Huys et al. (2002) recognized that some diarrheal isolates, while closely related to A. hydrophila, show atypical metabolic activities for urocanic acid (+), L-fucose (-), and L-arabinose (-). On these bases, these strains were classified into a subspecies known as A. hydrophila subsp. dhakensis (Huys et al., 2002). Then, in 2013, A. hydrophila subsp. dhakensis was recognized to be synonymous to A. aquariorum and both were combined under the name A. dhakensis, a species that is functionally divergent from A. hydrophila, based on multilocus phylogenetic analyses and phenotypic characteristics (Beaz-Hidalgo et al., 2013). Studies on the virulence factors expressed by the diarrheal isolate SSU, previously considered to be affiliated to A. hydrophila and now know to be A. dhakensis, are regarded as the seminal literature on molecular pathogenesis of Aeromonas (Grim et al., 2014). Given the turbulent nature of classification within Aeromonas spp., this review aims to clarify which virulence factors have been characterized within current members of A. hydrophila (Supplementary Table S1) by auditing the body of knowledge on the molecular understanding of these genes so that future research can progress from a more solid foundation.

Cascades of genetic regulation that lead to situational expression of virulence factors are known to occur in Aeromonas spp., but these interactions remain a relatively uncharted area of research in phylogenetically confirmed A. hydrophila strains. For example, outbreaks of A. hydrophila are generally thought to be linked with changes in host susceptibility caused by environmental changes, such as hypoxic conditions and excessive nitrite levels in farmed fish, as well as increases in temperature, which are linked with the production of virulence factors, such as cyototoxins and hemolysins (Swann and White, 1991; Mateos et al., 1993; Janda and Abbott, 2010). To exploit changes in host susceptibility due to increases in temperature, Aeromonas spp. virulence factors have also evolved temperature-dependent expression (Merino et al., 1992; Gonzalez-Serrano et al., 2002). For example, clinical strains of A. hydrophila can grow at temperatures greater than the isolate’s optimal growth temperature of 28°C (Popoff and Veron, 1976); however, when temperatures increase to 37°C, protease activity decreases and cytotoxin and hemolysin activity increases (Yu et al., 2007). In contrast, environmental isolates are well adapted to low temperatures and can grow uninhibited at temperatures as low as 4°C, a temperature that restricts growth of clinical isolates (Mateos et al., 1993). Some of the better studied regulatory effects are the linkage between quorum sensing and biofilm formation which was shown to not only mediate the expression of virulence factors, but also regulate cell density (Swift et al., 1997; Lynch et al., 2002; Janda and Abbott, 2010). In addition, while polar flagella in A. hydrophila are constitutively expressed, there are well-described regulators that trigger lateral flagella expression such as surface contact and viscosity (Wilhelms et al., 2011, 2013). Another class of regulatory effects includes the upregulation of virulence factors through lysogenic conversion; however, to-date no experimental data has been published on this phenomenon within A. hydrophila. Considering the broad effects that these regulatory factors have on disease, experimental studies that resolve these interactions are fundamental to the advancement of knowledge for the field of A. hydrophila as a whole. A review of known virulence factors and the respective regulatory effects that have been evaluated in Aeromonas spp. and are genetically present within A. hydrophila are presented in Figure 1.

Biofilms provide bacteria with resistance to antimicrobial agents and host defenses (Costerton et al., 1995; Lynch et al., 2002). Aeromonas spp. evolved multiple regulatory mechanisms for biofilm formation that are intimately linked with the production of virulence factors. The quorum sensing response regulator of the reclassified isolate A. piscicola A1 (formerly A. hydrophila), ahyRI, produces LuxRI homologs, N-(butanoyl)-L-homoserine lactones (BHL), and N-hexanoyl-L-homoserine lactones (AHL); autoinducers that regulate cell division (Swift et al., 1997). In A. dhakensis SSU ΔahyRI mutants, T6SS effectors Hcp and Vgr are unable to be secreted which results in decreased biofilm formation (Khajanchi et al., 2009). Interestingly, some strains transcribe ahyRI (e.g., A. hydrophila ATCC 7966), but lack AHL/BHLs, which may indicate an alternate function of ahyRI that has yet to be described (dos Reis Ponce-Rossi et al., 2016). Similarly, the recently characterized autoregulatory two-component signal transduction system QseBC is a widely conserved system within A. hydrophila and was first described in A. dhakensis SSU, as mutants with an inactive response regulator (QseB) have reduced swimming and swarming motility, form thicker biofilms, and secrete fewer virulence factors, which leads to attenuation of virulence. When the gene aha0701h is overexpressed in ΔqseB mutants, biofilm formation decreases, presumably due to dysregulation of genes fleN (regulates flagellar number) and vpsT (transcriptional response regulator; Khajanchi et al., 2012; Kozlova et al., 2012).

A. hydrophila isolates produce lateral flagella for surface movement/swarming and polar flagella for movement in suspension. Polar flagella production has been studied within A. piscicola AH-3, with mutations in flaAB, flaH, fliA, fliM, maf-1, and flrC abolishing production of polar flagella and resulting in decreased adherence and biofilm formation (Canals et al., 2006). Considering that flagellar glycosylation was shown to be linked with the ability to form biofilms as well as adhere to Hep-2 cells, it is important to mention that there are notable differences within Aeromonas species. In addition to having only a single lateral flagellin, polar and lateral flagella are glycosylated in A. piscicola AH-3 whereas A. hydrophila AH-1 has two lateral flagellins and only the polar flagellum is glycosylated. When pseudaminic acid biosynthesis genes pseB and pseI were mutagenized, the result was an inability to produce both polar and lateral flagella in A. piscicola AH-3, but only affected polar flagella production in A. hydrophila AH-1. Therefore, lateral flagella production was unaffected in glycosylation negative A. hydrophila AH-1 mutants (Fulton et al., 2015). Similarly, in the diseased eel isolate A. hydrophila W (no genome submitted), mutations in flgE, flgN, flhA, fliJ, flmB, lafK, and maf-5 result in loss of lateral flagella, which causes decreased motility, biofilm formation, and mucosal adherence (Jiang et al., 2015). While polar and lateral flagella transcriptional hierarchies, regulation, and contribution to virulence are well-described in other species, as the date of this publication, no member of A. hydrophila with a publically accessible genome has undergone genetic manipulations to evaluate the contribution of polar or lateral flagella for virulence.

The A. hydrophila gene cluster tapABCD is responsible for type IV pilus biogenesis and is an integral part of the extracellular secretory pathway. To test for function, the A. hydrophila Ah65 (genome unavailable) gene tapD gene was used to successfully complement a strain of Pseudomonas aeruginosa that lacks PilD (an ortholog of TapD; Pepe et al., 1996). Another type IV pilus is the bundle-forming pilus, which is encoded by bfp and acts an important internal colonization factor for multiple species of Aeromonas (A. hydrophila Ah65 was observed expressing both bfp and tap; Barnett et al., 1997; Kirov et al., 2000). Taken with the observation that TapD is required for secretion of virulence factors, such as aerolysin and proteases, these genes appear to be fundamental for pathogenicity.

Capsules, O-antigens, and S-layer proteins provide mechanisms to evade host defenses. Within Aeromonas spp., capsules also show anti-phagocytic activity, increase resistance to the complement system, and increase adherence (Martinez et al., 1995; Merino et al., 1997). O-antigens are a class of structurally diverse lipopolysaccharides that act as colonization factors. At 20°C O-antigen is produced by A. piscicola AH-3, but not at 37°C, resulting an O-antigen-deficient strains that are unable to colonize hosts and have reduced expression of T3SS components (Merino et al., 1996; Vilches et al., 2009). Across A. hydrophila, eight distinct O-antigen gene clusters are present, with all epidemic strains isolated from channel catfish (Ictalurus punctatus) sharing a homologous O-antigen gene cluster (Hossain, 2012). In A. hydrophila TF7 (genomic data unavailable), the S-layer protein gene (ahsA) encodes an external paracrystalline layer that is lost upon insertional mutagenesis of spsD (S-protein secretion; Thomas and Trust, 1995). Another study of S-layer proteins in five pathogenic human and eel isolates of A. hydrophila (A19, AH290, E37, E40, and TW1; genomic data unavailable) shows that serogroups of A. hydrophila other than O:11 contain S-layer proteins O:14 and O:81 (Esteve et al., 2004).

Hemolysins are a diverse group of multifunctional enzymes that play a central role in A. hydrophila pathogenesis (Wadstrom et al., 1976; Asao et al., 1984). The extracellular heat-labile hemolysin (AHH1) is the most abundant of several widely distributed hemolysins (AerA, AHH1, AhyA, and Asa1), with the most cytotoxic genotype being a synergistic combination of aerA and ahh1 (Hirono and Aoki, 1991; Wang et al., 2003). In A. media A6 (formerly A. hydrophila) Aerolysin A (aerA) and Hemolysin A (hlyA) comprise another two-component hemolytic system in which virulence is attenuated only when both hlyA and aerA activity is abolished (Wong et al., 1998). In A. dhakensis SSU, the iron dependent, fur and gidA-regulated, enterotoxin Act is the most cytotoxic virulence factor of and a core gene within A. hydrophila, with studies in A. dhakensis SSU demonstrating that Act induces multiple effects including hemolytic, cytotonic, and cytotoxic activities, but unlike other virulence factors exported via the T3SS or T6SS, Act is exported through the T2SS (Chopra et al., 2000; Sha et al., 2001, 2004, 2005; Fadl et al., 2006; Erova et al., 2012).

A. hydrophila spp. express diverse degradative enzymes that can contribute to virulence including collagenase, elastase, enolase, lipase, metalloprotease, and serine protease. A. piscicola AH-3 contains a collagenase, which has sequence similarity to the open reading frame AHA_0517 of A. hydrophila ATCC 7966^T, and was shown to be cytotoxic to Vero cells, with loss of this enzyme resulting in a 5–15% increase in cell viability; however, this mutation did not result in complete reduction of cytotoxicity (Duarte et al., 2015). The ahpAB genes of A. hydrophila AG2 (genomic data unavailable) produce potent virulence factors: an extracellular protease that is not essential for virulence, but is present in the most virulent pathotypes (AerA⁺Alt⁺Ahp⁺) along with a secreted elastase with caseinolytic and elastolytic activity that correlates with an LD₅₀ 100 times more virulent than AhpB^- mutants when assayed in rainbow trout (Oncorhynchus mykiss; Rivero et al., 1991; Cascon et al., 2000; Li et al., 2011). Another extracellular protease (epr) was discovered in the soft-shell turtle isolate A. hydrophila AH1 and found to be present in the most common pathotype in diseased fishes (Aer⁺Alt⁺Act⁺EprCAI⁺Ahp⁺; Chang et al., 1997; Hu et al., 2012). In the rainbow trout isolate A. hydrophila B32, a novel serine protease (ser) was found that exhibits cytotoxic properties and is thermostable, both of which are characteristics that differentiate this protease from known A. hydrophila α-hemolysins and β-hemolysins (Rodriguez et al., 1992). While four times less active than serine protease, the virulent A. hydrophila EO63 (genomic data unavailable) was shown to produce a thermostable metalloprotease with enzymatic activity on casein and elastin, an optimal pH of 8.0, and an LD₅₀ of 3.5 μg/g (Esteve and Birbeck, 2004). Enolase, a secreted and surface-expressed glycolytic enzyme, was identified as a virulence factor in A. dhakensis SSU, based on binding to human plasminogen which leads to production of plasmin (degrades blood plasma proteins), with previous reports showing that enolase functions as a heat-shock protein and a regulator of transcription by binding host chromatin/cytoskeletal structures as well as being necessary for viability (Sha et al., 2009).

In general, lipases have diverse functions, but are linked with virulence in numerous pathogens (Stehr et al., 2003). An extracellular lipase (EC3.1.1.3) is produced by A. piscicola AH-3 (formerly A. hydrophila); however, the link between virulence and this gene is speculative in A. hydrophila (Anguita et al., 1993). Conversely, the heat-labile lipase Alt and the heat-stable lipase Ast are important cytotonic enterotoxins in the pathogenicity of A. dhakensis SSU, with both being able to cause significant fluid secretion, with only the previously described cytotoxic enterotoxin Act having a greater effect on fluid secretion (Sha et al., 2002; Li et al., 2011). Based on comparative genomics, Alt and Ast are core elements of A. hydrophila; however, no experiments have been performed within existing members of this species to characterize these toxins. Two additional lipases, phospholipase A1 (pla) and phospholipase C (plc), were explored in A. piscicola AH-3, with the finding that pla lacks a significant effect on virulence while plc (lecithinase) was cytotoxic and has LD₅₀ values 10 times more virulent than plc-deficient mutants (Merino et al., 1999).

The range of virulence factors encoded by A. hydrophila includes adherence proteins, catalysts, nucleases, and toxins that may be expressed differently depending upon the respective environment. The role of the adhesin minD in virulence is its ability to mediate mucosal adherence, increase biofilm formation, and facilitate cell division as well as motility (Huang et al., 2015). The enzyme 5-enolpyruvylshikimate 3-phosphate synthase (encoded by aroA) is required for folate availability and contributes to A. hydrophila AG2 (genomic data unavailable) viability in intraperitoneally injected rainbow trout, with aroA mutants no longer recoverable from fish internal organs because environmentally derived folate is scarce (Hernanz Moral et al., 1998). Another element of host evasion is the nuclease encoded by the ahn gene of A. hydrophila J-1 which shows no significant change in hemolytic activity or growth in vitro; however, when Δahn mutants are introduced into fish and mice models, virulence is attenuated (Ji et al., 2015). Another conserved gene of A. hydrophila that has only been characterized in A. dhakensis SSU is vacB, which encodes RNase R; an exoribonuclease with multiple functions that include permitting growth at 4°C (cold-shock protein) and supporting motility. Isogenic mutants of RNase R show a 70% attenuation in virulence (Erova et al., 2008).

Another virulence factor that is shown to increase host evasion for A. dhakensis SSU is the pore-forming RTX toxin RtxA that requires contact with host cells and is regulated by the rtxACHBDE operon so that production of RtxA coincides with regulation of other cytotoxins, such as aerolysins and hemolysins and acts to covalently cross-link host cytoskeletal-actin, resulting in host cells having a rounded phenotype that leads to apoptosis (McCoy et al., 2010; Suarez et al., 2012). Another conserved virulence factor across all A. hydrophila strains in GenBank is the ToxR-regulated lipoprotein (TagA) of A. dhakensis SSU, which cleaves the complement C1-esterase inhibitor, thereby increasing serum resistance and decreasing erythrocyte lysis (Pillai et al., 2006).

The introduction of virulence factors as well as their effects on the alternate regulation within Aeromonas spp. is a recurring theme of crucial importance, yet these elements remain understudied. Plasmids are a confirmed source of multidrug resistance in Aeromonas spp. and have been shown to have the potential to be conjugally transferred between known human pathogens, such as Acinetobacter baumannii AYE and A. hydrophila (Del Castillo et al., 2013). In addition to being used as a “molecular map” to identify ancestral lineages, prophage that contain putative cis-acting elements and trans-acting factors were found to be conserved within hypervirulent strains of A. hydrophila which strongly implies that the differential regulation of virulence factors (and therefore the dramatic increase in virulence) may be caused by the lysogenic conversion of this conserved A. hydrophila lineage by these mobile genetic elements (Hossain et al., 2013). Of note, while A. dhakensis SSU contains the majority of virulence factors that are present within confirmed members of A. hydrophila, all other members of A. dhakensis with fully sequenced genomes appear to lack these genes (Supplementary Table S1). Results of numerous core genome phylogenies and average nucleotide identity analyses support the grouping of A. hydrophila SSU within A. dhakensis (Colston et al., 2014; Beaz-Hidalgo et al., 2015; Rasmussen-Ivey et al., 2016), if virulence factors are introduced or controlled by mobile genetic elements, then taxonomy and functionality demand separate analyses.

In Aeromonas spp., as with all pathogens, disease is the result of complex molecular interactions between bacterium, environment, and host; however, the literature on A. hydrophila remains limited by the lack of experimental data on validated members of A. hydrophila. While there numerous virulence factors shared between members of A. hydrophila, A. dhakensis SSU, and A. piscicola AH-3 there are also key examples in the literature that show conflicting data between virulence factors, secretion systems, quorum sensing, and their effect on pathogenicity. This inconsistency is illustrated by the highly virulent catfish isolate A. hydrophila ML09-119 that acts as a primary pathogen when other members of this species act as secondary pathogens (Griffin et al., 2013; Zhang et al., 2013). With that in mind, the thorough research conducted on A. dhakensis SSU by researchers, such as Dr. Ashok Chopra, still holds relevance for A. hydrophila, but future research should be mindful of the phylogenetic reclassification for strains AH-3 and SSU and that there may be significant differences in the molecular determinants of virulence for A. hydrophila.

To compare isolates of A. hydrophila, biochemical, morphological, and molecular techniques are required (Abbott et al., 2003; Martinez-Murcia et al., 2011; Beaz-Hidalgo et al., 2015). As of 2016, few strains exist that have enough supporting data to facilitate comparative studies. There are many sources of uncertainty when comparing A. hydrophila strains, including genetic heterogeneity, the lack of natural models of infection, and reclassification of bacterial strains as new data emerges. Future research should aim to couple typing techniques (e.g., genome sequencing) with experimental data on virulence determinants so that there is a clear phylogenetic context for these studies.

When considering known virulence factors, the definitive biological separation of A. hydrophila, A. piscicola AH-3, and A. dhakensis SSU has yet to be established. In A. hydrophila as in these other species, disease is the result of a molecular symphony, with each virulence factor contributing to a cumulative effect (Figure 1). Research studying novel virulence factors and regulatory effects will help unveil the determinants that allow for infection and what differentiates A. hydrophila from other aeromonads. To better understand A. hydrophila pathogenesis it is imperative that future research develops natural models of infection, assesses the role of mobile genetic elements in virulence, and quantifies the interplay between virulence factors and host response in concert with molecular genetic approaches.

CR-I, MF, DM, and ML all contributed to the conception, writing and editing of this manuscript.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.