Clinical Relevance of the Recently Described Species Aeromonas aquariorum

Strains and molecular identification.

A total of 150 biochemically identified extraintestinal clinical isolates (including 86 A. hydrophila, 28 A. sobria, 20 A. caviae, and 16 Aeromonas spp.) were received in our laboratory from Wen-Chien Ko (Cheng Kung University Hospital, Tainan, Taiwan) to be genetically identified with a previously described 16S rRNA gene-restriction fragment length polymorphism (RFLP) method (6, 11). In brief, the method consists of an initial digestion of 1,503 bp of the amplified 16S rRNA genes with enzymes AluI and MboI, which produce species-specific patterns for 10 Aeromonas species, including the most relevant clinical species (A. caviae, A. veronii, and A. hydrophila). Only 138 (92%) isolates were identified as Aeromonas species: 73 (52.9%) as A. caviae, 17 (12.3%) as A. hydrophila, 16 (11.6%) as A. veronii, 5 (3.6%) as A. media, and 3 (2.1%) as A. trota. The remaining 24 (17.5%) isolates showed an RFLP pattern similar to that described for A. caviae (Fig. 1). To confirm the identities of all these strains, the rpoD and gyrB genes were sequenced using previously described primers and conditions (23). Two of these isolates were proposed as two new species, A. taiwanensis and A. sanarellii (A. Alperi et al., submitted for publication), and the remaining 22 (15.9%) were shown to belong to A. aquariorum (Fig. 2). Three Spanish strains (MDC562, MDC573, and MDC671) isolated in the Hospital Comarcal Vega Baja (Orihuela, Spain) were also recognized, using the same genes, to belong to A. aquariorum (Fig. 2). In agreement with the results of previous studies (6, 10, 22) only a few strains (17 A. hydrophila and 15 A. caviae) were originally correctly identified using biochemical methods.

• Open in new tab
• Download powerpoint

FIG. 1.

Polyacrylamide gel showing almost the same 16S rRNA gene-RFLP pattern for some of the investigated A. aquariorum strains, the type strain of this species, A. hydrophila subsp. dhakensis, and A. caviae. Lanes: 1st and 13th, pBR322 DNA/BsuRI (HaeIII) marker 5; 2nd, A. hydrophila subsp. dhakensis LMG19559; 3rd, A. hydrophila subsp. dhakensis LMG 19562^T; 4th, A. aquariorum DSM18362^T; 5th, A. caviae CECT838^T; 6th to 10th, A. aquariorum clinical strains; 6th to 8th, strains isolated from feces (MDC562, MDC563, and MDC671); 9th, strain from a wound infection after a flame burn (A2-004); 10th, strain isolated from ascites in a case of acute appendicitis with peritonitis (A2-96); 11th, A. taiwanensis; and 12th, A. sanarellii. In the 7th, 11th, and 12th lanes, notice the atypical pattern showing some extra bands.

• Open in new tab
• Download powerpoint

FIG. 2.

Unrooted neighbor-joining phylogenetic tree based on the combined stretch of gyrB (450 bp)-rpoD (790 bp) gene sequences showing the relationship within the genus Aeromonas of 25 clinical A. aquariorum strains. The sources of isolates are given. Type strains are in bold font. Numbers at nodes indicate bootstrap values (1,000 replicates). The scale bar represents 1 nucleotide substitution per 50 nucleotides.

Antimicrobial susceptibility.

The susceptibility of the 25 strains identified as A. aquariorum was evaluated against a panel of 27 antimicrobial agents by using drugs and conditions described previously (17). All the isolates were resistant to amoxicillin (amoxicilline), cephalothin (cephalotin), and cefoxitin, but they varied in their resistance to other antimicrobial agents (Table 1). These responses were similar to those found in the environmental strains of A. aquariorum (17), with the exception of the 68% resistance to amoxicillin-clavulanic acid that contrasted with the 32% encountered in the environmental strains. It has been demonstrated that empirical treatment for Aeromonas bacteremia is normally inappropriate in ca. 20% of the cases and has an important impact on mortality (10).

Detection of virulence genes.

The strains were also screened for the presence of virulence genes, including the aerolysin/hemolysin genes, ast, alt, act, the TTSS genes (ascF-ascG), and aexT, using primers and conditions described earlier (2, 7, 8, 12). The genes that encode the pore-forming toxin aerolysin/hemolysin were present in all of the strains (Table 2). The two cytotoxic enterotoxin genes, the heat-stable ast and the heat-labile alt, were present in 84% and 28% of the strains, respectively, ast being much more prevalent than was found in previous studies (2). The three strains isolated from feces of patients with diarrhea and a considerable number of the strains from wounds (4/6, 66.6%) and blood (7/11, 63.3%) possessed the ascF-ascG genes. The aexT gene, which encodes one of the best-characterized toxins secreted by this system (27), was present in all stool strains and in ca. 30% of the blood and wound strains (Table 2). Interestingly, eight of the strains simultaneously possessed the act and the TTSS genes, which probably reflects the synergistic role of these genes in quorum sensing (20).

Differentiation of A. aquariorum from other species.

The 16S rRNA gene-RFLP pattern obtained for the Taiwan and Spanish isolates was compared with those of reference strains of A. aquariorum, i.e., the type strain (DSM18362^T); two strains previously identified as A. hydrophila subsp. dhakensis (LMG19559 and LMG15962^T [15], now considered a synonym of A. aquariorum [16]); and the type strain of A. caviae (CECT838^T). All the strains produced either the A. caviae RFLP pattern or a slightly similar pattern with extra bands (Fig. 1). This made us suspect that those strains from Taiwan identified as A. caviae by the 16S rRNA gene-RFLP method or other clinical strains of this species identified at our laboratory and received from Spanish hospitals could in fact belong to A. aquariorum. To evaluate this, we screened the 73 A. caviae strains from Taiwan, as well as 65 strains received from Spanish hospitals, that showed the A. caviae RFLP pattern for the production of acid from d-cellobiose (positive for A. caviae, negative for A. aquariorum [17]). None of the 73 Taiwan strains and only 6 of the Spanish strains, recovered from the feces of patients with diarrhea, were suspected to be A. aquariorum due to their negative production of acid from d-cellobiose. However, the rpoD gene sequences of these strains (data not shown) confirmed only two strains (33%) as belonging to A. aquariorum, while the remaining four belonged to A. caviae. These results showed that some A. caviae strains did not produce acid from d-cellobiose (1, 18) and, furthermore, that A. aquariorum could be hidden under A. caviae strains. The positive production of lysine decarboxylase, gas from d-glucose, and H₂S from cysteine of A. aquariorum are also traits that differentiate it from A. caviae, which is negative for these tests. Separation from A. hydrophila can be obtained by the fact that the new species does not utilize l-lactate (17). Strains with such characteristics should be suspected to belong to A. aquariorum.

The sequences of 16S rRNA genes were considered a reliable and straightforward tool for characterizing Aeromonas spp. (10) and have increasingly been used for identifying clinical strains (3, 13, 18, 24). However, we now know that this gene is unreliable for identification below the genus level (5 and A. Alperi et al., submitted). This is due to the high similarity between the sequences of the 16S rRNA genes of the different species and to the existence of nucleotide polymorphisms among the rrn operons of this gene (i.e., microheterogeneities) reported to occur in 8% of strains (5). Furthermore, these microheterogeneities are more prevalent in clinical strains of A. caviae and A. veronii than in environmental strains (5). In fact, the 16S rRNA gene sequence of the type strain of A. aquariorum only differs in three nucleotides from that of A. caviae (17), and microheterogeneities have been described in all three of these positions for strains of the latter species (5). In the present study, these limitations have been overcome by the use of housekeeping genes, such as rpoD and gyrB.

Epidemiologically related strains.

Three A. aquariorum isolates (A2-126, A2-131, and A2-96) showed identical gyrB and rpoD sequences (Fig. 2), indicating that they may be derived from the same clone. Genotyping with enterobacterial repetitive intergenic consensus-PCR using the primers and conditions described previously (21) confirmed that they belonged to the same genotype (data not shown). These strains also bear the same virulence genes and showed resistance to amoxicillin, amoxicillin-clavulanic acid, cephalothin, cefoxitin, erythromycin, and tetracycline. In a retrospective search of the origin of those strains, we noticed that two of them (A2-126 and A2-131) were isolated from a 69-year-old patient with acute cholangitis, one from a bile sample obtained from endoscopic nasobiliary drainage and the other 14 days later from a surgical wound of the abdominal wall after a cholecystectomy. This represents a new case of surgical site infection (SSI). Aeromonas SSIs have been poorly studied but are not rare, as we demonstrated in a recent review (25). The third strain (A2-96) was isolated from an ascites culture 7 months earlier in a 70-year-old patient who presented acute appendicitis complicated with peritonitis. The fact that both patients show the same genotype suggests a probable common source of infection, but the epidemiological relationship of the patients could not be established. Cholangitis is the most common Aeromonas infection of the hepatobiliary systems, with an incidence of isolation from bile samples ranging from 1.3 to 2.9% (10). Cases of secondary Aeromonas peritonitis due to appendix perforation have also been described (10, 14).

The present study is the first description of the implication of the new species, A. aquariorum, in clinical extraintestinal infections. Previously, as was mentioned above, this species had been reported from patients with diarrhea in Bangladesh under the name A. hydrophila subsp. dhakensis (15).

The medical relevance of these findings lies in the fact that the clinical strains of A. aquariorum possess many virulence genes that can potentially play an important role in the development of the infection. Clinicians should be aware of this new important Aeromonas species, which can be associated with diarrhea; bacteremia; wound infections, including SSI; and several other extraintestinal infections, as can be seen from the diversity of clinical samples in which it was recovered. It is likely that A. aquariorum could be distributed globally, probably unrecognized under A. hydrophila or A. caviae. However, more studies will be necessary to confirm this.