Chironomid egg masses harbour the clinical species Aeromonas taiwanensis and Aeromonas sanarellii

Authors


Correspondence: María J. Figueras, Facultat de Medicina i Ciències de la Salut, Departament de Ciènces Médiques Bàsiques, Unitat de Microbiologia, IISPV, Universitat Rovira i Virgili, Reus, Spain. Tel.: +34 977759321; fax: +34 977759322; e-mail: mariajose.figueras@urv.cat

Abstract

Bacteria of the genus Aeromonas are found worldwide in aquatic environments and may produce human infections. In 2010, two new clinical species, Aeromonas sanarellii and Aeromonas taiwanensis, were described on the basis of one strain recovered from wounds of hospitalized patients in Taiwan. So far, only four environmental isolates of A. sanarellii and one of A. taiwanensis have been recorded from waste water in Portugal and an additional clinical strain of A. taiwanensis from the faeces of a patient with diarrhoea in Israel. In the present study, strains belonging to these two species were identified from chironomid egg masses from the same area in Israel by sequencing the rpoD gene. This represents a new environmental habitat for these novel species. The first data on the virulence genes and antibiotic susceptibility are provided. The isolates of these two new species possess multiple virulence genes and are sensitive to amikacin, aztreonam, cefepime, cefoxatime, ceftazidime, ciprofloxacin, gentamicin, piperacillin–tazobactam, tigecycline, tobramycin, trimethoprim–sulfamethoxazole and imipenem. The key phenotypic tests for the differentiation of these new species from their closest relative Aeromonas caviae included the utilization of citrate, growth at 45 °C in sheep blood agar and acid production of cellobiose.

Introduction

Aeromonas are primarily inhabitants of aquatic environments, able to cause gastroenteritis, bacteraemia and wound or soft tissue infections in humans (Figueras, 2005; Janda & Abbott, 2010). Transmission to humans can occur through open wounds or by consumption of contaminated water or food (Figueras, 2005; Janda & Abbott, 2010; Khajanchi et al., 2010; Pablos et al., 2010). Several studies have provided further evidence that Aeromonas infections are waterborne because identical genotypes (clonal isolates) have been found in drinking water and in stools of patients with diarrhoea (Khajanchi et al., 2010; Pablos et al., 2010). These results are in agreement with some previous studies (Martínez-Murcia et al., 2000) and contradict others (Borchardt et al., 2003). In 2007, Aeromonas was discovered for the first time to be able to inhabit chironomid egg masses, like Vibrio cholerae does (Halpern et al., 2007; Senderovich et al., 2008). Chironomids are nonbiting midges that can infest drinking water systems and thus can be a source of Aeromonas transmission to humans (Halpern et al., 2007; Senderovich et al., 2008). Senderovich et al. (2008) surveyed bacterial communities able to degrade chironomid egg masses. About 4% of the isolates (45 out of 1018) degraded the egg masses, and of those, 43 were identified as Aeromonas caviae (n = 33), Aeromonas veronii (n = 9) and Aeromonas hydrophila (n = 1) on the basis of partial sequences of the 16S rRNA gene. Considering that the latter gene is not a reliable tool for the identification of all Aeromonas spp., Figueras et al. (2011c) re-identified those strains by sequencing the rpoD gene, which is considered more reliable (Figueras et al., 2011b). While the studied isolates of the species A. hydrophila and A. veronii were correctly identified, those of A. caviae proved to belong to the recently described novel species A. aquariorum (Figueras et al., 2011c). All these species have been involved in clinical infections and may bear several virulence genes, like those encoding Shiga toxins (stx1 and stx2), the type III secretion system (TTSS) (ascF-G, ascV), flagella (fla) as well as several toxins (ast, act, alt, aexT) among others (Chacón et al., 2004; Aguilera-Arreola et al., 2005; Fehr et al., 2006; Chopra et al., 2009; Alperi & Figueras, 2010; Senderovich et al., 2012).

Two new clinical species, Aeromonas taiwanensis and Aeromonas sanarellii, recovered from wound infections of hospitalized patients in Taiwan (although phenotypically misidentified as A. hydrophila and A. caviae, respectively) were recently discovered by sequencing the rpoD gene (Alperi et al., 2010a). Both species were described on the basis of a single strain (their type), and these were the only known strains until two recent publications reported four isolates of A. sanarellii and one of A. taiwanensis in waste water in Portugal (Figueira et al., 2011), and a strain of A. taiwanensis recovered from the faeces of a female patient with diarrhoea in Israel (Senderovich et al., 2012). Isolates of the species A. sanarellii and A. taiwanensis were recorded in the course of a new study that investigated the prevalence of Aeromonas populations in chironomid egg masses by culture and by real-time PCR methods (unpublished data). Considering the clinical relevance of these species, the present study describes for the first time the virulence genotypes and antibiotic susceptibility of these new species recovered from this new habitat and provides key phenotypic traits for their identification.

Materials and methods

Aeromonas sampling, identification and typing

Sampling for Aeromonas spp. populations was carried out in chironomid egg masses found in a waste stabilization pond in northern Israel between April and September 2009 using previously described procedures (Senderovich et al., 2008). Crushed egg masses were spread on M-Aeromonas agar (Biolife, Italy) for 24 h at 30 °C. Yellow, smooth, rounded colonies that were suspected Aeromonas species were then subcultured on Luria broth (LB) agar (Himedia, India). For each sample, about 15 Aeromonas isolates were identified to the species level using rpoD gene sequencing, according to Soler et al. (2004). To observe the existence or not of clonally related isolates, DNA typing was carried out with the enterobacterial repetitive intergenic consensus PCR (ERIC-PCR) technique using the primers and conditions described by Versalovic et al. (1991). Patterns with one or more different bands were considered different genotypes.

Phenotypic characterization

In all A. sanarellii and A. taiwanensis strains, 24 phenotypic tests (Supporting information, Tables S1 and S2) were evaluated using conventional methods at 30 °C for 24 h up to 7 days as previously described (Abbott et al., 2003; Alperi et al., 2010b) with the exception of utilization of citrate, which was determined using the Simmons's method (Cowan & Steel, 1993), and nitrate reduction (MacFaddin, 1976). Acid production from the carbohydrates d-mannose, d-sorbitol, d-cellobiose, d-raffinose, salicin and arbutin was evaluated as described in Alperi et al. (2010b). These, and additional carbohydrate fermentations (Table S2), were also carried out using the API 50CH system (BioMérieux) for 48 h at 30 °C. In addition, the β-galactosidase activity, production of hydrogen sulphide from cystein and the use of several carbohydrates as sole carbon and energy sources (Table S2) were evaluated using the API 20NE and 20E systems (BioMérieux) for 24 h at 30 °C.

Detection of putative virulent genes

The genes that encode the flagella (fla), lateral flagella (lafA), elastase (ahpB), cytotoxic and cytotonic enterotoxins (act, ast, alt), lipase (pla/lipH3/apl-l/lip), aerolysin/haemolysin (aerA) and serine protease (serine) were screened for all strains of both species using the conditions and primers described previously (Kingombe et al., 1999; Chacón et al., 2004; Sen & Rodgers, 2004; Aguilera-Arreola et al., 2005). The TTSS genes ascF-G and ascV and the genes encoding the toxins delivered by this system, that is, AexT (aexT) and AopP (aopP), were investigated using conditions and primers previously described (Braun et al., 2002; Chacón et al., 2004; Fehr et al., 2006). Aeromonas strains known to be positive were used as controls for all reactions. Additionally, some positive and negative PCR results were confirmed by repeating the experiment, and some positive results were also verified by sequencing the obtained amplicon.

Susceptibility to antimicrobial agents

Susceptibility testing of the strains was carried out using 19 antimicrobials listed in Table 2 using the MicroScan WalkAway-40 automated method.

Results and discussion

Aeromonas identification and typing

A total of eight (6.2%) isolates of A. sanarellii and 3 (2.3%) of A. taiwanensis were identified by sequencing the rpoD gene among the characterized 129 Aeromonas isolates (unpublished data) recovered from chironomid egg masses found at a waste stabilization pond in northern Israel (Fig. 1). This finding adds more knowledge to the diversity of Aeromonas present in this specific ecological habitat, as only the species A. aquariorum, A. caviae, A. veronii and A. hydrophila have been found previously in association with chironomids (Senderovich et al., 2008; Figueras et al., 2011c). Only two of the eight A. sanarellii isolates were clonally related (identical rpoD sequence and ERIC profile) (Fig. 1 and Fig. S1). Although some isolates (11A9B, 16A19C, 16A21C) showed an identical rpoD sequence, their ERIC and virulence profiles (Fig. S1 and Table 1) were different, indicating that they belonged to different strains. As only a fragment of 524 nucleotides (nt) of the rpoD gene was analysed, the nonclonally related isolates with an identical rpoD sequence could exhibit variations in other nonsequenced regions of the gene. Interspecies similarity (based on the 524 nt of the rpoD gene) between A. taiwanensis and A. sanarellii strains was 92.8–95.0%, while intraspecies similarity was 97.5–99.8% and 98.3–100%, respectively. These results agree with the interspecies cut-off value of 97% proposed for this gene by Soler et al. (2004). The rpoD sequence of the A. taiwanensis strain H53AQ1 recovered from faeces of a patient living in the same area (Senderovich et al., 2012) grouped with the environmental strains (Fig. 1) but it differed in 16–23 nucleotides, which indicates that they were not clonally related. Although no epidemiological relationship could be established in this case, the same Aeromonas clone that caused diarrhoea had been isolated from drinking water in other studies (Khajanchi et al., 2010; Pablos et al., 2010). Recently, four A. sanarellii and one A. taiwanensis isolates were recovered from waste water in Portugal (Figueira et al., 2011), which could have originally come from human faeces similar to the A. taiwanensis strain reported by Senderovich et al. (2012) in Israel. Considering this, waste water could have been the dispersion route of both bacterial species to natural water environments such as those inhabited by chironomids. Either these nonbiting midges or the waste water could be the source of the contamination of drinking water with Aeromonas.

Figure 1.

Unrooted neighbor-joining phylogenetic tree derived from rpoD gene sequences (524 nt) showing the relationships of Aeromonas isolates from chironomid egg masses to the strains A. taiwanensis CECT 7403T, H53AQ1 (clinical isolate), A. sanarellii CECT 7402T and other currently known species of Aeromonas. Numbers at nodes indicate bootstrap values (percentages of 1000 replicates above 40%). Bar, 0.02 substitutions per nucleotide position.

Table 1. Prevalence of virulence genes in A. taiwanensis and A. sanarellii strains
Isolate pla/lipH3/apl-1/lip serine asc-V ascF-G lafA aexT
  1. a

    Two of the eight A. sanarellii isolates were clonally related (equal rpoD sequence and ERIC profile).

  2. CECT (Colección Española de Cultivos Tipo, Spanish-type culture collection).

  3. All the strains were positive for ahpB, aerA and fla genes and negative for ast, act, alt and aopP genes.

A. taiwanensis
CECT 7403T++++++
10A4A
4A21C+++++
4A7A+++++
A. sanarellii
CECT 7402T+
11A9B
16A19C++
16A21C+++
16A15C++
15A7A+
10A3Aa+
10A1Aa+
15A5A

Genetic identification on the basis of the rpoD gene has revealed that the most abundant species in patients suffering from diarrhoea in Israel were as follows: A. caviae (65%), A. veronii (29%) and A. taiwanensis (6%) (Senderovich et al., 2012). This identification approach provides results equal to those obtained when using two or more housekeeping genes (Figueira et al., 2011; Figueras et al., 2011a, b), and once more, it has proven to be a reliable method. More studies from other geographical regions using a similar reliable approach will help to establish the true prevalence of these still poorly known Aeromonas species.

Phenotypic characterization

The biochemical traits observed for A. taiwanensis and A. sanarellii, which include both variable and stable characters when compared with those originally described only on the basis of the type strains, enabled the phenotypic diversity of these two species to be defined for the first time. In addition, it reveals which of the tests is more valuable for their recognition. Among the tests carried out, acid production of d-cellobiose and growth at 45 °C in sheep blood agar were the ones that differentiated both of these species from their closest relative A. caviae (Table S1). However, based on previously published results, it must be considered that only about 85% of A. caviae strains produce acid from cellobiose (Figueras et al., 2009). Furthermore, the use of citrate as a sole carbon source might discriminate A. sanarellii from A. taiwanensis and A. caviae. Both A. sanarellii and A. taiwanensis can also be differentiated from A. hydrophila by the Voges–Proskauer test, gas production from glucose and growth at 45 °C in sheep blood agar, all positive for A. hydrophila but negative for the other two species (Table S1). Despite diagnostic characteristics being provided, unequivocal identification can only be granted on the basis of some housekeeping genes, such as the rpoD, as previously demonstrated for these and other Aeromonas species (Alperi et al., 2010a; Figueras et al., 2011b, c). Some characteristics, such as lysine decarboxylase for A. sanarellii and acid production from raffinose for A. taiwanensis, were originally described as negative and positive, respectively, on the basis of the type strains (Alperi et al., 2010a). However, this has proven to vary after testing more strains of each species. Other variable phenotypic responses were observed among strains of the same species (Table S2). According to the API database, the isolates might belong to A. hydrophila/A. caviae/A. sobria, with a 67–98.4% certainty (Table S2).

Virulence genotypes

All A. sanarellii and A. taiwanensis strains carried the genes that encode aerolysin/haemolysin (aerA), elastase (ahpB) and flagella (fla), but lipase genes (pla/lip/lipH3/apl-1/lip) were only detected in 75% and 66.6% of A. taiwanensis and A. sanarellii strains, respectively (Table 1). The lateral flagella (lafA) and serine genes were detected in 75% and 25% of the A. taiwanensis strains but did not amplify in any of the A. sanarellii strains (Table 1). The TTSS genes (ascF-G and ascV), which are considered to encode an important virulence factor, were detected in 75% of A. taiwanensis strains, as did the aexT gene encoding the AexT toxin delivered by this system. Only 33.3% of the A. sanarellii strains possessed the TTSS genes but none of the strains were positive for the aexT gene (Table 1). The PCR results showed that the virulence genotype depended upon the strain despite A. taiwanensis strains bearing more virulent genes than A. sanarellii. None of the strains of either species were positive for the cytotoxic (act) and cytotonic enterotoxin genes (ast, alt) or for the gene of the AopP toxin secreted by the TTSS (Table 1). The same results were obtained for the act, alt, ast and fla genes with the A. taiwanensis stool strain H53AQ1 previously isolated in Israel (Senderovich et al., 2012). In a previous study, we discovered strains of an important and emergent clinical species, A. aquariorum, in chironomid egg masses, and they were found to have a higher prevalence of the alt (90.9%), act (27.3%) and TTSS genes (81.3%) but a lower prevalence of ahpB (81.8%), lipase (54.4%) and fla (27.4%) genes (Figueras et al., 2011c) than the currently investigated strains of A. sanarellii and A. taiwanensis. These strains do not have the alt and ast genes, but all harbour the ahpB and fla genes and have a prevalence of 33.3% and 75% for the TTSS genes and 66.6% and 75% for the lipase gene, respectively (Table 1).

Detection of virulence genes by PCR only provides an indication of the presence or absence of genes; further studies are required to prove whether such genes are indeed functional and contribute to their virulence. Furthermore, PCR assays may also produce false-negative reactions due to inhibition or to the presence of variability in the targeted DNA sequence, although the positive and negative controls included in each assay in the present study gave the expected results (unpublished data).

Susceptibility to antimicrobial agents

All strains were sensitive to 12 of the 19 antimicrobials tested and were resistant to ampicillin, as expected, but also to cefalotin (Table 2). Both species showed a varying susceptibility to several antimicrobials ranging from 25 to 77.7% and a similar susceptibility against all the antimicrobials tested except for cefazolin for which 44% of A. sanarellii were susceptible and all strains of A. taiwanensis were resistant (Table 2). This is the first antimicrobial susceptibility data presented for the species A. sanarellii and A. taiwanensis. The results of this study agree with previous reported data that indicated that most Aeromonas clinical strains, belonging to several species, were sensitive to amikacin, gentamicin, aztreonam, cefepime, ceftazidime, cefotaxime and ciprofloxacin (Overman & Janda, 1999; Vila et al., 2003; Tena et al., 2007; Awan et al., 2009; Senderovich et al., 2012), those therefore being the most active antibiotics for A. sanarellii and A. taiwanensis. The 100% sensitivity to imipenem found for the new species agrees with the data previously reported for other Aeromonas species (Vila et al., 2003; Senderovich et al., 2012) and was higher than results (65–67%) found by Overman & Janda (1999). In fact, in a recent study, we discovered that imipenem-resistant strains showed an over-expression of the imiS gene, encoding a chromosomal carbapenemase, and this was probably induced in vivo after treatment of a urinary tract infection with amoxicillin–clavulanic acid (Sánchez-Céspedes et al., 2009). Furthermore, strains in this study showed a susceptibility to cefoxatin (69.2%) and amoxicillin–clavulanic acid (30.8%) that was similar (70% and 27%, respectively) to the results reported by Senderovich et al. (2012) for the Aeromonas strains responsible for causing diarrhoea, among which A. taiwanensis was reported. Susceptibility to ciprofloxacin, cefalotin and trimethoprim–sulfamethoxazole was the characteristic antimicrobial profile of the group of 15 Aeromonas isolates that embraced those of A. sanarellii (n = 4, but three from the same genotype) and A. taiwanensis (n = 1) obtained from waste water in Portugal (Figueira et al., 2011), results which agree with those from the chironomid strains.

Table 2. Susceptibility of A. sanarellii and A. taiwanensis strains to 19 antimicrobial agents
Antimicrobial agent% SusceptibilityBreakpoints
A. sanarellii (n = 9)A. taiwanensis (n = 4)SR
  1. a

    Intermediate resistance.

  2. S, sensitive; R, resistant.

Amikacin100100≤16≥64
Amoxicillin–clavulanic acid33.325≤8/4≥32/16
Ampicillin00≤8≥16
Aztreonam100100≤8≥32
Cefalotin00≤8≥32
Cefazolin440≤8≥32
Cefepime100100≤8≥32
Cefotaxime100100≤8≥64
Cefoxatin66.675≤8≥32
Ceftazidime100100≤8≥32
Cefuroxime77.775≤8≥32
Ciprofloxacin100100≤1≥4
Ertapenem0a0a≤2≥8
Gentamicin100100≤4≥16
Piperacillin–tozobactam100100≤16/4≥128/4
Tigecycline100100≤1≥4
Tobramycin100100≤4≥16
Trimethoprim–sulfamethoxazole100100≤2/38≥4/76
Imipenem100100≤4≥16

In conclusion, this study shows the presence of A. sanarellii and A. taiwanensis strains in chironomid egg masses, from where they might disseminate to humans through the drinking water supply. Strains of both species bear TTSS genes, among other virulent determinants, and antibiotics such as amikacin, aztreonam, cefepime, cefotaxime, ciprofloxacin, cefalotin, trimethoprim–sulfamethoxazole, gentamicin, ceftazidime and imipenem should be considered potential candidates in the fight against infection produced by these species.

Acknowledgements

The authors thank C. Núñez for her technical assistance and I. Pujol for advice. This work was supported in part by the project with reference AGL2011-30461-C02-02 by the Ministerio de Ciencia e Innovación (Spain).

Ancillary