Diversity of Clavulanic Acid-Inhibited Extended-Spectrum β-Lactamases in Aeromonas spp. from the Seine River, Paris, France

Abstract

Environmental Aeromonas sp. isolates resistant to ceftazidime were recovered during an environmental survey performed with water samples from the Seine River, in Paris, France, in November 2009. Selected isolates were identified by sequencing of the 16S rRNA and rpoB genes. PCR and cloning experiments were used to identify broad-spectrum-β-lactamase-encoding genes and their genetic context. Clavulanic acid-inhibited extended-spectrum-β-lactamase (ESBL) genes were identified in 71% of the Aeromonas sp. isolates. A variety of ESBL genes were detected, including bla_VEB-1a, bla_SHV-12, bla_PER-1, bla_PER-6, bla_TLA-2, and bla_GES-7, suggesting an aquatic reservoir of those ESBL genes. Moreover, the repeated elements and different insertion sequences were identified in association with the bla_PER-6 and the bla_VEB-1a genes, respectively, indicating a wide diversity of mobilization events, making Aeromonas spp. a vehicle for ESBL dissemination.

Bacteria with intrinsic or acquired resistance to antibiotics are commonly found in aquatic environments, where Pseudomonas, Serratia, and Aeromonas are commonly identified (2, 17). Recent studies reported that the bla_TEM-1 gene was detected in most of the Gram-negative isolates resistant to ampicillin recovered from lakes in Brazil or from wastewater treatment plants in China (24, 40). Furthermore, several studies identified extended-spectrum β-lactamases (ESBLs) from water samples. For instance, the clinically relevant ESBL PER-1 was identified from Aeromonas media in a Swiss lake (32), and CTX-M-14 and TEM-52 were detected in Escherichia coli recovered from a South Korean river (21). Besides, nonclinically relevant Ambler class A β-lactamases TLA-2 and PER-6 were identified from uncultured bacteria (43, 12) and Aeromonas allosaccharophila isolates, respectively (15).

In addition, other studies reported on the identification of carbapenem-hydrolyzing β-lactamases in the environment, such as the Ambler class A β-lactamases IMI-2 and BIC-1 from Enterobacter asburiae (1, 5) and Pseudomonas fluorescens (14), respectively; the Ambler class B β-lactamase VIM-2 from Pseudomonas pseudoalcaligenes (41); and the class D β-lactamase OXA-23 from Acinetobacter baumannii (13).

The present study was designed to evaluate the diversity of broad-spectrum β-lactam resistance determinants in Aeromonas species in water samples from the Seine River, Paris, France. The working hypothesis was that Aeromonas spp. could be an important reservoir for ESBLs.

MATERIALS AND METHODS

Bacterial strains and plasmids.

Three water samples (100 ml) were collected from approximately 0.2 m below the water surface in sterile bottles in November 2009 and filtered through nitrocellulose membranes (pore size, 0.45 μm; Millipore), and the filters were resuspended in 3 ml of sterile water. Aliquots (0.1 ml) were plated on MacConkey agar plates supplemented with ceftazidime (2 μg/ml), and 10-fold dilutions were plated (0.1 ml) in parallel on antibiotic-free MacConkey agar plates for bacterial counts. Samples were processed on the day of collection. E. coli DH10B (Invitrogen, Life Technologies, Cergy-Pontoise, France) was used as the host for cloning and transformation experiments as previously described (42). Azide-resistant E. coli J53 was used as the recipient in mating-out assays (32). The kanamycin-resistant plasmid pBKCMV (Invitrogen) was used as a cloning vector. Bacterial cultures were grown in Trypticase soy (TS) broth at 37°C for 18 h.

Antibiotic susceptibility testing.

The antibiotic susceptibility profiles of 80 ceftazidime- or imipenem-resistant isolates were determined by the agar dilution method, and the results were interpreted according to the CLSI guidelines (7). AmpC overproducers were identified by testing susceptibility to ceftazidime on Mueller-Hinton agar plates supplemented with cloxacillin (250 μg/ml). Detection of ESBL production was carried out by the double-disk synergy test (DDST) (19).

Identification of ceftazidime-resistant isolates and genotyping.

Isolates were identified by the API 32GN system (bioMérieux, Marcy l'Etoile, France) and/or sequencing of the 16S RNA and/or rpoB or gyrB genes by using previously described primers (for 16S, primers 27F and 1492R [16]; for gyrB, primers gyrB3F and gyrB14R [17]; for rpoB, primers rpoBLAPS and LAPS27 [26]). Amplicons were sequenced and their sequences were compared to those in the nucleotide data library using leBIBI software (http://umr5558-sud-str1.univ-lyon1.fr/lebibi/lebibi.cgi) in order to determine their closest phylogenetic relatives. The genetic diversity of the ESBL-producing Aeromonas sp. isolates was assessed by pulsed-field gel electrophoresis (PFGE) using XbaI, as previously described (21). The chromosomal fingerprints were compared by eye and assigned to PFGE types and subtypes (44).

PCR amplification for detection of ESBL genes, analysis of genetic environment, and sequencing.

Specific primers were used under standard PCR conditions to detect ESBL-encoding genes, namely, bla_TEM (9), bla_SHV (30), bla_CTX-M (30), bla_GES (33), bla_PER-1/-2 (31, 35), bla_PER-6 (15), bla_VEB (39), and bla_BEL (34). The genetic environments of the bla_PER-1, bla_TLA-2, and bla_SHV genes were determined by PCR mapping using previously published specific primers together with primers designed to anneal at the 5′ and the 3′ conserved segments (CSs) of class 1 integrons (35, 12). The genetic environments of bla_PER-6, bla_VEB-1a, and bla_GES-7 were precisely determined by cloning those genes into plasmid pBKCMV. For that purpose, BamHI-, SacI-, ScaI-, or SalI-restricted DNAs were ligated into restricted plasmid pBKCMV and introduced into E. coli DH10B by electroporation as described previously (42). Recombinant plasmids were selected on TS agar plates containing ceftazidime (2 μg/ml) and kanamycin (30 μg/ml). All enzymes for DNA manipulations were used according to the recommendations of the supplier (GE Healthcare, Orsay, France). Recombinant plasmids were sequenced by using combinations of universal T3 and T7 primers and specific gene primers (Table 1), designed on the basis of sequences obtained on an Applied Biosystems sequencer (ABI 3130).

TABLE 1.

Oligonucleotides used for PCR amplification and sequencing

Primer	Sequence (5′ to 3′)	Reference or source
VEB-inv1	GTCAGTTTGAGCATTTGAATAC	37
VEB-inv2	AAGCGTATTTGTTGCAGAGTC	37
IS26out	CCCACAGAATGATGTCACGC	This study
aadB-B	CGCATATCGCGACCTGAAAGC	This study
Av9_1 A	AAGTGCCTAGAGATTGTTTCC	This study
TnpF-A	AACTACTGCTGGAGTATATGG	This study
IS4B	TAAATCATCAGGCTACAGGTC	This study
IS6100-A	ATTTCTACCTGTCGCCGACC	This study

Genetic support of β-lactamase-encoding genes.

Plasmid extraction was performed by the Kieser technique (20). E. coli NCTC 50192, harboring four plasmids of 154, 66, 38, and 7 kb, was used as a size marker for plasmids. Transformation assays were performed by electroporation with plasmid extracts from the identified positive isolates, and E. coli DH10B was used as the recipient strain. Selection was performed on agar plates supplemented with ceftazidime (2 μg/ml). Conjugation experiments were carried out as described previously (32). The I-CeuI restriction enzyme (Ozyme; New England Biolabs, Saint-Quentin-en-Yvelines, France), which digests a 26-bp sequence in the rrn genes for the 23S large-subunit rRNA, was used to determine whether the β-lactamase gene had a chromosomal location, and the fragments were separated by pulsed-field gel electrophoresis (36). DNA-DNA hybridization of plasmid extracts was performed after a Southern transfer onto a Hybond N+ nylon membrane (GE Healthcare), as previously described (42). Labeling of the probe and signal detection were carried out using an enhanced chemiluminescence labeling and detection kit, according to the manufacturer's instructions (GE Healthcare). PCR-based replicon-typing (PBRT) analysis was performed as described by Carattoli et al. (6). The 17 primer pairs targeting the FIA, FIB, FIC, HI1, HI2, I1-I, L/M, N, P, W, T, A/C, K, B/O, X, Y, and FII replicons were used in separate PCRs.

Nucleotide sequence accession numbers.

The nucleotide sequence data reported in this paper will appear in the GenBank nucleotide database under accession numbers HM453325, HM370390 to HM370393, HM453326, HM453327, and HM626463.

RESULTS

Bacterial counts and susceptibility testing.

The total bacterial count on MacConkey agar was 1.5 × 10⁵ CFU/liter of Seine River water, of which 2.6% grew on ceftazidime-agar plates (3.6 × 10³ CFU/liter). A total of 73 isolates were selected from ceftazidime-containing plates and subjected to identification. Fifty-six percent (n = 41) of the isolates that were resistant or of reduced susceptibility to ceftazidime were identified as Aeromonas spp., and the others were mostly Pseudomonas spp. and Stenotrophomonas maltophilia. Among those 41 ceftazidime-resistant Aeromonas spp., 29 (71%) displayed an ESBL phenotype, as evidenced by DDST. These isolates were identified as A. punctata, A. allosaccharophila, A. veronii, and A. media. Our study then focused on those 29 ESBL-producing Aeromonas isolates. The MICs of β-lactams, the patterns of resistance to other antimicrobial agents, and the molecular typing were used to evaluate that collection, and the findings are presented in Table 2. The 29 ceftazidime-resistant isolates were resistant to ticarcillin and had reduced susceptibility to cefotaxime and cefepime, but none of them were resistant to imipenem (Table 2). Addition of clavulanic acid significantly decreased the MICs of ceftazidime and aztreonam (Table 2). More than 86% of these isolates were resistant to sulfonamides and to nalidixic acid, although Aeromonas spp. are naturally fully susceptible to quinolones (Table 2) (22).

TABLE 2.

Resistance pattern of ESBL-producing Aeromonas sp. isolates

Isolate	Aeromonas species	Clone	ESBL identified	MIC (μg/ml)b									Non-β-lactam(s) to which isolate was resistant
				CTX	FEP	CAZ	CAZ + CLAV	TIC	TIC + CLAV	ATM	ATM + CLAV	IPM
1	A. punctata	A1a	VEB-1a	8	2	32	0.25	128	16	8	<0.06	0.12	KAN, TOB, GEN, AMK, TET, NAL, CIP, FOF, CHL, SXT
2	A. punctata	A2	VEB-1a	8	4	32	0.25	128	1	8	<0.06	0.12	KAN, TOB, GEN, AMK, TET, NAL, CIP, FOF, CHL, SXT
3	A. punctata	A3	VEB-1a	4	1	32	0.06	128	0.5	2	<0.06	<0.06	KAN, TOB, AMK, TET, NAL, CIP, SXT
4	A. punctata	A3	VEB-1a	8	2	32	0.12	64	2	4	<0.06	<0.06	KAN, TOB, TET, NAL, CIP, SXT
5	A. punctata	A4	VEB-1a	4	2	16	0.25	128	128	2	<0.06	0.12	SXT
6	A. punctata	A4	VEB-1a	8	4	32	0.25	256	128	8	<0.06	0.12	SXT
7	A. allosaccharophila	B1	VEB-1a	4	2	16	0.06	32	2	2	<0.06	0.25	KAN, TOB, TET, NAL, SXT
8	A. allosaccharophila	B2	VEB-1a	8	2	16	0.06	128	16	4	<0.06	0.12	KAN, TOB, TET, NAL, CIP, SXT
9	A. veronii	C1	VEB-1a	0.25	0.12	2	0.06	16	4	0.5	<0.06	0.25	SXT
10	A. veronii	C2	VEB-1a	1	0.25	4	0.06	32	4	0.25	<0.06	0.5
11	A. media	D1	VEB-1a	32	4	32	0.5	512	512	16	<0.06	<0.06	NAL
12	A. allosaccharophila	B3	SHV-12	0.25	0.25	2	0.06	128	1	1	<0.06	0.25	NAL
13	A. allosaccharophila	B3	SHV-12	1	1	4	0.06	512	2	2	<0.06	0.25	NAL
14	A. allosaccharophila	B4	SHV-12	4	0.25	4	0.06	256	2	2	<0.06	0.25	NAL
15	A. allosaccharophila	B4	SHV-12	1	0.5	4	0.12	512	2	2	<0.06	0.25	NAL
16	A. allosaccharophila	B4	SHV-12	1	0.25	4	0.06	256	2	2	<0.06	0.25	NAL
17	A. allosaccharophila	B4	SHV-12	1	0.25	4	0.06	256	2	1	<0.06	0.25	NAL
18	A. veronii	C3	SHV-12	4	1	4	0.06	512	16	4	<0.06	0.25	NAL
19	A. veronii	C3	SHV-12	1	1	4	0.06	512	8	4	<0.06	0.25	NAL
20	A. veronii	C3	SHV-12	4	2	8	0.06	128	8	2	<0.06	0.25	NAL
21	A. media	D2	SHV-12	8	2	32	0.5	512	256	16	<0.06	<0.06	NAL, RIF, SXT
22	A. punctata	A5	PER-1	8	4	32	0.25	256	64	16	<0.06	0.12	KAN, NAL, SXT
23	A. punctata	A6	PER-1	8	4	32	0.25	128	128	16	<0.06	0.12	NAL, SXT
24	A. media	D3	PER-1	16	4	64	0.25	128	16	64	<0.06	0.12	KAN, NAL
25	A. veronii	C4	PER-6	4	2	64	0.25	128	128	32	<0.06	0.12	TET, NAL, SXT
26	A. allosaccharophila	B5	PER-6	4	0.5	2	0.06	512	64	4	<0.06	0.5	TET, NAL, SXT
27	A. allosaccharophila	B5	PER-6	4	0.5	8	2	512	512	8	0.25	0.5	TET, NAL
28	A. veronii	C4	GES-7	2	0.5	16	2	64	32	0.25	<0.06	0.25	KAN, TOB, NAL
29	A. allosaccharophila	B6	TLA-2	2	1	8	0.12	128	64	1	<0.06	0.12	NAL, RIF

^aThe letter A was arbitrarily attributed to A. punctata, B to A. allosaccharophila, C to A. veronii, and D to A. media. Different numbers correspond to different clones.

^bAbbreviations: AMK, amikacin; ATM, aztreonam; CHL, chloramphenicol; CAZ, ceftazidime; CIP, ciprofloxacin; CLA, clavulanic acid; CTX, cefotaxime; FEP, cefepime; FOF, fosfomycin; GEN, gentamicin; IPM, imipenem; KAN, kanamycin; NAL; nalidixic acid; RIF, rifampin; SXT, trimethoprim-sulfamethoxazole; SSS, sulfonamide; TET, tetracycline; TIC, ticarcillin; TOB, tobramycin.

Identification of acquired β-lactamase genes.

PCRs and sequencing performed on the 29 Aeromonas sp. isolates exhibiting an ESBL phenotype, using primers specific for β-lactamase genes, identified the bla_VEB-1a (n = 11), bla_SHV-12 (n = 10), bla_PER-1 (n = 3), bla_PER-6 (n = 3), bla_TLA-2 (n = 1), and bla_GES-7 (n = 1) genes (Table 2).

Clonal relationship.

PFGE analysis performed on the 29 Aeromonas sp. isolates showed 19 distinct genotypes, including 4 clonally related A. allosaccharophila isolates harboring an identical bla_SHV-12 gene (Table 2; data not shown). The ESBL genes were therefore identified in a large diversity of Aeromonas species (Table 2).

Genetic support of bla_SHV-12 gene.

PCR mapping and sequencing revealed that the bla_SHV-12 gene, identified in 10 Aeromonas sp. isolates, was preceded by insertion sequence (IS) IS26, as previously reported (10). Southern blot hybridization of plasmid DNA extracted from the bla_SHV-12-positive Aeromonas sp. isolates using the corresponding probe did not give any signal, suggesting a chromosomal location of that ESBL gene (data not shown).

Genetic support of bla_TLA-2 gene.

PCR mapping to determine the genetic environment of the bla_TLA-2 gene in the single A. allosaccharophila isolate was performed with combinations of primers located in the 5′ CS, bla_TLA-2, and 3′ CS sequences. The bla_TLA-2 gene was not part of a gene cassette but was located in a class 1 integron, located on a ca. 50-kb plasmid. Mating-out transfer of the plasmid harboring the bla_TLA-2 gene to E. coli remained unsuccessful, suggesting that this plasmid has a narrow host range or was nonconjugative (43).

Genetic support of bla_PER-1 gene.

The bla_PER-1 gene identified in three Aeromonas sp. isolates was preceded by insertion sequence ISPa12 and was followed by ISPa13, thus being part of a composite transposon, Tn1213, as previously reported (35). Southern blot hybridization of plasmid DNA extracts recovered from the bla_PER-1-positive Aeromonas sp. isolates using the corresponding probes did not give any signal, suggesting a chromosomal location of the bla_PER-1 gene (data not shown).

Genetic support of bla_PER-6 gene.

The bla_PER-6 gene was identified in an A. veronii isolate and in clonally related A. allosaccharophila (isolates 26 and 27). Southern blot hybridization of plasmid DNA extracted from these isolates, using a bla_PER-6-specific probe, did not give any positive signal, suggesting a chromosomal location of this gene (data not shown). Analysis of the sequences surrounding the bla_PER-6 gene by PCR mapping showed that they were different from those initially identified in A. allosaccharophila recovered from the Seine River in 2009 (15). Thus, cloning of the bla_PER-6 gene from A. veronii isolate 25 was performed, and sequencing revealed that the bla_PER-6 gene interrupted the Tn1721 feature, including the tetA, tetR, and tnpAΔ5 genes, as previously identified in the A. allosaccharophila AL-1 isolate (15) (Fig. 1). Interestingly, the bla_PER-6 gene was bracketed by 135-bp repeated elements (Res; named Re1Av and Re2Av for A. veronii) sharing 84 and 80% nucleotide sequence identities with the reverse complementary sequence of the previously reported Re1 that bracketed the bla_VEB-1a gene (3). Re1-like sequences have been identified in association with bla_VEB-like genes in P. aeruginosa (3), Providencia stuartii (4), and Proteus mirabilis (48). In silico analysis of previously reported sequences surrounding the bla_PER-6 gene in A. allosaccharophila AL-1 showed a downstream-located Re1-like copy (namely, Re1Aa, for A. allosaccharophila) (Fig. 1). Re1Aa shares 80% nucleotide sequence identity with Re1Av.

FIG. 1.

Schematic representation of the genetic environment of β-lactamase genes. The genetic environment of the bla_PER-6 gene in A. veronii isolate 25 and A. allosaccharophila isolates 26 and 27 (A) is compared to the previously identified structure in A. allosaccharophila AL-1 (B) (15). Arrows indicate the direction of transcription of the coding regions. Repeated elements ReAv and ReAa are indicated as short black arrows. The open reading frame named ASA_0310 is a sugar-phosphate isomerase from A. salmonicida; TnpAΔ5 is a truncated relaxase/helicase. (C) Comparison of the sequences of Re1Av, Re2Av, and Re1Aa. Dashes indicate gaps introduced to optimize alignment, and periods indicate nucleotides identical to those in the Re1Av sequence.

Genetic support of bla_GES-7 gene.

Southern blot hybridization of plasmid DNA extracted from A. veronii isolate 28, using a probe specific for bla_GES-like genes, showed that the bla_GES-7 gene was located on a ca. 60-kb plasmid. Partial sequencing of this natural plasmid (named Tf28) showed that the bla_GES-7 gene was located inside an original structure in which the 3′ extremity of the class 1 integron was absent (Fig. 2). The bla_GES-7 gene was in the form of a gene cassette, located at the first position of a class 1 integron and followed by the aacA4 gene. Downstream of the aacA4 gene, there was no 3′ CS sequence identified, but instead, the orf15 and mobA gene arrays were identified. The structure containing orf15, mobA, orf2, orf3, and orf4 was previously identified in plasmid pRSB101 (43). Between the aacA4 and orf15 genes, the terminal inverted repeat (IRt) of Tn402 was identified. The terminal inverted repeat of class 1 integrons (IRi) is present upstream of the gene encoding the IntI1 integrase and has the same boundary with the adjacent sequence as in pRSB101 (43). In order to determine whether the backbone of the Tf28 plasmid was the same as that of pRSB101, PCR amplification was performed with primers located in the transporter module (orf11, encoding a resistance-nodulation-cell division [RND] efflux membrane fusion protein) and in the repA gene. The absence of PCR amplification suggested a different backbone. Mating-out assays between A. veronii isolate 28 and E. coli DH10B, performed at 37°C or at 30°C, remained unsuccessful, suggesting that the Tf28 plasmid was not self-transferable.

FIG. 2.

Genetic environment of the bla_GES-7 gene in A. veronii isolate 28. Arrows indicate the direction of transcription of the coding regions. The orf2, orf3, orf4, orf15, and mobA genes previously identified in plasmid pRSB101 (43) are indicated as gray arrows. IRt of the Tn402-like sequence and IRi are also shown.

Genetic support of bla_VEB-1a gene.

The bla_VEB-1a gene was identified in 11 Aeromonas sp. isolates. Plasmid extraction and Southern hybridization showed that the bla_VEB-1a gene was chromosomally located in A. punctata isolates 1 and 2, whereas it was plasmid located in all the other Aeromonas sp. isolates. The sizes of the plasmids harboring the bla_VEB-1a gene varied from 30 kb in A. media isolate 11 to 50 kb in A. allosaccharophila isolates 7 and 8 and up to 170 kb in all other isolates. Investigation of the genetic context with primers designed from the previously identified structures failed. The identification of bla_VEB-1-like genes has been reported from the family Enterobacteriaceae, P. aeruginosa, and A. baumannii recovered worldwide (27, 36, 39). The bla_VEB-1 gene is often part of a gene cassette located in class 1 integrons. However, in some cases, the bla_VEB-1 and bla_VEB-1a genes have been identified in association with Res in P. aeruginosa and the Enterobacteriaceae (3, 28, 38). More recently, the bla_VEB-1a gene has been identified in association with two copies of insertion sequence ISCR2 (37). PCR assays revealed that none of these previously described genetic structures was bracketing the bla_VEB-1a gene in the Aeromonas isolates recovered in the present study. Sequence analysis of the obtained recombinant or natural plasmids harboring the bla_VEB-1a gene revealed some similarities among all genetic environments, as shown in Fig. 3. In particular, partial sequences of a gene encoding a TnpF-like putative integrase from Acinetobacter genomospecies 3 (GenBank accession no. GQ926879) and of insertion sequences IS6100 and IS26 were identified (Fig. 3). Mating-out assays performed with Aeromonas isolates 3 to 11 as donors and E. coli J53 as the recipient, either at 37°C or at 30°C, remained unsuccessful, suggesting that these plasmids were not self-transferable. PBRT analysis showed that the bla_VEB-1a-bearing plasmids did not belong to any of the tested Inc groups that correspond to plasmids identified in the Enterobacteriaceae (6).

FIG. 3.

Genetic environment of the bla_VEB-1a gene in A. punctata isolates 1 to 6, A. allosaccharophila isolates 7 and 8, A. veronii isolates 9 and 10, and A. media isolate 11. The coding genes are represented by arrows indicating their translation orientation, the left and right inverted repeats of IS elements are shown by filled and empty triangles, respectively, and the core sites are indicated as circles. IS6100, IS26, and tnpF-like genes (gray rectangles) are recovered in most structures. TnpF-like is a putative integrase from Acinetobacter genomospecies 3 (GenBank accession no. GQ926879). TnpA-11 (arrow with gray squares) from A. punctata 1 to 4 shares 83% amino acid sequence identity with the sequence of transposase_11 from A. punctata (GenBank accession no. YP_067863). Av9_1 and Av9_2 from A. veronii 9 are proteins of unknown function.

DISCUSSION

Aeromonas spp. were the predominant bacteria recovered in this study, representing 56% of the colonies obtained on ceftazidime-containing plates. The ceftazidime resistance that we have observed was mainly related to the expression of ESBLs (71%). Many ESBLs that are not frequently identified among clinically relevant Gram-negative isolates in Paris (29), namely, VEB-1a, GES-7, TLA-2, and PER-1 and PER-6, have been identified in this screening. It was actually unexpected to identify such an important diversity of class A ESBLs in Aeromonas spp., which are species known to already possess intrinsic Ambler class B, C, and D β-lactamase genes. This raises the question of why Aeromonas spp. are so often hosts for those antibiotic resistance determinants. This waterborne location may be one part of the explanation. Rivers are subjected to the effects of several human activities. We considered this environment to be a good model for studying the prevalence and molecular diversity of genes that might represent a potential risk for human health. Aeromonas spp. living in water were demonstrated to be a potent reservoir of antibiotic resistance genes.

The bla_VEB-1a gene was identified in 38% of Aeromonas spp. resistant to or with reduced susceptibility to ceftazidime recovered from the Seine River, whereas bla_VEB-1 has been identified in France only once in E. coli (25) and several times in A. baumannii (27, 36). The wide dissemination of this gene among so diverse Aeromonas species is of interest, since it likely indicates that the natural reservoir of that clinically relevant ESBL gene is waterborne or raises the possibility of Seine River contamination. In addition, our study emphasizes that the dissemination of bla_VEB-1-like genes is related to a variety of genetic structures.

In addition, other rare β-lactamase-encoding genes were identified here, in particular, the bla_TLA-2, bla_PER-6, and bla_GES-7 genes. The bla_TLA-2 gene had already been identified on plasmid pRSB101, which had been recovered from wastewater treatment plants in Germany but which has never been reported from clinical samples (43). The bla_PER-1 gene is mainly reported in Europe and Asia (45, 46, 23) and to a lesser extent in North Africa (18), but it has very rarely been identified in France (35). The bla_PER-6 gene has been identified once from the Seine River in A. allosaccharophila (15). Similarly, the bla_GES-7/IBC-1 gene has been detected only from clinical Enterobacteriaceae in Greece (11, 47) and in Brazil (8) but had never been identified in France.

Analysis of the genetic structures surrounding all these ESBL genes in environmental Aeromonas isolates revealed an important diversity of genetic supports and genetic environments. Noticeably, the identification of ReAv repeated elements in association with the bla_PER-6 gene further underlines that those still underinvestigated genetic elements might play a relevant role in antibiotic resistance gene dissemination and mobilization.