INTRODUCTION

Top Abstract Introduction Materials and Methods Results Discussion References

The motile aeromonads represent a group of ubiquitous aquatic microorganisms which are not generally considered to be primary human pathogens (15). Some species, however, have been isolated from local or generalized human infections (7, 15, 20). By contrast, many members of the group are recognized as primary pathogens to a wide range of cold-blooded animals, in particular to fish (4, 6, 40). In temperate regions with mainly salmonid production, motile Aeromonas species are not commonly associated with disease outbreaks in aquaculture and are thus not directly targeted by treatment with antimicrobial agents.

However, freshwater fish farming does seem to have an impact on environmental Aeromonas spp., as indicated by a previous investigation of antimicrobial resistance at four Danish rainbow trout farms (35). Water, sediment, and fish samples were examined, and because of their ubiquitous distribution in the freshwater environment, the motile aeromonads were selected as bacterial indicators. In addition, members of the genus Aeromonas readily develop single or multiple antimicrobial resistance phenotypes (10, 12, 18, 23-25), and R-plasmids are commonly found (1, 3, 5, 13, 19, 28). Thus, they were well suited for monitoring the incidence of antibiotic resistance, as well as for investigating the conjugative spread of resistance genes in these settings.

Significantly higher proportions of antibiotic-resistant motile aeromonads were detected in the effluent of fish farms than in their inlets (35). In particular, many of the isolates were resistant to high levels of oxytetracycline (OTC), a combination of sulphadiazine/trimethoprim (S/T), or both. Potentiated sulphonamides and oxolinic acid (OXA) are the only antimicrobials licensed for therapeutic use in Danish aquaculture, while the usage of OTC is restricted, requiring dispensation in every case. The small amounts of OTC used in the fish farms thus did not appear to correlate with the observed high OTC resistance frequencies among environmental Aeromonas isolates (35). By contrast, the administration of S/T in fish farms during bacterial disease outbreaks might have promoted the emergence of S/T resistance in the vicinity of the farms.

Both types of resistance have been reported to be encoded by transferable plasmids within the genus Aeromonas (1, 3, 7, 14). Several classes of tetracycline resistance determinants have been described on R-plasmids within this group of bacteria (1, 2, 5, 8, 9, 30). Tetracycline resistance genes are frequently part of transposons, which are able to change their location within the cell, and thus achieve increased mobility, e.g., by inserting into conjugative plasmids (41, 43). Consequently, we decided to investigate the occurrence and distribution of the tetracycline resistance determinant classes A to E among the motile Aeromonas isolates, comparing OTC-resistant isolates with different plasmid profiles and origins.

Furthermore, the aeromonads were screened for the presence of class 1 integrons in order to elucidate the genetic background of the high S/T resistance level. The 3' conserved segment of the integron includes the sul1 and qacE1 genes, encoding sulfonamide- and quaternary ammonium compound resistance, respectively. The 5' conserved segment contains the genes encoding the integration of various numbers of gene cassettes, comprising the variable part of the structure. The gene cassettes usually are antimicrobial resistance genes, and dihydrofolate reductase (dhfr) genes conferring trimethoprim resistance are common (22, 29).

Only few studies have to date addressed the prevalence of class 1 integrons among environmental bacteria (27, 31), and to our knowledge, this is the first report of class 1 integrons within the widely distributed motile Aeromonas species.

Even though class 1 integrons are transposition defective, they are often plasmid borne as they are mobilizable in association with a functional transposon or by transposition proteins supplied in trans (1, 8, 9, 32, 41, 43). Thus, their presence might indicate whether horizontal gene transfer has occurred among the aeromonads found in and around the sampled fish farms, and associated antibiotic resistance genes could explain some of the observed resistance patterns. Subsequent conjugation assays were used to assess the transferability of both tetracycline resistance determinants and class 1 integrons on conjugative plasmids.

	MATERIALS AND METHODS

Top Abstract Introduction Materials and Methods Results Discussion References

Bacterial isolates and MIC testing. Between October 1997 and February 1999, we sampled four fish farms situated along a Danish stream on eleven occasions with monthly intervals (35). Farm 1 was located farthest upstream and thus received no effluents from other fish farms. However, the stream had previously received effluents from a sewage treatment plant and some agricultural areas. Farm 2 was situated close to the outlet of the adjacent upstream fish farm, and farm 4 was one of the last farms downstream. Water and sediment samples were collected from the inlet, outlet, and a pond of each farm as earlier described (35). Samples from gills and skin mucus of two to four fish per farm were also included.

1

1

1

1

1

1

Detection of tetracycline resistance determinants. We decided to test all OTC-resistant isolates for tetracycline resistance determinants A to E because they include the classes that are most commonly described in aeromonads (1, 2, 5, 8, 9) and that are frequently associated with R-plasmids (1, 8, 30). A multiplex PCR assay was used according to the method described by Guardabassi et al. (12). Escherichia coli strains containing the respective tetracycline resistance genes were included (class A, NCTC/50078; class B, HB101/pRT11; class C, DO7/pBR322; class D, C600/pSL106; and class E, HB101/pSL1504). Primers are listed in Table 1. Table 2 shows the distribution of identified tetracycline resistance determinants relative to the origin of the isolates.

Analysis of antimicrobial resistance genes associated with integrons. All isolates and transconjugants were screened for the presence of class 1 integrons with specific primers targeting the conserved 5' and 3' segments of the structure as previously described (34). Thus, the size of a PCR product depends on the number and size of the inserted gene cassettes (Fig. 1 and 2). Salmonella enterica serovar Typhimurium DT104 (9616368) was the positive control strain. PCR products were purified (S-400 HR MicroSpin Columns; Amersham Pharmacia Biotech, Uppsala, Sweden) and sequenced. The nucleotide sequence was determined in both senses of the DNA, using cycle sequencer 373A (Applied Biosystems, Perkin-Elmer, Foster City, Calif.) as reported earlier (34). With five different sizes ranging between about 150 bp (no insert) and 2,900 bp (Fig. 1), two to three representatives of each amplicon were sequenced. Subsequently, a suitable DNA restriction enzyme (MboII) was employed to determine whether equally sized integron amplicons contained the same gene cassettes. Restriction cutting and purification of DNA were performed as described (33), and the resulting fragments were separated by agarose gel electrophoresis.

View larger version (68K): [in this window] [in a new window]   FIG. 1.   Different sizes of PCR products obtained with a primer pair targeting the conserved segments of class 1 integrons in motile aeromonads. Class 1 integrons wereif presentinvariably cotransferred to E. coli on OTC resistance plasmids. Lanes a and m, 100-bp DNA marker; lane b, positive control S. enterica serovar Typhimurium DT104 with 1,000- and 1,200-bp amplicons; lane c, negative control; lane d, "empty" integron with no gene inserts between conserved ends; lanes e and f, isolate 1-75 and the corresponding E. coli transconjugant with 1,400-bp amplicons; lanes g and h, isolate 4-229 and transconjugant with 1,550-bp amplicons; lanes i and j, isolate 2-280 and transconjugant with 2,100-bp PCR products; lanes k and l, isolate 4-221 and transconjugant with 2,900-bp products. kb, kilobases.

View larger version (18K): [in this window] [in a new window]   FIG. 2.   Schematic view of MboII restriction sites () used for comparison of gene cassette content of different sizes of amplicons obtained with primers targeting the conserved segments (shaded areas) of class 1 integrons. Out of 141 integron-positive isolates, 74 belonged to profile I, 13 to profile II, 22 to profile III, and 9 to profile IV. Twenty-three isolates had "empty" integrons with no genes inserted into the variable region. UI, unidentified gene insert.

Plasmid profiling. All aeromonads and E. coli transconjugants were screened for their plasmid content by applying the alkaline lysis method described by Kado and Liu (17) followed by agarose gel electrophoresis. We used the 4.0 version of GelCompar (Applied Maths, Sint-Martens-Latem, Belgium) to analyze the resulting plasmid profiles.

Conjugational gene transfer. All OTC-resistant isolates with large plasmids were included as putative donors in a filter mating assay in order to detect the transfer of R-plasmids to a rifampin-resistant E. coli strain, CSH26Rf (37). Overnight cultures of donor and recipient were adjusted to an optical density at 600 nm of 0.5 with fresh veal infusion broth. Equal volumes (50 µl) of all cultures were mixed on a sterile 0.2-µm-pore-size nitrocellulose filter (Sartorius AG, Göttingen, Germany), which was placed on a veal infusion agar plate (Difco) and incubated at 20°C overnight. Cells were washed off the filter by vortexing in 10-ml sterile 0.9% NaCl solution, and appropriate 10-fold dilutions were prepared. From each dilution, 100-µl aliquots were spread on selective agar plates containing 20 µg of OTC ml¹, 100 µg of rifampin (Bie & Berntsen, Rødovre, Denmark)/ml¹, or both. Donor and recipient were also placed on the double selective plates for mutant detection. All assays were run in duplicate. The identity of transconjugants was confirmed biochemically (oxidase production, amino acids, and Voges-Proskauer reaction), and they were screened for the presence of plasmids as described above.

Statistical methods. The proportions of antibiotic resistance and the respective detected genes among the Aeromonas isolates were computed for each fish farm and all sampling sites. A logistic regression model was employed (proc genmod in SAS version 6.12; SAS Institute Inc., Cary, N.C.) to detect differences between resistance rates from inlet and outlet as well as differences between fish farms. In a few cases where the logistic regression model did not describe the data well, ² tests (with continuity correction) were performed instead.

	RESULTS

Top Abstract Introduction Materials and Methods Results Discussion References

Resistance phenotypes. Three hundred thirteen motile, mesophilic aeromonads were investigated. The most prevalent phenospecies identified were Aeromonas hydrophila (35.3%), A. bestiarum (19%), and A. veronii biovar sobria (15.3%). Fifteen percent of the isolates could not be reliably assigned to one phenospecies. Antibiotic resistance patterns did not vary significantly between the identified phenospecies (data not shown).

Antibiotic resistance genes. Table 2 sums up the distribution of the identified tetracycline resistance determinants and OTC-resistant Aeromonas isolates. It appeared that, while 69% of the aeromonads were OTC resistant, only 30% were carrying one or two of the five tetracycline resistance determinants included in the screening (Table 2). A total of 19 tetA-positive, 39 tetE-positive, and 6 tetD-positive isolates was found. Three isolates contained two tetracycline resistance determinants: tetA and -E, tetA and -D, and tetE and -D. Tetracycline resistance determinants B and C were not detected.

Plasmid profiles and conjugative transfer of resistance genes. The plasmid content of the 313 aeromonads did not seem to vary between the different phenospecies (data not shown). One hundred forty-four (46%) isolates did not contain any plasmids, while 16% did harbor at least one large (>30 kb) plasmid (profile A) (Table 3). The GelCompar analysis yielded three additional clusters: profile B with one small to medium-sized plasmid (2.3 to 20 kb; 42 isolates), profile C with two plasmids between 6.5 and 15 kb (43 isolates), and profile D with three to nine plasmids between 3 and 25 kb (34 isolates). The different plasmid profiles did not seem to vary according to sample matter (water, sediment, and fish) or origin of the isolate (farms or sample site) (data not shown).

	DISCUSSION

Top Abstract Introduction Materials and Methods Results Discussion References

The high OTC resistance levels (69%) detected among the aeromonads in this study were unexpected, considering that this agent has been rarely used for therapeutic purposes in Danish aquaculture since a change in legislation in 1994. DePaola et al. (8) found similarly high proportions of OTC-resistant aeromonads from catfish and their environments (58 to 83%), where the drug was routinely used in medicated feed. Other comparable investigations of motile aeromonads from different freshwater environments report considerably lower tetracycline resistance levels (7, 10, 11, 23, 28, 42). One explanation may be that, once acquired, the resistance genes are maintained within the population, protecting the bacteria from tetracyclines produced by other members of the microflora or residues in agricultural or domestic effluents. Goni-Urriza et al. (10) detected an increase of tetracycline resistance levels of Aeromonas isolates from 0 to 27% in a stream before and after the stream passed a wastewater discharge point. However, this resistance appeared to be entirely chromosomally mediated (10).

More than three different tetracycline resistance determinants occurred among OTC-resistant isolates, and in a few instances more than one determinant was detected in a single strain. DePaola et al. (8) classified over 90% of tetracycline-resistant A. hydrophila isolates from catfish farms as either tetA or tetE positive, and TetA, TetD, and TetE are considered to be the main tetracycline resistance determinants among motile aeromonads (1, 5, 8, 9). However, the majority of OTC-resistant isolates in this study (70%) did not belong to classes A to E (Tables 2 and 3). Thus, the genetic background of OTC resistance among motile aeromonads from this habitat appeared to be rather diverse and to vary locally, possibly as a response to various physical conditions or differences in local genetic exchange processes. Although probably not the only mechanism of horizontal resistance gene transfer, the transfer of R-plasmids is thought to play a major role in the dissemination of OTC resistance in the fish farming environment (5, 30, 39, 40, 41). In our study, there was a positive correlation between OTC-resistant motile aeromonads harboring large plasmids and the presence of tetA. Moreover, the majority of successfully transferred R-plasmids carried the tetA resistance gene. This contrasts with DePaola's findings, in which most donors in successful matings had unidentified tetracycline resistance determinants (8, 9).

Human as well as environmental Aeromonas isolates do often contain conjugative R-plasmids, some of which have been assigned to incompatibility groups C and U (14, 30). Both groups have wide host ranges, yet IncU plasmids have only been detected occasionally in genera other than Aeromonas (14, 44).

Rhodes et al. (30) reported that IncU OTC resistance plasmids in mesophilic aeromonads from hospital effluents and fish farms were closely related to IncU OTC resistance plasmids isolated from the fish pathogen Aeromonas salmonicida and a human E. coli strain. The predominant tetracycline resistance determinant determinant was tetA, and the presence of a complete or truncated form of tetracycline resistance transposon Tn1721 was demonstrated on several of the R-plasmids, thus illuminating yet another mechanism of horizontal dissemination of antimicrobial resistance in these settings (30, 36, 41). The present work demonstrated the common occurrence of class 1 integrons and their resistance gene cassettes on OTC resistance plasmids. Still, there was no evidence of a correlation between the occurrence of integrons and a certain plasmid type or tetracycline resistance determinant. Conversely, there was a close association of class 1 integrons and S/T resistance, as all S/T-resistant isolates contained an integron with a sul1 gene within the 3' conserved region and a dhfr gene cassette insert. Aeromonads from the inlets compared to those from the ponds and outlets of the fish farms were less likely to be OTC, OXA, and S/T resistant (35) and consequently integron positive. Furthermore, Aeromonas isolates from farm 1 were significantly more likely to be S/T resistant and integron positive than isolates from the farms farther downstream, perhaps due to a more frequent use of the drug at this farm during the trial period (35). It may be speculated that the distribution of these genetic elements is enhanced by the frequent use of potentiated sulphonamides in the fish farming environment, including occasional additional resistance gene cassettes like ant(3'')1a or catB2. If such integrons are mobilized onto R-plasmids, indirect selection could contribute to the maintenance of tetracycline resistance genes within the population.

Only few researchers so far have addressed the incidence and spread of class 1 integrons and integron-associated resistance genes in environmental microorganisms (27, 31, 41). Even in clinical settings, the epidemiology of integrons and gene cassettes is not resolved (22, 29, 32, 38, 43). Rosser and al. (31) detected the incidence of class 1 integrons to be 3.6% among gram-negative bacteria from an estuarine environment. Unlike in this study, almost half of the integrons lacked a sul1 gene and about the same proportion did not have any genes inserted into the variable region. Rosser et al. proposed that gene cassettes are excised from the structure in the absence of antibiotic selective pressure or that "empty" integrons represent ancestral elements which have not yet acquired gene cassette inserts.

The abundance of class 1 integrons and the inserted dhfr genes among the motile aeromonads correlates with transient selective pressures exerted by the administration of combined sulphonamide/trimethoprim drugs in freshwater fish farms. Aminoglycosides and chloramphenicol, on the other hand, are not used in Danish aquaculture. Possibly, aquatic motile aeromonads have acquired the gene cassettes by interacting with other microorganisms in soil or domestic effluents (31, 41). Aminoglycoside resistance gene cassettes may also be more stably integrated within integrons, explaining their frequent occurrence and persistence in many bacterial species and habitats (33).

In conclusion, our results point towards a significant effect of aquaculture on the motile aeromonads in and around the investigated fish farms, leading to an increase of OTC, S/T, and OXA resistance levels within this group. High levels of multiresistance (48%) indicated that the horizontal spread of resistance genes has contributed to the dissemination of, in particular, OTC-S/T resistance phenotypes (35), and the finding of class 1 integrons in 45% of the isolates further supported this hypothesis. The diversity of the isolates and characterized genes does not support the clonal spread of antibiotic-resistant aeromonads in and around the fish farms.

Conjugation assays demonstrated that different types of class 1 integrons were cotransferred to E. coli recipients on OTC resistance plasmids. However, it was evident that other genes and transfer mechanisms besides conjugation are probably involved, which should be further investigated in the future. The significance of antibiotic-resistant environmental bacteria has been much disputed (21, 30, 31, 41). Although ubiquitous in freshwater, motile Aeromonas species are not commonly retrieved from human or animal disease in temperate climate zones (15). However, extensive antimicrobial resistance within this group might provide a pool of resistance genes capable of transfer to other water-borne bacteria or fish pathogens. A better understanding of such processes in natural environments is crucial in order to assess the risk of antibiotic resistance among ubiquitous aquatic bacteria such as the motile aeromonads.