Molecular characterization of tetracycline- and quinolone-resistant Aeromonas salmonicida isolated in Korea.

The antibiotic resistance of 16 Aeromonas (A.) salmonicida strains isolated from diseased fish and environmental samples in Korea from 2006 to 2009 were investigated in this study. Tetracycline or quinolone resistance was observed in eight and 16 of the isolates, respectively, based on the measured minimal inhibitory concentrations. Among the tetracycline-resistant strains, seven of the isolates harbored tetA gene and one isolate harbored tetE gene. Additionally, quinolone-resistance determining regions (QRDRs) consisting of the gyrA and parC genes were amplified and sequenced. Among the quinolone-resistant A. salmonicida strains, 15 harbored point mutations in the gyrA codon 83 which were responsible for the corresponding amino acid substitutions of Ser(83)→Arg(83) or Ser(83)→Asn(83). We detected no point mutations in other QRDRs, such as gyrA codons 87 and 92, and parC codons 80 and 84. Genetic similarity was assessed via pulsed-field gel electrophoresis, and the results indicated high clonality among the Korean antibiotic-resistant strains of A. salmonicida.

Introduction

Aeromonas (A.) salmonicida is a pathogen that causes furunculosis and bacterial septicemia in a broad variety of fish, and is thus responsible for significant economic losses in the global aquaculture industry [37]. Recently, antibiotic-resistant A. salmonicida strains have been recognized as a serious concern owing to their potential health risks to humans and animals [32,33]. Among the antibiotics utilized in the treatment of furunculosis, both tetracycline and quinolone resistance have been widely documented [10,31]. Tetracycline-resistant strains of A. salmonicida are suspected to be the source of tetracycline resistance dissemination in the aquatic environment because tet genes, the determinants of tetracycline resistance, are generally encoded on plasmids [1,32,33]. Quinolone resistance is a potential public health threat since quinolones are also utilized for the treatment of Aeromonas infections in humans [15,16]. Quinolone resistance in Gram-negative bacteria is primarily attributable to mutations in the quinolone-resistance determining regions (QRDRs) consisting of the gyrA and parC genes, which are the subunits of the target enzymes of quinolones, DNA gyrase subunit A and topoisomerase IV, respectively [2]. The presence of the qnr gene which is associated with the plasmid-mediated quinolone-resistance, or efflux pumps is also known to be associated with mid- to low-level quinolone resistance [6,30].

Antibiotic resistance has been previously reported in several aquatic bacteria isolated in Korea including Edwardsiella tarda [17], Streptococcus iniae, and Streptococcus parauberis [29]. However, the antibiotic resistance of Aeromonas spp. has not previously been addressed. Therefore, in this study we evaluated the antimicrobial susceptibility and clonal relationship in A. salmonicida isolated from both cultured fish and the environmental water in Korea. In particular, the genetic determinants of tetracycline and quinolone resistance were assessed via (i) the detection of tetA to E, (ii) the detection of plasmid-encoded qnr genes, and (iii) the analysis of point mutations in QRDRs.

Materials and Methods

Bacterial isolation and culture conditions

Between 2006 and 2009, sixteen strains of A. salmonicida were isolated from a variety of samples from fish and sewage water from two private aquariums and three salmonid farms in Korea (Table 1). Two reference strains were purchased from the American Type Culture Collection (ATCC, USA): A. salmonicida subsp. salmonicida ATCC 33658 (ASS) and A. salmonicida subsp. masoucida ATCC 27013 (ASM). A. salmonicida isolates were first screened using a Vitek System 2 (bioMérieux, France). All strains of A. salmonicida were stored in tryptic soy broth (Difco, USA) with 10% glycerol at -80℃ and sub-cultured for 48 h on tryptic soy agar (Difco, USA) at 22℃. To assess strain purity, single colonies were selected and sub-cultured three times, and the resulting bacterial cells were harvested for further experiments.

Table 1

Aeromonas (A.) salmonicida strains used in this study

Antimicrobial susceptibility test

Antimicrobial susceptibility tests were conducted via broth micro-dilution methods according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI), and ASS was utilized as a quality control bacterial strain [7,8]. Since cut-off values have not been determined for all antibiotics, three references [7,8,24] were used for interpretation, as was the case in other previous reports [2,5,29]. Seven antimicrobials were diluted in following ranges: ampicillin (0.06 to 32 µg/mL), enrofloxacin (0.002 to 4 µg/mL), florfenicol (0.12 to 64 µg/mL), gentamicin (0.06 to 32 µg/mL), oxolinic acid (0.004 to 8 µg/mL), oxytetracycline (0.03 to 16 µg/mL), and trimethoprim-sulfamethoxazole (0.03/0.6 to 2/38 µg/mL). All antimicrobials were purchased from Sigma-Aldrich (USA). The antimicrobials were serially diluted two-fold in cation-adjusted MHB (CAMHB; Difco, USA) and 100 µL volumes of the dilutions were placed into 96-well micro-titer plates. The inoculations were prepared as follows: 18 strains of A. salmonicida were adjusted to a McFarland value of 0.5 and diluted 10-fold with CAMHB. With the addition of 5 µL of inocula into each micro-titer wells, the final cell densities were adjusted to 5 × 105 CFU/mL. In all cases, two control wells without antimicrobials or inocula were maintained. After 44 to 48 h of incubation at 22℃, the lowest concentration of antibiotics that visibly inhibited bacterial growth was defined as the minimal inhibitory concentration (MIC). The MIC results of A. salmonicida subsp. salmonicida were used to classify the strains as resistant or sensitive in accordance with the cut-off values established by Miller et al. [24] and the guidelines of M49-A [7] and M31-A3 [8].

DNA extraction and polymerase chain reactions (PCR)

Genomic DNA was extracted by harvesting the cells with sterile water followed by 10 min of boiling. After 3 min of centrifugation at 10,000 ×g, the supernatants were collected and 1 : 100 dilutions in sterile water were utilized as a PCR template. All isolates were confirmed to be A. salmonicida using Fer-3 and Fer-4 PCR primers [3]. Subspecies were determined by A. salmonicida subsp. salmonicida-specific PCR with MIY1 and MIY2 primers [4,26] and confirmed by 16S rRNA sequencing at Macrogen (Korea). Two multiplex PCR procedures were conducted as previously described to amplify the five tetracycline resistant genes (tetA to E) [27] and to detect the qnr genes [5]. The QRDRs of the gyrA and parC genes were detected using the following primers: ASGYRA1, ASGYRA2, ASPARC3, and ASPARC4 [10]. The primers used in this study are shown in Table 2.

Table 2

PCR primers used in this study

Sequence analysis

Sequencing was conducted by Macrogen (Korea) and the sequences were analyzed with the AlignX tool in the Vector NTI program (Invitrogen, USA). BLAST searches were conducted using both the blastn and blastx algorithms provided by the National Center for Biotechnology Information (NIH, USA).

Pulsed-field gel electrophoresis (PFGE)

Harvested bacterial cells were diluted with cell suspension buffer (100 mM Tris-HCl and 100 mM EDTA, pH 8.0) up to an optical density of 1.0 at 600 nm. A cell suspension volume of 100 µL was mixed with an equal volume of 1.6% SeaKem Gold agar (FMC Corporation, USA) and solidified in a 100 µL plug mold. The plugs were then incubated for 2 h with 1 mg/mL of lysozyme (Sigma-Aldrich, USA) at 37℃ and treated with 1 mg/mL of proteinase K (Sigma-Aldrich, USA) at 50℃ for 8 h. DNA in the plugs was digested for 18 h with 30 U of SpeI (New England Biolabs, USA) at 37℃ and electrophoresed in 1.0% SeaKem Gold agarose gel with a CHEF-Mapper III PFGE system (Bio-Rad, USA). The running conditions were 6 V/cm at 14℃ for 22 h, and the pulse times were 1.5 to 25 sec. The Lambda ladder PFG marker (New England Biolabs, USA) was included as a size marker. The gels were stained with ethidium bromide and photographed under UV trans-illumination. The genetic relationships among isolates were analyzed with Bionumerics software (Applied Maths, Belgium) and the clusters were determined using the UPGMA algorithm with the 70% Dice-coefficient of similarity (2.0% position tolerance).

Results

Bacterial identification

The 16 isolates were successfully identified as A. salmonicida using Vitek System 2 and species-specific PCR (Table 1). Among the various categories of biochemical tests in Vitek System 2, four (D-glucose, D-mannitol, and sucrose fermentation tests, and H2S production test) which were known to be different between subspecies of A. salmonicida [14] were focused on in this study. The ASS strain and 14 isolates (AS01 to AS15 except AS03) were positive in the D-glucose and D-mannitol fermentation tests and negative in the sucrose fermentation and H2S production tests. On the other hand, AS03 was positive only in a sucrose fermentation test and ASM was positive in all the four tests. Additionally, the AS16 strain showed different biochemical characteristics compared to the others; this strain was positive in the D-glucose, D-mannitol and sucrose fermentation tests, and negative in the H2S production test.

Among the 16 isolates, 14 strains were confirmed by PCR to be A. salmonicida subsp. salmonicida. The other two strains were confirmed to be A. salmonicida subsp. achromogenes (AS03) and A. salmonicida subsp. flounderacida (AS16), as their 16S rRNA sequences showed 100% homology with the 16S rRNA gene of A. salmonicida subsp. achromogenes strain 870626-1/1C (GenBank accession No. AM296505.1) and A. salmonicida subsp. flounderacida strain HQ010320-1 (GenBank accession No. AY786177.1), respectively. All the isolates and reference strains of A. salmonicida used in this study are shown in Table 1.

MICs

The MIC values of the A. salmonicida isolates are shown in Table 3. Among the 16 Korean isolates, eight oxytetracycline-resistant strains and sixteen oxolinic acid-resistant strains were detected. Enrofloxacin resistance was noted in only one isolate (AS16). A total of 9 multidrug-resistant (MDR) strains were observed: 7 strains (AS09 to AS15) that were resistant to oxytetracycline and oxolinic acid, one strain (AS03) that was resistant to ampicilin and oxolinic acid, and one strain (AS16) that was resistant to five antibiotics (ampicillin, gentamicin, oxytetracycline, enrofloxacin, and oxolinic acid). Moreover, strain AS16 exhibited a high level of resistance to both enrofloxacin (≥4 µg/mL) and oxolinic acid (≥8 µg/mL) although the other 15 quinolone-resistant strains were susceptible to enrofloxacin (≤0.03 µg/mL) and showed low-level oxolinic acid resistance (1~2 µg/mL).

Table 3

Minimal inhibitory concentrations (MICs), tetracycline resistance (tet) genes, mutations in quinolone-resistance determining regions (QRDRs) in the A. salmonicida strains

*Nucleotide changes and corresponding amino acid substitutions are shown in bold. †Not amplified. AM: ampicillin, GM: gentamicin, SXT: trimethoprim-sulfamethoxazole, FFC: florfenicol, OCT: oxytetracycline, ENR: enrofloxacin, OA: oxolinic acid, R: resistance.

tet genes in A. salmonicida isolates

The tetA gene (211 bp) was detected in seven isolates (AS09 to AS15) while the tetE gene (744 bp) was detected in strain AS16 (Fig. 1). The amplified PCR products were sequenced and aligned with the tet gene sequences from GenBank. All amplifed tetA fragments in this study showed 100% homology with the tetA gene of pRAS1, a drug resistance plasmid of A. salmonicida (GenBank accession No. AJ517790.2). The tetE gene, which was detected in AS16, showed 100% homology with the tetE gene of A. salmonicida subsp. salmonicida A449 plasmid 4 (pAsa4; GenBank accession No. CP000645.1) and A. salmonicida plasmid pYA90644 (GenBank accession No. DQ366299.1).

Fig. 1

Multiplex PCR assay of tetracycline resistance genes (tetA of 211 bp and tetE of 744 bp) in two reference strains and 16 isolates of Aeromonas (A.) salmonicida. Lane M: molecular mass marker; lane 1 to 18: strains AS01, AS02, AS03, AS04, AS05, AS06, AS07, AS08, AS09, AS10, AS11, AS12, AS13, AS14, AS15, ASS, ASM, and AS16, respectively. Marker sizes (bp) are indicated.

qnr genes and codon mutations in the QRDRs of A. salmonicida isolates

The gyrA (663 bp) and parC (418 bp) genes of QRDRs were successfully amplified from all 16 isolates and the ASS reference strain (Table 3). The amplified products were sequenced and their corresponding amino acid sequences were aligned with the sequences of gyrA (GenBank accession No. L42453.1) and parC (GenBank accession No. AF473701.1) of A. salmonicida ATCC 14174. AS03 and ASS strains had no point mutations while the AS16 strain harbored a Ser83→Asn83 substitution in gyrA codon 83 (Table 3). Other 14 isolates showed Ser83→Arg83 substitutions in the same loci. Additionally, AS16 had a single nucleotide mutation (AAA→AAG) at the parC codon 80 without an amino acid substitution. No other substitutions were detected on gyrA codon 87 (Asp87) and 92 (Leu92), or parC codon 80 (Lys80) and 84 (His84). The qnr gene was not detected in any of the A. salmonicida strains in this study.

Strain typing by PFGE

All A. salmonicida strains utilized in this study were clustered into four types based the PFGE results (Fig. 2). The ASM, AS03, and AS16 were divided into type A, B, and C, respectively. The other 14 A. salmonicida subsp. salmonicida isolates and ASS were classified into the same cluster designated as type D.

Fig. 2

Pulsed-field gel electrophoresis profiles of 18 A. salmonicida strains and UPGMA dendrogram. The vertical dotted line denotes a hypothetical node of 70% Dice coefficient of similarity (2.0% position tolerance).

Discussion

Based on Bergey's Manual of Determinative Bacteriology [14], at least three subspecies were identified among all 16 isolates using the biochemical results of Vitek System 2. All of these interpretations were in concordance with the subspecies-specific PCR and 16S rRNA sequencing results. Interestingly, the AS16 strain that showed distinctive biochemical characteristics (D-glucose (+), D-mannitol (+), sucrose (+) and H2S production (-)) was identified as the recently-reported A. salmonicida subsp. flounderacida (GenBank accession no. AY786177.1). Based on these results, we were able to confirm the presence of three subspecies of A. salmonicida (subsp. salmonicida, subsp. achromogenes and subsp. flounderacida) among those Korean isolates.

Considering the widespread use of tetracycline and quinolones in the aquaculture industry [12], the resistance of A. salmonicida to the two antibiotic classes was the focus of this study. According to the epidemiological cut-off values for A. salmonicida for oxytetracycline and oxolinic acid [24], eight oxytetracycline-resistant strains were detected, and sixteen isolates were oxolinic acid-resistant. Although enrofloxacin is one of the quinolones, like oxolinic acid [8], only one isolate (AS16) noted resistance to it. Interestingly, ampicillin resistance was detected only in three isolates (AS03, AS16, and ASM) although there have been some reports showing that A. salmonicida is naturally resistant to narrow-spectrum β-lactams [7]. One isolate showed resistance to gentamicin, and all strains were found to be susceptible to florfenicol and trimithoprim-sulfamethoxazole. Tetracycline resistance in A. salmonicida was strictly related to the presence of the tetA and tetE genes. These genes were also detected in other A. salmonicida strains from a variety of fish species in other countries [25,36]. The sequenced tetA and tetE genes in this study showed 100% homology to tetA in pRAS1and tetE in pAsa4. Since the pRAS1 and pAsa4 plasmids can be transferred into or replicate within certain strains of Escherichia coli [31,35], it has been suspected that tetracycline resistance has been disseminated between various bacterial species. The location and transferability of the tetA and tetE genes in A. salmonicida clearly warrants further investigation.

Despite the high levels of activity of quinolones against Aeromonas species [16,18], the number of quinolone-resistant Aeromonas strains has increased [10,34]. In this study, all 16 isolates showed resistance to oxolinic acid, and the AS16 strain was resistant to enrofloxacin. Genetic analysis through sequencing QRDRs revealed that these resistant strains except AS03 harbored point mutations in gyrA codon 83. The AS01 to AS15 (except AS03) strains had Ser83→Arg83 substitutions and the AS16 strain harbored a Ser83→Asn83 substitution. In particular, the AS16 strain having a Ser83→Asn83 substitution exhibited high-level resistance to both oxolinic acid and enrofloxacin. It is well known that quinolone resistance is principally related to mutations of the QRDRs, particularly in gyrA codons 83, 87, and 92, and in parC codons 80 and 84 [10,11]. In addition to previous studies that found Ser83→Ile83 and Ser83→Val83 substitutions on gyrA codon 83 in strains of Aeromonas [2,11,28], here we report two more putative substitutions that might contribute to quinolone resistance. Moreover, based on our results, amino acid substitutions on gyrA codon 83 may affect the level and spectrum of quinolone resistance in A. salmonicida.

In addition to the point mutations in the QRDRs, acquisition of qnr genes or efflux pumps also contribute to quinolone resistance [6,30]. In this study, the AS03 strain showed low-level resistance to oxolinic acid without mutations in the QRDRs. Since no qnr genes were detected, further investigation should be performed for determining whether this strain contains other quinolone-resistant mechanisms such as the efflux pump. The other possibility that could explain the differences in quinolone resistance between the isolates might be subspecies-specific natural resistance mechanisms that have yet to be elucidated; AS03 (A. salmonicida subsp. achromogenes) and AS16 (A. salmonicida subsp. flounderacida) were different subspecies from A. salmonicida subsp. salmonicida strains.

The PFGE results of this study typed isolates at the subspecies level. The genetic heterogeneity between subspecies was consistent with previous reports that described differences between typical and atypical A. salmonicida strains [9,13]. Interestingly, 14 A. salmonicida subsp. salmonicida strains isolated from Korea were found to be distinct from ASS although they were included in the same cluster; this result suggests geographical differences in the distribution of A. salmonicida.

Besides the subspecies classification, most of biochemical results and MDR patterns of the A. salmonicida strains also concurred with the PFGE types. For example, AS16 (type B) was identified as A. salmonicida subsp. flunderacida and resistant to the highest number of antibiotics including ampicillin, gentamicin, oxytetracycline, enrofloxacin, and oxolinic acid. Strains in other PFGE types (type A and D) were identified as A. salmonicida subsp. achromogenes or subsp. salmonicida that were resistant to one or two antibiotics. On the other hand, the similar PFGE pattern of the tetracycline-resistant and susceptible A. salmonicida subsp. salmonicida strains within type D appears to imply horizontal transfer of tet genes among these isolates. These results suggest a potential risk of the spread of MDR strains or dissemination of antimicrobial resistance genes in the Korean aquatic industry.

Thus far, only a few antibiotics are approved for use in animals in the worldwide aquatic industry [19,24]; nevertheless, antibiotic resistance is expected to continue to become more frequent [20-23]. The detection of MDR in Korean strains of A. salmonicida suggests that antibiotic resistance in aquaculture can pose a risk to both humans and animals. Thus, stricter guidelines for the use of tetracycline and quinolones will be necessary to prevent the dissemination and acquisition of antibiotic resistance in aquaculture.