First Report of Plasmid-Mediated Quinolone Resistance qnrA1 Gene in Klebsiella pneumoniae Isolate of Animal Origin

Introduction

S ince the introduction of quinolones, resistance of the Enterobacteriaceae to these agents has emerged independently many times and has become common and worldwide (Strahilevitz et al., 2009). Plasmid-mediated quinolone resistance (PMQR), which always coexisted with horizontally transmissible genetic elements and facilitated the emergence of mutational quinolone resistance, has made extensive inroads among organisms of clinical importance for both humans and animals (Strahilevitz et al., 2009). Different PMQR mechanisms have been recognized: qnr (qnrA, qnrB, qnrS, qnrC, and qnrD), aac (6′)-Ib-cr, and qepA, and their current prevalence is a problem (Strahilevitz et al., 2009). Investigations of PMQR determinants in veterinary clinical isolates have been previously recorded in China (Yue et al., 2008; Ma et al., 2009), but their prevalence in healthy animal commensal bacteria, which can be reservoirs of antibiotic resistance genes transferable to pathogens, remains largely unknown. qnrA was rarely identified from animals (Strahilevitz et al., 2009). To our knowledge, the presence of qnrA in Klebsiella pneumoniae of animal origin has not been recorded.

One (0.51%) qnrA1-carrying K. pneumoniae isolate, strain GDKA1 was detected in a total of 195 Enterobacteriaceae isolates from healthy chicken in five Guangdong farms in China in 2007. In this study, we characterized the isolate and the genetic environment of qnrA1.

Materials and Methods

Results and Discussion

Susceptibility testing showed that strain GDKA1 was resistant to ampicillin, sulfamethoxazole, spectinomycin, chloramphenicol, and tetracycline (Table 1). In contrast, this strain was susceptible to cefotaxime (MIC, 0.047 μg/mL) and ciprofloxacin (MIC, 0.19 μg/mL) and intermediate to nalidixic acid (MIC, 12 μg/mL). A transconjugant T_GDKA1 was selected during the mating experiment, and hybridization with qnrA1-specific probes revealed that the qnrA1 gene was located in pGDKA1 (data not shown). Tranconjugant T_GDKA1 showed a little higher quinolone MICs (Table 1) than E. coli J53 Az^R but not as high as to be defined as resistant according to the MIC susceptible breakpoint ≤1 mg/L and resistant breakpoint ≥4 mg/L for CLSI. This could be explained by the fact that qnrA1 gene just confers resistance to nalidixic acid and decreased suscpetibility to fluoroquinolones. The donor strain GDKA1 showed more resistance to nalidixic acid than its transconjugant T_GDKA1, indicating that there must be some other mechanism employed by the donor. Sequencing of quinolone resistance determining regions revealed that the donor had an S83I mutation in gyrA, whereas the recipient, transconjugant strain did not contain mutations associated with quinolone resistance. Resistances to spectinomycin, ampicillin, tetracycline, and sulfamethoxazole were cotransferred (Table 1).

A 7, 522-bp continuous DNA sequence obtained by digestion of pGDKA1 with EcoRI revealed that similar to many other plasmids reported (Wang et al., 2003; Mammeri et al., 2005; Rodríguez-Martínez et al., 2007), qnrA1 was located in a sul1-type integron structure and was embedded downstream of an insertion sequence ISCR1 element. The quinolone resistance gene was located immediately upstream of a truncated version of qacEΔ1/sul1, which is identical to that identified in pQR-1 of E. coli Lo from a patient at Bicêtre Hospital (Mammeri et al., 2005). Downstream of the qacEΔ1/sul1 structure, insertion sequence IS26 was found between a kanamycin resistance aph(3')-I gene with a relic of the Tn1696-like transposase tnpA1696 gene and a truncated version of the samB gene with plasmid partition genes parB and parA.

Amplification experiments using primers targeting intI1 and qnrA1 detected an integron on pGDKA1. Upstream of qnrA1, this integron harbored two gene cassettes, dfrA27 and aadA2, encoding resistance to trimethoprim and aminoglycosides, respectively. dfrA27 was first reported from E. coli in China (Wei et al., 2009), and this is the first identification of dfrA27 in Enterobacteriaceae isolates of animal origin. DNA analysis showed that the dfrA27 and aadA2 cassettes were expressed from a weak version of Pc promoter region, and low level of transcription was speculated. Genes flanking qnrA1 in pGDKA1 and other representative plasmids are shown in Figure 1.

OPEN IN VIEWER

FIG. 1.

Genes flanking qnrA1 in pGDKA1 and other representative plasmids. Open reading frames are shown as arrows or horizontal boxes with an arrow indicating the orientation of the coding sequence with the gene name above the corresponding boxes. (GenBank accession numbers for pQR1, In36, pGM252, and pQKp274H (Lavilla et al., 2008) are AY655485, AY259085, DQ831140, and EF682136, respectively.)

To our knowledge, the qnrA1 determinants previously described for complex sul1-type integrons were all inserted between ISCR1 and a second copy of qacEΔ1/sul1 in the same orientation as the transposes of ISCR1 (Strahilevitz et al., 2009). In pGDKA1, the genetic environment of qnrA1 was generally homologous to In36 except for the gene cassettes in the upstream integron and absence of ampR and the downstream qacEΔ1, which was truncated. The same truncated qacEΔ1 downstream of qnrA1 was also identified in plasmid pQR1 (Mammeri et al., 2005). In contrast, the genes following the second copy of qacEΔ1/sul1 were variable. In pMG252 carried by a K. pneumoniae strain from Birmingham, AL (Jacoby, 2005), a second copy of ISCR1 was present, whereas in In36 on the plasmid pHSH1 of E. coli isolated from China (Wang et al., 2003), orf5 was identified instead. Further, many other resistance genes, truncated genes, gene relics, and mobile elements such as IS6100 and IS26 were often found in these genome regions, thus indicating frequent insertion and deletion events. Similar genetic environments of qnrA1 in Enterobacteriaceae isolates of both human and animal origin might demonstrate a similar target site of recombination events mediating incorporation of the quinolone resistance determinant. This could, to some extent, reflect homological mechanisms of qnrA1 distribution. Additionally, qnr genes providing low-level quinolone resistance facilitate the recovery of chromosome-encoded target mutants resulting in higher-level resistance (Martínez-Martínez et al., 1998), so they could at least partly be responsible for the increasing quinolone resistance in field strains.

In many Enterobacteriaceae strains, genes for antibiotic resistance and pathogenicity are usually plasmid borne. Thus, it will be a further threat to public health when such plasmids are conjugative (Lawley et al., 2004). Additionally, intestinal bacteria not only exchange resistance genes among themselves but might also cause bacteria passing through the colon to acquire and transmit antibiotic resistance genes (Salyers et al., 2004). Antibiotic resistant bacteria of animal origin may pass antibiotic resistance genes to bacteria that normally reside in humans (Salyers and Shoemaker, 2006). The present of qnrA1 on a conjugative plasmid of health animal commensal bacteria widened the dissemination of qnrA1 gene and should be worthy of note.