First detection of plasmid-mediated, quinolone resistance determinants qnrA, qnrB, qnrS and aac(6')-Ib-cr in extended-spectrum beta-lactamase (ESBL)-producing Enterobacteriaceae in Budapest, Hungary.

Sir, Quinolone resistance in Enterobacteriaceae usually results from mutations in genes coding for chromosomally encoded type II topoisomerases, efflux pumps or porin-related proteins. Recently, transferable, plasmid-borne, quinolone resistance genes—qnrA, qnrB, qnrS, cr variant of aac(60)-Ib, qepA and oxqB—have been observed in clinical isolates, more frequently among strains producing plasmid-mediated, extended-spectrum b-lactamases (ESBLs). The objective of this study was to determine the prevalence of qnrA, qnrB, qnrS and aac(60)-Ib-cr genes in ESBL-producing clinical isolates of 70 Escherichia coli, 101 Klebsiella pneumoniae, 5 Citrobacter freundii and 61 Enterobacter spp., isolated in two microbiological laboratories in Budapest by collecting samples from seven different hospitals and clinics from 2002 until 2006. The prevalence of ESBL production among Enterobacteriaceae was between 1% and 28% during this period. All of the ESBL-producing organisms were tested for the presence of qnrA, qnrB, qnrS and aac(60)-Ib genes by PCR. To distinguish the qnr gene alleles, restriction enzymes were used and the result was further confirmed by sequencing. The aac(60)-Ib-cr was identified by digestion with BstF5I. The prevalence of qnr determinants was relatively high among K. pneumoniae (8%) isolates and low in the case of E. coli (1.4%) and in Enterobacter spp. (0%). One ESBL-producing C. freundii isolate (out of five) was qnr-positive. qnrA1, qnrB2 and qnrS1 were detected and qnrA1 was the most prevalent (3%) (Table 1). Sixty-three (26.6%) of all ESBL-producing isolates were positive for aac(60)-Ib, of which 19 (8% of all)—16 K. pneumoniae and 3 E. coli—had the aac(60)-Ib-cr variant. None of the qnr-positive strains harboured the aac(60)-Ib-cr variant. The MICs of quinolones were determined by a broth microdilution method according to the CLSI. Conjugation assays were performed for all of the qnr-positive isolates. The qnrA1-positive strains—six K. pneumoniae and one E. coli— were resistant to nalidixic acid, norfloxacin, ciprofloxacin and levofloxacin with very high MIC values (Table 1). The conjugation assay, which was successful in three qnrA1-positive strains, confirmed that the presence of qnrA1 gene alone causes only low-level quinolone resistance. The MIC values for the transconjugants were more than 8-fold higher for nalidixic acid, 132-fold higher for norfloxacin and 66to 132-fold higher for ciprofloxacin, compared with E. coli J53 Rif (nalidixic acid, 4 mg/L; norfloxacin, 0.06 mg/L; and ciprofloxacin, 0.12 mg/L). However, the transconjugants still showed only low-level quinolone resistance, suggesting that the high-level quinolone resistance in the original isolates may be related to the coordinating action of several resistance mechanisms. The qnrB gene was detected in two isolates: one K. pneumoniae and one C. freundii. K. pneumoniae M95 harboured both qnrA1 and qnrB2 genes, but only qnrA1 was transferable by conjugation, suggesting that the two genes have different genetic backgrounds. The ciprofloxacin and levofloxacin MIC values were the highest in K. pneumoniae M95, compared with the other qnrA1-positive strains, which might be explained by the parallel qnrA1 and qnrB2 production of the original isolate. The transfer of qnrA1 alone did not result in transconjugants with greatly altered quinolone susceptibility. The C. freundii M141 harbouring only the qnrB2 gene was susceptible to the fluoroquinolones tested with MICs at least 8-fold lower than the susceptibility breakpoints recommended by the CLSI. Other authors have also observed that QnrB is present in isolates with a wide range of quinolone MIC values, including full susceptibility. The wide range of MIC values for the qnr-positive strains highlights that the detection of the qnr genes in clinical strains might be very difficult phenotypically. The qnrS1 gene was detected only in a K. pneumoniae strain, which was resistant to most of the tested fluoroquinolones. However, the norfloxacin (16 mg/L) and ciprofloxacin (4 mg/L) MIC values were the breakpoints recommended by the CLSI. The levofloxacin MIC value (4 mg/L) was between the two CLSI breakpoints (8 and 2 mg/L). Of the 19 aac(60)-Ib-cr-positive isolates, 6 (32%) were susceptible to both ciprofloxacin and levofloxacin. The presence of aac(60)-Ib-cr does not result in large increases in MIC values of quinolones; however, it substantially increases the frequency of selection of chromosomal mutants. The characteristics of the b-lactamase enzymes—isoelectric points, PCR and sequencing data—produced by the qnr-positive strains were determined. In our study, the presence of qnrA1 gene was connected to SHV-5, SHV-12 and CTX-M-15 enzymes (Table 1). The presence of qnrB2 was connected with SHV-12. We could not determine the genotype of the ESBL enzyme in the qnrB2-positive C. freundii strain. The qnrS1 gene was found together with SHV-2 ESBL. We report the first appearance of plasmid-mediated quinolone resistance in Hungary. The prevalence was low: 3% for qnrA, 0.8% for qnrB, 0.4% for qnrS and 8% for aac(60)-Ib-cr genes, and K. pneumoniae contained most plasmid-borne resistance alleles. Research letters