Microbiol Immunol 2010; 54: 123–128 doi:10.1111/j.1348-0421.2009.00200.x
ORIGINAL ARTICLE
The prevalence of plasmid-mediated quinolone resistance determinants among clinical isolates of ESBL or AmpC-producing Escherichia coli from Chinese pediatric patients Chenxi Han1 , Yonghong Yang1 , Minggui Wang2 , Aihua Wang1 , Quan Lu3 , Xiwei Xu1 , Chuanqing Wang4 , Lan Liu5 , Qiulian Deng6 and Xuzhuang Shen1 1
Beijing Children’s Hospital, Affiliated to Capital Medical University, Beijing, 100045, 2 Division of Infectious Diseases, Huashan Hospital, Affiliated to Fudan University, Shanghai, 200040, 3 Shanghai Children’s Hospital, Affiliated to Shanghai Jiao Tong University, Shanghai, 200040, and 4 The Children’s Hospital of Fudan University, Shanghai, 201102, 5 Chongqing Children’s Hospital Affiliated to Chongqing Medical University, Chongqing, 400014, and 6 Guangzhou Children’s Hospital, Affiliated to Guangzhou Medical College, Guangzhou, 510120, China
ABSTRACT Three kinds of PMQR determinants (qnr genes, aac(6’)-Ib-cr, and qepA) have been discovered and shown to be widely distributed among clinical isolates. To characterize the prevalence of PMQR determinants in ESBL or AmpC-producing E. coli clinical isolates in Chinese children, a total of 292 ESBL or AmpCproducing E. coli clinical isolates collected from five children’s hospitals in China from 2005 to 2006 were screened for PMQR determinants by PCR. Twenty (6.8%) of the 292 isolates were positive for PMQR determinants. A total of 12 (4.1%) isolates were positive for qnr genes, comprising three positive for qnrA (1.0%), three for qnrB (1.0%), and six for qnrS (2.1%). Twenty-four (8.2%) isolates were positive for aac(6’)-Ib, of which 10 (3.4% of 292) had the –cr variant. There was no qepA gene detected in the isolates. Conjugation revealed that qnr, aac(6’)-Ib-cr, and ESBL-encoding genes were transferred together. Key words ESBL or AmpC-producing Escherichia coli, Pediatrics, PMQR, quinolone resistance.
Quinolones are an important group of antibiotics which are used broadly in adult patients because of their excellent bactericidal activity. Over the past few years however, resistance to these antibiotics has sharply risen in China due to their wide use (1). It is believed that quinolone resistance can only be acquired through a chromosomal mutation that includes the gyrA, gyrB, parC, and parE genes. The gyrA and parC genes are major factors for the mutation of the QRDR. Recently, three PMQR mechanisms have also been described. The first mechanism comprises qnr genes that encode target protection proteins of the pentapeptide repeat family (2, 3). The second
mechanism is the aac(6’)-Ib-cr gene, which encodes a new variant of the common aminoglycoside acetyltransferase. Two single-nucleotide substitutions at codons 102 and 179 in the wild-type allele aac(6’)-Ib enable the gene product to be capable of acetylating and reducing the activity of some quinolones, including norfloxacin and ciprofloxacin (4). The third mechanism involves qepA, a new plasmidmediated gene which encodes an efflux pump belonging to the major facilitator subfamily and is responsible for reduced fluoroquinolone susceptibility (5, 6). Quinolones are restricted for use among children because they are associated with a variety of adverse side
Correspondence Xuzhuang Shen, Beijing Children’s Hospital, Affiliated to Capital Medical University, 56 South Lishi Road, Beijing, 100045, China. Tel/fax: + 86 10 6802 9020; email:
[email protected] Received 31 May 2009; revised 19 November 2009; accepted 15 December 2009. List of Abbreviations: CLSI, Clinical and Laboratory Standards Institute; E. coli, Escherichia coli; ESBL, extended-spectrum beta-lactamase; K. pneumoniae, Klebsiella pneumoniae; LB, Luria-Bertani; MIC, minimum inhibitory concentration; PFGE, pulsed-field gel electrophoresis; PMQR, plasmid-mediated quinolone resistance; QRDR, quinolone resistance-determining region.
c 2010 The Societies and Blackwell Publishing Asia Pty Ltd
123
C. Han et al.
effects. However, our previous study showed there was both a high resistance rate against the quinolones and a high prevalence of PMQR determinants qnr gene among ESBL or AmpC beta-lactamase producing clinical K. pneumoniae isolates from Chinese pediatrics patients (7). The objective of this study was to screen for the presence of PMQR determinants in clinical isolates of ESBL or AmpCproducing E. coli from pediatric patients in China. We found a low carriage rate for qnr genes in those strains. However, we also found a close relationship between PMQR genes and beta-lactamase genes, as well as a high incidence of resistance rate to ciprofloxacin.
MATERIAL AND METHODS Bacterial isolates
From 2005 to 2006 a total of 292 clinical isolates of phenotypic positive ESBL or AmpC beta-lactamase E. coli were collected at five children’s hospitals located in China. Each isolate was obtained from a single patient. The isolates were obtained from patients admitted to the Department of Internal Medicine (240 strains, 82.2%), the Intensive Care Unit (20 strains, 6.8%), and the Department of Surgery (18 strains, 6.2%), and from patients seen by the Outpatient Service (14 strains, 4.8%). These isolates were obtained from the lower respiratory tract (60.6%), urine (21.9%), blood (3.4%), pus (3.1%), and other areas (10.9%). E. coli J53AzR , which is resistant to azide, was used as the recipient strain in the conjugation experiments (8). Screening and confirmation of ESBL and AmpC
All isolates were screened for ESBL production using the double-disk synergy method (9), the results of which were confirmed by the CLSI ESBL disk confirmatory method (CLSI 2003). Insensitivity to cefoxitin was preliminarily regarded as AmpC positive, and a confirmatory experiment was further performed according to the 3-aminophenylboronic acid procedure (Sigma-Aldrich, Milwaukee, WI, USA) and double disk diffusion method (10) (Oxoid, Basingstoke, UK). Quality control was performed by testing E. coli ATCC25922 and K. pneumoniae ACTT700603. Antibiotic sensitivity experiment
MIC of amikacin, ampicillin, ampicillin/clavulanic acid, cefepime, cefotaxime, ceftazidime, ciprofloxacin, gentamicin, and imipenem were determined by the agar dilution method according to the CLSI guidelines. Quality control was performed by testing E. coli ATCC25922. The 124
cefotaxime sodium was from Sigma USA, while all the other antibiotics and the Mueller-Hinton medium were purchased from Oxoid, England. Detection of PMQR determinants, gyrA, parC, gyrB, parE and beta-lactamase-encoding genes
The isolates were investigated for the presence of qnrA, qnrB, qnrS, aac(6’)-Ib, qepA, gyrA, parC, gyrB, parE, bla TEM , bla SHV , and bla CTX−M genes by PCR amplification using the primer sets described previously (5, 11–17). Because only a portion of the QRDR genes was sequenced, the sequenced range of the relevant nucleotide positions of gyrA, parC, gyrB, and parE genes were 193–536, 160– 347, 1051–1424, and 1071–1496, respectively, compared with the DNA sequence of the QRDR genes of E. coli. The Genebank accession numbers are DQ447134.1, NC˙011601.1, NC˙012892.1, and NC˙000913.2, respectively. Amplification of AmpC beta-lactamase was performed according to Perez-Perez and Hanson (18). The colonies were transferred to water through an Eppendorf tube and boiled to prepare DNA templates for PCR. The results of the DNA sequences were compared with the BLAST online search engine from GenBank at the National Center for Biotechnology Information Web site (http://www.ncbi.nlm.nih.gov/BLAST/). Conjugation experiments and plasmid detection
Conjugation experiments were carried out in an LB broth with E. coli J53AzR as the recipient as previously described (19). The PMQR-positive isolates were used for the conjugation experiments. Transconjugants were selected on trypticase soy agar plates containing sodium azide (150 μg/ml) for counterselection and sulfamethoxazole (300 μg/ml) to select for plasmid-encoded resistance. To determine if quinolone resistance was co-transferred, colonies were plated in replica onto trypticase soy agar plates with and without ciprofloxacin (0.06 μg/ml) .The MIC for the transconjugants were measured by an Epsilometer test, and the transconjugants carrying PMQR genes were then confirmed by PCR. Plasmid DNA of the donors and the transconjugants were extracted by alkaline lysis. E. coli V517 (plasmid sizes, 54, 5.6, 5.1, 3.9, 3.0, 2.7, and 2.1 kb) and E. coli J53 plac (plasmid size152 kb) were used as standards. Pulsed-field gel electrophoresis
The genomic DNA of PMQR-producing isolates were analyzed by PFGE after digestion with XbaI (Sangon, Shanghai, China). The DNA fragments were separated by c 2010 The Societies and Blackwell Publishing Asia Pty Ltd
PMQR determinants of E.coli
electrophoresis in 1% agaroseD-5 (TaKaRa, Shiga, Japan) in a 0.5 X tris-borate-EDTA buffer with the clamped homogenous electric fields apparatus (CHEF Mapper XA, Bio-Rad, Hercules, CA, USA) at 14◦ C, 6 volts/cm and with alternating pulses at a 120◦ angle in a 5–35 s pulse time gradient for 19 hr.
RESULTS Prevalence of PMQR determinants and gyrA, parC, gyrB, and parE genes
Twenty (6.8%) of the 292 E. coli isolates were positive for PMQR determinants. A total of 12 (4.1%) of the 292 isolates were positive for qnr genes, comprising three positive for qnrA (1.0%), three for qnrB (1.0%), and six for qnrS (2.1%). A total of 24 (8.2%) isolates were positive for aac(6’)-Ib, of which 10 (3.4% of 292) had the –cr variant. There was no isolates positive for qepA. There was one isolate having both the qnrA and aac(6’)-Ib genes, one with both the qnrB and aac(6’)-Ib-cr genes, one carrying both the qnrS and aac(6’)-Ib-cr genes, and one carrying both the qnrS and aac(6’)-Ib genes. QRDR PCR products were sequenced on both strands, and DNA sequences of gyrA, parC, gyrB, and parE genes for each of the PMQR-positive isolates were compared with the QRDR DNA sequences of E. coli. Among the PMQR-positive strains, gyrA mutations were observed in 5/20 (25%), of which three strains had both Asp87Asn and Ser83leu, one only Asp87Asn, while the third had only Ser83leu. In parC, the mutations showed up in four strains in the same Ser80Ile substitution. There were no mutations in gyrB and parE. (The Genebank accession numbers are DQ447134.1 for gyrA and NC˙011601.1for parC.) Detailed information on these PMQR determinantpositive isolates is presented in Table 1.
Antibiotic susceptibility
Of the 292 ESBL or AmpC-producing E. coli isolates, the rate of resistance to ciprofloxacin was 55% and the intermediate rate was 5%. The rates of resistance to cefotaxime, cefepime, and ceftazidime were 65.4%, 47.6%, and 25.3%, respectively. The rates of resistance to gentamicin, amikacin, and ampicillin/clavulanic acid were 54.8%, 6.2%, and 22.3%, respectively. No isolates were resistant to imipenem. All the tested strains were resistant to ampicillin, and the MIC 50 and MIC 90 were all >512 μg/ml. The MIC 50 /MIC 90 of cefepime, cefotaxime, ceftazidime, ciprofloxacin, gentamicin, amikacin, ampicillin/clavulanic acid, and imipenem were 32/256, 128/512, 4/64, 32/512, 16/256, 4/16, 16/32, and 0.125/0.25 μg/ml, respectively. Conjugation experiments and plasmid detection
The PMQR genes could be transferred by conjugation in 15 of the 20 qnr and aac(6’)-Ib-cr-positive donors. Two isolates harboring more than one PMQR gene (x25 harbored qnrA1 and aac(6’)-Ib; x37 harbored qnrB4 and aac(6’)-Ib-cr) transferred only a single gene to the recipient (x25 transferred qnrA1 and x37 transferred aac(6’)Ib-cr). The MIC of ciprofloxacin for 15 transconjugants were 0.032–2 μg/ml, representing a 2- to 125-fold increase compared with the recipient (MIC 0.016 μg/ml) (Table 1). Plasmid DNA was extracted from the 20 donors and the 15 transconjugants. Each of the positive donors contained one large plasmid (>152 kb) and one to four small plasmids (