|Home | About | Journals | Submit | Contact Us | Français|
In this work, we have characterized two small ColE-like plasmids (pECY6-7, 2.7 kb in size, and pECC14-9, of 3.0 kb), encoding the QnrB19 quinolone resistance determinant, that were carried by several clonally unrelated quinolone-resistant commensal Escherichia coli strains isolated from healthy children living in different urban areas of Peru and Bolivia. The two plasmids are closely related to each other and carry the qnrB19 gene as the sole resistance determinant, located in a conserved genetic context between the plasmid RNAII sequence (which controls plasmid replication) and the plasmid Xer site (involved in plasmid dimer resolution). ISEcp1-like or other putative insertion sequences are not present in the qnrB19-flanking regions or elsewhere on the plasmids. Since we previously observed a high prevalence (54%) of qnrB genes in the metagenomes of commensal enterobacteria from the same population of healthy children, the presence of pECY6-7- and pECC14-9-like plasmids in those qnrB-positive metagenomes was investigated by PCR mapping. Both plasmids were found to be highly prevalent (67% and 16%, respectively) in the qnrB-positive metagenomes, suggesting that dissemination of these small plasmids played a major role in the widespread dissemination of qnrB genes observed in commensal enterobacteria from healthy children living in those areas.
Qnr proteins are small pentapeptide repeat proteins that bind and protect type II DNA topoisomerases from inhibition by fluoroquinolones (29-31). They represent the first discovered transferable mechanism of resistance to quinolones, and their dissemination has been associated with the increase of fluoroquinolone resistance rates in clinical isolates of the Enterobacteriaceae (17, 24). qnr-carrying isolates have been reported worldwide (17, 24), and five different lineages of Qnr proteins have been described so far: QnrA, QnrB, QnrS, and more recently QnrC and QnrD (6, 15, 17, 32).
In previous studies, we have observed a remarkable rate of quinolone resistance in commensal Escherichia coli from healthy children living in urban areas of Peru and Bolivia (1) and a high prevalence of qnrB genes (mostly qnrB19) in commensal enterobacteria from the same population of healthy children (19).
In this work, we have characterized two small ColE-like plasmids encoding QnrB19, carried by several clonally unrelated quinolone-resistant commensal E. coli strains isolated from children living in different areas, and we have demonstrated that the dissemination of those plasmids apparently played a major role in the widespread dissemination of qnrB genes observed in commensal enterobacteria from that population.
(These results were presented in part at the joint 48th Interscience Conference on Antimicrobial Agents and Chemotherapy/46th Annual Meeting of the Infectious Diseases Society of America, Washington, DC, 2008, and at the 19th European Congress of Clinical Microbiology and Infectious Diseases, Helsinki, Finland, 2009.)
The 107 E. coli isolates investigated in this study for the presence of qnr genes represented a random selection of the 1,053 ciprofloxacin-resistant commensal E. coli isolates collected during a survey on commensal E. coli carried out in 2005 on 3,193 healthy children living in four urban areas of Latin America (Camiri, Santa Cruz Department, and Villa Montes, Tarija Department, Bolivia; and Yurimaguas, Loreto Department, and Moyobamba, San Martin Department, Peru) (1).
Phylogenetic grouping of E. coli isolates was determined by the multiplex PCR-based method, which allows identification of the four major phylogenetic groups (A, B1, B2, and D) (7). Randomly amplified polymorphic DNA (RAPD) genotyping was performed using, separately, the decamer primers 1290 and 1254, as previously described (18). RAPD patterns were considered to be different when the profiles differed by at least one band.
DNA extraction from E. coli was performed as described by Sambrook and Russell (27). The presence of qnr genes (qnrA, qnrB, and qnrS) was investigated by dot blot hybridization using the DIG system according to the manufacturer's instructions (Roche Diagnostics SpA, Milan, Italy). Specific probes used in hybridization experiments were generated by PCR as described previously (qnrA and qnrS  and qnrB ). Sequence analysis of qnrB genes was determined on both strands of PCR amplification products at an external facility (Macrogen, Seoul, Korea), as described previously (19).
Antimicrobial susceptibility was determined by disk diffusion testing according to the guidelines of the Clinical and Laboratory Standards Institute (CLSI) (8, 10). MICs were determined by broth microdilution testing according to CLSI guidelines (9-10). Antibiotics were from Sigma-Aldrich (St Louis, MO). E. coli ATCC 25922 was always used for quality control purposes.
Plasmid extraction from E. coli was carried out by the alkaline lysis method as described by Sambrook and Russell (27). Southern blot analysis was carried out on nylon membranes as described for the dot blot hybridization. qnrB-harboring plasmids were transferred into E. coli HB101 (F− hsdS20 recA13 ara-14 proA2 lacY1 galK2 rpsL20 [Strr] xyl-5 mtl-1 supE44 leuB6 thi-1) by electroporation, using a Gene Pulser apparatus (Bio-Rad Laboratories, Richmond, CA) and approximately 500 ng of the plasmid DNA preparation, under the conditions recommended by the manufacturer. Mueller-Hinton agar plates containing nalidixic acid (8 μg/ml) (NAL-MH) were used for selection of transformants. The presence of qnrB genes in the transformants was confirmed by PCR. To exclude the occurrence of chromosomal mutations in transformants following selection on NAL-MH, sequence analysis of gyrA and parC was carried out with selected transformants, as described previously (26). Replicon typing was carried out by the PCR-based method described by Carattoli et al. (3). Plasmid restriction profiles were analyzed by agarose gel electrophoresis after digestion with SacII, HaeIII, EcoRI, and HindIII (Promega, Madison, WI). The nucleotide sequence of pECCY6-7 and pECC14-9 was determined on both strands by using a primer-walking technique with purified plasmid preparations at an external facility (Macrogen). Analysis and comparisons of nucleotide sequence were carried out using programs available at the NCBI web interface (http://www.ncbi.nlm.nih.gov).
The 167 metagenomes of commensal enterobacteria analyzed for the presence of pECY6-7- and pECC14-9-like plasmids represented all the qnrB-positive metagenomes detected in a previous study, where the presence of qnr genes was investigated in a sample of 310 metagenomic DNAs of commensal enterobacteria from the same population of healthy children from Bolivia and Peru (19). The presence of pECY6-7- and pECC14-9-like plasmids in the metagenomic DNAs was screened for by a PCR mapping approach. PCRs were always carried out in a 50-μl volume, using 30 pmol of each primer, 200 μM deoxynucleoside triphosphates, 1.5 mM MgCl2, and 1.5 U of the enzyme GoTaq (Promega) in the reaction buffer provided by the enzyme manufacturer. Reaction parameters were as follows, unless otherwise specified: initial denaturation at 94°C for 5 min; denaturation at 94°C for 45 s, annealing at 55°C for 45 s, and elongation at 72°C for 1.30 min, repeated for 35 cycles; and a final extension at 72°C for 10 min. In the first step of the PCR mapping approach, carried out with all metagenomes, primers designed on qnrB19 (qnrB19_RV, 5′-CGGCACCTGAAAAATCGCAG) and on the replication origin (ColEPB_FW, 5′-CTGACACTCAGTTCCGCGA) were used to target amplification of the conserved region containing qnrB19 and the ColE-like backbone (qnrB19-ColEPB PCR). In the second step, carried out with metagenomes positive for the qnrB19-ColEPB PCR, primers designed in divergent orientation on qnrB19 (qnrB19_FW, 5′-TGGATGGGGACTCAGGTACT, and qnrB19_RV, 5′-CGGCACCTGAAAAATCGCAG) were used to target amplification of the whole plasmids (qnrB19circ PCR) (PCR elongation time of 4 min). The identity of PCR amplicons was confirmed by Southern blotting using a probe targeting the replication origin of pECY6-7 and pECC14-9 (ColEPB probe, generated with primers ColEPB_FW [5′-CTGACACTCAGTTCCGCGA] and ColEPB_RV [TGCTGCCAGTGGCGATAAGT]) and restriction profiling with HaeIII. Selected amplicons (representative of amplicons of different sizes and origins) were analyzed by sequencing.
The complete circular nucleotide sequences of pECY6-7 and pECC14-9 have been deposited in the GenBank sequence library and assigned the accession numbers GQ374156 and GQ374157, respectively.
The presence of qnr genes was investigated by PCR in 107 ciprofloxacin-resistant E. coli isolates, representing a random sample of the 1,053 ciprofloxacin-resistant commensal E. coli isolates collected in a previous survey carried out with 3,193 healthy children living in four urban areas of Peru and Bolivia (1). A total of six out of the 107 isolates (6%) were found to be positive for qnrB genes, while qnrA and qnrS genes were not detected. Amplicon sequencing identified the qnrB19 allele (5, 23) in all of them. The qnrB19-positive isolates were found in each urban area of Peru and Bolivia (Table (Table11).
Multiplex PCR showed that all the qnrB19-positive isolates belonged to phylogenetic group A (Table (Table1),1), which is usually associated with commensalism and minor virulence (7, 20), supporting the notion that commensal E. coli can act as a reservoir of qnr genes. However, RAPD genotyping showed different profiles (Table (Table1),1), indicating a likely plasmid-mediated nonclonal dissemination (Table (Table1).1). All the qnrB19-positive isolates exhibited a multidrug resistance (MDR) phenotype which variably included tetracycline, ampicillin, kanamycin, streptomycin, trimethoprim, and sulfonamides (Table (Table11).
Southern blot experiments, carried out with plasmid preparations from the six qnrB-positive isolates, showed that the qnrB19 genes were located on low-molecular-weight plasmids, with the exception of one isolate from Peru (E. coli M4-6) (data not shown). The small qnrB19-harboring plasmids were transferred by electroporation into E. coli HB101 and further characterized. Restriction analysis with several enzymes (SacII, HaeIII, EcoRI, and HindIII) revealed two different plasmid profiles for the qnrB19-harboring plasmids from Peru (estimated size, 2.7 kb) and Bolivia (estimated size, 3.0 kb), respectively (Table (Table1).1). Although quinolone MICs were increased, transformants remained susceptible to fluoroquinolones, indicating that the ciprofloxacin resistance phenotype of the donors was contributed by other resistance mechanisms, which were not further investigated in this study (e.g., mutations in the genes encoding DNA topoisomerases II, upregulation of native efflux pumps, and/or decreased expression of outer membrane porins). No other resistance trait present in the original isolate was cotransferred. The small qnrB19-harboring plasmids could not be assigned to a replicon type using the multiplex PCR-based method (3).
A representative of each small qnrB19-harboring plasmid, pECY6-7 (from E. coli Y6-7) and pECC14-9 (from E. coli C14-9) (Table (Table1),1), was subjected to complete DNA sequencing.
pECY6-7 and pECC14-9 are small ColE-like plasmids of 2,699 and 3,071 bp, respectively. They share a 2,132-bp common region (99% identical) and differ by a unique region of 567 bp and 939 bp in pECY6-7 and pECC14-9, respectively. The common region contains a ColE-like backbone and the qnr region (Fig. (Fig.1).1). The ColE-like backbone includes putative regions for plasmid replication (oriV, RNAII, and RNAI ) and mobilization (oriT ) and showed similarity with the corresponding region of the ColE-like plasmids p15A (89%) (28) and p9123 (86%) (a sul2-harboring plasmid which has been shown to enhance host fitness) (12). The nature of the plasmid backbone explained why it could not be identified by PCR-based replicon typing. The qnr region is located between the putative −35 promoter of RNAII (11) and the plasmid Xer site, which is known to be involved in plasmid dimer resolution (2). The unique regions, located between the Xer site and oriT, are not related to each other: the 939-bp region in pECC14-9 showed 93% identity with the corresponding region in p9123 (12), while an internal part of the 567-bp region in pECY6-7 (378 bp) exhibit 85% similarity with plasmid pMGD2 from an environmental isolate of Klebsiella spp. (GenBank accession no. AY033498). Homology between pECY6-7 and pECC14-9 abruptly ends in correspondence of the Xer site, suggesting that the Xer site-specific recombination might have played a role in the evolution of these plasmids.
The regions flanking qnrB19 show 100% identity with those present in the qnrB19-harboring transposons Tn5387 from a Klebsiella pneumoniae clinical isolate from the United States (245 bp upstream and 225 bp downstream) (23) and Tn2012 from an E. coli clinical isolate from Colombia (155 bp upstream and 225 bp downstream) (5) (Fig. (Fig.2).2). Upstream of qnrB19, the homology with Tn5387 ends in correspondence of a putative imperfect IRR2 for ISEcp1 located within the pspF gene, which has been found in the genetic environment of various qnrB genes (22), suggesting that ISEcp1-like elements could have been involved in the mobilization of qnrB19 to these plasmids. This hypothesis is also supported by the presence, upstream of the putative IRR2, of a 5-bp A+T-rich motif (TTATA) which was shown to represent a potential target site for ISEcp1 (21). However, no ISEcp1-like sequence or other insertion sequences are present downstream of qnrB19 or elsewhere on the plasmids. To explain this finding, one could hypothesize that the ISEcp1-like element involved in qnrB19 mobilization could have undergone a subsequent excision. The presence of a putative IRR2 in the sequence of Tn5387 and Tn2012 immediately downstream of the end of homology with the qnr region in pECY6-7 and pECC14-9 would support this hypothesis (Fig. (Fig.22).
ColE-like plasmids have been associated with the dissemination of qnrS genes in enterobacteria (13, 16), but to the best of our knowledge, this is the first study reporting a qnrB variant in such small ColE-like plasmids.
Finding of the closely related pECY6-7 and pECC14-9 plasmids in clonally unrelated strains of commensal E. coli from children living in different urban areas of Peru and Bolivia suggested that these plasmids could have a broad diffusion and contribute to the high prevalence of qnrB genes that was observed in commensal enterobacteria from the same population of healthy children (19). To investigate this possibility, we screened for the presence of pECY6-7- and pECC14-9-like plasmids in the 167 metagenomes of commensal enterobacteria from children that had previously tested positive for qnrB genes (1). Screening was performed by a two-step PCR mapping approach. In the first step, carried out with all metagenomes, amplification of the conserved ColE-like backbone and qnr region was targeted by the primers ColEPB_FW and qnrB19_RV (qnrB19-ColEPB PCR); in the second step, carried out with the metagenomes that tested positive, amplification of the entire plasmid was targeted by the primers qnrB19_FW and qnrB19_RV (qnrB19circ PCR) (Fig. (Fig.1).1). The identities of amplicons were confirmed by Southern blot hybridization, restriction mapping with HaeIII, and sequencing (with selected samples).
Of the 167 qnrB-positive metagenomes analyzed, 128 (77%) were positive with the qnrB19-ColEPB PCR, revealing the presence of pECY6-7- and pECC14-9-like plasmids. Of these 128 metagenomes, 124 (97%) were positive with the qnrB19circ PCR, revealing the presence of pECY6-7 in 97 cases, of pECC14-9 in 12 cases, and of both plasmids in 15 cases (Table (Table2).2). Sequence analysis of 4 selected amplicons (2 amplicons of different sizes from each country) confirmed 100% identity with pECY6-7 and pECC14-9, respectively.
The qnrB19 genes previously detected in commensal enterobacteria from the population of healthy children were also found in species other than E. coli (19). In this work, we investigated the presence of pECY6-7- and pECC14-9-like plasmids in all the qnrB19-carrying non-E. coli isolates from that study (two K. pneumoniae isolates, one Klebsiella oxytoca isolate, and one Escherichia hermannii isolate) (19) by PCR mapping and sequencing. Plasmids identical to pECY6-7 were found in a K. pneumoniae isolate from Peru (K. pneumoniae Y1) and in an E. hermannii isolate from Bolivia (E. hermannii C1). These results confirmed that pECY6-7 is also capable of spreading among enterobacterial species other than E. coli.
Dissemination of plasmid-mediated quinolone resistance determinants, such as the qnr genes, has been associated with the worldwide increase in fluoroquinolone resistance rates in clinical isolates of Enterobacteriaceae (17, 24). Recently we observed that qnrB genes (mainly qnrB19) are highly prevalent in the commensal enterobacteria of healthy children living in urban areas of Peru and Bolivia, suggesting that the intestinal microbiota could represent an important reservoir of similar resistance determinants (19). In this work, we demonstrated that two small ColE-like plasmids, closely related to each other and carrying the qnrB19 gene as the sole resistance determinant, are highly prevalent among commensal enterobacteria in children living in those settings and apparently played a major role in the widespread dissemination of qnrB genes observed in that area.
The plasmid pECY6-7 was the most prevalent, and it was found to be disseminated in different species of enterobacteria (E. coli, K. pneumoniae, and E. hermannii). While the manuscript was in revision, a plasmid identical to pECY6-7 was identified in a Salmonella enterica serovar Typhimurium strain of human origin isolated in The Netherlands (GenBank accession no. FN428572) (14), suggesting that the dissemination of this plasmid could be even more widespread.
The reasons accounting for the high prevalences of pECY6-7 and pECC14-9 in the commensal enterobacteria of the study population remain unclear. Data collected about household use of antibiotics excluded previous use of fluoroquinolones in the children investigated (1). Moreover, the absence of other resistance genes in pECY6-7 and pECC14-9 excluded the possibility that selection of these plasmids could be related to exposure to other antibiotics. Further studies of the mobilization and fitness impact of the plasmids pECY6-7 and pECC14-9 could provide important information for understanding their remarkable propensity for such widespread dissemination.
This study was carried out as a follow-up of research activities of the ANTRES project (a project on antibiotic use and resistance in Latin America, supported by the European Commission within the INCO-DEV program of the FP5).
Published ahead of print on 14 December 2009.