Multiple Copies of qnrA1 on an IncA/C₂ Plasmid Explain Enhanced Quinolone Resistance in an Escherichia coli Mutant

TEXT

The plasmid-mediated quinolone resistance gene qnrA1 causes a modest decrease in quinolone susceptibility but facilitates the selection of mutants with higher levels of quinolone resistance through additional chromosomally encoded mechanisms (1, 2). Since the report of the first plasmid-mediated quinolone resistance (PMQR) determinant, qnrA1, in 1998, PMQR genes have been identified worldwide (3, 4).

In our previous study, in which mutants with decreased ciprofloxacin susceptibility were selected from J53/pMG252 with qnrA1 (1), we found that qnrA expression was 2.2-fold higher in mutant J53 Cip^r (ciprofloxaxin resistance) 8-2 carrying what we term plasmid pMG252A than in the parental strain, with no changes detected in the qnrA promoter region. This mutant also had a 3.1-fold increase in qnrA copy number without a change in copy number for bla_FOX-5, another marker on the plasmid, indicating that the qnrA copy number increase was gene specific and potentially responsible for increasing the ciprofloxacin MIC from 0.25 μg/ml in the parental strain to 2 μg/ml in the mutant. Furthermore, no mutations enhancing ciprofloxacin resistance were found in chromosomal genes of mutant J53 Cip^r 8-2, and the enhanced resistance could be transferred by conjugation and eliminated along with pMG252A, indicating that plasmid pMG252A was directly involved.

Therefore, to confirm an increase in the number of copies of qnrA1 and the nature of the duplication, we sequenced both pMG252 and pMG252A from the J53 Cip^r 8-2 mutant.

In a first attempt, we extracted both plasmids using the Qiagen large construct kit and obtained some short-read sequences with the Illumina platform, performed by the Center for Computational and Integrative Biology (CCIB) DNA Core Facility at Massachusetts General Hospital (Cambridge, MA). Libraries were pooled in equimolar concentrations for multiplexed sequencing on the Illumina MiSeq platform with a 2 × 150-bp read run parameter. Purified plasmids were sequenced, producing 154 × 10⁵ reads and 1.28 × 10⁵ reads for plasmids pMG252 and pMG252A, respectively. Ninety-five percent of the reads had a mean sequence quality (Phred score) of >30. The total depth of coverage after deduplication for each plasmid was 108× for pMG252 and 98× for pMG252A. Because of the large size of pMG252 (∼180 kb) (2) and the challenges imposed by insertion elements, repetitive sequences, and duplications, de novo assembly from these data was unable to produce a circularized plasmid sequence or an unambiguous location and copy number for qnrA1. To overcome these challenges, we used the MinION platform (Oxford Nanopore Technologies) to produce long sequence reads via Day Zero Diagnostics (Boston, MA). Total DNA was extracted using the Genomic DNA kit (Genomic-tip 100/G; Qiagen), prepped with SQK-RBK004, run using an R9.4.1 flow cell on a MinION Mk1B sequencer controlled by MinKNOW 2.0, and base called using Albacore v2.3.1. A larger number of reads and higher read lengths were achieved, and circularized plasmids were obtained. The final assembly was polished using Illumina reads (Fig. 1).

• Open in new tab
• Download powerpoint

FIG 1

Genetic map of parental pMG252 plasmid. The figure was constructed with Geneious Prime 2019.0.4.

The circularized sequence of plasmid pMG252 was 185,643 bp long, while that of pMG252A was 201,521 bp. The two plasmids had the common backbone of IncA/C plasmids (5, 6). They were identified as IncA/C₂ type 2 (5). Additionally, plasmid MLST (PMLST) typing assigned the pMG252 plasmid to sequence type 3 (ST3), which is the most common group of IncA/C plasmids (6). By the core gene PMLST (cgPMLST) scheme (7), pMG252 belongs to core-genome ST 3.3 (cgST3.3).

pMG252, which is the first plasmid identified with a qnr gene, was found in a Klebsiella pneumoniae clinical isolate (8). It had an average G+C content of 52.8%. It contained a large antibiotic resistance island (41,056 bp) located between the two ends of the rhs2 gene. This module included a class 1 integron (termed region S1) with the following gene cassette array: bla_CARB-2-aadA2-cmlA1-catB11-qacEΔ1-sul1-ISCR1-qnrA1-qacEΔ1-sul1-ISCR1-dfrA19, followed by another class 1 integron harboring an aadB-catB3 array. Downstream from the integrons were msr(E), encoding a macrolide efflux pump, and mph(E), encoding a macrolide 2′-phosphotransferase, both associated with insertion sequence IS26. Different insertion sequences separated the msr(E)-mph(E) module from a mercury resistance operon that was bracketed by the end of the rhs2 gene (Fig. 1).

Other plasmid-encoded functions that have been previously annotated in sequenced IncA/C₂ plasmids (5) have been also identified in pMG252 (Fig. 1). These functions included genes for methyltransferases (dcm1, dcm2, and dcm3), DNA metabolism (nuc, topB, kfrA, uvrD, ter, int, and pri), protein export and folding (sppA, dsbA, dsbC, and yacC), and plasmid stability (parA-parB and stbA) (Fig. 1).

In a comparison of the assemblies of pMG252 and pMG252A, they were found to be 88.9% identical, which was ascertained using the multiple-sequence alignment program MAFFT (Geneious). The major difference between them was in region S1 that indeed contained multiple copies of qnrA1 (Fig. 2). In most qnrA1-containing plasmids, a single copy of ISCR1 is found downstream from qnrA1, but in plasmid pMG252, qnrA1 is surrounded by two copies of ISCR1. The qnrA1-ISCR1 complex is inserted into sul1-type integron-containing cassettes that confer resistance to β-lactams (bla_CARB-2), chloramphenicol (cmlA1 and catB11), streptomycin (aadA2), sulfonamide (sul1), and trimethoprim (dfrA19) (2). The assembled sequence of plasmid pMG252A revealed the presence of four additional copies of qnrA1-qacEΔ1-sul1-ISCR1, thus validating our findings of an increased qnrA1 copy number.

• Open in new tab
• Download powerpoint

FIG 2

(A and B) Genetic environment of qnrA1 in plasmids pMG252 (A) and pMG252A (B) in region S1. Purple arrows show genes conferring resistance to antimicrobials or biocides, dotted purple arrows indicate the multiple copies of qnrA1, and green arrows denote elements involved in recombination and/or transposition.

There are numerous studies describing the prevalence of PMQR harboring qnrA1 in Enterobacteriaceae (9, 10). qnrA1 has been associated in the same plasmid with other PMQR genes, such as qnrB, qnrS, and qepA (11), but it is unusual to find more than one copy of the same PMQR gene within a plasmid. An exception is plasmid pSZ50 from Mexico, in which the entire integron containing ISCR1, qnrA1, and other resistance genes was duplicated in tandem, although quinolone resistance was not increased (12). In our study, plasmid pMG252A with 5 copies of qnrA1 had an 8-fold higher MIC to ciprofloxacin than did its pMG252 parent, reaching an MIC of 2 μg/ml ciprofloxacin and thus exceeding the CLSI breakpoint for resistance (13). The elevated resistance was transmissible by pMG252A on conjugation. Unlike other qnr types, qnrA1 is not inducible by quinolones (14). The duplication of ISCR1 around qnrA1 in pMG252 and related plasmids creates the structure of a composite transposon; this facilitates enhancement of the qnrA1 copy number and, consequently, quinolone resistance under conditions of quinolone stress, and it adds to the repertoire of mechanisms that can increase quinolone resistance to clinically important levels.

The genetic organization of bla_FOX-5 in pMG252 was similar to that previously published (GenBank accession number CP007732). Upstream of the bla_FOX-5 gene, there was a region with 98% DNA homology to the ISAs2 transposase gene from Aeromonas salmonicida, a member of the IS30 family (15), followed by a bleomycin resistance gene. Immediately in the 3′ direction, there was a multidrug transporter gene (mdrL), followed by a transcriptional regulator gene (lysR) and another copy of ISAs2.

In summary, we determined by plasmid sequencing the novel presence of 5 copies of qnrA1 in plasmid pMG252A, as a result of transposition via insertion sequence ISCR1 located downstream from qnrA1 within a class 1 integron; we also characterized pMG252 as a sequence type 3.3 (ST3.3) member of the IncA/C₂ type 2 family.