Co-transfer of plasmid-mediated blaAmpC and uoroquinolone resistance genes in Klebsiella pneumoniae isolates causing nosocomial urinary tract infection

Background: Klebsiella pneumoniae is a pathogen that frequently causes nosocomial urinary tract infection (UTI), and the prevalence of plasmid-mediated resistance determinants among clinical isolates of K. pneumoniae leads to the appearance of resistance to antibiotics. The aim of this study was to investigate the prevalence of plasmid-mediated quinolone resistance (PMQR) genes in acquired AmpC (ac-AmpC) β ‐ lactamase ‐ producing K. pneumoniae isolates from patients with nosocomial UTI and to characterize the transmissibility of plasmids co-harbouring bla AmpC and PMQR genes. Methods: From January 2017 to June 2018, we collected 46 AmpC-producing K. pneumoniae isolates causing nosocomial UTI from a tertiary care hospital in China. β-lactamase, PMQR and virulence genes were detected by PCR and sequencing. Clonal relatedness was assessed using ERIC-PCR and multilocus sequence typing (MLST). Plasmids carrying multiple bla AmpC and PMQR genes were characterized by PCR-based replicon typing (PBRT) and S1-PFGE. Conjugation and electroporation experiments were carried out to assess resistance transfer mediated by plasmids. Overlapping PCR was used to map the genetic context of the bla AmpC genes. Results: In the studied isolates, non-susceptibility of third-generation cephalosporin and uoroquinolone was very high (>80%). bla CMY-2 , bla DHA-1 , and quinolone resistance gene (qnr) were detected in 11, 41 and 33 isolates, respectively. Among the isolates, 6 strains co-harboured multiple AmpC and qnrB genes. The bla AmpC and qnrB genes from these six isolates were co-transferrable to recipients via conjugation or electroporation, with IncFIA, IncFIB and IncA/C being the dominant replicons (sizes from ~78 to 217 kb). Forty-six isolates were categorized into 25 ERIC types, and the 6 isolates harbouring multiple bla AmpC and

consecutively from patients with nosocomial UTI (one per patient) and were selected according to a resistance phenotype compatible with AmpC production: resistance or reduced susceptibility, as determined according to the Clinical and Laboratory Standards Institute (CLSI) breakpoints, to amoxicillin/clavulanic acid (< 18 mm, > 8/4 µg/ml) and cefotaxime (< 26 mm, > 1 µg/ml), ceftazidime (< 21 mm, > 4 µg/ml) or aztreonam (< 21 mm, > 4 µg/ml). The presence of the ampC gene was con rmed by multiplex PCR as described by Pérez-Pérezetal [21]. The study protocol was approved by the Ethics Committee of Dalian Medical University. Detection of β-lactamase and PMQR genes Total DNA preparations were obtained by thermolysis of isolates as described previously [21]. Multiplex PCR assays were performed on all isolates for the detection of ampC family (bla MOX , bla CMY , bla FOX , bla DHA , bla ACC , bla ACT , and bla MIR ) and PMQR genes (qnrA, qnrB, qnrS, aac(6')-Ib and qepA). The primers used for multiplex PCR ampli cation are listed in Table 1. aac(6')-Ib-cr was differentiated from its wildtype allele by digestion with the enzyme BtsCI (New England Biolabs, Massachusetts, USA). Other βlactamase genes (bla SHV , bla TEM bla CTX−M and bla OXA ) were analysed by PCR using speci c primers and conditions we described previously [22]. were analysed using BioNumerics 7.6 software (Applied Maths, Kortrijk, Belgium). PCR ngerprint pro les were analysed using the Dice (similarity) coe cient. Cluster analysis was performed as previously described based on the unweighted pair group method with arithmetic averages (UPGMA) with a position tolerance of 0.05 [23].
Resistance transfer ability of multiple AmpC-producing K. pneumoniae isolates Conjugation experiments were performed using a broth mating protocol. Multiple AmpC-producing K.
When a multiple ampC-harbouring plasmid could not be transferred by conjugation, we tried electroporation to harvest transformants. E. coli HB101 transformants were obtained by electroporation of plasmid DNA extracted with the MiniBEST Plasmid Puri cation Kit (Takara, Dalian, China) in an electroporator (Eppendorf, Eppendorf AG, Germany) and selected on LB agar supplemented with cefotaxime (4 µg/ml). PCR ampli cation, antimicrobial susceptibility testing and plasmid replicon typing were performed for all transconjugants/transformants to identify resistance determinants, antibiotic phenotypes and incompatibility groups, respectively. Molecular Characterization Of Plasmids Carrying Multiple Ampc Genes S1-nuclease (Takara, Dalian, China) digestion as well as pulsed-eld gel electrophoresis (S1-PFGE) analysis was performed for donor strains and the corresponding transconjugants or transformants. For plasmid size estimation, comparison with the molecular weight marker Salmonella braenderup H9812 was performed. Plasmid replicons were determined using the PCR-based replicon typing (PBRT) scheme with 18 pairs in PCR for detecting F, FIA, FIB, FIC, HI1, HI2, I1-Iγ, L/M, N, P, W, T, A/C, K, B/O, X, Y and FII replicons, as described by Carattoli et al. [24].
Genetic context analysis of bla ac−AmpC genes in multiple AmpC-producing K. pneumoniae isolates The anking regions of ac-AmpC genes were identi ed by overlapping PCR as previously described [26]. For bla CMY−2 , the presence of a composite genetic structure (bla CMY−2 -blc-sugE) originating from the Citrobacter freundii chromosome and ISEcp1 gene (responsible for the transfer of the bla CMY−2−like -blc-sugE region) was explored. For the genetic organization of bla DHA−1 , the searched genes were ISCR1, IS26, orf2 (conserved region of unknown function in Morganella species), ampR, qacEΔ1 and sul1. PCR products were sequenced and compared with those available in GenBank (www.ncbi.nih.gov/BLAST).

Statistical analysis
Fisher's exact test was used to compare the distribution of PMQR genes between simplex and multiplex ac-AmpC-producing K. pneumoniae isolates. P-values < 0.05 were considered statistically signi cant. SPSS 13.0 Statistics software (Stats Data Mining Co., Ltd., Beijing, China) was used for analyses.

Resistance Transfer And Molecular Characterization Of Plasmids
Plasmids carrying ac-AmpC and qnrB genes were successfully transferred via conjugation (n = 2) and/or electroporation (n = 4) from the 6 multiple AmpC-producing K. pneumoniae isolates. The resistance pro les of the transconjugants/transformants were similar to those of the K. pneumoniae donor strains, demonstrating the transfer of antimicrobial resistance, including resistance to β-lactam-based antimicrobial compounds and uoroquinolones. S1-PFGE plasmid pro les and PCR replicon typing identi ed 3 replicons, IncFIA, IncFIB and IncA/C (ranging from ~ 78 to 217 kb), which were found in both donors and transconjugants/transformants and were associated with the transfer of bla DHA−1 , bla CMY−2 , qnrB and several other β-lactamase genes (Table 4 and Fig. 1). Virulence Factors Six K. pneumonia strains carrying multiple ac-AmpC and qnrB genes showed important virulence pro les, including the entB (n = 4), mrkD (n = 6) and ybtS (n = 6) genes, and had the capsular serotype K5 (n = 4). Two-three virulence factors were found per isolate and the results are presented in Table 5.

Genetic Context Of Ac-ampc-encoding Genes
Analysis of the genetic context of the 6 strains that harboured multiple ac-AmpC genes revealed that the bla CMY−2 gene was associated with ISEcp1, which is responsible for the transfer of the bla CMY−2−like -blc-sugE region from the chromosome of Citrobacter freundii to plasmids. In our study, six strains contained ISEcp1, the outer membrane lipoprotein gene blc and the drug e ux channel sugE upstream and downstream of the bla CMY−2 gene. However, truncation at the 5' end of ISEcp1 was observed in the isolate Kp7 (Fig. 2a).
We found a structure composed of the bla DHA−1 , ampR, qacEΔ1 and sul1 genes in 6 bla DHA−1 -carrying isolates. All isolates had the ISCR1 element upstream of bla DHA−1 , but no IS26 element was found upstream or downstream (Fig. 2b).

Clonal Relationships
The clonal relatedness of all isolates was determined by ERIC-PCR. By analysing the ERIC-PCR pro les (Fig. 3), the 46 isolates were categorized into two main phylogenetic groups including 25 ERIC types. In the major phylogenetic subgroup, 95.65% (44/46) of the isolates were clustered. The 25 distinct types were then labelled A01 to A25. We found that 6 isolates harboured multiple ac-AmpC β-lactamase and qnrB genes belonging to ST1 or STnew1 (Table 5).

Discussion
The spread of antimicrobial resistance is primarily caused by the dissemination of large plasmids carrying multiple antibiotic resistance genes [27]. In the last decade, the prevalence of plasmid-mediated AmpC-producing K. pneumoniae isolates has increased in nosocomial acquired and healthcareassociated infections [28,29]. Plasmids carrying ampC genes are often co-harboured with different antibiotic resistance markers, such as ESBL and PMQR genes [30]. Co-transfer of these resistance genes may contribute to the emergence of multidrug-resistant strains, increasing the risk for antibiotic treatment failure. Our study focused on the dissemination of K. pneumoniae strains carrying multiple AmpC and PMQR genes in patients with nosocomial UTI. To the best of our knowledge, this is the rst report on the co-occurrence of various plasmid-borne bla AmpC and PMQR genes in Enterobacteriaceae causing nosocomial infections.
In the studied isolates, non-susceptibility of both third-generation cephalosporin and uoroquinolone was very high (CTX, 100%; CAZ, 89.13%; CRO, 97.83%; LEV, 82.61%; and CIP, 89.13%), as non-susceptibility to at least one third-generation cephalosporin was one of the major criteria adopted in this study. However, with the increasing use of uoroquinolone-class drugs in community-and hospital-acquired-UTI patients, the rates of resistance to uoroquinolones have increased [31]. In this study, the resistance rates of two uoroquinolones (LEV and CIP) in K. pneumoniae isolates were prominently higher than those reported in a prior Asia-Paci c survey (51-55%) of pathogens causing UTI [31]. And recent report mentioned above suggested that uoroquinolones are not recommended as a UTI therapy empirically unless in vitro data show susceptible results [31].
The most commonly occurring gene encoding an ac-AmpC enzyme in this study was DHA-1, which was found in 89.13% of the K. pneumoniae isolates and is the most prevalent plasmid-mediated AmpC in China and other areas of Asia; this study also found a low prevalence of bla CMY−2 (23.91%), which has been mostly reported in Europe. Notably, the simultaneous production of multiple AmpC β-lactamases in a single isolate (as observed in 6 of the 46 isolates), complicates the detection of both enzymes and treatment of infections caused by these isolates. The multiplex AmpC-producing isolates exhibited a higher percentage (100%) of PMQR prevalence than the simplex AmpC-producing isolates (82.5%), but the difference was not statistically signi cant. Multiplex PCR also revealed that qnrB had the highest prevalence (84.78%), followed by qnrA (4.35%) and qnrS (4.35%). Earlier studies also indicate that qnrB is the most prevalent PMQR gene in China and that its distribution is closely related to the presence of acquired AmpC β-lactamases [32,33].
Co-transfer of PMQR genes along with bla AmpC in large individual plasmids co-harbouring many other resistance genes has been shown in other studies [18]. The successful transfer of bla AmpC along with qnrB genes was studied in the 6 isolates carrying multiple ac-AmpC β-lactamase genes. However, this study showed that not all multiple AmpC genes were co-transferred with the qnrB gene, which indicates that these genes are likely located on different plasmids. Previous studies on plasmid replicon types show that the IncI1, IncA/C, IncFII and IncX1 plasmids carry bla AmpC [34,35]. PMQR genes are associated with the IncN, IncL/M, IncFII, IncHI1, IncI1, IncR, and colE types [36]. In the present study, we found that in K. pneumoniae, plasmids carrying both bla AmpC and PMQR genes were of the replicon type IncA/C, followed by IncFIB and IncFIA. IncF-group plasmids are highly conjugative and are widely distributed in Enterobacteriaceae, and the presence of any gene in this group of plasmids will escalate the spread of the gene to other organisms [36]. Furthermore, we found that bla AmpC genes were generally located on plasmids of various sizes ranging from ~ 78 to 217 kb. Our results are in agreement with reports of ac-AmpC genes found to be encoded by plasmids of sizes varying from 7 to 180 kb [16].
Subsequently, we analysed the regions surrounding the ac-ampC genes. The genetic organization of bla CMY−2 was highly conserved. All the isolates carried the mobile element ISEcp1 (ISEcp1-bla CMY−2 -blc-sugE), as documented in previous reports [26,37]. A well-conserved structure was also found in all isolates co-harbouring the bla DHA−1 and bla CMY−2 genes. However, although IS26 or ISCR1 elements are commonly related to the transmission of DHA-1 enzymes [38], none of our isolates harboured the IS26 insertion sequence.
Based on the distinct patterns observed with ERIC-PCR typing, the dissemination of AmpC-producing K. pneumoniae strains is multifactorial and cannot be fully attributed to a predominant clone, as was reported by Thouraya et al [37]. In addition, we found that the six isolates co-harbouring multiple acquired bla AmpC genes belonged to the ST1/STnew1 clone. This ST has been described recently in a clinical K.
pneumoniae isolate of swine origin [39]. Moreover, we found that clonal K. pneumoniae strains had the capsular serotype K5 and showed important virulence pro les, including the three genes entB, mrkD and ybtS. mrkD and ybtS are widely distributed in K. pneumoniae and encode type 3 mbriae and phenolatetype siderophores.

Conclusions
The present study indicates that third-generation cephalosporin and uoroquinolone resistance is high in clinical isolates causing nosocomial UTI. PMQR genes were highly prevalent in AmpC-producing K. pneumoniae isolates, and qnrB was predominant. AmpC genes were co-transferred with qnrB and several other β-lactamase genes, conferring resistance to most antibiotics tested. All transconjugants/transformants were associated with IncFIA, IncFIB and IncA/C plasmids and a conservative genetic environment. This study has several limitations. First, this study was a retrospective analysis, and only limited UTI patient information was available. Another limitation of the study was the small number of multiple AmpC-producing K. pneumoniae isolates obtained from a single medical centre. However, the study focused on the transfer of plasmid-mediated resistance genes among clinical AmpCproducing K. pneumoniae strains to better guide the prevention and clinical treatment of nosocomial infections. Further studies are needed to address these limitations. Abbreviations UTI, urinary tract infection; PMQR, Plasmid-mediated quinolone resistance; ac-AmpC, acquired AmpC; ERIC-PCR, enterobacterial repetitive intergenic consensus-PCR; PBRT, PCR-based replicon typing; PFGE, Pulsed-eld gel electrophoresis; ST, sequence type; AMP, ampicillin; KZ, cefazolin; CXM, cefuroxime; CTX, cefotaxime; CAZ, ceftazidime; CRO, ceftriaxone; FEP, cefepime; FOX, cefoxitin; ATM, aztreonam; AMC, amoxicillin/clavulanic acid; IPM, imipenem; LEV, levo oxacin; CIP, cipro oxacin; AMK, amikacin.

Declarations
Ethics approval and consent to participate The study protocol was carefully reviewed and approved by the institutional Ethics Committee of the Dalian Medical University.

Consent for publication
Not Applicable.

Competing interests
The authors declare that they have no competing interest.