Distribution of quinolone resistance gene (qnr) in ESBL-producing Escherichia coli and Klebsiella spp. in Lomé, Togo

Background Qnr genes are known to confer a low-level resistance to fluoroquinolone in Enterobacteriaceae. They are often found on the same resistance plasmids as extended spectrum β-lactamase (ESBL) and constitute the most common antibiotic resistance mechanism. This study aimed to detect the presence of qnr genes in ESBL-producing E. coli and Klebsiella spp. Methods From May 2013 to July 2015, 91 E. coli and 64 Klebsiella spp. strains with phenotypic resistance to quinolone were collected from several specimens and analyzed for the detection of qnrA, qnrB, qnrS genes and the β-lactamase resistance genes (blaCTX-M, blaTEM, blaSHV) using simplex and multiplex PCR. Results In the present study, 107 (69%; 61 E. coli and 46 Klebsiella spp.) of 155 bacterial strains tested were found harboring at least one qnr gene consisting of 74 (47.74%) qnrB, 73 (47.10%) qnrS and 4 (2.58%) qnrA. Of the 107 strains encoding qnr genes, 102, 96 and 52 carried CTX-M1, TEM and SHV type ESBL respectively. Conclusion This study identified quinolone resistance (qnr) gene in ESBL-producing E. coli and Klebsiella spp. in Togo. These finding which suggest a possible resistance to quinolone are of high interest for better management of patients and control of antimicrobial resistance in Togo. Electronic supplementary material The online version of this article (10.1186/s13756-019-0552-0) contains supplementary material, which is available to authorized users.


Background
Quinolones and β-lactams are classes of extensively used molecules worldwide in the treatment of many infectious diseases [1]. Quinolones are synthetic antibiotics used for infections involving Gram-negative bacteria such as Enterobacteriaceae. Fluoroquinolones have broad-spectrum intrinsic activity greater than quinolones [2].
Three main mechanisms of quinolone resistance have been described: i) the accumulation of mutations in the genes encoding quinolone target DNA gyrase and topoisomerase IV; ii) a decrease of intracellular concentration of fluoroquinolones by porins down-regulation or modification of the efflux pumps activity, iii) the acquisition of plasmid resistance genes [2].
The acquisition of plasmid-mediated quinolone resistance genes (PMQR) leads to the protection of quinolone's targets by qnr proteins belonging to the pentapeptide repeat (PRP) family and the hydrolysis of quinolones by the aac (6′)-Ib-cr protein [2,3]. Mechanism of plasmid-mediated quinolone resistance leads to a low level of fluoroquinolone resistance and facilitates the selection of mutant strains with a high level of fluoroquinolone resistance [3,4].
Since the discovery of plasmid quinolone resistance genes, a large number of qnr alleles have been found on plasmids or bacterial chromosome. About 100 qnr genes variant have been described mainly from Enterobacteriaceae, and grouped into 5 distinct families: qnrA, qnrB, qnrC, qnrD and qnrS [3,5].
In Togo, results from previous studies revealed the presence of beta-lactamase gene CTX-M1 (95.73%), TEM (82.31%) and SHV (45.12%) in ESBL-producing E. coli and Klebsiella spp. The production of ESBL was associated with high co-resistance to fluoroquinolone (93% for ciprofloxacin), aminoglycosides (76.36% for gentamicin) and trimethoprim/sulfamethoxazole (95.65%) [15,16]. This high level of multidrug resistance suggests acquisition of plasmid-mediated antibiotic resistance factors in these strains. PMQR determinants as qnr genes was also usually found in multidrug resistance plasmid among Enterobacteriaceae producing-ESBL especially E. coli and species of Klebsiella [3,6,17,18]. In this study we are interested in fluoroquinolone resistance. Here, we report the frequency of qnrA, qnrB and qnrS genes in ESBL-producing E. coli and Klebsiella spp.

Samples collection and identification
Well characterized Escherichia coli and Klebsiella spp. strains were collected during a prospective study from May 2013 to July 2015 in the bacteriology laboratory of the National Institute of Hygiene (INH) in Lomé, Togo. This public health institute is specialized in biomedical analysis, epidemiological surveillance, immunization, water, and food quality control. Strains were isolated from various pathological specimens including urine, vaginal swabs, pus, and sperm samples. Standard microbiological methods were used to isolate and purify bacterial strains on Mac-Conkey or Eosin Methylene Blue (EMB) media. Strains were identified using the API 20E identification system (API 20 E, Identification System for Enterobacteriaceae and others non-fastidious Gram-negative rod; BioMérieux, Marcy-Etoile, France). The API 20 E system is a standardized technique allowing only the biochemical identification of an Enterobacterial strain using an isolated colony.

Susceptibility test and ESBL phenotype detection
Antibiotic susceptibility test was performed and interpreted according to the 2014 recommendations of Antibiogram Committee of the French Society of Microbiology [19].
E. coli ATCC 25922 strain was used as a control for antibiotic susceptibility testing.
Escherichia coli and Klebsiella spp. strains resistant to at least one third generation cephalosporin (cephalosporin, ceftazidim, ceftriaxon or cefotaxim) were collected in a storage medium (trypticase soy broth TCS) and stored at − 80°C. Samples were then sent under strict transportation conditions (in triple packaging boxes with ice packs), to Molecular Biology Laboratory of CERBA/LABIOGENE in Ouagadougou, Burkina Faso for qnr and ESBL genes detection on in January 2018.

Extraction of bacterial DNA
Rapid DNA extraction was performed using a boiling technique. Shortly, strains from TCS broth were reactivated on TCS agar for 18-24 h and two or three isolated colonies were inoculated in Luria Bertani (LB, 2 mL). After 18-24 h of overnight culture, LB broth were centrifuged at 10000 rpm/min for 10 min and the pellet suspended in 500 μL of phosphate buffer (100 mM, pH 7) to cell-wall weakening. The mix was heated at 100°C for 15 min in a water bath to release bacterial nucleic acid.
DNA was then precipitated in 250 μL of absolute ethanol, washed twice in 1000 μL of ethanol 75%, dried and re-suspended in 100 μL of sterile water.
Qnr genes (qnrA, qnrB and qnrS) amplification was performed using the following thermal cycling profile: 32 cycles consisting of 45 s at 95°C for denaturation, 45 s at 53°C for annealing and 60 s at 72°C for extension.
For the blaTEM and blaSHV ESBL genes, multiplex PCR amplification conditions were as follows: initial denaturation step at 94°C for 10 min; 30 cycles of denaturation at 94°C for 40s, annealing at 60°C for 40s, extension at 72°C for 1 min, followed by a final extension step at 72°C for 7 min. However, the amplification of BlaCTX-M-G1 was carried out as previously described by Pagani [23] in 25 μL reaction mixture according to the following PCR program: initial denaturation at 96°C for 10 min. 35 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min and extension at 72°C for 1 min. Final extension at 72°C for 10 min. Negative (DNA from E. coli ATCC 25922) and positive controls (DNA from qnr B and S genes positive strains) were used to check potential unspecific amplification. Specific sequences primers provided by Applied Biosystems (California, USA) are shown in Table 1. DNA fragments were analyzed by electrophoresis in a 2% agarose gel at 100 V for 1 h in TBE 1X containing ethidium bromide using 100-bp DNA ladder (Promega, USA) as a size marker.

Statistical analysis
Statistical analysis was performed using Epi Info Version 7.1.1.14 software. Fisher's exact test was used for comparison and the difference was statistically significant when p < 0.05.

Bacterial strains
A sample of 155 strains, 91 E. coli and 64 Klebsiella spp.

Antibiotic susceptibility profile
All E. coli strains were resistant to ceftriaxon and cefotaxim and 97.80% to ceftazidim. The resistance rates to other β-lactam antibiotics were 2.20% for imipenem (very low levels), 17.59% for piperacillin-tazobactam and 25.35% for cefoxitin. Quinolones, nalidixic acid and ciprofloxacin were very inactive with rates of 96.67% and 94.51% respectively. Among aminoglycosides, the strains were more resistant to gentamicin (75.82%) in contrast to amikacin which showed very low levels of resistance (3.30%).
All Klebsiella spp. strains were resistant to cefepim. Resistance to ceftazidime and ceftriaxone was 98.44% (63/ 64) each one, and to cefotaxime 96.88% (62/64). Nalidixic acid, ciprofloxacin, doxycycline and trimethoprimsulfamethoxazole were also inactive antibiotics with at least a resistance rate of 90%. Only imipenem, amikacin and fosfomycin were very active on Klebsiella spp. strains with a low resistance rate (< 5%). The resistance profile to other beta-lactams and other antibiotics for all isolates is presented in Table 2.

Distribution of qnr genes
Most isolates that were resistant to ciprofloxacin and nalidixic acid encoded qnrB and qnrS alone or in association but no qnr genes was detected in 29 E. coli and 15 Klebsiella spp. strains resistant to ciprofloxacin and nalidixic acid. Nevertheless, isolates encoding qnrB or qnrS were also identified among nalidixic acid-susceptible (10/ 105) and ciprofloxacin-susceptible strains (7/107) ( Table 4).   Results revealed that qnr subtypes (A, B, S) could coexist alone or in association with blaCTX-M1, blaTEM and blaSHV. Among E. coli strains, qnrBS combinations was most frequently associated with TEM/CTX-M1 combinations and among Klebsiella spp., the most frequent association was qnrB plus TEM/SHV/CTX-M1 (Table 4).

Discussion
Plasmid-mediated quinolone resistance may facilitate the spread and increase frequency of quinolone-resistant strains. Until now qnr genes have been widely detected in different parts of the world. Such data are not available in Togo. This is the first study which reports the frequency and diversity of qnr genes among ESBLproducing Enterobacteriaceae in Togo.
In Morocco, qnrB was found at 23%, qnr A at 10% and qnrS at 3% in 50% Klebsiella spp. and 18.7% E. coli [9]. In 2014 in Moroccan community enterobacteria, the prevalence of qnr gene was 2.6% (1.7% qnrS1 and 0.9% qnrB) [25]. These genes are usually plasmid mediated and can easily spread among the members of Enterobacteriaceae, through gene transfer mechanisms [3,6,17,18]. Results of plasmid isolation test and conjugation experiments in different studies indicated that these qnr gene were carried by conjugative plasmid of high molecular weight. These determinants can be transferred between bacteria, thus realizing the epidemic spread of quinolone resistance through horizontal gene transfer [11,13,[25][26][27][28]. However, due to the lack of financial resources, conjugation experiments or hybridization confirming the presence of target genes on plasmids were not performed in the present study.
Among strains encoding the qnr gene in our study, 90.48% were resistant to nalidixic acid, 93.46% to ciprofloxacin, 98.13% to ceftazidim and 99.07% to ceftriaxon. These rates are higher than those observed in Morocco (57% for nalidixic acid and 78% ciprofloxacin, 100% for ceftazidim and 71% for cefotaxim) [9]. In Mexico, the resistance rate among qnr positive pediatric strains was 41.1% for nalidixic acid, 29.4% for ciprofloxacin, 82.3% for ceftazidim and 100% for cefotaxim [13]. In this study, higher resistance rate of the qnr positive strains against nalidixic acid and ciprofloxacin could be explained by the concomitant presence of two or three qnr gene groups (43/107, 40.20%), also found in Algeria and in Vietnam, thus inducing an additive effect on the minimal inhibiting concentration (MICs) of these different molecules. In addition, qnr positive isolates showed more resistance to gentamicin (82.24%). This may be explained by the fact that plasmid-mediated quinolone resistance is associated with integrons bearing resistance determinants to several other antibiotics such as betalactams and aminoglycosides [3,4,27,29]. The qnr genes were identified among isolates which were susceptible to nalidixic acid and ciprofloxacin.
This result has clinical implications since the acquisition of the qnr genes by quinolone susceptible ESBLproducing strains could lead to selection of ciprofloxacin and cephalosporin resistant strains an increasing the Table 4 Co-existence of qnr gene and bla gene in E. coli and Klebsiella spp. mutant prevention concentration (MPC) [6,17]. However no qnr genes was detected in 29 E. coli and 15 Klebsiella spp. strains resistant to ciprofloxacin and nalidixic acid, can suggest the presence of another mechanism of resistance to quinolones such as mutations in the gyrase and topoisomerase IV genes [2,6,30]. The presence of ESBL and some of the quinoloneresistant genes in the same mobile genetic elements could explain the co-resistance to beta-lactams and fluoroquinolones. Our results showed that all E. coli and Klebsiella spp. strains qnr positive were ESBL-producing. Among qnr-positive strains, 102 produce CTX-M1, 52 SHV and 96 produce TEM.
Our results also revealed that qnr subtypes could coexist alone or in association with beta-lactamase genes.
Among E. coli strains, qnrBS combinations was most frequently associated with blaTEM/CTX-M1 combinations and among Klebsiella spp., the most frequent association was qnrB plus blaTEM/SHV/CTX-M1. The blaCTX-M1, plasmid-mediated class A ESBL [34][35][36] expression was observed to be currently more frequent in a double or triple combination with blaTEM and blaSHV. This genes combination suggests a progressive consolidation of resistance genes on a single mobile genetic element (plasmids, integrons, etc.).
These findings raise the hypothesis that the qnr genes detected in these strains could also have the same plasmid location. Plasmid isolation and conjugation experiments should be then investigated to confirm the presence of target genes on plasmids. Our results also suggested the community emergence of PMQR determinants (qnr gene) that contributed to the development and spread of fluoroquinolone resistance in E. coli and Klebsiella spp. isolates in Togo. The presence of these determinants in the outpatient is worrisome, due of the potential spread of plasmids in a scenario of uncontrolled oral quinolone usage, which can compromise therapeutic options and therefore concern for public health.
The potential limitations of this study were the absence of data on minimal inhibiting concentration (MIC) for nalidixic acid and ciprofloxacin to determine the level of bacterial resistance to these antibiotics and also the bias of including only ESBL strains and the absence of molecular typing of strains and sequence analysis of the different genes.
Altogether, the results of the present study underline the frequency of qnr determinants associated to fluoroquinolones resistance among E. coli and Klebsiella spp ESBLproducing strains in Togo and identifies the presence of qnr genes in quinolone-susceptible strains which could lead to in vivo selection of ciprofloxacin-resistant strains.

Conclusion
This first report of qnrA, qnrB and qnrS gene among ESBL-producing E. coli and Klebsiella spp. from Togo, extends upon similar finding in many countries supporting the wide distribution of qnr genes.
The results revealed a high rate of qnrB and qnrS alone or in combination and a higher association with blaTEM/CTX-M1 and blaTEM/SHV/CTX-M1 combinations. These qnr genes positive strains were highly resistant to nalidixic acid, ciprofloxacin, ceftazidim, ceftriaxon and gentamycin. However, they remain susceptible to imipenem, amikacin and fosfomycin.
The plasmid-mediated quinolone resistance genes and their association with cephalosporin resistance mediated by ESBL contribute to the spread of multidrug resistance due to their easy transfer between bacteria. Their wide dissemination impairs treatment outcome of common infections in community and hospitals settings. These finding suggest the strengthening of the public health policies in Togo in order to prevent, monitor and control antimicrobial resistance through the implementation of an antibiotic resistance surveillance system. Further studies on sequence analysis of the ESBL gene amplicons are also needed to determine different resistance profile of ESBL-producing bacteria in Togo.