Prevalence of Plasmid-Mediated Quinolone Resistance (PMQRs) Determinants and Whole Genome Sequence Screening of PMQR-Producing E. coli Isolated from Men Undergoing a Transrectal Prostate Biopsy

Fluoroquinolones (FQs) are recommended as prophylaxis for men undergoing transrectal prostate biopsy (TRUS-Bx). Recent studies suggest a significant share of FQ-resistant rectal flora in post-TRUST-Bx infections. Methods: 435 Enterobacterales isolates from 621 patients attending 12 urological departments in Poland were screened by PCR for PMQR genes. PMQR-positive isolates were tested for quinolone susceptibility and investigated by whole genome sequencing (WGS) methods. Results: In total, 32 (7.35%) E. coli strains with ciprofloxacin MIC in the range 0.125–32 mg/L harbored at least one PMQR gene. qnrS and qnrB were the most frequent genes detected in 16 and 12 isolates, respectively. WGS was performed for 28 of 32 PMQR-producing strains. A variety of serotypes and sequence types (STs) of E. coli was noticed. All strains carried at least one virulence gene. AMR genes that encoded resistance against different classes of antibiotics were identified. Additionally, five of 13 ciprofloxacin-susceptible E. coli had alterations in codon 83 of the GyrA subunits. Conclusion: This study provides information on the common presence of PMQRs among E. coli, which may explain the cause for development of post-TRUS-Bx infections. High numbers of virulence and antimicrobial resistance genes detected show a potential for analysed strains to develop infections.


Introduction
Fluoroquinolones (FQs) are an important class of synthetic broad-spectrum antibacterial agents used in medicine for treatment of multiple infections. However, their increased use has led to increasing bacterial resistance, mainly among different species of Enterobacterales, creating a challenge for the effectiveness and safety of therapies used against infections caused by these bacteria.
Recently, the emerging threat of infectious complications after transrectal ultrasoundguided prostate biopsy (TRUS-Bx) has been reported [1 -3]. The most common pathogen in the setting of post-biopsy infectious complications is Escherichia coli, with causative strains probably originating from the patient's own endogenous flora [4]. Moreover, the prevention and management of TRUS-Bx infectious due to E. coli has become more complicated due to widespread emergence of FQ-resistant E. coli isolates. Such strains may be selected as a result of any FQ therapy or by FQ prophylaxis conventionally used prior to biopsy [5]. Studies suggest that pre-biopsy screening for such FQ-resistant pathogens, with subsequent "tailored" prophylaxis based on antimicrobial susceptibility results, may be an effective way to reduce infectious complications [1, 6,7]. Therefore, there is a growing need to understand FQ-resistant strains by monitoring their prevalence and identifying predicting factors.
The main mechanism conferring high-level FQ resistance is the result of spontaneous point mutations within DNA gyrase and topoisomerase IV coding target enzymes for quinolones. A point mutation results in amino acid substitutions within the quinolone resistance determining region (QRDR) of gyrA and/or parC chromosomal genes encoding subunit A gyrase (GyrA) and subunit C (ParC) topoisomerase IV, respectively [8]. Although considerably less frequent, mutations in subunit B (GyrB) of DNA gyrase and subunit E (ParE) of topoisomerase IV have also been found to confer resistance to FQ. Meanwhile, in recent decades, plasmid-mediated quinolone resistance (PMQR) among Enterobacterales has become a global phenomenon [9]. PMQR mechanisms encoded by plasmid genes include Qnr peptides (capable of protecting DNA gyrase and topoisomerase IV from quinolones), aminoglycoside acetyl transferase aac(6 )-ib-cr (modifying quinolones with a piperazinyl substituent), and the quinolone efflux pumps-QepA and OqxAB [9]. The co-existence of mutations in QRDR and PMQR determinants in FQ-resistant Enterobacterales was described in previous works [9][10][11]. However, in the absence of chromosomally mediated quinolone resistance mechanisms (mutations), the acquisition of plasmid-mediated quinolone resistance mechanisms (PMQRs) leaves fluoroquinolones' MIC values in the susceptible category, according to the EUCAST breakpoints criteria. It should also be noted that the presence of PMQRs has been found to promote QRDR mutations, increasing FQ resistance rates [9]. In addition, PMQRs transmitted on plasmids are capable of rapid spread between bacteria of different species, therefore giving rise to potential complications for the use of fluoroquinolones as therapy or prophylaxis. Consequently, the presence of PMQR-positive bacteria within rectums of patients undergoing biopsy are considered an important risk factor for complications and may pose a serious challenge to the treatment of infections.
Some studies have previously investigated the range of post-TRUS-Bx complications in the context of FQ resistance. Recently, our team published a study on the molecular mechanisms of FQ-resistance including the gyrA, gyrB and parC, parE mutations and PMQRs of Enterobacterales from rectal swabs of 48 men undergoing transrectal prostate biopsy [12]. Until recently, little has been known about the molecular properties of E. coli as causative agents of post-biopsy infections. Thus, we aimed to determine the prevalence of PMQR determinants in Enterobacterales strains isolated from rectal swabs of 621 men undergoing transrectal prostate biopsy at 12 urological hospitals in Poland. Then, we investigated PMQR-producing E. coli strains by WGS for antibiotic resistance genes, virulotype, and genotype, that potentially could be responsible for infections associated with post-TRUS-Bx complications in patients in Poland.

Bacterial Isolates
A total of 435 Enterobacteriaceae strains were cultured from 621 rectal swabs of men undergoing prostate biopsy procedure, collected at 12 urological departments participating in the presented study. E. coli was the most prevalent gram-negative bacteria isolated. The main criterion for the selection of E. coli isolates for testing was the presence of PCR-product of the most commonly PMQRs i.e., qnrA, qnrB, qnrS, aac(6 )-Ib-cr, and qepA.

Whole Genome Sequencing Analysis of PMQR-Positive E. coli Strains
Data of whole genome sequencing analysis of 28 PMQR-positive E. coli isolated from rectal swabs of men undergoing transrectal prostate biopsy are presented in Table 2. Notably, nearly all genomes (except one) sequenced using the Nanopore platform were assemble into one large circular chromosome with few plasmids, which were usually also circular. For these strains, the chromosome size ranged from 4,544,229 bp to 5,258,810 bp (with an average of 4,895,805 bp). All these strains had 1-13 plasmids (with an average of five plasmids) and on average one of them was large (>100 kb). Among all 28 PMQRpositive E. coli tested, seven (IncFI, IncFII, IncX, IncI, IncY, IncN, Col types) plasmid replicons were identified with different frequencies ( Table 2).
Hypothetical pathogenic potential of analysed strains were confirmed by the Pathogen-Finder tool, with minimal and maximal probability of being a human pathogen estimated at 0.919 and 0.943 respectively (with an average of 0.93).
All genomes were deposited in the GenBank database under BioProject no PRJNA861130. Genotypic analysis of fluoroquinolone resistance. Point mutations involving amino acid substitutions among PMQR-positive E. coli were detected in 21 isolates out of 28 sequenced ( Table 2). The substitutions were observed in two codons of GyrA: 83 (Ser → Leu) and 87 (Asp → Asn); in three codons of ParC: 56 (Ala → Thr), 80 (Ser → Ile), and 84 (Glu→Gly, Val); in three codons of ParE: 416 (Leu → Phe), 458 (Ser → Ala) and 529 (Ile→Leu), respectively. No substitution was observed in the GyrB subunit in any of the isolates tested. The Ser83 → Leu mutation in the GyrA was the most common in PMQR-positive E. coli isolates, and was found in 20 isolates with a ciprofloxacin MIC at 0.125 → 32 mg/L, including five isolates of 13 susceptible to ciprofloxacin with ciprofloxacin MIC < 0.5 mg/L. The second most widespread gyrA mutation was Asp87 → Asn, found among 11 isolates. In parC, the most frequent mutation was Ser80→Ile, found in 12 isolates. Other identified mutations were observed infrequently. The PMQR-positive E. coli isolates resistant to ciprofloxacin with MIC > 4 mg/L had two or more substitutions, found in GyrA, ParC, and ParE (Table 2). Among genes encoding plasmid-mediated FQ resistance, qnrS1, qnrB19, and aac(6 )-Ib-cr were identified in 12, two, and three tested isolates, respectively.

Discussion
The genetic profiling of clinical strains can provide useful information about their potential for causing disease and resistance to treatment. This study investigated the prevalence of PMQR determinants and the WGS of PMQR-positive Enterobacterales strains isolated from men undergoing prostate biopsy procedures, collected at 12 urological departments in Poland. These isolates exhibited a diverse range of serotypes, sequence types (ST), virulotypes, and antimicrobial-resistant gene profiles.
In general, current data shows the prevalence of PMQR determinants in 7.35% of Enterobacterales isolates, with some qnrB and qnrS genes found in E. coli isolates susceptible to FQs (ciprofloxacin MICs 0.125-0.25 mg/L). These findings correspond with the popular hypothesis that PMQR determinants may promote mutations in the quinolone resistancedetermining region (QRDR) and, consequently, increase quinolone resistance in clinical settings, so that their presence may significantly contribute to the occurrence of postbiopsy complications [8][9][10]. Moreover, the findings strongly confirm the need to perform microbiological diagnostic tests (including PCR on PMQRs) before a biopsy, which has also been suggested by other authors [13][14][15].
According to the European Association of Urology (EAU), it is highly recommended to use antimicrobial prophylaxis in men prior to a transrectal prostate biopsy (TRUS-Bx), to minimise the risk of bacterial infection after the procedure [5]. All patients in our study from whom PMQR-producing isolates were collected received ciprofloxacin orally 2 h prior to the procedure. It should be noted that the PMQR resistance traits found in FQ-susceptible E. coli strains from patients receiving ciprofloxacin prophylaxis prior to TRUS-Bx indicated that FQ prophylaxis or treatment of post-biopsy complications may be ineffective in patients who carry PMQR-positive strains, and in consequence may promote development of high-level FQ resistance.
In our data, qnrS1 was dominant (78.6%) among ciprofloxacin-susceptible E. coli isolates, while qnrB was detected in 35.7% of such isolates. In the ciprofloxacin-resistant E. coli group with ciprofloxacin MIC ≥ 1.5 mg/L, qnrB and aac(6 )-Ib-cr determinants were dominant and were the same types of PMQRs reported for FQ resistant strains in other reports, [10,16]. These differences in proportions of different qnrs may provide additional premises for two probable means of selection of PMQR-positive strains playing essential roles in acquisition of quinolone resistance-foodborne and environmental (qnrS), or medical care (qnrB, aac(6 )-Ib-cr). Studies have suggested that the agricultural use of antimicrobial agents increased the number of human infections caused by drug-resistant bacteria [17]. Furthermore, consumption of food contaminated by PMQR-positive strains or use of FQs in ambulatory and hospital therapy may promote pre-selection factors for FQ resistance [12].
In many studies of the causes of post-biopsy complications, collections of FQ-resistant isolates have been investigated [1][2][3]18]. In those studies, FQ-resistant E. coli were found to occur frequently in strains isolated from men undergoing TRUS-Bx. In this prospective study, the only selection criterion was prevalence of PMQR. However, we found that 82 (18.8%) of the investigated E. coli strains were resistant to ciprofloxacin, with MIC > 0.5 mg/L. Differences in ciprofloxacin-resistance rates between patients from 12 urological centres were observed. The ciprofloxacin MICs for the 82 selected ciprofloxacinresistant E. coli were in the range 0.75->32 mg/L. The vast majority (n = 44; 53.7%) of ciprofloxacin-resistant isolates had ciprofloxacin MIC > 32 mg/L (data not shown).
WGS showed a huge variation among E. coli strains tested, as assessed by multilocus sequence typing (MLST) and serotyping. In the present study, E. coli strains O25:H4 ST131 were found to predominate among other 19 STs detected, as shown in Table 2. One strain ST131 had a different serotype within it i.e., O16:H5. However, ST131 E. coli isolates of serotype O16:H5 have also been identified in Japan, Denmark, Australia, and France [19][20][21][22]. Nowadays, ST 131 has been identified as the predominant E. coli lineage among extraintestinal pathogenic E. coli (ExPEC) isolates worldwide. E. coli ST131 isolates have been reported to produce extended-spectrum β-lactamases (ESBL), such as with them could be transferred genes encoding resistance to antimicrobials, ions, and other chemical components.
A limitation of our research was the performance of WGS analysis four years after susceptibility testing and PCR. This could have significantly influenced on the observed lack of correlation between the PMQR genes detected by the PCR method and the WGS analysis. qnr genes are localized on mobile elements and their presence can be affected by over-stabilizing the DNA-helicase complexes. Thus, they could be lost during laboratory manipulations without antimicrobial pressure. Such a situation has also been noticed by other scientists (personal communication during ECDC/EFSA fourth joint meeting on AMR for the FWD-Network and EURL-AR Network). This remains a hypothesis, and its verification needs further investigation.
In conclusion, by using whole genome sequencing, we obtained valuable information about bacterial strains and demonstrated that WGS can be a valuable early warning system for emerging resistance mechanisms and potential therapeutic failure. Our data also suggest that antimicrobial resistance is a potentially more important trait than virulence in terms of increased risk of post-biopsy infection. This is in line with the findings of other authors [33]. Meanwhile, the molecular characteristics including virulence and presence of antimicrobial resistance genes show the considerable potential of the investigated E. coli in the development of infection.

Bacterial Strains, Antimicrobial Susceptibility Testing and PCR
A prospective study was conducted between April and December 2016 at 12 urological departments in Poland. Ethics approval was obtained from the local ethics committee of the National Institute of Public Health National Institute of Hygiene (Research Ethics Committee No. 4/2015 from 2 December 2015), Warsaw, Poland. Rectal swabs were collected from 621 patients undergoing transrectal prostate biopsy. The samples were suspended in BHI broth with 20% glycerol, frozen at −20 • C and transported frozen to our laboratory. Rectal swabs were streaked onto a MacConkey agar plate (Oxoid Ltd., Basingstoke, UK) and were incubated at 37 • C for 24 h in aerobic conditions. From each plate with gram-negative bacterial growth, a representative colony of each distinct morphotype was selected for species identification by classical biochemical tests. Screening for qnrA, qnrB, qnrS, aac(6 )-Ib-cr, and qepA genes as more common PMQRs was carried out for all selected strains by using PCR as described previously [10].
All PMQR-positive isolates were screened with ciprofloxacin (5 µg) and nalidixic acid (30 µg) (Oxoid Ltd., Basingstoke, UK) on Mueller-Hinton agar (Oxoid Ltd., Basingstoke, UK). In addition, the E-test (bioMerieux, Marcy l'Etoile, France) method was used in all PMQR-positive isolates to determine the minimum inhibitory concentration (MIC) of ciprofloxacin. The susceptibility results were interpreted according to European Committee on Antimicrobial Susceptibility (EUCAST) criteria (http://eucast.org, accessed on 1 January 2022), and a zone of inhibition around the ciprofloxacin disc (<22 mm) along with MIC of ciprofloxacin >0.5 mg/L were classified as resistance to FQs. The MIC for each isolate was measured at least twice.

Whole Genomes Sequencing (WGS)
Twenty-eight of the PMQR-positive isolates recovered in this study were subjected to whole genome sequencing analysis (WGS). All steps of the WGS (DNA purification, library preparation and sequencing) were performed at the National Institute of Public Health NIH-National Research Institute. Libraries for 13 strains were prepared using an Illumina DNA Prep Kit (Illumina Inc., San Diego, CA, USA). The libraries were sequenced on an Illumina MiSeq sequencer (Illumina Inc.) using 150-bp paired-end reads. Libraries for another 15 strains were prepared using a Nanopore Rapid Barcoding Kit, and the whole genome sequencing was performed on GridION (Oxford Nanopore Technologies, Oxford, UK). Genomes were assembled using CLC Genomics Workbench 22 (for strains sequenced on Illumina) and the NanoForms server (for strains sequenced on Nanopore [34]). Species confirmation, multilocus sequence typing (MLST), serotyping, virulence and AMR (Antimicrobial Resistance) genes, and plasmid replicons were analysed using SpeciesFinder 2.0 [35], MLST 2.0 [36], SerotyperFinder 2.0 [37], VirulenceFinder 2.0 [38], ResFinder 4.1 [39], respectively, available on the Centre for Genomic Epidemiology website (http://www. genomicepidemiology.org/, accessed on 1 January 2022). Additionally, the hypothetical pathogenic potential was estimated by proteome analysis using PathogenFinder 1.1 [40].
Author Contributions: K.P. was responsible for the design of this study and choice of methods, data analysis, check and manuscript draft. K.Z. was responsible for phenotypic analysis, including strains isolation and identification, antibiotic susceptibility testing and PCR experiments and data analysis, performed WGS experiments. T.W. was responsible for performed WGS experiments, data analysis and participation in writing of the WGS part of the manuscript. R.G. was the main reviewer and editor of the final version of the manuscript. All authors have read and agreed to the published version of the manuscript. Informed Consent Statement: Informed consent was obtained from all subjects involved in the study.
Data Availability Statement: All data are presented within the article. The authors confirmed that personal identity information of the patients was unidentifiable from this report.