Characterization of Integrons and Quinolone Resistance in Clinical Escherichia coli Isolates in Mansoura City, Egypt

Escherichia coli is a common pathogen in both humans and animals. Quinolones are used to treat infections caused by Gram-negative bacteria, but resistance genes emerged. Only scarce studies investigated the association between plasmid-mediated quinolone resistance (PMQR) genes and integrons in clinical isolates of E. coli. The current study investigated the prevalence of quinolone resistance and integrons among 134 clinical E. coli isolates. Eighty (59.70%) isolates were quinolone-resistant, and 60/134 (44.77%) isolates were integron positive with the predominance of class I integrons (98.33%). There was a significant association between quinolone resistance and the presence of integrons (P < 0.0001). Isolates from Urology and Nephrology Center and Gastroenterology Hospital were significantly quinolone-resistant and integron positive (P ≤ 0.0005). Detection of PMQR genes on plasmids of integron-positive isolates showed that the active efflux pump genes oqxAB and qepA had the highest prevalence (72.22%), followed by the aminoglycoside acetyltransferase gene (aac(6′)-Ib-cr, 66.67%) and the quinolone resistance genes (qnr, 61.11%). Amplification and sequencing of integrons' variable regions illustrated that no quinolone resistance genes were detected, and the most predominant gene cassettes were for trimethoprim and aminoglycoside resistance including dfrA17, dfrB4, and dfrA17-aadA5. In conclusion, this study reported the high prevalence of PMQR genes and integrons among clinical E. coli isolates. Although PMQR genes are not cassette-born, they were associated with integrons' presence, which contributes to the widespread of quinolone resistance in Egypt.


Introduction
Despite being a well-known member of the normal flora for both humans and animals [1], the pathogenic strains of Escherichia coli cause a wide variety of infections including the skin and soft tissue, urinary tract, gastrointestinal tract, and central nervous system [2]. e severity of infections can vary from mild to life-threatening conditions based on the virulence capacity and antimicrobial resistance of the bacterium.
Fluoroquinolones are broad-spectrum synthetic antibiotics that can successfully treat infections caused by Enterobacteriaceae. Unfortunately, resistance to fluoroquinolones is increasing that is mediated chromosomally by modifications in their targets such as DNA gyrase and topoisomerase IV [3]. Plasmid-mediated quinolone resistance (PMQR) is now spreading among Gram-negative bacteria. ree main mechanisms are employed for PMQR resistance. e first is the modification of quinolones' targets by qnr proteins [encoded by quinolone resistance genes (qnr genes)]. Secondly, fluoroquinolones are modified through acetylation by the aminoglycoside acetyltransferase enzyme (encoded by the aac(6′)-Ib-cr gene). Lastly, excretion of hydrophobic fluoroquinolones by active efflux pumps such as QepA and OqxAB (encoded by qep, oqxA, and oqxB genes) [4].
Integrons are genetic elements that account for the dissemination of antimicrobial resistance genes in different species of bacteria. ey are found on plasmids or transposons, which facilitate their transfer between bacterial cells [5]. Two main parts constitute an integron. e first part is made up of an integrase gene (intI), an integration site (attI), and a promoter. e second part includes an attachment site (attC) for a gene cassette. is cassette has a single gene and mostly an open reading frame but no promoter. When attI and attC sites recombine, genes carried by gene cassettes can be expressed through the integron promoter [6]. Gene cassettes integrated into integrons carry antimicrobial resistance genes that facilitate dissemination of antimicrobial resistance upon transfer of plasmids or transposons that include the integrons [7,8]. Although different classes of integrons have been discussed before, class I and II integrons remain the most prevalent ones among resistant clinical isolates [9,10].
Although the association between PMQR and integrons among environmental isolates has been reported in several studies [8,[11][12][13], only few studies investigated the association between them in clinical E. coli isolates, and the investigations of the location of these genes on the variable region gene cassettes were scarce. erefore, our study aimed to investigate the prevalence and association of quinolone resistance and integrons among clinical E. coli isolates. Additionally, the prevalence of PMQR genes among integron-positive isolates was studied.  [14]. e study fulfills the ethical guidelines approved by " e Research Ethics Committee, Faculty of Pharmacy, Mansoura University" that follow the Code of Ethics of World Medical Association (Declaration of Helsinki involving the use and handling of human subjects; Code Number: 2021-9).

Genomic and Plasmid DNA Extraction.
Genomic DNA was obtained from all isolates using the modified boiling technique [16]. e filtrate was stored at −20°C and used as a template for polymerase chain reaction (PCR) amplification of integrase genes. e selection of isolates for plasmid DNA extraction was based on the results obtained from PCR amplification of integrons on total DNA in addition to quinolone sensitivity testing. Plasmid DNA extraction was done using the Gen-eJET Plasmid Miniprep kit (K0502, EU, Lithuania).

PCR Amplification of Integrase
Genes, Gene Cassettes, and (PMQR) Genes. Multiplex PCR was used for the detection of class I, II, and III integrons in all isolates using primers intI1, intI2, and intI3, respectively (Table 1). e variable regions of class I and II integrons and PMQR genes were amplified by PCR on plasmids extracted from selected 43 isolates using the primers listed in Table 1. PCR reaction mixture and cycling conditions were performed as previously described.

Restriction Fragment Length Polymorphism (RFLP) and
Characterization of Gene Cassettes. To investigate the similarity between the gene cassettes of the different isolates, RFLP of purified PCR products of variable regions of class I and II integrons (extracted from the agarose gel using GenJet Gel extraction kit (K0691, EU, Lithuania)) was detected after digestion by restriction enzyme HinfI (CutSmart, New England Biolabs) [21] and sequenced on an automated sequencer (ABI Prism 3100). e resulted sequences were assembled using the CodonCode Aligner (version 9.0.1). BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) against the GenBank database was performed for sequence comparison and annotation.

Nucleotide Sequence Accession Numbers.
e integron gene cassette nucleotide sequences identified in this study have been recorded in the GenBank database (GenBank accessions no. MW770320 to MW770348).

Statistical Analysis. GraphPad InStat (Fisher's exact and
Chi-square tests) was used to analyze the results statistically. Results with P-value ≤0.05 were considered of statistical significance.
It was found that urine isolates were significantly resistant to the tested quinolones (P < 0.0001). Besides, these isolates were significantly associated with UNC (P < 0.0001) and GEH (P � 0.0005).
PMQR genes were detected on plasmid DNA of the 36/ 43 isolates (83.72%) that showed phenotypic quinolone resistance and harbored integrons. e results in Table 2 showed that aac(6′)-Ib, oqxB, qnrS, and qepA were the most prevalent genes. None of the test isolates carried the qnrC gene. Seven isolates did not harbor any of the tested genes. e distribution of PMQR genes among different sources was illustrated in Table 2. It showed that all genes were significantly detected in urine samples (P < 0.0001), while isolates from a throat swab and blood carried one and two genes, respectively. Figure 1 illustrates the obtained 23 gene profiles. Isolates that carried 5 genes were all isolated from urine samples from UNC. Nine profiles were shown by more than 1 isolate (Pr1-Pr4, Pr6, Pr8, Pr11, Pr12, and Pr19). e most predominant profile was Pr4 (5 isolates, 11.62%). qnrA, qnrB, qnrD, and oqxA genes were not detected as a single-gene profile.
e qnrB gene was associated with profiles that carried 5 genes. e genes coexisted with each other without any significant association.   (5) -  regions of 1,600 bp (P < 0.0001) and 600 bp were more prevalent among isolates. e digestion of the amplified variable regions of integrons I and II with HinfI restriction enzyme resulted in seven different patterns (P1-P7) and two different patterns (P8 and P9), respectively. Figure 2 illustrates the restriction patterns corresponding to each size. e variable region of 1,600 bp gave the pattern P4 and was found in eighteen isolates. P7 that represents a 600 bp amplicon was found in eight isolates.
Gene sequencing of the amplified variable regions showed that seven different gene cassettes were detected among 23 class I integron-positive isolates and two gene cassettes in 2/4 of class II integron-positive isolates. Five gene cassettes had one gene either of dihydrofolate reductase (dfr family) or aminoglycosides resistance (aad family).

Discussion
Since the discovery of PMQR genes in 1998, a wide distribution of these determinants, especially in members of Enterobacteriaceae, was observed. is may be due to the broad use of fluoroquinolone antibiotics in human medicine, veterinary medicine, and the environment [22]. In the current study, fluoroquinolones' resistance was highly prevalent in the tested isolates (57.46-59.70%). A similar result was reported by a study performed in Egypt [23]. In contrast, previous studies from several countries [24][25][26][27][28] reported that most of the tested isolates were susceptible to fluoroquinolones (70-90%). e main causes behind antibiotic resistance are the misuse of antibiotics, improper prescription by physicians, and unawareness of patients who do not follow dosage regimens [29]. Furthermore, the use of quinolones in agriculture may account for an increase in PMQR genes in Enterobacteriaceae [30]. Fluoroquinolone resistance and E. coli isolated from urine were found to have a significant correlation, where fluoroquinolones were one of the most important therapeutic regimens used in the treatment of uropathogens [31]. erefore, this may explain the high prevalence of fluoroquinolone resistance among isolates from urinary tract infections, particularly in developing countries.
Integrons have been described as an acquired resistance mechanism by capturing, excising, and expressing gene cassettes through site-specific recombination, thus aiding in the spread of antibiotic resistance [32]. In this study, integrons were found in 44.77% of the tested isolates, which is similar to other studies [10,12]. Several studies, on the other hand, found a higher prevalence rate of integrons [33][34][35]. e present study revealed the predominance of integron I among quinolone-resistant isolates (59/80 isolates, 73.75%). In previous studies, integron I was more frequent among Gram-negative bacteria [36][37][38].     [39,40]. Other studies that investigated PMQR genes recorded variations in their prevalence rate even in the same country [41][42][43][44]. is variation may be due to the difference in study populations, types of isolates, selection criteria, and geographical distribution [45]. e active efflux pump genes, oqxAB and qepA, had the highest prevalence (72.22%), followed by aac(6′)-Ib-cr gene (66.67%) and qnr genes (61.11%). A study conducted in Egypt recorded aac(6′)-Ib-cr as the most frequent PMQR gene (61%), followed by qnrS (43.3%) [39]. Another study in China indicated a similar incidence rate of acc(6′)-Ib-cr and qepA genes [46]. Among the detected qnr genes, qnrS International Journal of Microbiology showed the highest prevalence, which agrees with previous studies [47][48][49]. Although detected in only 4.65% of our isolates, qnrB gene was reported as one of the most abundant genes in other studies [39,45,50]. qnrC gene was absent in all tested isolates, which agrees with other studies [45,51,52]. e qepA distribution rate in our results was similar to what was reported [52]. Several studies reported that aac(6′)-Ib-cr was detected more than qnr genes in Enterobacteriaceae [39,44,[53][54][55]. is may be attributed to the gene coding for this enzyme is highly prevalent in fluoroquinolone-resistant phenotypes. Besides, upon exposure to ciprofloxacin, this enzyme eases the selection of highresistant ciprofloxacin chromosomal mutants. In addition, the low level of fluoroquinolone resistance mediated by this enzyme can be converted to high-level resistance when coexisting with qnr proteins in the same bacterial isolates [56]. Seven isolates that showed quinolone resistance phenotypically did not harbor any of the tested genes on their plasmid. eir quinolone resistance may be attributed to other mechanisms that were not investigated in this study like chromosomal mutations of gyrA or parC genes or due to impermeability of the membranes [39].
In the current study, the 36 isolates harboring the PMQR genes showed 23 different gene profiles.
is result was consistent with what was reported earlier [49,57]. Concerning the coexistence of PMQR genes, no significant association was found. Similar results were reported previously [49,58]. Most isolates (23/36 isolates, 63.88%) showed more than 1 mechanism of PMQR, where 36.11% (13/36 isolates) of them had genes that encode the 3 PMQR mechanisms. A recent study in China has demonstrated the coexistence of three PMQR genes, including aac(6′)-Ib-cr, qnrS2, and oqxAB, on a multiple resistance plasmid [59]. A significant association was found between resistance mechanisms encoding efflux with qnr and aac(6′)-Ib-cr genes (P-value � 0.0179 and 0.0023, respectively). No significant association between resistance mechanisms encoded by aac(6′)-Ib-cr and qnr genes was found, which contrasts with other studies [39,60].
Among the 43 integron-positive E. coli isolates, gene cassettes were amplified in 32 isolates (74.41%), which is comparable to the results reported by Zhang et al. (72.7%) [61]. Although the isolates carried the integrase genes, the gene cassettes were not amplified in 26.19% and 50% of class I and II integron positive isolates, respectively. e lack or variation of a 3′-conserved segment [62] and the variation in the binding site of the primer or the extensive size of gene cassette [63,64] may explain the absence of gene cassettes in these isolates. e most abundant variable regions were of size 600 and 1,600 bp as they were found in 8 and 18 isolates, respectively. Similar results were reported earlier [65,66]. Identical restriction patterns of the same sized variable regions may be indicative of the presence of the same gene cassettes that were confirmed by sequencing.
Several gene cassettes belonging to class I integron have been described. Resistance to several antibiotics is mainly due to these cassettes [67]. In the current study, 23 out of 42 class I integron positive isolates carried 7 different gene cassettes (dfrA7, dfrA12, dfrA17, dfrB4, aadA22, dfrA17-aadA5, and aadA2-orfF-dfrA12). e reported cassettes encode for resistance to trimethoprim and aminoglycosides and have been recorded among class I integrons in previous studies [10,13,61,68,69]. e diversity of class I integron cassettes is comparable to previous studies [10,61] that reported 10 and 7 different gene cassettes among the tested isolates, respectively. In contrast to class I integrons, few different resistance cassettes have been associated with class II integrons [70]. In class II integron positive isolates of the current study, dfrA17-aadA5 and dfrA1-sat2-aadA1 gene cassettes were found. Zhang et al. [61] reported dfrA1-sat2-aadA1 as a single gene cassette among class II integron positive isolates. e gene cassettes in the present study showed the dominance of dfr genes within class I and II integrons. Similar results were reported previously [10,34,69]. is may be explained by the extensive use of trimethoprim as a first-line treatment for urinary tract infections in both hospital and community settings [34]. dfrA17, dfrB4, and dfrA17-aadA5 gene cassettes were found in nine, eight, and five isolates, respectively. is may be due to the interspecies transfer of the integrons. Integrons disseminate in hospitals through the cross-transmission of integron-carrying clones from one patient to another [34,71].
Concerning the variable region of 1,600 bp, four and seven variable regions carried dfrA17-aadA5 and dfrA17 gene cassettes, respectively. It is worth mentioning that these seven variable regions had also the aadA5 gene, which had a frameshift mutation that resulted in several internal stop codons and gaps. Similar results were reported by Chen et al. [66] where the 1,600 bp amplicon harbored dfrA17-aadA5 genes. is cassette array was reported in other bacteria and other regions around the world [9,[72][73][74][75]. is array can spread all over the world by self-transferable plasmids in humans or animals [76].
No PMQR genes were detected in the amplified gene cassettes. is indicates that quinolone resistance genes are not cassette-born. is coincides with the study of Kubomura et al. who found that except for dfrA and aadA genes, the antibiotic resistance genes are mostly found outside the integrons [69].
is may also suggest that resistance is related to other mobile genetic elements such as plasmids and transposons [77].

Conclusion
is study indicates the high prevalence of quinolone resistance and integrons in E. coli isolates, especially those from urine. PMQR genes were widely distributed among the tested isolates, implying lateral gene transfer. Gene sequencing of the amplified variable regions of integrons revealed that the most predominant gene cassettes were for trimethoprim and aminoglycoside resistance. Although PMQR genes are not cassette-born, they were associated with integrons' presence on the plasmids. Future studies to explain this association phenomenon should be performed. Continued surveillance of PMQR and integrons should be 8 International Journal of Microbiology conducted to control their spread and the associated health risks.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflict of interest.

Authors' Contributions
e authors contributed equally to this work. All the authors contributed to the conception of the study, design of the work, analysis, and drafting the work and read and approved the final manuscript.