Drug Resistance Determinants in Clinical Isolates of Enterococcus faecalis in Bangladesh: Identification of Oxazolidinone Resistance Gene optrA in ST59 and ST902 Lineages

Enterococcus faecalis is one of the major causes of urinary tract infection, showing acquired resistance to various classes of antimicrobials. The objective of this study was to determine the prevalence of drug resistance and its genetic determinants for E. faecalis clinical isolates in north-central Bangladesh. Among a total of 210 E. faecalis isolates, isolated from urine, the resistance rates to erythromycin, levofloxacin, and gentamicin (high level) were 85.2, 45.7, and 11.4%, respectively, while no isolates were resistant to ampicillin, vancomycin and teicoplanin. The most prevalent resistance gene was erm(B) (97%), and any of the four genes encoding aminoglycoside modifying enzyme (AME) were detected in 99 isolates (47%). The AME gene aac(6′)-Ie-aph(2”)-Ia was detected in 46 isolates (21.9%) and was diverse in terms of IS256-flanking patterns, which were associated with resistance level to gentamicin. Tetracycline resistance was ascribable to tet(M) (61%) and tet(L) (38%), and mutations in the quinolone resistance-determining region of both GyrA and ParC were identified in 44% of isolates. Five isolates (2.4%) exhibited non-susceptibility to linezolide (MIC, 4 μg/mL), and harbored the oxazolidinone resistance gene optrA, which was located in a novel genetic cluster containing the phenicol exporter gene fexA. The optrA-positive isolates belonged to ST59, ST902, and ST917 (CC59), while common lineages of other multiple drug-resistant isolates were ST6, ST28, CC16, and CC116. The present study first revealed the prevalence of drug resistance determinants of E. faecalis and their genetic profiles in Bangladesh.


Introduction
Enterococcus is ubiquitously distributed in the environment and constitutes the normal flora of the intestinal tract in humans and animals. However, among this genus, particularly Enterococcus faecalis and Enterococcus faecium have been recognized as the common opportunistic pathogens implicated in urinary tract infections, wound/surgical site infections, and catheter-associated infections of the bloodstream [1]. E. faecalis is far more frequently isolated from clinical specimens than E. faecium, causing more intractable infectious disease, which is related to an increased level of drug resistance and the ability to form biofilm [2].
E. faecalis has an ability to acquire resistance to several antimicrobials such as aminoglycoside (high-level), penicillins, glycopeptides, quinolones, tetracyclines, and macrolides, via transposons, plasmids, or mutations, while it exhibits intrinsically reduced susceptibility to cephalosporins, aminoglycoside, lincosamide, trimethoprim-sulphamethoxazole [3]. Traditionally, resistance to penicillin, aminoglycoside (high-level), glycopeptide has been the major issue of clinical practice in the treatment of enterococcal infections. For such drug-resistant E. faecalis strains, represented by those with vancomycin resistance, newer compounds including linezolid, daptomycin, and tigecyclines have been used recently. Linezolid, a substance of oxazolidinone, has been a promising agent against infections caused by vancomycin-resistant enterococci (VRE), methicillin-resistant S. aureus (MRSA), and penicillin-resistant pneumococci since the turn of the twenty-first century [4]. However, during the past decade, resistance to oxazolidinones and daptomycin has been reported in many parts of the world, posing a concern for a serious limitation in the treatment of human infections caused by VRE and MRSA [5][6][7]. Linezolid resistance is attributable to mutations in the 23S rRNA and rplC/rplD encoding the 50S ribosomal proteins L3/L4, or the acquisition of optrA encoding an ATP-binding cassette (ABC)-F protein and cfr variant coding for 23S rRNA methyltransferase [5]. Daptomycin resistance is associated with increased cell wall thickening and other structural alterations of the cell wall caused by mutations in various genes responsible for call wall biosynthesis [5,6].
While aminoglycoside (gentamicin) combined with ampicillin has been the standard regimens of treatment for infective endocarditis due to enterococcus [8], recently designed regimens also include adaptomycin, tigecycline, and ceftaroline, together with ampicillin, gentamicin, or fosfomycin [2]. Moreover, the increased expression of tetracycline resistance determinants (tet(L), tet(M)) was revealed to confer tigecycline resistance to E. faecium [5,9]. Accordingly, for the control of enterococcal infections, it is essential to understand the prevalence of resistance to conventional drugs as well as new compounds, and also their resistance mechanisms.
In Bangladesh, only limited information is available for the drug resistance of Enterococcus compared with S. aureus and Gram-negative bacteria [10]. Enterococcus was described as the second most frequent bacterial species from puerperal infections [11] and the third most common isolate from urinary tract infections [12]. In our previous study of puerperal infections, though the number of clinical isolates was low, E. faecalis was totally resistant to gentamycin, and approximately half of the isolates showed resistance to minocycline, erythromycin, and levofloxacin [11]. Similarly, high resistance rates to gentamicin and ciprofloxacin were observed for E. faecalis isolates from urinary tract infections [12,13], suggesting the spread of drug-resistant enterococcal strains in this country. However, genetic mechanisms of the resistance have not yet been studied. In the present study, we investigated the prevalence of resistance to clinically important antimicrobials and genetic mechanisms of the resistance for E. faecalis clinical isolates, to delineate the comprehensively drug resistant traits of this species in Bangladesh. We describe here the first identification of oxazolidinone resistance gene optrA in a novel fexA-optrA cluster in the country, in addition to characteristics of the genetic features of high-level gentamicin resistance gene, and other determinants of the conventional drug resistance of E. faecalis.

Bacterial Isolates
A total of 210 E. faecalis clinical isolates from the urine specimens of patients with urinary tract infections were analyzed. These isolates were collected in Mymensingh Medical College (MMC) hospital and Swadesh private hospital in Mymensingh, Bangladesh, consecutively, for a 15 month period starting from January 2018. The MMC hospital is only a tertiary referral medical center in the Mymensingh division, having 1000 beds, and the Swadesh private hospital is a 30 bedded private care facility. The age range of patients with a urinary tract infection was 1-80 years, while the sex ratio (male/female) was 1.3 (120/90). Only one isolate per patient was included in this study. Urine specimens were inoculated onto a chromogenic agar plate (HiCrome UTI Agar, HiMedia Laboratories), followed by incubation at 37 • C for 48 h aerobically. Bacterial colonies on the agar plates, which were morphologically assigned to E. faecalis, were picked up and further examined by Gram staining, catalase test, oxidase test and nitrate reduction test [4]. All the isolates were confirmed as E. faecalis by the PCR targeting PBP5 gene, with the use of primers and conditions reported previously [14]. For some isolates which could not be identified as E. faecalis by the PCR, the species was confirmed by the determination of the 16S rRNA gene sequence through direct sequencing with PCR product amplified by specific primers, as previously described [15]. Individual isolates were stored in Microbank (Pro-Lab Diagnostics, Richmond Hill, ON, Canada) at −80 • C, and were recovered when they were analyzed. This study was approved by the Institutional Review Board of MMC.

Genetic Analysis of aac(6 )-Ie-aph(2 )-Ia
For all the isolates having aac(6 )-Ie-aph(2")-Ia genes, IS256-flanking pattern (A, B, C or D) was assigned by PCR using the primers reported by Watanabe et al. [24]. For selected isolates with different IS256-flanking patterns and MIC to GEN, full-length gene sequences of aac(6 )-Ie-aph(2")-Ia were determined directly from PCR products using the BigDye Terminator v. 3.1 Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) on an automated DNA sequencer (ABI PRISM 3100). The primers used for sequencing are shown in Table S1.

Genetic Determinants of Oxazolidinone and Daptomycin Resistance
Isolates exhibiting non-susceptibility to linezolid (MIC, ≥4 µg/mL) were further analyzed for the presences of cfr, fexA, optrA and mutation in 23S rRNA, L3-and L4-encoding genes as described previously [22,[25][26][27]. The nucleotide sequence of the fexA-optrA gene cluster was determined for LZD-non-susceptible isolates by PCR and direct sequencing using the primers designed in this study (Table S1). In addition, for the isolates exhibiting a different MIC to daptomycin, the pgsA gene, which encodes phosphatidyl glycerophosphate synthase [6] was also sequenced with primers designed in this study (Table S1). The multiple alignment of nucleotide/amino acid sequences determined in the present study and those retrieved from the GenBank database was performed by Clustal Omega program (https://www.ebi.ac.uk/Tools/msa/clustalo/), which was also used for the calculation of sequence identity.

Multilocus Sequence Typing (MLST)
For selected isolates having different drug resistance profiles and IS256-flanking pattern of aac(6 )-Ie-aph(2")-Ia, the sequence type (ST) based on the MLST scheme [28] was identified using the web-based genotyping tool PubMLST (https://pubmlst.org/efaecalis/). The MLST data were further assigned to the clonal complex (CC) by BURST analysis available in the PubMLST website.

GenBank Accession Numbers
The nucleotide sequences of fexA-optrA and its cluster, pgsA, and aac(6 )-Ie-aph(2")-Ia were deposited in the GenBank database under the accession numbers listed in Table S2.

Prevalence of Antimicrobial Resistance and Resistance Determinants
The resistance rates to twelve antimicrobials and the detection rates of resistance determinants of 210 E. faecalis isolates are shown in Table 1. Resistance to ERY was found in 85.2% of all isolates, with erm(B) being highly prevalent (97.1%). Approximately 60% of isolates exhibited TET resistance, associated with tet(M) (61%) and/or tet(L) (38.1%). High-level resistance to gentamicin (GEN-HLR) was detected in 24 isolates (11.4%), while aac(6 )-Ie-aph(2")-Ia was identified in 46 isolates (21.9%). Among all the E. faecalis, 99 isolates (47%) harbored any of the four AME genes, among which aph(3 )-IIIa was the most commonly detected (34.8%). Resistance to LVX was found in 45.7% of isolates, which had mostly double mutations, S84I in GyrA and S82I in ParC. Resistance to NIT was found in 10.5%, and no isolates were resistant to AMP, IPM, VAN, TEC, and FOF.
optrA and fexA were identified in five isolates (2.4%), which showed non-susceptibility to LZD (MIC, 4 µg/mL), and also resistance to CHL and FFC, showing a higher MIC than optrA-negative isolates, although the increased MIC of TDZ was not observed (Table S3). To the patients with these isolates, LZD had never been administered for treatment. In these isolates, the cfr gene was not detected, and no mutation was identified in the 23S rRNA gene (V domain) and L3-and L4-encoding genes (data not shown). Only a LZD-non-susceptible isolate (SJ116) showed non-susceptibility to DAP (MIC, 8 µg/mL). These five isolates had identical sequence of pgsA, of which the deduced amino acid sequences of the protein products were identical to those of E. faecalis strains S613 and R712, which were described as DAP-susceptible and resistant, respectively [6]. Single amino acid substitution, which was reported to be involved in DAP resistance in S. aureus [6], was not found in the five optrA-positive E. faecalis isolates ( Figure S1).
Nucleotide sequences of aac(6 )-Ie-aph(2")-Ia and its 5 -and 3 -end regions were determined for nine isolates representing the IS256 flanking patterns A through D, and their genetic structures are shown schematically in Figure 1. The five isolates with pattern C examined were assigned to three different types (C1-C3). Although isolates showing GEN-HLR had C1 and C2 types which had intact aac(6 )-Ie-aph(2")-Ia with IS256 at different positions in its downstream, two GEN-susceptible isolates had a pseudogene of aac(6 )-Ie-aph(2")-Ia, which was truncated by IS1216 and lacked 5 -end region including start codon of this AME gene (C3 type). Two isolates of pattern D possessed intact aac(6 )-Ie-aph(2")-Ia, despite showing non-GEN-HLR.

ST of Isolates with Different Characteristics
ST was identified for 39 E. faecalis isolates with different profiles of drug resistance and resistance genes, including those with different IS256-flanking patterns of aac(6 )-Ie-aph(2")-Ia and those with fexA-optrA (Table 2). Isolates showing GEN-HLR were generally resistant to multiple classes of antimicrobials (macrolide, quinolone, tetracycline) and mostly belonged to ST6, ST28, and STs of CC28 and CC116. In contrast, isolates with pattern D-aac(6 )-Ie-aph(2")-Ia were resistant to lesser number of the drugs examined and mostly assigned to CC16. Isolates without aac(6 )-Ie-aph(2")-Ia and optrA had less resistance determinants showing resistance to less antimicrobials, and included the isolates of ST28 and its relevant STs (ST945, ST919). Five isolates having fexA-optrA were classified into ST59 (two isolates), ST902 (triple-locus variant of ST21, two isolates), and ST917 (single-locus variant of ST59, one isolate). Although these isolates had similar profile of resistance genes showing multiple drug resistance, CC59 isolates were resistant to LVX associated with QRDR mutations in GyrA and ParC, unlike ST902 isolates. Table 2. Genotypes, antimicrobial resistance profile and resistance genes/genetic determinants of selected E.faecalis isolates (n = 39).

Genetic Background of fexA and optrA
The five isolates with non-susceptibility to LZD were analyzed for the fexA-optrA cluster, which comprises fexA, a short open reading frame, and optrA, with the same orientation. Nucleotide sequences of the fexA-optrA cluster from the five isolates were identical to that reported for E. faecalis strain 743,142 and 981,649 (Taiwan, clinical isolates) plasmids (GenBank accession nos. MF443377 and MF443386, respectively) by BLAST search. The deduced amino acid sequences of optrA (655 amino acids) of the five E. faecalis isolates have three divergent amino acids (K3E, Y176D, G393D) compared with the OptrA prototype in E. faecalis strain E349 (GenBank accession no. KP399637) [22], which corresponds to the "EDD variant" [29]. FexA (475 amino acids) of Bangladeshi isolates was different from that of strain E349 by four amino acids ( Figure S2). When sequences of this gene of E. faecalis, and other enterococcal species and staphylococcus were retrieved from GenBank database, at least ten variants of FexA were identified, having divergent amino acids at 15 positions with 97.7-99.8% identity ( Figure  S2).
Strain SJ82 was further analyzed for the broader region containing fexA-optrA cluster ( Figure 2) and compared with this region, reported for various strains previously [30]. This region of SJ82 contained tnpB, tnpC, spc, erm(A) upstream from fexA, having an identical sequence to those in the E. faecalis strains 743,142 and 981,649. Although similar genetic structure is seen in also E. faecalis strain TZ2 and E1731 (China), sequences of the connecting portion between fexA and erm(A) were different from SJ82. In the downstream of optrA, three genes (cspC, RNase J, and efrA) are located as seen in E. faecalis strains 981,649 (Taiwan), TZ2 (China), 973,450 (France), C54 (China), D32 (Denmark), and NCTC8745 [30]. The nucleotide sequence of RNase J was identical to that of strains C54, D32, and NCTC8745, while it was slightly different from strains 981,649 and 743,142 (sequence identity 98.9-99.2%) with divergent positions being dispersed in this gene ( Figure S3). The genetic organization of the fexA-optrA cluster was quite distinct from that of the prototype of optrA in pE349 (GenBank accession no. KP399637) [22]. As a whole genetic organization of the fexA-optrA cluster region of SJ82 was similar to that of the E. faecalis strain 981649, except for RNase J which was identical to that of other strains (e.g., C54). (uppermost) and the genetic organization or components similar to that of SJ82 in other strains reported previously [30] or available in GenBank database. Prototype of the fexA-optrA cluster in the pE349 of E. faecalis strain E349 [22] is shown at the bottom. Arrows indicate the transcription direction of genes. Arrows of RNase J are shown in black and blue, representing different sequences. Gene names are shown above arrows, and the strain names are indicated on the right.

Discussion
The present study first characterized E. faecalis from urinary tract infections in Bangladesh for the situation of antimicrobial susceptibility and genetic factors related to drug resistance. While resistance to penicillins and glycopeptides was absent, higher resistance rates were noted for ERY, TET, and LVX. The prevalence of ERY resistance (85.2%) was higher than in our previous study of puerperal infections in Bangladesh [11], and comparable to that reported for blood isolates in the US and Europe [31], and hospital isolates in the middle East [32]. The resistance rate to LVX (45.7%) was similar to that in our previous studies in Bangladesh [11] and the Middle East [32], but appears to be (uppermost) and the genetic organization or components similar to that of SJ82 in other strains reported previously [30] or available in GenBank database. Prototype of the fexA-optrA cluster in the pE349 of E. faecalis strain E349 [22] is shown at the bottom. Arrows indicate the transcription direction of genes. Arrows of RNase J are shown in black and blue, representing different sequences. Gene names are shown above arrows, and the strain names are indicated on the right.

Discussion
The present study first characterized E. faecalis from urinary tract infections in Bangladesh for the situation of antimicrobial susceptibility and genetic factors related to drug resistance. While resistance to penicillins and glycopeptides was absent, higher resistance rates were noted for ERY, TET, and LVX. The prevalence of ERY resistance (85.2%) was higher than in our previous study of puerperal infections in Bangladesh [11], and comparable to that reported for blood isolates in the US and Europe [31], and hospital isolates in the middle East [32]. The resistance rate to LVX (45.7%) was similar to that in our previous studies in Bangladesh [11] and the Middle East [32], but appears to be higher than the global average [33]. TET resistance rate was much less prevalent than in the US and Europe [31]. The susceptibility rate of E. faecalis to nitrofurantoin, which is commonly prescribed for urinary tract infections, was higher in our study (89.5%) than in previous reports in Bangladesh and India (78-86%) [12,13,34].
The prevalence of GEN-HLR in the present study (11.4%) was substantially lower than in recent reports from India, the Middle East, Australia, and Japan (22-55%) [32,[34][35][36][37]. However, aac(6 )-Ie-aph(2")-Ia, which is responsible for GEN-HLR, was more prevalent (21.9%) than the phenotypically detected GEN-HLR. This discordance is considered to be in part related to diversity in IS256-flanking patterns and also the genetic alteration of this gene, as observed in our present study. Although GEN-HLR was found in the isolates with pattern A and B, a portion of pattern C and D isolates was not highly resistant to GEN. The presence of IS256 at both sides of aac(6 )-Ie-aph(2")-Ia (pattern A), which represents Tn4001 [38], is relevant to GEN-HLR via elevated transcription of the AME gene due to the presence of IS256 [39]. In contrast, aac(6 )-Ie-aph(2")-Ia might not be transcribed efficiently in most of pattern D isolates which lacked adjacent IS256, moreover, this AME is unlikely to be produced in pattern C isolates having a pseudogene lacking a start codon of this gene. In the present study, pattern D was the most frequent, and accounted for half of the isolates harboring aac(6 )-Ie-aph(2")-Ia. Similarly, the dominance of pattern C and D representing truncated transposon structures of this gene was observed among the clinical isolates in other countries [24,40,41]. In addition, some variations in pattern C were also found in other reports [40,41]. It is suggested that intact Tn4001-like transposon containing this AME gene might have evolved through recombination events [41], and the resultant variants of this AME gene associated with non-GEN-HLR might have been persisting probably due to less exposure of GEN to E. faecalis in current antibiotic therapy.
It was remarkable that oxazolidinone resistance-determinant optrA was identified in five E. faecalis isolates showing non-susceptibility to LZD, although these isolates were derived from urinary tract infections without the use of this antimicrobial for treatment. The protein product of optrA belongs to the ATP-binding cassette (ABC)-F protein superfamily that mediates resistance through ribosomal protection [42], which is more of a common cause of oxazolidinone resistance in enterococci [30,43,44], compared with other genetic determinants cfr or mutations in 23S rRNA and ribosomal proteins. LZD-resistant E. faecalis have been reported worldwide, particularly more frequently in China [7]. Although the non-susceptibility rate to the LZD of enterococci was reported as <1% at the global level [7,30], the prevalence of LZD-resistance/non-susceptibility or optrA is evidently higher in China (1-4%) [22,[44][45][46][47], with a higher rate in E. faecalis than in E. faecium. Moreover, in China, the increase in optrA-positive rate in enterococci was documented from 2004 (0.4%) to 2014 (3.9%) [46], and E. faecalis harboring optrA was isolated from the stool samples of healthy adults and children with a carriage rate of 3.5% [29]. The present detection rate of optrA in E. faecalis in Bangladesh (2.4%) may be comparable to those reported in China, implying relatively high prevalence among clinical isolates in this country, and may suggest also the potential distribution among the healthy population since LZD has been rarely used. The occurrence of optrA-positive E. faecalis was reported also in patients having no history of LZD treatment in Korea [48]. Because the five LZD-non-susceptible isolates in our study exhibited resistance to multiple drugs including NIT, it is possible that these isolates might have occurred by selection with ordinary antimicrobial treatment.
The five optrA-positive E. faecalis isolates had an "EDD" variant which represents one of the 25 types of OptrA amino acid sequence [29]. The isolates with the "EDD" variant in China showed intermediate resistance to LZD (4 µg/mL), as observed in our present study, and accounted for 24% and 13% of the optrA-positive isolates from healthy humans and clinical isolates, respectively [29,46]. In contrast, a higher MIC to LZD (≥8 µg/mL) was evident for the isolates with wild-type optrA and some variants, e.g., "RDK" [22,29]. Although E. faecalis, with intermediate resistance to LZD from urinary tract infections, as those detected in our study, may not cause issues in treatment, such isolates are suggested to increase the potential risk for opportunistic cross-infections in a healthcare setting.
Various genetic backgrounds containing optrA have been documented in the context of plasmid or chromosome of enterococci [30]. The fexA-optrA cluster, which was detected in E. faecalis in Bangladesh, is one of the major genetic structures in clinical isolates from humans, including the prototype strain E349 [29], as well as isolates from animals (chicken, pig) and retail meat [18,30,53]. tnpB and tnpC were associated with the fexA-optrA region of strain SJ82, which suggested that this optrA cluster may be located on a Tn554 family transposon (e.g., Tn6674) as described for that containing optrA in E. faecalis [54]. In the present study, the fexA-optrA cluster and its upstream region including the erm(A) and spc of the Bangladeshi strain SJ82 were found to be genetically distinct from those of E349, but identical to those of E. faecalis clinical isolates in Taiwan, and highly similar to those of E. faecalis from humans and animals in China [30]. In contrast, the NRase J gene located downstream from optrA was identical to other strains in China and other countries [30]. The identification of such a novel genetic background of optrA in strain SJ82 suggests the occurrence of recombinations in optrA-containing regions among E. faecalis distributed in humans and/or animals in Asia. Since this optrA region contains other resistance genes (erm(A), spc, and fexA), it is possible that the selective dissemination of optrA may be caused by the use of macrolides and spectinomycin to humans, or florfenicol to animals. Among the five isolates with optrA, an isolate SJ116 exhibited non-susceptibility to DAP. In the present study, no mutation was detected in the pgsA of strains SJ116 and also previously reported DAP-non-susceptible strains, although mutation in pgsA was involved in DAP resistance in S. aureus and Corynebacterium striatum [6]. Although various genes are revealed to be implicated in DAP resistance in enterococci [5], it was suggested that a mutation in pgsA may not mediate the primary role in decreased susceptibility to DAP.

Conclusions
The present study on E. faecalis clinical isolates in Bangladesh revealed considerably high resistance rates to ERY and LVX, while a low rate of GEN-HLR and none with glycopeptide and penicillin resistance among them. LZD-non-susceptible isolates harboring optrA were first identified in this country, indicating the need for further epidemiological investigation to determine its prevalence and clinical significance.