Emergence of fluoroquinolone resistance and possible mechanisms in clinical isolates of Stenotrophomonas maltophilia from Iran

Stenotrophomonas maltophilia exhibits wide spectrum of fluoroquinolone resistance using different mechanisms as multidrug efflux pumps and Smqnr alleles. Here, the role of smeDEF, smeVWX efflux genes and contribution of Smqnr alleles in the development of fluoroquinolone resistance was assessed. Ciprofloxacin, levofloxacin and moxifloxacin resistance were found in 10.9%, 3.5%, and 1.6% of isolates, respectively. More than four-fold differences in ciprofloxacin MICs were detected in the presence of reserpine and smeD, F, V expression was significantly associated with ciprofloxacin resistance (p = 0.017 for smeD, 0.003 for smeF, and 0.001 for smeV). Smqnr gene was found in 52% of the ciprofloxacin-resistant isolates and Smqnr8 was the most common allele detected. Fluoroquinolone resistance in S. maltophilia clinical isolates was significantly associated with active efflux pumps. There was no correlation between the Smqnr alleles and ciprofloxacin resistance; however, contribution of the Smqnr genes in low-level levofloxacin resistance was revealed.

. Furthermore, S. maltophilia harbors a novel quinolone resistance gene, namely Smqnr which is encoded by the chromosome, rather than plasmid-mediated qnr genes 15 . Therefore, development of resistance to quinolones and the relevant resistance mechanisms are not fully described in S. maltophilia 16 .
Despite the increasing prevalence of antibiotic resistance and a considerable resistance against fluoroquinolones (0-20%) in clinical isolates of S. maltophilia in Iran, mechanisms of fluoroquinolone resistance in Iranian isolates of S. maltophilia were not completely studied 17,18 . Here, the genetic background of resistance to fluoroquinolones including the role of active efflux pumps and their gene expression, the effect of reserpine as an efflux pump inhibitor on minimum inhibitory concentrations (MICs), and association of Smqnr alleles with fluoroquinolone resistance in clinical isolates of S. maltophilia in Iran was sought.
Fluoroquinolone susceptibility of S. maltophilia strains. According to the disc diffusion method, ciprofloxacin resistance was found in 41 (10.9%) strains and among the remaining strains, 113 (30.1%) showed intermediate susceptibility to ciprofloxacin and 221 (58.9%) were ciprofloxacin-susceptible. Thirteen (3.5%) and 6 (1.6%) strains showed resistance or intermediate susceptibility to levofloxacin. Majority of the strains was susceptible to moxifloxacin (369, 98.4%), one strain was intermediate susceptible and only five strains (1.3%) were resistant to moxifloxacin. Table 1 shows the susceptibility profile of the S. maltophilia strains to ciprofloxacin and levofloxacin. Based on the MICs, 48 ciprofloxacin-and 4 levofloxacin-resistant strains were identified with the MICs equivalent or greater than 2 and 8 µg/mL, respectively. Furthermore, MIC90 of ciprofloxacin was ≤ 32 µg/ mL while that was ≤ 2 µg/mL for levofloxacin. The presence of smeDEF and smeVWX genes. Among the 48 strains ciprofloxacin-resistant strains, smeE and smeF genes were not detected in 3 and 8 strains, respectively using PCR. The remaining strains yielded amplicons for smeD, smeE, smeF, smeV, smeW and smeX genes (Supplementary Table S1). There were not significant differences in the MICs of ciprofloxacin/levofloxacin among the smeDEF-positive and smeDEF-negative S. maltophilia strains (p > 0.05).

Effect of reserpine on ciprofloxacin
Expression of smeD, smeF and smeV genes. Twenty-nine out of 48 ciprofloxacin-resistant strains were detected with ≥ threefold expression of smeD gene. Compared to the S. maltophilia ATCC13637, overexpression of smeF gene was found in 24 strains and expression level of smeV gene was ≥ 3 folds in 9 strains. Overexpression of the three efflux pump genes tested were noted in 3 out of 4 levofloxacin-resistant strains and one levofloxacinresistant strain showed overexpression for smeD and smeF genes but not smeV. The expression of smeD, F, V genes was significantly correlated with higher MICs of ciprofloxacin and this correlation was also found between smeV gene and levofloxacin, compared to ATCC13637 standard strain. The expressions level of smeD, smeF, and smeV genes are demonstrated in Fig. 2.    Co-effect of efflux pumps and Smqnr alleles on ciprofloxacin MICs. Based on the expression of efflux pump genes and/or presence of Smqnr alleles, the isolates were classified into the following sets: (a) the isolates which harbored Smqnr alleles and had smeDEF overexpression (14 isolates); 57.1% (4 isolates) of these isolates had MIC ≥ 4 µg/mL but the differences of ciprofloxacin MICs among the isolates in this group was not statistically significant compared to the isolates without efflux pump genes overexpression and Smqnr alleles (p = 0.08), (b) the isolates having smeVWX overexpression and Smqnr alleles (four isolates); all these were among the resistance isolates and there were no significant differences in the ciprofloxacin MICs of isolates with smeVWX overexpression and Smqnr alleles in comparison with the isolates which did not have efflux pump genes overexpression and Smqnr alleles (p < 0.05), (c) the isolates with overexpression of smeDEF and smeVWX and Smqnr alleles(four isolates); all the isolates in this category showed MIC ≥ 4 µg/mL. The comparison of ciprofloxacin MICs between this recent group and the isolates with no overexpression of efflux pump genes and Smqnr alleles was significant (p = 0.002).

Discussion
Intrinsic resistance nature of S. maltophilia against multiple antimicrobial agents and limited therapeutic options made great concern to control the increasing S. maltophilia nosocomial infections. However, resistance rate of S. maltophilia strains varies depending on different geographical areas. In this study, a large series of isolates from Tehran and a neighboring province were studied and we found a susceptibility rate of fluoroquinolones (89% for ciprofloxacin and 96.5% for levofloxacin) similar to the previous study (84.1% for ciprofloxacin and 99.4% for levofloxacin) in Iran which studied 44 and 45 isolates 18,19 . The resistance to ciprofloxacin ranged from 13 to 96% globally 3 and fluoroquinolone resistance in neighboring countries of Iran was as follows: an increasing rate of resistance from 7.8% (1998)(1999)(2000)(2001)(2002)(2003) 20 to 89% (2006-2013) 21 in Turkey, 64-100% in Pakistan 22 and 23% (2003-2009) in Saudi Arabia 23 . Overexpression of smeDEF and smeVWX genes might contribute to increased MICs of multiple antibiotics and developing multi-drug resistant S. maltophilia strains 24 . In this study, the average of increased MICs of ciprofloxacin in smeD, F, V overexpressed strains was statistically significant compared to the isolates with no overexpression. The differences of MICs of levofloxacin were statistically significant only when the smeV gene is overexpressed. Furthermore, the differences of MICs of ciprofloxacin in the presence of reserpine confirmed that ciprofloxacin resistance was affected by the smeD, F, V overexpression. Therefore, the significant role of active efflux in fluoroquinolone resistance of S. maltophilia strains was demonstrated in the present study. These findings are supported by the results from previous studies [24][25][26] . In contrast, Wu et al. demonstrated that smeDEF did not considerably contribute to fluoroquinolone resistance and implication of efflux pumps in resistance to fluoroquinolone might have overestimated 27 .
There were geographical differences of 47-70% in Smqnr frequency 28 and little correlation between the Smqnr alleles and resistance to fluoroquinolones in the S. maltophilia isolates was assumed 29 . Nonetheless, almost half (n = 25, 52.1%) of the 48 ciprofloxacin-resistant S. maltophilia strains (MIC ≥ 2 µg/mL) were found carrying Smqnr genes in the present study. Two remaining strains with intermediate susceptibility to ciprofloxacin (MIC = 1 µg/mL) was Smqnr-negative. Four strains with high level resistance to both ciprofloxacin (MIC = 8, 32 (2 isolates), and 128 µg/mL) and levofloxacin (MIC = 8 µg/mL) showed overexpression for smeDF genes harboring Smqnr9 (n = 2), Smqnr new variant 4 (n = 1), and a Smqnr-negative strain. The frequency of 65.9% Smqnr alleles were also previously reported inS. maltophilia strains from Iran which studied only 44 strains 18 . The most common Smqnr allele in the current study was Smqnr8, followed by Smqnr11, Smqnr9, Smqnr30, and Smqnr35. Therefore, there was no significant association between ciprofloxacin resistance and Smqnr alleles in the strains examined (p = 0.2). However, significant differences of levofloxacin MICs among the Smqnr-positive and Smqnrnegative strains were found (p = 0.008). Significant difference in resistance to levofloxacin of Smqnr-positive isolates was previously demonstrated by Kanamori et al. using a MIC of ≥ 2 µg/mL for levofloxacin but not a MIC of ≥ 8 µg/mL and they highlighted the role of Smqnr genes in low-level fluoroquinolone resistance 28 28 . The role of Smqnr genes remains obscure and high-level fluoroquinolone resistance in S. maltophilia isolates might be associated with mechanisms other than Smqnr as described previously 29,30 .
Comparison of the MICs of ciprofloxacin in the presence of both efflux and Smqnr alleles showed that the higher MICs were noted when overexpression of the two smeDEF and smeVWX efflux pump genes alongside with the Smqnr alleles detected. Totally, the higher MIC levels more related with two parameters; the number of overexpressed genes and the level of expression (higher levels of expression in more genes result in higher MICs). Therefore, according to these findings and comparison with the results of the isolates only having Smqnr alleles or active efflux pumps, overexpression of smeDEF and smeVWX genes were more important in resistance development and can lead to high level fluoroquinolone resistance. High-level fluoroquinolone resistance due to the overexpression of multi-drug efflux pump semDEF and low-level fluoroquinolone resistance by qnrD were already reported by Cavaco et al. and Valdezate et al. 31,32 .
In conclusion, this study revealed that active efflux pumps can significantly contribute in fluoroquinolone resistance in S. maltophilia isolates. No correlation between the Smqnr alleles and ciprofloxacin resistance in the clinical isolates of S. maltophilia was found, but Smqnr alleles were mostly associated with lower MICs of levofloxacin. Therefore, efflux pumps were largely linked to higher MICs of fluoroquinolone than the Smqnr alleles. Further studies are required to assess the contribution of Smqnr to the fluoroquinolone susceptibility of S. maltophilia isolates.

Materials and methods
Bacterial isolates and identification. A total of 385 clinical isolates of S. maltophilia were collected during the period between September 2010 and August 2017 from six hospitals (H1-H4, H6, H11) in Iran. Phenotypic identification of the isolates was done using different biochemical tests including, oxidase, catalase, DNase, nitrate reduction, citrate, esculin hydrolysis, gelatin liquefaction, lysine decarboxylase and sugar fermentation on triple sugar iron (TSI) agar 5 . Genomic DNA was prepared from single colony of each isolate using the standard phenol-chloroform method 33 and species-specific PCR (SS-PCR) using the following primers; SM1 5′-CAG CCT GCG AAA AGTA-3′ and SM4 5′-TTA AGC TTG CCA CGA ACA G-3′ was applied to target the 23S rRNA gene 34  Antimicrobial susceptibility testing. Antibiotic susceptibility testing of the isolates against ciprofloxacin (5 µg), levofloxacin (5 µg) and moxifloxacin (5 µg) (Mast Group Ltd, UK) was determined using the disk diffusion method on the Mueller-Hinton agar (Merck, Germany) plates according to Clinical and Laboratory Standards Institute (CLSI) guidelines 35 . In addition, minimum inhibitory concentrations (MICs) of ciprofloxacin and levofloxacin were determined by broth microdilution method for the isolates which were considered as ciprofloxacin-intermediate/-resistant according to the disk diffusion method. In other words, 50 isolates were selected for this experiment, in which 41 isolates were ciprofloxacin resistant and the remaining nine isolates were among the ciprofloxacin-intermediate-susceptible isolates which were selected based on isolation date, sources, the presence of resistance genes and MICs). In brief, twofold serial dilutions of ciprofloxacin and levofloxacin were prepared in 96-well microplates containing Mueller-Hinton broth (Merck, Germany) to obtain the concentration ranging from 0.25 to 128 μg/mL. The 0.5 MacFarland bacterial suspensions were used and the final concentration was equal to 5 × 10 5 CFU/mL. The plates were sealed and incubated for 20-24 h at 35 °C. The critical breakpoints of ciprofloxacin for Pseudomonas aeruginosa were used for interpretation of the results because of no breakpoints for S. maltophilia were recommended by the CLSI 35  Genomic detection of smeDEF and smeVWX genes. To confirm the presence of the genes encoding efflux pumps including, smeDEF and smeVWX, PCR was done using specific primers for smeD, smeE, smeF, smeV, smeW, and smeX genes ( Table 2). A representative PCR amplicon of each gene was sequenced to ensure the specific amplification.

Quantitative reverse transcription PCR (RT-qPCR).
A single colony of each isolates were cultured in Luria-Bertani (LB) broth (Merck, Germany) and placed in a 37 °C shaking incubator at 180 rpm until the growth reached logarithmic phase (OD600 = 0.5). The log-phase bacterial cell were used to extract total RNA by the RNeasy mini kit (Qiagen, Valencia, CA, USA) according to the manufacturer's instructions. Then, total RNA was treated with RNase free DNase I (Ambion, Austin, TX, U.S.A.) to further eliminate genomic DNA. After that, the quantity and quality of yielded RNA were evaluated using the Nanodrop (Thermo Scientific, Waltham, MA, USA) and RNA integrity verification was done on 1% agarose gel. Finally, the purified RNA was confirmed by PCR using gyrA primers ( Table 3). The StepOnePlus™ real-time PCR System (Applied Biosystems, Foster The RT-qPCR data analysis was carried out using the 2 −∆∆CT method to evaluate expression level of smeD, smeF and smeV genes by normalization to the gyrA housekeeping gene as well as compared to the S. maltophilia ATCC 13637 as a reference strain. Efflux pump expression greater than 3 folds was considered as overexpression 27 .

PCR detection and sequence analysis of Smqnr alleles.
To amplify a 811 bp fragment of Smqnr gene, the following primer set was used; forward primer; 5′-ACA CAG AAC GGC TGG ACT GC-3′ and reverse primer; 5′-TTC AAC GAC GTG GAG CTG T-3′ 29 . PCR was performed using the 10 µL of Pfu PCR PreMix, (Bioneer, Korea), 10 pM of each primer, 50 ng of template DNA and 6 µL of distilled water to reach final volume of 20 µL PCR reaction mix. PCR products were sequenced with the corresponding PCR primers and translated to amino acid sequences using the Expasy translate tool (http:// web. expasy. org/ trans late/). The obtained sequences were compared to the previously deposited Smqnr sequences in GenBank and alleles with one or more amino acid substitution were considered as new variants 6 . Multiple alignments of all available Smqnr sequences in GenBank till October 14, 2020 were performed to analyze the phylogenetic relationships of Smqnr alleles. The phylogenetic tree of Smqnr genes was constructed using Molecular Evolution and Genetic Analysis (MEGA) version 7.0.14 (http:// www. megas oftwa re. net/). Statistical analysis. The SPSS version 23.0 was used to analyze the obtained results. Descriptive statistics of the data was conducted by frequencies and crosstabs. The Pearson Chi-Square test was used to analyze the reduction of ciprofloxacin MICs under reserpine treatment. The effect of the presence or absence of efflux pump genes on antibiotic resistance was evaluated using Kruskal-Wallis test. The correlation of the antibiotic MICs with the efflux pump's expression or Smqnr alleles was evaluated using Pearson Chi-Square test. co-effect of efflux pumps and Smqnr alleles on ciprofloxacin MICs was assessed by Chi-Square test. The MIC 50 (MIC required to inhibit the growth of 50% of organisms) and MIC 90 (MIC required to inhibit the growth of 90% of organisms) of ciprofloxacin and levofloxacin of the strains were calculated. A p value of < 0.05 was considered significant.

Nucleotide accession numbers.
Ethics approval. This study was approved by the Ethics Committee of Tehran University of Medical Sciences "IR. TUMS. MSP. SPH. REC.1396.4388".