Genomic alterations involved in fluoroquinolone resistance development in Staphylococcus aureus

Aim Fluoroquinolone (FQ) is a potent antibiotic class. However, resistance to this class emerges quickly which hinders its application. In this study, mechanisms leading to the emergence of multidrug-resistant (MDR) Staphylococcus aureus (S. aureus) strains under FQ exposure were investigated. Methodology S. aureus ATCC 29213 was serially exposed to ciprofloxacin (CIP), ofloxacin (OFL), or levofloxacin (LEV) at sub-minimum inhibitory concentrations (sub-MICs) for 12 days to obtain S. aureus -1 strains and antibiotic-free cultured for another 10 days to obtain S. aureus-2 strains. The whole genome (WGS) and target sequencing were applied to analyze genomic alterations; and RT-qPCR was used to access the expressions of efflux-related genes, alternative sigma factors, and genes involved in FQ resistance. Results A strong and irreversible increase of MICs was observed in all applied FQs (32 to 128 times) in all S. aureus-1 and remained 16 to 32 times in all S. aureus-2. WGS indicated 10 noticeable mutations occurring in all FQ-exposed S. aureus including 2 insdel mutations in SACOL0573 and rimI; a synonymous mutation in hslO; and 7 missense mutations located in an untranslated region. GrlA, was found mutated (R570H) in all S. aureus-1 and -2. Genes encoding for efflux pumps and their regulator (norA, norB, norC, and mgrA); alternative sigma factors (sigB and sigS); acetyltransferase (rimI); methicillin resistance (fmtB); and hypothetical protein BJI72_0645 were overexpressed in FQ-exposed strains. Conclusion The emergence of MDR S. aureus was associated with the mutations in the FQ-target sequences and the overexpression of efflux pump systems and their regulators.


Introduction
In addition, FQ resistance is also mediated by the overexpression of endogenous efflux systems that pump drugs out of the cell could significantly reduce the accumulation of drugs inside the bacterial cells [34,35]. Among chromosomal efflux proteins in S. aureus, NorA, NorB, and NorC which are encoded by norA, norB, and norC respectively, played an important role in decreased susceptibility to antibiotics especially FQs [36][37][38]. These proteins are belonged to the major facilitator superfamily (MFS) and negatively regulated by MgrA [37].
In this study, the in vitro-induced FQ resistance in the fully susceptible S. aureus ATCC 29213 was generated for investigating the mechanism associated with intrinsic FQ resistance development. Whole genome sequencing of in vitro-induced FQ-resistant S. aureus strains as well as target sequencing of important genes were analyzed to associate the mutations in the target enzymes with the obtained resistant phenotypes. Additionally, the expression of effluxrelated genes, alternative sigma factors, and genes involved in FQ resistance were evaluated using RT-qPCR.

Selection of FQ-resistant strains
S. aureus ATCC 29213 (initial S. aureus), a fully susceptible strain, was cultured in Mueller-Hinton Broth (MHB) that contained sub-minimum inhibitory concentrations (sub-MICs) values of FQ which is either ciprofloxacin (CIP), ofloxacin (OFL) or levofloxacin (LEV) following a previously published procedure [39]. The experiment was repeated until no increase in MIC to the FQ used in exposure was observed. At the endpoint, the obtained CIP-, LEV-, and OFLexposed S. aureus (S. aureus-1) strains were sent to NK Biotechnology Company (HCM, Vietnam) for 16S rRNA sequencing. These selected resistant strains were sub-cultured for another 10 days in an antibiotic-free medium with the daily examination of MIC in order to obtain CIP-, LEV-, and OFL-revertant S. aureus (S. aureus-2) strains. Repetitive sequence-based PCR (Rep-PCR) amplification was established to confirm the genetic relatedness of initial S. aureus ATCC 29213 and those of selected strains. The test was adapted from Vito et al. [40]. The strains at day 12 th of FQ exposure (S. aureus-1 including CIP-, OFL-and LEV-1) and the strains at day 10 th of antibiotic-free culture (S. aureus-2 including CIP-, OFL-and LEV-2) were then used for other experiments.

Antimicrobial susceptibility testing
The micro-dilution method on 96-well plates, instructed by EUCAST guidelines (Eucast.org, version 11.0), was applied on seven bacterial strains (initial S. aureus, S. aureus-1 and -2) to determine the MIC value of the strains to CIP, LEV, OFL, moxifloxacin, nalidixic acid, ampicillin, amoxicillin, chloramphenicol, cefalexin, doxycycline, erythromycin, lincomycin, oxacillin, and tetracycline. At the same time, MICs of the strains were also measured under the presence of an efflux pump inhibitor, reserpine (Sigma-Aldrich, USA), at a concentration of 20 mg/L.

Whole genome sequencing
DNA extraction was carried out using the GeneJET Genomic DNA Purification Kit (Thermo Scientific, USA) according to the manufacturer's instructions. Sequencing libraries were prepared using the TruSeq Nano DNA LT Library Prep Kit (Illumina, Singapore). Pooled libraries were then sequenced using an Illumina NextSeq 500 sequencer with 2 x 151 bp reads. The resulting FASTQ files were mapped to the ATCC 29213 genome (Genbank GCF_001879295.1) using bwa (version 0.7.10) (PMID 19451168); indel realignment and SNP (single nucleotide polymorphism) calling were performed using Lofreq* (version 2.1.2) with default parameters (PMID 23066108).

Expression evaluation of efflux-related genes, alternative sigma factors, and genes involved in FQ resistance
Total RNA extraction and cDNA Synthesis of seven S. aureus strains were performed using Monarch 1 Total RNA Miniprep Kit (NEB, UK) and SensiFAST cDNA Synthesis Kit (Bioline, UK), according to the manufacturer's instructions. Each real-time qRT-PCR reaction consisted of 20 μL of reagents, including Luna 1 Universal qPCR Master Mix (NEB, UK), one primer pair (Table 1; Reference gene: 16S rRNA), cDNA template, and nuclease-free water. The thermal cycle of the RT-qPCR machine was set up based on NEB #M3003 instruction manual (NEB, UK). Transcription values (Ct) are analyzed as described in [42].

Statistical analysis
The RT-qPCR experiments were performed in triplicates. After obtaining the transcriptional values (Ct) from amplicon-based fluorescence thresholds, the Ct values of the target genes were normalized to that of the 16S rRNA transcripts to obtain a ΔCt. The 2 -ΔΔCt method was then used to compare the relative expression patterns between FQ-exposed strains and the initial S. aureus [42]. Fold change was calculated against initial S. aureus and visualized on a log scale, with gene expression of initial S. aureus as 1. Values greater than one were considered up-regulated, while values smaller than 1 were down-regulated. Fold change and confidence level 95% CI (error bar) were calculated in MS Excel according to the standard practice [43]. Expression data of three biological replicates were analyzed by one-tail Student's t-test to identify the statistical significance of differential expression between FQ-exposed S. aureus strains

Serial exposure to FQs leading to FQ-resistant phenotype
After being serially exposed to FQs, S. aureus ATCC 29213 turned from FQ-sensitive to FQresistant phenotype. The MIC values of the FQ-exposed strains reached peaks at day 8 th -10 th of exposure (Fig 1), and no further increase in MIC values was observed after the peaks even if these strains were continuously exposed to the antibiotics. After 12-day serial exposure to FQs, the MIC values of obtained S. aureus-1 were 128 times higher than that of the initial strain for CIP, and 32 times higher for OFL and LEV. Impressively, S. aureus-2 strains did not revert to antibiotic-sensitive phenotype after being cultured in an antibiotic-free medium for 10 days (remained 8-32 times higher than the initial strain). Gram staining, 16S rRNA sequencing, and Rep-PCR fingerprint results proved no contamination during the antibiotic exposure process (S1 Fig). In addition, molecular genotyping results showed that S. aureus-1 and -2 strains were identical to the initial S. aureus ATCC 29213.

Susceptibility of S. aureus strains to different antibiotics
The FQ-exposed S. aureus ATCC 29213 turned into a multidrug-resistant phenotype, which was not only resistant to other FQs but also to other antibiotics of unrelated groups ( Under the presence of reserpine, S. aureus ATCC 29213 was not affected while susceptibility of S. aureus-1 strains was reduced by 2-8 folds for some antibiotics such as ampicillin, amoxicillin, doxycycline, erythromycin, and lincomycin (S1 Table).

Whole genome sequencing of S. aureus ATCC 29213, S. aureus-1 and S. aureus-2 strains
Whole genome sequencing was used to identify mutations associated with the changes in resistance (S2 Table). Overall, a total of 42 positions where at least one strain carried a variant relative to the initial strain were discovered. Of these 42 variant positions, 10 were common in all strains sequenced, indicating differences in our parental strain from the publicly available genome sequence. Among 32 remaining variant positions, 24 were SNPs, 6 were deletions, and 2 were insertions. A total of 11 variants were seen in all 6 of the S. aureus-1 and S. aureus-2 strains relative to our re-sequenced ATCC 29213 (Table 3). Of these 11 SNPs, 4 were found in predicted protein-coding genes: an amino acid deletion in SACOL0573 (encoding a PIN domain-containing protein); an amino acid insertion in rimI (an alanine acetyltransferase); a missense mutation in grlA, encoding DNA topoisomerase IV subunit A, and a synonymous mutation in hslO, which encodes heat shock protein Hsp33. The remaining 7 variants located in the upstream region of norA and gene coding hypothetical protein BJI72_0645 were not found in annotated protein-coding sequences. The initial S. aureus was exposed to CIP, OFL, or LEV for 12 days to obtain FQresistant S. aureus strains (CIP-1, OFL-1, and LEV-1). These exposed S. aureus strains were then continuously subcultured for another 10 days in an antibiotic-free medium to obtain CIP-2, OFL-2, and LEV-2 with the daily examination of MICs values evaluated. Day 0: MIC; day 1-12: 12 FQ exposed days; R1-R10: 10 days in antibiotic-free medium. https://doi.org/10.1371/journal.pone.0287973.g001

Mutations observed in the Quinolone-resistance determining regions (QRDRs)
Target sequencing of QRDRs indicated that there were some mutations in both grlA (S80F) and gyrB (T451S and/or R450S) (Table 4), among those, mutations of both genes were found in CIP-and OFL-1 while only one mutation in gyrB (T451S) was found in LEV-1. Additional sequencing of the S. aureus strains at days 4 th , 6 th , 8 th , 10 th of CIP-, LEV-and OFL-exposure revealed the grlA mutation (S80F) to appeared in earlier steps than the ones in gyrB, suggesting the primary role of grlA mutation in FQ resistant phenotype ( Table 4).

Overexpression of alternative sigma factors in FQ-exposed strains
The data in Fig 2 illustrates that there were upregulations of sigB and sigS genes in most FQexposed strains, except for OFL-2 and LEV-2. The expression of sigB and sigS genes were recorded highest in LEV-1, at a mean fold change expression of 2.39 and 3.80 respectively, while the most significant downregulation of both genes was recorded in OFL-2 which were about 0.41 and 0.55 fold decrease. In terms of the S. aureus-2 group, the upregulation of both genes in CIP-2 made this the only strain in the group that experienced a rise in gene expression of both alternative sigma factors, σ B and σ S in comparison with initial S. aureus. In the statistical analysis between two groups of S. aureus-1 and S. aureus-2 for the same antibiotic, it can be seen that there was a decrease in gene expression in both sigB and sigS genes of S. aureus-2 strains compared with S. aureus-1 strains.

Efflux expression increased in FQ-exposed strains
The effect of FQ exposure on the expression level of the efflux pump genes was checked and compared to the initial S. aureus. As shown in Fig 3, all S. aureus-1 and 2 strains increased the expression of mgrA, norA, norB, and norC. Regarding to the expression level of mgrA, which is involved in the efflux pump regulation of S. aureus, LEV-1 exhibited the highest expression level (about 12.63 folds change), followed by CIP-1 and LEV-2 ( Fig 3A). Besides, norA was also upregulated in all FQ-exposed strains, especially in CIP-1 and CIP-2 of which the expression increased by about 57.61 and 70.78 folds respectively compared to initial strains ( Fig 3B). Among S. aureus-1 strains, OFL-1 showed the lowest expression in mgrA, norA, norB, and norC. A similar result was also found in OFL-2 when compared with the remaining S. aureus-2 strains. In addition, almost all strains belonging to S. aureus-1 group had higher expression in mgrA, norA, norB, and norC compared to the strains which were treated with the same antibiotic in S. aureus-2 group.

Overexpression of genes related to FQ-resistant development in S. aureus
According to the WGS results, there were genomic mutations located in the sequence of rimI and fmtB in both S. aureus-1 and -2 strains. Besides, 7 SNPs were also found in all FQ-exposed strains which were located in the upstream region of norA and the gene encoding hypothetical protein BJI72_0645 (Table 3). The results suggested that these variants might be related to FQresistant development in S. aureus. Therefore, investigating the expression of rimI, fmtB, and the gene encoding hypothetical protein BJI72_0645 should be carried out to better understand the relationship between these mutations and the FQ-resistant development in S. aureus. Under the effects of FQ exposure, rimI was overexpressed in all FQ-exposed strains, especially the expression increased by 30.43 folds in LEV-1 (Fig 4A). The data also indicated that fmtB was upregulated in fmtB-mutant strains (Fig 4B). Besides, although there was no fmtB mutation in CIP-1 and CIP-2 which were resistant to ciprofloxacin, fmtB was also overexpressed in both strains. Regarding the expression level of the gene encoding hypothetical protein BJI72_0645, this gene was also overexpressed in all FQ-exposed strains, especially LEV-1  ( Fig 4C). In addition, this study also found that the expression of rimI, fmtB, and the gene encoding hypothetical protein BJI72_0645 in S. aureus-1 strains was higher than that in corresponding S. aureus-2 strains in all cases.

Sub-MIC exposure to FQ altered the antibiotic susceptibility profile of S. aureus
After 12 days of exposure to the sub-MICs of FQs, S. aureus increased its MIC values and became resistant to exposed antibiotics. Serial exposure to sub-MICs can create positive selection pressure that drove the development of resistant phenotypes, which is in agreement with previous literature [48][49][50]. Although S. aureus had a decrease in MIC value after 10 days of culturing in an antibiotic-free medium, indicating some reversion of resistance trait, S. aureus-2 strains kept their resistance to FQs. Besides, after being serial exposed to CIP, OFL, and LEV, S. aureus became resistant not only to the exposed antibiotics and other FQs but also to RT-qPCR analysis of sigB and sigS in S. aureus ATCC 29213 and its exposed strains. * indicated a significant difference in gene expression between initial S. aureus and FQ-exposed S. aureus strains. https://doi.org/10.1371/journal.pone.0287973.g002

PLOS ONE
unrelated antibiotics such as ampicillin, amoxicillin, doxycycline, erythromycin, and lincomycin. The results suggested that serial FQ exposure resulted in cross-resistance to unrelated antibiotics and the emergence of multidrug-resistant phenotypes in S. aureus. It should be well noted that this cross-resistance development under CIP exposure was well observed in Grampositive bacteria such as Actinobacillus pleuropneumoniae [51], S. aureus [52], Mycobacterium tuberculosis [53]; but not in Gram-negative, ones like Salmonella enterica [54]; Pseudomonas aeruginosa [55].

Genomic alterations involved in FQ-resistant development in S. aureus
It has been shown that the acquisition of FQ resistance is mainly due to the mutations in target enzymes, topoisomerase IV (GrlA/B) as well as DNA gyrase (GyrA/B) [56]. In S. aureus, the mutations occurred more frequently in QRDRs of grlA (S80F or Y, E84K, and A116E or P) and gyrA (S84L or A, S85P, and E88K) which were described as the primary FQ resistance mechanism [56]. Other studies have found that the mutations in QRDRs of gyrB including D437N, R458E, D432N, and N470D also contributed to FQ resistance [57,58]. In our in vitro-induced FQ-resistant model, we found mutations in both grlA (S80F) and gyrB (T451S and/or R450S), among those, mutations of both genes were found in CIP-and OFL-1 while only one mutation in gyrB (T451S) was found in LEV-1. Additional sequencing of the S. aureus strains at days 4 th , 6 th , 8 th , 10 th of CIP-, LEV-and OFL-exposure revealed the grlA mutation (S80F) to appeared in earlier steps than the ones in gyrB, suggesting the primary role of grlA mutation in FQ resistant phenotype. These findings suggested that the mutation in grlA (S80F) primarily induced the emergence of FQ resistance and might support the mutations in gyrB (T451S); which was required for high-level FQ resistance in S. aureus.
In addition, we observed that mutations in both grlA and gyrB occurred in all S. aureus-1 strains. These mutations resulted in amino acid alterations in both DNA gyrase and topoisomerase IV, which were FQ targets. Consequently, that led to the emergence of FQ resistance in S. aureus-1 due to the alteration of binding sites between FQ and target enzymes. However, interestingly, all mutations in grlA disappeared in S. aureus-2 without a significant MIC decrease in LEV-2 and only a four-time reduction in CIP-2. With OFL-2, it is impressive that no mutation in gyrA, gyrB, grlA, and grlB QRDRs was retained while it kept most of its resistant ability. These data suggested that mutations in FQ targets are important in the presence of FQs but they may not be essential mechanisms to mediate FQ and multidrug-resistant phenotype. One previous study on Enterococcus faecium also showed no mutation detected in QRDRs of gyrA, gyrB, parC, and parE genes in the strain selected from 2-step FQ-resistant mutant induction [59], suggesting the importance of other mechanisms.

Changes in expression of alternative sigma factors associated with the adaptive response of S. aureus under serial FQ exposure
Alternative sigma factors (σ) are an essential component of core RNA polymerase (RNAP) that helps determine promoter selectivity [60]. The association of appropriate alternative sigma factors with core RNAP can help bacteria perform specialized cellular functions or stress-adaptive  of rimI (A), fmtB (B), and the gene encoding hypothetical protein BJI72_0645 (C) in S. aureus ATCC 29213 and its exposed strains. * indicated a significant difference in gene expression between initial S. aureus and FQ-exposed S. aureus strains.
https://doi.org/10.1371/journal.pone.0287973.g004 responses through redirection of transcription initiation [61]. The stress-response alternative sigma factors σ B and σ S contribute to bacterial survival under multiple stresses. σ B affects cellular processes including oxidative stress resistance, drug resistance, stress adaptation, and biofilm formation [62]. Conversely, σ S is activated in response to stresses such as stress resistance, DNA and cell wall damage, cell morphology alterations, metabolism, virulence, and lysis [63].
In this study, the transcriptional analysis of sigB and sigS genes showed that both genes increased their expression in S. aureus-1 strains in responding to FQ exposure but reduced in the S. aureus-2 strains when FQs were withdrawn. It is understandable as sigma factors reacted quickly to extracytoplasmic stresses in S. aureus in which under standard conditions, sigma factor transcription remains low but can be upregulated in response to external stimuli, especially chemicals leading to DNA damage and cell wall disruption such as antibiotics [47]. In addition, sigma factors enable bacteria to rapidly detect antimicrobial signals, activate resistance genes, and repair mechanisms specific to the antimicrobials [64]. In our case, both sigB and sigS seemed to play an important role in S. aureus fitness and adaptation to FQs. It could be the reason for the marked change in proteomic profiles of the bacteria in responding to FQs [65].

Sub-MIC exposure to FQs led to overexpression of the efflux pump and its regulator
Several specific efflux pumps have been associated with antibiotic resistance in clinical isolates of S. aureus. With respect to FQ, norA, norB, and norC are the most important efflux pumps localized on the cytoplasmic membrane of S. aureus [45]. These pumps can extrude many chemically and structurally dissimilar compounds, namely hydrophilic and hydrophilic FQs (such as norfloxacin, ciprofloxacin, moxifloxacin, and garenoxacin), dyes (like ethidium bromide and rhodamine) and biocides (tetraphenylphosphonium and cetrimide) [37,66]. It has been shown that mgrA functions as a positive regulator of norB and a negative regulator of norA and norC [67].
Regarding the expression level of mgrA, the results in this study were consistent with our previous study on proteomic analysis [65]. In order to test the effects of overexpression of mgrA on multidrug efflux pump norA, norB, and norC, RT-qPCR was applied for determining their expression. In analysis, the expression of norA, norB, and norC was increased in all tested strains. It means that efflux pumps were activated and could be one of the mechanisms leading to the occurrence of multidrug-resistant S. aureus strains. Our results are in agreement with the previous study, in which overexpression of MgrA has led to a change in the expression level of norA and norB [66]. However, the increased expression of norC in this study was not associated with the regulation of mgrA, and it might result from mutational alterations in uncharacterized loci that affect the expression of this gene.
The decrease in efflux expression in S. aureus-2 compared to those of S. aureus-1 suggests that under the FQ stressor in the environment, the efflux activity of S. aureus was promoted as an adaptive response to resist the antibiotics. However, the efflux activity could decrease when S. aureus-1 was cultured continuously in antibiotic-free media even though it still kept its resistance to FQ. This is consistent with the result that S. aureus turned to an FQ-resistant phenotype after 12-day serial exposure to FQ, and the MIC values witnessed a slight decrease after 10-day continuous culturing S. aureus-1 in antibiotic-free media.

Sub-MIC exposure to FQs affected the protein acetylation and multi-drug resistant phenotype of S. aureus
Protein acetylation is one of the major post-translational modifications (PTMs) which play an important role in cell signaling and occurs when the cell encounters specific environmental stress conditions [68][69][70]. RimI encoded by rimI is a ribosomal protein S18 acetyltransferase which catalyzes the acetylation of the N-terminal alanine of ribosomal protein S18; and also acts as an N-epsilon-lysine acetyltransferase that catalyzes acetylation of several proteins [71,72]. In this study, an insertion mutation (S33_S34insC) in rimI was found in all S. aureus-1 and 2 that might affect the RimI activities and the protein acetylation in S. aureus. In terms of expression analysis, rimI was overexpressed in these mutant strains compared to the initial S. aureus. These results suggested that the mutation in rimI, as well as its overexpression in FQexposed S. aureus, was an adaptive response under antibiotic stressors. This might be associated with acetylation which alters mRNA translation efficiency in several proteins in S. aureus to maintain its survival in an antibiotic-stress environment.
The fmtB gene codes for FmtB, a~263 kDa cell wall-anchored protein [73]. Although there is limited information about the function of fmtB, it has been proven to contribute to the methicillin-resistant phenotype in S. aureus [74,75]. According to the antibiotic susceptibility profile of seven S. aureus strains, after being exposed to FQ, S. aureus enhanced its resistance to FQ and other antibiotics including B-lactam. Besides, the mutation K1283delinsRG (CTT26104C) in fmtB sequence was also found in OFL-and LEV-exposed strains. For RT-qPCR analysis, fmtB was upregulated in all mutant strains. Although there was no mutation on fmtB found in CIP-1 and CIP-2, which were resistant to ciprofloxacin, fmtB was also overexpressed in both strains. It was suggested that the in vitro-induced FQ resistance cooperating with the acetylation modification somehow affects the fmtB expression and the antibiotic resistance of S. aureus to FQs and other antibiotics of unrelated groups leading to a multi-drug resistance phenotype.
According to whole genome analysis, 7 SNPs were found in the upstream region of both norA and the gene encoding hypothetical protein BJI72_0645. The RT-qPCR results indicated that the variants might play a crucial role in regulating the transcription of these genes because both were overexpressed in all FQ-exposed S. aureus.

From genetic alterations to protein expression
The development of antibiotic resistance might involve not only transcriptional regulation and genetic modifications but also translational regulation. In the proteomic study of our research group, under the FQ stressor in the environment, there were 147 unique proteins in S. aureus-1 and S. aureus-2 which changed their expression in comparison to the initial strain [65]. Regarding the molecular function of differently expressed proteins, 93 proteins were responsible for binding, 89 proteins for catalytic activity, 28 proteins for the structural constituent of ribosome, and 10 other proteins were involved in transcription factor activity (6), transporter activity (2) and antioxidant activity (2). Proteins involved in SOS/stress response and antibiotic resistance and pathogenesis were upregulated upon FQ exposure. Among them, RecA and MgrA are global regulators which are strongly implicated in antibiotic resistance development [76,77]. The presence of RecA leads to the upregulation of SOS genes which involve in DNA repair and promote the autoproteolysis cleavage of LexA (an SOS gene repressor) [78,79]; while MgrA affects multiple S. aureus genes which encoded proteins involving multidrug resistance, autolytic activity, and virulence [80,81]. The fluctuation in protein expression depends on many factors including extracellular agents such as harsh temperatures, chemical and antibiotic stress in the environment, and intracellular agents such as modifications in the genetic materials. Previous studies have proven that the elements belonging to genetic code including amino acid [82], untranslated regions [83][84][85][86], length [86], GC content [87][88][89], and mRNA secondary structure [90] affected the regulation of protein expression. However, under the serial exposure of S. aureus to FQs, the direct relationship between protein expression and genetic modifications found in this study was not proved. They only suggested that resistance to multiple drugs including FQs under drug exposure generally requires the contribution of multiple proteins and processes. In short, the genetic mutations found in FQ-exposed S. aureus and overexpression of global regulators may be the key to the expression change of a range of proteins which assist this bacterium adapt the antibiotic-containing environments.

Conclusion
Exposure to sub-MICs of FQs provided positive selection pressure for the development of resistance traits leading to alterations to the antibiotic susceptibility profile and transcription of the pathogen. These alterations were reflected in the DNA mutations and the change in mRNA expression levels of efflux pumps, especially alternative sigma factors. Multidrug resistance or antibiotic-resistant development is a complicated and multifactorial process. The found mutations might be spontaneous, but their combination is clearly a cause of the resistance status observed in the bacterium.
FQs presently play a vital role in saving lives and are employed in treating a broad spectrum of infectious diseases or even cancers. Nonetheless, the extensive utilization of FQs in human and animal healthcare has led to a growing number of antibiotic-resistant pathogens. The ease of S. aureus to develop FQ resistance emphazised the prudent and appropriate use of antibiotics in preventing the selection of resistant bacteria. Additionally, molecules with ability to interfere the resistance development would be highly recommended in FQ combination therapy.  Table. The statistical analysis of fold change in gene expression between initial S. aureus strain and FQ-exposed S. aureus strains (mean ± standard error).