Prevalence and molecular characterization of methicillin-resistant Staphylococcus aureus with mupirocin, fusidic acid and/or retapamulin resistance

Background The data on the prevalence of resistance to mupirocin (MUP), fusidic acid (FA) and retapamulin (RET) in methicillin-resistant Staphylococcus aureus (MRSA) from China are still limited. This study aimed to examine these three antibiotics resistance in 1206 MRSA clinical isolates from Eastern China. Phenotypic MUP, FA and RET resistance was determined by minimum inhibitory concentrations (MICs), and genotypic by PCR and DNA sequencing of the mupA/B, fusB-D, cfr, vgaA/Av/ALC/B/C/E, lsaA-C/E and salA and mutations in ileS, fusA/E, rplC, and 23S RNA V domain. The genetic characteristics of resistance isolates were conducted by pulsed field gel electrophoresis (PFGE) and multilocus sequence typing (MLST). Results Overall MRSA MUP, FA and RET resistance was low (5.1, 1.0 and 0.3%, respectively). MupA was the mechanism of high-level MUP resistance. All low-level MUP resistance isolates possessed an equivocal mutation N213D in IleS; of these, 2 reported an additional V588F mutation with an impact on the Rossman fold. FusA mutations, such as L461K, H457Q, H457Y and V90I were the primary FA mechanisms among high-level resistance isolates, most of which also contained fusC; however, all low-level resistance strains carried fusB. Except lsaE gene detected in one isolate, no other resistance mechanisms tested were found among RET-resistant isolates. Additionally, sixteen PFGE types (A-P) were observed, among which type B was the most common (49/76, 64.5%), followed by types E and G (4/76, 5.3% each) and types C and M (3/76, 3.9% each). All resistant strains were divided into 15 ST types by MLST. ST764 (24/76, 31.6%), ST630 (11/76, 14.5%), ST239 (9/76, 11.8%) and ST5 (7/76, 9.2%) were the major types. PFGE type B isolates with the aforementioned STs were mainly found in mupirocin resistant isolates. Conclusions MUP, FA and RET exhibited highly activity against the MRSA isolates. Acquired genes and chromosome-borne genes mutations were responsible for MUP and FA resistance; however, the mechanism for some RET-resistant isolates remains to be further elucidated. Also, the surveillance to MUP in MRSA should be strengthened to prevent elevated resistance due to the expansion of clones.


Background
Methicillin-resistant Staphylococcus aureus (MRSA) is a major pathogen responsible for various hospitalacquired and community-associated infectious worldwide [1]. Because of strong resistance to antibiotics, treatment of MRSA infections is challenging in clinical anti-infective therapy, leading to high risk of mortality and expensive medication [1].
Skin and soft tissue infections (SSTIs) constitute the common diseases caused by S. aureus, including MRSA [2], which can be treated with topical antibiotics mupirocin (MUP) and fusidic acid (FA) [3]. However, unreasonable long-term use of these drugs leads to the emergence of resistance, which is a significant public health concern [3]. Therefore, novel topical antimicrobial agent retapamulin (RET) is developed for the treatment of S. aureus SSTIs [3].
MUP resistance in staphylococcus is divided into two phenotypes: high-level (MuH, minimum inhibitory concentration (MIC) ≥ 512 μg/mL) and low-level (MuL, MIC = 8-256 μg/mL) [3,4]. The MuH is mediated by gene mupA or mupB, and the MuL is related to point mutations in the chromosomal isoleucyl-tRNA synthetase gene (ileS) [3,5]. Previous studies showed that mutations in fusA or fusE in chromosome confer high-level FA resistance (FAH), and acquired fusB-D genes that mediate low-level resistance (FAL) [3,6,7]. RET has been licensed in USA and Europe for the topical treatment of SSTIs caused by methicillin-sensitivity S. aureus and Streptococcus pyogenes [8]. The RET resistance in S. aureus is often mediated by the point mutations of ribosomal protein L3 (encoded by rplC) or the 23S rRNA V domain, or efflux pumps VgaA/Av/A LC /B/C/E, LsaA-C/E and SalA, or methylation of the 23S rRNA subunit (methylated by methyltransferase encoded by chloramphenicol-florfenicol resistance (cfr) gene) [3,9,10].
Several previous studies reported the resistance of MUP and FA in Eastern China [11][12][13][14]. However, since Eastern China is a region with a vast territory, the antibiotic resistance spectrum and the resistance mechanisms may be diversified in different hospitals. Therefore, the data on the resistance of both drugs in clinical isolates of S. aureus are limited. To the best of our knowledge, there is no information on RET resistance in China.
In this study, we determined the prevalence of MUP, FA and RET resistance among MRSA isolates from Shanghai and Zhejiang province in Eastern China, and analyzed the underlying resistance mechanisms. Furthermore, PFGE and MLST analysis were also carried out for the genetic characterization of resistant isolates.

Prevalence of MUP, FA and RET resistance
A total of 1206 MRSA isolates were screened using broth microdilution assay, and 49 MuH, 12 MuL, 6 FAH, 6 FAL, 2 RET-resistant and 1 MuH-RET-resistant isolates were identified. The detailed MIC data of the MUP, FA and RET resistance isolates were listed in Fig. 1. Although 75.1% isolates in this collection were obtained from respiratory samples, the resistance rates of these strains to MUP, FA and RET did not differ significantly as compared to those obtained from wound secretion, the second most common specimen type ( Table 1).

Mechanisms of MUP resistance
A total of 49 MuH and 1 MuH-RET resistant isolates contained the mupA gene, and no isolates were mupB positive ( Table 2). The sequences of Smr, Mrm and Lmr DNA fragments were compared to the known ileS gene of S. aureus (Gene bank accession no. X74219). The findings showed that the 12 MuL isolates possessed N213D mutation in the Smr fragment, and 2 MuL isolates had V588F mutation in the Mrm fragment. No mutations were identified in the Lmr fragment. The mutations were accompanied by different MICs: N213D/ V588F, 8 and 32 μg/mL (1 isolate each); N213D, 8 (1 isolate), 32 (2 isolates), 64 (1 isolate) and 128 μg/mL (6 isolates) ( Fig. 1 and Table 2). There was not significant difference in the distribution of MIC values between isolates with the single mutation N213D and the double mutation N213D/V588F (P = 0.077). In addition, no mupA or mupB gene was detected in MuL isolates.

Mechanisms of FA resistance
To uncover the mechanisms of FA resistance among 6 FAH isolates, the full-length of fusA and fusE genes were sequenced and compared to those of S. aureus ATCC 25923. Herein, we identified 2 isolates that contained H457 missense mutation (H457Q and H457Y, accompanied by MICs of 128 and 16 μg/mL, respectively), while 3 harbored the L461K mutation (MIC = 128 μg/ mL), and 1 (MIC = 64 μg/mL) simultaneously possessed L461K, E8K and V90I mutations in FusA ( Fig. 1 and Table 2). All the FA resistance isolates were evaluated for fusB, fusC and fusD. Among the 6 FAH isolates, 4 carried the fusC gene. Furthermore, all the 6 FAL isolates were fusB-positive, and only one carried fusC gene ( Table 2). No mutations were found in fusE, and no isolates were fusD positive. Mechanism of RET resistance 1/3 isolates with resistance to RET harbored the lsaE gene, and the remaining two displayed negative findings for all the resistance mechanisms tested ( Fig. 1 and Table 2).

PFGE
The 76 isolates with MUP, FA and/or RET resistance were divided into 16 patterns: type A-P ( Fig. 1). Among 49 MuH alone strains, 40 belonged to type B, 4 were type E, and each of the remaining 5 belonged to types A, F, J, K and N, respectively. One MuH-RET resistant isolate was also type B. Among 12 MuL isolates, 9 were type B, and 3 belonged to type A, M and N, respectively. Type C was the most frequent type in FAH strains (3/6; 50%). Type G was the most common pattern in FAL strains (4/6; 66.7%). Two MRSA resistant to RET alone belonged to types O and H, respectively.

Discussion
MUP is effective for the prevention and treatment of MRSA SSTIs. However, the resistance (including MuL) is beneficial for MRSA treatment and eradication failure [15,16]. The prevalence rate of MUP resistance in MRSA clinical isolates varies from 0.5-10.1% for MuH and 2.4-8.6% for MuL in USA, 0-75% for MuH and 0-46.7% for MuL in Asia, and from 0.8-98% for MuH and 0-31.2% for MuL in Europe [15]. In the present study, the isolation rates of MuH and MuL were low: 4.1% (50/ 1206) and 1.0% (12/1206), respectively. Recent studies displayed that the prevalence of MuH is mediated by plasmid-borne mupA gene [15], which is the same as our results. Although mupB, also a plasmid-borne gene, is correlated with MuH [5], the mechanism is rarely examined in staphylococci, including the isolates investigated in this study. The point mutations in the ileS gene, resulting in amino changes in MUP-binding site (located in amino acids 450-650, also named Rossman fold), are the main mechanisms determining MuL [15,17]. V588F and V631F are well identified frequent mutations in IleS responsible for MuL [15]. In this study, only two MuL isolates (PT300 and wu9) contained the V588F mutation, and no MuL isolates harbored the V631F mutation. Notably, all MuL isolates harbored the N213D mutation that was located in a hotspot amino acid sequence between 200 and 350, as described by Lee et al. [17]. The N213D mutation has been previously reported and are considered to have no impact on the sensitivity of MUP [18]. Although the mupA gene located on the chromosome is also associated with MuL [3], we did not detect the gene in our MuL isolates. Also, no other mutations in IleS were found. Lee et al. [17] reported that a mutation of S634F that confers phenotype of susceptibility or MuL in diverse isolates. In view of the above reasons, the contribution of N213D mutation to MuL should be evaluated further.
FA is a steroidal antimicrobial agent that suppresses the production of bacterial proteins by stopping the dissociation of elongation factor G (EF-G) from ribosome [6,19]. Clinically, the main application of topical FA is for the treatment of SSTIs and decolonization of S. aureus, including MRSA; this method is similar to that of MUP [3]. The prevalence of FA resistance reported by recent large studies varies in MRSA isolates from USA (0-0.3%), Australian (4.1-5.1%), Denmark (17.8%), Greece (57.0%) and other European countries (9.9%) [3]. In China, the resistance levels in MRSA are also different in different areas, for example, 3.0-5.3% in MRSA from Beijing, Shanghai, Shenyang and Shenzhen cities [12,20], and 27.1% in MRSA from Wenzhou city [11]. Compared to the aforementioned data from China, our results showed a very low resistance rate (12/1206, 1.0%).
In S. aureus, the mutations in fusA (encoding EF-G) or fusE (coding for ribosome protein L6, RplF) lead to a decreased affinity of FA for the EF-G ribosome complex [3,21]. About > 30 point mutations in FusA sequence were described; however, only a few were experimentally verified to play a role in FA resistance [3,22,23]. The V90I mutation in domain I (amino acids 1-280) and H457Y, H457Q and L461K   mutations in domain III (amino acids 404-480) were observed in this study, and have been previously identified causing FAH in S. aureus [21,22,24]; also, L461K is the most prevalent mechanism among clinical FAH S. aureus strains [3]. L461K also existed in the majority (4/6, 80%) FAH isolates, leading to FA MICs ≥64 μg/mL in the current study. One novel substitution with E8K in domain I was identified in addition to V90I and L461K mutations in the FAH isolate. However, whether the novel mutation is associated with FA resistance is not yet clarified.
Protection of EF-G by FusB family molecules is another mechanism conferring the resistance (low-level) of FA [3]. FusB family proteins (including FusB, FusC, and FusD) can restore the translation of protein by binding to EF-G in the presence of FA [3]. Previous studies showed that fusB was the most prevalent in Netherlands and mainland China [11,12,25], and fusC primarily existed in isolates from Taiwan, Australia, USA and European [26][27][28]. In our isolates with FA resistance, fusB existed in all FAL isolates (6/6), and fusC was most prevalent in FAH isolates (4/6, 66.7%). The fusD gene was identified in Staphylococcus saprophyticus. The gene is related to the "intrinsic resistance of FA" among this species [19]. Hitherto, this determinant is rarely detected in S. aureus strains.
RET is a semisynthetic pleuromutilin drug that represses the synthesis of bacterial proteins by interacting with domain V of 50S ribosomal subunit [3]. This drug has a potency to act as an alternative to MUP to eradicate the S. aureus colonization, except when used for the treatment of S. aureus-induced SSTIs [8,29]. Currently, limited data are available among clinical S. aureus strains worldwide. The resistance rates of 664 UK S. aureus (74% are MRSA), 155 USA MRSA, 403 USA MRSA, and 400 USA S. aureus from several studies were 0.15, 2.6, 0.25, and 9.5%, respectively [8,[29][30][31]. In this study, the prevalence of RET resistance was very low (3/1206, 0.24%). In the UK and USA, the RET resistance among S. aureus or MRSA with MUP resistance was < 1-2.6% [8,31]. In the present study, only one MRSA isolate was observed to have resistance to both RET and MUP. For RET resistance in staphylococci, except mutation and/or methylation of ribosomal protein and rRNA, the ABC-F proteins (efflux pumps) encoded by three types of genes (vga, lsa and sal) were alsocritical mechanisms [32]. However, in our 3 RET-resistant isolates, only one was confirmed to contain the salE gene. The salE gene confers the resistance to pleuromutilins, lincosamides and streptogramin A [32]. Fortunately, the gene was localized on the chromosome or non-conjugative plasmid (pV7073) [32], which limits its transmission among staphylococci. For two strains we could not clarify phenotypic resistance to the detected resistance determinants, implying that there possibly exists other gene or gene variant leading to RET resistance.
ST239 and ST5 are two predominant sequence types in China. However, the strains identified in this study mainly belonged to ST764 (31.6%), which was more than the total percentage of ST239 and ST5 (16/76, 21.1%). Early investigation showed that most of FA-resistant MRSA belong to ST239 and ST5 [7,27]. However, our findings showed no FA-resistant strains were these two types. ST764 MRSA, first reported in Japan, is a single-locus of ST5 nosocomial MRSA clone with or without the arginine catabolic mobile element (ACME, a feature of CA-MRSA) [33,34]. In recent years, several studies reported the S. aureus clone with ST764 in China [13,35]. In this study, ST630 (14.5%) was the second most common type, which also became a prevalent S. aureus clone causing SSTIs in Wenzhou region, Zhejiang province, China [36]. Figure 1 shows that 66.7% (4/6) FAL isolates belonged to ST630 (PFGE G-ST630), which was similar to the previous report [9]. Notably, 50% (6/12) MUL isolates and 39.0% (16/41) MUH isolates with MIC ≥1024 μg/mL were PFGE B-ST764, and were identified from the same hospital (Shanghai General Hospital) (Fig. 1), indicating clone transmission. In addition, the spread of other MRSA clones with different genetic patterns, such as PFGE B-ST1821, PFGE B-ST239 (4 isolates each), PFGE B-ST5 and PFGE B-ST630 (3 isolates each) were also responsible for the increased MUH in this hospital. It should be noted that some resistant isolates had the same PFGE pattern (such as PFGE B, C, E, G or M), although the ST types were diverse from each other (Fig. 1). These finding hint that these isolates might be clonally related.

Conclusions
In this study, the MRSA isolates exhibited a low prevalence of resistance to MUP, FA and RET, especially to the latter two, and cross-resistance to the three antibiotics was rare. The mupA gene mechanism mediated MuH. The contribution of the N213D mutation in IleS found in our MuL isolates that decreased the resistance of MUP is yet unclear. FusA mutations, FusB and FusC were the frequent genetic mechanisms that mediate FA resistance. Phylogenetic detection showed the transmission of multiple clones, especially PFGE B-ST764 clone that made a major contribution to the increased MUP resistance in this collection of isolates. Owing to the concern of resistance development and clonal dissemination in healthcare, continuous surveillance for the resistance of these topic antibiotics in S. aureus is essential in China.

Bacterial isolates
A total of 1206 non-duplicate MRSA isolates from various clinical specimens were collected from 8 hospitals in Table 3 Primers and PCR conditions for detecting mupirocin, fusidic acid and retapamulin resistance genes in this study  PCR was used to detect the factors mediating MUP (mupA, mupB and ileS (amplifying three fragments of Smr, Mrm and Lmr, which might possess the mutations mediating mupirocin resistance)), FA (fusA, fusB, fusC, fusD and fusE) and RET (rplC, cfr, vgaA/Av/A LC /B/C/E, lsaA-C/E, salA and 23S RNA V) resistance [5, 7-10, 26, 29, 38-41]. The primers and programs for the amplification of genes are presented in Table 3. The DNA sequencing of one randomly selected PCR product for mupA, mupB, fusB-D, cfr, vgaA/Av/A LC /B/C/E, lsaA-C/ E, and salA was used for the identification of target fragments. All the PCR products for ileS-Smr, ileS-Mrm, ileS-Lmr, fusA, fusE, rplC and 23S RNA V were sequenced to determine the putative mutations.

Pulsed field gel electrophoresis (PFGE)
PFGE was performed for MUP, FA and/or RET resistant strains, as described previously [42]. BioNumerics software 7.0 was used for the analysis of DNA fingerprint profiles. An 80% cutoff value was set to assess the similarity.

Statistical analysis
The difference of MUP, FA and RET-resistant rates of isolates from major samples was analyzed by Pearson's Chi-square test or Fisher's exact test using SAS 8.0 (SAS Institute, Cary, NC, USA). Student's t test was used to determine the difference in the distribution of MIC values between isolates with different mutations. A P value < 0.05 (two-tailed) indicated significance.