BMC Infectious Diseases BioMed Central Research article

Background: Mupirocin is a topical antimicrobial agent which is used for the treatment of skin and postoperative wound infections, and the prevention of nasal carriage of methicillin-resistant Staphylococcus aureus (MRSA). However, the prevalence of mupirocin resistance in S. aureus, particularly in MRSA, has increased with the extensive and widespread use of this agent in hospital settings. This study characterized low-and high-level mupirocin-resistant S. aureus isolates obtained from Nigeria and South Africa.


Background
The treatment of infections caused by antibiotic-resistant bacteria especially methicillin-resistant Staphylococcus aureus (MRSA) has become a worldwide problem in hospital and community settings. Infection control programmes implement measures to contain the dissemination of MRSA which include efforts to eradicate carriage of S. aureus [1]. The antibiotic mupirocin used topically has been shown to possess potent activity against staphylococci and is used for the treatment of skin and postoperative wound infections, and the prevention of nasal carriage of MRSA [2,3]. However the widespread use of mupirocin led to resistance in S. aureus, which has been reported worldwide [4][5][6][7][8][9].
Mupirocin-resistant strains are divided into two groups: low-and high-level resistance (MIC 8-256 and >256 mg/ L, respectively) [2]. In most cases, low-level resistance to mupirocin is related to alterations in the host isoleucyl-tRNA synthetase (IRS) [10,11]. Until recently, chromosomal mupirocin resistance was considered clinically unimportant [2,12]. However, low-level mupirocin resistance appears to be more prevalent in clinical isolates than high-level resistance [13][14][15], and the emergence of lowlevel mupirocin resistance has been shown to increase failure rates for nasal decolonization of MRSA [16][17][18]. High-level mupirocin resistant strains cannot be eradicated with mupirocin and constitute a serious clinical problem, especially when they are resistant to methicillin [19]. The clinical isolates exhibiting high-level resistance to mupirocin contain two distinct IRS enzymes: endogenous IRS plus an additional IRS encoded by the mupA gene [20], which is carried on plasmids that vary in size, restriction patterns and their ability to be transferred in conjugation experiments [4,[20][21][22][23]. The mupA gene has also been reported in the genomic DNA of a few S. aureus isolates expressing low-level resistance, suggesting that the mupA gene may be located in the chromosome [24,25]. Moreover, the chromosomal location of the mupA gene in S. aureus expressing high-level mupirocin resistance has been described [26].
The previous studies conducted by these investigators indicated that the prevalence of S. aureus resistance to mupirocin in South Africa and Nigeria was 7% and 0.5% respectively [7,27]. This study reports on the phenotypic and molecular characterization of mupirocin-resistant S. aureus isolates in Nigeria and South Africa.

Antimicrobial susceptibility testing and PCR detection of the mupA gene
A total of 17 S. aureus isolates were investigated based on their resistance to mupirocin from two studies on antibiotic susceptibility patterns of S. aureus obtained from clin-ical samples in Nigeria and South Africa [7,27]. Susceptibility to various antibiotics was based on the disk diffusion method according to the National Committee for Clinical Laboratory Standards (now Clinical Laboratory Standards Institute) guidelines [28]. Susceptibility to heavy metals (cadmium acetate, mercuric chloride) and nucleic-acid binding compounds (ethidium bromide and propamidine isethionate) was performed on the isolates using disks prepared in the laboratory with the indicated concentrations (10 μl): cadmium acetate (50 μg), propamidine isethionate (50 μg), mercuric chloride, (109 μg) and ethidium bromide (60 μg). Interpretation of zone diameters were considered as follows: ≤ 9 mm (resistance), 10-12 mm (intermediate) for cadmium acetate; ≤ 25 mm (resistance) for mercuric chloride; ≤ 10 mm (resistance), 11-14 mm (intermediate) for propamidine isethionate; and ≤ 9 mm (resistance), 10-14 mm (intermediate) for ethidium bromide, as published previously [29]. Minimum inhibitory concentration (MIC) of mupirocin was determined using E-test strips (AB Biodisk, Solna, Sweden) according to the manufacturer's instructions. The high-level mupirocin-resistant isolates (based on the disk diffusion and E-test methods) were confirmed by PCR detection of the mupA gene as described previously [7,27].
The low-level mupirocin-resistant isolates were obtained from eleven wound samples and one isolate each from blood, urine samples, and endotracheal aspirate. Furthermore, the high-level mupirocin-resistant isolate (A15) from South Africa was obtained from a wound sample while 35 IBA, the isolate from Nigeria was recovered from a blood sample. Information on the source of the methicillin-susceptible S. aureus (MSSA) from South Africa (P1929) was not available.

Plasmid DNA isolation
Plasmid DNA was isolated by the cetyl trimethyl ammonium bromide method (CTAB) as earlier reported [30]. Plasmids were analysed by agarose (0.6% w/v) gel electrophoresis in 1 × TAE buffer (pH 7.2) at 25 V for 16 hr. The Staphylococcus aureus strain WBG 4483, which has 4 plasmids (40.3 kb, 22.5 kb, 4.4 kb and 3.5 kb) served as the plasmid molecular size standard. The approximate plasmid sizes (closed circular forms) were estimated by visual inspection and using the GeneTools program (SynGene Bioimaging System, Cambridge, United Kingdom).

Curing experiments
Two of the three high-level mupirocin-resistant isolates (A15 -South Africa and 35 IBA -Nigeria) were selected for curing and conjugation experiments based on their antibiotic resistance profile. The mupirocin-resistant isolate P1929 from South Africa was not included in the experiment because its antibiotic susceptibility pattern was similar to 35 IBA. The loss of resistance determinants (plasmids) was investigated as previously reported [31]. The isolates were subcultured on Brain Heart Infusion Agar (BHIA, Biolab, South Africa) and incubated at 43.5°C for 24 hr. Sub-culturing on freshly prepared BHIA was performed twice and incubated as stated above. Thereafter, serial dilutions of the culture were plated on BHIA plates, and incubated at 37°C for 24 and 48 hrs. Single colonies were replica plated on BHIA plates as control and selection plates of BHIA containing mupirocin (10 mg/L) and erythromycin (5 mg/L) for A15, and mupirocin (10 mg/L) for 35 IBA. Single colonies which grew on the control plate but did not grow on the selection plates were noted and antibiotic susceptibility testing was performed on them to verify loss of resistance. Thereafter, colonies were screened for loss of the resistance determinants by plasmid analysis and visualized after electrophoresis on 0.6% agarose gels. Cured strains (susceptible to mupirocin) were confirmed for loss of resistance by a negative result for the mupA gene by PCR.

Conjugation Experiments
Conjugation experiments were performed in Brain Heart Infusion Broth (BHIB, Biolab, South Africa) with 40% polyethylene glycol as previously reported [31]. Strains A15 and 35 IBA were used as donors and S. aureus WBG541 (fusidic and rifampicin-resistant) was the recipient strain. Donor and recipients controls were set up with each test. Transfer was considered to have occurred when growth was observed on the selection plates from the donor-recipient mixtures and not from the control experiments. The transconjugants (using A15 as the donor strain) were screened on BHIA plates containing mupirocin (10 mg/L), and fusidic acid (5 mg/L); erythromycin (5 mg/L) and fusidic acid (5 mg/L). Selections of transconjugants for 35 IBA were made on BHIA containing mupirocin (10 mg/L) and fusidic acid (5 mg/L); tetracycline (5 mg/L) and fusidic acid (5 mg/L). Transfer frequency was expressed as the number of transconjugants per number of donor cells. Single colonies of transconjugants were screened on BHIA containing appropriate antibiotics by the replica plating method and antibiotic susceptibility testing was performed on transconjugants. Furthermore, the transconjugants were screened to confirm plasmid content by agarose gel electrophoresis. In addition, transconjugants exhibiting highlevel resistance to mupirocin were confirmed by the MIC values and detection of the mupA gene.

PCR-RFLP of the coagulase gene
The 3' end region of the coagulase gene was amplified by PCR and restriction fragment polymorphisms (RFLPs) of the amplicons were determined by digestion with AluI (Fermentas, UK) as previously described [7].

PFGE typing
PFGE typing of SmaI (Fermentas, UK) digested DNA was carried out by a modification of the protocol previously described [32]. The banding patterns were interpreted visually and strains showing the same PFGE pattern were classified into pulsotypes using an alphabet (e.g. A, B, C etc). Numeric sub-codes were used to represent < 3 band difference (subtypes, e.g. A1, B1, etc) based on a previous report [33].

Resistance pattern of low and high-level mupirocin resistant S. aureus isolates to antimicrobial agents
The resistance profiles of the mupirocin-resistant S. aureus isolates to various antimicrobial agents are presented in Table 1. A total of 14 isolates from six health care institutions in KZN, South Africa exhibited low-level resistance to mupirocin (MIC: 8-24 mg/L). They were also resistant to methicillin, gentamicin, tetracycline and trimethoprim. A total of 12 low-level mupirocin resistant isolates were resistant to erythromycin, 10 isolates were resistant to ciprofloxacin, and four isolates each were resistant to rifampicin and chloramphenicol. Resistotyping revealed that six low-level mupirocin resistant isolates expressed resistance to cadmium acetate, propamidine isethionate, mercuric chloride and ethidium bromide. Four isolates were resistant only to mercuric chloride and two isolates were resistant to cadmium acetate and mercuric chloride. An MRSA isolate was susceptible to the heavy metals and nucleic-acid binding compounds. The mupA gene was not detected in the low-level mupirocin resistant strains while the high-level mupirocin resistant isolates with MIC of >1024 mg/L tested positive for the mupA gene. Furthermore, the high-level mupirocin-resistant isolates were resistant to cadmium acetate but susceptible to mercuric chloride and nucleic acid binding compounds (Table 1).  (Figure 2). The strains in type A were observed in two health care institutions located in Durban and a health care facility in Pietermaritzburg, Greytown and Empangeni, while type B was noted in two health care institutions in Durban and in Eshowe. Based on their PFGE patterns, the high-level mupirocin-resistant strains from the two countries were unrelated (Figure 3).

Curing experiment
The isolates 35 IBA and A15 expressing high-level mupirocin resistance were subjected to curing experiments as a first step to determining the genetic location of their resistance determinants. Results of the curing experiments with 35 IBA showed that 6 of 293 colonies (2.0%) screened for loss of resistance were susceptible to mupirocin. Agarose gel electrophoresis indicated that this was associated with the loss of c. 35 kb plasmid (Figure 4). Of the 294 colonies screened on mupirocin and erythromycin selection plates for A15, three and six colonies lost resistance to erythromycin and mupirocin respectively. Resistance to these antibiotics in the cured strains was lost together with plasmids of c. 2.3 kb and 38 kb respectively ( Figure 4). The mupA gene was not detected in the cured strains (susceptible to mupirocin) by PCR.

Conjugation experiment
Transconjugants were not observed on mupirocin selection plates when 35 IBA was used as the donor strain. However, transconjugants were obtained on mupirocin selection plates with A15 as the donor strain. Replica plating of 98 transconjugants on BHIA selection plates with cadmium (5 mg/L), erythromycin (5 mg/L) and mupirocin (10 mg/L) yielded one, three and 98 colonies respectively. The susceptibility pattern of the parent and cured    The emergence and spread of low-level mupirocin-resistant S. aureus should be of concern among health care workers in South Africa because of recent reports on the increase in treatment failure rates for nasal decolonization of MRSA due to the emergence of low-level mupirocin resistance [16][17][18].
Plasmid profiles of parent and cured strains Plasmid-mediated resistance to antimicrobial agents among pathogenic bacteria constitutes a major clinical and economic problem worldwide. In this study, the genetic location of the high-level mupirocin resistance determinant was resolved in two isolates by plasmid analysis, involving curing and conjugation experiments. Three features were identified in the transfer experiments and plasmid analysis. The first feature was the transfer of the 41.1 kb plasmid encoding high-level mupirocin resistance. High-level mupirocin resistance has been found in self-transmissible and non-self transmissible plasmids in different countries [4,[38][39][40], and this study has demonstrated what appears to be the first report of a conjugative

Conclusion
This study has demonstrated the clonal dissemination of multi-resistant MRSA strains exhibiting low-level resistance to mupirocin in KZN and the emergence of highlevel mupirocin resistant S. aureus in Nigeria and South Africa. Although the high-level mupirocin resistant S. aureus strains from the two countries are genetically unrelated, the ability of the 41.1 kb mupirocin plasmid to transfer the resistance determinant indicate the potential for spread to mupirocin-susceptible MSSA and MRSA isolates under the selective pressure of mupirocin. It is also recommended that routine testing of MSSA and MRSA for mupirocin resistance be conducted even in facilities where mupirocin is not administered. This will facilitate the early detection of resistance and assist in the control and spread of mupirocin-resistant S. aureus.
EcoRI restriction pattern (plasmids) of transconjugants derived from the methicillin/mupirocin resistant S. aureus (A15) strain from South Africa Figure 6 EcoRI restriction pattern (plasmids) of transconjugants derived from the methicillin/mupirocin resistant S. aureus (A15) strain from South Africa. M: Molecular weight standard, phage lambda DNA digested with HindIII; Lane 1 -TransMup (resistant to mupirocin); Lane 2 -TransCad (resistant to mupirocin and cadmium EEU planned the plasmid analysis and JL supervised the project. All authors participated in the preparation of the final manuscript.