Professor A. D. Russell, Welsh School of Pharmacy, Cardiff University, Cardiff CF1 3XF, UK (e-mail: russellD2@cardiff.ac.uk).
Mupirocin resistance could be transferred from highly resistant clinical isolates of Staphylococcus aureus to highly sensitive recipients of Staph. aureus, Staph. epidermidis and Staph. haemolyticus. Transconjugants of the latter two organisms could transfer this resistance into mupirocin-sensitive Staph. aureus. Moderately resistant strains did not transfer this resistance to sensitive recipients, nor did strains with high-level mupirocin resistance developed by serial transfer or habituation. The inhibitory effects of mupirocin on crude isoleucyl-tRNA synthetases (IRS) isolated from mupirocin-sensitive and -resistant strains of Staph. aureus have been determined. Drug concentrations needed to produce 50% inhibition, I50 values, were very low against IRS from a highly sensitive strain, somewhat higher against IRS from moderately resistant strains, much higher against enzyme from strains trained in vitro to high-level resistance, and considerably higher still against IRS extracted from clinical isolates possessing high-level mupirocin resistance and from the transconjugates of such strains resulting from crosses with mupirocin-sensitive strains. It is concluded that high-level resistance in clinical isolates is plasmid-mediated involving a second, mupirocin-resistant IRS whereas in moderately resistant strains, and in strains trained in vitro to high-level resistance, chromosomal mutations are likely to be responsible for decreasing IRS sensitivity.
Mupirocin (pseudomonic acid A) is a narrow-spectrum antibiotic originally isolated from Pseudomonas fluorescens (Sutherland et al. 1985; Hill et al. 1988). It is active predominantly against staphylococci and more permeable Gram-negative species such as Hemophilus and Neisseria. Enterobacteriaceae are intrinsically resistant due to a permeability barrier (Al-Masaudi et al. 1988). In staphylococci, mupirocin-resistant isolates which show a high level of resistance to this antibiotic have occasionally been clinically isolated. Moderately resistant strains of Staph. aureus have also been isolated (Rahman et al. 1987, 1989, 1990, 1993). Mupirocin is used topically as an ointment against methicillin-resistant Staph. aureus (MRSA), usually for periods of not less than 10_d.
The mechanism of antibacterial action of mupirocin involves specific inhibition of bacterial isoleucyl tRNA synthetase (IRS; Hughes & Mellows 1978a,b, 1980; Capobianco et al. 1989). The mupirocin structure resembles the isoleucyl adenylate complex and thus, the target for its activity is the first part of the aminoacylation reaction in which isoleucyl adenylate is formed.
In this paper, the development and transferability of mupirocin resistance is described.
Materials and methods
Chemicals and antibiotics
Mupirocin was a gift from SmithKline Beecham, Brockham Park, Betchworth, Surrey. Chemicals were purchased from Sigma.
Bacterial strains and culture
The strains used are listed in Table 1 together with their resistance phenotypes, which were used as markers in the transfer experiments. The strains were routinely cultured on nutrient agar (Oxoid) and their identity confirmed by conventional tests and the API STAPH (BioMerieux). Diagnostic sensitivity test (DST) agar (Oxoid) was used instead of nutrient agar when culturing coagulase-negative staphylococci, due to their poor growth on nutrient agar.
Table 1. Staphylococcal strains† fused in transfer experiments or isoleucyl-tRNA synthetase studies
All strains are Staphylococcus aureus except for UHW 1, UHW 2, UHW 4 (Staph. epidermidis) and UHW 7, UHW 8 (Staph. haemolyticus).
Determination of minimum inhibitory concentrations (MIC)
A 2 ml volume of the appropriate concentration of antibiotic was added to 18_ml molten agar to give a series of doubling concentrations from 0·13 to 512 μg ml−1. The plates were poured and overdried before use. Overnight cultures of test bacteria were diluted 1 : 100, and 10 μl were spotted onto the plates using the Denley multipoint inoculator (Denley, Billingshurst, UK) and incubated at 37 °C for 48_h. The MIC was taken as being the lowest concentration of antibiotic that prevented growth.
Stepwise development of mupirocin resistance
Cultures containing a series of increasing concentrations of mupirocin in 10_ml volumes of nutrient broth (Oxoid) were inoculated with 100 μl of an overnight culture of moderately mupirocin-resistant Staph. aureus. The cultures were incubated at 37 °C until there was a clear differential of growth between two dilutions. The culture with the highest concentration of antibiotic which permitted growth was used to inoculate a further series of media containing increasing concentrations of mupirocin. This procedure was repeated for a maximum of 10 sub-cultures. Mupirocin resistance thus produced was found to be stable. The resistant variants selected were tested for cross-resistance to other antibiotics.
Antibiotic disc susceptibility testing
A sterile swab dipped into an overnight culture of test organism was used to inoculate evenly DST agar. When the inoculum had dried, the antibiotic sensitivity discs (Oxoid) were placed on the inoculated plates (no more than six to a plate) and incubated overnight at 37 °C. Staphylococcus aureus NCTC 6571 acted as control organism.
Donor and recipient cells were grown overnight at 37 °C in 10_ml nutrient broth; 1_ml of the donor cell culture and 3_ml of the recipient cell culture were mixed, and the mixture filtered (Millipore filter, pore size 0·4 μm). The filter was placed, with the bacterial cells facing upwards, onto a nutrient agar plate (DST agar was used for transfers involving coagulase-negative staphylococci as recipients). After incubation overnight at 37 °C, the filter was removed and vortexed in 1_ml nutrient broth. The suspended mating mixtures were serially diluted and dilutions spread onto plates containing appropriate selective antibiotics at a concentration of 5 μg ml−1. Donor and recipient controls were always plated separately. Colonies were counted after incubation for 48_h at 37 °C. The transfer frequencies were expressed as the number of transconjugants per recipient cell. The direction of transfer was determined by replica plating and utilizing secondary markers not used for selection in the original cross. All crosses were repeated three times unless stated otherwise, and all the results are presented as mean values.
Extraction of crude IRS and determination of activity
A 10_ml aliquot of an overnight broth culture of the test organism was inoculated into 500_ml nutrient broth in a 2 litre flask and incubated at 37 °C in a shaking water-bath at 200–220 rev min−1. Cell growth was monitored spectrophotometrically at an absorbance of 660 nm. When the culture was at the top of the exponential phase, the cells were harvested.
The culture was harvested in the RC5C centrifuge (Dupont, Stevenage, UK) at 5500 rev min−1 and 4 °C for 10_min. The cell pellet was washed with phosphate-buffered saline (PBS) and centrifuged twice before determining the weight of the pellet. The cells were resuspended in lysis/ sonication buffer (20_mmol l−1 Tris, 2_mmol l−1 dithiothreitol (DTT), 5_mmol l−1 MgCl2, 0·5 mmol l−1 EDTA, pH 8·2) to a final density of 0·4_g cells ml−1 of buffer. Lysostaphin and DNAse 1 were added to final concentrations of 150 μg ml−1 and 15 μg ml−1, respectively, and incubated overnight at 4 °C. The suspension was sonicated for five cycles of 20_s on and 30_s off at 18–20 microns amplitude. The broken cell suspension was centrifuged at 19 000 rev min−1 for 20_min. The supernatant fluid was removed and glycerol added to a final concentration of 30% (v/v). The sample was stored at −20 °C until required for use. The dilute crude IRS in enzyme buffer (Tris 0·138_g in 9 ml H2O, pH 8·2) was placed on ice; 2_ml 100_mmol l−1 DTT solution were then added and the solution made up to 100_ml with water. From this, 1 : 2·5, 1 : 5, 1 : 10, 1 : 20 and 1 : 50 dilutions were prepared. A 50 μl volume of each dilution was added to 100 μl reagent mixture of the following composition: buffer 5_ml, ATP 2·5_ml (30_g l−1), tRNA 2·5_ml (20_g l−1), 14C isoleucine 1·0_ml, water 9_ml. This was incubated for 10_min; 2_ml of 7% trichloracetic acid (TCA) were then added and the samples left on ice for 30_min to allow the precipitate to form.
Determination for I50 for crude IRS sample
Aliquots (10 μl) of the appropriate dilution of mupirocin or water were placed in reaction vials and equilibrated at 37 °C. Diluted IRS (50 μl) was added and incubated for 5_min, then 100 μl of warmed reagent mixture were added. After incubation for 15_min, 2_ml 7% TCA were added and the samples left on ice for 30_min to allow the precipitate to form.
Filtration and counting
The reaction mixture from the vial was deposited onto a Millipore filtration manifold. The vial was rinsed with 2_ml TCA and the washings deposited on the same filter. The filter was washed with 2 × 10_ml aliquots of TCA and 2 × 10_ml aliquots of ethanol. The filters were then placed in a scintillation vial in 6_ml Optisafe scintillation fluid. The vials were then counted for 2_min. The percentage inhibition of IRS activity at each concentration was calculated as follows:
% inhibition = (C - S) × 100/C
where C is the control count and S is the sample count for the concentration of enzyme. The percentage inhibition was then plotted against mupirocin concentrations for each crude IRS sample, and the I50 calculated as the concentration (μg l−1) of mupirocin that reduced the activity of IRS by 50%. IRS determinations were carried out in duplicate.
Training to mupirocin resistance
Three mupirocin-sensitive strains of Staph. aureus (NCTC 6571, 11561 and RN2677) did not readily gain mupirocin resistance, a frequency of <1 × 10−7 cells spontaneously mutating to give a resistance (MIC) of >5 μg ml−1. Conversely, three strains of Staph. aureus with moderate resistance to the antibiotic (MICs 8–32 μg ml−1, i.e. Clarke, G1217 and K227) developed high resistance (MICs >512 μg ml−1) when trained to grow on media containing the drug. An example is provided in Fig. 1. These new strains were termed Staph. aureus Clarke (T10), G1217 (T10) and K227 (T8), where the figure in brackets indicates the number of transfers for each strain to become highly mupirocin-resistant.
Moderately resistant strains and highly resistant trained strains possessed stable mupirocin resistance. By comparison, the highly resistant Staph. aureus strains L2 and L8 became sensitive to mupirocin (MICs <0·25 μg ml−1) when exposed to unfavourable conditions of incubation at 40 °C or growth in non-selective media.
Transferability of mupirocin resistance between Staph. aureus strains
The initial series of transfer experiments detected transfer of high-level mupirocin resistance to mupirocin-sensitive Staph. aureus strains. Transconjugants were detected on plates containing mupirocin and a selective antibiotic for recipient, i.e. streptomycin for Sau 2, rifampicin for 11561 (R), RN450 (R) and 6571 (R), and novobiocin for RN2677. Neither donor nor recipient cultures alone grew on these media.
Mupirocin resistance was transferred at a frequency of 9·7 × 10−6 from highly resistant Staph. aureus strains L2, L8, Kelesh, Eagles and C7, but no transfer was demonstrated from strain F89. Resistance was transferred into two (Sau 2 and RN2677) of five recipient strains used ( Table 2). Confirmation that mupirocin resistance had been transferred was obtained by testing the susceptibility of the possible transconjugants against secondary markers, in each case 50 colonies for each cross being tested. The donor strains (L2, L8, Kelesh and Eagles) were all resistant to erythromycin but the transconjugants from crosses with either Sau2 or RN 2677 were sensitive. All the transconjugants from crosses involving Sau2 were resistant to rifampicin. Staphylococcus aureus strain C7 is resistant to mupirocin only, but the transconjugants involving this strain with Sau 7 or RN2677 were resistant to at least three antibiotics. It can thus be concluded that mupirocin resistance was transferred from the donors to the recipients.
Table 2. Mupirocin transfer frequencies from highly resistant strains of Staphylococcus aureus to sensitive strains of Staph. aureus
Figures are transfer frequencies.
(R), trained to rifampicin resistance.
4·6 × 10−6
2·1 × 1·0−6
9·7 × 10−6
4·4 × 10−6
<1·9 × 10−6
3·2 × 10−6
<3·9 × 10−9
<2·2 × 10−9
<7·7 × 10−10
<6·7 × 10−10
<2·6 × 10−9
<8·1 × 10−10
<1·7 × 10−9
<5·8 × 10−10
<4·5 × 10−10
<2·6 × 10−9
<6·4 × 10−10
<4·3 × 10−10
<4·8 × 10−10
<6·5 × 10−10
<4·9 × 10−10
<1·8 × 10−9
<2·1 × 10−9
4·4 × 10−6
1·2 × 10−6
4·0 × 10−6
1·6 × 10−6
<1·2 × 10−9
1·0 × 10−6
Mupirocin resistance was not transferred above detectable levels to mupirocin-sensitive strains from the moderately resistant Staph. aureus strains, or from the strains trained to high-level resistance.
Interspecies transfer of mupirocin resistance
Donors (Staph. aureus strains Kelesh and Eagles) which showed a high level of mupirocin resistance (MICs >512 μg ml−1) were used to study interspecies transferability. Transconjugants were selected on DST agar with mupirocin and gentamicin (an antibiotic to which the non-Staph. aureus strains were resistant; Table 1). The results ( Table 3) demonstrate that mupirocin resistance was transferred into three (Staph. epidermidis UHW2, and Staph. haemolyticus UHW7 and UHW8) of the five recipient strains employed. The transconjugants were confirmed by their secondary markers. Strain Kelesh was a less efficient donor than Eagles. Transconjugants (K2 and E2) from the crosses of Staph. aureus Kelesh and Eagles with UHW 2 were resistant to rifampicin and sensitive to erythromycin, and those (K7, E7, K8 and E8) from the crosses with UHW7 and UHW8 were resistant to ampicillin and sensitive to erythromycin. Using the API STAPH system, transconjugants were confirmed as Staph. epidermidis or Staph. haemolyticus, thus demonstrating conclusively that mupirocin resistance had been transferred.
Table 3. Mupirocin transfer frequencies from highly resistant strains of Staphylococcus aureus to mupirocin-sensitive coagulase-negative staphylococci
Recipient and strain
Staph. epidermidis UHW 1
<2·9 × 10−9
<2·4 × 10−6
Staph. epidermidis UHW 2
7·4 × 10−7
1·0 × 10−6
Staph. epidermidis UHW 4
<3·2 × 10−9
<1·9 × 10−9
Staph. haemolyticus UHW 7
2·5 × 10−8
6·8 × 10−7
Staph. haemolyticus UHW 8
4·2 × 10−8
7·9 × 10−7
These mupirocin-resistant transconjugants of Staph. epidermidis and Staph. haemoloyticus transferred mupirocin resistance to mupirocin-sensitive Staph. aureus ( Table 4).
Table 4. Transfer frequencies of mupirocin resistance from mupirocin-resistant coagulase-negative staphylococci (transconjugants) to mupirocin-sensitive strains of Straphylococcus aureus
The degrees of inhibition of IRS by different concentrations of mupirocin are shown in Fig._2. In the donors, Kelesh and Eagles, and their respective transconjugants, the I50 values of mupirocin are high ( Table 5) whereas in the recipient, Sau 2, used in the transconjugant experiments, the I50 value of mupirocin is only 1/1000 of that value.
Table 5. I50 concentrations of mupirocin against IRS and sensitivity of Staphylococcus aureus strains to the antibiotic
I50 (ng ml−1)
MIC (μg ml−1)
Transconjugants from Staph. aureus Kelesh × Staph. aureus Sau 2.
Transconjugants from Staph. aureus Eagles × Staph. aureus Sau 2.
Some moderately resistant strains, and strains trained to high-level resistance, were also investigated. I50 values of mupirocin against IRS from the parent strains ranged from 5·6 to 17·8 μg ml−1 and from 224 to 891 ng ml−1 against IRS from the trained strains (see Table 5; Fig._3 exemplifies the results obtained with K227 and K227 (T8)). I50 values against these highly resistant strains are thus considerably below those described for clinical isolates and their transconjugants.
Highly sensitive strains of Staph. aureus (MICs <1 μg l−1) could not be trained to mupirocin resistance, although moderately resistant strains (mupirocin MIC 8–16 μg ml−1) could be trained to give a stable, high-level resistance (MIC >512 μg ml−1). This trained resistance could not, however, be transferred to mupirocin-sensitive recipients, which suggests that such resistance is chromosomally mediated rather than located on a plasmid. In this context, it is interesting to note the findings of Antonio et al. (1995) who concluded that moderate mupirocin resistance resulted from a point mutation on the Staph. aureus chromosome.
Studies with two strains (L2 and L8) of Staph. aureus showing high-level resistance to mupirocin demonstrated that growth at 40 °C in a non-selective medium produced a loss of resistance (MICs <0·25 μg ml−1), implying that high-level resistance is located on a plasmid. Noble et al. (1988) and Rahman et al. (1987, 1989, 1990, 1993) have amply demonstrated the association of plasmids with high-level resistance.
Al-Masaudi et al. (1991) and Kloos & Lamb (1991) showed that gene transfer between staphylococcal populations is possible at room temperature. In the experiments described in the present paper, it was found that mupirocin resistance is transferred from some (but not all) highly mupirocin-resistant strains of Staph. aureus to some highly sensitive strains of this organism ( Table 2) and of other staphylococcal species (Table 3). Furthermore, mupirocin resistance from such transconjugants of Staph. epidermidis and Staph. haemolyticus could then be transferred into sensitive Staph. aureus recipients, interestingly at a higher transfer frequency ( Table 4). Non-transferable high-level mupirocin resistance, e.g. from Staph. aureus F89 ( Table 2) may be due to a narrow host-range plasmid or to the incorporation of the gene responsible into the chromosome to give permanent high-level resistance. It is also possible that F89 has become trained to high-level resistance in the environment, although this is considered to be unlikely due to the favourable conditions needed to generate a stable high resistance.
The exchange of genetic material between Staph. aureus and Staph. epidermidis is probably associated with the high degree of DNA homology between the species involved (Kloos 1980). Such a genetic exchange could provide a reservoir of resistance genes for Staph. aureus, as both species occupy the same niche on the human skin (Mellows 1985).
Early studies with cell-free systems from Escherichia coli demonstrated that the target site of mupirocin is isoleucyl-tRNA synthetase (IRS), the enzyme which charges the appropriate tRNA with isoleucine (Farmer et al. 1992).
The I50 values of mupirocin were low against crude IRS enzyme isolated from mupirocin-sensitive strains, rather higher against mupirocin moderately resistant strains, much higher against strains trained to a high mupirocin resistance, and considerably higher still against highly resistant clinical isolates ( Table 5). Farmer et al. (1992) reported that a general increase in IRS I50 values corresponded to a general increase in MIC values.
The transferable mupirocin resistance gene is located on a large plasmid (Rahman et al. 1987, 1989, 1990, 1993) and is responsible for producing a different IRS enzyme from that present on the Staph. aureus chromosome (Gilbart et al. 1993; Cookson 1998). This second, plasmid-specified enzyme may compete more strongly for the active site in the acylation process than the chromosomal IRS, and will thus require a higher mupirocin concentration to reduce its activity. The IRS from transconjugants had similar I50 values to those from the donor strains, which suggests strongly that such an IRS enzyme is responsible for mupirocin resistance and that high-level mupirocin resistance is located on a plasmid.
In the trained strains, resistance to the drug is likely to be due to chromosomal mutations being selected at increasing mupirocin concentrations, as also suggested by Cookson (1998). Such mutations are one-point mutations in the gene (Antonio et al. 1995), the IRS produced having a lower affinity for the binding process than the plasmid-encoded IRS.
Characteristics of mupirocin resistance in Staph. aureus are summarized in Table 6, which examines the types of response to mupirocin, MIC values, how sub- and supra-MIC levels affect growth, and the I50 values. High-level and moderate-level mupirocin resistance poses a significant clinical problem.
Table 6. Characteristics of mupirocin resistance in Staphylococcus aureus strains