Resistance to telithromycin is conferred by msr(A), msrC and msr(D) in Staphylococcus aureus

Sir, The antibiotic resistance determinants Msr(A), MsrC and Msr(D) are all members of a group of ‘incomplete’ or Class 2 ABC transporters, characterized by the presence of two fused nucleotidebinding domains but lacking any identifiable transmembrane domains, and are typically involved in cellular processes other than transport. The msr(A) gene is carried on large staphylococcal plasmids, and has been shown to confer high-level inducible resistance to 14and 15-membered ring macrolides and type B streptogramins (MSB antibiotics) in Staphylococcus aureus in the absence of any additional plasmid-encoded determinants. A single copy of msrC is found on the chromosome of Enterococcus faecium. Insertional inactivation of msrC in this species resulted in increased susceptibility to MSB antibiotics. 3 We recently reported the cloning of msrC into S. aureus, where it was found to confer a greatly enhanced level of protection against MSB antibiotics compared with that in E. faecium. Both determinants confer high-level MSB-resistance in S. aureus, whether cloned on a high copy number vector or integrated in single copy into the chromosome. The homologousmsr(D) determinant, found on large chromosomal genetic elements in streptococci, was recently shown to confer low-level resistance to 14-membered ring macrolides and ketolides in Streptococcus pneumoniae. Msr(D) shares a high level of amino acid sequence similarity with Msr(A) and MsrC (75% and 78%, respectively). We report here the cloning of msr(D) into S. aureus RN4220 on a multicopy shuttle vector. The msr(D) gene and its promoter were amplified from Streptococcus pyogenes ATCC BAA-946, cloned into the PCR cloning vector pDrive (Qiagen Ltd) and subcloned into the KpnI site of the staphylococcal vector pSK265. The cloning vectors pDrive and pSK265 carry ampicillinand chloramphenicol-resistance determinants, respectively. All molecular cloning was carried out in Escherichia coli JM109. Transformants carrying this shuttle vector construct, designated pMsrD, were selected on LB agar containing 50 mg/L ampicillin and 5 mg/L chloramphenicol. S. aureus RN4220 was transformed with pMsrD by electroporation, with selection of transformants on agar containing 5 mg/L chloramphenicol. S. aureus RN4220 transformed with pMsrD was found to have enhanced resistance to erythromycin, a 14-membered ring macrolide (MIC 4–8mg/L), and telithromycin, a ketolide antibiotic (MIC 0.125–0.25 mg/L), representing 16-fold and 4-fold increases in MICs, respectively (Table 1). The MIC of telithromycin was increased further to 1 mg/L upon induction with 1 mg/L erythromycin (Table 1). However, pMsrD transformants remained susceptible to 16-membered ring macrolides, lincosamides and to streptogramins A and B, even following induction with erythromycin. Daly et al. recently reported that msr(D) similarly confers resistance to macrolides and inducible resistance to ketolides in S. pneumoniae. Determination of telithromycin MICs gave rise to the unexpected finding thatmsr(A) andmsrC also confer inducible resistance to this ketolide antibiotic in S. aureus (Table 1). MICs of telithromycin were increased by up to 128-fold following induction with 5 mg/L of erythromycin for plasmid-borne msr(A) and msrC in S. aureus RN4220. To the best of our knowledge, this is the first report of ketolide resistance conferred by msr(A) or msrC in any species. The msr(A) gene is preceded by a control region containing a number of inverted repeat sequences and an open reading frame encoding a short ‘leader peptide’ (LP). Deletion of this control region resulted in constitutive resistance to MSB antibiotics in S. aureus. We found that MICs of telithromycin for S. aureus RN4220 carrying msr(A) which was deleted for its control region (2 mg/L) were not affected by the addition of erythromycin, indicating that the msr(A) control region may also be required for inducible resistance to telithromycin. A similar control region has been identified upstream of the msrC gene in E. faecium, and is likely to be responsible for the inducible nature of resistance conferred by this determinant. We located an open reading frame preceded by a Shine–Dalgarno sequence 110 nt upstream of the msr(D) coding sequence that could potentially encode a short LP (Met-Tyr-Leu-Ile-Phe-Met). A series of inverted repeat sequences were identified in this putative control region, but none of these was found to encompass the ribosome binding site preceding msr(D). Moreover, consensus –10 and –35 promoter sequences were identified within the open reading frame of the proposed LP. No promoter sequences were found upstream of the LP. The nucleotide sequence upstream of msr(D) did not closely resemble the control regions upstreamofmsr(A) andmsrC, which are thought to be involved in the regulation of gene expression by translational attenuation. It is therefore unlikely that msr(D) expression is regulated by the samemechanism. Althoughmsr(D) was expressed from its own promoter in S. aureus in this study, it may be noted that other researchers have demonstrated the co-transcription of mef(A) and msr(D). In this case, the LP sequence upstream of msr(D) would also be co-transcribed, and its expression could potentially regulate translation of the msr(D) transcript. It remains unclear how such ‘incomplete’ ABC transporters provide the bacterial cell with protection against antibiotics. It was originally proposed that active efflux of the drugs could be orchestrated by ‘hijacking’ and altering the specificity of