Jeremy I. Ross

Inducible erythromycin resistance in staphlyococci is encoded by a member of the ATP‐binding transport super‐gene family

A Staphylococcus epidermidis plasmid conferring inducible resistance to 14‐membered ring macrolides and type B streptogramins has been analysed and the DNA sequence of the gene responsible for resistance determined. A single open reading frame of 1.464kbp, preceded by a complex control region containing a promoter and two ribosomal binding sites, was identified. The deduced sequence of the 488‐amino‐acid protein (MsrA) revealed the presence of two ATP‐binding motifs homologous to those of a family of transport‐related proteins from Gram‐negative bacteria and eukaryotic cells, including the P‐glycoprotein responsible for multidrug resistance. In MsrA, but not these other proteins, the two potential ATP‐binding domains are separated by a Q‐linker of exceptional length. Q‐linkers comprise a class of flexible inter‐domain fusion junctions that are typically rich in glutamine and other hydrophilic amino acids and have a characteristic spacing of hydrophobic amino acids, as found in the MsrA sequence. Unlike the other transport‐related proteins, which act in concert with one or more hydrophobic membrane proteins, MsrA appears to function independently when cloned in a heterologous host (Staphylococcus aureus RN4220). MsrA might, therefore, interact with and confer antibiotic specificity upon other transmembrane efflux complexes of staphylococcal cells. The active efflux of [14C]‐erythromycin from cells of S. aureus RN4220 containing msrA has been demonstrated.

16S rRNA Mutation Associated with Tetracycline Resistance in a Gram-Positive Bacterium

ABSTRACT A genetic basis for tetracycline resistance in cutaneous propionibacteria was suggested by comparing the nucleotide sequences of the 16S rRNA genes from 16 susceptible and 21 resistant clinical isolates and 6 laboratory-selected tetracycline-resistant mutants of a susceptible strain. Fifteen clinical isolates resistant to tetracycline were found to have cytosine instead of guanine at a position cognate with Escherichia coli 16S rRNA base 1058 in a region important for peptide chain termination and translational accuracy known as helix 34. Cytosine at base 1058 was not detected in the laboratory mutants or the tetracycline-susceptible strains. The apparent mutation was recreated by site-directed mutagenesis in the cloned E. coli ribosomal operon, rrnB, encoded by pKK3535. E. coli strains carrying the mutant plasmid were more resistant to tetracycline than those carrying the wild-type plasmid both in MIC determinations and when grown in tetracycline-containing liquid medium. These data are consistent with a role for the single 16S rRNA base mutation in clinical tetracycline resistance in cutaneous propionibacteria.

EFFECTS OF BENZOYL PEROXIDE AND ERYTHROMYCIN ALONE AND IN COMBINATION AGAINST ANTIBIOTIC-SENSITIVE AND -RESISTANT SKIN BACTERIA FROM ACNE PATIENTS

Topical formulations of erythromycin and benzoyl peroxide are popular and effective treatments for mild to moderate acne vulgaris. Use of the former is associated with resistance gain in both skin propionibacteria and coagulase‐negative staphylococci, whereas use of the latter is not. We evaluated the efficacy of a combination of erythromycin and benzoyl peroxide against a total of 40 erythromycin‐sensitive and ‐resistant strains of Staphylococcus epidermidis and skin propioni‐ bacteria in vitro. Using the checkerboard technique, five erythromycin resistant strains of Propionibacterium acnes were inhibited synergistically or additively by the combination. Complete mutual indifference was exhibited between the drugs against the remaining 35 strains. However, erythromycin resistant staphylococci and propionibacteria were inhibited by the same concentration of benzoyl peroxide as erythromycin‐sensitive strains. These results suggest that, although the combination of erythromycin and benzoyl peroxide is not synergistic against the majority of erythromycin‐resistant staphylococci and propionibacteria, the concomitant therapeutic use of both drugs should counteract the selection of erythromycin‐resistant variants and reduce the number of pre‐existing resistant organisms on the skin of acne patients.

Msr(A) and related macrolide/streptogramin resistance determinants: incomplete transporters?

The gene msr(A) confers inducible resistance to 14-membered-ring macrolides and type B streptogramins (MS(B) resistance) in staphylococci. The encoded hydrophilic protein (Msr(A)) is 488 amino acids and contains two ATP-binding motifs characteristic of the ABC transporters. The classical organisation of ABC transporters requires interaction between the two cytoplasmically located ATP-binding domains with two hydrophobic domains positioned in the membrane. Msr(A) appears to mediate drug efflux and yet contains no hydrophobic membrane spanning domains. In addition, Msr(A) functions in previously sensitive heterologous hosts such as Staphylococcus aureus in the absence of other plasmid encoded products. Current research on Msr(A) and related determinants in Gram-positive cocci and in antibiotic producing organisms is reviewed. Alternative hypotheses for the mechanism of action of Msr(A) (i.e. active transport vs. ribosomal protection) are discussed. Evidence indicating Msr(A) may have a role in virulence in addition to conferring antibiotic resistance is also considered.

Identification of a chromosomally encoded ABC-transport system with which the staphylococcal erythromycin exporter MsrA may interact

The energy-dependent efflux of erythromycin (Er) in staphylococci is due to the presence of msr A, which encodes an ATP-binding protein. MsrA is related to the multi-component ATP-binding cassette (ABC) transporters which characteristically also contain membrane-spanning domains. Since MsrA functions in a heterologous host in the absence of other plasmid-encoded products, the requirement for a transmembrane (TM) complex might be fulfilled by hijacking a chromosomally encoded protein. Two genes, stpA and smpA, were identified upstream from msrA on the original Staphylococcus epidermidis plasmid, encoding an ATP-binding protein and a hydrophobic TM protein, respectively. Sequences highly similar to stpA and smpA (stpB and smpB) were also found adjacent to a chromosomal copy of msrA in S. hominis. In Southern blots, internal fragments of stpA or smpA hybridized to the chromosome of the Ers S. aureus RN4220. Cloning and sequence analysis of the region identified revealed the presence of two genes, stpC and smpC, related to stpA and smpA. The deduced amino-acid sequences of the gene products showed that StpA and StpC were 85% identical, whereas SmpA and SmpC were 65% identical. A gene similar to msrA was not present in the S. aureus chromosome. There was no further sequence similarity outside these conserved regions. These results indicate that the chromosomes of S. hominis and S. aureus contain sequences encoding a potential TM protein with which MsrA might interact.

MINIMAL FUNCTIONAL SYSTEM REQUIRED FOR EXPRESSION OF ERYTHROMYCIN RESISTANCE BY MSRA IN STAPHYLOCOCCUS AUREUS RN4220

Previous studies have suggested that inducible erythromycin (Er) resistance in staphylococci mediated by the plasmid-borne ABC-transporter msrA is dependent on additional unidentified chromosomally encoded transmembrane (TM) domains. The requirement for two S. aureus candidate sequences, stpC and smpC, highly similar to sequences adjacent to msrA on the original S. epidermidis plasmid was investigated. Deletion of the sequences by allelic replacement was accomplished by electroporation of S. aureus RN4220 with a nonreplicating suicide vector. S. aureus strains carrying a delta(stpC-smpC) mutation showed an identical ErR phenotype to those arising from single crossover events and unmutated RN4220 containing msrA. This proves that neither stpC nor smpC is required for ErR. To further define the minimal functional unit required for MSR, the control region within the leader sequence of msrA was deleted. This resulted in constitutive resistance to Er and type B streptogramins (Sg), proving that SgR does not require the presence of Er. Deletion constructs containing the N- or C-terminal ABC regions of MsrA did not confer ErR in RN4220 singly or in combination.

Resistance to erythromycin and clindamycin in cutaneous propionibacteria is associated with mutations in 23S rRNA

The skin bacterium Propionibacterium acnes has been implicated in the pathogenesis of inflamed lesions of the multifactorial skin disease acne vulgaris [1]. Agents, including some antibiotics, which reduce the numbers of this organism in vivo are therapeutic and failure to respond to treatment has been associated with the presence of erythromycin-resistant P. acnes [2]. Combined resistance to erythromycin and clindamycin in cutaneous propionibacteria was first reported in 1979 in the USA [3]. Subsequently erythromycin-resistant organisms have been isolated from antibiotic-treated acne patients in several countries [2, 4, 5]. The incidence of such resistant organisms is increasing, and in 1996 49% of acne patients attending the Leeds General Infirmary carried erythromycin-resistant strains [unpubl. data]. Generally bacteria acquire resistance genes on plasmids, transposons or phages. The most common form of aquired resistance to erythromycin is via methylation of Escherichia coli equivalent base A-2058 in 23S rRNA by methylases encoded by erm genes. This leads to reduced binding of three classes of antibiotics, macrolides (erythromycin, clarithromycin and tylosin), lincosamides (clindamycin) and type B streptogramins (virginiamycin S, pristinamycin IA), known collectively as the MLS antibiotics. All these drugs inhibit protein synthesis and have overlapping binding sites in 50S ribosomal subunits [6]. Other acquired mechanisms of resistance to erythromycin include enzymatic inactivation and active efflux [7]. We previously classified 54 isolates of erythromycin-resistant P. acnes collected from antibiotic-treated acne patients in Leeds between 1982 and 1988 into four phenotypic classes based on their cross-resistance patterns to eight MLS antibiotics [8] (table 1). Class II resistance was only detected in Propionibacterium granulosum and is not considered in this report. Attempts to elucidate the mechanism of resistance in class I, III and IV strains by cloning and hybridization studies using erm gene probes were unsuccessful and we could detect no evidence of plasmids in P. acnes. Laboratory-selected spontaneous erythromycin-resistant mutants of the sensitive P. acnes strain P37 demonstrated constitutive MLS resistance (class I) which resembled the phenotype of clinical isolates. This evidence and reports of clinical resistance to the macrolide clarithromycin associated with mutation of bases in the peptidyl transferase