Erythromycin resistance by ribosome modification.

Erythromycin inhibits protein synthesis by its effect on ribosome function (14, 118, 119). The metabolic modifications that enable cells to cope with the inhibitory action of erythromycin fall under major headings that include (i) target site alteration, (ii) antibiotic modification, and (iii) altered antibiotic transport. This minireview concentrates on target site alteration, which for erythromycin is the 50S subunit of the ribosome. The first clinical isolates of macrolide-resistant staphylococci were described in reports from France, England, Japan, and the United States shortly after the introduction of erythromycin into clinical practice in 1953. On the basis of current understanding of the biochemistry of erythromycin’s action, resistance in most of the strains that were described in early reports can be ascribed to a posttranscriptional modification of the 23S rRNA by an adenine-specific N-methyltransferase (methylase) specified by a class of genes bearing the name erm (erythromycin ribosome methylation). The last decade has seen the isolation and characterization of approximately 30 erm genes from diverse sources, ranging from clinical pathogens to actinomycetes that produce antibiotics; for many of these genes, both the respective nucleotide sequences that encode the methylases as well as the flanking sequences that control their expression have been determined. A tabulation of the erm genes that have been described is presented in Table 1. Any discussion of mechanisms of resistance to macrolide antibiotics must include the chemically distinct, but functionally overlapping, lincosamide and streptogramin B families as well. This type of resistance has therefore also been referred to as MLS resistance. Members of the MLS antibiotic superfamily include, among the macrolides, carbomycin, clarithromycin, erythromycin, josamycin, midecamycin, mycinamicin, niddamycin, rosaramicin, roxithromycin, spiramycin, and tylosin; among the lincosamides, celesticetin, clindamycin, and lincomycin; and among the streptogramins, staphylomycin S, streptogramin B, and vernamycin B. The streptogramin family is subdivided into A and B groups or alternatively into M and S groups, respectively. Methylation of A2058 confers resistance to the Band S-group streptogramins but not to the Aand M-group streptogramins. The reason for this grouping was originally based on empirical observations from clinical bacteriology that resistance to one class often involved resistance to the other two classes (11, 16, 35, 39, 41, 135); however, (i) the three classes of antibiotics interact competitively when binding to the 50S subunit, and only one antibiotic molecule can bind per 50S subunit (129, 130); this suggests that the binding sites for these antibiotics overlap or at least functionally interact. (ii) Nucleotide alterations in 23S rRNA, both mutational and posttranscriptional, that confer coresistance to MLS antibiotics appear to cluster in the peptidyltransferase region in 23S rRNA domain V, providing a physical basis and a common location for their sites of action (50, 101–104, 109, 110, 128) (Fig. 1 and Table 2), and (iii) footprinting experiments show that the nucleotides in 23S rRNA domain V are protected by bound MLS antibiotics against modification by agents such as dimethyl sulfate (DMS) and kethoxal that can derivatize purine and pyrimidine bases in single-stranded DNA or RNA (26, 76) (Table 3). The erm family of genes is not alone in conferring clinical resistance to macrolide antibiotics. A notable early exception to the established MLS resistance pattern was the MS pattern reported by Jánosy and coworkers (58, 59), who described clinical isolates that were coresistant to erythromycin and streptogramin B but that remained susceptible to lincosamide antibiotics. The molecular basis for resistance in these strains was subsequently shown by Ross et al. (94) to involve the active efflux of erythromycin and streptogramin B but not clindamycin. Additional mechanisms of macrolide resistance, all associated with the acquisition of new genetic information, including structural modification of erythromycin by phosphorylation (82), glycosylation (60), and lactone ring cleavage by erythromycin esterase (2, 83), have been added to the list. Mechanisms involving mutational alteration of genes that normally reside in the host and that encode either ribosomal protein or rRNA have also been described and will be discussed below in detail. Reviews of erythromycin resistance that relate to material covered in the present work have been presented previously (4, 18, 20, 21, 28, 29, 133). Recent developments in the synthetic chemistry of semisynthetic macrolides, including the biological and clinical aspects of their actions, have been reviewed by Kirst (65, 66). A forthcoming review covers the inducible nature of MLS resistance and its implications for the mechanism of action of erythromycin (134).