Macrolides are composed of 14 (erythromycin and clarithromycin)-, 15 (azithromycin)-, or 16 (josamycin, spiramycin, and tylosin)-membered lactones to which are attached amino and/or neutral sugars via glycosidic bonds. Erythromycin was introduced in 1952 as the first macrolide antibiotic. Unfortunately, within a year, erythromycin-resistant (Emr) staphylococci from the United States, Europe, and Japan were described (101). Erythromycin is produced bySaccharopolyspora erythraea, while the newer macrolides are semisynthetic molecules with substitutions on the lactone. The newer derivatives, such as clarithromycin and azithromycin, have improved intracellular and tissue penetration, are more stable, are better absorbed, have a lower incidence of gastrointestinal side effects, and are less likely to interact with other drugs. They are useable against a wider range of infectious bacteria, such as Legionella, Chlamydia,Haemophilus, and some Mycobacterium species (notM. tuberculosis), and their pharmacokinetics provide for less frequent dosing than erythromycin (21, 47, 96,97). As a result, the usage of the newer macrolides has increased dramatically over the last few years, which has led to increased exposure of bacterial populations to macrolides (101-103,107).
Macrolides inhibit protein synthesis by stimulating dissociation of the peptidyl-tRNA molecule from the ribosomes during elongation (101,103). This results in chain termination and a reversible stoppage of protein synthesis. The first mechanism of macrolide resistance described was due to posttranscriptional modification of the 23S rRNA by the adenine-N 6 methyltransferase (101-103). These enzymes add one or two methyl groups to a single adenine (A2058 in Escherichia coli) in the 23S rRNA moiety. Over the last 30 years, a number of adenine-N 6-methyltransferases from different species, genera, and isolates have been described. In general, genes encoding these methylases have been designated erm(erythromycin ribosome methylation), although there are exceptions, especially in the antibiotic-producing organisms (see Tables 1 and 3) (103). As the number of erm genes described has grown, the nomenclature for these genes has varied and has been inconsistent (Table 1). In some cases, unrelated genes have been given the same letter designation, while in other cases, highly related genes (>90% identity) have been given different names.
rRNA methylase genes involved in MLSB resistance
The binding site in the 50S ribosomal subunit for erythromycin overlaps the binding site of the newer macrolides, as well as the structurally unrelated lincosamides and streptogramin B antibiotics. The modification by methylase(s) reduces the binding of all three classes of antibiotics, which results in resistance against macrolides, lincosamides, and streptogramin B antibiotics (MLSB). The rRNA methylases are the best studied among macrolide resistance mechanisms (47, 101-103). However, a variety of other mechanisms have been described which also confer resistance (Table2). Many of these alternative mechanisms of resistance confer resistance to only one or two of the antibiotic classes of the MLSB complex.
Efflux and inactivating genes
In this review, we suggest a new nomenclature for naming MLS genes and propose to use the rules developed for identifying and naming new tetracycline resistance genes (51, 52). This system, with a few recent modifications, was originally designed because of the ability of two genes to be distinguished uniquely by DNA-DNA probe methodology (51). It was generally found that two genes with <80% amino acid sequence identity provided enough variability in nucleotide sequence to permit distinct probes to be designed. Although many investigators are likely to sequence new genes, the use of probe technology allows rapid identification of isolates containing potentially new genes, as well as a reliable way to screen populations and determine the frequency of any one resistant determinant. Therefore, we continued this paradigm by assigning two genes of ≥80% amino acid identity to the same class and same letter designation, while two genes that show ≤79% amino acid identity are given a different letter designation. Table 1 shows the results of the classification, with some classes having members with little variability, while others, like classes A and O, show a greater range of homology at both the DNA and amino acid levels. As new gene sequences emerge, ideally they will need to be compared by oligonucleotide probe hybridization and/or sequence analysis against the bank of known genes before a new designation is assigned. If multiple genes are available in any one class, especially when there is a range as in class A, then all representative members of the class should be examined, not just one. To confirm that the proposed name or number for the newly discovered resistance determinant has not been used by another investigator, please contact M. C. Roberts for this information. A similar request has been made for newtet genes (52).
RRNA METHYLASES
Over the last 30 years, a large number of different rRNA methylase genes (erm) have been isolated from a variety of bacteria that range from E. coli to Haemophilus influenzaein gram-negative species and from Streptococcus pneumoniaeto Corynebacterium spp. in gram-positive species (Table3). In addition, a variety of gram-positive and gram-negative anaerobes, and even spirochetes such as Borrelia burgdorferi and Treponema denticola, have all been shown to carry erm genes (Table 3) (36, 77, 78). All erm enzymes methylate the same adenine residue, resulting in an MLSB phenotype (9, 100-103). This adenine (A2058) or one of the adjacent residues in the peptidyltransferase region (A2057 or A2059) is changed to another nucleotide by mutation in macrolide-resistant Mycobacterium intracellulare,Mycobacterium avium, Propionibacterium spp., andHelicobacter pylori (58, 84, 100-103).
Location of antibiotic resistance genesa
Differences between the various erm genes are seen in the regulation of expression of the phenotype. Some of the enzymes are inducibly regulated by translational attenuation of a mRNA leader sequence; in the absence of erythromycin, the mRNA is in an inactive conformation due to a sequestered Shine-Dalgarno sequence, preventing efficient initiation of translation of theerm transcripts. Mutational analyses of theerm(C) leader peptide suggested that the peptide, (FS)IFVI, is critical for induction (103). However, when theerm peptides from the erm genes are compared, little sequence similarity is apparent (103). Recently, a second mechanism of regulation has been described in which the lack of erythromycin prevents the complete synthesis of the mRNA due to rho factor-independent termination. This type of regulation has been described for the erm(K) system (20), and by homology, we hypothesize that it may also exist for erm(D), as well as erm(J), because they are highly related and have been grouped together under class D (Table 1). In either system, inducible isolates, when tested, may appear to be susceptible or intermediately resistant to macrolides and susceptible to lincosamides. Erythromycin is generally a good inducer in most species; in animal or human streptococcal isolates, lincosamides and/or streptogramin B may be good inducers (47, 76). Good overviews of regulation of the erm genes can be found in recent reviews by Weisblum (100-103).
Inducible strains predominated in the 1960s to 1970s. However, today it is more common in many geographical areas to find isolates that constitutively produce the rRNA methylase without preexposure to antibiotics. Constitutive erm gene expression is usually associated with structural alterations in the ermtranslational attenuator, including deletions, duplications, and point mutations in erm(C) (104). They can be distinguished from inducible isolates by the stable MICs for them regardless of whether they are pregrown with or without an inducer (76, 102).
Many of the erm genes are associated with conjugative or nonconjugative transposons which tend to reside on the chromosomes, although some have been found in plasmids. They are often associated with other antibiotic resistance genes, especially tetracycline resistance genes. The erm(F) gene is often linked with thetet(Q) gene, while the erm(B) gene is often linked with the tet(M) gene (24, 86, 95). These conjugative transposons can have a wide host range, which may explain why clinical isolates of many different bacterial species have been found to carry these erm genes (Table 3). The ermgenes in general have low G+C contents (31 to 34%), while the overall chromosomal G+C contents found in gram-negative species are ≥50% and ∼35% in gram-positive species.
It has been common practice for investigators to give theirerm gene a new name regardless of the DNA and predicted amino acid sequence similarity to previously characterizederm genes and without regard to whether the gene resides in a different isolate, species, or genus. The result has been that, over the years, the names of these erm genes have become confusing, and often a complex table is required to remember which genes are closely related (Table 1). In the worst cases, genes for unrelated enzymes have been given the same name (erm(A), causing confusion in the literature and GenBank listings (Table 1). The opposite also has occurred where very closely related or virtually identical enzymes have been given a variety of different names. For example, erm(F) (GenBank no. M14730) is found on theBacteroides transposons Tn4351 and Tn4000 (71), erm(FS) (no. M17808) is on Bacteroides transposon Tn4551 (91), and erm(FU) (no. M62487) (32) is also fromBacteriodes. All three enzymes share ≥97% DNA and amino acid identity (Table 1). Since there are no phenotypic differences between the three erm(F) genes and distinguishing them by any method other than sequencing is problematic, we propose that all three should be known as class F: the Erm(F) protein and theerm(F) gene (Table 1).
The situation is even worse with class B, which is composed of a larger number of genes, including erm(AM), erm(B),erm(BC), erm(BP), and erm(Z), whose sequences share ≥98% homology (Table 1). Because the normal gene designation is to use a single letter (26) and the possibility of confusion between erm(A) anderm(AM), we propose that this group be known as class B: theerm(B) genes and the Erm(B) protein (Table 1). Recent dendrograms of many of the erm genes can be found in articles by Seppälä et al. (88) and Matsuoka et al. (56) and support this grouping of all of these genes within the class B designation.
To help those in the field, GenBank numbers or references for sequences that have not been deposited are listed in Table 1. If a new gene sequence shows ≥80% amino acid homology to any member of a gene class and confers a similar phenotype to the host, we propose that the new gene be placed in the existing group and not be given a new letter or number designation. Thus, with classes that show a wide range of homologies, like class A (81% amino acid homology) or class O (84% amino acid homology), multiple members must be compared to the new gene. Note that the class designation is based on the amino acid sequence of the structural gene only and does not include the various regulatory sequences that can occur upstream of the gene. These guidelines are intended to apply to all of theN-methyltransferases, regardless of whether the gene was originally identified in pathogenic, opportunistic, normal flora bacteria or an antibiotic-producing species. Once all of the single capital letters have been used to identify new erm genes, we recommend naming genes as follows: erm(30),erm(31), etc. This system has been proposed for naming of new tet genes [tet(30), etc.] (52). Furthermore, a similar set of guidelines should be adopted for the genes that encode other mechanisms of resistance to any of the MLS antibiotics (Table 1). Class Y for gene erm(GM), class S for gene erm(SF), class T for geneerm(GT), class V for gene erm(SV), class X for genes erm(CD), erm(CX), and erm(A), and class 2 for gene srm(D) are new class designations that conform to the single-letter designation (Table 1).
There are a number of other methylase genes, most often found in methylase-producing organisms which have not been given ermdesignations, such as tlr(D), car(B),myr(B), and smr(A). All are from species which confer resistance to a 16-membered ring macrolide (Table 1). We have grouped and renamed them classes H for car(B), I formdm(A), N for tlr(D), O for genes lrmand srm(A), U for lmr(B), and W formyr(B). The clr gene could not be classified, because there is no sequence in the database or literature available. Less work has been done to determine if these genes are found outside their respective antibiotic producers (Table 3). erm genes are often linked with tet genes, and since genes conferring resistance to oxytetracycline, originally found in antibiotic-producing streptomycetes, are now found in some clinical Mycobacteriumisolates, it is certainly possible that some erm genes have also moved into Mycobacterium spp. and other genera (68).
To prevent two unrelated genes from being given the same designation, we propose to establish a reference center, as has recently been recommended for tetracycline resistance genes. By using the guideline presented above in governing the identification of new ermgenes, surveys can be conducted in bacterial populations to examine the spread of particular MLSB-resistant determinants. A single internal DNA fragment or oligonucleotide probe or a PCR assay that detects all members of a gene class can be established to screen large numbers of isolates. Not only will the adoption of a uniform naming system reduce the number of new erm gene names, but it will hopefully prevent confusion over unrelated genes being given the same designation and also prevent highly related genes from having different gene designations.
EFFLUX SYSTEMS
A number of different antibiotic resistance genes code for transport (efflux) proteins. These do not modify either the antibiotic or the antibiotic target, but instead pump the antibiotic out of the cell or the cellular membrane, keeping intracellular concentrations low and ribosomes free from antibiotic. Many of these proteins [mef(A), mef(E), andlmr(A)] have homology to the major facilitator superfamily (MFS) of efflux proteins. Others [car(A),msr(A), msr(B), ole(B),ole(C) srm(B), tlr(C), vga, and vga(B)] are putative members of the ABC transporter superfamily (70). In early years, most macrolide resistance was mediated by the presence of erm genes. However, more recently, other mechanisms of macrolide resistance have been found in increasing frequency in certain gram-positive populations (23, 27,41, 43, 44, 92, 93, 106). Three different efflux systems which confer resistance have been described for gram-positive cocci [msr(A) (macrolide and streptogramin B resistant),mef(A) (macrolide efflux), and vga andvga(B) (virginiamycin factor A)] (4) (Table 2). Besides the academic interest in these genes, their presence in an erythromycin-resistant bacterial pathogen of interest may also have implications in terms of therapeutic choices. If an isolate carries amef gene, clindamycin can be considered, whereas the presence of an erm(B) gene would preclude consideration of a lincosamide. Recently, we and others have identifiedStreptococcus pneumoniae strains which carry bothmef and erm(B) genes and, as expected, have the MLSB phenotype (41, 53).
The mef genes have been found in a variety of gram-positive genera, including corynebacteria, enterococci, micrococci, and a variety of streptococcal species (30, 43, 53, 90) (Table3), suggesting a much wider distribution of this group of genes than originally imagined. Many of these genes are associated with conjugative elements located in the chromosome and are readily transferred conjugally across species and genus barriers (43,53).
Two mef genes have been characterized in the literature:mef(A) (23) and mef(E) (94). The mef(A) gene was described inStreptococcus pyogenes, while the mef(E) gene was found in S. pneumoniae. Since the two genes share 90% DNA and 91% amino acid homology (Table 2), we recommended that these two genes be considered a single class, A: mef(A) gene and Mef(A) protein (Table 2).
The msr(A), msr(SA), msr(SA)′, andmsr(B) group differs from the mef genes because they confer resistance to both macrolide and streptogramin B antibiotics (MS) (13, 55-57). The msr(B) gene is roughly half the size of msr(A), but very homologous to it. Though this gene is significantly shorter than the msr(A) gene sequence, we placed it with the other msr genes (Table2).
In antibiotic producers, there are efflux pumps specific for MLSB antibiotics that generally belong to the ABC transporter superfamily (87). They include car(A) from Streptomyces thermotolerans (87),ole(B) from Streptomyces antibioticus (7,80), srm(B) from Streptomyces ambofaciens(73), lmr(C) from Streptomyces lincolnensis (70), and tlr(C) fromStreptomyces fradiae (87). In addition to themsr(A) efflux pumps, there are two efflux systems identified in staphylococci that confer resistance to streptogramin A antibiotics,vga and vga(B) (4). Besidesmef(A), other efflux proteins that appear to be fueled by the proton motive force have been described for MLSBantibiotics. A lincomycin-specific efflux pump encoded bylmr(A) has been described in S. lincolnensis(110).
OTHER MECHANISMS
A variety of other mechanisms which usually confer resistance to only one of the three classes (M, L, or S) or one component such as streptogramin A, but not streptogramin B, have been described (103) (Table 2). These proteins modify the antibiotic rather than the rRNA target or serve as pumps that shuttle the antibiotic out of the bacterial cell. Enzymes which hydrolyze streptogramin B [vgb (virginiamycin factor B hydrolase), vgb(B) genes] or modify the antibiotic by adding an acetyl group (acetyltransferases) to streptogramin A [vat(virginiamycin, factor A acetylation), vat(B),vat(C), sat(A), and sat(G) genes] have been described (1-6) (Table 2). Many of these genes are plasmid borne, and often these vat-related genes [vat, vat(B), and vat(C) genes] are downstream of other genes encoding resistance to streptogramins [vgb, vga(B), and vgb(B) genes, respectively] in staphylococci (2), but not in enterococci (72). The acetyltransferase genes are related, in the active site region, to a novel chloramphenicol acetyltransferase family of enzymes. We have renamed sat(A) as vat(D) andsat(G) as vat(E) to simplify the nomenclature (Table 2).
Unlike most of the other genes described in this review, both theere (erythromycin esterification) and mph(macrolide phosphotransferase) genes (Table 2) were first described inE. coli rather than gram-positive cocci (8, 63, 64, 66,67). According to our guidelines, mph(K) has been reassigned to mph(A), because there are only 10 amino acid (1%) differences between the two proteins. mph(BM) andmph(C) (66a) are grouped under Mph(C), because these genes are nearly identical to each other and distinct frommph(A) and mph(B) (Table 2). Several lincomycin nucleotidyltransferases have been identified: lin(A) inStaphylococcus haemolyticus (16),lin(A)′ in Staphylococcus aureus (17), and lin(B) in Enterococcus faecium(14). We propose changing lin(A) andlin(B) to lnu(A) and lnu(B) (for lincomycin nucleotidyltransferase), because the former letters have already been used for gamma BHC dehydrochlorinase and cyclohexadiene hydrolase genes. It is suggested that prior to naming a new gene class, it is necessary to determine if the proposed three-letter designation has been used for other previously characterized genes.
CONCLUSIONS
With the introduction of the newer, more stable macrolides with enhanced properties, there has been a significant increase in macrolide usage. Macrolides like azithromycin and clarithromycin are recommended for prophylactic use to prevent Mycobacterium avium complex disease in human immunodeficiency virus patients. As macrolide use increases, so does its exposure to bacterial populations, increasing the opportunity for bacteria to acquire macrolide or MLS resistance. Given that intragenic transfer of macrolide-resistant determinants is possible (15), it is likely that all of the genes described in this review will spread into new species and that new genes will be identified. Therefore, it is important to clarify the nomenclature of these resistance genes for their expanding audience.
ACKNOWLEDGMENTS
We thank M. Matsuoka for providing unpublished material; B. Weisblum for discussions; J. Davies, C. J. Smith, and S. Schwarz for reading the manuscript; and S. Lerner for doing sequence comparisons.
- Copyright © 1999 American Society for Microbiology
REFERENCES
- 1.↵
- 2.↵
- 3.↵
- 4.↵
- 5.↵
- 6.↵
- 7.↵
- 8.↵
- 9.↵
- 10.
- 11.
- 12.
- 13.↵
- 14.↵
- 15.↵
- 16.↵
- 17.↵
- 18.
- 19.
- 20.↵
- 21.↵
- 22.
- 23.↵
- 24.↵
- 25.
- 26.↵
- 27.↵
- 28.
- 29.
- 30.↵
- 31.
- 32.↵
- 33.
- 34.
- 35.
- 36.↵
- 37.
- 38.
- 39.
- 40.
- 41.↵
- 42.
- 43.↵
- 44.↵
- 45.
- 46.
- 47.↵
- 48.
- 49.
- 50.
- 51.↵
- 52.↵
- 53.↵
- 54.
- 55.↵
- 56.↵
- 57.↵
- 58.↵
- 59.
- 60.
- 61.
- 62.
- 63.↵
- 64.↵
- 65.
- 66.↵
- 66a.↵
- 67.↵
- 68.↵
- 69.
- 70.↵
- 71.↵
- 72.↵
- 73.↵
- 74.
- 75.
- 76.↵
- 77.↵
- 78.↵
- 79.
- 80.↵
- 81.
- 82.
- 83.
- 84.↵
- 85.
- 86.↵
- 87.↵
- 88.↵
- 89.
- 90.↵
- 91.↵
- 92.↵
- 93.↵
- 94.↵
- 95.↵
- 96.↵
- 97.↵
- 98.
- 99.
- 100.↵
- 101.↵
- 102.↵
- 103.↵
- 104.↵
- 105.
- 106.↵
- 107.↵
- 108.
- 109.
- 110.↵
