Macrolide Resistance in Staphylococcus aureus: Inducers of

Transcription

1 ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Apr. 1977, p Copyright 1977 American Society for Microbiology Vol. 11, No. 4 Printed in U.S.A. Macrolide Resistance in Staphylococcus aureus: Inducers of Macrolide Resistance NORRIS E. ALLEN The Lilly Research Laboratories, Eli Lilly and Company, Indianapolis, Indiana Received for publication 12 October 1976 Several macrolide-, lincosamide-, and streptogramin B-type (MLS) antibiotics were tested as inducers of erythromycin A (EM)-resistant [C]leucine incorporation. Only -membered-ring macrolides having a glycosidically linked 6-deoxy sugar at the C-3 position of the lactone ring and the structurally dissimilar lincosamide, celesticetin, showed inducer activity. Modifications of EM at the C- 4" position of cladinose can apparently destroy the inducer property but do not affect the inhibitory properties of the antibiotic. The findings clearly show that inducer and inhibitor activities can be dissociated and are consistent with the concept that distinct binding/receptor sites are utilized for inhibition of ribosome function and induction of resistance. Macrolide resistance in certain strains of Staphylococcus aureus can be induced by exposure to subinhibitory concentrations of erythromycin A (EM) (1, 19). Several antibiotics besides EM reportedly induce macrolide resistance in S. aureus. For example, oleandomycin (7, 8), lankamycin (7), desacetyllankamycin (7), leucomycin (11), carbomycin (), and lincomycin () have been reported to induce resistance in strains isolated from diverse sources. Among clinical isolates, published descriptions of inducibly resistant S. aureus indicate only EM, oleandomycin, the lankamycins, and megalomycin as inducers of macrolide resistance (5, 7-9, ). These compounds are all -membered-ring macrolide antibiotics. In contrast, mutants of clinical strains that can be induced by diverse MLS antibiotics, including - and 16-membered-ring macrolides (11) and lincomycin (), have been isolated in the laboratory. One mutant is described as inducible by high temperature (). Apparently, both strain and drug specificities are determining factors in the induction of macrolide resistance. EM and its derivatives are the only reported inducers of macrolide resistance in S. aureus 1206 (16, 19). It was recently reported () that a mutant of this strain had lost EM inducibility but could be induced by either lincomycin or carbomycin. This report, coupled with the observation that celesticetin and several macrolides other than EM induced zone distortion when placed adjacent to lincomycin (1), suggested that several antibiotics might induce macrolide resistance in S. aureus MLS antibiotics and EM derivatives were surveyed for ability to induce EM-resistant [0]leucine incorporation. The results indicate 669 that several -membered-ring macrolides and celesticetin induce resistance in S. aureus Furthermore, a single modification of the EM molecule appears to eliminate inducer capability without affecting the inhibitory properties of the antibiotic. MATERIALS AND METHODS The inductive capacity of MLS antibiotics was measured by growing S. aureus 1206 to early stationary phase in synthetic broth as described (1). Cells were washed once and resuspended in fresh broth to a cell density of approximately 0.24 mg of protein per ml. After preincubation at 370C for 30 min, inducers were added at the desired concentration. Incubation of the cell suspension was continued for 2 h, and [C]leucine incorporation in the presence and absence of 0,uM EM was measured on duplicate aliquots exactly as described before (1). Each of the antibiotics tested for inducer activity was also assayed for inhibition of [Clleucine incorporation as described (1), using cells incubated for 2 h without added antibiotics. Sources of the antibiotics used have been cited previously (1), and EM derivatives were kindly furnished by E. Massey of the Lilly Research Laboratories. Antibiotic stock solutions were routinely prepared in methanol-water at 1 to 5 mg/ml. The maximum level of methanol present during induction was 0.1%, and there was no inhibition of [C]leucine incorporation or induction by methanol at this concentration. The EM derivatives were generally less soluble than the other antibiotics and were dissolved in undiluted methanol. Equivalent amounts of methanol were added to the untreated controls. RESULTS MLS antibiotics. MLS antibiotics were tested at 0.,, and,um for their ability to induce EM-resistant [0]leucine incorporation

2 670 ALLEN (Table 1). The inhibitory activity of each compound was simultaneously determined at,, and 0 AM. Induced EM-resistant incorporation is expressed as a percentage of the uninhibited rate since exposure to many compounds for 120 min resulted in lowered rates of ['4C]leucine incorporation. The inhibitory effects of 16-membered-ring macrolides at,tm precluded any estimate of induction. The data in Table 1 were accumulated from None TABLE 1. Inducer activity of MLS antibiotics ANTIMICROB. AGENTS CHEMOTHER. five separate experiments. In each experiment, EM was tested as an inducer along with an uninduced control, and standard errors are included with these data. In the absence of exogenous inducer, EM-resistant incorporation was approximately 16% of the uninhibited rate. An increase in resistant incorporation above this level indicates induction by the given antibiotic. Among the -membered-ring macrolides, Antibiotic Concn Inducer 1 Inhibitor Antibiotic 1 Concn InducerT A.M) activitya activity" (AM) activitya Inhibitor activity" -membered-ring crolides Erythromycin A Oleandomycin Megalomycin Lankamycin Desacetyllankamycin Narbomycin Pikromycin ma c ± ± ± ma- 12-membered-ring crolides Methymycin ma- 16-membered-ring crolides Carbomycin Spiramycin Tylosin Niddamycin Triacetyloleandomycin Streptogramin B types Vernamycin Ba Lincosamides Lincomycin Celesticetin d ainducer activity is the rate of EM-resistant ['4Clleucine incorporation expressed as a percentage of the uninhibited rate. Each antibiotic listed was tested for inducer activity at three concentrations as described in Materials and Methods. After the 120-min induction, cells were washed free of inducer and incorporation of [C]leucine was measured in the presence and absence of 0 FM EM as described (1). The rate of [C]leucine incorporation in the presence of 0 IAM EM, expressed as a percentage of the rate of incorporation in the absence of EM, reflects the extent of induction by each antibiotic. Uninduced, control cells were incubated 120 min without added antibiotics. [C]leucine incorporation by these cells in the presence of 0 um EM was % of the uninhibited rate; i.e., incorporation was inhibited by %. binhibitor activity is the percentage of inhibition of [C]leucine incorporation. Eaeh antibiotic listed was tested for inhibition of protein synthesis at three concentrations as described (1), using cells incubated for 120 min without added inducer. e Standard errors from five determinations. d, No measurement of EM-resistant [C]leucine incorporation due to inhibitory effects of inducer. 7 11,

3 VOL. 11, 1977 EM, oleandomycin, triacetyloleandomycin, megalomycin, lankamycin, and desacetyllankamycin showed inducer activity. Griseomycin was not tested. EM was clearly the most potent inducer and was also the most active inhibitor of [C]leucine incorporation. The decreased levels of induced resistance at EM concentrations exceeding 0.,uM resulted from inhibition of the induction process (). Narbomycin and pikromycin both failed to induce resistance, but they also failed to inhibit [C]leucine incorporation. Induced resistance to both compounds was readily demonstrated by the disk test (1), implying that they inhibit ribosome function as do other MLS antibiotics (). Additional experiments with narbomycin (data not presented) showed no inductive capacity even at 700,uM. Methymycin, a 12-membered-ring macrolide, also failed to induce resistance or inhibit [0]leucine incorporation, but higher concentrations were not evaluated. No 16-membered-ring macrolide induced macrolide resistance. At,uM, most of these compounds destroyed greater than 80% of the [C]leucine-incorporating activity. Because these antibiotics were such potent inhibitors of protein synthesis, it was possible that even at the lowest concentration tested their capacity to inhibit exceeded their capacity to induce. In further testing, lower noninhibitory concentrations (0.00 and 0.0 j,m) of carbomycin and tylosin were nevertheless unable to induce resistance (data not shown). Neither vernamycin Ba nor lincomycin induced resistance, but celesticetin, a derivative of lincomycin (6), showed good inducer activity. Although not as active as an inducer as EM at 0.,uM, celesticetin appeared to be a more effective inducer at higher concentrations since its ability to inhibit protein synthesis and thus induction was less than that of EM. EM derivatives. Eighteen derivatives of EM were compared for inducer activit using the disk method (). The derivatives are listed in Table 2 and the structure ofem is given in Fig. 1. Each of the derivatives demonstrated antibacterial activity when tested against Sarcina lutea by the disk diffusion test (data not shown). All but two compounds induced macrolide resistance in S. aureus 1206; the 4"- tosyl and 4"-mesyl derivatives of the cyclic carbonate were the only compounds unable to induce noticeable zone distortion. Several derivatives, including the 4"-tosyl cyclic carbonate, were compared as inducers of EM-resistant [C]leucine incorporation (Table 3). The unmodified cyclic carbonate was a strong inhibitor of protein synthesis and was also a potent inducer. Further modification INDUCERS OF MACROLIDE RESISTANCE 671 yielding the 4"-tosyl derivative preserved the inhibitory properties of this molecule but resulted in a derivative unable to induce resistance at the concentrations tested. The comparable mesyl derivative was not tested further, but all other derivatives tested showed inducer activity. DISCUSSION MLS antibiotics constitute a diverse and structurally dissimilar group of compounds. The macrolides contain a characteristic 12-, -, or 16-membered-ring lactone with one, two, or three 6-deoxy sugar substitutions. The size of the lactone ring and the number and position of the attached deoxy sugars are used to classify these antibiotics. With the exception of possible similarities between EM and lincomycin based on a comparison of molecular models (16), the macrolides are structurally distinct from the lincosamides and the streptogramin B-type antibiotics. Although structurally diverse, the MLS antibiotics demonstrate functional relatedness. Competition between lincomycin and EM for binding to the same, overlapping, or related sites on the ribosome is well established (2, 6). In S. aureus, induced EM resistance is accom. panied by resistance to all MLS antibiotics (1, 16), and a constitutive mutation confers resistance to the same drugs (1). The present study reveals that several macrqlide antibiotics plus celesticetin can induce EM resistance in S. aureus Thus, the functional relatedness of these antibiotics is extended to include inducer as well as inhibitor function. Work from other laboratories (8, 10, 12) has indirectly suggested that the inductive capacity of macrolide antibiotics may be due, in part, to the structure and position of the 6-deoxy sugars. The results from the present study support this idea and further suggest a critical role for the sugar attached at the C-3 position of the lactone ring, e.g., cladinose in EM (Fig. 1). The evidence for the importance of this sugar and the size of the lactone ring in induction of resistance ins. aureus by macrolide antibiotics is considerable. First, only -membered-ring macrolides containing a C-3-linked 6-deoxy sugar were active as inducers. No 12- or 16-membered-ring macrolide tested induced zone distortion (1) or EM-resistant [C]leucine incorporation (Table 1). The 12- and 16-membered-ring macrolides have different sugar arrangements and no C-3- linked deoxy sugar. Second, the two -membered-ring macrolides that did not induce, narbomycin and pikromycin, are unique in not containing an attached 6-deoxy sugar at the C-3

4 672 ALLEN ANTIMICROB. AGENTS CHEMOTHER. DESOSAMINE CH3CH2 Erythromycin A Erythromycin B Erythromycin C N-oxide Des-N-methyl 8-Hydroxy 9-Hydrazone 9-Ketoxime TABLE 2. CH ( -'OCH3 FIG. 1. Structure of erythromycin A. 9(s)-Dihydro 9(s)-Erythromycylamine 9(r)-Erythromycylamine Erythromycin derivatives 6,9-Hemiketal 6,9,12-Spiroketal 11,12-Cyclic carbonate 11,12-Cyclic carbonate- 8-hydroxy 2'-Benzoyl 3'-De(dimethylamino) 4"-Tosyl-11,12-cyclic carbonate 4"-Mesyl-11,12-cyclic carbonate CLADINOSE position. These two antibiotics differ only by a single hydroxyl on the lactone ring. It should be pointed out that interpretation of the negative inducer response is complicated by the inability of both antibiotics to inhibit [Clleucine incorporation, although neither compound behaved as an inducer in the disk test. Third, both lankamycin and desacetyllankamycin showed inducer activity, but the desacetyl form was a more active inducer than the acetylated antibiotic. Lankamycin is acetylated at the C-4" position of arcanose, which is the 6- deoxy sugar attached to the lactone ring at the C-3 position. Omura et al. (7) reported identical findings for these two antibiotics using another method to measure induction. Fourth, using the disk method to detect inducer activity among EM derivatives, all but two compounds were strong inducers. The 4"- tosyl and 4"-mesyl derivatives of EM cyclic carbonate were unable to induce zone distortion. The tosylated derivative failed to induce EM-resistant [C]leucine incorporation, whereas the unmodified cyclic carbonate was as active an inducer as EM (Table 2; 8). Furthermore, inhibition of protein synthesis by the cyclic carbonate was equal to or greater than that shown by EM. Since the 4"-tosyl cyclic carbonate retains inhibitory activity, it is clear that reduced transport cannot account for the reduction in inducer activity. Thus, the loss of inductive capacity can be attributed to the modification at the C-4" position of cladinose (see Fig. 1). Other workers have shown that the monoglycosides of EM (minus cladinose or desosamine) TABLE 3. Inducer activity ofem derivatives Derivative Concn (MM) Inducer Inhibitor activitya activityb None 0 Erythromycin A '-Benzoyl '-De(dimethylamino) ,12-Cyclic carbonate '-Tosyl-11,12-cyclic 0. carbonate a. b See Table 1.

5 VOL. 11, 1977 and the aglycone (minus both sugars) were either inactive or had severely diminished inducer activity (8, 10, 12). Moreover, Ono et al. (8) reported that introduction of an acetyl group in the C-4" position of EM resulted in decreased inducer activity but had a negligible effect on antibacterial activity. Taken together, the data clearly indicate that inducer and inhibitor activities of EM can be dissociated. Fifth, induction by leucomycin using the mutant described by Saito et al. (11) was affected by modification of mycarose at the C-4" position. Although leucomycin is a 16-membered-ring macrolide and the sugar arrangement is not the same as in EM, removal of mycarose or deacetylation at the C-4" appeared to eliminate inducer activity. Celesticetin's ability to induce macrolide resistance was unexpected since this antibiotic is structurally dissimilar to EM and other - membered-ring macrolides. Whether a structural similarity exists between celesticetin and EM analogous to the similarity between lincomycin and EM proposed to explain identical or related ribosome binding sites (16) remains to be determined. At any rate, induction by celesticetin means that inducibility in S. aureus is not limited to a narrow spectrum of closely related structural types and proves that induction of resistance in natural isolates of S. aureus is not strictly a property of macrolides. A closer similarity with inducibly macrolide-resistant streptococci, some of which can be induced by numerous MLS antibiotics (3, 4, 20), is suggested. The fact that only MLS antibiotics will induce raises the interesting possibility that the binding or receptor site(s) on the hypothetical regulatory protein may recognize the same molecular structures as does the ribosomal binding site(s). Tanaka and Weisblum () described a mutant of S. aureus that had lost EM inducibility but that could be induced by either lincomycin or carbomycin. They theorized that this phenotype may have resulted from a reversal of the relative potency these antibiotics had for inhibition and induction. In other words, other MLS antibiotics may be potential inducers, but fail because their capacity to inhibit is so much greater. The finding that celesticetin can induce resistance in S. aureus is consistent with this proposal, since although higher concentrations of this antibiotic are required for induction, celesticetin's capacity to inhibit ribosome function appears to be much less than that of EM. The inability of the 4"-tosyl cyclic carbonate of EM to induce resistance is further evidence that modification can alter the relative INDUCERS OF MACROLIDE RESISTANCE 673 potency an antibiotic has for induction and inhibition. Finally, clinically isolated inducibly macrolide-resistant staphylococci have been classified on the basis of whether they could be induced by EM only or by EM and oleandomycin (5, 8). EM was previously reported as the only inducer in S. aureus 1206 (16). It can be concluded from the studies reported here that numerous MLS antibiotics can potentially induce resistance but go undetected because they have a much stronger affinity for ribosome inhibition or because detection may require a sensitive measure of induction. The possibility that mutations can lead to a changed inducer spectrum coupled with the disclosure of inducible macrolide resistance in streptococci (3, 4, 20) warns that any classification of strains based solely on response to inducers would be an oversimplification at this time. ACKNOWLEDGMENTS I thank Ken Christy and Janet Epp for their assistance in portions of this work. LITERATURE CITED 1. Allen, N. E Macrolide resistance in Staphylococcus aureus: induction of macrolide-resistant protein synthesis. Antimicrob. Agents Chemother. 11: Chang, F. N., and B. Weisblum The specificity of lincomycin binding to ribosomes. Biochemistry 6: Hyder, S. L., and M. M. Streitfeld Inducible and constitutive resistance to macrolide antibiotics and lincomycin in clinically isolated strains ofstreptococcus pyogenes. Antimicrob. Agents Chemother. 4: Malke, H Genetics of resistance to macrolide antibiotics and lincomycin in natural isolates of Streptococcus pyogenes. Mol. Gen. Genet. 5: Mitsuhashi, S., H. Oshima, M. Inoue, and Y. Yamaguchi Resistance to macrolide antibiotics and lincomycin in Staphylococcus aureus, p In S. Mitsuhashi (ed.), Drug action and drug resistance, vol. I: Macrolide antibiotics and lincomycin. University Park Press, Baltimore. 6. Monro, R. E., R. Fernandez-Munoz, M. L. Celma, and D. Vazquez Mode of action of lincomycin and related antibiotics, p In S. Mitsuhashi (ed.), Drug action and drug resistance in bacteria, vol. I: Macrolide antibiotics and lincomycin, University Park Press, Baltimore. 7. Omura, S., S. Namiki, M. Shibata, T. Muro, and J. Sawada Studies on the antibiotics from Streptomyces spinochromogenes var. kujimyceticus. V. Some antimicrobial characteristics of kujimycin A and kujimycin B against macrolide resistant staphylococci. J. Antibiot. 23: Ono, H., M. Inoue, J. C. H. Mao, and S. Mitsuhashi Drug resistance in Staphylococcus aureus. Induction of macrolide resistance by erythromycin, oleandomycin and their derivatives. Jpn. J. Microbiol. 19: Osono, T., and H. Umezawa Josamycin, a new

6 674 ALLEN ANTIMICROB. AGENTS CHEMOTHER. macrolide antibiotic of resistance non-inducing type, p In S. Mitsuhashi (ed.), Drug action and drug resistance in bacteria, vol. I: Macrolide antibiotics and lincomycin. University Park Press, Baltimore. 10. Pestka, S., R. Vince, R. LeMahieu, F. Weiss, L. Fern, and J. Unowsky Induction of erythromycin resistance in Staphylococcus aureus by erythromycin derivatives. Antimicrob. Agents Chemother. 9: Saito, T., H. Hashimoto, and S. Mitsuhashi Macrolide resistance in Staphylococcus aureus. Isolation of a mutant in which leucomycin is an active inducer. Jpn. J. Microbiol. : Saito, T., M. Shimizu, and S. Mitsuhashi Macrolide resistance in staphylococci, p In S. Mitsuhewhi (ed.), Drug action and drug resistance in bacteria, vol. I: Macrolide antibiotics and lincomycin. University Park Press, Baltimore.. Shimizu, M., T. Saito, and S. Mitsuhashi Macrolide resistance in Staphylococcus aureus. Relation between spiramyciir-binding to ribosome and inhibition of polypeptide synthesis in a heat inducible-resistant mutant. Jpn. J. Microbiol. : Tanaka, T., and B. Weisblum Mutant ofstaphylococcus aureus with lincomycin- and carbomycininducible resistance to erythromycin. Antimicrob. Agents Chemother. 5: Weaver, J. R., and P. A. Pattee Inducible resistance to erythromycin in Staphylococcus aureus. J. Bacteriol. 88: Weisblum, B Macrolide resistance in Staphylococcus aureus, p In S. Mitsuhashi (ed.), Drug action and drug resistance in bacteria, vol. I: Macrolide antibiotics and lincomycin. University Park Press, Baltimore.. Weisblum, B., and J. Davies Antibiotic inhibitors of the bacterial ribosome. Bacteriol. Rev. 32: Weisbium, B., and V. Demohn Erythromycininducible resistance in Staphylococcus aureus: survey of antibiotic classes involved. J. Bacteriol. 98: Weisbium, B., C. Siddhikol, C. J. Lai, and V. Demohn Erythromycin-inducible resistance in Staphylococcus aureus. Requirements for induction. J. Bacteriol. 106: Yagi, Y., A. E. Franke, and D. B. Clewell Plasmid-determined resistance to erythromycin: comparison of strains of St;eptococcus faecalis and Streptococcus pyogenes with regard to plasmid homology and resistance inducibility. Antimicrob. Agents Chemother. 7: