Derivative, 4'-Deoxy, 6'-N-Methylamikacin

Transcription

1 ANTIMICROBLAL AGENTS AND CHEMOTHERAPY, Apr. 1981, p /81/ $02.00/0 Vol. 19, No. 4 Resistance to Antibiotic Synergism in Streptococcus faecalis: Further Studies with Amikacin and with a New Amikacin Derivative, 4'-Deoxy, 6'-N-Methylamikacin STEPHEN B. CALDERWOOD, CHRISTINE WENNERSTEN, AND ROBERT C. MOELLERING, JR.* Infectious Disease Unit, Department ofmedicine, Massachusetts General Hospital, Boston, Massachusetts 02114,1 and Harvard Medical School, Boston, Massachusetts Received 7 October 1980/Accepted 26 January 1981 Streptococcus faecalis strains may resist penicillin-aminoglycoside synergy by the production of plasmid-mediated aminoglycoside-modiffying enzymes. One, of these enzymes, aminoglycoside 3'-phosphotransferase, has been shown to have a broad range of substrate specificity, including amikacin. We have studied a derivative of amikacin, 4'-deoxy, 6'-N-methylamikacin (BB-K311), against 11 clinical blood isolates of S. faecalis. Minimal inhibitory concentrations of BB- K311 were quite similar to those of amikacin, ranging from 125 to 1,000,tg/ml. In assays for antibiotic synergy, penicillin and amikacin produced enhanced killing compared with penicillin alone only against those three strains which lacked the phosphotransferase enzyme. The other eight enzyme-positive strains actually demonstrated significant antagonism between penicillin and amikacin. In contrast, BB-K311 showed enhanced killing with penicillin against all 11 strains, regardless of enzyme production. Analysis of substrate profiles with crude preparations of the aminoglycoside 3'-phosphotransferase enzyme confirmed that BB-K311 was a very poor substrate for modification, as expected from the synergy studies. Use of other aminoglycoside analogs confirmed the 3'- site of modification. These findings suggest that removing the 4'- group in amikacin effectively blocks 3'- phosphorylation by the S. faecalis enzyme. Enterococci differ from other streptococci in two important aspects of antibiotic response. The minimal inhibitory concentrations (MICs) of beta-lactam antibiotics for these organisms are substantially higher than for other streptococci (18, 29). In addition, even at antibiotic concentrations in excess of the MIC, enterococci demonstrate tolerance to the usual killing effect of all cell wall-active agents (16). Because of these two properties of antibiotic response, treatment of serious enterococcal infections, particularly bacterial endocarditis, is often ineffective with a single agent (8, 17). The combination of penicillin and streptomycin was shown to produce a synergistic bactericidal effect against enterococci more than three decades ago (9, 11). Many of the newer aminoglycosides have also been shown to produce this effect when combined with penicillin (12, 20, 21). Not all enterococcal strains, however, demonstrate such antibiotic synergy (1, 28, 31). Recent work on the mechanisms by which enterococci may resist penicillin-aminoglycoside synergy has confirmed the presence of plasmid-mediated aminoglycoside-modifying enzymes in such strains (3, 14, 15). Streptococcus faecalis was shown to produce an aminoglycoside adenylyltransferase, mediating resistance to penicillinstreptomycin synergy, and an aminoglycoside 3'- phosphotransferase [APH(3')], mediating resistance to penicillin-kanamycin and penicillin-amikacin synergy. We have examined the activity of an amikacin derivative, 4'-deoxy, 6'-N-methylamikacin (BB- K311), against 11 blood isolates of S. faecalis. We were interested in determining whether such a modification would block phosphorylation at the 3' site and enable this compound, when combined with penicillin, to produce synergistic killing of organisms resistant to penicillin-amikacin synergy. During the course of the study, we encountered an unexpected phenomenon, the first known example with a clinical isolate of antagonism between penicillin and an aminoglycoside. This report describes the results of these experiments. (This work was presented at the 20th Interscience Conference on Antimicrobial Agents and Chemotherapy, September, 1980, New Orleans, La.) MATERIALS AND METHODS Strains and media. Eleven blood isolates of S. faecalis were obtained from patients hospitalized at 549

2 550 CALDERWOOD, WENNERSTEN, AND MOELLERING ANTIMICROB. AGENTS CHEMOTHER. the Massachusetts General Hospital, Boston. These were stored as stock cultures in soft agar medium without antibiotics at -70 C. Strain JH2-7 was kindly provided by Alan Jacob (10). Identification of isolates as enterococci was by standard bacteriological techniques, and identification to the species level was according to the recommendations of Facklam (6). Before use, the stock cultures were subcultured on brucella agar containing 5% horse blood, and singlecolony isolates were used for all further steps. Unless specifically indicated, all incubations were at 37 C. Antibiotics. 4'-Deoxy, 6'-N-methylamikacin (BB- K311) was kindly provided by Kenneth Price and Felix Leitner, Bristol Laboratories Division, Bristol-Myers Co., Syracuse, N.Y. Antibiotics used for MIC determinations included potassium penicillin G (Pfizer, Inc., New York, N.Y.), streptomycin sulfate (Eli Lilly and Co., Indianapolis, Ind.), kanamycin A sulfate and amikacin sulfate (Bristol Laboratories Division, Bristol- Myers Co.), and gentamicin sulfate (Schering Corp., Schering-Plough Corp., Kenilworth, N.J.). Purified aminoglycoside preparations used in the enzyme assays included neomycin B sulfate (The Upjohn Co., Kalamazoo, Mich.); lividomycin A sulfate, butirosin (85% A, 15% B), ribostamycin sulfate, kanamycin B sulfate, kanamycin sulfate (clinical laboratory standard), amikacin (clinical laboratory standard), and dideoxykanamycin B (Bristol Laboratories Division, Bristol-Myers Co.); gentamicin Cl, Cla, and C2 (Schering Corp., Schering-Plough Corp.); and tobramycin (clinical laboratory standard; Eli Lilly and Co.). MICs. MICs were determined in dextrose phosphate broth (GIBCO Diagnostics, Madison, Wis.) by a microtiter broth dilution method utilizing an Automatic Diluter (Cooke Laboratory Products, Alexandria, Va.). The inoculum was 105 to 106 colony-forming units per ml. The lowest concentration of antibiotic preventing visible growth after overnight incubation was defined as the MIC. Synergy experiments. Tests for antibiotic synergy were performed as described previously (19). The starting concentration was 107 colony-forming units per ml in dextrose phosphate broth. Penicillin was added at 10 U/ml, and amikacin or BB-K311 was added at 20.tg/ml. Each experiment included a control with no added antibiotic. Samples were removed at 0, 4, and 24 h, and residual viable counts were determined. The concentrations of aminoglycoside used were below the MICs for the organisms and produced no significant growth inhibition in the absence of penicillin. Four strains were also tested at an inoculum of 105 colony-forming units per ml, and for two of these the lower inoculum yielded more enhancement of penicillin killing by BB-K311 than at the higher inoculum. For these two strains, the results with the 105 inoculum were used in analysis of the data. Aminoglycoside-modifying enzyme assay. The assay for aminoglycoside phosphotransferase was performed largely as described previously (7, 15). After overnight incubation of a single-colony isolate, 2.0 ml was inoculated into 100 ml of dextrose phosphate broth and incubated at 37 C with shaking for 4 h. Late-log-phase cells were harvested by centrifugation (4,500 x g for 15 min at 4 C), and the pellets were washed twice in TMN buffer [10 mm magnesium acetate, 10 mm tris(hydroxymethyl)aminomethane, 70 mm NH4Cl, 1 mm dithiothreitol, ph 7.6] and then resuspended in 2.0 ml of TMN. After sonic disruption, intact cells and other debris were removed by centrifugation (4,500 x g for 15 min at 4 C), and the supernatant was used as a crude enzyme preparation. Phosphotransferase activity was detected by the phosphocellulose paper binding assay. The reaction mnixture consisted of 10 p1 of buffer (ph 7.1), 10 pl of 32P-labeled adenosine 5'-triphosphate solution, 10 pl of aminoglycoside solution (0.4 mg/ml), and 10 pl of crude enzyme preparation. The mixture was incubated at 35 C for 20 min when substrate profiles were being determined and for 10 to 180 min when the amount of phosphorylation over time was being measured. Enzyme activity versus different substrates was quantitated relative to neomycin. Appropriate controls utilized TMN buffer in place of the crude enzyme preparation or distilled water in place of the aminoglycoside. Adenosine 5'-[y-32P]triphosphate (1,000 to 3,000 Ci/ mmol) was obtained from New England Nuclear Corp., Boston, Mass. For use in the enzyme assay, this was diluted with distilled water and unlabeled adenosine 5'-triphosphate to a specific activity of 65 mci/mmol. Adenosine 5'-triphosphate was obtained from Sigma Chemical Co., St. Louis, Mo. RESULTS Antimicrobial susceptibility. The MICs of various antibiotics for the 11 strains tested are shown in Table 1. Eight of the organisms had high-level resistance to streptomycin and kanamycin and have previously been shown to produce plasmid-mediated APH(3') and adenylyltransferase (14, 15). Three of the organisms did not have high-level resistance, and the MICs of streptomycin and kanamycin for these organisms were much lower. For all of the organisms, the MICs of amikacin and of BB-K311 were virtually identical. Despite the previous demonstration that amikacin is a substrate for the phosphotransferase of these organisms (15), the MICs of amikacin for the enzyme-positive strains were only twofold higher than those of the enzyme-negative strains. Synergy experiments. The combination of penicillin (10 U/ml) and BB-K311 (20 ug/ml) produced enhanced killing compared with penicillin alone in all 11 strains tested (Table 2). This was true whether or not the strain produced the aminoglycoside-modifying enzymes. Of the eight enzyme-positive strains, five showed enhanced killing by the combination ranging from 101- to 102-fold, and three showed >102-fold enhancement. In contrast, penicillin and amikacin produced enhanced killing only against those strains which were enzyme negative. For those strains which produced a phosphotransferase, penicillin and amikacin were actually antagonistic, rather than indifferent as predicted (Table 2). In the eight

3 VOL. 19, 1981 AMIKACIN AND BB-K311 AGAINST S. FAECALIS 551 TABLE 1. Antimicrobial susceptibility of 11 strains of S. faecalis MIC Strain Penicillin Streptomycin Kanamycin Amikacin BB-K311 Gentamicin G (U/ml) (pg/ml) (Ag/ml) (pg/ml) (pg/ml) (g/mil) EBC-2 2 >2,000 >2,000 1,000 1, EBC-3 2 >2,000 >2, EBC-5 2 >2,000 >2, EBC-9 2 >2,000 >2, EBC-15 2 >2,000 >2, EBC-22 1 >2,000 >2,000 2,000 1, EBC-38 2 >2,000 >2, , >2,000 >2, JH TABLE 2. Summary of results of synergism experiments at 24 h Magnitude of enhanced killing by penicillin-aminoglycoside APH(3') compared with No. of strains Aminoglycoside production Noofsrm penicillin alone' molcsd io2lo 12-1o-I l 10l- 1O1O10 lo,,-lo, 1o1O 101_1O2 0o_o2,o >102 Yes 8 Amikacin 3 5 Yes 8 BB-K No 3 Amikacin 1 2 No 3 BB-K 'Numbers in columns refer to number of strains. accepted clinical indications for the use of combination antimicrobial therapy (26,27). The success of such regimens, however, depends largely on the production of a synergistic effect against enzyme-positive strains, penicillin-amikacin produced a mean of 0.99 logio units (9.8-fold) less killing at 24 h than did penicillin alone (P < 0.002, two-tailed paired t test). Representative examples of synergy experiments with an enzyme-positive and an enzyme-negative strain are shown in Fig. 1 and 2, respectively. Aminoglycoside phosphotransferase assays. The three strains which lacked high-level streptomycin and kanamycin resistance produced uniformly negative results in the assays for phosphotransferase activity. The eight enzyme-positive strains were tested against a total of 13 aminoglycoside substrates (Fig. 3). Of note, kanamycin B was an excellent substrate, whereas tobramycin (3'-monodeoxy kanamycin B) showed no modification. Kanamycin A and amikacin were also good substrates, but BB- K311 showed very little activity. In an experiment in which aminoglycoside was incubated for 3 h with or without crude enzyme, tobramycin was not significantly phosphorylated by the enzyme compared to the control without enzyme (P > 0.10, two-tailed paired t test). In contrast, BB-K311, although modified quite slowly, showed significant phosphorylation by enzyme compared to control (P < ). DISCUSSION Serious enterococcal infections, particularly bacterial endocarditis, are one ofthe most widely the infecting strain. Much recent interest has focused on the mechanisms by which bacteria may resist antibiotic synergy. Early studies with enterococci demonstrated ribosomal resistance to streptomycin and associated failure of penicillin-streptomycin synergy in a laboratory mutant (33). More recent studies in clinical isolates of enterococci, however, have focused attention on the production of aminoglycoside-modifying enzymes which confer resistance to the synergistic effects of penicillin in combination with those aminoglycosides which are modified (3, 14, 15). In S. faecalis, high-level resistance to streptomycin and kanamycin (MICs of >2,000,ug/ml) correlates with the production of two aminoglycoside-modifying enzymes, a phosphotransferase and an adenylyltransferase. The phosphotransferase confers resistance to synergy with both penicillin-kanamycin and penicillin-amikacin combinations. In S. faecium, production of an acetyltransferase likewise confers resistance to synergy when penicillin is combined with kanamycin, tobramycin, sisomicin, or netilmicin (32). Previous studies in this laboratory have suggested that the site of aminoglycoside modification by the S. faecalis phosphotransferase is the 3'-hydroxyl group (15). The studies reported

4 552 CALDERWOOD, WENNERSTEN, AND MOELLERING ANTIMICROB. AGENTS CHEMOTHER AMIK (20/Lg/ml) CZ7 2 U/O Ulm))\ BB-K311 (20,g/ml) O I \ HOURS FIG. 1. Effect ofamikacin and BB-K311 alone and in combination with penicillin against strain EBC-9, an organism which contains APH(3). AMIK, Amikacin; PEN, benzylpenicillin G. 9' 8 1Z 7 I 6 5 ) L CONTROL AMIK (204,g/ml) BB-K311 (20,Lg/ml) \ \PEPEN (10U/ml) _ \A E\I *PEN 0 2U/ml)2 + _ K. N/BB-K311(20II g/ml) PEN (10MU/mlm)+i AMIK (20}.g/ml) _\ 4I HOURS FIG. 2. Effect ofamikacin and BB-K311 alone and in combination with penicillin against strain 447, an enzyme-negative isolate. AMIK, Amikacin; PEN, benzylpenicillin G. here confirm those earlier observations. Kanamycin B (Fig. 4) is among the best substrates for this phosphotransferase. Tobramycin, on the other hand, which differs from kanamycin B only in the absence of a 3'-hydroxyl group, is not a substrate for modification (Fig. 3). The activity of the phosphotransferase against butirosin and lividomycin confirms that this enzyme belongs to the group APH(3')-III (4, 5). Amikacin, a semisynthetic derivative of kanamycin A (Fig. 5), is generally resistant to modification by phosphotransferases active at the 3'- hydroxyl site (25). Gram-negative bacilli carrying the enzymes APH(3')-I and APH(3')-II usually fail to modify amikacin and remain susceptible to that drug. A laboratory mutant strain of Escherichia coli has been described with an APH(3')-II that is active against amikacin (24). APH(3')-III enzymes have been previously described in Pseudomonas aeruginosa and Staphylococcus aureus, as well as S. faecalis (4, 30). The activity of the P. aeruginosa enzyme against amikacin is not known. The S. aureus phosphotransferase is able to modify amikacin in the radioenzymatic assay; however, these organisms remain susceptible to amikacin because of the relatively high Km of the enzyme for this antibiotic (2, 4). Previous studies in this laboratory have shown that the S. faecalis phosphotransferase is active against amikacin and that strains with this enzyme resist penicillin-amikacin synergy (15). The present work demonstrates that these strains not only resist synergy but also actually show antagonism with the penicillin-amikacin combination (Table 2). The amount of this antagonism is quite significant (nearly 10-fold) and was seen in all eight enzymepositive strains tested. To our knowledge, this is the first report with clinical isolates of antagonism between a penicillin and an aminoglycoside. The in vivo significance of this effect is presently being tested in the rabbit model. The mechanism of antagonism is as yet unknown. One approach to the problem of antibiotic resistance mediated by aminoglycoside-modifying enzymes has been the creation of semisynthetic aminoglycoside derivatives which resist such modification. Amikacin is an example of such a semisynthetic derivative currently in clinical use. Previous work with the 4'-deoxy derivative of kanamycin A has shown that this compound resists certain phosphotransferases active at the 3'-hydroxyl site (22). 4'-Deoxy, 6'-N-methylamikacin (BB-K311, Fig. 5) is a semisynthetic derivativer of amikacin designed in the hopes of resisting all known aminoglycoside-modifying enzymes (13, 23). Since previous work had shown amikacin to be a substrate for the S. faecalis phosphotransferase, we were interested in determining whether BB-K311 would resist

5 VOL. 19, 1981 o_. 200 EBC EBC-3 _0...o.... E..... AMIKACIN AND BB-K311 AGAINST S. FAECALIS EBC-15 EBC-22 n _ _... ~ioeo 20 EBC _.,.. 0 ei-z E BC ,, 200 EBC V..... FIG. 3. Substrate profiles ofphosphotransferases from eight strains of S. faecalis with high-level resistance to streptomycin and kanamycin (MICs, >2,000 pg/ml). Activities are calculated relative to neomycin as 100%o. Neo, Neomycin; Liv. A., lividomycin A; But, butirosin; Rib, ribostamycin; Km. A, kanamycin A; Km. B, kanamycin B; Amik, amikacin; BBK, BB-K311; Gm Cl, gentamicin Cl; Gm Cla, gentamicin Cla; Gm C2, gentamicin C2; DKB, dideoxykanamycin B; Tm, tobramycin. H HC- NH2 NH2 HO KANAMYCIN B TOBRAMYCIN FIG. 4. Structural formulas of kanamycin B and tobramycin. this enzyme. The MICs of both amikacin and BB-K311 against our strains were similarly high (Table 1). Tests for antibiotic synergism, however, showed that BB-K311 enhanced the activity of penicillin against both enzyme-negative and enzyme-positive strains, as opposed to amikacin, which produced synergy only in enzymenegative strains (Table 2). The radioenzymatic assay for phosphorylation confirmed that BB- K311 was a very poor substrate for the phosphotransferase, in contrast with amikacin (Fig. 3). This suggests that removing the 4'-hydroxy group in amikacin effectively blocks 3'-phosphorylation by the S. faecalis enzyme, although the mechanism of blockade is as yet unknown. The results of the prolonged incubation experiment prove that BB-K311 does undergo enzymatic phosphorylation at a rate significantly higher than control, although the amount of phosphorylation is too low to be picked up in the standard assay system. This is consistent with the hypothesis that the site of phosphorylation is still present within the molecule (i.e., the 3'- hydroxy site) but is effectively blocked by struc-

6 554 CALDERWOOD, WENNERSTEN, AND MOELLERING ANTIMICROB. AGENTS CHEMOTHER. H HC-NH2 CH20H 0 0 KANAMYCIN A NH HO AMIKACIN HC-NHCH3 NH2 0 NHC-CH -CH2 -CH2 X<o < NH2 CH20H 0 0 HO NH BB-K311 FIG. 5. Structural formulas of kanamycin A, amikacin, and BB-K311. tural changes elsewhere in the compound. BB- K311 thus may resist inactivation by all known aminoglycoside-modifying enzymes. Further testing in vitro and in vivo seems warranted. ACKNOWLEDGMENTS S.B.C. was supported by Public Health Service research fellowship grant AI from the National Institute of Allergy and Infectious Disease. We gratefully acknowledge the assistance of Florence Larson in the preparation of the manuscript. LITERATURE CITED 1. Calderwood, S. A., C. Wennersten, R. C. Moellering, Jr., L. J. Kunz, and D. J. Krogstad Resistance to six aminoglycosidic aminocyclitol antibiotics among enterococci: prevalence, evolution, and relationship to synergism with penicillin. Antimicrob. Agents Chemother. 12: Courvalin, P., and J. Davies Plasmid-mediated aminoglycoside phosphotransferase of broad substrate range that phosphorylates amikacin. Antimicrob. Agents Chemother. 11: Courvalin, P. M., W. V. Shaw, and A. E. Jacob Plasmid-mediated mechanisms of resistance to aminoglycoside-aminocyclitol antibiotics and to chloramphenicol in group D streptococci. Antimicrob. Agents Chemother. 13: Davies, J., and D. I. Smith Plasmid-determined resistance to antimicrobial agents. Annu. Rev. Microbiol. 32: Dowding, J., and J. Davies Mechanisms and origins of plasmid-determined antibiotic resistance, p In D. Schlessinger (ed.), Microbiology American Society for Microbiology, Washington, D.C. 6. Facklam, R. R Recognition of group D streptococcal species of human origin by biochemical and physiological tests. Appl. Microbiol. 23: Haas, M. J., and J. E. Dowding Aminoglycosidemodifying enzymes. Methods Enzymol. 43: Hook, E. W., Ill, R. B. Roberts, and M. A. Sande Antimicrobial therapy of experimental enterococcal endocarditis. Antimicrob. Agents Chemother. 8: Hunter, T. H Use of streptomycin in treatment of bacterial endocarditis. Am. J. Med. 2: Jacob, A. E., and S. J. Hobbs Conjugal transfer of plasmid-borne multiple antibiotic resistance in Streptococcus faecalis var. zymogenes. J. Bacteriol. 117: Jawetz, E., J. B. Gunnison, and V. R. Colman The combined action of penicillin with streptomycin and chloromycetin on enterococci in vitro. Science 111: Korzeniowski, 0. M., C. Wennersten, R. C. Moellering, Jr., and M. A. Sande Penicillin-netilmicin synergism against Streptococcus faecalis. Antimicrob. Agents Chemother. 13: Kresel, P. A., T. A. Pursiano, K. E. Price, M. Misiek, and F. Leitner BB-K311: a derivative of amikacin resistant to inactivation by all known aminoglycoside-modifying enzymes, p In J. D. Nelson and C. Grassi (ed.), Current chemotherapy and infectious disease: proceedings of the 11th International

7 VOL. 19, 1981 Congress of Chemotherapy and the 19th Interscience Conference on Antimicrobial Agents and Chemotherapy. American Society for Microbiology, Washington, D.C. 14. Krogstad, D. J., T. R. Korfhagen, R. C. Moellering, Jr., C. Wennersten, and M. N. Swartz Plasmid-mediated resistance to antibiotic synergism in enterococci. J. Clin. Invest. 61: Krogstad, D. J., T. R. Korfhagen, R. C. Moellering, Jr., C. Wennersten, M. N. Swartz, S. Perzynski, and J. Davies Aminoglycoside-inactivating enzymes in clinical isolates of Streptococcus faecalis. J. Clin. Invest. 62: Krogstad, D. J., and A. R. Parquette Defective killing of enterococci: a common property of antimicrobial agents acting on the cell wall. Antimicrob. Agents Chemother. 17: Mandell, G. L, D. Kaye, M. E. Levison, and E. W. Hook Enterococcal endocarditis. Arch. Intern. Med. 125: Moellering, R. C., Jr., B. K. Watson, and L J. Kunz Endocarditis due to group D streptococci. Am. J. Med. 57: Moellering, R. C., Jr., C. Wennersten, and A. N. Weinberg Studies on antibiotic synergism against enterococci. L. Bacteriologic studies. J. Lab. Clin. Med. 77: Moellering, R. C., Jr., C. Wennersten, and A. N. Weinberg Synergy of penicillin and gentamicin against enterococci. J. Infect. Dis. 124(Suppl.):S207- S Moellering, R. C., Jr., C. Wennersten, and A. J. Weinstein Penicillin-tobramycin synergism against enterococci: a comparison with penicillin and gentamicin. Antimicrob. Agents Chemother. 3: Naito, T., S. Nakagawa, Y. Abe, K. Fujisawa, and H. Kawaguchi Aminoglycoside antibiotics. VIII. Synthesis and activity of 4'-deoxykanamycin A. J. Antibiot. 27: Naito, T., S. Nakagawa, S. Toda, K. Fujisawa, and H. Kawaguchi Aminoglycoside antibiotics. XIII. Synthesis and activity of 4'-deoxy-6'-N-methylamikacin AMIKACIN AND BB-K311 AGAINST S. FAECALIS 555 and related compounds. J. Antibiot. 32: Perlin, M. H., and S. A. Lerner Amikacin resistance associated with a plasmid-borne aminoglycoside phosphotransferase in Escherichia coli. Antimicrob. Agents Chemother. 16: Price, K. E., M. D. DeFuria, and T. A. Pursiano Amikacin, an aminoglycloside with marked activity against antibiotic-resistant clinical isolates. J. Infect. Dis. 134(Suppl.):S249-S Rahal, J. J., Jr Antibiotic combinations: the clinical relevance of synergy and antagonism. Medicine 57: Sande, M. A., and W. M. Scheld Combination antibiotic therapy of bacterial endocarditis. Ann. Intern. Med. 92: Standiford, H. D., J. B. demaine, and W. M. M. Kirby Antibiotic synergism of enterococci. Arch. Intern. Med. 126: Toala, P., A. McDonald, C. Wilcox, and M. Finland Susceptibility of group D streptococcus (enterococcus) to 21 antibiotics in vitro, with special reference to species differences. Am. J. Med. Sci. 258: Umezawa, Y., M. Yagisawa, T. Sawa, T. Takeuchi, and H. Umezawa Aminoglycoside 3'-phosphotransferase III, a new phosphotransferase. J. Antibiot. 28: Watakunakorn, C Penicillin combined with gentamicin or streptomycin: synergism against enterococci. J. Infect. Dis. 124: Wennersten, C. B., and R. C. Moellering, Jr Mechanism of resistance to penicillin-aminoglycoside synergism in Streptococcus faecium, p In J. D. Nelson and C. Grassi (ed.), Current chemotherapy and infectious disease: Proceedings of the 11th International Congress of Chemotherapy and the 19th Interscience Conference on Antimicrobial Agents and Chemotherapy. American Society for Microbiology, Washington, D.C. 33. Zimmermann, R. A., R. C. Moellering, Jr., and A. N. Weinberg Mechanism of resistance to antibiotic synergism in enterococci. J. Bacteriol. 105: