Characterization of Penicillin-Binding Protein 2 of Staphylococcus

Transcription

1 ANTIMICROBIAL AGENTS AND CHEMOTHERAPY, Mar. 1992, P /92/ $02.00/0 Copyright 1992, American Society for Microbiology Vol. 36, No. 3 Characterization of Penicillin-Binding Protein 2 of Staphylococcus aureus: Deacylation Reaction and Identification of Two Penicillin-Binding Peptides HENRY F. CHAMBERS* AND CATHLEEN MIICK Medical Service, San Francisco General Hospital, San Francisco, Califonia 94110, and Department of Medicine, University of California San Francisco, San Francisco, California Received 22 July 1991/Accepted 2 January 1992 Penicillin-binding protein (PBP) 2 is the major PBP of five that have been identified in susceptible strains of Staphylococcus aureus. Beta-lactam antibiotic binding to PBP 2 is important for the antibacterial effect. Antibiotic binding to PBP 2 in strain 209P was examined with sodium dodecyl sulfate-polyacrylamide gel electrophoresis in competition assays using [3H]penicillin as the radiolabel. Clavulanic acid, which is specifically bound by PBP 2, and cefaclor, which is specific for PBP 3, were studied. Cefaclor, which alone appeared not to bind PBP 2, in combination inhibited PBP 2 binding of clavulanic acid. By varying the temperature during radiolabeling with [3Hlpenicillin in cefaclor competition assays and in direct radiolabeling assays with [3H]cefaclor, it was shown that cefaclor was bound by PBP 2 with high affinity (50%o inhibitory concentration, <0.1,ug/ml) and that the apparent low-affinity binding (50%Yo inhibitory concentration, >10,ug/ml) in competition assays performed at 37 C was due to rapid deacylation. Two penicillin-binding peptides of PBP 2 also were identified in fluorographs of PBPs separated by nonequilibrium ph gradient gel and two-dimensional electrophoresis. Rapid deacylation for some antibiotics and the presence of two penicillinbinding peptides are two properties of PBP 2 that should be considered when correlating results of binding assays with effects of beta-lactam antibiotics on S. aureus. Five penicillin-binding proteins (PBPs), 1, 2, 3, 3', and 4, with molecular masses of 85, 81, 75, 70, and 45 kda, respectively, have been identified in susceptible strains of Staphylococcus aureus (8, 19). PBPs are membrane-bound proteins that catalyze carboxypeptidase and transpeptidase reactions of bacterial cell wall synthesis (18). PBPs are targets of beta-lactam antibiotics. These antibiotics are structural analogs of the natural PBP substrate and inhibit cell wall synthesis by covalently binding to PBP enzymatic sites (11, 18). Fluorography of PBPs radiolabeled with penicillin has been the primary experimental method used to assay binding of beta-lactam antibiotics. PBPs differ in binding affinities among the several protein species and among antibiotics. Difference in binding affinities is the basis of experimental designs utilizing antibiotics to study PBP function. For example, if a specific PBP binds an antibiotic in association with a particular morphological disturbance, then that PBP is a candidate enzyme for the reaction that governs normal morphology. If a PBP does not bind an antibiotic, and the antibiotic is bactericidal nevertheless, then that PBP would not be considered an important target for the antibiotic. Current evidence indicates that PBPs 1, 2, and 3 are the major targets of beta-lactam antibiotics (3, 6, 12). They are also required for viability. It has not been possible to produce mutants lacking either PBP 2 or 3. Mutants deficient in PBPs 1 and 4 have been reported (4, 5), although there is a question about whether PBP 1 was lacking, and PBP 1 also appears to be necessary for survival (12). The experiments described below grew out of results of binding studies in which the role of PBP 2 in mediating a beta-lactam antibiotic effect in S. aureus was examined. * Corresponding author. 656 Abundant data indicate that PBP 2 is a critical antibiotic target, yet cephalexin (7) and cefaclor (3) appeared not to bind to PBP 2 but were effective antibacterial agents, suggesting that PBP 2 is not an important target of these drugs. In pursuit of an explanation for this inconsistency, two unrecognized properties of PBP 2 were identified. MATERLILS AND METHODS Experimental strains. Strains used for these studies were P-lactamase negative. Strain 209P is a methicillin-susceptible laboratory strain of S. aureus. Strain 27R is a homogeneous methicillin-resistant laboratory strain. Strains ID-7 and 67-0 are heterogeneous methicillin-resistant clinical isolates. Strain 29 is a clinical isolate with low-level resistance to methicillin (2). Strain 1-63 is a methicillin-susceptible clinical isolate. Growth curves. An overnight culture was diluted 1:1,000 into 100 ml of prewarmed Trypticase soy broth, which was incubated at 37 C on a shaking-platform incubator. Growth was measured by densitometry in a Sequoia-Turner spectrophotometer at 570 nm. Once exponential growth was achieved, 10-ml culture samples at an absorbance of 0.1 were dispensed into individual 50-ml flasks. Antibiotic was added to the sample, which was incubated at 37 C, and absorbance was measured hourly. Sample preparation and radiolabeling of PBPs. Both membrane preparations and whole cells were used as samples for radiolabeling of PBPs. Membranes were prepared from exponentially growing cells in Trypticase soy broth (BBL)-4% NaCl. Cells were harvested at a density of 2 x 108 to 5 x 108 CFU/ml, washed, and then mechanically disrupted. Membranes were separated by differential centrifugation and suspended in 10 mm Tris-HCI, ph 7.0, to a final protein concentration of 10 mg/ml as assayed by a protein

2 VOL. 36, 1992 assay kit (Bio-Rad, Richmond, Calif.). Samples of 10,ug of protein were incubated with nonradioactive antibiotic and then radiolabeled with 5,ug of [3H]penicillin per ml. A reference sample was prepared by radiolabeling without prior incubation in nonradioactive active antibiotic. Temperature, times of incubation, nonradioactive antibiotic, and concentration were varied, depending upon the experiment. The reaction was stopped by addition of denaturation buffer for electrophoresis. For whole-cell preparations, cells in the exponential growth phase were grown to a density of approximately 2 x 108 to 5 x 108 CFU/ml. Nonradioactive antibiotic, or buffer only, for the reference sample, was added to a 1-ml volume. After incubation for 30 min at 37 C, cells were collected by centrifugation for 5 min at 8,000 x g and the pellet was suspended in 15,ul of 10 mm Tris-HCI-10 mm MgCl2 buffer, ph 7.6, containing 15,ug of DNase per ml, 15,ug of RNase per ml, 200,ug of lysostaphin per ml, and 5,ug of [3H]penicillin per ml. After incubation for 30 min, the reaction was stopped by addition of denaturation buffer. Electrophoretic separation of PBPs. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was performed in a minigel apparatus (Bio-Rad) using an acrylamide/bisacrylamide ratio of 10:0.13%. The gel was stained with Coomassie blue. Radiolabeled PBPs were detected by fluorography with prefogged Kodak X-Omat film (9). Nonequilibrium ph gradient gel electrophoresis (NEPH GE) was performed by the denaturing isoelectric focusing method of Robertson et al. (13). The final concentration of Triton X-100 was 4%, and ampholytes with ph ranges of 6 to 8 and 5 to 7 (Bio-Rad) were each used at a final concentration of 0.1%. PBPs were detected by fluorography. Two-dimensional electrophoresis was performed by the method of O'Farrell (10) in a minigel system using 1.5-mm spacers. After proteins had been separated by NEPHGE, the lane was sliced from the gel, incubated in equilibration buffer, and laid on top of the stacking gel for SDS-PAGE. Measurement of antibiotic binding to PBPs. PBP binding of antibiotics in competition assays was expressed in terms of a percent saturating concentration, i.e., percent reduction in radiolabeled protein, measured by scanning densitometry, relative to the reference sample. Percent density was plotted versus the log1o concentration, and IC50 and IC90, the concentrations that reduced radiolabeling by 50 and 90%, respectively, were calculated from the equation for the linear portion of the curve, derived by least-squares analysis. RESULTS PBP binding. Clavulanic acid was bound by PBP 2 in competition assays with high binding affinity and specificity (Fig. 1). After 15 min of incubation at 37 C, 1,ug of clavulanic acid per ml reduced the density of radiolabeled PBP 2 in fluorographs by more than 90%. PBP 3 was 16% saturated by clavulanic acid at this concentration. Under the same conditions, cefaclor was specifically bound by PBP 3 with more than 90% saturation at 0.1,ug/ml. PBP 2-binding sites were 9 and 34% saturated at cefaclor concentrations of 1 and 10,ug/ml, respectively. The correlation between PBP 2 binding and antibacterial activity of cefaclor for strain 209P was examined by comparing lysis in cells exposed to cefaclor and clavulanic acid (Fig. 2). By the macrodilution method with Mueller-Hinton broth, the MICs of cefaclor and clavulanic acid were 1 and 64,ug/ml, respectively. Cefaclor was used at concentrations of 0.05, 0.125, and 0.25,ug/ml, concentrations which in PBP 2 OF S. AUREUS 657 Clay Cefa1lor l 17i FIG. 1. Fluorograph of PBPs 1, 2, and 3 bound by clavulanic acid (Clav) or cefaclor. PBPs in membrane preparations were radiolabeled with 5,ug of [3H]penicillin per ml for 15 min at 37 C after incubation for 15 min at 37 C with either clavulanic acid or cefaclor at concentrations of 0, 0.1, 1, and 10,ug/ml. The numbers at the left indicate PBPs. competition assays specifically bound PBP 3, yielding percent saturations ranging between 50 and 100%, but did not bind PBP 2. Clavulanic acid was used at a concentration of 5,ug/ml, which produced greater than 90% saturation of PBP 2 but less than 50% saturation of PBP 3. Neither cefaclor at concentrations of '0.125,ug/ml nor clavulanic acid alone produced cell lysis, but clavulanic acid significantly enhanced lysis in cells treated with 0.05 jig of cefaclor per ml. This suggested that PBP 2 binding was important for cell lysis. Cefaclor alone, at a concentration of 0.25,ug/ml, at which PBP 2 appeared not to be bound, however, produced cell lysis similar that of the combination, suggesting that PBP 2 is not an important target. Binding of PBPs by cefaclor and clavulanic acid was examined next in whole cells. Paradoxically, more PBP 2 was radiolabeled in samples that had been incubated with cefaclor plus clavulanic acid (the latter at concentrations which, alone, saturated PBP 2 99%) than in samples incubated only with clavulanic acid (Fig. 3). PBP 2 became less saturated as the cefaclor concentration was increased from 0.1 to 1,ug/ml. The binding assay was repeated by using membrane samples, and similar results were obtained, indicating that induction of a PBP was not the explanation for the paradoxical binding pattern produced by cefaclor and clavulanic acid in combination. These experiments suggested that cefaclor was bound to PBP 2 but that binding was not detected. Perhaps release of bound cefaclor freed the active site for binding of [3H]penicillin during radiolabeling. Experiments were performed to characterize more fully the antibiotic-binding properties of PBP 2. Rapid deacylation of PBP 2 with release of bound cefaclor was examined as a possible explanation for failure to detect cefaclor binding. Cefaclor binding to PBP 2 was examined both in a competition assay, by its inhibition of [3H]penicillin binding, and in a direct binding assay. Radiolabeling in the competition assays was performed both at 4 C and at 37 C. The lower temperature was chosen to slow deacylation and prevent release of bound cefaclor from PBP 2 during radiolabeling. Binding of [3H]penicillin to PBP 2 at a concentration of 5 jig/ml was shown not to be significantly affected over this temperature range. Binding was essentially complete after only 1 min at 4 or 37 C and remained constant over 1 h (the coefficient of variation ranged from 17 to 21%). Hence, any inhibition of [3H]penicillin binding to PBP 2 in the presence of cefaclor could not

3 658 CHAMBERS AND MIICK ANTiMICROB. AGENTS CHEMOTHER. No drug Clav C a o aż 50. 8% 1 a.0 r Goa Cecbor ~~~~~~~~~~~~~~Cofaclor Clay + Ceclor I I " FIG. 2. Effect of 5 plg of clavulanic acid (Clav) per ml and cefaclor (Ceclor) on the growth of strain 209P. The concentration of cefaclor used is indicated with each growth curve. The antibiotic was added to exponentially growing cultures in Trypticase soy broth at 0 h, and the cultures were incubated at 37'C. Growth was measured by densitometry. be attributed to altered binding kinetics of [3H]penicillin at 40C. Cefaclor binding to PBP 2 was then examined in competition assays in which samples were preincubated with 1,ug of cefaclor per ml for 30 min at 37 C and then radiolabeled at either 4 or 37 C for 1 to 120 min. Amounts of radiolabeled PBP 2 increased with time at 37 C but not at 4 C (Fig. 4), indicating that cefaclor had bound to PBP 2 but was released at 37 C. The half-life of cefaclor acyl-pbp 2 was approximately 10 min at 37 C, compared with > 120 min at 4 C. The half-life of penicillin acylated PBP 2 at 37 C (determined by sampling of [3Hlpenicillin-radiolabeled PBP 2 incubated in FIG. 3. Effect of cefaclor on binding of clavulanic acid Clav) to PBP 2 in whole cells. All incubations with antibiotic and [ H]penicillin were for 30 min at 37 C. (A) Effects of no clavulanic acid (0) and 5 or 10 jg of clavulanic acid per ml on radiolabeling. (B) Effects of no cefaclor or clavulanic acid (0); cefaclor alone at a concentration of 0.1, 0.5, or 1 p,g/ml; and cefaclor in combination with 5 or 10,ug of clavulanic acid per ml. 25 FIG C Minutes Percent saturation of PBP 2 by 1,ug of unlabeled cefaclor per ml. A membrane sample was incubated with 1 1zg of cefaclor per ml for 30 min at 37 C and then radiolabeled with [3Hlpenicillin at either 37 or 4'C for 1 to 120 min. The density of PBP 2 was measured, and percent saturation was calculated relative to that of a sample radiolabeled with 5.g of [3H]penicillin per ml for 1 min at 40C. 1,000,ug of unlabeled penicillin per ml) was 45 min, in agreement with its reported half-life (17). Binding of cefaclor to PBPs was also determined in direct assays using [3H]cefaclor (gift of Eli Lilly & Company, Indianapolis, Ind.). Membrane samples were radiolabeled for 15 min at 37 C. PBPs 1, 2, and 3 were 90% saturated at concentrations of 0.06, 0.3, and pg/ml, respectively (Fig. 5). Thus, the affinity of PBP 2 for cefaclor, expressed as percent saturation, was 100 times greater when measured directly instead of indirectly in a competition assay. Separation of PBP 2 bands by isoelectric point. The possibility that the unusual antibiotic-binding properties of PBP 2 result from the presence of more than one penicillin-binding peptide at the PBP 2 position in SDS-PAGE gels was also examined by fluorography of PBPs separated by NEPHGE. Two distinct peptides corresponding to PBP 2 were identified (Fig. 6). Correspondence of the two more-acidic radiolabeled peptides to PBP 2 and the more-basic peptide with PBP 3 was shown by their differential binding of cefaclor and clavulanic acid. The PBP 2 peptides specifically bound clavulanic acid but not cefaclor at 37 C, and PBP 3 bound cefaclor but not clavulanic acid. Two-dimensional electrophoresis confirmed that the two acidic, clavulanic acidbinding peptides migrated at the PBP 2 position and that the basic cefaclor-binding protein migrated with PBP 3 in SDS- PAGE gels (Fig. 7). The migration of PBPs 1 and 4 with NEPHGE and two-dimensional electrophoresis was not systematically investigated. PBP 1, which can be difficult to resolve in standard SDS-PAGE systems, did not focus over the ph range used to separate PBPs 2 and 3 and was not

4 VOL. 36, 1992 PBP 2 OF S. AUREUS 659 FIG. 5. Fluorograph of [3H]cefaclor-labeled PBPs. PBPs 1, 2, and 3 are indicated in the left margin. Concentrations of 0.025, 0.1, 0.5, and 1 plgiml were used to label PBPs directly in membrane samples incubated for 15 min at 370C. identified. A low-molecular-wei'ght PBP, possibly PBP 4, which could be visualized only with prolo'nged exposure of the fluorograph, appeared to migrate with a ph similar to that of PBP 3. 3/7A? FIG. 6. Fluorograph of PBP 2 peptides and PBP 3 radiolabeled with 5 p~g of [3H]penicillin per ml and separated by NEPHGE (A). The top of the fluorograph was the acidic side of the gel slab. (B) Flu'orograph of NEPHGE slab gel samples incubated with only 5 plg of[3h]penicillin per ml, indicated by 0; cefaclor at 1 g/ml and then [3H]penicillin; or clavulanic acid (Clay) at 1 plg/ml and then [3H]penicillin. Each incubation was for 15 min at 370C. FIG. 7. Two-dimensional gel of a [3H]penicillin-labeled sample separated by NEPHGE and then SDS-PAGE. a and b are the acidic and basic PBP 2 peptides, respectively. A sample of radiolabeled PBPs 1, 2, 3, and 4 separated by SDS-PAGE without NEPHGE was used as size markers. Similar results were obtained with NEPHGE by using whole cells instead of membranes, suggesting that the two peptides were not an artifact of membrane preparation. Five strains of S. aureus, 29, 27R, 1-63, ID-7, and 67-0, were also examined. In all five, two peptides corresponding to PBP 2 were separated by NEPHGE. Binding of antibiotics to the two PBP 2 peptides was compared. The two peptides were similar for each of the four drugs tested for samples radiolabeled at 4 C (Table 1). Methicillin and cefaclor at 37 C were bound less well (i.e., higher IC50 and IC90 values) by the acidic peptide than by the basic peptide. DISCUSSION Two properties of PBP 2 of S. aureus that are potentially important for understanding of the effects of antibiotics on cells were identified. (i) PBP 2 can be rapidly deacylated with some antibiotics. If the time of radiolabeling in a competition assay exceeds the half-life of the acyl-pbp, then PBP binding of the antibiotic will be significantly underestimated. (ii) PBP 2 resolved into two distinct peptides in NEPHGE gels. The paradoxical effect of increased radiolabeling of PBP 2 in the presence of cefaclor plus clavulanic acid was the TABLE 1. Antibiotic binding to acidic and basic PBP 2 peptides separated by NEPHGE Drug Radiolabeling IC50 (ilg/ml) IC90 (.g/ml) temp ( C) Acidic Basic Acidic Basic Cefaclor Cefaclor Cefoxitin Cefoxitin Clavulanic acid Clavulanic acid Methicillin Methicillin

5 660 CHAMBERS AND MIICK original observation suggesting that cefaclor bound PBP 2 with relatively high affinity, despite results of competition assays indicating an IC50 of >10,ug/ml. Direct radiolabeling assays indicated that the 50% saturating concentration was on the order of 0.1,ug/ml. Cefaclor's high-affinity binding with relatively rapid deacylation (acyl-pbp half-life of 10 min) probably accounts for the paradoxical effect of cefaclor on clavulanic acid binding. If PBP 2 which is exposed to saturating concentrations of both cefaclor and clavulanic acid has bound clavulanic acid, which has a relatively stable acyl-pbp derivative, then it will not be available for radiolabeling. If, instead of clavulanic acid, PBP 2 has bound cefaclor, which has an acyl derivative half-life of 10 min, then during 30 min of radiolabeling with a saturating concentration of [3H]penicillin, approximately 90% (the amount deacylated over three 10-min half-lives) of the cefaclor-bound PBP 2 could become free to bind [3H]penicillin. As the concentration of cefaclor increased, more PBP 2 would be bound by it rather than clavulanic acid. Consequently, the amount of PBP 2 available for deacylation and radiolabeling would increase, causing apparent inhibition of clavulanic acid binding of PBP 2 by cefaclor. The ability of PBP 2, or other PBPs, to deacylate rapidly has important implications for interpretation of experiments in which binding of PBPs is correlated with physiological or antibacterial effects on cells. Failure to detect binding because of deacylation could lead to erroneous conclusions about antibiotic effects. The observation that cefaclor was antibacterial at concentrations at which binding of PBP 2 seemed not to occur was inconsistent with other experimental data suggesting that PBP 2 is an important target for the beta-lactam antibiotic effect (3, 6, 12A. When binding was determined by direct labeling with [ H]cefaclor or under conditions that inhibited deacylation, the affinity of both PBP 2 peptides for cefaclor actually was found to be relatively high (IC50, on the order of 0.1,ug/ml). Thus, the antibacterial effect of cefaclor is associated with binding of PBP 2, a result consistent with its importance as a target of beta-lactam antibiotics. Cephalexin, which is structurally very similar to cefaclor, also has been reported to be bound by PBP 2 with low affinity and bound by PBP 3 with high affinity (7). Septation abnormalities observed in cells treated with cephalexin have been attributed to specific binding only to PBP 3, suggesting a role for this PBP in normal septation (7). At the 2-,ug/ml concentration used in those experiments, however, it is possible that both PBPs 2 and 3 were bound and that PBP 2 binding was not detected because of deacylation. Thus, ascribing a specific function to PBP 3 on the basis of these experiments may not be correct. Deacylation occurs with PBPs 1, 3, and 4, as well as PBP 2 (14, 17). Binding assays should be designed to take into account the possibility of deacylation as an explanation for low binding affinity of an antibiotic before inferences about PBP functions are made on the basis of these assays. The finding of two PBP 2 peptides is intriguing because it suggests that another, previously unrecognized, antibiotic target, perhaps with a special function, is present in S. aureus. The two peptides may be distinct PBPs encoded on separate genes or result from postranscriptional or posttranslational modification of the product of the same gene. Data from the binding studies suggest that the two peptides are related: both peptides had qualitatively similar binding and deacylation activities for methicillin, cefaclor, clavulanic acid, and cefoxitin. If a single gene does encode PBP 2, then posttranslational processing seems likely, because bac- ANTIMICROB. AGENTS CHEMOTHER. teria are not known to modify mrna before translation. The additional PBP 2 peptide also could be a degradation product of the complete protein. An extra PBP 2 has been reported previously from this laboratory (2) and by Berger-Bachi et al. (1). Interestingly, both reports concerned strains expressing low-level resistance (i.e., the strains did not produce PBP 2a, the extra PBP produced by typical methicillin-resistant strains) to methicillin, one naturally (2) and the other after antibiotic selection (1). We speculated that the extra peptide in this strain could be an important target of penicillin and might be responsible for its low-level resistance. Berger-Bachi speculated that an altered PBP 2 with relatively lower binding affinity for beta-lactam antibiotics may have been derived by modification of one of the two PBP 2 peptides. These observations are particularly interesting in light of the potential role of PBP 2 in mediating low-level resistance to beta-lactam antibiotics (15, 16). Increased deacylation of PBP 2 is a possible mechanism for reduced antibiotic binding and, therefore, for resistance. Further studies are needed to determine whether this is likely to be the case and to determine whether PBP 2 peptides are encoded by one gene or two. ACKNOWLEDGMENT This work was supported by USPHS grant Al REFERENCES 1. Berger-Bachi, B., A. Strassle, and F. H. Kayser Natural methicillin resistance in comparison with that selected by invitro drug exposure in Staphylococcus aureus. J. Antimicrob. Chemother. 23: Chambers, H. F., G. Archer, and M. Matsuhashi Lowlevel methicillin resistance in strains of Staphylococcus aureus. Antimicrob. Agents Chemother. 33: Chambers, H. F., and M. Sachdeva Binding of betalactam antibiotics to penicillin-binding proteins in methicillinresistant Staphylococcus aureus. J. Infect. Dis. 161: Curtis, N. A., and M. V. Hayes A mutant of Staphylococcus aureus H lacking penicillin-binding 1 is viable. FEMS Microbiol. Lett. 10: Curtis, N. A., M. V. Hayes, A. W. Wyke, and J. B. Ward A mutant of Staphylococcus aureus H lacking penicillin-binding protein 4 and transpeptidase activity in vivo. FEMS Microbiol. Lett. 9: Frere, J. M., and B. Joris Penicillin-sensitive enzymes in peptidoglycan biosynthesis. Crit. Rev. Microbiol. 11: Georgopapadakou, N. H., B. A. Dix, and Y. R. Mauriz Possible physiological functions of penicillin-binding proteins in Staphylococcus aureus. Antimicrob. Agents Chemother. 29: Georgopapadakou, N. H., and F. Y. Liu Binding of beta lactam antibiotics to penicillin-binding proteins of Staphylococcus aureus and Streptococcus faecalis: relation to antibacterial activity. Antimicrob. Agents Chemother. 18: Laskey, R. A., and A. D. Mills Quantitative film detection of 3H and 14C in polyacrylamide gels by fluorography. Eur. J. Biochem. 56: O'Farrell, P. H High resolution two-dimensional electrophoresis of proteins. J. Biol. Chem. 250: Park, J. T., and J. L. Strominger Mode of action of penicillin. Biochemical basis for the mechanism of action of penicillin and its selective toxicity. Science 125: Reynolds, P. E The essential nature of staphylococcal penicillin-binding proteins, p In P. Actor, L. Daneo- Moore, M. L. Higgins, M. Salton, and G. D. Shockman (ed.), Antibiotic inhibition of bacterial cell surface assembly and function. American Society for Microbiology, Washington, D.C. 13. Robertson, E. F., H. K. Dannelly, P. J. Malloy, and H. C.

6 VOL. 36, 1992 Reeves Rapid isoelectric focusing in a vertical polyacrylamide minigel system. Anal. Biochem. 167: Sumita, Y., F. Fukasawa, and T. Okuda Affinities of SM-7338 for penicillin-binding proteins and its release from these proteins in Staphylococcus aureus. Antimicrob. Agents Chemother. 34: Tomasz, A., H. B. Drugeon, H. M. de Lencastre, D. Jabes, L. McDougal, and J. Bille New mechanism for methicillin resistance in Staphylococcus aureus: clinical isolates that lack the PBP 2a gene and contain normal penicillin-binding proteins with modified penicillin-binding capacity. Antimicrob. Agents Chemother. 33: Tonin, E., and A. Tomasz Beta-lactam-specific resistant PBP 2 OF S. AUREUS 661 mutants of Staphylococcus aureus. Antimicrob. Agents Chemother. 30: Waxman, D. J., and J. L. Strominger Cephalosporinsensitive penicillin-binding proteins of Staphylococcus aureus and Bacillus subtilis active in the conversion of ['4C]penicillin G to [14C]phenylacetylglycine. J. Biol. Chem. 254: Waxman, D. J., and J. L. Strominger Penicillin-binding proteins and the mechanism of action of beta-lactam antibiotics. Annu. Rev. Biochem. 52: Wyke, A. W Isolation of five penicillin-binding proteins from Staphylococcus aureus. FEMS Microbiol. Lett. 22: