High level oxacillin and vancomycin resistance and altered cell wall composition in Staphylococcus aureus carrying the staphylococcal mecA and the enterococcal vanA gene complex.

Recently, for the first time in the history of this bacterial species, methicillin-resistant Staphylococcus aureus (MRSA) carrying the enterococcal vanA gene complex and expressing high level resistance to vancomycin was identified in clinical specimens (CDC (2002) MMWR 51, 565-567). The purpose of our studies was to understand how vanA is expressed in the heterologous background of S. aureus and how it interacts with the mecA-based resistance mechanism, which is also present in these strains and is targeted on cell wall biosynthesis. The vanA-containing staphylococcal plasmid was transferred from the clinical vancomycin-resistant S. aureus (VRSA) strain HIP11714 (CDC (2002) MMWR 51, 565-567) to the methicillin-resistant S. aureus (MRSA) strain COL for which extensive genetic and biochemical information is available on staphylococcal cell wall biochemistry and drug resistance mechanisms. The transconjugant named COLVA showed high and homogeneous resistance to both oxacillin and vancomycin. COLVA grown in vancomycin-containing medium produced an abnormal peptidoglycan: all pentapeptides were replaced by tetrapeptides, and the peptidoglycan contained at least 22 novel muropeptide species that frequently showed a deficit or complete absence of pentaglycine branches. The UDP-MurNAc-pentapeptide, the major component of the cell wall precursor pool in vancomycin-sensitive cells was replaced by UDP-MurNAc-depsipeptide and UDP-MurNAc-tetrapeptide. Transposon inactivation of the beta-lactam resistance gene mecA caused complete loss of beta-lactam resistance but had no effect on the expression of vancomycin resistance. The two major antibiotic resistance mechanisms encoded by mecA and vanA residing in the same S. aureus appear to use different sets of enzymes for the assembly of cell walls.

Until the late 1990s, clinical isolates of Saccharomyces aureus have retained a uniform high sensitivity to vancomycin with minimal inhibitory concentration (MIC) 1 values in the vicinity of 1 g/ml. Beginning with the late 1990s, isolates with reduced susceptibility to vancomycin began to be reported from several countries (1), but the MIC value of these vancomycin intermediate-resistant S. aureus isolates was limited to the range of 8 -16 g/ml of the antibiotic. The first highly vancomycin-resistant S. aureus (VRSA) isolates (MIC over 32 g/ml) were only detected during the last year in two hospitals in the United States (2,3). The appearance of such VRSA strains in clinical specimens is of obvious and grave concern, because the spread of VRSA isolates may seriously jeopardize the chemotherapy of multidrug-resistant S. aureus disease and raise the specter of untreatable staphylococcal infections (4,5).
The VRSA strain HIP11714 recovered from a dialysis patient in Detroit, Michigan has acquired the vanA gene complex most likely from a vancomycin-resistant Enterococcus faecalis strain coinfecting the diabetic wound and catheter insertion site of the patient (1). HIP11714 was also methicillin-resistant: it carried the heterologous mecA gene, the key component of wide spectrum ␤-lactam resistance. Clinical aspects of this case (1), the genetic background of HIP11714 (4), and the structure of the S. aureus plasmid pLW1043 into which the transposon Tn1546 integrated 2 are the subject of separate publications. The purpose of the study described here was different: we wanted to examine how two very different antibiotic resistance mechanisms, based either on mecA or on vanA, are expressed in the same S. aureus cell and what impact the enterococcal vanA gene complex may have on the structure of the staphylococcal cell wall. The mechanism of vanA-mediated vancomycin resistance, at least when it is expressed in enterococci, involves the production of a cell wall precursor with abnormal chemical structure (for recent review see Ref. 6). The mechanism of mecA-based ␤-lactam resistance involves the production of a new cell wall biosynthetic enzyme, the penicillin-binding protein 2A (PBP2A) (7)(8)(9). To understand interactions between these two resistance mechanisms each targeting the bacterial cell wall, we transferred the vanA-carrying plasmid from the clinical strain HIP1174 to the MRSA strain COL, in which extensive biochemical and genetic information on cell wall structure (10 -12) and biosynthesis (13)(14)(15) is available. Strain COL does not carry chromosomal or plasmid-born regulatory elements that often control the transcription of mecA in clinical MRSA isolates (16,17). Strain COL is one of the fully sequenced MRSA strains (18) in which a large body of genetic information is also available concerning genetic determinants, auxiliary genes (19 -22), that are essential for the optimal expression of ␤-lactam resistance. These genetic determinants may provide useful tools for the dissection of the vanA-and mecA-mediated resistance mechanisms.

EXPERIMENTAL PROCEDURES
Construction of Strain COLVA by Filter Mating-In the filter mating experiments the recipient used was the MRSA strain COL, and the donor was the clinical VRSA isolate HIP11714, which carried transposon Tn1546 on pLW1043. Overnight cultures of donor and recipient grown in brain heart infusion (BHI) broth (Difco, Detroit, MI) were mixed in a 10 to 1 ratio in favor of the donor cells, filtered onto a Nalge 22-g filter under vacuum followed by incubation of the filters, "face up," on antibiotic-free BHI agar at 37°C for 18 h. After incubation, the bacteria were eluted into BHI broth and plated on agar containing 50 g/ml vancomycin plus 4 g/ml tetracycline (selective marker for the recipient). One of the five transconjugants was purified by two consecutive single colony isolations on the selective agar and was transferred to the Laboratory at the Rockefeller University where the purity and stability of the drug resistance trait was retested and the transconjugant was named strain COLVA.
Strains and Growth Conditions-S. aureus strains were grown in tryptic soy broth (TSB) (Difco) at 37°C with aeration. Growth was monitored by measuring the optical density at 600 nm with an LKB spectrophotometer (Amersham Biosciences, Uppsala, Sweden). In some experiments, peptidoglycan and cell wall precursor analyses were performed on bacteria grown in the presence of 125 or 250 g/ml vancomycin and also in 10 g/ml oxacillin. In these experiments the overnight cultures grown in TSB supplemented with the antibiotics were diluted into fresh TSB medium containing the same antibiotic concentration.
Testing Antibiotic Susceptibility-Antibiotic resistance level and synergistic activity of antibiotics were determined by the E-test following the recommendations by the manufacturer (Ab Biodisk, Solna, Sweden) and by the method of population analysis (23).
Preparation and Analysis of Peptidoglycan-Cell wall peptidoglycan was prepared, purified, and solubilized by enzymatic hydrolysis using previously described methods (10,11,13). The muropeptides liberated by the enzymatic hydrolysis were reduced by borohydride and separated by reversed-phase high performance liquid chromatography (HPLC) with a Shimadzu LC-10A HPLC system as described before (13). The peaks of interest were isolated and desalted by HPLC (13), and MS analysis was performed.
Preparation of UDP-linked Peptidoglycan Precursors and Analysis by HPLC-Cytoplasmic pools of the UDP-linked precursors were extracted by a method previously described (24) with some modification (25). In strains COL and COLVA grown in antibiotic-free medium, bacitracin (100 g/ml or 2ϫ MIC) was added to the cultures 30 min before harvesting the cells to amplify the amounts of cell wall precursors. Recent experiments in our laboratory have shown that bacitracin, an inhibitor of the regeneration of bactoprenylphosphate in the membrane transfer reaction of cell walls synthesis (26,27) can be successfully used instead of vancomycin (28) for the amplification of the cell wall precursor pool. Analysis was performed with a Shimadzu LC-10A HPLC system. Samples were applied to a 250-ϫ 4.6-mm reversed-phase HPLC column (ODS-Hypersil, 3 m, Keystone Scientific, Bellefonte, PA). The column was eluted at a flow rate of 0.5 ml/min with a linear gradient of 5% (v/v) FIG. 2. Differences in the muropeptide composition of the vancomycin resistant S. aureus strain COLVA grown in antibioticfree and in vancomycin-containing medium. Strain COLVA was grown in antibiotic-free TSB (B), TSB containing 250 g/ml vancomycin (C), and in TSB containing 10 g/ml oxacillin (D). As control strain COL was grown in TSB free of antibiotic (A). Cell wall peptidoglycan was prepared and enzymatic hydrolysates separated by HPLC. The numbers refer to main muropeptide components of strain COL (13), and the numbers followed by the letter v refer to the novel muropeptide components of the vancomycin-resistant strain COLVA. The arrows in B indicate muropeptide species that become dominant in the vancomycin grown bacteria (C). methanol in 100 mM ammonium formate (pH 3.5) to 30% (v/v) methanol in 100 mM ammonium formate (pH 3.5) in 30 min. The pH of buffers was adjusted by formic acid. Column temperature was 40°C. The eluted compounds were detected by absorption at 254 nm. This method was used to analyze precursors by HPLC coupled to mass spectrometer (LC-MS) (29,30). The flow rate was 0.5 ml/min for the HPLC separation, and it was reduced to 50 l/min before the ion source. To improve the separation of early precursors (UDP-MurNAc and UDP-MurNAc-Ala) the method was modified: the pH of the 100 mM ammonium formate was adjusted to pH 2.5 by HCl.
Mass Spectrometry Analysis-Samples of muropeptides were isolated by HPLC, lyophilized, and dissolved in H 2 O:CH 3

Expression of High Level Resistance to Oxacillin and Vancomycin in
Strain COLVA-Strain COLVA grown overnight either in TSB or in TSB supplemented with 100 g/ml vancomycin were tested for susceptibility to oxacillin and vancomycin by the E-test. The cultures grown in TSB showed high level resistance to oxacillin (MIC, 800 g/ml) (Fig. 1A) and the culture grown in vancomycin showed high-level vancomycin resistance (MIC: 512 g/ml) (Fig. 1B).
Composition of the Cell Wall Peptidoglycan of Strain COLVA Grown in Antibiotic-free and Antibiotic-containing Media-The HPLC elution profile of muropeptides recovered from the peptidoglycan of strain COL was compared with that of strain COLVA with both strains grown in antibiotic-free TSB. The two elution profiles were virtually identical, except that in strain COLVA next to many of the familiar muropeptide peaks of strain COL (13) (Fig. 2, top panel) there was often a small secondary peak eluting from the HPLC column with slightly shorter retention times (see arrows in Fig. 2, second panel). In COLVA grown in the presence of vancomycin these novel peaks became the dominant muropeptide species of the peptidoglycan (Fig. 2, third panel). Growth of COLVA in the presence of oxacillin (10 g/ml) caused a virtual disappearance of all but one (peak 3v) of the novel peaks and the peptidoglycan composition changed to a profile characteristic of strain COL grown in the presence of ␤-lactam antibiotics (13) (Fig. 2, bottom panel).
Structure of the Muropeptide Components in Strain COLVA Grown in Vancomycin-containing Medium-The novel muropeptide peaks that became dominant in the vancomycingrown cells were purified and subjected to mass spectrometric analysis. Table I shows the experimental and theoretical molecular masses obtained for the 22 new muropeptide species detected. The proposed chemical structures derived from the data are shown in Fig. 3. The most striking and novel features of these muropeptides are (i) the complete replacement of pentapeptides by tetrapeptides and (ii) the frequent deficit and/or complete lack of pentaglycine branches in muropeptides particularly in muropeptide monomers and in the biosynthetically "first" donor components of oligomeric muropeptides. Also apparent were structural modifications in some muropeptides that we interpret as evidence for lack of amidation of the stempeptide glutamic acid residues (muropeptide 4v) and the presence of deacetylated (16v and 17v) and O-acetylated (6v) hexosamines. The percent representation of various muropeptide peaks in bacteria grown under different conditions is illustrated in Table II.
Composition of the Cell Wall Precursor Pool- Fig. 4 shows the composition of the cytoplasmic cell wall precursor pool in strains COL and COLVA grown either in antibiotic-free medium or in media supplemented with antibiotics. These are the same growth conditions that were tested for their effect on the composition of the peptidoglycan in Fig. 2. In strains COL and COLVA grown in antibiotic-free TSB, the HPLC separation combined with mass spectrometric analysis identified six components in the precursor pool that were common and were a Muropeptide peak numbers are as in Fig. 1. The letters in parenthesis indicate multiple structures within the HPLC peak. b Experimental molecular masses were calculated from single-, double-, triple-, and tetra-charged ions of the reduced compounds. c Molecular masses were calculated from suggested amino acid composition of the reduced muropeptides.  Table I).

Expression of vanA in S. aureus
MurNAc-L-Ala-D-Glu (peak 4), and the pentapeptide derivative of UDP-MurNAc, UDP-MurNAc-L-Ala-D-Glu-L-Lys-D-Ala-D-Ala (peak 6) (35), the latter representing close to 80% of the precursor material. An additional minor peak 7 was identified in both COL and COLVA as a derivative of the UDP-MurNAcpentapeptide carrying a single glycine substitute on the ⑀amino group of the lysine residue. The precursor pool of COLVA (but not COL) also contained an additional minor component (peak 8), which had HPLC retention time and molecular mass characteristic of the UDP-MurNAc-depsipeptide, UDP-MurNAc-L-Ala-D-Glu-L-Lys-D-Ala-D-Lac, described in vancomycin-resistant enterococci (36,37). Radically different results were obtained when the cell wall precursor pool was analyzed from COLVA grown in vancomycin-containing medium (Fig. 4, third panel). Although peaks 1-4 appeared to be present unchanged, peak 6, the major component in the bacteria grown without vancomycin, was diminished and was replaced by two new major components (peaks 5 and 8). In addition several minor components were also present. Mass spectrometric analysis of peak 5 identified it as the UDP-MurNAc-L-Ala-D-Glu-L-Lys-D-Ala tetrapeptide. Peak 8 had a retention time of 21 min on the HPLC elution profile and a molecular mass that differed by 1 mass unit from that of the UDP-MurNAc-pentapeptide.
The composition of the cell wall precursor pool in strain COLVA grown in TSB supplemented with 10 g/ml oxacillin (Fig. 4, bottom panel) was indistinguishable from the precursor profile of COLVA grown in drug-free TSB, including even the presence of small amounts of the UDP-MurNAc-depsipeptide (peak 8). The percentage representation of various compounds in the cytoplasmic cell wall precursor pool in COL and COLVA grown under different conditions is shown in Table III. Fragmentation of the molecular ion at m/z 1151.6 gave two product ions at m/z 747.5 and 562.4 that corresponded to the loss of UDP and UDP-GlcNAc from the precursor ion, respectively (Fig. 5, A and C). The ion at m/z 562.4 corresponds to the structure Lac-L-Ala-␥-D-Glu-L-Lys-D-Ala-D-Lac. The first lactate at the N-terminal in this fragment came from the cleavage of muramic acid between N-acetylglucosamine and lactate.
To obtain the sequence of the peptide, further fragmentation (Fig. 5B)    the result of the loss of HCOOH and H 2 O from the peak 258.2, as was shown before (38). Taken together, the MS/MS data provide a rigorous proof that the UDP-pentapeptide isolated from the COLVA strain of S. aureus grown in the presence of vancomycin contains D-lactate instead of D-alanine as the Cterminal residue of the cell wall precursor molecule. the agar medium also contained oxacillin at a constant concentration of 40 g/ml. The number of bacterial colonies was determined after incubation at 37°C for 48 h. All cells grew on the agar plates containing 250 g/ml vancomycin, and the number of colonies dropped sharply on the plates containing 512 g/ml vancomycin. The presence of oxacillin reduced the vancomycin MIC value from 512 to 12 g/ml in 99.9% of the bacteria and generated a heterogeneous population profile (Fig. 6A). Fig. 6A was performed to test the effect of sub-MIC concentrations of vancomycin on the oxacillin resistance of strain COLVA. Inclusion of 50 g/ml vancomycin in the oxacillin plates reduced the oxacillin MIC of strain COLVA from 800 to 10 g/ml in 99.9% of the bacteria, and the culture became heterogeneous in the oxacillin-resistant phenotype (Fig. 6B).

Inhibition of the Expression of Oxacillin Resistance by Sub-MIC Concentrations of Vancomycin-An experiment similar to the one illustrated in
Effect of Inactivation of mecA on Vancomycin Resistance and Cell Wall Composition of Strain COLVA-The Tn551-inactivated mecA (39) was transduced into strain COLVA to generate mutant COLVA-⌬mecA. The results of E-test showed that COLVA-⌬mecA has retained unaltered high level resistance to vancomycin but resistance to oxacillin was reduced to 10 g/ml (Fig. 1B). Strain COLVA-⌬mecA was grown in TSB containing 250 g/ml vancomycin, and cell walls and precursor pool prepared and analyzed as described. The cell wall precursor pool showed a composition very similar to that of COLVA grown in vancomycin-containing medium (Table III). The composition of the cell wall peptidoglycan is shown in Table II. It is apparent that COLVA with an inactivated mecA produced a muropeptide HPLC profile that was very similar to the muropeptide profile of strain COLVA that carried an intact mecA. affinity for ␤-lactams (7-9). Vancomycin-resistant enterococci contain a cell wall precursor with altered C-terminal residue (depsipeptide), the bactoprenyl derivative of which (Lipid II) is only poorly bound by the antibiotic (6) (Fig. 7). Strain COLVA carries both types of drug resistance, and the primary purpose of the studies described here was to examine how these two mechanisms, one based on the mecA and the other on the vanA gene complex, are expressed in the same staphylococcal cell.
The results of E-tests and population analysis clearly demonstrate that strain COLVA carrying both mecA and vanA can produce cultures composed of homogeneous populations of bacteria that are resistant to high concentrations of either oxacillin (MIC: 800 g/ml) or vancomycin (MIC: 512 g/ml).
Analysis of the cytoplasmic cell wall precursor pool of COLVA is consistent with the full expression of the vanA gene complex in the presence of vancomycin. The UDP-MurNAcpentapeptide, which was the prominent component of the cytoplasmic pool of both strains COL and COLVA grown in antibiotic-free medium (Fig. 4), was diminished in quantity and was replaced by two new components in COLVA grown in vancomycin-containing medium. These two components were identified as the tetrapeptide and depsipeptide derivatives of UDP-MurNAc (Fig. 4C). These compositional shifts and the appearance of tetra-and depsipeptide compounds are similar to the altered chemical composition of the cell wall precursors described in vancomycin-resistant enterococci (40 -42).
A superficial glance at the HPLC profile of peptidoglycan hydrolysates generated from COL and COLVA indicates that the basic structural principles of S. aureus wall were retained. Thus, the relative amounts of high oligomers (eluting from the HPLC column with retention times longer than 100 min) were comparable: 38% in COL, 38.5% in COLVA grown in drug-free medium, and 37.4% in COLVA grown in the presence of vancomycin. The representation of monomeric and lower oligomeric muropeptides was also similar in COL (13.2 and 48.8%, respectively) and in COLVA grown in TSB (13.9 and 47.6%, respectively). However, in COLVA grown in vancomycin-containing medium, these percentages shifted in favor of the monomeric fraction (24.4%).
A closer comparison of the muropeptide profile of strains COL and COLVA reveals several major differences. The HPLC profile of COLVA grown in drug-free TSB shows a number of minor peaks that are absent from strain COL grown under the same conditions. These minor peaks become the dominant muropeptide species in the peptidoglycan of COLVA grown in the presence of vancomycin. Analysis of the new muropeptide species by mass spectrometry showed that they lacked the fifth alanine residue from both the monomeric and also from the cross-linked muropeptide species; they were also deficient in the pentaglycine side chains, which are the dominant feature of the cell wall of COL and other S. aureus strains (13). The structures drawn in Fig. 3 suggest that the deficit in the pentaglycine branches was particularly frequent among monomers (see structures 1v through 9v in Fig. 3) and among muropeptides that may have served as the biosynthetically "first" donors during the formation of muropeptide dimers, trimers, and oligomers (see structures 11v, 12v, and 14v in Fig. 3). Yet another type of structural abnormality not observed before in strain COL was the occasional lack of amidation of the stem peptide D-glutamic acid residues (4v or 6va); the excess of 42 molecular weight units that we interpreted as the possible presence of an O-acetyl group (6va) and suggestive evidence for deacetylated hexosamines (16va and 17va).
Although strain COLVA was clearly capable of expressing high level resistance to both oxacillin and vancomycin, these two inhibitors had strong and mutual antagonist affects on the expression of the drug-resistant phenotypes. As little as 40 g/ml oxacillin reduced the vancomycin MIC value of strain COLVA from 512 to 12 g/ml. Similarly, as little as 50 g/ml vancomycin reduced the oxacillin MIC of COLVA from 800 to 10 g/ml. In both cases the homogeneous phenotypes were also converted to heterogeneous ones (Fig. 6).
The experiments illustrated in Figs. 6 and 1 provide important insights into differences in the mechanism of cell wall synthesis in bacteria expressing either the ␤-lactam or the vancomycin resistance genes. It has been shown that in strain COL the assembly of peptidoglycan in the presence of oxacillin is catalyzed by two PBPs: the transpeptidase activity of PBP2A and the transglycosylase activity of PBP2 (14,15). The inhibition of expression of vancomycin resistance by oxacillin in strain COLVA suggests that the low affinity PBP2A, which is the only transpeptidase that remains active in the presence of oxacillin (13), is not able to utilize the depsipeptide cell wall precursors. The inhibition of vancomycin resistance by penicillin in enterococci has been interpreted in a similar manner. It was proposed that PBP5, which has low affinity to ␤-lactam, is unable to utilize depsipeptide wall precursors (43). Thus, the polymerization of UDP-MurNAc-depsipeptide precursor and its cross-linking in the peptidoglycan of strain COLVA must be catalyzed by one of the highly oxacillin sensitive native PBPs (Fig. 7). This interpretation is consistent with the composition of peptidoglycan of COLVA grown in the presence of oxacillin: except for the monomeric muropeptide 3v, all the other muropeptides were the "normal" structures characteristic of S. aureus grown in the presence of transpeptidase inhibitors (13). An direct experimental evidence for this interpretation came from the insertional inactivation of mecA in strain COLVA, which had no affect on the vancomycin MIC while it completely inactivated resistance to oxacillin (Fig. 1).
The sensitivity of the oxacillin-resistant phenotype to sub-MIC concentrations of vancomycin may be explained by the same mechanism. Induction of the vanA gene complex by vancomycin would lead to the destruction of the normal pentapeptide cell wall precursor (peak 6 in Fig. 4) and would flood the cytoplasmic cell wall precursor pool with the abnormal depsipentapeptide (peak 8 in Fig. 4). This abnormal cell wall precursor appears to be an inadequate substrate for PBP2A, which is the only transpeptidase available for wall synthesis in the presence of oxacillin in the medium. The inability of PBP2A to utilize depsipeptide precursors is reminiscent of early observations that documented the high degree of sensitivity of this protein to modification of the structure of the C termini of cell wall precursors (44,45).
Our findings indicate that the vanA gene complex can be fully expressed in S. aureus with several features of the mechanism that were already identified in vancomycin-resistant enterococci (46). The novel aspect of our study was that we could follow the vanA-induced changes in cell wall metabolism from the composition of wall precursor pool all the way to the composition of the peptidoglycan. Earlier studies found no difference in the composition of the cell walls of vancomycinsusceptible and vancomycin-resistant E. faecium (47,48). However, a comparison of the cell wall structure of vancomycinsusceptible and vancomycin-resistant E. faecalis has noted the complete absence of pentapeptide components (either D-alanine or D-lactate) from the peptidoglycan of the resistant bacteria (38,49). It was proposed (49) that the D-lactate terminating acceptor peptides may be substrates of a DD-decarboxypeptidase the activity of which would remove the C-terminal Dlactate residues from the peptidoglycan of the resistant cells.
The complete replacement of pentapeptides by tetrapeptides in the peptidoglycan of COLVA grown in vancomycin-containing medium may be explained by a similar mechanism. However, the simultaneous presence of both normal (pentapeptide containing) and abnormal tetrapeptide-containing muropeptides in COLVA grown in antibiotic-free medium (Fig. 2B) implies that the enzyme involved must have a high degree of selectivity for the removal of terminal D-lactic acid residues.
The frequent absence or less than complete pentaglycine branch in muropeptide monomers of bacteria grown in the presence of vancomycin suggests that the depsipeptide containing Lipid II is not an ideal substrate for the FemX, -A, and -B proteins (50). Our data indicate that such abnormal depsipeptides carrying incomplete pentaglycine branches may nevertheless move forward in the biosynthetic pathway and incorporate preferentially as muropeptide monomers into the peptidoglycan. Such incomplete pentaglycine branches are also found in oligomeric components as the biosynthetically "first" donor muropeptides. This observation suggests that muropeptides defective in pentaglycine branches are not favored as acceptors in the transpeptidation reaction.
It was shown before that the sortase-dependent attachment of proteins occurs preferentially to pentaglycine branches in the cell wall of S. aureus (51). It remains to be seen if the deficit in pentaglycine branches in vancomycin resistant S. aureus also reduces attachment of surface-exposed proteins and the pathogenic potential of such strains.
The undiminished vancomycin resistance of the mutant strain of COLVA in which mecA was insertionally inactivated shows clearly that PBP2A, the key protein of ␤-lactam resistance, does not participate in the expression of vancomycin resistance but is dedicated to the biosynthesis of cell walls in the presence of ␤-lactam antibiotics. Polymerization of the depsipeptide-containing muropeptide must be catalyzed by some of the native PBPs of S. aureus. The two major antibiotic resistance mechanisms encoded by mecA and vanA residing in the same S. aureus appear to use different sets of enzymes for the assembly of cell walls.