The Basis for Resistance to (cid:1) -Lactam Antibiotics by Penicillin-binding Protein 2a of Methicillin-resistant Staphylococcus aureus *

Penicillin-binding protein 2a (PBP2a) of Staphylococcus aureus is refractory to inhibition by available (cid:1) -lac-tam antibiotics, resulting in resistance to these antibiotics. The strains of S. aureus that have acquired the mecA gene for PBP2a are designated as methicillin-re-sistant S . aureus (MRSA). The mecA gene was cloned and expressed in Escherichia coli , and PBP2a was purified to homogeneity. The kinetic parameters for interactions of several (cid:1) -lactam antibiotics (penicillins, cephalosporins, and a carbapenem) and PBP2a were evaluated. The enzyme manifests resistance to covalent modification by (cid:1) -lactam antibiotics at the active site serine residue in two ways. First, the microscopic rate constant for acylation ( k 2 ) is attenuated by 3 to 4 orders of magnitude over the corresponding determinations for penicillin-sensitive penicillin-binding proteins. Second, the enzyme shows elevated dissociation constants ( K d ) for the non-covalent pre-acylation complexes with the antibiotics, the formation of which ultimately would lead to enzyme acylation. The two factors working in concert effectively prevent enzyme acylation by the antibiotics in vivo , giving rise to drug resistance. Given the oppor-tunity to form the acyl enzyme species in in vitro experiments, circular dichroism measurements revealed that the in gel, using Fluorimager. contri-bution each the (cid:1)

Emergence of bacterial strains designated as methicillinresistant Staphylococcus aureus (MRSA) 1 from the 1960s to the present has created clinical difficulties for nosocomial infections worldwide (1). The genetic determinant for this resistance is mecA, which is not native to S. aureus but has been acquired by it many times over the past 40 years from unknown sources (2). The gene product of mecA is a penicillin-binding protein (PBP) designated PBP2a. S. aureus normally produces four PBPs (3), enzymes that are anchored on the cytoplasmic membrane, the functions of which are the assembly and regulation of the latter stages of the cell wall biosynthesis (4,5). Whereas these four PBPs are susceptible to modification by ␤-lactam antibiotics, an event that leads to bacterial death, PBP2a is refractory to the action of all available ␤-lactam antibiotics. PBP2a is capable of taking over the functions of the four typical PBPs of S. aureus in the face of the challenge by ␤-lactam antibiotics.
The pharmaceutical industry responded by initiating research programs in the discovery of novel ␤-lactams that will inhibit PBP2a. A few cephalosporins have been identified, of which a handful has now advanced into clinical trials for MRSA treatment (6 -8). Also, the clinical urgency has been met in the past few years by the introduction of Synercid (a combination of quinupristin and dalfopristin) (9), daptomycin (10), and linezolid (an oxazolidinone) (11) for treatment of MRSA. However, resistance to all of these agents exists, and the recent emergence of variants of MRSA resistant to linezolid (12) and glycopeptide antibiotics (13)(14)(15)(16) has created a situation in which certain strains of S. aureus are either treatable only with a single class of antibiotics or are simply not treatable, which is a disconcerting situation clinically.
As ␤-lactams (penicillins, cephalosporins, carbapenems, etc.) arguably remain the most important antibiotics clinically (55% of all antibiotics used globally belong to this class), it is imperative to understand the molecular mechanisms for resistance to these antibiotics. This mechanistic knowledge, along with structural information (17), should prove indispensable in devising strategies to circumvent the clinical problem presented by MRSA. In this vein, we report herein our cloning, expression, and purification of PBP2a of S. aureus. Kinetics were carried out with a series of ␤-lactam antibiotics to explore their interactions with PBP2a. Furthermore, we report on our results in understanding the incremental steps in the catalytic process of PBP2a. The enzyme undergoes a substantial conformational change in the course of its interactions with ␤-lactam antibiotics, which should have implications for the catalytic events of PBP2a in cross-linking of the bacterial cell wall.

EXPERIMENTAL PROCEDURES
Cloning of PBP2a-Chromosomal DNA of S. aureus ATCC706986 was isolated using a DNeasy tissue kit (Qiagen). The mecA gene was amplified without the sequence encoding its 23-amino acid-long Nterminal anchoring region using two oligonucleotide primers, SAPBP-2a-1D (TAATCCATGGCTTCAAAAGATAAAGAAAT) and SAPBP2a-R (TAATAAGCTTCTGTTTTGTTATTCATCTATAT). These primers contain recognition sequences for the restriction endonucleases NcoI and HindIII (italicized) that were utilized for cloning of the PCR product into the corresponding sites of expression vector pET24d(ϩ). Recombinant plasmid was initially used to transform competent cells of Escherichia coli JM83. Both DNA strands of the mecA gene from several transformants were sequenced, and recombinant plasmid was used to transform competent cells of E. coli BL21(DE3).
Site-directed Mutagenesis-To produce mutants of PBP2a, the mecA gene was recloned from the pET24d(ϩ) vector into the smaller vector pUC19 using sites for the restriction endonucleases XbaI and HindIII. The QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) was utilized to produce three mutant derivatives of PBP2a. To obtain the K406A mutant two mutagenic primers, MecS/A-d (CCAGG-TTCAACTCAAGCAATATTAACAGCAATG) and MecS/A-r (CATTGCT-GTTAATATTGCTTGAGTTGAACCTGG), that include the codon GCA (in bold) for alanine instead of that for lysine (AAA) were used. Two other mutagenic primers, SaurY519PD (GCTGATTCAGGTTTCGGAC-AAAGTGAAAT) and SaurY519PD (ATTTCACTTTGTCCGAAACCTG-AATCAGC), that contain the TTC codon for phenylalanine (in bold) were used to introduce the Y519F mutant derivative of PBP2a. The double mutant enzyme, K406A/Y519F, was produced by introducing a second substitution into the K406A mutant derivative. After mutagenesis the nucleotide sequence for each of the genes producing mutant enzymes was verified, and these genes were recloned between the NcoI and HindIII sites of the pET24d(ϩ) expression vector.
Expression of Wild-type PBP2a and Its Mutant Variants K406A, K406A/Y519F, and Y519F in E. coli-The wild-type PBP2a and K406A, K406A/Y519F, and Y519F mutant variants were each expressed using the same method. E. coli BL21 (DE3) was transformed with the plasmid pET24d(ϩ), which contained the wild-type and mutant mecA gene in its multiple cloning site. A 3-ml overnight seed culture was used to inoculate 500 ml of the LB medium supplemented with kanamycin (30 g/ml). Cells were grown at 37°C with shaking (120 rpm) until the A 600 reached ϳ0.8 (about 6 h) followed by the addition of 0.4 mM isopropyl-␤-D-thiogalactopyranoside to induce expression. The bacterial culture was then incubated at 25°C for another 20 h. Cells were harvested by centrifugation at 5500 ϫ g for 10 min at 4°C, and the pellet was suspended in 10 mM Tris/HCl buffer, pH 8 (buffer A).
Purification of Wild-type and Mutant PBP2a-The wild-type PBP2a and its mutants K406A, K406A/Y519F, and Y519F were each purified using the same three-step purification protocol with an LP chromatography system (Pharmacia) at 4°C. Cells were disrupted by 30 cycles of sonication (20 s of burst and 20 s of rest for each cycle) using a Branson sonifer. The resulting supernatant was then centrifuged at 14,000 ϫ g for 25 min using a Beckman-Coulter centrifuge. Pelletting, suspension in buffer A, and sonication were each repeated three times to ensure a high yield. The resultant cell-free extract was loaded at 2 ml/min onto a Q-Sepharose column (2.5 ϫ 30 cm; 80 ml of High Q support resin, Bio-Rad) equilibrated with buffer A. The proteins were eluted with a linear gradient of 0 -0.3 M NaCl in buffer A at 4 ml/min (total volume of 800 ml). PBP2a eluted at ϳ0.10 -0.15 M NaCl as determined by SDS-PAGE.
The fractions containing PBP2a were combined, concentrated, and brought to 1.5 M (NH 4 ) 2 SO 4 in buffer A. The combined solution was then loaded at 2 ml/min onto a phenyl-agarose column (2.5 ϫ 30 cm; 60 ml of phenyl-agarose resin, Sigma) equilibrated with 1.5 M (NH 4 ) 2 SO 4 in buffer A. The protein was eluted with a linear gradient of 1.5-0.5 M (NH 4 ) 2 SO 4 in buffer A at 4 ml/min (total volume of 600 ml). The fractions containing PBP2a eluted at ϳ1.2-0.8 M (NH 4 ) 2 SO 4 and were identified by SDS-PAGE.
The protein fractions were combined and concentrated, and the buffer was exchanged to 0.2 M NaCl in buffer B (50 mM sodium phosphate, pH 7.0) and loaded at 1.0 ml/min onto a Sepharose column (2.5 ϫ 50 cm; 160 ml of High S support resin, Bio-Rad) equilibrated with buffer B. The protein was eluted with a linear gradient of 0.2-1.0 M NaCl in buffer B at 1.5 ml/min to a final volume of 1500 ml. PBP2a was eluted from the column at ϳ0.6 -0.8 M NaCl. The fractions were combined, dialyzed against 25 mM Hepes in 1 M NaCl, pH 7.0. The protein concentration was determined with the BCA protein assay kit (Pierce). The yield from a 500-ml cell culture of either the wild-type PBP2a, the K406A mutant, or the Y519F mutant was ϳ20 mg. The double mutant K406A/Y519F yielded ϳ10 mg from a 500-ml cell culture. Each was concentrated to ϳ12 mg/ml. The wild-type and mutant proteins used in our experiments were all homogenous (data not shown). 13 C NMR Experiments-The wild-type PBP2a protein (5 mg) was dialyzed against several changes of degassed 25 mM sodium acetate buffer (pH 4.5) and then against degassed 50 mM sodium phosphate, 0.15 mM NaCl, 0.1 mM EDTA, pH 7.0. Subsequently, the protein was dialyzed against buffer containing 50 mM sodium phosphate, 0.15 mM NaCl, 10% D 2 O, and 20 mM NaH 13 CO 3 (the source of CO 2 ). The protein was concentrated to 0.15 mM. The 13 C NMR spectrum of the wild-type PBP2a protein indicated no modification of the protein by 13 C-labeled carbon dioxide. The procedure did not affect the quality of the protein because the pseudo first-order rate constants for acylation of the protein by nitrocefin with and without this treatment remained the same.
Determination of the Kinetic Parameters for Interactions of ␤-Lactam Antibiotics with the PBP2a Protein-PBP2a experiences acylation at the active site serine, and the acyl enzyme species slowly undergoes deacylation according to Equation 1.
E represents PBP2a, EI is the non-covalent pre-acylation complex, E-I is the covalent acyl enzyme species, and P denotes the product of hydrolysis of the ␤-lactam antibiotic. The first-order rate constants for protein acylation were determined for different ␤-lactam compounds using a Cary 50 UV spectrophotometer (Varian Inc.) at room temperature. The parameters for the reaction between PBP2a and nitrocefin were determined directly by monitoring the formation of the acyl enzyme species at 500 nM (⌬⑀ 500 ϭ ϩ15,900 cm Ϫ1 M Ϫ1 ). The experiments were carried out in 25 mM Hepes, 1 M NaCl (pH 7.0) buffer. The observed first-order rate constants (k obs ) were measured at a protein concentration of 2.5 M with different concentrations of nitrocefin (20 -120 M) and monitored for 45 min each, at which time the protein was invariably acylated.
Nitrocefin (120 M) was used as the reporter molecule to determine the apparent first-order rate constants for acylation by other nonchromogenic (or poorly chromogenic) ␤-lactams at varying concentrations in competition experiments (18).
The deacylation rate constants for wild-type PBP2a were determined using BOCILLIN FL as a reporter molecule (19). A typical reaction mixture (60 l) contained 15 M of PBP2a and a ␤-lactam antibiotic concentration at least 2-fold higher than its K d value. The mixture was incubated at room temperature for 45 min in 25 mM Hepes, 1 M NaCl (pH 7.0) buffer. The excess ␤-lactam was removed by passing the mixture through a Micro Bio-Spin®6 column (Bio-Rad). An aliquot (3 l) of the mixture was diluted 5-fold with the same buffer and incubated at room temperature for different time intervals. The amount of the free protein, liberated from the acyl protein species, was assayed by the addition of BOCILLIN FL to afford a final concentration of 40 M and incubated for an additional 45 min at room temperature. SDS sample buffer (15 l) was added to the reaction mixture, which was then boiled for 3 min. The samples (30 l in total) were loaded onto a 10% SDSpolyacrylamide gel, which was developed and then scanned using a Storm840® Fluorimager.
Circular Dichroism-The CD spectra of the wild-type PBP2a (6 M in 25 mM Hepes, 1 M NaCl, pH 7.0) were recorded on a Jasco J-600 instrument (Easton, MD, 5-mm path length) in the absence and presence of 30 M oxacillin or 30 M ceftazidime. The ␤-lactams had negligible CD readings compared with the protein. Regardless, the contribution of the ␤-lactam substrate was subtracted in each case. The proteins were incubated with the ␤-lactam antibiotics at 25°C.

RESULTS AND DISCUSSION
PBP2a has been cloned and studied by others (20 -25). We have cloned this protein for our studies as well. The mecA gene was PCR-amplified from the chromosomal DNA of S. aureus ATCC706986 without the 69 base pairs in the 5Ј-region encoding the 23-amino acid-long N-terminal membrane anchor. The gene was cloned between the NcoI and HindIII sites of the expression vector pET24d(ϩ), and PBP2a was produced intracellularly after induction with isopropyl-␤-D-thiogalactopyranoside. The gene was recloned into the XbaI and HindIII sites of the smaller plasmid pUC19 to produce mutant variants of the mecA gene efficiently. Subsequent to mutagenesis, the corresponding genes were recloned back between the NcoI and HindIII sites of the pET24d(ϩ) vector, and the enzymes were produced intracellularly by isopropyl-␤-D-thiogalactopyranoside induction. PBP2a was purified to apparent homogeneity in three chromatographic steps. We typically obtained ϳ40 mg of pure protein from a liter of culture.
We have evaluated the kinetics of interactions of three cephalosporins (nitrocefin, cefepime and ceftazidime), two penicillins (ampicillin and oxacillin), and one carbapenem (imipenem) with PBP2a (Table I). PBP2a, as with virtually all other known PBPs, undergoes acylation with its peptidoglycan substrate at an active site serine (Ser-403) for its transpeptidase activity (cell wall cross-linking). The acyl enzyme species then undergoes the transpeptidation reaction with another strand of the peptidoglycan. ␤-Lactam antibiotics subvert this process by undergoing the enzyme acylation process, but the resulting complex is often stable such that the enzyme is inactivated, and the organism is deprived of its vital function.
The processes for interactions of ␤-lactam antibiotics with PBP2a were sufficiently slow that the need for stopped-flow rapid kinetics was obviated. It is noteworthy that manifestation of resistance is because of both a slow rate of enzyme acylation (k 2 effect) as well as an absence of high affinity of the enzyme for ␤-lactams in general (K d effect). The t1 ⁄2 for enzyme acylation was in the range of 3 to 12 min with these antibiotics. This contrasts dramatically to t1 ⁄2 values of low milliseconds for typical penicillin-sensitive PBPs (26,27). The elevated dissociation constants for the pre-acylation complexes ranged between 180 and 1618 M, resulting in second-order rate constants (k 2 /K d ) of 1-19 M Ϫ1 s Ϫ1 . The rate constants for deacylation (k 3 ) of the acyl enzyme species were exceedingly poor, giving t1 ⁄2 values in the range of 26 to 77 h. Considering that S. aureus doubles its population size in 20 -30 min under favorable growth conditions, the formation of the acyl enzyme species is irreversible for practical purposes. In essence, the non-covalent encounters between the antibiotics and PBP2a are not favorable (high K d ), and the rate constants for enzyme acylation are exceedingly poor (slow k 2 ). Hence, formation of the acyl enzyme species would not take place in vivo for these two reasons. The fact that the acyl enzyme species with ␤-lactam antibiotics is extremely stable is irrelevant to the resistance problem, as the species would simply not form in vivo. Considering that PBP2a fulfills the critical physiological needs of the bacterium in the presence of ␤-lactam antibiotics, the set of events that led to the evolution of this important resistance enzyme to antibiotics is quite remarkable (28).
We hasten to add that the kinetic parameters that we report herein are somewhat different from those reported by Lu et al. (29), who used a mass spectrometric approach for analysis of kinetics. Whereas our K d values are sufficiently high to preclude enzyme acylation when considering the in vivo situation, the corresponding numbers by Lu et al. (29) were substantially higher than ours (high millimolar range).
The issue of activation of the active site serine is of interest. As will be discussed below, Ser-403 is well sheltered within the active site and its side chain hydroxyl is in contact with the side chain of Lys-406 (8). This arrangement of Ser-X-X-Lys for PBPs and related ␤-lactamases is understood to be important for the mechanisms of these enzymes (30). We have shown that when the corresponding lysine is mutated to alanine in the OXA-10 ␤-lactamase, the enzyme cannot undergo acylation by its substrate (31). A similar mutation in the penicillin-binding protein BlaR from S. aureus was shown to attenuate the rate of protein acylation by 6730-fold (32). The K406A mutant variant of PBP2a underwent extremely sluggish acylation. The effect was mostly on k 2 , which was attenuated by 80-to 130-fold for the mutant variant (Table II). Whereas the magnitude of the effect is relatively small, this level of attenuation reduces the already sluggish rate of acylation to the range of 10 Ϫ5 s Ϫ1 , which is the basal level that was attained for the BlaR protein and not far from the undetectable levels seen for the same mutation in the OXA-10 enzyme. Hence, in the cases of the BlaR and the OXA-10 proteins the acylation rate constants were higher, so the drops in their magnitudes were also larger on mutation. However, the basal level that we have observed for the lysine to alanine mutant variants in all three proteins were essentially the same.
Tyr-519 is another potentially basic residue within the active site. It could potentially provide the activation if it were unprotonated in the side chain and if the side chain were to undergo rotation from the position seen in the x-ray structure. Mutant enzyme variants Y519F and K406A/Y519F gave kinetic properties similar to the wild-type and to the K406A mutant, respectively. Therefore, this tyrosine residue does not play a role in catalysis, and Lys-406 is the basic residue that promotes the active site serine for enzyme acylation (Table II).
It is noteworthy that at least one penicillin-binding protein is now shown to be carboxylated in the side chain of its active site lysine (product of carbon dioxide addition to the lysine side chain amine) (16). In light of the reversibility of lysine carboxylation in proteins, there are known examples of lysine-carboxylated proteins that were identified by x-ray crystallography in their non-carboxylated forms. Hence, there was a possibility that PBP2a might be carboxylated at Lys-406. We carried out the diagnostic 13 C NMR experiment for detection of protein lysine carboxylation with PBP2a, as reported for other proteins previously (15,16). The experiment showed that PBP2a is not carboxylated at any lysine, and thus the crystal structure depicts the correct structure for Lys-406.
As shown in Fig. 1A, the x-ray structure of PBP2a reveals that the active site of the enzyme is not an open cleft. Indeed, the access to the active site is not obvious from the x-ray structure. Lim and Strynadka (17) have shown that the acyl enzyme species with ␤-lactam antibiotics largely maintains the active site in the same conformation with small movements within the immediate vicinity of the ligand away from that seen in the native enzyme (Fig. 1, B and C). A conformational change to open the active site would appear to be necessary both for the turnover events with the peptidoglycan substrate and for interactions with inhibitors such as ␤-lactam antibiotics.
The relatively slow nature of the kinetics of the interactions of ␤-lactam antibiotics with PBP2a indicated to us that these interactions might be studied by circular dichroism spectroscopy to explore the possibility of such protein conformational changes. We carried out these studies with oxacillin (a penicillin) and ceftazidime (a cephalosporin). Incubation of PBP2a with either oxacillin or ceftazidime resulted in dramatic conformational changes in the protein (Fig. 2), most readily observed at the minima at 208 and 222 nm, which are because of ␣-helices. As revealed in Fig. 2, A and C, the helix content decreased on exposure to the antibiotic, and a set of conforma-  tional changes was noted within the first four t1 ⁄2 values for acylation (for virtually complete protein acylation). These conformational changes continued for the duration of the monitoring for 3 days. In essence, the monitoring of the two wavelengths in the course of the experiments (Fig. 2, B and D) indicated that substantial conformational flexibility exists in the protein. The details of conformational changes were not identical in the two cases, reflecting the differences in the structures of the penicillin and cephalosporin used for these experiments. A fuller understanding of these differences should await structural-biological studies in the future. Whereas ϳ30% of the enzymic activity was lost at the end of 3 days of the CD experiment, the conformational state of the enzyme returned largely to the native state in both CD experiments. The relatively subtle conformational change seen for x-ray structures of the acyl enzyme species compared with the native structure (8) would not account for our observations in the CD experiments. Hence, the x-ray structure shows a complex that has settled, conformationally speaking, close to the native state, such as the species that we observed near the middle of the CD determinations (ϳ700 min for oxacillin and 1400 min for ceftazidime). Based on the k 3 values (Table I), by the end of the CD experiment, the acyl enzyme species are expected largely to have undergone hydrolysis to return to the native state. We underscore that these conformational changes are expected to be operative during the typical turnover events by this enzyme as well in light of the closed nature of the active site. A volume in excess of 1000 Å 3 is needed for the sequestration of the two peptidoglycan residues within the active site for the transpeptidase activity (33). The requisite conformational change would be expected to create this space for the catalytic events. Furthermore, these conformational changes must take place substantially more rapidly for the case of the peptidoglycan substrate. Although we cannot predict at the present what may precipitate these conformational changes, it is inherently intuitive that the polymeric peptidoglycan sub-FIG. 1. Active site of PBP2a from the x-ray structure. A, a stereo view to the active site environment from the x-ray structure of PBP2a is rendered as a solvent-accessible surface (Connolly surface, green), whereas important residues in the active site are shown in a capped sticks representation. A dotted Connolly surface (purple) is used to demonstrate the surface of the regions that cover the active site opening. B, a stereo view of the secondary structures (orange tube representation) and various important residues for the native enzyme structure is depicted. C, the penicillin G/PBP2a acyl enzyme complex is shown. (Penicillin G is shown in yellow, and capped sticks are color-coded according to atom types; oxygen, nitrogen, and carbon are shown in red, blue, and white, respectively.) The perspectives are the same for all three panels. strate would bind at a site outside of the immediate active site to initiate the processes.
A pertinent question on activity should be whether truncation by removal of the membrane anchor would affect activity. The conclusion from studies by others is that there is no consequential difference on activity with the loss of the membrane anchor (29). This is also entirely in accordance with the x-ray structure for PBP2a, which indicates that the point of insertion into the membrane by the membrane-spanning portion is quite distal to the catalytic domain (17).
In this study we have described the kinetics of interactions of six ␤-lactam antibiotics with the PBP2a of S. aureus. We also documented dramatic conformational changes for the protein in the presence of these antibiotics within the time scale for these turnover events. The function of PBP2a would appear to be more complex than previously appreciated. In light of the clinical importance of this protein to resistance to ␤-lactam antibiotics, a more complete understanding of these processes at the structural level is required. It is with such fuller understanding of these events that we may be able to conceive of strategies for inhibition of this deleterious bacterial enzyme in the near future.