An alkaline D-stereospecific endopeptidase with beta-lactamase activity from Bacillus cereus.

We purified a novel extracellular D-stereospecific endopeptidase, alkaline D-peptidase (D-stereospecific peptide hydrolase, EC 3.4.11.-), to homogeneity from the culture broth of the soil bacterium Bacillus cereus strain DF4-B. The Mr of the enzyme was 37,952, and it was composed of a single polypeptide chain. The optimal pH for activity was ∼10.3. The enzyme was strictly D-stereospecific toward oligopeptides composed of Dphenylalanine such as (D-Phe)3 and (D-Phe)4. The enzyme also acted to a lesser extent on (D-Phe)6, Boc-(D-Phe)4 (where Boc is tert-butoxycarbonyl), Boc-(D-Phe)4 methyl ester, Boc-(D-Phe)3 methyl ester, Boc-(D-Phe)2, (D-Phe)2, and others, but not upon their corresponding peptides composed of L-Phe, (D-Ala)n (n = 2-5), (D-Val)3, and (D-Leu)2. The mode of action of the enzyme was clarified with synthetic substrates ((D-Phe)2-D-Tyr and D-Tyr-(D-Phe)2) and eight stereoisomers of (Phe)3. The enzyme had β-lactamase activity toward ampicillin and penicillin G, although carboxypeptidase DD and D-aminopeptidase activities were undetectable. The gene coding for alkaline D-peptidase (adp) was cloned into plasmid pUC118, and a 1164-base pair open reading frame consisting of 388 codons was identified as the adp gene. The predicted polypeptide was similar to carboxypeptidase DD from Streptomyces R61, penicillin-binding proteins from Streptomyces lactamdurans and Bacillus subtilis, and class C β-lactamases. Thus, the enzyme was categorized as a new “penicillin-recognizing enzyme.”

Some peptidases act on peptides containing D-amino acids. Soluble Streptomyces carboxypeptidase DD catalyzes not only the transpeptidation reaction on the peptide intermediate in peptidoglycan biosynthesis, but also the hydrolysis of N ␣ , N ⑀diacetyl-L-lysyl-(D-Ala) 2 in water (1). A D-peptidase has been purified and characterized from an actinomycete, although it is not strictly specific toward peptides containing D-amino acids (2). In Enterococcus, the vanX gene product, (D-Ala) 2 hydrolase, plays a role in vancomycin resistance (3). The chemically synthesized "D-enzyme" of human immunodeficiency virus type 1, in which all of the amino acids were replaced with the corresponding D-amino acids, displays D-stereospecificity (4). We discovered D-aminopeptidase (EC 3.4.11.19) from Ochrobactrum anthropi and found that its primary structure is similar to the ␤-lactamases and penicillin-binding proteins (5,6). The enzyme acts mostly on peptides with D-Ala at the NH 2 terminus to yield D-amino acids and does not act on D-amino acid derivatives with bulkier substituents. We proposed that Daminopeptidase is a new "penicillin-recognizing enzyme" (1), based on its primary structure, inhibition by ␤-lactam compounds, and the ability to catalyze peptide bond formation in organic solvents, although the enzyme does not show ␤-lactamase activity (6,7).
In this paper, we describe the screening of soil microorganisms for D-stereospecific endopeptidases using a synthetic peptide ((D-Phe) 4 ), characterization of the new enzyme alkaline D-peptidase (ADP), 1 as well as cloning and sequencing of the adp gene from Bacillus cereus strain DF4-B.
Synthesis of Substrates-Peptide substrates used to screen and test the substrate specificity were synthesized from D-and L-Phe. NH 2 and COOH termini were protected by Boc (8) and methyl groups, respectively. Isobutyl chloroformate (9) and carbodiimide (10) condensed the monomer to a dimer and the dimer to a tetramer, respectively. The following peptide derivatives were synthesized: Boc-D-Phe, D-Phe tertbutyl ester, (D-Phe) 2 ⅐HCl, Boc-(D-Phe) 2 , (D-Phe) 2 methyl ester⅐HCl, Boc-(D-Phe) 2 methyl ester, (D-Phe) 3  The details will be reported elsewhere. 2 Screening for (D-Phe) 4 -degrading Microorganisms-We screened the ability of microorganisms to hydrolyze (D-Phe) 4 in LB medium (11) in enriched cultures at 30°C. (D-Phe) 4 was dissolved in Me 2 SO (10%, w/v) and then added to 2 ml of LB medium containing soil samples. The mixture was then aerobically shaken for 2 days. A loopful of the culture broth was transferred to the same medium and aerobically incubated under the same conditions. A small portion of the culture broth was streaked onto a plate of the same medium containing 1.5% agar and incubated at 30°C overnight. Strains forming clear zones around the colonies were isolated, and (D-Phe) 4 degradation in the liquid culture was monitored by TLC (chloroform/methanol/acetic acid ϭ 8:2:1 or 10:2:1) and visualized with ninhydrin. (D-Phe) n (n ϭ 1-4) were well separated by TLC with the solvent. A bacterial strain isolated from a soil of Kanagawa Prefecture, Japan completely degraded the substrate by forming (D-Phe) 2 .
Identification of the Microorganisms That Degraded (D-Phe) 4 -Taxonomical studies of strain DF4-B showed that it was B. cereus because * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank TM  the Gram-positive rods were aerobic, spore-forming, motile, catalasepositive, egg yolk-positive, and lysozyme-resistant, and they formed acid from glucose (12). Details will be reported elsewhere. 2 Enzyme Assay and Definition of Units-ADP activity was routinely assayed at 30°C by measuring the production of (D-Phe) 2 from (D-Phe) 4 . The reaction mixture was composed of 1 mol of (D-Phe) 4 , 10 l of Me 2 SO, 50 mol of Tris-HCl, pH 9.0, and 1 mol of MgSO 4 , and the assay was started by addition of the enzyme in a total volume of 500 l. The reaction was terminated with 50 l of 2 N HClO 4 , and the amount of (D-Phe) 2 formed was estimated with a Waters 600E HPLC apparatus equipped with a Cosmosil 5C18-MS reverse-phase column (4.6 ϫ 150 mm) at a flow rate of 1.0 ml/min, using 35% methanol in 5 mM KH 2 PO 4 / H 3 PO 4 buffer, pH 2.9. Absorbance of the eluate was monitored at 254 nm. One unit of enzyme activity is defined as the amount of enzyme that catalyzes the formation of 2 mol of (D-Phe) 2 from 1 mol of (D-Phe) 4 .
The substrate specificity was examined qualitatively by thin-layer chromatography first and then was quantitatively assayed by the following methods. (a) The enzyme activity toward peptide substrates was measured as described above with 2.5 units of enzyme. The amounts of (D-Phe) 2 , D-Phe, L-Phe-D-Phe, Boc-D-Phe, and Boc-(D-Phe) 2 were quantitatively assayed by HPLC on a Cosmosil 5C18-MS reverse-phase column (4.6 ϫ 150 mm) at a flow rate of 1.0 ml/min, using the following solvent system of 5 mM KH 2 PO 4 /H 3 PO 4 buffer, pH 2.9: methanol ϭ 13:7 or 9:11 (v/v). To determine the kinetic constants of the enzyme for the peptide substrate, a reaction mixture containing 0.1-12.5 mol of the substrate, 50 mol of Tris-HCl, pH 9.0, 1 mol of MgSO 4 , 10 l of Me 2 SO, and 100 l of the enzyme solution in a total volume of 500 l was used. (b) The enzyme activity toward ␤-lactam compounds was determined by measuring the consumption of the substrate in a reaction mixture containing 100 mol of potassium phosphate buffer, pH 7.0, 2.5 mol of ␤-lactam compound, and 0.46 units of the enzyme in a total volume of 1 ml. After incubation at the same temperature, 100 l of the reaction mixture was added to 900 l of 100 mM KH 2 PO 4 /H 3 PO 4 buffer, pH 2.9, to stop the reaction. After centrifugation (15,000 ϫ g, 10 min), the amount of ␤-lactam compounds in the supernatant was determined by HPLC on an ODS-80Ts column (46 ϫ 150 mm) with a solvent of 25% methanol in 5 mM KH 2 PO 4 /H 3 PO 4 buffer, pH 2.9, and of 50 mM potassium phosphate buffer, pH 7.0, for ampicillin and penicillin G, respectively. To determine the kinetic constants for ␤-lactam compounds, potassium phosphate buffer, pH 7.0, was used in place of Tris-HCl, pH 9.0, and MgSO 4  Purification of ADP from the Culture Broth of B. cereus Strain DF4-B-Column chromatographies were done at temperatures lower than 5°C. Potassium phosphate buffer, pH 7.0, containing 0.1 mM EDTA was used throughout the purification. The strain was aerobically cultivated at 30°C for 48 h and shaken at 150 rpm in 352 batches in 2-liter flasks, each containing 400 ml of nutrient broth, which consisted of 1% meat extract, 1% polypepton, and 0.5% NaCl, in tap water, pH 7.0 (total of 140 liters). Cells were removed by centrifugation at 9500 ϫ g for 10 min. The supernatant was concentrated with an Amicon Hollow Fiber cartridge system (0.4 kg/cm 2 ), and a crude enzyme was obtained. The pellet formed between 50 and 80% ammonium sulfate saturation was dissolved in 0.01 M buffer and dialyzed with the same buffer. The enzyme solution was added to the first DEAE-Toyopearl column (6.0 ϫ 35 cm), which had been equilibrated with 0.01 M buffer. The enzyme was eluted with 0.1 M buffer containing 0.1 M NaCl, and the active fractions were concentrated by ammonium sulfate. The dialyzed enzyme solution was added to the second DEAE-Toyopearl column (5 ϫ 20 cm), which had been equilibrated with 0.01 M buffer. The column was eluted with a linear gradient of 0.01 M buffer (1.2 liters) to 0.1 M buffer containing 0.1 M NaCl (1.2 liters). The active fractions were then brought to 30% ammonium sulfate saturation. The enzyme solution was added to the first Butyl-Toyopearl 650 M column (5 ϫ 20 cm). The active fractions were eluted with a linear gradient of ammonium sulfate (30 to 0% saturation) in 0.01 M buffer. The active fractions were brought to 30% ammonium sulfate saturation. The enzyme solution was added to the second Butyl-Toyopearl 650 M column (5 ϫ 12 cm). The active fractions that eluted with a linear gradient of ammonium sulfate (30 to 0% saturation) in 0.01 M buffer were combined and brought to 80% ammonium sulfate saturation. The mixture was centrifuged at 28,000 ϫ g for 20 min, and the resulting pellet was dialyzed against the same buffer. The enzyme solution was added to the third DEAE-Toyopearl column (5 ϫ 12 cm). The enzyme was eluted with a linear gradient of 0.01-0.1 M buffer containing 0.2 M NaCl. The active fractions were combined, dialyzed, concentrated by ultrafiltration, and applied to a column of Superdex 200 equilibrated with 0.05 M buffer containing 0.1 M NaCl. The column was eluted by fast liquid protein chromatography (Pharmacia) at 1.0 ml/min, and the active fractions were collected and concentrated by ultrafiltration.
Analytical Methods-1 H NMR spectra were measured with a JEOL EX 400 apparatus. Protein was assayed, SDS-polyacrylamide gel electrophoresis was done, and the M r of the enzyme was estimated as described previously (6). Purified ADP (720 g) was digested with lysyl-end peptidase (5.0 g; Wako, Tokyo, Japan) at 30°C for 16 h. The digest was separated by HPLC on a reverse-phase column (ODS-80Ts) in a 10 -80% linear gradient of acetonitrile containing 0.1% trifluoroacetic acid at a flow rate of 0.5 ml/min with continuous monitoring of the absorbance at 215 nm. To determine the NH 2 -terminal amino acid sequence, the enzyme samples were covalently bound to Sequelonarylamine and Sequelon-diisothiocyanate membranes and then analyzed with a Prosequencer 6625 automatic protein sequencer (Milligen/ Biosearch). The molecular mass of the enzyme was estimated with a PE-Sciex API III triple quadrupole mass spectrometer equipped with an ionspray ion source in the positive ion mode (Sciex, Thornhill, Ontario, Canada).
Cloning of the adp Gene and Construction of pADP1-B. cereus genomic DNA was isolated as described (6). Escherichia coli cells were cultured in LB medium with 100 g/ml ampicillin. Plasmids were purified using the QIAGEN plasmid purification kit. B. cereus genomic DNAs were partially digested with MboI and fractionated by sucrose density gradient ultracentrifugation (5-25%; 100,000 ϫ g, 16 h). DNA fragments of ϳ3-6 kb were purified and ligated into BamHI-digested and dephosphorylated pUC118 by T4 ligase. Ampicillin-resistant transformants expressing ADP activity were directly selected by monitoring halo formation from (D-Phe) 4 because the host E. coli JM109 cells did not show ADP activity. One of ϳ17,000 transformants exhibited ADP activity, and it harbored a plasmid designated pBDP2 with a 5-kb DNA insert. For subcloning, pBDP2 was digested with EcoRI and SalI, and the resulting 1.3-kb fragment was introduced into the EcoRI and SalI sites of pUC118 to yield pBDP22.
B. cereus genomic cDNAs were completely digested with EcoRI and SalI and then size-fractionated by sucrose density gradient ultracentrifugation to give DNA fragments of ϳ1.8 kb. These fragments were ligated by T4 ligase into pUC118 that had been digested with EcoRI and SalI and dephosphorylated and then introduced into E. coli JM109. Ampicillin-resistant transformants were screened with the DNA insert of pBDP22 as a probe by colony hybridization. One clone (designated pADP1) with an ϳ1.8-kb DNA insert was selected for further analysis.
DNA Hybridization and Sequencing-Colony hybridization and Southern blot hybridization using 32 P-labeled probes prepared by nick translation were performed as described by Maniatis et al. (11). The DNA sequence was determined by the dideoxy chain termination procedure (28) using ␣-35 S-dCTP or with an automatic gene sequencer (ALF red sequencers, Pharmacia).

Screening for the (D-Phe) 4 -degrading Peptidase Producer and
Its Growth Rates-B. cereus was isolated from a soil sample and considered as a likely source of the enzyme. When the strain was cultured in 400 ml of nutrient broth, the enzyme activity was detected in the culture broth, and its formation was associated with the growth of the microorganism. We harvested the supernatant at 48 h to avoid lowering the specific activity of the enzyme by cell lysis.
Purification of Alkaline D-Peptidase and Molecular Weight-Since the enzyme was constitutive, the substrate (D-Phe) 4 was omitted from the medium in large-scale culture. The enzyme isolated from the supernatant of 140 liters of culture broth was electrophoretically pure. A summary of the purification procedure for the enzyme is shown in Table I. The enzyme was purified ϳ300-fold with an 8% yield.
The enzyme was judged to be homogeneous by SDS-polyacrylamide gel electrophoresis and HPLC on a TSK G-3000 SW column, as each of these procedures yielded a single band or a single peak. Fig. 1 (A and B) shows the results of SDS-polyacrylamide gel electrophoresis and the estimation of the M r by HPLC, respectively. The M r of the subunit calculated was ϳ36,000 as determined by SDS-polyacrylamide gel electrophoresis. That of the native enzyme was ϳ37,000 according to gel filtration chromatography, indicating that the native enzyme was a monomer. Mass spectrophotometry revealed that the M r of the enzyme was 37,952. The absorption of the purified enzyme in 0.01 M potassium phosphate buffer, pH 7.0, was maximal at 281 nm.
Substrate Specificity and Kinetic Properties-The substrate specificity of the enzyme was examined as shown in Table II. The enzyme was active toward (D-Phe) 3  These properties indicate that the enzyme is an endopeptidase that acts D-stereospecifically upon peptides composed of aromatic D-amino acids. On the other hand, a dimer was formed when D-Phe methyl ester and D-Phe amide were the substrates. Eight stereoisomers of the phenylalanine trimer were synthesized, and their effectiveness as substrates for the enzyme was tested. The enzyme recognized the configuration of the second D-Phe of tripeptides and catalyzed the hydrolysis of the second peptide bond from the NH 2 terminus. The calculated V max /K m values for the peptides containing L-Phe were lower than those for (D-Phe) 3 . The enzyme also showed ␤-lactamase activity toward ampicillin and penicillin G. The calculated V max values of the enzyme for ␤-lactam compounds were about the same as those for (D-Phe) 3 and (D-Phe) 4 , while the K m values were several hundred times larger. On the other hand, carboxypeptidase DD (13) and D-aminopeptidase (6) activities were undetectable.
The following compounds were inert as substrates: (L-Phe) 4 , (L-Phe) 3 (26,600), and trypsin inhibitor (20,100)); lane 2, purified ADP (10 g). B: elution profile of the enzyme and determination of the M r of the enzyme. The protein standards (E) in order of decreasing M r were glutamate dehydrogenase (290,000), lactate dehydrogenase (142,000), enolase (67,000), adenylate kinase (32,000), and cytochrome c (12,400). Absorbance at 280 nm is expressed as the relative absorbance. The relative absorbance represents the percentage of full-scale deflection on the recorder expressed from 0 to 1.0 on the ordinate.
Time course of (D-Phe) 4 Hydrolysis and Mode of Action of the Enzyme-We measured the time course of the (D-Phe) 4 degradation. As shown in Fig. 2, (D-Phe) 4 was hydrolyzed to (D-Phe) 2 and D-Phe. No (D-Phe) 3 was detected. These results coincide with the kinetic properties of the enzyme described above. The mode of action of the enzyme was examined with the synthetic substrates D-Tyr-(D-Phe) 2 and (D-Phe) 2 -D-Tyr as shown in Fig.  3. When D-Tyr-(D-Phe) 2 was the substrate, D-Phe was released first, and then D-Tyr was slowly formed. When (D-Phe) 2 -D-Tyr was used as the substrate, D-Tyr was released first, and then D-Phe was slowly formed. In both reactions, the second peptide bond from the NH 2 terminus of the substrate was hydrolyzed first. These results show that the enzyme acts as a D-stereospecific dipeptidyl endopeptidase.
Cloning and Nucleotide Sequencing of the adp Gene-E. coli  (D-Phe) 6 1.8 a (D-Phe) 4 100 a 0.398 199 500 (D-Phe) 3 90 JM109 transformants expressing ampicillin resistance and ADP activity were screened by halo formation on LB agar containing (D-Phe) 4 . Plasmid pBDP22 with a 1.3-kb insert conferred ADP activity in E. coli and was found to contain a lacZ-adp gene fusion, which encodes a ␤-galactosidase (1-12 amino acids)-ADP ( 250 GAT to the COOH terminus) hybrid protein, by sequence analysis. Genomic Southern hybridization using the DNA insert of pBDP22 as a probe indicated that the 1.8-kb EcoRI-SalI fragment contained the full-length adp gene. One clone (pADP1) that transformed E. coli into expressing ADP and that carried a 1.8-kb insert was further analyzed. Transcription in the plasmids appeared to be governed by an extant promoter of the adp gene because ADP activity was expressed when the direction of the transcription of the insert was opposite that of the original plasmid with pUC119. The nucleotide sequence of the 1.8-kb EcoRI-SalI fragment revealed a single open reading frame (ORF) that probably initiates at the 106 ATG codon preceded by a potential ribosomebinding site ( 91 GGAG). Translation of the ORF encoded a predicted protein of 388 amino acids with an M r of 42,033, with an amino acid sequence identical to those obtained by NH 2 -terminal amino acid sequencing of the six peptides prepared from purified ADP as shown in Fig. 4. Considering that ADP was secreted in the culture broth and that the M r of the predicted ORF (42,033) was larger than those estimated by SDS-polyacrylamide gel electrophoresis and HPLC, ADP is synthesized with a signal peptide. In fact, the predicted ORF exhibits a positively charged NH 2 terminus, followed by a hydrophobic stretch with a high leucine content. This domain closely resembles those of the signal peptides of exported proteins in Bacillus species (14). The NH 2 -terminal amino acid was suggested to be serine ( 223 AGT) based on the mass spectrometry results (M r 37,952) with purified ADP. This ORF ( 223 AGT to 1267 AAG) encodes a protein with a calculated M r of 37,926, which is in agreement with those estimated by other methods. The observed difference of 26 mass units between the M r deduced from the primary structure and that calculated by mass spectrometry was probably caused by the formation of an oxazolidinone ring at the NH 2 -terminal Ser. However, the exact molecular structure of the NH 2 terminus is not clear.
Deduced Primary Sequence and Similarities to Other Proteins-Alignment by the SWISS-PROT and NBRF-PIR data bases using the BLAST, FASTA, and DNASIS programs showed that the deduced primary structure of ADP is similar to FIG (20), and esterase from Pseudomonas sp. (30.5% identical over 154 amino acids) (21). Fig. 5 shows the results of the alignment of the primary structure of ADP with that of Streptomyces R61 carboxypeptidase DD. The sequence Ser-Xaa-Xaa-Lys is perfectly conserved in this class of enzymes, as shown in Fig. 6. A triad sequence similar to the His-Tyr-Gly commonly found in ␤-lactamases could not be found (22). DISCUSSION In this study, we synthesized (D-Phe) 4 in eight steps to screen for a new microbial endopeptidase. A strain degrading (D-Phe) 4 , indicated by a clear zone on LB agar plates, was isolated and identified as B. cereus. We purified the enzyme 300-fold to homogeneity from 140 liters of culture broth in 8% yield. We designated the enzyme as alkaline D-peptidase. It is the first D-stereospecific endopeptidase isolated with ␤-lactamase activity.
Since the NH 2 -terminal amino acid of the natural enzyme could not be determined, probably because it was modified, the M r of the enzyme was determined by mass spectrometry. The observed M r of the enzyme was 37,952. Serine ( 223 AGT) appears to be the first residue of the mature protein for the following reasons. Bacillus species seem to prefer small neutral amino acid residues at positions Ϫ1 and Ϫ3 of the signal peptide (14), especially at Ala Ϫ3 -Xaa-Ala Ϫ1 , where cleavage occurs after the carboxyl-terminal alanine. However, both of the alanine residues are occasionally substituted by other amino acid residues with a short side chain. In fact, ␤-mannosidase produced by Bacillus sp. has the sequence Ser Ϫ3 -Met-Ser Ϫ1 -Ser ϩ1 (23), which is similar to the sequence Ser Ϫ3 -Val-Ser Ϫ1 -Ser ϩ1 found in ADP (Fig. 4). However, assuming that the NH 2 -terminal amino acid of the secreted mature protein is serine ( 223 AGT), the signal peptide of ADP (39 amino acids) is longer than those of other exported proteins of bacilli (18 -35 amino acids) (14). Therefore, ADP might be synthesized as a propeptide with a few extra residues between the signal peptide and mature regions, like the ␣-amylase from B. subtilis (24) and the ␤-lactamases from B. cereus (25) and Bacillus licheniformis (14).
We synthesized several derivatives of D-Phe oligopeptides. The enzyme showed D-stereospecific dipeptidyl aminopeptidase and endopeptidase activities. The calculated V max /K m values for (D-Phe) 3 and (D-Phe) 4 were ϳ1000 times higher than for (D-Phe) 2 4 , forming (D-Phe) 2 , which was then slowly hydrolyzed to D-Phe. The product (D-Phe) 3 was not formed. The enzyme had an optimal pH of ϳ10.3, and it was stable in a relatively wide pH range. The enzyme was inhibited by phenylmethylsulfonyl fluoride, indicating that a serine residue is responsible for exerting the enzyme activity. Thus, the enzyme was named alkaline Dpeptidase (D-stereospecific peptide hydrolase, EC 3.4.11.-). Although the enzyme showed ␤-lactamase activity toward ampicillin and penicillin G, carboxypeptidase DD (13) and D-aminopeptidase activities were not detected. The kinetic measurements have clearly shown that the enzyme prefers (D-Phe) 3 and (D-Phe) 4 to the penicillins. ADP activity similar to that described in this study has not been reported, although some ␤-lactamases and carboxypeptidase DD are known to be active toward acyclic depsipeptides (26,27). Because the enzyme acts on aromatic ring-containing substrates and is excreted into the medium, the enzyme might have evolved to be active toward the synthetic substrates ampicillin and penicillin G.
We found that ADP (350 amino acids after modification) is homologous to the sequence of carboxypeptidase DD from Streptomyces R61 (13), penicillin-binding proteins, class C ␤-lactamases, and D-aminopeptidase from O. anthropi (6). As shown in Fig. 6, the sequence Ser-Xaa-Xaa-Lys is perfectly conserved in this class of enzymes, and the consensus sequence is located ϳ60 residues from the NH 2 termini of most of the enzymes. Thus, we propose that alkaline D-peptidase from B. cereus could be categorized as a new member of the penicillinrecognizing enzymes, which include penicillin-binding proteins, ␤-lactamases, and D-aminopeptidase.