Multiple Enzymatic Activities of the Murein Hydrolase from Staphylococcal Phage φ11

Bacteriophage muralytic enzymes degrade the cell wall envelope of staphylococci to release phage particles from the bacterial cytoplasm. Murein hydrolases of staphylococcal phages φ11, 80α, 187, Twort, and φPVL harbor a central domain that displays sequence homology to knownN-acetylmuramyl-l-alanyl amidases; however, their precise cleavage sites on the staphylococcal peptidoglycan have thus far not been determined. Here we examined the properties of the φ11 enzyme to hydrolyze either the staphylococcal cell wall or purified cell wall anchor structures attached to surface protein. Our results show that the φ11 enzyme has d-alanyl-glycyl endopeptidase as well asN-acetylmuramyl-l-alanyl amidase activity. Analysis of a deletion mutant lacking the amidase-homologous sequence, φ11(Δ181–381), revealed that the d-alanyl-glycyl endopeptidase activity is contained within the N-terminal 180 amino acid residues of the polypeptide chain. Sequences similar to this N-terminal domain are found in the murein hydrolases of staphylococcal phages but not in those of phages that infect other Gram-positive bacteria such as Listeria or Bacillus.

Bacteriophage muralytic enzymes degrade the cell wall envelope of staphylococci to release phage particles from the bacterial cytoplasm. Murein hydrolases of staphylococcal phages 11, 80␣, 187, Twort, and PVL harbor a central domain that displays sequence homology to known N-acetylmuramyl-L-alanyl amidases; however, their precise cleavage sites on the staphylococcal peptidoglycan have thus far not been determined. Here we examined the properties of the 11 enzyme to hydrolyze either the staphylococcal cell wall or purified cell wall anchor structures attached to surface protein. Our results show that the 11 enzyme has D-alanyl-glycyl endopeptidase as well as N-acetylmuramyl-L-alanyl amidase activity. Analysis of a deletion mutant lacking the amidase-homologous sequence, 11(⌬181-381), revealed that the D-alanyl-glycyl endopeptidase activity is contained within the N-terminal 180 amino acid residues of the polypeptide chain. Sequences similar to this N-terminal domain are found in the murein hydrolases of staphylococcal phages but not in those of phages that infect other Gram-positive bacteria such as Listeria or Bacillus.
The cell wall envelope of Gram-positive bacteria is a macromolecular, exoskeletal organelle that is assembled and turned over at designated sites (1). The cell wall functions not only to protect bacteria from osmotic lysis but also serves as a surface organelle that allows Gram-positive pathogens to interact with their environment, most notably the infected tissues of the host (2). To understand the mode of action of antibiotics and to identify new targets of antibacterial therapy, Staphylococcus aureus has been employed as a model organism to study the physiology of the cell wall of Gram-positive bacteria for almost 50 years (3)(4)(5). The characterization of muralytic enzymes has been instrumental in understanding cell wall turnover as well as determining the structure of peptidoglycan (6).
The peptidoglycan of S. aureus consists of a repeating disac-charide, N-acetylmuramic acid-(␤1-4)-N-acetylglucosamine (MurNAc 1 -GlcNAc) (4,7). The D-lactyl moiety of N-acetylmuramic acid is amide-linked to the short peptide component of peptidoglycan (8 -10). Wall peptides are cross-linked with other peptides attached to neighboring glycan strands, thereby generating a three-dimensional network that surrounds the staphylococcal cell (11)(12)(13). During cell wall synthesis the peptidoglycan precursor molecule, lipid II (C 55 -PP-MurNAc-(L-Ala-D-iGln-L-Lys-(NH 2 -Gly 5 )-D-Ala-D-Ala-COOH)-GlcNAc; C 55 -PP is undecaprenyl pyrophosphate), is incorporated into the peptidoglycan network via transglycosylation and transpeptidation reactions (14 -17). Whereas transglycosylation leads to the polymerization of the glycan strands, the transpeptidation reaction results in the cross-linking of the peptide backbone of the cell wall (18). During this reaction, the terminal D-Ala of the pentapeptide precursor (L-Ala-D-iGln-L-Lys-(NH 2 -Gly 5 )-D-Ala-D-Ala-COOH) is removed and the carboxyl of D-Ala at position four is linked to the free amino of the pentaglycine cross-bridge within cell wall peptides of neighboring peptidoglycan strands (19). Fig. 1 shows the structure of the staphylococcal peptidoglycan and the cleavage sites of muralytic enzymes. Several investigators employed biochemical techniques to isolate and characterize staphylococcal murein hydrolases, leading to the identification of N-acetylmuramyl-L-alanine amidase, N-acetylglucosaminidase, D-Ala-Gly endopeptidase, as well as Gly-Gly endopeptidase activities (6, 20 -23). The isolation of staphylococcal transposon variants that are defective in autolysis has permitted the assignment of specific functions to individual genes as well as sequence comparison between murein hydrolases from many different bacterial species (24,25). Murein hydrolases can be grouped into two separate classes of enzymes. The first group is the autolysins, for example S. aureus N-acetylmuramyl-L-alanine amidase (Atl), Nacetylglucosaminidase (Atl) (24), as well as the presumed Gly-Gly endopeptidase (LytM) (26). Autolysins are secreted from the bacterial cytoplasm by N-terminal signal peptides and degrade the cell wall peptidoglycan at specific sites during physiological growth and/or during stationary phase. Bacteriophage-encoded murein hydrolases, also called endolysins, are exported by holins, small polypeptides that are inserted into the cytoplasmic membrane of the bacterial host (27). Endolysins function to completely hydrolyze the cell wall peptidoglycan as a means to release bacteriophage particles from the bacterial cytoplasm.
The gene encoding the murein hydrolase of staphylococcal phage 11 has been sequenced (25). Sequence comparison revealed that this endolysin has a modular organization consisting of an N-terminal, a central, and a C-terminal domain. The central domain of the 11 enzyme displays sequence similarity with known N-acetylmuramyl-L-alanine amidases, whereas the C-terminal domain is homologous to the cell wall targeting signal of lysostaphin, a staphylolytic bacteriocin secreted by Staphylococcus simulans biovar staphylolyticus (23,28,29). Targeting signals are thought to function as binding domains that direct muralytic enzymes to specific receptors on the bacterial surface (28,30). The N-terminal domain of 11 murein hydrolase displays sequence similarity to endolysins of other aureophages (187, 11, 80␣, Twort, and PVL); however, a specific murein hydrolase function has thus far not been assigned to this domain (Fig. 1C).
Surface proteins of S. aureus are covalently attached to the peptidoglycan by a mechanism requiring a C-terminal sorting signal and conserved LPXTG motif (31,32). After export from the cytoplasm via an N-terminal signal peptide, the sorting signal is cleaved between the threonine and the glycine of the LPXTG motif. The carboxyl of threonine is linked to the amino group of the pentaglycine cross-bridge in the staphylococcal cell wall (33,34). Our previous work sought to characterize the cell wall anchor structure of staphylococcal surface proteins after their solubilization with murein hydrolases (32,35,36). When the peptidoglycan is cut with lysostaphin, a glycyl-glycine endopeptidase (29,37), purified surface protein migrates uniformly on SDS-PAGE (32). The solubilized polypeptides have two or three glycine residues linked to the carboxyl of threonine (34). In contrast, when solubilized with N-acetylmuramidase or N-acetylmuramyl-L-alanine amidase, surface proteins migrate as a spectrum of fragments on SDS-PAGE with increasing mass (35). This can be explained as surface protein linked to peptidoglycan with different amounts of cross-linked cell wall subunits.
We have used purified recombinant 11 murein hydrolase to solubilize staphylococcal surface proteins from the bacterial cell wall (36). When analyzed on SDS-PAGE, surface protein migrated as two distinct species, closely resembling the uniform migration of lysostaphin-released species but not the heterogeneous pattern observed for N-acetylmuramidase-or Nacetylmuramyl-L-alanine amidase-solubilized surface protein.
This was a surprising result, because the 11 murein hydrolase is known to display sequence homology with known N-acetylmuramyl-L-alanine amidases. Characterization of the 11 murein hydrolase-released cell wall anchor structures of surface proteins revealed the presence of linked cell wall tetrapeptide (L-Ala-D-iGln-L-Lys-(surface protein-Gly 5 )-D-Ala-COOH) and disaccharide-tetrapeptide (MurNAc-(L-Ala-D-iGln-L-Lys-(surface protein-Gly 5 )-D-Ala-COOH)-GlcNAc) (36). These observations suggested that the 11 murein hydrolase may cleave the peptidoglycan at two sites, the amide bond between N-acetylmuramyl and L-Ala and the peptide bond between D-Ala and Gly. We show here that the 11 murein hydrolase displays D-Ala-Gly endopeptidase as well as N-acetylmuramyl-L-alanine amidase activity. Furthermore, we demonstrate that the N-terminal domain of 11 murein hydrolase functions as a D-Ala-Gly endopeptidase.

EXPERIMENTAL PROCEDURES
Strains and Materials-Staphylococcal strain OS2 has been previously described (31) and was used as a host for the recombinant plasmid pHTT4 encoding the recombinant surface protein Seb-MH 6 -Cws (36). Escherichia coli strain BL21(DE3) (38) was used as a host for the expression of the 6His-tagged 11 and 11(⌬181-381) enzymes. The coding sequence of the 6His-tagged 11 amidase was contained on plasmid pHTT2 (36). The truncated mutant protein, 11(⌬181-381), was generated by the ligation of DNA fragments encoding the N-  (20)) illustrating the sites of hydrolysis for the 11 murein hydrolase, N-acetylmuramyl-L-alanine amidase (amidase), muramidase, glucosaminidase, and lysostaphin. B, schematic diagram of the murein hydrolases from staphylococcal phage 11 (GenPept accession number 113675 (25)), 80␣ (accession number 1763243 (64)), Twort (accession number 2764981 (43)), PVL (accession number 3341932 (65)), and 187 (accession number 2764983) (43). Also shown is a schematic diagram of the deletion construct 11(⌬181-301), generated by removal of the central amidase portion of the full-length 11 enzyme. C, sequence alignment of the predicted N-terminal domains of the enzymes shown in B. Optimal sequence alignment was carried out via the World Wide Web using the Multiple Sequence Alignment version 2.1 (66), provided by the Institute for Biomedical Computing at Washington University, St. Louis, MO.
terminal 180 amino acids and C-terminal 100 amino acids of the fulllength 11 protein. DNA encoding the C-terminal domain of the 11 hydrolase was polymerase chain reaction-amplified from pHTT2 DNA using the primers LA-Kpn (28) and 11-Bam6His (AAGGATCCCTAG-TGATGGTGATGGTGATGACTGATTTCTCCCCATAAGTC) (36). DNA encoding the N-terminal domain was polymerase chain reactionamplified from pHTT2 DNA using the primers 11-Nde (AACATATG-CAAGCAAAATTAACTAAAAAT) (36) and 11-Kpn2 (AAAGGTACCT-TCTACTGCTTTAGGTTGTGG). The resulting polymerase chain reaction products were digested with BamHI and KpnI or NdeI and KpnI and subsequently cloned in a three-way ligation into the NdeI and BamHI sites of the vector pET9a (38) to generate pWil54. Tryptic soy broth was purchased from Difco. Mutanolysin and isopropyl-1-thio-␤-D-galactopyranoside were purchased from Sigma. All other chemicals were purchased from Fisher unless otherwise noted.
Isolation of Seb-MH 6 -Cws from Staphylococcal Cell Walls-The purification of Seb-MH 6 -Cws surface protein from cell walls of S. aureus OS2 harboring the plasmid pHTT4 was carried out as described previously (35). Briefly, sedimented cells were washed in 50 mM Tris-HCl, pH 7.5, and were broken with a Bead-Beater instrument (Biospec Products, Bartlesville, OK). Walls were sedimented by ultracentrifugation and suspended in 50 -70 ml of wash buffer (100 mM potassium phosphate, pH 7.5, 1% Triton X-100). The suspension was stirred for 3-12 h at 4°C, and cell walls were sedimented by centrifugation for 15 min at 32,500 ϫ g. The walls were washed three times in 100 mM sodium phosphate, pH 6.0, followed by a single wash with water prior to digestion with either mutanolysin, 11, or 11(⌬181-381) enzymes (see below). Digestion of cell walls with mutanolysin was carried out in 100 mM sodium phosphate buffer, pH 6.0, as described previously (35). Digestion of cell walls with 11 or 11(⌬181-381) was carried out in 50 mM Tris-HCl, pH 7.5, 100 mM NaCl with approximately 1 mg of enzyme per liter of staphylococcal culture from which the walls were isolated. Buffers throughout the procedure were supplemented with 1 mM PMSF and 1 M pepstatin to prevent proteolysis of the Seb-MH 6 -Cws protein. Solubilized Seb-MH 6 -Cws protein was fast protein liquid chromatography purified by affinity chromatography on nickel resin as described previously (35).
Purification of 6His-tagged 11 and 11(⌬181-381) Enzymes-E. coli BL21(DE3) harboring either pHTT2 and pLysS (38) or pWil54 were grown overnight in LB medium containing either chloramphenicol (30 g/ml) and kanamycin (50 g/ml) or chloramphenicol alone, respectively. The culture was diluted 1:50 into fresh medium and grown with shaking at 37°C to an A 600 of 0.2 whereupon the expression of T7 polymerase was induced by the addition of isopropyl-1-thio-␤-D-galactopyranoside to a final concentration of 1 mM. Cells were harvested after 4 h by centrifugation at 8,000 ϫ g. Preliminary experiments revealed that both full-length and truncated 11 enzymes were contained in inclusion bodies, and subsequent purification steps were therefore carried out under denaturing conditions. The E. coli cell sediments were suspended in 50 ml of buffer A (6 M guanidine hydrochloride, 0.1 M NaH 2 PO 4 , 0.01 M Tris-HCl, pH 8.0); cells were lysed by passage through a French pressure cell at 14,000 pounds/square inch, and insoluble material was removed by centrifugation at 30,000 ϫ g for 20 min. Enzymes were fast protein liquid chromatography purified from the supernatant over a 1-ml bed volume column of nickel nitrilotriacetic acid resin (Qiagen). The column was pre-equilibrated with buffer Q1 (8 M urea, 100 mM NaH 2 PO 4 , 0.01 M Tris-HCl, pH 7.3) and, after loading with sample, was washed with an additional 10 ml of buffer Q1. The column was washed with 15 ml of 30% Q1 and 70% Q2 (8 M urea, 100 mM NaH 2 PO 4 , 0.01 M Tris-HCl, pH 4.2). Protein was eluted from the column in a pH-dependent manner by a step gradient to 100% Q2 buffer.
The purified enzymes, in approximately 5 ml of Q2 buffer, were folded by dialysis against 1 liter of dialysis buffer (50% glycerol, 50 mM sodium phosphate, 10 mM MgCl 2 , 2 mM dithiothreitol, pH 6.8) for 4 -6 h at 4°C without stirring and again against fresh dialysis buffer overnight with slow stirring. Insoluble protein was removed by centrifugation, and the concentration of soluble enzyme was measured by a modified Lowry assay (Bio-Rad) against a standard of bovine serum albumin. Enzyme preparations were stored at Ϫ20°C and used directly from stock for all assays.
Preparation of Staphylococcal Cell Walls-Crude cell walls used for Morgan-Elson and DNFB assays (see below) were prepared from a culture of S. aureus OS2 grown to post-log phase in tryptic soy broth. Cells were harvested by centrifugation for 15 min at 8,000 ϫ g, suspended in 50 mM Tris-HCl, pH 7.5, and broken with glass beads in a Bead-Beater instrument (5 pulses of 3 min each interspersed with 5 min cooling on ice). Crude cell walls were sedimented by ultra-centrifugation (100,000 ϫ g for 30 min) and suspended in 20 ml of a 1% SDS solution and boiled for 30 min. Crude walls were sedimented by centrifugation at 30,000 ϫ g and washed five times with water.
Highly purified staphylococcal peptidoglycan employed for Edman degradation and mass analysis of 11 and 11(⌬181-381)-solubilized muropeptides was prepared essentially as described by de Jonge et al. (39). S. aureus OS2 cells were grown to mid-log phase in tryptic soy broth, and culture was chilled to 4°C on ice. Cells were harvested by centrifugation, suspended in a solution of 4% sodium dodecyl sulfate (SDS), and boiled for 30 min. Cells were centrifuged for 10 min at 30,000 ϫ g and washed six times with water. The cells were broken with 0.2-mm glass beads in a Bead-Beater instrument as described above. Glass beads were removed by centrifugation at 200 ϫ g for 5 min. Broken cell walls were collected by centrifugation at 30,000 ϫ g for 15 min, and the pellet was suspended in 50 mM Tris-HCl, 10 mM MgCl 2 , pH 7.5. Walls were treated with ␣-amylase (Sigma, 100 g/ml), DNase (Sigma, 10 g/ml), and RNase A (Qiagen, 50 g/ml) for 2 h at 37°C. CaCl 2 was added to a final concentration of 10 mM, and protein was removed from the peptidoglycan by the addition of trypsin (Sigma, 100 g/ml) and incubating the walls overnight at 37°C. Walls were collected by centrifugation, washed two times with water, once with 8 M LiCl, once with 100 mM EDTA, twice more with water, and finally with acetone. Walls were suspended in water and stored at Ϫ20°C.
Preparation of C-terminal Anchor Peptides-Preparation of C-terminal anchor peptides from Seb-MH 6 -Cws was carried out as described previously (35). Purified Seb-MH 6 -Cws was methanol/chloroform-precipitated, dried under vacuum in a Speed-Vac concentrator (Savant), FIG. 2. RP-HPLC analysis of murein linked to mutanolysin-released anchor peptides redigested with 11 murein hydrolase. A, diagram of the hybrid surface protein Seb-MH 6 -Cws used in these studies. Labeled arrows indicate the specific cleavage sites for signal peptidase, cyanogen bromide, and sortase. B, mutanolysin-solubilized anchor peptides were reduced with sodium borohydride and digested with the 11 murein hydrolase. Enzyme was precipitated with 10% trifluoroacetic acid, and the acid-soluble muropeptides were separated by RP-HPLC. Absorbance was monitored at 206 nm, and fractions corresponding to major absorption peaks (A, B, and C) were analyzed by ESI-MS (see text). C, SDS-PAGE of Seb-MH 6 -Cws after solubilization from the cell wall with either mutanolysin (Mutano.) or mutanolysin solubilization followed by a second digest with the full-length 11 hydrolase (Mutano. ϩ 11). and suspended in 1-3 ml of 70% formic acid. Five mg of cyanogen bromide was added, and the cleavage reaction was incubated for 16 -18 h at room temperature in the dark. The cleaved peptides were dried under vacuum, washed twice with water, and suspended in 1-2 ml of buffer A. Samples were loaded onto a column packed with 1 ml of nickel nitrilotriacetic acid pre-equilibrated with 10 ml of buffer A. The Column was washed with 10 ml of each buffer A, buffer B (8 M urea, 100 mM NaH 2 PO 4 , 0.01 M Tris-HCl, pH 8.0), and buffer C (same as buffer B, but pH 6.3). Peptides were eluted with 2 ml of buffer D (6 M guanidine hydrochloride, 0.5 M acetic acid, pH 4.3) and subjected to rpHPLC on C18 column (2 ϫ 250 mm, C18 Hypersil, Keystone Scientific). Separation was carried out at 40°C at a flow rate of 0.2 ml/min with a linear gradient starting 10 min after injection from 99% H 2 O (0.1% trifluoroacetic acid) to 60% H 2 O and 40% CH 3 CN in 35 min followed by a steeper gradient to 99% CH 3 CN in 5 min. Elution of peptides was monitored at 215 nm, and 1-min fractions were collected.
HPLC Separation of Muropeptides-For the analysis of muropeptides removed from surface protein by digestion with the 11 enzyme, anchor peptides were solubilized with mutanolysin and, after purification, were redigested with the 11 murein hydrolase at a concentration of 400 g/ml overnight at 37°C. Digestion was terminated by the addition of trifluoroacetic acid to a concentration of 10%. Samples were placed on ice to precipitate the 11 enzyme, which was separated from the soluble muropeptides by centrifugation at 15,000 ϫ g for 15 min. Trifluoroacetic acid-soluble compounds (muropeptides) were desalted over a C18 cartridge (Analtech) and dried under vacuum. Dried muropeptides were suspended in 200 l of water. To the peptides, 200 l of 0.5 M sodium borate buffer, pH 9.0, was added followed by the immediate addition of 1-3 mg of solid sodium borohydride. Reduction of muropeptides was previously shown to be essential for effective sepa-ration by rpHPLC (40). The reaction was incubated for 30 min at room temperature and quenched by the addition of 20 l of 20% phosphoric acid.
The purification and rpHPLC analysis of muropeptides generated by the digestion of intact peptidoglycan with the 11 and 11(⌬181-381) enzymes was carried out using highly purified peptidoglycan as a substrate. Approximately 2 mg of pure staphylococcal peptidoglycan (see above) was digested with approximately 400 -500 g of enzyme in 1 ml of 50 mM Tris-HCl, 100 mM NaCl, pH 7.5, for 4 h at 37°C. Enzyme was precipitated by the addition of trifluoroacetic acid to 10% and centrifugation after 15 min on ice. Removal of linked teichoic acids from the soluble muropeptides was achieved by heating the soluble muropeptides to 60°C for 14 h. Muropeptides were desalted over a C18 cartridge and dried. The muropeptides were resuspended and reduced with sodium borohydride as described for the anchor muropeptides.
Separation of muropeptides by rpHPLC was carried out using a C18 column and a method devised for E. coli wall peptides (40) that was later modified for the separation of S. aureus muropeptides (39). The C18 column employed for the separation of muropeptides from anchor peptides (2 ϫ 250 mm, C18 Hypersil, Keystone Scientific) contained a slightly different resin than the C18 column employed for the separation of muropeptides from purified peptidoglycan (2 ϫ 250 mm, C18 Aquasil, Keystone Scientific), resulting in a discrepancy in the retention times observed between the two experiments. A linear gradient was started immediately after injection of 5% (v/v) methanol in 100 mM NaH 2 PO 4 (pH 2.5) to 30% (v/v) methanol in 100 mM NaH 2 PO 4 (pH 2.8) in 100 min. Azide was not added to the start buffer, and base-line drift was accounted for by subtracting the chromatogram of a blank run. Eluted muropeptides were monitored at 206 nm, and positive fractions were desalted using a C18 cartridge and dried under vacuum prior to analysis by ESI-MS or Edman degradation.
Chemical Analysis of Solubilized Muropeptides by Modified Morgan-Elson and DNFB Assays-Crude staphylococcal cell walls were resuspended in water (20 mg wet weight/ml) to make a stock solution. Into each reaction tube, 400 l of wall stock was mixed with 500 l of 100 mM NaH 2 PO 4 , pH 7.5 (or pH 6.0 for mutanolysin reaction), and 100 l of NaCl to give an initial A 600 of 1.1. Lysostaphin (10 g/ml), 11 enzyme (100 g/ml), 11(⌬181-301) (100 g/ml), or mutanolysin (50 units/ml)  were added, and the crude cell walls were digested overnight to an A 600 Ͻ0.1. Insoluble material was removed by centrifugation for 30 min at 13,000 ϫ g, and soluble material was transferred to another Eppendorf tube.
For the measurement of soluble N-acetylhexosamines by a modified Morgan-Elson reaction (22), 100 l of the soluble digestion products were mixed with 100 l of 2% K 2 B 4 O 7 in water. The samples were boiled for 30 min and allowed to cool to room temperature. p-Dimethylaminobenzaldehyde (Ehrlich's reagent) was dissolved into a total volume of 9.5 ml of acetic acid, and 0.5 ml of concentrated HCl was added to give a stock reagent. 1 ml of stock reagent was diluted per 7 ml of acetic acid to generate the color reagent. To the boiled muropeptides, 900 l of color reagent was added, and the mixture was kept at 37°C for 20 min. Color was read spectrophotometrically at 585 nm, and samples were compared with standards of GlcNAc of known concentration. GlcNAc standards were boiled in K 2 B 4 O 7 for 7 min instead of 30 min to compensate known discrepancies between the extinction coefficient of free N-acetylhexosamines and those that exist as disaccharides or are amidically linked to wall peptides.
The liberation of free amino groups was measured by mixing 100 l of the soluble wall digestion products with 100 l of 2% K 2 B 4 O 7 in water (22). To this, 20 l of DNFB solution (130 l of DNFB in 10 ml of ethanol) was added, and the mixture was heated to 60°C for 30 min. The reaction was stopped by the addition of 800 l of 2 N HCl, and absorbance was read at 420 nm. All experiments were performed in triplicate and compared with a standard of L-alanine.
ESI-MS of Muropeptides-Dried muropeptides were dissolved in 30 l of water:CH 3 CN:formic acid (50:50:0.1). A Perkin-Elmer Sciex API III triple quadrupole mass spectrometer was tuned and calibrated by flow injection (10 l/min) of a mixture of PPG 425, 1000, and 2000 (3.3 ϫ 10 Ϫ5 , 1 ϫ 10 Ϫ4 , and 2 ϫ 10 Ϫ4 M, respectively) in water:methanol (1:1) containing 2 mM ammonium formamide and 0.1% CH 3 CN. Calibration across the m/z range 10 -2400 was achieved by multiple ion monitoring of eight PPG solution signals (singly charged ions at m/z 58.99, 326.25, 906.67, 1254.92, 1545.13, 1863.34, and 2010.47 and the doubly charged ion at m/z 520.4). The ion spray voltage was operated at 4.5 kV using hydrocarbon-depleted air for the spray nebulization, and spectra were generated with a curtain gas produced from the vapors of liquid nitrogen. Samples were introduced into the ionization source by flow injection. ESI-MS spectra were obtained at instrument conditions sufficient to resolve the isotopes of the PPG/NH 4 ϩ singly charged ion at m/z 906 with 40% valley, an orifice voltage of 60, and step size during data collection of 0.3 Da. Daughter ion spectra were obtained using degraded mass resolution to improve sensitivity of detection, and a step size of 1 Da was used for data collection. Under these conditions, the isotopes of the PPG/NH 4 ϩ singly charged ion at m/z 906 were not resolved from one another.  (36). This observation suggested that the 11 murein hydrolase may cleave the peptidoglycan at two sites, the amide bond between N-acetylmuramyl and L-Ala and the peptide bond between D-Ala and Gly. If so, 11 murein hydrolase digestion of muramidase-released surface protein should yield L-Ala-D-iGln-L-Lys-(surface protein-Gly 5 )-D-Ala-COOH. Indeed, the mobility of the doubly digested surface protein on SDS-PAGE was found to be identical to surface protein solubilized directly from the staphylococcal peptidoglycan with the 11 enzyme (Fig. 2C), indicating that the attached peptidoglycan subunits had been removed.

11
To examine the structures of the muropeptides removed from the surface protein by the 11 enzyme, we employed a strategy previously devised to analyze the cell wall anchor structures of staphylococcal surface proteins (35,36). Seb-MH 6 -Cws is a fusion between staphylococcal enterotoxin B and the cell wall sorting signal of protein A ( Fig. 2A). At the fusion site between these domains a methionine followed by six histidines has been inserted. After solubilization of the staphylococcal cell wall with the muramidase mutanolysin, Seb-MH 6 -Cws was purified by affinity chromatography on nickel resin and cleaved by cyanogen bromide close to the anchoring point of the polypeptide with the cell wall. The C-terminal anchor peptides were purified by a second round of affinity chromatography and were then digested with 11 murein hydrolase. When separated by rpHPLC, the doubly digested cell wall anchor peptides generated three major and several minor peaks of absorption at 206 nm (Fig. 2B). The major peaks (designated A, B, and C) were desalted, concentrated, and analyzed by ESI-MS.
The sample that eluted at 7.2% methanol (peak A) generated an ion with m/z 702.4, a measurement that is in close agreement with the structure of a peptidoglycan cleavage fragment  These ESI-MS data further corroborated our hypothesis that the 11 hydrolase may cut muramidase-released cell wall anchor structures at two positions, the amide bond between Nacetylmuramyl and L-Ala and the peptide bond between D-Ala and Gly. Nevertheless, the ESI-MS data cannot distinguish between D-Ala-Gly or Gly-Gly endopeptidase activity, as the compound mass of 702.74 could be explained as peptides with the structure NH 2 -L-Ala-D-iGln-L-Lys-(NH 2 -Gly 5 )-D-Ala-COOH, NH 2 -L-Ala-D-iGln-L-Lys-D-Ala-Gly 5 -COOH, or any permutation thereof. To distinguish between these possibilities we examined the structure of the ion with m/z 702.4 by CID in an MS/MS experiment (Table I and Fig. 3). We observed a fragmentation product ion at m/z 286, corresponding to an intact pentaglycine cross-bridge. We also observed a daughter ion of mass 614, a product of removal of a single D-alanine from position 4 of the stem peptide, indicating that the D-Ala at position four of the wall peptide was not substituted and that the pentaglycyl was linked to the ⑀-amino of lysine. See Table  I for a listing of the observed daughter ions and an interpretation of their structure.
We would also expect that Edman degradation of the sample in peak A and peak B should yield five consecutive cycles of phenylthiohydantoin glycine if the 11 enzyme displayed D-Ala-Gly endopeptidase activity. This was tested, and the concentration of phenylthiohydantoin glycine obtained after Edman degradation remained similar during five consecutive cleavage cycles. Equimolar amounts of phenylthiohydantoin alanine and glycine were released in cycle 1 of the peak A sample, whereas no such release of alanine was observed for material that eluted in peak B (Table II). This observation is consistent with the presence of an N-terminal alanine in the peak A but not in the peak B sample.
Purification and Characterization of the 11(⌬181-301) Enzyme-To test whether the N-terminal domain of 11 hydrolase contained D-Ala-Gly endopeptidase or any other muralytic activity, we generated a recombinant enzyme, 11(⌬181-301), in which the 201 amino acid residues spanning the amidase domain were removed (Fig. 1A). Like the full-length 11 hydrolase, the 11(⌬181-301) enzyme was purified via a C-terminal six histidyl tag by affinity chromatography on nickel-Sepharose (see "Experimental Procedures"). As measured with both crude and purified cell walls, the 11(⌬181-301) enzyme degraded the staphylococcal peptidoglycan similar to fulllength 11 hydrolase (Fig. 4).
Alignment of the N-terminal domains from the phage enzymes (Fig. 1C) revealed the presence of several regions of homology between all four polypeptides, including 18 invariant residues. Of particular note is the presence of one conserved cysteine and three conserved aspartic acids, potentially catalytic residues that leave open the possibility that D-Ala-Gly endopeptidase cleavage occurs by a mechanism similar to cysteine or aspartate proteases. To address this possibility we FIG. 4. Digestion of crude staphylococcal cell walls with the 11(⌬181-301) enzyme. Crude staphylococcal cell walls were prepared by lysing staphylococci with glass beads and sedimenting the walls by centrifugation. The wall pellet was boiled in 2% SDS in water for 20 min and washed several times with water alone to remove excess detergent. The crude walls were resuspended to a final absorbance of 0.7 (600 nm) in the lysis buffer and distributed to tubes containing either EDTA (⅙, 5 mM), p-hydroxymercuribenzoic acid (Ⅲ, 1 mM), a mixture of pepstatin (2.5 M) and PMSF (2 mM, ▫), or water (no inhibitor, ‚/no enzyme, छ). Reactions were started by the addition of purified enzyme (except in the no enzyme tubes), and aliquots were taken at various time points, and lysis of walls was measured as a decrease in absorbance at 600 nm. The experiment was performed in triplicate using the identical cell wall/lysis buffer preparation, and results were averaged. Absorbances at all data points are displayed as an average of the three experiments, and the results did not vary between the experiments by more than 3%.  Fig. 2) Compounds eluted after HPLC chromatography as shown in Fig. 2 were desalted and analyzed by ESI-MS or subjected to protein sequence analysis. No amino acids other than glycine and alanine were detected in significant amounts during the sequencing reaction. Numbers in parentheses indicate the concentration of cleaved phenylthiohydantoin residues in picomoles.  (22) Ala (12) Ala (10) Ala (9) Ala (10) tested the staphylolytic activity of the 11(⌬181-301) enzyme in the presence of protease inhibitors, and we found that the activity could be completely abolished in the presence of 1 mM p-hydroxymercuribenzoic acid, an organic mercurial known to form stable complexes with thiol moieties (cysteine) (Fig. 4) (41). The purified enzyme was partially inhibited in the presence of 5 mM EDTA; however, a mixture of pepstatin and PMSF, inhibitors of aspartate and serine proteases, had no effect on the cell wall hydrolysis activity of the 11(⌬181-301) enzyme. These data suggest that the conserved cysteine residue is required for the D-Ala-Gly endopeptidase activity of the N-terminal portion of 11 murein hydrolase. Many muralytic enzymes possess glycan hydrolase activity as indicated by the release of reducing N-acetylhexosamines (22). We analyzed the 11 and 11(⌬181-301) enzymes for such an activity against crude staphylococcal cell walls using a modified Morgan-Elson procedure. As indicated in Table III, neither the 11 nor 11(⌬181-301) enzyme released significant hexosamine from staphylococcal cell walls. As a control, mutanolysin, an N-acetylmuramidase, released reducing hexosamines, whereas lysostaphin, a glycyl-glycine endopeptidase, did not. We measured the activity of murein hydrolases to release free amino groups from the staphylococcal cell wall with the Sanger reagent (DNFB), a chromogenic compound reactive with free amino groups. 11 murein hydrolase as well as the 11(⌬181-301) enzyme released significant amounts of free amino acid from the staphylococcal cell wall. As a control, mutanolysin liberated only small amounts, whereas lysostaphin released large amounts of free amino groups. These data indicate that 11 murein hydrolase as well as the 11(⌬181-301) enzyme display endopeptidase but no glycan hydrolase activity on the staphylococcal cell wall.
To reveal the identity of amino groups liberated by digestion with the 11 enzymes, highly purified staphylococcal peptidoglycan was digested, and the resulting muropeptides were subjected to Edman degradation. The 11 hydrolase-digested sample contained similar amounts of phenylthiohydantoin al-anine and glycine during the first cycle, whereas the next four cycles yielded predominantly glycine. In contrast, the 11(⌬181-301) enzyme-digested sample contained mostly phenylthiohydantoin glycine during five consecutive cleavage cycles (Table IV).
Purification of Peptidoglycan Cleavage Products of the 11 Enzymes-To analyze the digestion products of the 11 or 11(⌬181-301) enzymes by rpHPLC, digested staphylococcal peptidoglycan was further degraded with mutanolysin to cleave the glycan strands and liberate muropeptide monomers. After reduction of muropeptides with sodium borohydride, the sample was subjected to rpHPLC using the same procedure employed for the separation of muropeptides removed from the staphylococcal surface protein. The 11 hydrolase-digested sample yielded one major peak that eluted at 5.6% methanol (Fig. 5A). Analysis of this sample by ESI-MS revealed a compound with m/z 702.3, corresponding to the mass of a singly charged muropeptide of the structure NH 2 -L-Ala-D-iGln-L-Lys-(NH 2 -Gly 5 )-D-Ala-COOH (predicted mass 702.7). rpHPLC analysis of the 11(⌬181-301) enzyme-digested sample also generated one major peak that eluted at 8.0% methanol (Fig. 5B). ESI-MS of the eluted sample revealed a single ion of m/z 1182.6. This measurement is in agreement with the predicted m/z of a singly charged muropeptide of the structure MurNAc-(L-Ala-D-iGln-L-Lys-(NH 2 -Gly 5 )-D-Ala-COOH)-(␤1-4)-GlcNAc (calculated ion mass 1182.2). Taken together the data indicate that the 11 murein hydrolase has N-acetylmuramyl-L-Ala amidase as well as D-Ala-Gly endopeptidase activity, whereas the 11(⌬181-301) enzyme displays only D-Ala-Gly endopeptidase activity.
Solubilization of Surface Protein with the 11(⌬181-301) Enzyme-If the 11(⌬181-301) enzyme functions as a D-Ala-Gly endopeptidase, digestion of the staphylococcal cell wall with this enzyme should solubilize surface proteins linked to multiple muropeptide subunits that are attached along a single glycan strand. This was tested and 11(⌬181-301)-solubilized Seb-MH 6 -Cws was purified and subjected to SDS-PAGE. In TABLE III Liberation of amino groups and hexosamines from digested staphylococcal peptidoglycan Crude cell walls were digested with the enzymes as indicated. Soluble aliquots were removed and measured for the presence of free amino groups or hexosamines (see text). Data represent an average of three independent experiments. The standard deviation of these measurements is indicated as Ϯ.  (22) by measuring the absorbance at 585 nm and comparing with a set of GlcNAc standards of known concentration. As differences between the extinction coefficients of free (unsubstituted) hexosamines and those substituted with wall peptides have been observed (22), we have attempted to reduce the discrepancy by boiling the standards in borate buffer for 7 min instead of 30 min for the solubilized muropeptides (see "Experimental Procedures" for a detailed protocol).
b The concentration of amino groups was determined by modification with the Sanger reagent (DNFB) and measuring the absorbance of each sample compared to a set of L-alanine standards of known concentration (22).

TABLE IV
Edman degradation of muropeptides generated by the digestion of purified staphylococcal peptidoglycan Edman degradation of muropeptides generated from 11 and 11 (⌬181-301) digestion of purified staphylococcal cell walls (pmol released per cycle indicated in parentheses). Glycine and alanine were the only phenylthiohydantoin amino acids observed in significant amounts during sequencing. Ala (55) Ala (52) Ala (45) Ala (41) contrast to surface protein released with the full-length 11 enzyme, the 11(⌬181-301)-solubilized species migrated as a spectrum of fragments on SDS-PAGE, similar to mutanolysin and amidase-released surface protein. Redigestion of the 11(⌬181-301) solubilized surface protein with mutanolysin caused Seb-MH 6 -Cws to migrate uniformly on SDS-PAGE, suggesting that the heterogeneity in mass is caused by linked murein subunits that are tethered along glycan strands (Fig. 6).
To examine the structure of these surface proteins further, the C-terminal anchor peptides of 11(⌬181-301) solubilized Seb-MH 6 -Cws were obtained by cyanogen bromide cleavage and affinity purification. MALDI-MS revealed the presence of a heterogeneous population of ions spaced at regular intervals (Fig. 6). The ion signal at m/z 2757 likely corresponds to anchor peptide linked to a single wall tetrapeptide of the structure MurNAc-(L-Ala-D-iGln-L-Lys(NH 2 -H 6 AQALPET-Gly 5 )-D-Ala)-(␤1-4)-GlcNAc that is acetylated at the O-6 position of the muramic acid (calculated ion mass 2756.82). Ion groups centered at m/z 3919, 5152, and 6314 can be explained as anchor peptides linked to two, three, four, or five peptidoglycan subunits. A summary of the ions observed in each of these groups is given in Table V. The ion signals became more broadly dispersed with increasing mass and frequently diminished in intensity. These phenomena are likely due to the fact that the linked peptidoglycan subunits exist as either tetrapeptide (1162.16 Da) or pentapeptide subunits (1233.24 Da) that may or may not be acetylated at the O-6 position of N-acetylmuramic acid (42.03 Da). Thus, each additional linked peptidogly-can subunit increases the number of possible compounds by a factor of 4. Surface protein linked to three peptidoglycan subunits could theoretically exist as any one of 16 possible mass combinations, whereas a protein linked to four peptidoglycan subunits yields 25 different compounds. Hence a majority of the mass signals at m/z greater than 7000 overlapped and were difficult to interpret. Nonetheless, we were able to observe ion groups to an m/z of approximately 18,000, corresponding to anchor peptides linked to as many as 15 subunits (data not shown).
The ion with m/z 2235 corresponds to anchor peptide linked to a wall subunit of the structure NH 2 -L-Ala-D-iGln-L-Lys-(NH 2 -H 6 AQALPET-Gly 5 )-D-Ala-COOH, i.e. a single peptidoglycan subunit lacking the MurNAc-GlcNAc disaccharide (calculated MH ϩ 2235). Ion groups centered at m/z 3238, 4439, 5631, and 6384 can be explained as anchor peptides attached to two, three, four, or five wall subunits where one of the subunits lacks the peptide substituent (a loss of 683.72 Da). Apparently, each of these cell wall structures was generated by N-acetylmuramyl-L-Ala amidase hydrolysis of one of the murein subunits. We speculate that these compounds are generated by the endogenous autolysin amidase activity in crudely prepared staphylococcal cell walls from which the Seb-MH 6 -Cws protein was purified (42). Ions representing anchor peptides linked to multiple cell wall pentapeptides were not readily observed (usually the anchor peptides were linked to either a single or no pentapeptides). The reasons for this are most likely due to the fact that the staphylococcal peptidoglycan is highly crosslinked by transpeptidation with the majority of subunits existing as tetrapeptides. DISCUSSION In this paper we analyzed the enzymatic properties of 11 murein hydrolase using both staphylococcal peptidoglycan as well as solubilized surface protein as substrates. We conclude that the enzyme is bifunctional, having both an N-acetylmuramyl-L-alanyl amidase and a D-alanyl-glycine endopeptidase activity. By engineering a deletion that removed the central amidase domain of this enzyme, we determined that the endopeptidase activity is contained within the first 180 residues of the polypeptide chain. To our knowledge this is the first report of such an activity associated with the aureophage endolysins. The presence of a muralytic activity in the N-terminal portion of the enzyme is in agreement with a recent report for the related muralytic enzyme of staphylococcal phage Twort (43). Like the 11 enzyme, the Twort murein hydrolase was characterized as an amidase based on the ability of the fulllength enzyme to liberate DNFB reactive L-alanine from digested peptidoglycan (43). Because the N-terminal domain of the Twort enzyme is similar to that of the 11 hydrolase, we think it is likely that the Twort enzyme as well as other aureophages also display D-alanyl-glycine endopeptidase activity.
Recently it has been reported that the ␣and ␤-lytic proteases secreted by Achromobacter lyticus are capable of cleaving staphylococcal peptidoglycan at the pentaglycine crossbridge (44 -46). Both lytic proteases appear to hydrolyze not only the Gly-Gly bond but also D-Ala-Gly and Ala-Ala bonds, suggesting that these enzymes display a more relaxed substrate specificity than the 11 enzyme. The ␣-lytic protease also possesses amidase activity indicating that it may be functionally similar to the 11 hydrolase, although the gene for this enzyme has not yet been cloned (45). The ␤-lytic protease is a metalloenzyme that bears no primary sequence homology to the 11 enzyme (44,46). Crude enzymatic preparations capable of cleaving the staphylococcal wall peptide-cross-bridge junction were also isolated over 30 years ago from Streptomyces albus G (47) and Myxobacter AL-1 (10) culture filtrates. The FIG. 5. Comparison of muropeptides generated by digestion of purified staphylococcal peptidoglycan with the full-length 11 and 11(⌬181-301) enzymes by rpHPLC. A, staphylococcal peptidoglycan was solubilized with the 11 murein hydrolase, and glycan strands were degraded by subsequent digestion of the solubilized muropeptides with mutanolysin (see "Experimental Procedures"). Digestion products were reduced with sodium borohydride and analyzed by rpHPLC on a C18 column using a protocol developed for the separation of staphylococcal muropeptides (39). Products eluting with the indicated absorption peak were desalted and analyzed by ESI-MS. The observed m/z of the ion signal and the proposed structure is shown. B, analysis of products generated by the solubilization of staphylococcal peptidoglycan with purified 11(⌬181-301) enzyme.
Myxobacter enzyme was reported to have an additional, less efficient amidase activity, which may indicate that it is also functionally similar to the 11 hydrolase (10). Staphylococcal peptidoglycan is highly cross-linked (8,11,39,48), and disruption of the cross-bridges appears to be a highly efficient means by which to rapidly hydrolyze the cell wall. It is likely not a coincidence that enzymes whose function is to destroy the murein sacculus, for example lysostaphin and phage hydrolases, have chosen the staphylococcal cross-bridge as their target.
All aureophages described to date belong to the order of tailed phages (caudovirales) and are further subclassified into three families as follows: the Myoviridae, the Siphoviridae, or the Podoviridae. Whereas 187, 11, and 80␣ belong to the Siphoviridae family, Twort belongs to the Myoviridae (49). That each of these phages encodes proteins homologous to the D-Ala-Gly endopeptidase of 11 suggests that this activity will be found in several distantly related aureophages. Genes encoding phage hydrolases have been cloned and sequenced from many different tailed phages that infect a variety of Grampositive bacteria including Listeria sp. (50), Streptococcus pneumoniae (51)(52)(53), and Bacillus cereus (54). A comparison of the muralytic enzymes from Gram-positive bacteria and their phages has indicated that these enzymes are modular, containing separate catalytic and cell wall binding domains (targeting domains) (51,54,55). It has been proposed that the evolution of these enzymes has likely occurred through the swapping of these domains and that phage lytic enzymes have co-evolved with host autolysins (52,56,57). A majority of the muralytic enzymes cloned from the Gram-positive phages are amidases and bear significant homology to the central amidase portion of the 11 enzyme. Exceptions include the lytic enzymes of phages A118 (Ply118) and A500 (Ply500) of Listeria monocytogenes, which possess an L-alanyl-D-glutamate peptidase activity (50), and the Cpl glycosidase of the pneumococcal phage Cp-1 (51). Due to the near ubiquitous presence of MurNAc-Lalanine linkages in bacterial peptidoglycan, amidases frequently display activity against cell walls from a wide variety FIG. 6. Characterization of 11 (⌬181-301)-solubilized anchor peptides by MALDI-MS. Seb-MH 6 -Cws was solubilized from the staphylococcal cell wall with the 11(⌬181-301) enzyme, and C-terminal anchor peptides were prepared by treatment of the full-length protein with cyanogen bromide, affinity purification, and rpHPLC (see "Experimental Procedures"). HPLC fractions containing anchor peptides were analyzed by MALDI-MS. Ions indicated by * are apparently the result of partial amidase removal of peptide substituents from subunits along the linear glycan strand to which the anchor peptide is attached. Ions indicated in parentheses are approximate values due to a high degree of overlap with nearby ion signals (see text). Inset, Coomassie-stained 12% SDS-PAGE of Seb-MH 6 -Cws solubilized from staphylococcal peptidoglycan with mutanolysin or the 11(⌬181-301) enzyme. 11(⌬181-301)-solubilized Seb-MH 6 -Cws was redigested with mutanolysin, and surface protein products migrated uniformly on SDS-PAGE, indicating their linkage to murein subunits that are tethered along disaccharide (MurNac-GlcNac) moieties. Approximately 5-fold less protein was loaded in the double-digested sample (mutano. ϩ 11(⌬181-301)) to compensate for the overloading that occurred when the many species of various molecular weights were converted into a single species. The arrow designated (⌬) indicates the migration of the 11(⌬181-301) enzyme on SDS-PAGE. of bacterial species. In contrast, we suspect that the D-alanylglycine endopeptidase activity of the 11 enzyme may be specific toward only a small subset of the known types of peptidoglycan. Although D-alanyl-glycyl bonds have also been found in the peptidoglycan of strains within the genus Micrococcus (58), Thermus (59), and Deinococcus (60), sequences homologous to the N-terminal domain of the 11 hydrolase have thus far only been identified in enzymes isolated from staphylococci or their phages. The C-terminal domain of the 11 hydrolase bears significant homology to the wall targeting domain of lysostaphin and InlB, a protein found on the surface of L. monocytogenes (61). Targeting domains have been shown to help control both the specificity and the level of activity of muralytic enzymes (28). Thus far, at least two other types of targeting domains have been characterized in autolysins of Gram-positive bacteria. The first type to be identified was found in a set of bacterial and phage-encoded pneumococcal hydrolases that recognize specific choline moieties in the teichoic acid polymer associated with the envelope of S. pneumoniae (62,63). The other examples are the C-terminal repeat domains recently discovered on the multifunctional staphylococcal autolysin, Atl. In contrast to the other two binding domains, which likely bind a receptor found all over the surface of the target cell, the Atl targeting domains have been shown to direct reporter proteins specifically to the sites of cell division (30).