Isolation of the Penicillin-Binding Peptide from D-Alanine

Proc. Natl. Acad. Sci. USA Vol. 74, No. 3, pp. 1009-1012, March 1977 Biochemistry

Isolation of the penicillin-binding peptide from D-alanine carboxypeptidase of Bacillus subtilis (penicilloyl-carboxypeptidase/peptide fragments/ester bond) NAFSIKA GEORGOPAPADAKOU, SVEN HAMMARSTROM, AND J. L. STROMINGER The Biological Labhratories, Harvard University, Cambridge, Massachusetts 02138 Contributed by Jack L. Strominger, January 6, 1977

ABSTRACT The D-alanine carboxypeptidase of B. subtilis we report two such fragments resulting from a tryptic and is a membrane-bound enzyme which is inhibited by penicillins Pronase digestion, respectively. and binds them covalently. The enzyme has been labeled with [14C- or [a5SSpenicillin. After tryptic or Pronase digestion of the MATERIALS AND METHODS labeled, denatured, reduced, and carboxymethylated enzyme, a radioactive peptide was isolated in each case. The amino acid Materials. Deoxyribonuclease, lysozyme, and dithiothreitol compositions of these two peptides are reported. The Pronase were purchased from Sigma; L-1-tosylamido-2-phenylethyl peptide was a subset of the tryptic peptide. Neither contained chloromethyl ketone (TPCK)-trypsin from Worthington; a cysteine residue and the only amino acid in the Pronase pep- Pronase from Calbiochem; Sephadex G-50 from tide to which the penicillin could be bound was a serine resi- (fine) Phar- due. macia; guanidine hydrochloride from Heico; dansyl chloride, Chen-Ching polyamide sheets, standard amino acid mixture, The D-alanine carboxypeptidase (CPase) of Bacillus subtilis 6 M HCl, ninhydrin, fluorescamine, and triethylamine from is a membrane-bound enzyme which catalyzes the hydrolysis Pierce; [8-14C]penicillin G, potassium salt (specific activity: 54 of the terminal D-alanine residue of UDP-MurNAc-penta- mCi/mmol) from Amersham/Searle. [35S]Penicillin G, potas- peptide, the uridine nucleotide intermediate in cell wall pep- sium salt (specific activity: 1.6 mCi/mmol) was kindly syn- tidoglycan synthesis. The reaction is similar to the transpepti- thesized by Richard Thoma of E. R. Squibb and Sons. Pyridine dase reaction involved in cell wall biosynthesis, except that was purified by distillation after 2 hr of refluxing with ninhy- water rather than an amino group is the acceptor. Both reac- drin (5 g of ninhydrin per liter of pyridine). Iodoacetic acid was tions, it was suggested, proceed through an acyl-enzyme in- purchased from Eastman and purified by recrystallization from termediate (1), and both are inhibited by penicillins. CPase is petroleum ether. All other reagents were of analytical grade. a moderate-sized protein (50,000 molecular weight) (2), is the Preparation of Labeled CPase. CPase was isolated by af- major penicillin binding component in B. subtilis (3), and has finity chromatography (4) from B. subtilis (strain Porton) been purified to homogeneity by affinity chromatography (4). membranes obtained by lysozyme DNase treatment (11). In Its function is not known, but it might be involved in the reg- a typical experiment, 10 mg of enzyme was obtained from 300 ulation of crosslinking during cell wall biosynthesis. g of cells. The enzyme in 5 ml of 0.05 M Tris-HCI buffer (pH It was suggested earlier that the penicilloyl-enzyme complex 7.5) containing 1% (vol/vol) Triton X-100, 0.5 M NaCl and 1 might be a penicilloyl thioester involving a cysteine residue in mM 2-mercaptoethanol] was incubated with labeled penicillin the susceptible enzyme. This suggestion was based on the in- G ([8-14C]penicillin G, 25 ,uCi, or [35S]penicillin G, 0.8,Ci) for hibition of penicillin binding by thiol reagents and the release 10 min at 25°. The incubation was stopped by adding 4 volumes of a penicilloyl derivative by hydroxylamine and ethanethiol of cold acetone containing a 100-fold molar excess of unlabeled (5). However, the concentrations of thiol inhibitors which were penicillin G. The resulting suspension was centrifuged at 11,000 effective may have been too high to be specific (6). Further- X g for 20 min, and the pellet was dried under a stream of ni- more, an alternative interpretation of the release of penicilloyl trogen. by hydroxylamine has been obtained, namely, the release is an Denaturation, Reduction, Carboxymethylation, and enzymatically catalyzed transfer of penicilloyl to hydroxyl- Trypsin or Pronase Digestion. The precipitated carboxy- amine, and denatured penicilloyl enzyme is stable under con- peptidase, labeled with penicillin G (160 nmol, 1.1 X 107 dpm) ditions in which acyl thioesters do not migrate and are cleaved was dissolved in 2 ml of 6 M guanidine-HCI containing 2 mM by hydroxylamine (refs. 6-10; S. J. Curtis and J. L. Strominger, EDTA, and 0.2 M Tris-HCI (pH 8.2). After 6 mg of di- manuscript submitted for publication). The chemical reactivity thiothreitol had been added, the solution was incubated under of penicilloyl-enzyme, denatured under several conditions, was nitrogen at 370 for 1 hr. Then 20 mg of iodoacetate was added compatible with the presence of a serine ester (8). However, at and the solution was incubated under nitrogen and in the dark least in the case of Escherichia coli, D-alanine carboxypeptidase at room temperature for 1 hr. The solution was dialyzed against IA, a sulfhydryl group, presumably in the active site, is involved distilled water at 40 during which the carboxypeptidase pre- in the deacylation of the penicilloyl- and acyl-enzyme. Its cipitated. The resulting suspension was lyophilized. After lyo- substitution by thiol reagents results in stabilization of the philization, the carboxymethylated carboxypeptidase was penicilloyl-enzyme and in accumulation of an acyl-enzyme dissolved in 3 ml of 0.1 M NH4HCO3 and TPCK-trypsin was intermediate derived from a normal substrate (ref. 8; S. J. Curtis added to a final concentration of 10% (wt/wt) relative to the and J. L. Strominger, manuscript submitted for publication). carboxypeptidase. The hydrolysis was carried out at 370 for 1 These findings prompted a close examination of the peni- hr, and the reaction was stopped by freezing with dry ice-ace- cilloyl-enzyme linkage, by isolating and characterizing peptide tone followed by lyophilization. Pronase digestion was per- fragments containing bound penicillin. In this communication, formed under the same conditions as trypsin digestion. If longer incubation times were used, much of the bound penicillin was Abbreviation: CPase, D-alanine carboxypeptidase. released. 1009 Downloaded by guest on October 2, 2021 1010 Biochemistry: Georgopapadakou et al. Proc. Natl. Acad. Sci. USA (74) 1977

Carboxypeptidase [ 14C] penicillin G

[14C] penicilloyl-CPase

I 1. dithiothreitol 2. iodoacetate

carboxymethylated [ 14C] penicilloyl-CPase (50%) *

Pronase Trypsin

mixture of peptides mixture of peptides

1. G-50 chromatography 1. G-50 chromatography (30%) 2. pH 6.5 electrophoresis 2. pH 6.5 electrophoresis (15%) 3. descending chromatography 3. pH 3.5 electrophoresis (5%) 4. pH 3.5 electrophoresis I 4. descending chromatography (2%) Pronase [14C] penicilloyl-peptide tryptic [14C]penicilloyl-peptide FIG. 1. Flow diagram for the isolation of tryptic and Pronase peptides from 14C-labeled CPase. Approximate yields for [14C]penicilloyl-CPase, based on radioactivity, are given in parentheses. Isolation of Labeled Peptides. Fractionation was first water/pyridine (15/3/12/10) (13). Analytical chromatography performed on a Sephadex G-50 (fine) column (1.2 X 100 cm) was carried out in isoamyl alcohol/pyridine/water (6/6/7) (14). with 0.1 M NH4HCO3 as eluent at a flow rate of 10 ml/hr at The purifications are summarized in Fig. 1. 4°. Fractions of 2 ml were collected and monitored for both 14C Transfers of peptide between fractionations were performed content (by counting 50 ,l aliquots in Aquasol) and peptide by the "sewing" method (15). Peptides were located by spraying content (by measuring A23o nm). The fractions comprising the guide strips with triethylamine/CH3Cl (1/20, vol/vol) followed major radioactive peak were pooled and lyophilized. Further by fluorescamine/acetone (0.1/100, wt/vol) and then visual- purification was achieved by high voltage electrophoresis at pH izing under long-wavelength ultraviolet light. Radioactivity 6.5 and 3.5 and descending chromatography on Whatman was located by autoradiography. The concentration of the ra- 3MM paper (46 X 57 cm) prewashed with 1% acetic acid. The dioactive peptides was calculated from the radioactivity electrophoresis buffers (all vol/vol) were pyridine/acetic measurements and the specific activity of the label. acid/water (50/2/950; pH 6.5) and pyridine/acetic acid/water Performic Acid Oxidation and Amino Acid Analysis. The (1/10/89; pH 3.5) (12), and the electrophoreses were carried purified radioactive peptides (usually 1 nmol), eluted from the out on a flat plate and in a tank, respectively. Chromatography paper with distilled water, were hydrolyzed in 6 M HC1 in of the peptides was carried out in 1-butanol/acetic acid/ evacuated tubes at 1100 for 20 hr. When analyses for cysteic acid were performed, separate samples were subjected to performic acid and oxidation by the procedure of Moore (16) prior to hydrolysis with acid. Amino acid analyses were performed on a Durrum d-500 or a Beckman 121 M amino acid ana- lyze. NH2-terminal Amino Acid Determination. Approximately 0.2-0.5 nmol of peptide was dansylated by the procedure of Gray and Hartley (17). The dansyl amino acids were identified 20 40 60 Fraction no. ' by comparison with a standard dansylated amino acid mixture SEPHADEX G-50 CHROMATOGRAPHY , / pH 6.5 ELECTROPHORESIS after two-dimensional polyamide thin-layer chromatography. RESULTS cm D. Properties of the penicilloyl-enzyme linkage o- -Origin _ _ _ _ _ Labeled penicillin was bound to the B. subtilis D-alanine car- 10-o4. . boxypeptidase under conditions that led to complete inhibition -10- [35S] of the enzyme (18). After carboxymethylation, there was 0.5 -11~~~ -['4c+35SI mol of The 30- [k4C] 0 - Origin mol of labeled penicillin per enzyme. penicilloyl- Solvent enzyme complex was stable near neutral pH (electrophoresis Front 10 at pH 6.5 and descending chromatography) but only partially DESCENDING CHROMATOGRAPHY pH 3.5 ELECTROPHORESIS so in basic (Sephadex G-50 chromatography, pH 8-9) or acidic FIG. 2. Purification of the tryptic penicilloyl-peptide by Se- (electrophoresis at pH 3.5) conditions. This might account for phadex G-50 chromatography (A), high voltage electrophoresis at pH the low overall yield of radioactive peptide (typically, 2%; Fig. 6.5 (B) and pH 3.5 (C), and descending chromatography (D). Ra- 1). At pH 8.5 and 370, the half-life of the penicilloyl-enzyme dioactive bands (-) were localized by autoradiography while peptide linkage is 4 hr (S. Hammarstrom, unpublished). Thus, after bands (...) were localized by staining a guide strip with fluorescamine. preliminary experiments, the trypsin or Pronase digestions were In each step, the labeled peptide is indicated by a heavy arrow. The out for 1 hr with a trypsin or Pronase to carboxy- form of radioactivity is given in brackets. In the case of the 35S-labeled carried high peptide, the order of (D) and (C) has been reversed (compare at bro- peptidase ratio. Finally, during the isolation of the Pronase ken arrows). peptide, the electrophoresis step at pH 3.5 was used last, this Downloaded by guest on October 2, 2021 Biochemistry: Georgopapadakou et al. Proc. Natl. Acad. Sci. USA (74) 1977 1011

Table 1. Amino acid composition of labeled tryptic and i.2[-A. Pronase peptides 48 ' ~Q8 32 B Residues per mole

04 I I 16 t. Amino acid Tryptic Pronase I I I -p:nIQ,I; -' Asn* 1.2 (1) 1.0 (1) 20 40 60 Thr 0.9 (1) Fraction No. Ser 1.4 (1) 1.3 (1) SEPHADEX G-50 CHROMATOGRAPHY Glu 1.2 (1) Pro 0.8 (1) Gly 2.5 (2) 2.5 (2) e Ala 1.1 (1) 1.2 (1) cm Ile 0.7 (1) 1.4 (1) -30- Leu 1.0 (1) -0 Lys 1.0 (1) -10- 11 6 Total residues 0- -Origin * Based on the electrophoretic mobility of the Pronase peptide (see 0- Discussion). L.-J t The assumed integral values are given in parentheses. pH 3.5 ELECTROPHORESIS DESCENDING CHROMATOGRAPHY FIG. 3. Purification of the Pronase penicilloyl-peptide by Se- phadex G-50 chromatography (A), electrophoresis at pH 6.5 (B), being the harshest condition of the isolation procedure (Fig. 3). descending chromatography (C), and pH 3.5 electrophoresis (D). In At pH 3.5, the penicilloyl moiety itself might decompose; in a each step, the labeled peptide is indicated by a heavy arrow; (-) companion experiment using [35S]penicillin, the radioactivity radioactive bands; (...) peptide bands. did not migrate with the peptide at this step (Fig. 2). These properties of the penicilloyl-enzyme linkage are consistent with (solvent front at 35 cm). The final purification step was elec- either an ester or thioester type of bond. However, as will be trophoresis at pH 3.5 (2 hr; 40 V/cm) (Fig. 3D) at which the discussed later, amino acid analysis of the tryptic peptide after mobility of the radioactive peptide was -3 cm/hr. The purified performic acid oxidation revealed no cysteic acid (Fig. 2), thus peptide had glycine as its NH2-terminal amino acid (isoleucine eliminating the possibility of a thioester bond. and aspartic acid/asparagine being also present but in lesser amounts). The amino acid composition of the Pronase peptide Purification and characterization of labeled tryptic is also shown in Table 1. peptide The radioactive peptide present in the tryptic digest of labeled DISCUSSION and carboxymethylated CPase was first subjected to gel fil- Previous studies in this laboratory have indicated that penicillin tration as described in Materials and Methods. The major ra- acts on the susceptible enzymes by binding covalently to them. dioactive fractions eluted from the Sephadex G-50 column (Fig. In most of these studies, D-alanine carboxypeptidase from B. 2A) contained a single labeled peptide with an anode mobility subtilis has been chosen as a model system because of its ac- of -2.5 cm/hr on electrophoresis at pH 6.5 (30 min; 60 V/cm). cessibility. Nevertheless, even in this case, the mode of attach- Three more radioactive bands with no peptide content were ment of penicillin and its spatial relationship to the binding of observed at -4, +5, and +11 cm (Fig. 2B) when the 14C label the substrate is poorly understood. It is likely that the ,B-lactam was in the side chain of penicillin, but were absent when the `5S bond of penicillin is opened by a nucleophilic attack on the label was in the nucleus. This suggests that these products were carbonyl group because this is the most reactive part of the degradation products of the penicilloyl moiety (compare refs. penicillin molecule. In a previous report, it was suggested that 19 and 20). The radioactive peptide was further purified by the nucleophile is an SH group because penicillin bound to electrophoresis at pH 3.5 (Fig. 2C) (2 hr; 40 V/cm, mobility of D-alanine carboxypeptidase could be released with hydroxyl- -11 cm/hr) and descending chromatography (RF = 0.56 with amine or ethanethiol (5). More recently, it was shown that solvent front at 35 cm) (Fig. 2D). The distribution of radioac- penicillin release by hydroxylamine or ethanethiol was pre- tivity and fluorescamine-positive material is shown in Fig. 2. vented by denaturation of penicilloyl-carboxypeptidase under The peptide obtained was homogeneous as judged by analytical a variety of conditions (7-9). These reactions are enzymatically chromatography and had glycine as its sole NH2-terminal catalyzed. Thus, the present evidence does not support a amino acid. The amino acid composition of the purified tryptic thioester bond between penicillins and D-alanine carboxy- peptide is shown in Table 1. peptidase (reviewed in refs. 6 and 9). The possibilty of an ester bond, involvng a serine in the en- Purification and characterization of labeled Pronase zyme, has also been examined. Surprisingly, the enzyme was peptide insensitive to diisopropylfluorophosphate [up to 10 mM (N. H. The radioactive peptide obtained from Pronase digestion was Georgopapadakou, unpublished)], an organophosphorus re- smaller than the tryptic peptide as judged by its elution profile agent that generally inhibits serine hydrolases [alkaline phos- in Sephadex G-50 chromatography (Fig. 3A). The penicilloic phatase being a notable exception (21, 22)]. It is also insensitive acid peak, a discrete peak in the case of the tryptic peptide, was to phenylmethylsulfonyl fluoride, another, although less potent, buried under the major radioactive peak in the case of the inhibitor of several serine hydrolases (23). Pronase peptide. On electrophoresis at pH 6.5 (30 cm; 60 V/cm) The present study indicates the binding of penicillin to D- (Fig. 3B), the labeled peptide had a mobility of -3 cm/hr and alanine carboxypeptidase from B. subtilis involves a serine on descending chromatography (Fig. 3C) an RF value of 0.82 residue (Table 1). Cysteine was excluded on the basis of the Downloaded by guest on October 2, 2021 1012 Biochemistry: Georgopapadakou et al. Proc. Natl. Acad. Sci. USA (74) 1977

absence of cysteic acid after performnic acid oxidation. The 3. Blumberg, P. M. & Strominger, J. L. (1972) J. Biol. Chem. 247, oxidation was necessary because unmodified cysteines do not 8107-8113. survive acid hydrolysis of peptide bonds whereas the penicilloyl 4. Blumberg, P. M. & Strominger, J. L. (1972) Proc. Natl. Acad. Sci. group could be hydrolyzed off rapidly under these conditions. USA 69, 3751-3755. Serine is the only group in the Pronase peptide to which the 5. Lawrence, P. J. & Strominger, J. L. (1970) J. Biol. Chem. 245, penicilloyl moiety could be attached. The stability of an acid 3660-3666. 6. Blumberg, P. M. & Strominger, J. L. (1974) Bacteriol. Rev. 38, anhydride of glutamic acid (present in the tryptic peptide but 291-335. not in the Pronase peptide) or aspartic acid would preclude its 7. Blumberg, P. M., Yocum, R., Willoughby, E. & Strominger, J. isolation (24), and the electrophoretic mobilities of the tryptic L. (1974) J. Biol. Chem. 249,6828-6835. and Pronase peptides can be accounted for only if the aspartic 8. Kozarich, J. W., Buchanan, C. E., Curtis, S. J., Hammarstrom, acid residue is an asparagine one. S. & Strominger, J. L. (1976) in Ninth Rochester International It should be emphasized that the fact that penicillin binds to Conference on Environmental Toxicity, in press. a serine residue does not necessarily imply a serine protease (i.e., 9. Kozarich, J. W. (1977) in Microbiology 1977, ed. Schlessinger, a protease with a "superactive" serine). The high reactivity of D. (American Society for Microbiology, Washington, D.C.), in the ring, which workers have press. f3-lactam early compared to that 10. Strominger, J. L. (1977) in Microbiology 1977, ed. Schlessinger, of an acid chloride (25), might compensate for the lack of re- D. (American Society for Microbiology, Washington, D.C.), in activity of serine. In this case, the question arises whether the press. same serine residue is involved in the acyl-enzyme complex 11. Bishop, D. G., Rutberg, L. & Samuelsson, B. (1967) Eur. J. Bio- formed from substrate (ref. 9; J. W. Kozarich and J. L. Strom- chem. 2, 448-453. inger, manuscript submitted for publication), i.e., whether, as 12. Ryle, A. P., Sanger, F., Smith, L. F. & Kitai, R. (1955) Biochem. has been suggested in the past (1), penicillin acts as a substrate J. 60, 541-556. analogue and binds at the same site. This can only be resolved 13. Light, A. L. & Smith, E. L. (1962) J. Biol. Chem. 237, 2537- by isolating tryptic acyl-peptide and comparing it to the pen- 2546. icilloyl-peptide. 14. Baglioni, C. (1961) Biochim. Biophys. Acta 48,392-396. 15. Brown, J. R. & Hartley, B. S. (1966) Biochem. J. 101, 214- During the preparation of this manuscript, the isolation of 228. a penicilloyl-peptide from a soluble D-alanine carboxypepti- 16. Moore, S. (1963) J. Biol. Chem. 238, 235-237. dase from Streptomyces was reported (26). In that case too, the 17. Gray, W. R. & Hartley, B. S. (1963) Biochem. J. 89,379-380. penicillin was bound to a serine residue. The isolation of peni- 18. Blumberg, P. M. & Strominger, J. L. (1971) Proc. Natl. Acad. Sci. cilloyl-proteins (6) and of these penicilloyl-peptides from both USA 68, 2814-2817. B. subtilis and Streptomyces strongly supports the mechanism 19. Hammarstrom, S. & Strominger, J. L. (1975) Proc. Natl. Acad. of inactivation of penicillin-sensitive proteins originally pro- Sci. USA 72,3463-346?. posed by Tipper and Strominger (1). 20. Hammarstrom, S. & Strominger, J. L. (1976) J. Biol. Chem. 251, 7947-7949. We wish to thank K. Okumura and M. Pamukcu for performing the 21. Dabich, D. & Neuhaus, 0. W. (1966) J. Biol. Chem. 241, 415- amino acid analyses; J. Venuti for growing B. subtilis; L. Waxman and 420. L. K. Drickamer for helpful advice; and A. Jornvall for helping us with 22. Morton, R. K. (1955) Biochem. J. 61, 232-240. preliminary purification and analysis of the tryptic peptide. This work 23. Fahrney, D. E. & Gold, A. M. (1963) J. Am. Chem. Soc. 85, was supported by research grants from the National Science Founda- 997-1000. tion (PCM 71-02110) and the National Institutes of Health (AI-10952) 24. Hartsuck, J. A. & Lipscomb, W. M. (1971) in The Enzymes, ed. and by an exchange visitor grant to S.H. (1974-1975) from the Swedish Boyer, P. D. (Academic Press, New York), Vol. III, 1-56. Medical Research Council. 25. Johnson, J. R., Woodward, R. B. & Robinson, R. (1949) in The Chemistry of Penicillin, eds. Clarke, H. T., Johnson, J. R. & 1. Tipper, D. J. & Strominger, J. L. (1965) Proc. Natl. Acad. Sci. Robinson, R. (Princeton University Press, Princeton, N.J.), pp. USA 54, 1133-1141. 440-454. 2. Umbreit, J. N. & Strominger, J. L. (1973) J. Biol. Chem. 248, 26. Frire, J.-M., Duez, C., Ghuysen, J.-M. & Vanderkhoce, J. (1976) 6759-6766. FEBS Lett. 70, 257-260. Downloaded by guest on October 2, 2021