Amidation and Cross-Linking of the Enzy-matically Synthesized Peptidoglycan of Bacillus stearothermophilus*

A particulate enzyme system from Bacillus sfearofher-mophilus catalyzes the synthesis of peptidoglycan from uridine nucleotide precursors cross-linking of the the amidated obtained of the peptidoglycan comparison with similar materials obtained of The the the into

enzyme system from Bacillus sfearofhermophilus which catalyzes the synthesis of peptidoglycan from uridine nucleotide precursors has been investigated. The preparation catalyzed extensive cross-linking of the enzymatically synthesized peptidoglycan.
In addition in the presence of ATP and ammonia or L-glutamine amidation of the peptidoglycan also occurred. The nature of the amidated products obtained on lysozyme digestion of the enzymatically synthesized peptidoglycan was elucidated by a comparison with similar materials obtained after lysozyme digestion of isolated vegetative cell walls of this organism.
The peptidoglycans of the vegetative cell walls of the Bacillaceae fall into two general types. In the more common type, chemotype 1 (I), meso-Dapl residues are cross-linked through their D-NH2 groups to the COOH-terminal n-alanine residues of adjacent tetrapeptide units, as has been shown for Bacillus cereus (2), Bacillus lichenijormis (3,4), Bacillus megaterium (5), Bacillus stearothermophilus (6, 7), Bacillus subtilis (8,9), and Bacillus thuringiensis (10). In the less common type the c-NH2 groups of &sine are cross-linked through either n-isoasparagine in B. sphaericus (11) and in one strain of B. pasteu+ or by P-Daspartyl-L-alanine (or L-serine) in the other strains of B. pasteuri to the COOH-terminal n-alanine residues of adjacent tetrapeptide substituents.
Efficient systems for peptidoglycan synthesis in vitro have been demonstrated in B. megaterium (12,13) and in B. stearothermophilus NCTC 10339 (12). Less efficient systems have been described in B. lichenijormis and B. subtilis (14,15). Such efficient biosynthetic systems are a prerequisite for investigations of the synthesis of the specialized kind of peptidoglycan found in the spore cortex (16,17) which is synthesized at Stage IV of sporulation (18)(19)(20).
In order to be able to recognize the sporu-* This work was supported by Research Grants from the National Science Foundation (GB-30690X) and the United States Public Health Service (AI-09576).
lation-specific peptidoglycan products, it is necessary to characterize completely the products produced by the particulate enzyme from vegetative cells.
In the present paper the products obtained on lysozyme digestion of enzymatically synthesized peptidoglycan of B. stearothermophilus will be compared to those found on similar treatment of isolated vegetative cell walls of this organism.

MATERIALS AND METHODS
Organism and Growth Conditions-B. stearothermophilus ATCC 15952 was grown to produce spores on a medium3 containing 2 g of glucose, 1 g of vitamin-free Casamino Acids (Difco), 1.0 g of KH2P04, 2.09 g of K2HP04, 1 g of NaCl, 1 g of KC1 per liter of distilled water to which was added a spore salt solution, sterilized separately, which contained per liter of medium 200 mg of KN08, 40 mg of MgC12, .6Hz0, 15 mg of CaC12.2Ht0, 2 mg of MnS04 HzO, 2 mg of ZnSOh, 1 mg of NH4 molybdate, and 0.2 mg of CaC12. A 15-liter culture containing silicone antifoam was grown at 60" with maximum aeration and maximum stirring in a Microferm fermenter (New Brunswick Corp.) for about 1.5 days, and then harvested by centrifugation.
The resulting pellet containing mainly spores with some vegetative and sporulating forms was washed twice with 200 ml of water at O", centrifuging at 2500 X g for 5 min. This spore suspension was mixed with 100 ml of water and 1 ml of CHC& and allowed to autolyze at 4" for 6 days. The resulting pellet was washed five times with water by centrifugation at 2500 x g to remove membranous material.
The cleaned spores were suspended in 25 ml of Hz0 and layered on top of a two-layer system formed from 15 ml each of 50% and 30% NaBr in 50 mM Na phosphate buffer, pH 6.9 in 6 x 40 ml polyallomer tubes. After centrifugation in the SB-110 rotor (IEC ultracentrifuge) at 23,000 rpm for 1 hour, membranes, whole cell ghosts, vegetative and sporulating cells were removed at the interface.
The resulting spore pellet free from contamination by vegetative forms was washed twice with water and then lyophilized for storage. Heat-activated spore suspensions in Hz0 (100" for 10 min) were used as the inoculum for vegetative growth in 5 g of tryptone, 3 g of yeast extract, 2 g of glucose, 3 g of K2HP04, 2 g of KHzPOl per liter of distilled water with silicone antifoam agent at 60" with maximum possible aeration and agitation. The cells were harvested in the log phase and frozen at -15".

Preparation of Particulate
Enzyme-The frozen cells were washed with 50 mM Tris-HCl, pH 8.5, 10 mM MgC& buffer and then broken in the Mini-mill (Gifford-Wood) with glass beads in the Tris-MgCl, buffer as described previously (13). Particulate enzyme from 6 g of frozen cells was suspended in 1 ml of the Tris-MgCl, buffer and stored frozen at -80" in small aliquots (25 mg per ml of protein).
Full activity was preserved for extended periods of time by this method.
Preparation of Cell Walls-Frozen vegetative cells (49 g) were broken in batches in the Mini-mill with water (13), cycling the unbroken cells twice more through the procedure to maximize breakage.
The 5000 x g supernatant containing cell wall and cytoplasmic membrane components was centrifuged at 48,000 x g for 20 min and the resulting pellet was washed with 2 x 150 ml of water at 0". The suspension in 100 ml of water was heated to 100" with stirring for 30 min to inactivate autolytic enzymes. When cool, the pellet was washed with 2 x 80 ml of water and then lyophilized extensively.
The yield of dry cell walls plus membranes was 1.1 g.
Lysozyme Degradation of Cell Walls-Five hundred milligrams (dry weight) of cleaned cell walls plus membranes were suspended in 50 ml of 10 mM NaP04 buffer, 2.5 mM EDTA, pH 6.9 and 0.5 ml of toluene and shaken gently at 37" with 50 mg of salt-free lysozyme (Worthington Biochemicals) for 7 hours. Digestion was continued overnight and for 7 hours longer with two additions of 25 mg of fresh lysozyme.
The mixture was heated at 100" for 2 min, cooled, and spun down at 20,000 x g for 20 min. The pellet was washed with 1 ml of Hz0 and then the combined supernatant and washings were lyophilized. The residue was dissolved in 4 ml of Hz0 and streaked directly onto 3 x 39 cm lines on Whatman No. 3MM paper for electrophoresis at pH 4.0 for 5 hours at 34 volts per cm. The papers were dried at room temperature.
Marker strips, 0.75 cm wide, were cut both internally and at the edges and were stained with ninhydrin in ethanol or with the alcoholic NaOH reagent of Sharon and Seifter (21). Appropriate bands were cut from the paper and eluted four times with 0.1 N acetic acid by centrifugation in aluminum foil.
Whatman No. 3MM paper was used as the support with the Gilson model D solvent cooled apparatus.
Purification of Substrates for Amidation Reaction-Cold and [14C]MurN-labeled UDP-MurNAc-L-Ala-n-Glu-meso-Dap-n-Ala-D-Ala-and UDP-[%]GlcNAc solutions were absorbed onto short columns of well washed DEAE-cellulose (DE 52, Whatman) previously equilibrated with 50 mM Tris HCl buffer, pH 7.5. The columns were washed well with Hz0 to remove NH4+ and then the nucleotides were eluted with small volumes of 30% triethylamine carbonate, pH 9. The buffer was removed at reduced pressure over HzS04 and KOH.
Deionized water was added to give the final solutions used. Radiochemicals Ala were prepared in this laboratory (13,22).
Counting Technique-Radioactive areas on paper were counted in 15 ml of toluene-based scintillation fluid at 60 to 70% efficiency for 14C. . The tube contents were mixed well and then incubated at 37" for 1 hour. The reaction was stopped by heating to 100" for 5 min, and the entire contents of the tube were incubated with 25 ~1 of 10 mg per ml of lysozyme in 10 mM sodium phosphate buffer, 2.5 mM EDTA pH 6.9 for 12 hours at 37". The incubation was continued for 1 day with two additions of 10 ~1 of freshly prepared 10 mg per ml lysozyme in buffer. After heating for 2 min at loo", cooling and centrifuging for 10 min at 1500 x g, the supernatant was removed. The pellet was washed twice with 50 ~1 of Hz0 and then the combined supernatant and washings were spotted on a 2-cm line for electrophoresis in pH 4.0 buffer at 33 volts per cm for 4 hours.
Amino Acid Analysis-Samples of approximately 50 nmoles were hydrolyzed with 100 ~1 of degassed 4 N HCl at 105" for 17 hours in sealed tubes. The HCl was removed on a rotary evaporator and the residue was dissolved in 5.0 ml of sample diluter buffer (Beckman, Spinco Div.). Aliquots of 1.0 ml were analyzed together with 1.0 ml of a solution of 10.0 pM 2-NHz-3guanidinopropionic acid and norleucine as internal standards for the short and long column, respectively.
A modified Beckman-Spinco Auto-Analyzer model 120B was used with the technique of Peterson and Bernlohr (23) modified as follows.
The first buffer for the long column was adjusted to pH 3.09 with HCl and the change to pH 4.26 buffer was timed to occur after alanine but before 2,6-diaminopimelic acid (87 min under the conditions used). GlcN was quantitated from both the long and the short columns, and a mean is presented under "Results." For NH, quantitation, samples of 5 to 10 nmoles were treated with 100 ~1 of 0.02 x NaOH and then lyophilized to remove free NHs in the sample.
Hydrolysis and analysis on the column were carried out as above, except that a blank was taken through the same procedure, and the NH3 values were adjusted accordingly.
Dinitrophenylation-The method of Jarvis and Strominger (24) was modified.
Aliquots of about 50 nmoles were dissolved in 25 ~1 of HzO, treated with 10 ~1 of 1 y0 triethylamine in ethanol and 25 ~1 of 0.1 M I-fluoro-2,4-dinitrobenzene in ethanol, and then incubated in the dark in closed tubes at 57" for 40 min. Solvents were removed under reduced pressure and the residues were hydrolyzed as above in the dark. The hydrolyzates were added directly to 2 ml of sample diluter buffer and analyzed on the long column as above. Mono-DNP-Dap could not be estimated accurately because of variable amounts of other material absorbing at 440 nm at a similar elution time (a broad peak where valine and leucine eluted).
Quantitation of the remaining diaminopimelic acid was used to estimate the amount derivatixed, with glutamine and alanine acting as internal standards in the hy-drolyzate.
A yield of 90% derivatization of diaminopimelic acid authentic standards obtained from lysozyme digestion of Ikhewas obtained with the known disaccharide-peptide monomers (Fl rich&z coli cell walls (compare C6 and C3 in Reference 25). Dcand F2, see "Results"), and this figure was used to correct the velopment of the electrophorcsis strips in the second dimension dinitrophenylation extent for unknowns. by paper chromatography with Solvent A yielded a pattern of products which was identical with that obtained from a B. RESULTS megaterium particulate enzyme system ( Fig. 1  in the presence of UDP-MurNAc-pentapeptide dimeric, and oligomeric disaccharide-peptide products (Table I) and UDI-'-[14C]GlcNAc.
The reaction was optimal at pH 8.5 in indicated that the enzyme system achieved about 42y0 cross-Tris-HCl buffer in the presence of 10 mM MgC&. 130th the linking (from the ratio of dimcric and oligomeric products to total initial rate of peptidoglycan synthesis and the over-all yield after monomers plus dimers plus oligomers) under the conditions 2 hours were optimal at 37", although appreciable synthesis of emp1oyed. peptidoglycan still occurred in vitro at 55" (Fig. 1). In B. Labeled compounds with less mobility toward the anode at stearothermophilus NCTC 10339, peptidoglycan synthesis in pH 4.0 than the usual monomeric or dimeric products and present vitro was optimal at 55" (12).
in variable amounts were also noted on electrophoretograms. Enzyme-Prelimithat label was found in the positions expected for both disaccha-nary experiments suggested that the presence of these labeled ride-peptide monomers (GlcNAc-i\lurNAc-peptides) and cross-products of lysozyme digestion might be dependent on ATP and linked dimers (bis(GlcNAc-MurNAc-peptides)) as compared with NH,.
After the nucleotide substrates required for peptidoglycan synthesis had been freed from NH4+ by DEAE-cellulose chromatography, a clearer pattern for the requirement of ATP and sources of NH, emerged.  digested with lysozyme and separated by electrophoresis at pH 4.0 (Fig. 2). The areas corresponding to Products A, 13, and C were cut out and counted. (The amount of Product D was small and was ignored in this experiment.) The results ( slightly more efficient than NH&l. The formation of Product C appeared to be dependent on NH&l alone rather than on ATP or glutamine. Penicillin G was utilized to investigate the formation of these products further (Table III and Fig. 3). As expected, penicillin G inhibited formation of the normal crosslinked dimer (Product G) (22,26,27). The formation of Product A was stimulated by penicillin G in the presence of ATP and NH4CI as would be expected for a monomeric disaccharidepentapeptide with COOH-terminal n-Ala-n-Ala, while the formation of Product B in the presence of ATP and NH&l was inhibited. The formation of Product C in the presence of ATP and NH&l appeared also to be inhibited but the changes were small and may not be significant. The results for Product B could be explained by inhibition of n-alanine carboxypeptidase by penicillin G and for Product C by inhibition of transpeptidase (26, 27), assuming the structures elucidated below.
Products A, B, C, and D were investigated further by a larger scale preparation from UDP-[14C]GlcNAc, followed by their isolation from the pH 4.0 electrophoretogram. Electrophoretic and chromatographic data for these products are shown in Table IV. Reduction of these products with NaBH4 and then electrophoresis at pH 4.0 gave only bands at the original positions for A, B, C, and D. The absence of products with much reduced anionic character at pH 4.0 showed the absence of muramic acid lactam in these products (16,17). Treatment of A, B, C, and D with Streptomyces amidase (a gift of Dr. J.-M. Ghuysen (1)) followed with separation by electrophoresis at pH 4.0 and chromatography in Solvent C (28) showed that only the disaccharide, GlcNAc-MurNAc, was released. No traces of the tetrasaccharide (GlcNAc-MurNAc-GlcNAc-MurNAc) the isomeric disaccharide (MurNAc-GlcNAc) , muramic acid lactam-containing oligosaccharides, or disaccharides lacking N-acetyl groups were detected.
The combined data suggested that Products A, 13, C, and D were normal disaccharide-peptides containing the original peptide chain (since these products were also obtained when UDP-MurNAc-Ala-Glu-Dap-[r4C]Ala-[r4C]Ala was the labeled substrate). The most probable explanation for their reduced anionic character at pH 4.0 was the masking of a COOH group by an amidation reaction, since amidated peptide units are known in other strains of B. stearothermophilus (6, 7) and other Bacillaceae (2-4, 8, 9).
To rule out the possibility that Products A, B, C, and D might be artifacts of the preparation procedure, a particulate enzyme preparation from E. coli prepared by grinding the cells with alumina (26, 27, 29) was used to synthesize peptidoglycan from UDP-[14C]GlcNAc in the presence of ATP and NH&l. The resulting peptidoglycan was degraded with lysozyme directly or with trypsin followed by lysozyme and the resulting supernatants were separated by electrophoresis at pH 4.0 (Fig. 4). No bands were found with mobilities similar to Products A, B, C, and D from B. stearothermophilus, except for a weak band at Moi,, 0.82 which was present only in the trypsin-pretreated lysoeyme digests. This weak band could represent the disaccharide tetrapeptide, GlcNAc-MurNAc-n-Ala-n-Glu -(L) -meso -Dap -(L) -Lys, which has been reported as a trypsin degradation product of the lipoprotein-peptidoglycan conjugate from E. coli (30,31 (27,28,30). The peptidoglycan-synthesizing system, in the presence of ATP and NH&l where applicable, and the lysozyme digestion technique were identical with those used for B. stearothermophilus ("Materials and Methods"). Where indicated the boiled peptidoglycan-synthesizing mixtures were treated with 1 mg per ml of trypsin for 2 hours at 25", heated to 100" for 5 min and then treated with lysoeyme as before. The electrophoretogram prepared as in Fig. 2 was autoradiographed for 4 days.
used for preparing the particulate enzyme. After washing well with water, the cell wall and membrane mixture was degraded with lysozyme. The digestion products were separated by preparative paper electrophoresis at pH 4.0. Bands which were ninhydrin positive and which showed amino sugar fluorescence (21) werefound with mobilities corresponding to the enzymatically synthesized Products B, C, and D as well as the nonamidated disaccharide-peptide Products E, F, and G. These latter correspond to products seen in lysozyme digests of E. coli cell walls. As will be shown below, E is the cyclic dimer corresponding to E. coli compound C4; F is the monomer, CS; and G is the normal dimer C3 (25). Only a very weak band was seen corresponding to enzymatically synthesized Product A. These bands were eluted from the paper and further separated by paper chromatography in Solvent A followed by two separations in Solvent B. The final separation utilized Whatman No. 3MM paper prewashed by continuous elution with 1% acetic acid and then dried to minimize contamination with extraneous amino acids from the paper. Product B gave a single major band in Solvent A, together with a small band at Roe 1.46 which corresponded to Product A (Table IV). Chromatography of Product B in Solvent B gave a poor separation into two compounds, Bl and B2. Products C, E, and G gave only single bands in Solvents A and B. Product D was separated in Solvent A into a major band (Dl) and a minor band (D2) with Roe 0.87 which was not further investigated. Product F separated further in both Solvents A and B into two broad, overlapping bands, Fl and F2. The mobilities of those products are shown in Table V. The purified products, Ul, B2, C, Dl, E, Fl, F2, and G, were hydrolyzed. The hydrolysates were analyzed for amino acid and amino sugar content.
The expected compounds, GlcN, MurN, alanine, glutamic acid, and diaminopimelic acid, were found with only negligible contamination from other amino acids (Table VI), except for the presence in the hydrolysate of U2 of an unknown ninhydrin-positive compound which will be discussed below. Quantitation of NH8 was accomplished separately on the amino acid analyzer with carefully controlled conditions (see "Materials and Methods").
The degree of cross-linking was estimated by dinitrophenylation of the intact disaccharide-peptide followed by amino acid analysis of the hydrolysate.
The formation of mono-DNP-Dap could not be accurately quantitated because of interference from other products absorbing at 440 nm. Therefore, derivatization of NH&erminal diaminopimelic acid was measured by loss of diaminopimelic acid after dinitrophenylation.
The analytical results (Table VI)  Products Fl and F2 were the nonamidated disaccharide tripeptide and tetrapeptide monomers, and G was the nonamidated bis (disaccharide-peptide) dimer (corresponding to product C3 from E. coli, see Reference 25). Product E gave essentially the same analytical results as G. Its mobility on electrophoresis and chromatography suggested that it was a tetrasaccharide with internal cross-linking between the diaminopimelic acid of one tripeptide or tetrapeptide unit and another tetrapeptide unit, i.e. a cyclic dimer equivalent to C4 from E. coli (25).
Compounds of this type are apparently formed through transglycosylation of disaccharide-peptide dimers by lysozyme (21,25). Bl appeared to be the monoamidated-tripeptide.
The NH8 value (Table VI) for I<1 suggested a diamide, but this is ruled out by its electrophoretic mobility at pH 4.0. This high NH% value may have been due to some contamination with B2 which yielded excessive amounts of NH3 even after prior treatment with dilute alkali and lyophilization, or by the difficulty of doing NHP analyses on a compound isolated in relatively small amounts. The NH3 contents of C and Dl together with their electrophoretie mobilities at pH 4.0 suggested that they were diamidated and monoamidated bis(disaccharide-peptide) dimers, respectively.
The assignment of Dl as the "internal" amide is based on the greater mobility of the minor component D2 separated on chromatographs in Solvent A, by analogy with the mobilities of the equivalent peptides obtained after autolysis of B. subtilis cell walls on thin layer chromatography in isobutyric acid-triethylamine-H20 (100:7:43 by volume) (9). It is of interest that a