In Vitro Synthesis of Cross-linked Murein and Its Attachment to Sacculi by PBP1A from Escherichia coli*

The penicillin-binding protein (PBP) 1A is a major murein (peptidoglycan) synthase in Escherichia coli. The murein synthesis activity of PBP1A was studied in vitro with radioactive lipid II substrate. PBP1A produced murein glycan strands by transglycosylation and formed peptide cross-links by transpeptidation. Time course experiments revealed that PBP1A, unlike PBP1B, required the presence of polymerized glycan strands carrying monomeric peptides for cross-linking activity. PBP1A was capable of attaching nascent murein synthesized from radioactive lipid II to nonlabeled murein sacculi. The attachment of the new material occurred by transpeptidation reactions in which monomeric triand tetrapeptides in the sacculi were the acceptors.

The penicillin-binding protein (PBP) 1A is a major murein (peptidoglycan) synthase in Escherichia coli. The murein synthesis activity of PBP1A was studied in vitro with radioactive lipid II substrate. PBP1A produced murein glycan strands by transglycosylation and formed peptide cross-links by transpeptidation. Time course experiments revealed that PBP1A, unlike PBP1B, required the presence of polymerized glycan strands carrying monomeric peptides for cross-linking activity. PBP1A was capable of attaching nascent murein synthesized from radioactive lipid II to nonlabeled murein sacculi. The attachment of the new material occurred by transpeptidation reactions in which monomeric tri-and tetrapeptides in the sacculi were the acceptors.
Most bacteria have a murein (peptidoglycan) sacculus that forms an uninterrupted net-like structure around the cytoplasmic membrane. The sacculus is essential for the osmotic stability of the cell (1) and consists of glycan strands that are crosslinked by short peptides (2). During growth and division, the sacculus is enlarged by the incorporation of lipid II precursor by a yet unknown mechanism, which involves the coordinated action of murein synthases, murein hydrolases, and probably regulatory proteins (3).
Escherichia coli contains six known murein synthases (4 -8). However, it is not known what particular reaction these enzymes catalyze during the enlargement of the sacculus and, hence, which precise function they have in cell wall growth. Furthermore, most of the murein synthases from E. coli and other species are poorly or not characterized so far, perhaps because of difficulties in obtaining pure enzyme and lipid II substrate in sufficient quantities.
The penicillin-binding proteins (PBPs) 2 PBP1A and PBP1B are the major bifunctional murein synthases in E. coli catalyzing both the oligomerization of the glycan strands and the forma-tion of the peptide cross-links (9 -11). Both PBPs form independent dimers in vivo (12). PBP1A and PBP1B are not essential for cell growth, but cells lacking both enzymes are not viable, indicating that both have a similar, essential function that cannot be taken over by other murein synthases (5,13). Yet, mutants lacking either PBP1A or PBP1B show particular phenotypes, indicating that these synthases may play distinct roles during cell growth and division. For example, mutants without PBP1B are more sensitive to ␤-lactam antibiotics than mutants without PBP1A (14). Furthermore, PBP1B-deficient mutants, but not PBP1A-deficient mutants, lose cell integrity upon inactivation of PBP2, PBP3, or the cell division protein FtsQ (15).
The enzymatic activities of PBP1B with lipid II or artificial substrates have been studied previously (16 -19). However, apart from two early studies from Matsuhashi and colleagues (20,21), the activities of PBP1A remain uncharacterized. In this article, we report that PBP1A is capable of producing in vitro a cross-linked murein with glycan strands of (on average) almost 20 disaccharide units and with 18 -26% of the peptides being part of cross-links. We have also studied the specificity of the PBP1A-catalyzed transpeptidation reaction. For the first time, we could demonstrate in vitro an attachment of nascent murein to isolated murein sacculi by a murein synthase, namely by PBP1A.
Cloning of a mrcA Expression Plasmid-The mrcA gene encoding for PBP1A (GenBank TM accession no. AAA58193) was amplified by PCR from chromosomal DNA from E. coli wild type MC1061 using the primers 1A_up (5Ј-GGACGAATTCCC-AAATGAAACTAAATGGG-3Ј) and 1A_down (5Ј-ATTGGAT-CCATCAGAACAATTCCTGTGCC-3Ј). The PCR fragment was digested with EcoRI and BamHI and cloned into the plasmid pJFK118EH (26), resulting in pTK1A. This plasmid was used as template in a PCR for the amplification of the mrcA gene with the primers PBP1Ahis_up (5Ј-GCGGCTAGCATGAAACTA-AATGGGAAAT-3Ј) and PBP1Ahis_down (5Ј-CCTGAATTC-ATCAGAACAATTCCTGTGC-3Ј). After digestion with NheI and EcoRI, the fragment was cloned into the overexpression plasmid pET28a(ϩ) (Novagen), resulting in pTK1Ahis. The correctness of the mrcA sequence in pTK1Ahis was confirmed by sequencing.
Purification of PBP1A Variants-The purification was performed in two chromatography steps. First, 20 ml of membrane extract (see above) was incubated with 2 ml of washed and equilibrated Ni 2ϩ -nitrilotriacetic acid Superflow beads (Qiagen) for 2 h at 6°C. The beads were spun down (4200 ϫ g, 30 min, 4°C) and washed twice with buffer A containing 5 mM imidazole. His-tagged PBP1A variants were eluted with buffer A containing 500 mM imidazole for 30 min at 6°C. The eluate was dialyzed against 25 mM Tris/HCl, 10 mM MgCl 2 , 0.02% NaN 3 , 1 M NaCl, 10% glycerol, pH 7.5, for 4 h. For the removal of the His tag, 16 units of thrombin (restriction grade, Novagen) were added, and dialysis continued for 16 -20 h. Then the sample was dialyzed for 1.5 h against 2 liters of buffer I (10 mM sodium acetate, 10 mM MgCl 2 , 0.02% NaN 3 , 500 mM NaCl, 10% glycerol, pH 4.8) and for 16 h against buffer I containing 200 mM NaCl. The sample was diluted 1-fold with 10 mM sodium acetate, 10 mM MgCl 2 , 0.02% NaN 3 , pH 4.8, and the protein was purified on a 1-ml HiTrap SP HP column  Then, the membrane was destained, and biotin was detected with streptavidin-horseradish peroxidase and enhanced chemoluminescence (right side). The lanes are numbered in the same way as in A. M, molecular weight marker. Only PBP1A and PBP1A(TG*), but not PBP1A(TP*), bound biotinylated ampicillin.
Transglycosylation and Transpeptidation Reaction of PBP1A with Lipid II-Lipid II (1.2 nmol, 11,000 dpm) was vacuum dried and dissolved in 5 l of methanol. The reaction was performed in a total volume of 50 l in 25 mM Tris/HCl, 10 mM MgCl 2 , 100 mM NaCl, 0.05% Triton X-100, pH 7.5 (buffer B), with 0.5-0.8 M PBP1A variant. In different experiments, the precursor UDP-MurNAc-tripeptide, UDP-MurNAc-tetrapeptide, or UDP-MurNAc-pentapeptide was added 1 mM. In one experiment, the transpeptidation domain of PBP1A was blocked with 50 g/ml penicillin G. The reaction was performed for 1 h at 30°C. The pH of the samples was adjusted to 4.8 prior to boiling for 5 min followed by a digestion with 50 g/ml cellosyl for 1 h at 37°C and boiling for 5 min.
Time Course Experiments-In a total volume of 325 l in buffer B, lipid II (24 M) and PBP1A (0.8 M) were incubated at 30°C. At different time points (0, 5, 10, 15, 30, and 60 min), 50 l-aliquots were taken and boiled for 5 min. The samples were chilled on ice, and cellosyl digestion was done as described above.
In a second experiment, glycan strands with monomeric peptide side chains were preformed with PBP1A(TP*) in a total volume of 425 l for 1 h. The sample was boiled for 10 min, and PBP1A (0.8 M) was added followed by an incubation for 10 min on ice. After the addition of lipid II (24 M), the time course experiment was performed as described above.
Attachment of Newly Synthesized Material to Sacculi-PBP1A (0.8 M) was preincubated for 10 min on ice with 0.5 mg/ml murein sacculi (isolated from MC1061) in buffer B. Lipid II (4.8 M) was added, and the reaction proceeded in a final volume of 100 l for 1 h at 30°C. The sacculi were sedimented by centrifugation (14,000 ϫ g, 10 min), and the pellet was washed twice with 200 l of buffer A containing 0.2% Triton X-100. The pellet samples were digested for 18 h at 37°C with 35 g/ml cellosyl. The radioactivity of each sample (digested pellet, two wash fractions, and supernatant) was determined by scintillation counting.
For HPLC analysis, the reaction was performed in a final volume of 200 l. PBP1A (0.8 M) was incubated for 10 min on ice with or without 1 mg/ml sacculi (from MC1061). Lipid II (12 M) was added, and the reaction was performed for 1 h at 30°C. After sedimentation (14,000 ϫ g, 10 min), the pellets were washed twice with 200 l of buffer A containing 0.2% Triton X-100 and resuspended in 200 l of 20 mM sodium phosphate, pH 4.8. The pH of the supernatant and of the washes was adjusted to 4.8, and the samples were digested with 17.5 g/ml cellosyl for 18 h at 37°C. Samples were boiled for 5 min.
Analysis of the Reaction Products-Prior to HPLC analysis, the samples were reduced with NaBH 4 as described (28). The reaction products were separated by HPLC according to a published method (25,28), using a radioactivity flow-through detector (Canberra). The separated muropeptides were identified by their retention times with the help of muropeptide standards (19,28). Compound 1 (Fig. 4C) could be dephosphorylated to compound 2 either by extended boiling or by treatment with an acidic phosphatase followed by reduction with NaBH 4 (19). with inactive transpeptidase domain (V ). The reaction products were boiled, digested with cellosyl, reduced with NaBH 4 , and separated by HPLC with detection of radioactivity. Native PBP1A oligomerized lipid II to glycan strands and formed peptide cross-links (product peaks 2 and 3; see Fig. 4 for structures). The cross-linked product was not formed upon inactivation of PBP1A with penicillin G or with PBP1A(TP*). PBP1A(TG*) was inactive.

RESULTS
PBP1A belongs to the class A PBPs and is anchored to the cytoplasmic membrane via a (predicted) transmembrane helix (amino acids 14 -34). Most of the enzyme (amino acids 35-858) is located in the periplasm, where it contains a transglycosylase and a transpeptidase domain. The modular domain organization of PBP1A is similar to that of PBP1B (Fig.  1). The five typical signature motifs in the transglycosylase domain and the three signature motifs in the transpeptidase domain were present in both enzymes. In PBP1B, Glu 233 and Ser 510 are essential for transglycosylation and transpeptidation activity, respectively (16). The corresponding residues in PBP1A are Glu 94 and Ser 473 (Fig. 1).
To study the enzymatic reactions of PBP1A, we cloned pET28a(ϩ)-based expression plasmids for the purification of PBP1A and of two variants with single amino acid exchanges, PBP1A(TG*) with a Glu 94 3 Gln exchange and PBP1A(TP*) with a Ser 473 3 Ala exchange. Purification involved the preparation of membrane fraction, Ni 2ϩ -nitrilotriacetic acid affinity chromatography, cleavage of the N-terminal oligo-histidine tag by thrombin, and cation exchange chromatography. Fig. 2 shows the purified proteins and a ␤-lactam binding assay. As expected, PBP1A and PBP1A(TG*), but not PBP1A(TP*), bound biotinylated ampicillin.
The activities of the PBP1A variants were studied with radioactive lipid II substrate employing a recently developed in vitro FIGURE 4. Specificity of the transpeptidation reaction catalyzed by PBP1A. PBP1A (0.8 M) was incubated with lipid II (24 M) in the absence or in the presence of different soluble murein precursors (1 mM) followed by cellosyl digestion, boiling for 10 min, and reduction with NaBH 4 . The reaction products were analyzed by HPLC. A, chromatograms obtained from the sample with PBP1A and lipid II alone (I); PBP1A, lipid II, and UDP-MurNActripeptide (II); PBP1A, lipid II, and UDP-MurNAc-tetrapeptide (III); and PBP1A, lipid II, and UDP-MurNAc-pentapeptide (IV). The different peaks are numbered. B, quantification of the reaction products (average of three independent experiments). Numbers I-IV correspond to the chromatograms I-IV in A. Light gray bars, peak 1; dark gray bars, peak 2; and white bars, peak 3. The black bars represent the percentage of compounds 5, 6, and 7. C, proposed structures of the muropeptides 1-6. G, GlcNAc; M(r), reduced MurNAc; and P, phosphate. Note that compounds 1 and 2 are monomeric, and compounds 3-6 are dimeric muropeptides. PBP1A was able to transfer the disaccharide tetrapeptide to tri-, tetra-, or pentapeptide structures yielding tetratri-, tetratetra-, or tetrapenta products. murein synthesis assay ( Fig. 3; see Ref. 19 for a scheme). Briefly, boiling of lipid II yielded the phosphorylated disaccharide pentapeptide (Fig. 4C, compound 1). Because of partial dephosphorylation during the boiling and reduction steps, a small fraction (0 -15%) of nonphosphorylated disaccharide pentapeptide was formed. Upon reaction with a transglycosylase, glycan strands were produced. Digestion with cellosyl and boiling yielded nonphosphorylated disaccharide pentapeptide (Fig. 4C, compound 2) and phosphorylated disaccharide pentapeptide, the latter originating from glycan chain ends and nonreacted lipid II. If a transpeptidase is present, the procedure also yields a dimeric (cross-linked) muropeptide, the bisdisaccharide tetrapentapeptide (Fig. 4C, compound 3). PBP1A oligomerized glycan strands with an average length of 19.6 disaccharide units and formed peptide cross-links (Fig. 3). In different experiments, between 18 and 26% of the peptides became part of crosslinks. PBP1A(TP*) produced glycan strands with an average of 10.4 disaccharide units and completely lacked peptide cross-linking activity, as did PBP1A that was blocked with penicillin G. PBP1A(TG*) did not produce glycan strands and did not form peptide cross-links. The specificity of the transpeptidation reaction was studied by adding to the reaction of PBP1A with lipid II a 42-fold molar excess (over lipid II) of artificial substrates, the murein precursors UDP-MurNAc-tripeptide, UDP-MurNAc-tetrapeptide, or UDP-MurNAc-pentapeptide (Fig. 4). Addition of the UDP compounds almost completely prevented crosslinking reactions between the two lipid II molecules, as demonstrated by the small fraction of the normal bisdisaccharide tetrapentapeptide (Fig. 4, compound 3). PBP1A could use tri-, tetra-, and pentapeptide substrates as acceptors in transpeptidation reactions, leading to the formation of "mixed" transpeptidation products (Fig. 4C, compounds 4, 5, and 6). As expected, the formation of the mixed compound 4 (Fig. 4C) could be completely prevented by the preincubation of PBP1A with penicillin G, and compound 4 was not formed from lipid II and UDP-MurNAc-tripeptide by PBP1A(TP*) (data not shown).
Subsequently, we studied the time course of the reaction of PBP1A with lipid II. Strikingly, only transglycosylation (but not transpeptidation) occurred within the first 15 min of the reaction under consumption of ϳ25% of the available lipid II (Fig.  5A) and with a rate of lipid II consumption of ϳ0.75 mol/(mol PBP1A ⅐ min). Between 15 and 60 min, transglycosylation and transpeptidation reactions proceeded until almost all lipid II was consumed. Peptide cross-links were formed at a rate of ϳ0.06 mol/(mol PBP1A ⅐ min). The absence of initial transpeptidation activity indicates that PBP1A requires the presence of oligomerized glycan strands carrying monomeric peptides for cross-linking activity. Therefore, the time course experiment was repeated but this time in the presence of oligomerized glycan strands that had been preformed from lipid II by PBP1A(TP*). Indeed, if glycan strands with monomeric peptides were present, the transpeptidation reaction started imme-  (Fig. 4C), substrate plus chain ends; closed diamonds, compound 2 (Fig. 4C), transglycosylation product; closed triangles, compound 3 (Fig. 4C), dimeric transglycosylation-transpeptidation product. Initially, only transglycosylation product was detected. The transpeptidation product appeared after 15 min of reaction time. B, the mixture contained transglycosylation product (preformed by PBP1A(TP*) and lipid II), radioactively labeled lipid II, and PBP1A. Sampling and sample processing occurred as described in A. Symbols, as defined in A. In the presence of murein glycan strands, the transpeptidation reaction started immediately. diately, with cross-links being formed with a rate of ϳ0.06 mol/ (mol PBP1A ⅐ min) (Fig. 5B).
The time course experiments indicated that polymeric murein and not lipid II precursor could be the preferred acceptor for the PBP1A-catalyzed cross-linking reactions. Therefore, we tested in a pull-down experiment whether PBP1A was able to attach newly synthesized murein to isolated sacculi (Fig. 6). In the absence of murein sacculi, the reaction product of PBP1A and lipid II was not sedimented by centrifugation (Fig. 6, sample  B). In the presence of sacculi, ϳ25% of the total radioactivity sedimented together with the sacculi (Fig. 6, sample C), indicating that a fraction of the newly synthesized murein had been covalently attached to the sacculi. No attachment to sacculi occurred with PBP1A(TG*) or with PBP1A(TP*) (Fig. 6, samples D and E). The PBP1A-catalyzed attachment of new material to sacculi could be blocked with penicillin G (Fig. 6, sample F), suggesting that the attachment occurred via transpeptidation reactions.
To determine the structure(s) of the attachment site(s), the reaction products were sedimented by centrifugation, and the supernatant, as well as the sedimented murein sacculi, was digested with cellosyl followed by reduction and HPLC analysis (Fig. 7). The reaction of PBP1A and lipid II was performed in the presence and absence of murein sacculi. In the absence of murein sacculi (Fig. 7A, left two bars), 98.1% of the radioactivity was present in the supernatant, and the produced murein had an average glycan strand length of 18.2 disaccharide units, with 26.4% of the peptides being part of cross-links. In the presence of murein sacculi (Fig. 7A, right two bars), ϳ44% of the total radioactivity was sedimented together with the sacculi, and ϳ56% of the total radioactivity was present in the supernatant. HPLC analysis revealed a different muropeptide composition in the supernatant and in the pellet (Fig. 7B); the quantification of the respective compounds is depicted in C. The supernatant (Fig. 7B, chromatogram III) contained a similar muropeptide profile as the sample without murein sacculi (chromatogram I), except that the average glycan strand length was decreased to 6.1 disaccharide units. In the pellet, the average glycan strand length was Ͼ40 disaccharide units, and it contained additional radioactive cross-linked compounds (Fig. 7B, chromatogram IV), the bisdisaccharide tetratripeptide and the bisdisaccharide tetratetrapeptide. These compounds could only be produced by cross-linking reactions between new pentapeptides (from radioactive lipid II) as donors and monomeric tri-and tetrapeptides in the sacculi as acceptors. Thus, the attachment of new material to sacculi by PBP1A occurred by transpeptidation reactions with monomeric (and not with cross-linked dimeric or trimeric) acceptors.

DISCUSSION
PBP1A formed in vitro a cross-linked, oligomeric murein from lipid II precursor. As predicted from sequence comparison (Fig. 1), the Glu 94 and Ser 473 residues are essential for the catalytic activities. Although the PBP1A(TP*) variant with inactive transpeptidase domain could still catalyze transglycosylation reactions, a PBP1A(TG*) variant with inactive transglycosylase domain lacked both transglycosylase and transpeptidase activities. Similar observations have been made with inactive variants of PBP1B from E. coli (16). Transpeptidation by PBP1A(TG*) occurred neither in the presence of an oligomeric reaction product produced by PBP1A(TP*) nor in a reaction mixture containing PBP1A(TG*), PBP1A(TP*), and lipid II (data not shown). Therefore, it is possible that the transpeptidase domain of PBP1A is activated by ongoing transglycosylation reactions and that the cross-linking occurs preferentially with pentapeptide donors in the glycan strands produced by the transglycosylase domain.
The time course experiments of the reaction of PBP1A with lipid II revealed that the cross-linking reaction was delayed and started only after glycan strands with monomeric pentapeptides had been produced by transglycosylation. If such material was added initially to PBP1A and lipid II, formation of crosslinks started immediately and proceeded until 21% of the newly added peptides became part of cross-links. On the other hand, only 2% of the peptides became part of cross-links if oligomerized glycan strands were incubated with PBP1A alone (without lipid II) for 60 min (not shown). Apparently, transpeptidation requires ongoing transglycosylation (see above) and, in addition, the presence of oligomerized glycan strands carrying monomeric peptides. Either binding of the enzyme to this oligomeric material activates the transpeptidase function or the oligomeric material is a substrate for the transpeptidation reaction (or both possibilities). We have obtained evidence that monomeric peptides in high molecular weight murein function as acceptors in PBP1A-catalyzed cross-linking reactions. First, a fraction of newly synthesized murein was covalently attached to murein sacculi, and this attachment was inhibited by penicillin G, an inhibitor of the transpeptidation reaction (Fig. 6). Second, we identified dimeric muropeptides with tetratripeptide and tetratetrapeptide cross-links as products, proving that cross-linking reactions had occurred between pentapeptides (from lipid II and/or newly synthesized glycan strands) as donors and tri-and tetrapeptides in the sacculi as acceptors (Fig. 7). The observed cross-linking reactions of PBP1A, with acceptor peptides in nascent murein or in murein sacculi, are depicted schematically in Fig. 8. Interestingly, the formation of  (12 M) in the absence and presence of nonlabeled murein sacculi (1 mg/ml). The samples were centrifuged, and the supernatant and pellet were treated with cellosyl. A, an aliquot of the samples was taken to quantify the radioactivity in the supernatants (gray bars) and pellets (black bars). B, the samples were reduced by NaBH 4 and analyzed by HPLC using a radioactivity detector. Chromatograms I-IV correspond to the samples I-IV in A. Chromatogram V shows the UV profile of a digestion of sacculi with cellosyl. Note that sacculi contained only minor amounts of pentapeptide compounds. C, quantification of the radioactive peaks in chromatograms I-IV. Dashed bar, peak 1; light gray bar, peak 2; white bar, peak 3; black bar, peak C; dark gray bar, peak B. D, proposed structures of peaks 1, 2, 3, A, B, and C. G, GlcNAc; M(r), reduced MurNAc; P, phosphate. In the absence of sacculi, the radioactive transglycosylase and transpeptidase products were present only in the supernatant. In the presence of sacculi, 44% of the newly synthesized murein was present in the pellet. Attachment of newly synthesized murein to sacculi occurred via cross-linking (transpeptidation) reactions to tri-and tetrapeptides in the sacculi, yielding the radioactive compounds B and C (chromatogram IV ).
cross-links between nascent murein and sacculi did not take place at the expense of the formation of crosslinks between new (pentapeptide) donors and acceptors; in the presence of sacculi, ϳ23% of the peptides became part of tetrapentapeptide cross-links (in the attached material), which was only slightly less than in the murein produced in the absence of sacculi (ϳ26%) (Fig.  7). In this experiment, a total of ϳ30% of the pentapeptides (originating from lipid II) had participated as donors in transpeptidation reactions, of which 11.5% reacted with pentapeptides (in lipid II) and 19% with tri-and tetrapeptides (in sacculi) as acceptors.
PBP1A and PBP1B are similar in their in vitro activities in that both can use artificial tri-, tetra-, and pentapeptides (in UDP-linked precursors) as acceptors for transpeptidation reactions (19). Such reactions have been observed before in ether-permeabilized cells (29,30). On the other hand, PBP1A and PBP1B are different with respect to several properties. (i) The transglycosylation and transpeptidation rates were significantly higher for PBP1B. (ii) Under optimum condition, PBP1B produced a murein with almost 50% of the peptides being present in cross-links. With PBP1A, we observed a maximum of 26% of the peptides becoming part of cross-links. (iii) PBP1B produced a trimeric compound by cross-linkage of a pentapeptide donor and a tetrapentapeptide acceptor. A previous study reported the production of a trimeric cross-linked product by PBP1A (21). We have observed in some experiments the formation of a putative trimeric product only in small quantities (Fig. 7B, chromatogram I, see peak at 100 min). It is possible that the differences in the enzyme preparation methods and, hence, in the purities of the enzyme account for this discrepancy. (iv) Although both PBP1A and PBP1B required the presence of Triton X-100 for solubilization, the activities of PBP1A, but not of PBP1B, significantly decreased if the concen- FIGURE 8. Schemes for TG and TP reactions catalyzed by PBP1A. PBP1A polymerizes lipid II by transglycosylation and simultaneously attaches the growing glycan strand by transpeptidation to monomeric peptides, present either in another newly synthesized glycan strand (A) or in a murein sacculus (B). The leaving groups (undecaprenyl pyrophosphate and D-alanine) are encircled by a dashed line. The cross-linking reaction with two pentapeptides (A) yields a tetrapentapeptide product, whereas the cross-linking reaction between a pentapeptide (donor) and a tri-or tetrapeptide acceptor yields a tetratri-or tetratetrapeptide cross-link. Free amino groups at meso-diaminopimelic acid residues that represent possible acceptor sites for transpeptidation reactions are indicated (NH 2 ). Glycan strands are shown as rods with gray (MurNAc) and white (GlcNAc) parts. Straight arrow, peptide; black circle, phosphate group; zigzag line, undecaprenyl residue.
trations of the detergent exceeded 0.2% (data not shown). (v) PBP1B was most active as a dimer, whereas dimerization of purified PBP1A was not observed. (vi) Unlike in the case of PBP1A, time course experiments did not reveal a preference of PBP1B for polymerized murein glycan strands with monomeric peptides as acceptors for the transpeptidation reaction. In addition, we have observed that PBP1B attached a significantly smaller fraction of nascent murein to sacculi in vitro as compared with PBP1A (data not shown). It is possible that differences in the activities of both enzymes account for the different phenotypes of mutants lacking either PBP1A or PBP1B (see the Introduction).
For the first time, we could demonstrate in vitro an attachment of nascent murein to sacculi, a reaction that occurs in living cells during the growth of the sacculus. Similar to our in vitro experiments, it has been shown that incorporation of new material in vivo involves the formation of new cross-links; during cell division, cross-links are formed mainly between two new peptides, whereas during cell elongation, cross-links are also formed between new peptides and peptides in the existing sacculus (31)(32)(33)(34). In vitro, PBP1A most likely attached new material without increasing the surface of the sacculi. The enlargement of the sacculus in vivo is a more complex process and presumably requires a coordinated interplay between different bifunctional and monofunctional murein synthases, murein hydrolases, and regulatory proteins. The precise role of PBP1A in the enlargement of the sacculus during cell elongation and division remains to be determined. Our results indicate that PBP1A might synthesize patches of nascent murein and attach them to the sacculus.