The MurC Ligase Essential for Peptidoglycan Biosynthesis Is Regulated by the Serine/Threonine Protein Kinase PknA in Corynebacterium glutamicum*

The Mur ligases play an essential role in the biosynthesis of bacterial cell-wall peptidoglycan and thus represent attractive targets for the design of novel antibacterials. These enzymes catalyze the stepwise formation of the peptide moiety of the peptidoglycan disaccharide peptide monomer unit. MurC is responsible of the addition of the first residue (l-alanine) onto the nucleotide precursor UDP-MurNAc. Phosphorylation of proteins by Ser/Thr protein kinases has recently emerged as a major physiological mechanism of regulation in prokaryotes. Herein, the hypothesis of a phosphorylation-dependent mechanism of regulation of the MurC activity was investigated in Corynebacterium glutamicum. We showed that MurC was phosphorylated in vitro by the PknA protein kinase. An analysis of the phosphoamino acid content indicated that phosphorylation exclusively occurred on threonine residues. Six phosphoacceptor residues were identified by mass spectrometry analysis, and we confirmed that mutagenesis to alanine residues totally abolished PknA-dependent phosphorylation of MurC. In vitro and in vivo ligase activity assays showed that the catalytic activity of MurC was impaired following mutation of these threonine residues. Further in vitro assays revealed that the activity of the MurC-phosphorylated isoform was severely decreased compared with the non-phosphorylated protein. To our knowledge, this is the first demonstration of a MurC ligase phosphorylation in vitro. The finding that phosphorylation is correlated with a decrease in MurC enzymatic activity could have significant consequences in the regulation of peptidoglycan biosynthesis.

The Mur ligases play an essential role in the biosynthesis of bacterial cell-wall peptidoglycan and thus represent attractive targets for the design of novel antibacterials. These enzymes catalyze the stepwise formation of the peptide moiety of the peptidoglycan disaccharide peptide monomer unit. MurC is responsible of the addition of the first residue (L-alanine) onto the nucleotide precursor UDP-MurNAc. Phosphorylation of proteins by Ser/Thr protein kinases has recently emerged as a major physiological mechanism of regulation in prokaryotes. Herein, the hypothesis of a phosphorylation-dependent mechanism of regulation of the MurC activity was investigated in Corynebacterium glutamicum. We showed that MurC was phosphorylated in vitro by the PknA protein kinase. An analysis of the phosphoamino acid content indicated that phosphorylation exclusively occurred on threonine residues. Six phosphoacceptor residues were identified by mass spectrometry analysis, and we confirmed that mutagenesis to alanine residues totally abolished PknA-dependent phosphorylation of MurC. In vitro and in vivo ligase activity assays showed that the catalytic activity of MurC was impaired following mutation of these threonine residues. Further in vitro assays revealed that the activity of the MurC-phosphorylated isoform was severely decreased compared with the non-phosphorylated protein. To our knowledge, this is the first demonstration of a MurC ligase phosphorylation in vitro. The finding that phosphorylation is correlated with a decrease in MurC enzymatic activity could have significant consequences in the regulation of peptidoglycan biosynthesis.
Due to the increasing number of antibiotic-resistant strains and the emergence of new pathogenic microorganisms, one of the biggest challenges for modern biomedical research is the continuous development of new antimicrobial drugs targeting bacterial essential mechanisms such as cell division or peptidoglycan (PG) 4 biosynthesis (1). The bacterial cell wall PG is a giant molecule that sustains the shape of the bacterial cell and contains the outward forces generated in maintaining an osmotic pressure gradient against the environment. Without this PG layer the cell integrity would be ruptured, and this could lead to cell death. Therefore the PG biosynthesis machinery represents a promising source of putative targets for antibacterial chemotherapy (2,3).
The biosynthesis of bacterial PG is a complex two-stage process (4). The first stage involves the assembly of the disaccharide peptide monomer unit by enzymes located in the cytoplasm or at the inner surface of the cytoplasmic membrane (3,5). The peptide moiety of the monomer unit is assembled stepwise by the successive additions of L-alanine, D-glutamic acid, meso-diaminopimelic acid or L-lysine, and D-alanyl-D-alanine to UDP-N-acetylmuramic acid (UDP-MurNAc). These steps are catalyzed by specific peptide synthetases (ligases), which are designated as MurC, MurD, MurE, and MurF, respectively, all participating in non-ribosomal peptide bond formation with the concomitant hydrolysis of ATP. The MurNAc-pentapeptide motif of the resulting nucleotide precursor is then transferred by the MraY translocase onto the undecaprenyl phosphate carrier molecule, generating the lipid intermediate I. The subsequent addition of the N-acetylglucosamine motif of UDP-GlcNAc onto lipid I generates lipid II in a reaction catalyzed by MurG (6). The second stage of the PG biosynthesis consists in the polymerization by transglycosylation and transpeptidation reactions of the disaccharide pentapeptide monomers, a reaction taking place in the periplasmic space and that is catalyzed by the penicillin-binding proteins.
The PG biosynthesis pathway enzymes, which are essential and specific for bacteria, represent important potential targets for screening novel antibacterial compounds. Due to the growing emergence of bacterial multiresistance to currently used antibiotics, the discovery of new therapeutic compounds has indeed become a necessity. In recent years, an extensive search for specific inhibitors interfering with the cytoplasmic steps of this pathway, and in particular with the four steps catalyzed by the Mur ligases, has been developed (2,7,8). The UDP-Mur-NAc:L-alanine ligase (MurC), encoded by the murC gene, represents such an interesting candidate for drug development (9,10). Recently, different phosphinic acid derivatives and substrate analogues have been identified as Mur ligase inhibitors (11,12).
Corynebacterium glutamicum is a rod-shaped non-pathogenic Gram-positive actinomycete widely used in the industrial production of amino acids such as L-lysine and L-glutamic acid (13). C. glutamicum has been extensively studied as a model microorganism due to the strategies employed by this actinomycete to achieve a rod-shaped morphology. In fact, the mechanisms taking place in C. glutamicum happened to be completely different from that of Escherichia coli or Bacillus subtilis (14,15), whereas the number of genes involved in cell division and PG biosynthesis in C. glutamicum is lower (16).
Interestingly, an earlier work on the phosphoproteome of C. glutamicum (17) identified MurC has being phosphorylated in vivo, suggesting that protein phosphorylation plays a much broader function in C. glutamicum than was previously expected. Recently, we described the characterization of the four STPKs from C. glutamicum ATCC 13869 and highlighted their role in cell division (18). Moreover, Thakur and Chakraborti (19) showed that MurD from Mycobacterium tuberculosis was phosphorylated by the Ser/Thr protein kinase (STPK) PknA, although no further characterization of the role of the phosphorylation on the MurD enzyme activity was investigated. Therefore, it was tempting to speculate that MurC in C. glutamicum could also be regulated by STPK phosphorylation.
The focus of this work is to study the regulation of MurC in C. glutamicum via phosphorylation. As a first step in deciphering the potential role/participation of the corynebacterial STPKs in the regulation of MurC activity, we confirmed its specific phosphorylation by the PknA kinase through a combination of in vitro phosphorylation assays and mass spectrometric identification of the different MurC phosphorylation sites. Moreover, we demonstrated that the murein ligase activity of MurC was negatively regulated upon its phosphorylation. To our knowledge, this work represents the first evidence of a Mur enzyme regulated by phosphorylation.

EXPERIMENTAL PROCEDURES
Bacterial Strains and Growth Conditions-Bacterial strains and plasmids are described in Table 1. Strains used for cloning and expression of recombinant proteins were E. coli TOP10 (Invitrogen) and E. coli BL21(DE3)Star (Stratagene), respectively. E. coli cells were grown and maintained at 37°C in LB medium supplemented with 100 g/ml ampicillin and/or 50 g/ml kanamycin, when required. The murC temperature-sensitive E. coli strain H1119 was grown at 30°C in 2YT (1.6% Bactotrypton, 1.0% Bactoyeast extract, 0.5% NaCl, pH 7.0) medium and was used for genetic complementation experiments with plasmids carrying wild-type or mutated copies of the murC gene. C. glutamicum cells were grown at 30°C in TSB (Trypticase soy broth, Oxoid) or TSA (TSB containing 2% agar) medium supplemented with 12.5 g/ml kanamycin. Plasmids to be transferred by conjugation from E. coli to corynebacteria were introduced by transformation into the donor strain E. coli S17-1. Mobilization of plasmids from E. coli S17-1 to C. glutamicum R31 was accomplished as described previously (20).
Cloning, Expression, and Purification of MurC Proteins-First, the murC gene was cloned to generate a recombinant MurC protein expressed in E. coli. Therefore, the murC gene was amplified by PCR using C. glutamicum ATCC 13869 genomic DNA as a template and the primers pair murC1/ murC2 (Table 2), containing NdeI and NheI restriction sites, respectively. The 1461-bp amplified product was digested by NdeI and NheI and ligated to the pETTev vector (Table 1) generating the pTEVmurC plasmid. E. coli BL21(DE3)Star cells transformed with this construction were used for expression and purification of His 6 -tagged MurC, as previously described (21). Finally, the purified His 6 -tagged MurC was treated with TEV protease according to the manufacturer's instructions (Invitrogen). Secondly, overexpression and purification of MurC from C. glutamicum cultures was performed using standard PCR strategies. The murC gene from C. glutamicum was amplified using the primers pair murC1/murHisNdeI2 ( Table 2). The PCR product carrying a His tag at its C-terminal end was digested with NdeI and subsequently cloned under the control of the Pdiv promoter into plasmid pEDiv ( Table 1). The resulting expression vector, named pEDivmurHis, was introduced by conjugation into C. glutamicum R31. Purification of the soluble His 6 -tagged MurC protein from C. glutamicum was performed as described previously (21).
In Vitro Kinase Assays-In vitro phosphorylation was performed with 2 g of MurC in 20 l of buffer P (25 mM Tris-HCl, pH 7.0, 1 mM dithiothreitol, 5 mM MgCl 2 , 1 mM EDTA) with 200 Ci/ml [␥-33 P]ATP corresponding to 65 nM (PerkinElmer Life Sciences, 3000 Ci/mmol), and 0.5 g of kinase. Plasmids pGEXA, pGEXB, pGEXL, and pTEVGfull (Table 1) were used for the expression and purification in E. coli of the four recombinant STPKs from C. glutamicum as previously described (18). After 15-min incubation, the reaction was stopped by adding sample buffer and heating the mixture at 100°C for 5 min. The reaction mixtures were analyzed by SDS-PAGE. After electrophoresis, gels were soaked in 20% trichloroacetic acid for 10 min at 90°C, stained with Coomassie Blue, and dried. Radioactive proteins were visualized by autoradiography using direct exposure films.
Analysis of the Phosphoamino Acid Content of Proteins-MurC sample (5 g) phosphorylated in vitro by the GST-tagged PknA, and unreacted [␥-33 P]ATP were separated by one-dimensional gel electrophoresis and electroblotted onto an Immobilon polyvinylidene difluoride membrane. The 33 P-labeled protein bands were detected by autoradiography and excised from the Immobilon blot and hydrolyzed in 6 M HCl for 1 h at 110°C. The acid-stable phosphoamino acids released were separated by electrophoresis in the first dimension at pH 1.9 (800 V/h) in 7.8% acetic acid, 2.5% formic acid, followed by ascending chromatography in the second dimension in 2-methyl-1-propanol/formic acid/water (8:3:4, v/v). After migration, radioactive compounds were detected by autoradiography. Authentic phosphoserine, phosphothreonine and phosphotyrosine were run in parallel and visualized by staining with ninhydrin.
Immunoblotting-Corynebacterial MurC purified either from E. coli or C. glutamicum was loaded on a 10% polyacrylamide gel, electrophoresed, blotted on polyvinylidene difluoride, and detected using either monoclonal mouse anti-phospho-threonine, -serine, -tyrosine, or polyclonal rabbit anti-His antibodies used at 1:100 and 1:10,000 dilution, respectively. Alkaline phosphatase-conjugated anti-mouse or anti-rabbit was used as a secondary antibody at a 1:5,000 dilution.
Complementation with C. glutamicum MurC-Plasmids pTEVmurC, pTEVmurC1T, pTEVmurC2T, pTEVmurC3T, pTEVmurC4T, pTEVmurC5T, and pTEVmurC6T were used as templates for PCR amplification of the murC gene with the primers pair murH1119-1/murH1119-2 (Table 2) containing a BspHI and BglII restriction site, respectively. The resulting 1461-bp products, carrying the wild-type and mutated gene copies, respectively, were digested with BspHI and BglII, and cloned into the plasmid vector pTrc99A (22) between the compatible NcoI and BamHI sites, generating pTrcmurC, pTrcmurC1T, pTrcmurC2T, pTrcmurC3T, pTrcmurC4T, pTrcmurC5T, and pTrcmurC6T, respectively. These plasmids allowing high level expression of the C. glutamicum MurC and MurC mutants under the control of the strong isopropyl 1-thio-␤-D-galactopyranoside-inducible trc promoter, were transformed in the E. coli MurC temperature-sensitive mutant strain H1119 (23). Transformants were grown at the permissive temperature (30°C) before being shifted to the restrictive temperature (42°C). Complementation was judged by the ability of the mutant to grow at the restrictive temperature.
UDP-MurNAc L-alanine Ligase Assay-The L-alanine-adding activity of wild-type and mutant MurC proteins was assayed according to Liger et al. (24) a Restriction sites are underlined and specified by brackets. b Mutagenized codons are shown in bold.
followed by lyophilization. To test the effect of the phosphorylation of MurC on its enzyme activity, the assay was performed as described above excepted that the MurC enzyme (20 g/ml) was preincubated overnight at 37°C with or without wild-type PknA (16 g/ml), or with PknA_K49M mutant (30 g/ml) in buffer A supplemented with 2 mM ATP and 5 mM MgCl 2 . The radioactive product UDP-MurNAc-L-[ 14 C]alanine and substrate L-[ 14 C]alanine were separated by reversed-phase highperformance liquid chromatography on a Nucleosil 5C18 column (4.6 ϫ 150 mm, Alltech France) using 50 mM ammonium formate, pH 3.9, as the eluent, at a flow rate of 0.6 ml/min. Detection was performed with a radioactive flow detector (model LB506-C1, Berthold France, La Garenne-Colombes, France) using the Quicksafe Flow 2 scintillator (Zinsser Analytic, Maidenhead, UK) at 0.6 ml/min. Quantitation was carried out with the Radiostar software (Berthold).

RESULTS AND DISCUSSION
MurC Is a Substrate of the C. glutamicum PknA-The peptide moiety of the PG monomer unit is assembled stepwise in the cytoplasm by the successive actions of four Mur ligases designated as MurC, MurD, MurE, and MurF. These enzymes share limited sequence identity but have several highly conserved regions that map primarily to the active site. Each of them comprises three structural domains: an N-terminal domain with a Rossmann-type fold primarily responsible for binding of the UDP-MurNAc(-peptide) substrate, a large cen-tral ATP-binding (ATPase) domain, and a C-terminal domain associated with binding of the amino acid substrate (25). Based on protein sequence alignments and on the three-dimensional structure of the E. coli MurC protein (26), these three domains in the C. glutamicum MurC may extend between Met 1 -Gly 118 , Ser 119 -Arg 334 , and Arg 335 -Asn 486 , respectively ( Figs. 1 and 4). By comparing the amino acid sequences of the MurC, -D, -E, and -F ligases from different bacterial genera, several conserved residues were identified. This analysis led to the characterization of the ATP-binding motif located in the conserved sequence (GXXGK(T/S)), and therefore named the GKT motif (Fig. 1). There are also a number of common conserved amino acids residues present in the different Mur ligases. Taking E. coli as a model, it was established that Asp 50 , Lys 130 , Glu 174 , and Asp 351 were essential for the catalytic process of these ligases, whereas residues His 199 , Asn 293 , Asn 296 , and Arg 327 were involved in the structure of the active site. The sequence of several MurC proteins from representative species of mycobacteria and corynebacteria were aligned using the ClustalW and Espript programs (Fig. 1). The invariant residues Asp 50 , Lys 130 , Glu 174 , Asn 296 , and Arg 327 appeared conserved in all the sequences aligned, thus confirming that MurC from C. glutamicum harbors most of the characteristics of a functional Mur ligase enzyme. Therefore, the recombinant MurC protein was expressed and purified from E. coli BL21(DE3)Star harboring the pTEVmurC. The protein contained an N-terminal His E_col, E. coli). The conserved GKT motif involved in ATP binding is represented by a shaded oval. Triangles represent PknA-dependent phosphorylation sites (Thr 51 , Thr 120 , Thr 133 , Thr 167 , Thr 362 , and Thr 365 ) of C. glutamicum MurC. Residues indicated by an asterisk in the E. coli MurC sequence correspond to residues that were previously identified as invariant within the whole Mur ligase family (3,25): the Asp 50 , Lys 130 (GKT motif), and Glu 174 residues involved in the catalytic process, and the Asn 296 and Arg 327 residues involved in the structural organization of the active site. Numbering of amino acids corresponds to the MurC protein from C. glutamicum. tag, which was subsequently removed after cleavage with the TEV protease. Analysis of the purified recombinant protein by SDS-PAGE revealed that the protein was expressed in a soluble form migrating slightly above its predicted molecular mass of 52 kDa ( Fig. 2A).
Previous work on the phosphoproteome of C. glutamicum (17) identified MurC as being phosphorylated in vivo. Moreover, we have recently reported the characterization of the four STPKs from C. glutamicum ATCC 13869 and highlighted their role in cell division (18). These results prompted us to investigate whether MurC would also represent a substrate for corynebacterial STPKs. A systematic approach was used to investigate whether STPKs of C. glutamicum (PknA, PknB, PknG, or PknL) could phosphorylate MurC. All these STPKs were expressed and purified from E. coli as described previously (18). The different STPKs migrate as diffuse bands reflecting the different levels of phosphorylation for each isoform, and this aberrant profile of migration of STPKs kinases has already been reported in earlier studies (27)(28)(29). Interestingly, when STPKs were incubated in the presence of recombinant MurC and [␥-33 P]ATP, phosphorylation of MurC was specifically observed with the PknA kinase, whereas PknB and PknL, which display autophosphorylation activity in vitro, did not phosphorylate MurC (Fig. 2A). The PknG kinase needs to be activated by the PknA kinase to trigger its autophosphorylation activity, therefore PknG was phosphorylated by PknA as previously described and then used to transphosphorylate MurC (18). As shown on Fig. 2A Fig. 2A).
We next determined the nature of the amino acid residues phosphorylated in MurC by analyzing the phosphoamino acid content of PknAphosphorylated MurC. The MurC protein (5 g) was labeled with [␥-33 P]ATP in vitro, separated by SDS-PAGE, excised, and subjected to acid hydrolysis. Fig. 2B shows that MurC was exclusively phosphorylated on threonine residues.
Together, these data suggest a specificity of substrate recognition by PknA toward MurC. Although a specific interaction between MurC and PknA may exist, it remains to be established whether this specific partnering also occurs in vivo.
The Activation Loop Thr 179 and Thr 181 Residues Are Essential for the Autophosphorylation Activity of C. glutamicum PknA-Diverse mechanisms of eukaryotic protein kinase regulation have been described. The transition between active and inactive forms may occur via control of access to the catalytic site and/or the substrate-binding site. These regulatory mechanisms involve phosphorylation/dephosphorylation via an autocatalytic mechanism or the action of other kinases and phosphatases.
The activation loop appears to be a major control element of an active/inactive conformational switch in numerous kinases (30), and the conformation of the activation loop often depends on the phosphorylation state (31). Based on structural studies, it is thought that the activation loop controls both the accessibility to the catalytic site and the binding of the substrate. Recently, Canova et al. (27) demonstrated that the phosphorylated residues Thr 173 and Thr 175 present in the activation loop were essential for the autophosphorylation activity of the M. tuberculosis PknL kinase, and that phosphorylation of the Thr 173 residue was also required for optimal PknL-mediated phosphorylation of its substrate Rv2175c. Interestingly, the corresponding threonine residues Thr 179 and Thr 181 found in the C. glutamicum PknA were recently identified as phosphorylation sites (18). The presence of such residues in PknA from C. glutamicum supports the concept that phosphorylation of the activation loop could play a regulatory role, as recently described for other mycobacterial STPKs (27, 32, 33). To determine a possible link between Thr 179 and Thr 181 and the autophosphorylation activity of PknA, the two residues were individually substituted by Ala to generate PknA_T179A and PknA_T181A, respectively. In addition, a double mutant was also created and designated PknA_T179A/T181A. These PknA mutants were expressed and purified from E. coli as previously described (18). The electrophoretic migration profile showed that the different PknA mutants migrated differently from the wild-type protein (Fig. 3A, upper panel). We reasoned that, because these proteins displayed different electrophoretic mobility properties, they must differ in their intrinsic phosphorylation states. This hypothesis was confirmed by labeling the mutant proteins in vitro with [␥-33 P]ATP. PknA_T179A gave rise to a lower radioactive signal than the wild-type enzyme (Fig. 3A, lower panel). Quantification of the signal intensity indicated an autophosphorylation activity of 47% with respect to the activity of the wild-type protein, which was arbitrarily placed at 100%. This result argues that Thr 179 is a key phosphorylation site of the activation loop, necessary for PknA autophosphorylation activity. The signal intensity generated by PknA_T181A was intermediate between those of the wild-type and Thr 179 mutant proteins (Fig. 3A, lower panel), representing 81% of residual activity with respect to the wild-type protein activity. Therefore, this result confirms the requirement of Thr 181 for optimal PknA autophosphorylation activity, although this residue appears less important than Thr 179 . When both Thr residues were mutated, the intensity of the signal was similar to the one of the single Thr 179 mutant (Fig.  3A, lower panel). Together, these results confirm that double phosphorylation of the activation loop residues Thr 179 and Thr 181 is necessary for full kinase activity and unambiguously demonstrate the role of both phosphothreonines.
To exclude the possibility of exogenous contamination that could explain labeling of PknA with [␥-33 P]ATP, Lys 49 was mutated to Met. Purified PknA_K49M, was incubated in the presence of [␥-33 P]ATP, and as expected, no signal could be detected as Lys 49 is essential for catalyzing the phosphorylation reaction, in agreement with previous reports (Fig. 3A, lower  panel) (34,35). This confirms that the radioactive signals observed for the different isoforms of the PknA kinase are only representing the autophosphorylation activity of PknA and not a contamination from the purification procedure.
Phosphorylation of MurC Is Dependent on the PknA Activation Loop Thr 179 Residue-Several recent publications indicated that phosphorylation of the threonine residues present in the STPK activation loop is not only necessary for controlling the kinase activity but is also required for recruitment and phosphorylation of its substrate (27,36). These results prompted us to analyze the contribution of the activation loop key residues of PknA (Thr 179 and Thr 181 ) in the transphosphorylation mechanism between the kinase and its substrate MurC. This was achieved by incubating recombinant MurC with PknA_K49M, PknA_T179A, PknA_T181A, or PknA_T179A/T181A in the presence of [␥-33 P]ATP (Fig. 3B). PknA_K49M was unable to phosphorylate C. glutamicum MurC, indicating that phosphorylation of PknA is a prerequisite to the transphosphorylation reaction (Fig. 3B, lower panel). More importantly, whereas the T179A mutation completely abolished the transphosphorylation reaction, replacement of Thr 181 by Ala did not alter PknA-dependent phosphorylation of MurC (Fig. 3B, lower panel). Furthermore, MurC could not be phosphorylated by PknA_T179A/ T181A, clearly demonstrating that phosphorylation of MurC is dependent on Thr 179 . The finding that MurC interacts only with the phosphorylated form of PknA and that this interaction is abolished by the T179A substitution suggests that this phosphopeptide recognition motif is involved in protein-protein interaction between the two corynebacterial partners, as previously described in its related actinomycete M. tuberculosis (27).
MurC Is Phosphorylated on Multiple Threonine Residues-The role of post-translational mechanisms like phosphorylation in regulatory processes or the effect of phosphorylation on a given substrate in the physiology and/or cell division represent key events to our understanding of the signaling mechanisms through serine/threonine phosphorylation. This usually requires the identification of the phosphorylated sites, which often remains very challenging. To identify which of the 31 threonine residues of the C. glutamicum MurC corresponded to the phosphorylated site(s), a mass spectrometry approach was used. This technique has recently been proven to be a method of choice for characterizing post-translational modifications such as phosphorylation (18,27,37). LC-ESI/MS/MS was applied for the identification of phosphorylated peptides and for the localization of phosphorylation sites in MurC. Purified MurC was subjected to in vitro phosphorylation by PknA with non-radioactive ATP, prior to resolving on SDS-PAGE, and in-gel digestion with either trypsin or chymotrypsin. Phos-phorylated amino acid residues were assigned by peptide fragmentation in MS/MS: y and b daughter ions containing one phosphothreonine were associated with a neutral loss of phosphoric acid (ϪH 3 PO 4 , i.e. Ϫ98 Da). As detailed in Table 3, analysis of tryptic and chymotryptic digests allowed the characterization of six phosphorylation sites in MurC corresponding to Thr 51 , Thr 120 , Thr 133 , Thr 167 , Thr 362 , and Thr 365 (Figs. 1 and 4). These residues were found in the three structural domains of  the MurC protein, two of them, namely Thr 120 and Thr 133 , being located very close to the ATP-binding site of the protein ( 129 GKT 131 ). Thus, MS profiling coupled to in vitro kinase assays unambiguously demonstrated the phosphorylation of MurC by the PknA kinase. The fact that threonines but not serine residues were identified was consistent with our phosphoamino analysis (Fig. 2B).
Definitive identification and localization of the six threonine residues identified by mass spectrometry was achieved by sitedirected mutagenesis by introducing mutations that prevent their specific phosphorylation. Thus, all the six threonine residues were sequentially replaced by alanine, yielding the MurC_1T to MurC_6T mutants. These mutant proteins were expressed, purified as His-tagged proteins in E. coli BL21(DE3) Star harboring the different mutant constructs (Table 1), and used in an in vitro kinase assay. After TEV protease cleavage of the His tag, the recombinant MurC phosphorylation site mutants were incubated with [␥-33 P]ATP and PknA. The proteins were separated by SDS-PAGE and analyzed by autoradiography. As shown in Fig. 4B (upper panel), equal amounts of MurC_WT or MurC mutants were used. Phosphorylation of the MurC_6T protein in which all six threonine residues were replaced by alanine appeared completely abolished, compared with phosphorylation of the intermediate mutants, as evidenced by the absence of a specific radioactive band (Fig. 4B,  lower panel). Thus, these results unambiguously demonstrated that MurC without its six threonine phosphorylation sites has lost its ability to be phosphorylated by PknA, thus confirming the identification of all the sites of phosphorylation. An additional round of mass spectrometry analysis was also performed directly on the MurC_6T protein, which failed to identify any additional phosphate group that could eventually have arisen from a compensatory mechanism to the loss of the specific phosphorylation sites.
This type of analysis provided us the essential groundwork for mechanistic/functional studies of MurC regulation and demonstrated the efficiency of combining genetics and mass spectrometry analyses, with precise identification of phosphoacceptors, a prerequisite for a further understanding of the MurC mode of action. In particular, strains with defined mutations within the phosphorylation sites will be extremely helpful to establish the role of MurC phosphorylation-dependent regulation in corynebacterial growth and physiology.
In Vivo Phosphorylation of MurC in C. glutamicum-To assess the relevance of in vitro phosphorylation, the in vivo phosphorylation of MurC was also investigated in C. glutamicum. Therefore, to determine whether phosphorylation of MurC occurs on threonine residues in vivo, Western blot analysis was performed using specific anti-phosphothreonine, antiphosphoserine, anti-phosphotyrosine, or anti-His antibodies. To overproduce and purify phosphorylated MurC, an His tag was attached at the C terminus of the MurC protein under the control of the C. glutamicum divIVA promoter (15) in plasmid pEDiv (Table 1), and the recombinant protein was expressed in C. glutamicum. Cultures of E. coli_pTEVmurC or C. glutami-cum_pEDivmurHis overexpressing the His-tagged MurC protein were collected and lysed, and the soluble MurC was then purified to homogeneity as previously described (21). It was found that MurC purified from the C. glutamicum strain was only phosphorylated on threonine residues (Fig. 5). This was consistent with our phosphoamino acid content analyses and the in vitro identification of the MurC phosphorylation sites when phosphorylated by PknA. It clearly establishes that the murein ligase MurC is highly phosphorylated in vivo on threonines. In contrast, the anti-phosphothreonine antibodies did not react with recombinant MurC protein purified from E. coli thus confirming that phosphorylation of the C. glutamicum MurC did not occur following heterologous expression in E. coli, which thus confirmed its specific phosphorylation in C. glutamicum (Fig. 5).
MurC Phosphorylation Sites Are Critical for in Vitro and in Vivo Activity-The importance of the phosphorylation sites for MurC ligase activity was investigated in vitro by site-directed mutagenesis. The six threonine residues were replaced by alanine, and the catalytic activity of the wild-type and mutant purified MurC proteins was assayed by following the incorporation of L-[ 14 C]alanine into the PG nucleotide precursor, as described under "Experimental Procedures." As shown in Fig. 6A, the activity of mutants MurC_1T, MurC_2T, and MurC_3T, corresponding to successive mutagenesis of Thr 362 , Thr 365 , and Thr 51 residues, was reduced by ϳ30 -40% as compared with the wild-type level. Remarkably, the activity of the MurC_4T (T362A/T365A/T51A/T120A), MurC_5T (T362A/T365A/ T51A/T120A/T167A), and MurC_6T (T362A/T365A/T51A/ T120A/T167A/T133A) mutants appeared much more drastically altered and almost completely abolished in the sixthreonine mutant (Fig. 6A). Moreover, to discern whether some threonines may be more important to murC activity than others individual mutagenesis was performed. The ligase activity of each individual mutant corresponding to MurC_T51A, MurC_T120A, MurC_T133A, MurC_T167A, MurC_T362A, and MurC_T365A, was of 72%, 29%, 44%, 79%, 26%, and 95% as compared with the wild-type level. Therefore, these results confirmed that Thr 120 , Thr 133 , and Thr 362 represent critical residues for MurC ligase activity, as already observed with the MurC_1T to MurC_6T mutants proteins (Fig. 6A).
These results confirmed that the overall activity of MurC relies on the presence of these threonine residues. In fact, one could imagine that replacing these critical residues present in all three characteristic domains of MurC could have a dramatic effect on the enzymatic activity of the ligase, and especially the mutagenesis of Thr 133 , a key residue closed to the GKT motif necessary for the binding of ATP.
The activity of the MurC proteins was also tested in vivo using a functional assay based on complementation of the E. coli temperature-sensitive murC mutant strain H1119. This strain, which grows normally at 30°C but lyses when shifted at the non-permissive temperature of 42°C, was transformed with the pTrc99A plasmid, or the pTrc99murC and pTrc99murC1T, 2T, 3T, 4T, 5T, and 6T derivatives (Table 1) expressing wild-type or mutated forms of MurC. As shown in Fig. 6B, all types of transformants grew normally at the permissive temperature of 30°C, but only those expressing the wildtype protein or the 1T, 2T, and 3T mutated proteins restored growth of the E. coli mutant at the temperature of 42°C. No complementation by the plasmids expressing the 4T, 5T, and 6T mutants was observed (Fig. 6B).
A perfect correlation was thus observed between the in vitro and in vivo data, i.e. between the catalytic activity of these proteins and their ability to complement the murC temperaturesensitive E. coli mutant. Combined together, the results obtained with the different phosphorylation site mutants stress the importance of these threonine residues for MurC activity. These data not only suggest that these residues are important for a fully expressed activity of the MurC protein, due probably to their localization in critical regions of the enzyme, but also indicate, if we put forward the analysis, that the introduction of a negative charge due to the phosphorylation of threonines may have an impact on the regulation of the MurC enzyme activity.
Phosphorylation Negatively Modulates MurC Ligase Activity-Therefore, we assessed whether phosphorylation by PknA could have a direct effect on the regulation of the ligase activity of MurC. The C. glutamicum MurC protein purified from E. coli was pre-treated with either the PknA_WT kinase or the PknA_K49M kinase, yielding the phosphorylated and nonphosphorylated isoforms of MurC, respectively, which were subsequently tested for activity (Fig. 7). The PknA_K49M kinase, which is unable to autophosphorylate (Fig. 3) and therefore transphosphorylate its substrate, represented here an appropriate control, because one could imagine that the kinase itself could be responsible for the effect on MurC activity. Interestingly, we found that the ligase activity of the phosphorylated isoform of MurC was severely inhibited, representing ϳ30% of the wild-type activity, whereas no decrease of activity was observed following treatment by the PknA_K49M kinase (Fig.  7). An apparent increase of the MurC activity was observed in the latter case, suggesting that the presence of the mutant PknA protein, although inactive, stabilized in some way the MurC protein (protein-protein interactions) and reduced the slight loss of activity observed for MurC during the preincubation period. These results were consistent with the in vitro and in vivo analyses performed with the MurC mutants indicating that specific threonine residues were important in terms of enzymatic activity. Therefore, the PknA-mediated phosphorylation of MurC seems to represent a key mechanism for regulating/ controlling its activity. Phosphorylation may rather modulate MurC activity on a fine-tuned level rather than a strict on/off mechanism. However, further work is needed to understand at a molecular level whether and how phosphorylation of MurC modifies the overall structure of this protein and negatively regulates the binding of L-alanine to UDP-MurNAc to generate the peptide moiety of the PG disaccharide peptide monomer unit.
We have previously shown that C. glutamicum PknA and PknB are key players in signal transduction pathways for the regulation of the cell shape and are both essential for sustaining corynebacterial growth (18). In this study, we extended our previous findings and demonstrated for the first time that PknA is able to specifically phosphorylate the essential murein ligase Purified MurC proteins were assayed as described under "Experimental Procedures." Three independent experiments were performed, yielding similar results. The activity of the wild-type protein (here represented as 100%) was 2650 nmol/min/mg of protein, and those of the 1T, 2T, 3T, 4T, 5T, and 6T MurC mutants were 875, 800, 1050, 75, 20, and 10 nmol/min/mg of protein, respectively. B, in vivo functional complementation assays using an E. coli temperature-sensitive murC mutant strain. The pTrc99A plasmid vector and derivative plasmids expressing wild-type or mutated versions of the C. glutamicum MurC protein were transformed into the E. coli thermosensitive murC mutant strain H1119. Functional complementation was assayed by following the growth of the transformants at the permissive temperature of 30°C (left panel), or at the non-permissive temperature of 42°C (right panel). enzyme MurC, both in vitro and in vivo, and that post-translational modifications appear to be a way to regulate its ligase activity. Moreover, due to the fact that corynebacteria are extensively used for industrial production of amino acids such as glutamate, these results may open the way to manipulation of the cell wall chemical structure to increase amino acid secretion. Furthermore, these findings could be useful to extend our present understanding of the dcw (division-cell wall biosynthesis) cluster in Gram-positive bacteria toward designing of new antimicrobial drugs targeting the cell wall metabolism processes like the one involving the Mur enzymes. Moreover, if these inhibitors are capable of interfering with these regulatory processes through selective inhibition of PknA or MurC phosphorylation, it may lead to the design of novel classes of inhibitors. Indeed, the development of STPK inhibitors and the expertise in designing such inhibitors for eukaryotic STPKs represent an intense research area and could be exploited for the development of new drugs, especially toward the pathogenic corynebacteria Corynebacterium diphtheriae.