Tyrosine phosphorylation of protein kinase Wzc from Escherichia coli K12 occurs through a two-step process.

In bacteria, several proteins have been shown to autophosphorylate on tyrosine residues, but little is known on the molecular mechanism of this modification. To get more information on this matter, we have analyzed in detail the phosphorylation of a particular autokinase, protein Wzc, from Escherichia coli K12. The analysis of the hydropathic profile of this protein indicates that it is composed of two main domains: an N-terminal domain, including two transmembrane alpha-helices, and a C-terminal cytoplasmic domain. The C-terminal domain alone can undergo autophosphorylation and thus appears to harbor the protein-tyrosine kinase activity. By contrast, the N-terminal domain is not phosphorylated when incubated either alone or in the presence of the C-domain, and does not influence the extent of phosphorylation of the C-domain. The C-domain contains six different sites of phosphorylation. Among these, five are located at the C-terminal end of the molecule in the form of a tyrosine cluster (Tyr(708), Tyr(710), Tyr(711), Tyr(713), and Tyr(715)), and one site is located upstream, at Tyr(569). The Tyr(569) residue can autophosphorylate through an intramolecular process, whereas the tyrosine cluster cannot. The phosphorylation of Tyr(569) results in an increased protein kinase activity of Wzc, which can, in turn, phosphorylate the five terminal tyrosines through an intermolecular process. It is concluded that protein Wzc autophosphorylates by using a cooperative two-step mechanism that involves both intra- and interphosphorylation. This mechanism may be of biological significance in the signal transduction mediated by Wzc.

Besides their capacity to function as tyrosine kinases, most of these proteins are also involved in the production and/or transport of exopolysaccharides. Because exopolysaccharides are important virulence factors, a possible relationship between tyrosine phosphorylation and bacterial pathogenicity has been proposed. This hypothesis is supported by a number of recent data (10,11,14).
Other proteins from Gram-negative bacteria exhibit striking similarity to Ptk and Wzc, such as protein EpsB from the phytopathogen Rastonia solanacearum (15), and protein AmsA from the other phytopathogen Erwinia amylovora (16). They also are involved in the metabolism of exopolysaccharides, but their tyrosine-kinase activity has not been clearly evidenced yet.
These various proteins share several common structural features specific to bacteria, namely the Walker A and B ATPbinding motifs (17), which are not usually found in the counterpart eukaryotic kinases, and a series of tyrosine residues clustered at the C-terminal end of the molecule. For Ptk and Wzc, the Walker A motif has been shown to be effectively employed for autophosphorylation of the protein on tyrosine, suggesting that bacteria utilize for phosphorylation a novel mechanism different from that of eukaryotes (10,17). On the other hand, it has been suggested that the tyrosine cluster would be the target sequence for autophosphorylation (10,18), but no accurate characterization of the concerned residues has been made and its function remains unknown. Concerning intracellular localization, these Gram-negative bacterial proteins are all anchored in the inner-membrane and, in the particular case of ExoP, membrane topology studies have indicated that it consists of two main domains: a C-terminal domain, comprising the tyrosine cluster, which is located in the cytoplasmic fraction, and an N-terminal domain, with two transmembrane ␣-helices, which is present in the periplasm (19). Interestingly, several Gram-positive bacteria also contain homologues of Ptk and Wzc, including Streptococcus pneumoniae (20 -22), Streptococcus agalactiae (23,24), and Staphylococcus aureus (25,26). However, in these bacteria, the N-and Cterminal domains are represented in two separate polypeptides encoded by two distinct genes (20,25). Thus, the CpsC protein of S. pneumoniae is equivalent to the N-domain of Wzc, and the CpsD protein is similar to the C-domain, which is phosphorylated at a tyrosine-rich sequence (18).
In this work, we have examined in detail the process of bacterial protein autophosphorylation on tyrosine by using protein Wzc from E. coli K12 as a model. We have analyzed the region of the protein modified by phosphorylation, the effect of the rest of the molecule on this reaction, and the precise num-ber and location of the different phosphorylation sites. In addition, the mechanism of autophosphorylation has been investigated to determine the respective account of intra-and interphosphorylation in the overall reaction.

EXPERIMENTAL PROCEDURES
Bacterial Strains, Plasmids, and Growth Conditions-E. coli JM109 strain (27) was used as template for PCR amplification of the wzc gene fragments and for generating wzc gene mutants. E. coli XL1-Blue strain (28) was used to propagate plasmids in cloning experiments. For expression experiments, either E. coli XL1-Blue strain or E. coli BL21(pREP4-groESL) (29) was used. Primers and plasmids used in this study are listed below in Tables I and II, respectively. All strains were grown and maintained in Luria-Bertani or 2TY medium at 37°C. When required, media were supplemented with antibiotics at the following concentrations: ampicillin (50 g/ml), kanamycin (25 g/ml), and tetracycline (15 g/ml).
DNA Manipulation-Plasmid isolation was carried out using a Qiaprep purification kit (Qiagen). All restriction and DNA modifying enzymes were used as recommended by the manufacturer (Promega). All amplification reactions (PCR) were performed using Pfu polymerase (Promega). PCR products and plasmid DNA fragments were purified using a QiaexII kit (Qiagen). Oligonucleotides were provided by Sigma-Genosys Ltd. Transformation of E. coli cells was performed by the method of Dagert and Ehrlich (30). DNA sequencing was carried out by Genome-Express Corp. The nucleotide sequence of all synthesized and mutated genes was checked to ensure proper base replacement and error-free amplification. DNA sequences were analyzed by the DNAid computer program (31). BLAST searches (32) and sequence alignments (33,34) were performed by using our laboratory site server (www.ibcp.fr).
Construction of His 6 Tag wzc Domain Expression Plasmids-The 824-bp wzc (amino acids 1339 -2163) and 773-bp wzc (nucleotides 1339 -2112) gene fragments, with appropriate sites at both ends, encoding the C-terminal domain of Wzc, respectively, with or without the tyrosine cluster, were synthesized by PCR amplification using genomic DNA from E. coli JM109 strain as a template and primer pairs 4/1 and 4/2, respectively ( Table I). The DNA fragment synthesized was restricted by BamHI and HindIII and ligated into pQE30 vector opened with the same enzymes. The resulting plasmids were termed pQE30-41 and pQE30-42 (Table II). Similarly, DNA fragments of 1356 and 1275 bp encoding the N-terminal domain of Wzc, respectively, with or without the second Wzc-predicted transmembrane helix encoding sequence, with appropriate sites at both ends, were synthesized by PCR amplification, using primer pairs NTwzc/1275 and NTwzc/1356 (see Table I). The amplified fragments were restricted by BamHI and Acc65I enzymes and inserted into the pQE30 vector opened with the same enzymes. The resulting plasmids were termed pQE30-NTwzc1275 and pQE30-NTwzc1356 (see Table II). These plasmids introduced an Nterminal His 6 -tag on each fragment and provided high level, IPTG 1inducible, expression from the lac promoter.
Site-directed Mutagenesis-Site-directed mutagenesis was carried out by using either the Transformer site-directed mutagenesis kit from CLONTECH, based on the method developed by Deng and Nickoloff (35), or PCR amplification.
The first strategy was applied to generate single mutation on the cytoplasmic domain of Wzc, i.e. His 6 -Wzc-(Ser 447 -Ala 720 ) and His 6 -Wzc-(Ser 447 -Ala 704 ). The primers used are listed in Table I. Concerning the selection primer, elimination of the pQE30 XbaI site was obtained by creating a new AvaII site. This procedure was applied directly to the pQE30-41 vector to generate substitution, either Y569F or K540M, in His 6 -Wzc-(Ser 447 -Ala 720 ) (see Table II), and to the pQE30-42 vector to generate substitutions Y467F, Y491F, Y569F, Y636F, Y668F, or K540M in His 6 -Wzc-(Ser 447 -Ala 704 ) (see Table II).
The second strategy was used to create substitutions of tyrosine for phenylalanine in the C-terminal tyrosine cluster of Wzc. PCR amplification, using pQE30-41-Y569F as a template (see Table II), was first performed with the primer pair 4/L6 (see Table I). The amplified DNA fragment, with appropriate sites at both ends, was restricted by BamHI and HindIII and ligated with pQE30 vector previously opened with the same enzymes. The resulting plasmid was termed pQE30-41Y569F-L6 (see Table II). By using the different primers F1L4, F2L3, F3L2, F4L1, and F5 (see Table I) in combination with primer 4, PCR amplification using pQE30-41Y569F-L6 as a template was carried out. The five different DNA fragments synthesized were restricted by BamHI and HindIII and ligated into pQE30 vector opened with the same enzymes. The resulting plasmids were termed pQE30-41Y569-F1L4, pQE30-41Y569-F2L3, pQE30-41Y569-F3L2, pQE30-41Y569-F4L1, and pQE30-41Y569-F5 (see Table II). To generate a plasmid overproducing His 6 -Wzc-(Ser 447 -Lys 720 ) with only tyrosine 705 conserved, a PCR amplification was carried out using pQE30-41Y569F as a template and the primer pair 4/L5 (see Table I). The DNA synthesized was restricted by BamHI and HindIII and ligated into pQE30 vector opened with the same enzymes. The resulting plasmid was termed pQE30-41Y569-L5.
Construction of Wild and Mutated GST-Wzc Cytoplasmic Domain Expression Plasmids-To construct plasmids expressing the cytoplasmic domain of Wzc fused with GST, PCR amplification was carried out using, on the one hand, pQE30-41 or pQE30-41K (see Table II) as templates with the primer pair 4/1bis and, on the other hand, pQE30-42 or pQE30-42K (see Table II) as templates with the primer pair 4/2bis. 824-bp wzc (amino acids 1339 -2163) and 764-bp wzc (nucleotides 1339 -2103) gene fragments, with appropriate sites at both ends, were obtained. The amplified fragments were restricted by BamHI and EcoRI enzymes, then ligated into pGEX-KT vector, previously opened with the same enzymes, to yield plasmids pGEX-KT-41, pGEX-KT-42, pGEX-KT-41K, and pGEX-KT-42K (Table I).
Overproduction and Purification of His 6 Tag Fusion Wzc Cytoplasmic Domain-E. coli XL1-Blue cells were transformed with pQE30 vector derivatives expressing wild or mutated Wzc cytoplasmic domain, His 6 -Wzc-(Ser 447 -Lys 720 ) and His 6 -Wzc-(Ser 447 -Ala 704 ) (see Table II). The purification procedure was the same for each Wzc cytoplasmic fragment, wild or mutated. An overnight cell culture was used to inoculate 100 ml of 2TY medium supplemented with ampicillin and tetracycline, and was incubated at 37°C under shaking until the A 600 reached 0.5. IPTG was then added at a final concentration of 0.5 mM, and growth was continued for 3 h at 37°C under shaking. Cells were harvested by centrifugation at 3000 ϫ g for 10 min, washed in 10 ml of buffer A (50 mM sodium phosphate, pH 8.0, 300 mM NaCl, 10 mM imidazole, 10% glycerol, 10 mM ␤-mercaptoethanol), and centrifuged again in the same conditions. The cell pellet was resuspended in buffer A containing deoxyribonuclease I (DNase I) and ribonuclease A (RNase A) at a final concentration of 5 g/ml each. Cells were disrupted in a French pressure cell at 16,000 lb/in 2 (p.s.i.). The resulting suspension was centrifuged at 4°C for 30 min at 30,000 ϫ g. The supernatant was added to Ni 2ϩ -NTA-agarose matrix and batch binding was allowed to proceed for 1 h at 4°C under gentle shaking. The lysate/ Ni 2ϩ -NTA-agarose mixture was loaded on a column and was first washed with buffer A, then with 20 mM imidazole in the same buffer until A 280 reached a basic line. Protein elution was carried out with buffer B (50 mM sodium phosphate, pH 8.0, 300 mM NaCl, 100 mM imidazole, 10% glycerol, 10 mM ␤-mercaptoethanol), and eluted fractions were analyzed by SDS-PAGE (36). Fractions containing purified His 6 -tagged proteins were pooled and dialyzed 2ϫ2 h at 4°C against a large volume of buffer C (50 mM sodium phosphate, pH 7.4, 150 mM NaCl, 10% glycerol, 5 mM MgCl 2 , 5 mM dithiothreitol) and stored at Ϫ20°C in the same buffer adjusted to 20% glycerol.
Overproduction and Purification of His 6 Tag Fusion Wzc N-terminal Fragment-E. coli BL21 (pREP4-groESL) cells were transformed with pQE30 vector derivatives pQE30-NTwzc1275 and pQE30-NTwzc1356 (see Table II), expressing Wzc N-terminal fragment His 6 -Wzc-(Thr 2 -Gly 425 ) and His 6 -Wzc-(Thr 2 -Gly 452 ), respectively. Cells from these strains were used to inoculate 100 ml of BL medium supplemented with ampicillin and kanamycin and were incubated at 37°C under shaking until A 600 reached 0.7. IPTG was then added at a final concentration of 0.5 mM, and growth was continued for 2 h at 30°C under shaking. Cells were harvested by centrifugation at 3000 ϫ g for 10 min, washed in 10 ml of buffer A, and centrifuged again in the same conditions. The cell pellet was resuspended in the same buffer A containing DNase I and RNase A at a final concentration of 5 g/ml each. Cells were then disrupted in a French pressure cell at 16,000 p.s.i. The resulting suspension was supplemented with Triton X-100 at a final concentration of 1% and centrifuged at 4°C for 30 min at 30,000 ϫ g. The supernatant was added to Ni 2ϩ -NTA-agarose, and batch binding was allowed to proceed for 1 h at 4°C under gentle shaking. The lysate/Ni 2ϩ -NTAagarose mixture was loaded on a column and was first washed with buffer A containing 1% Triton X-100, then with 50 ml of 20 mM imidazole in the same buffer. Protein elution was carried out five times with 1 ml of buffer B containing 0.1% Triton X-100 and 200 mM imidazole. Eluted fractions were collected and analyzed by SDS-PAGE (36). After dialysis against buffer C, the purest fraction was tested for in vitro phosphorylation assay.
Overproduction and Purification of GST-Wzc Cytoplasmic Domain-E. coli XL1-Blue cells were transformed with pGEX-KT vector derivatives producing either Wzc-(Ser 447 -Lys 720 ) and Wzc-(S 447 -R 701 ) polypeptides or substituted K540MWzc-(Ser 447 -Lys 720 ) and K540MWzc-(S 447 -R 701 ) polypeptides. The purification procedure was the same for all four polypeptides. Cells from these strains were used to inoculate 100 ml of BL medium supplemented with ampicillin and tetracycline and were incubated at 37°C under shaking until A 600 reached 0.5. IPTG was then added at a final concentration of 0.5 mM, and growth was continued for 3 h at 37°C under shaking. Cells were harvested by centrifugation at 3000 ϫ g for 10 min, washed in 10 ml of buffer D (10 mM sodium phosphate, pH 7.4, 150 mM NaCl, 1 mM EDTA, 10% glycerol), and centrifuged again in the same conditions. The cell pellet was resuspended in buffer D containing DNase I and RNase A at a final concentration of 5 g/ml each. Cells were disrupted in a French pressure cell at 16,000 p.s.i. The resulting suspension was centrifuged at 4°C for 30 min at 30,000 ϫ g. The supernatant was incubated with glutathione-Sepharose 4B matrix (Amersham Biosciences, Inc.), suitable for purification of glutathione S-transferase (GST) fusion proteins, for 30 min at 4°C under gentle shaking. The protein-resin complex was packed into a column for washing and elution steps. The column was washed with buffer D. Protein elution was carried out with buffer D containing 10 mM glutathione. After loading 1 ml of elution buffer, the column was incubated for 15 min at 4°C. This process was repeated three times. Eluted fractions were collected, analyzed by SDS-PAGE (36), and stored at Ϫ20°C in the same buffer adjusted to 20% glycerol.
Phosphorylation Assay-In vitro phosphorylation of about 2 g of the different purified wild or mutant His 6 -tagged and/or 2 g of GST-fused Wzc fragments was carried out for 10 min at 37°C in a reaction mixture (20 l) containing 25 mM Tris-HCl, pH 7.0, 1 mM dithiothreitol, 5 mM MgCl 2 , 1 mM EDTA, and 10 M ATP with 200 Ci/ml [␥-32 P]ATP. The reaction was stopped by addition of an equal volume of 2ϫ sample buffer, and the mixture was heated at 100°C for 5 min. One-dimensional gel electrophoresis was performed as described by Laemmli (36). After electrophoresis, gels were soaked in 16% trichloroacetic acid for 10 min at 90°C. They were stained with Coomassie Blue, and radioactive proteins were visualized by autoradiography using direct-exposure films.

Phosphorylation of Wzc Occurs Specifically in the C-terminal
Domain-Wzc is a 79344-Da protein of 720 amino acids bound to the inner membrane of E. coli K12 (9, 10). To get more information on its topography within the cell, we analyzed the hydropathic profile of its amino acid sequence according to the predictive method of Kyte and Doolittle (37). Two main domains in the protein (Fig. 1) could thus be detected: on the one hand, in the N-terminal part, a domain containing two transmembrane ␣-helices, termed TM1 and TM2, which include amino acids Trp 32 to Ala 52 and Leu 426 to Leu 445 , respectively, and appear to flank the periplasmic region of the protein; on the other hand, a C-terminal domain extending from Ser 447 to Lys 720 , which seems to correspond to the cytoplasmic region of the protein. The latter domain harbors in particular a Walker A ATP-binding motif, from Gly 533 to Val 543 , and a Walker B motif, from Asp 637 to Asp 642 (9). Such predicted organization of Wzc into two domains is in agreement with the previous data obtained with protein ExoP from S. meliloti showing that this Wzc homologue contains two transmembrane ␣-helices and a large C-terminal cytoplasmic domain (19). Concerning the distribution of the 18 tyrosine residues contained in total in Wzc, 7 of them are present in the N-domain (from Met 1 to Arg 446 ), and the other 11 are in the C-domain, namely in the form of a cluster of 6 residues, at the C-terminal end, between Ala 704 and Lys 720 .
From these observations, it seemed interesting to determine in what region(s) of the molecule the phosphorylation reaction would occur when Wzc undergoes autophosphorylation. For this, the C-terminal fragment of the protein, from Ser 447 to Lys 720 (approximate molecular mass of 31 kDa), was overproduced by using PQE30 vectors. It was expressed with a His 6tag, purified to homogeneity, incubated with [␥-32 P]ATP, and analyzed by SDS-PAGE and autoradiography. Fig. 2 (lane 1) shows that this His 6 -Wzc-(Ser 447 -Lys 720 ) fragment was significantly labeled in vitro, which indicates that it contains both an intrinsic protein kinase activity and specific phosphorylation sites.
A similar experiment was performed in the case of the Nterminal domain. Two types of pQE30 derivative vectors were constructed to express either the polypeptide fragment from Thr 2 to Gly 425 containing the N-domain with the transmembrane ␣-helix TM1, or the fragment from Thr 2 to Gly 452 containing both TM1 and TM2 ␣-helices. The idea was to check the possible effect of the second ␣-helix on the phosphorylation of the N-domain. These two fragments, obtained with a His 6 -tag, respectively, His 6 -Wzc-(Thr 2 -Gly 425 ) and His 6 -Wzc-(Thr 2 -Gly 452 ), were purified and incubated separately in the presence of radioactive ATP, then analyzed by SDS-PAGE and autora- diography. No labeling was observed, in either case, indicating that the N-domain of Wzc, bearing one or two ␣-helices, is unable to autophosphorylate (Fig. 2, lanes 2 and 3).
In another series of assays, the His 6 -Wzc-(Thr 2 -Gly 425 ) and His 6 -Wzc-(Thr 2 -Gly 452 ) fragments were incubated each in the presence of the C-fragment His 6 -Wzc-(Ser 447 -Lys 720 ) and analyzed in the same conditions as above. No phosphorylation of either N-fragment occurred, because no radioactivity was found in the corresponding zone of molecular mass, around 49 kDa (Fig. 2, lanes 4 and 5). This indicated that, not only is the N-domain of Wzc unable to autophosphorylate but it also cannot serve as a substrate in the phosphorylation reaction catalyzed by the C-domain. Interestingly, the analysis of the radioactive band intensity in the 30-kDa region of the gel showed that it was similar in the presence or absence of the N-domain (Fig. 2, lanes 4 and 5 versus lane 1). It therefore seems that the N-domain of Wzc has no effect on the extent of phosphorylation of the C-domain.
A Phosphorylation Site Is Located Outside the C-terminal Tyrosine Cluster of Wzc-The comparative analysis of various proteins known to autophosphorylate on tyrosine shows that they share a number of common structural features, namely a series of tyrosine residues, termed "tyrosine cluster," located in their C-terminal end (Fig. 3). From observations based on substitution or deletion experiments, a few reports have previously indicated that this cluster would be the target of the phosphorylation reaction in Gram-negative (10) as well as in Grampositive (18) bacteria, even though the nature and number of phosphorylation sites were not characterized. We re-examined this concept by measuring phosphorylation in a C-terminal fragment of Wzc deleted from its tyrosine cluster. By using a pQE30 derivative vector, a His-tagged polypeptide missing the 6 terminal tyrosine residues, Tyr 705 to Tyr 715 (Fig. 1), was synthesized and purified to homogeneity. This His 6 -Wzc-(Ser 447 -Ala 704 ) construct was incubated with [␥-32 P]ATP and analyzed by SDS-PAGE and autoradiography as already described. As shown in Fig. 4 (lane 1), a significant amount of radioactivity was then incorporated in a 30-kDa molecule corresponding to His 6 -Wzc-(Ser 447 -Ala 704 ), thus indicating that some phosphorylation occurred in the C-domain of Wzc outside the terminal tyrosine cluster. Upstream of the 6-residue tyrosine cluster, the C-domain of Wzc contains 5 different tyrosine residues, respectively at positions 467, 491, 569, 636, and 668 (Fig. 1). The next question was therefore to determine which tyrosine(s) among these five could be phosphorylated and account for the radioactive labeling of His 6 -Wzc-(Ser 447 -Ala 704 ). To answer, site-directed mutagenesis experiments were carried out by using the pQE30-42 vector (Table II) as a template. Each of the 5 relevant tyrosines was substituted individually for phenylalanine in His 6 -Wzc-(Ser 447 -Ala 704 ) to generate five different vectors: pQE30-42Y467F, pQE30-42Y491F, pQE30-42Y569F, pQE30-42Y636F, and pQE30-42Y668F (Table II). Each mutant polypeptide was overproduced, purified to homogeneity, incubated with radioactive ATP, and analyzed by gel electrophoresis and autoradiography. In all cases, except one, the mutant fragment appeared to keep incorporating radioactivity to the same extent as the non-mutant control, showing that the 4 corresponding tyrosine residues (467, 491, 636, and 668) are not involved in the phosphorylation reaction (Fig. 4). The only exception concerned the His 6 -Wzc-(Ser 447 -Ala 704 ) polypeptide substituted for phenylalanine at position 569 (Fig.  4, lane 4). This result demonstrated, for the first time, that a tyrosine residue located outside the C-terminal tyrosine cluster in a phosphorylatable protein represents an active phosphorylation site for endogenous protein-tyrosine kinase activity.
The C-terminal Tyrosine Cluster of Wzc Harbors Five Phosphorylation Sites-Experiments were undertaken to determine which tyrosines are phosphorylated in the C-terminal tyrosine cluster of Wzc. Because tyrosine 569 was previously identified as a phosphorylation site, site-directed mutagenesis was carried out with pQE30-41 (Table II) as a template to generate the pQE30-41Y569F vector expressing His 6 -Wzc-(Ser 447 -Lys 720 ) with the substitution Y569F. This procedure was chosen with the aim of eliminating the phosphorylation background due to Tyr 569 so as to allow specific analysis of the phosphorylation sites of the tyrosine cluster. First, we constructed a vector encoding a phosphorylation switched-off form of Wzc-(Ser 447 -Lys 720 ), i.e. with all six tyrosines Tyr 705 , Tyr 708 , Tyr 710 , Tyr 711 , Tyr 713 , and Tyr 715 substituted to phenylalanine in addition to the Y569F substitution. This vector was obtained by PCR mutagenesis and termed pQE30-41Y569F-L6 (Table II). Then, site-directed mutagenesis was performed to reverse, one by one, the Y 3 F substitutions in the tyrosine cluster, and to restore in each case one of the six different tyrosine residues. By doing so, six different vectors expressing each His 6 -Wzc-(Ser 447 -Lys 720 )-Y569F with only one phosphorylatable tyrosine remaining in the C-terminal cluster were obtained (Table II). The corresponding mutant forms of His 6 -Wzc-(Ser 447 -Lys 720 ) were overproduced, purified to homogeneity, and assayed individually for radioactive ATP incorporation. The non-mutated form His 6 -Wzc-(Ser 447 -Lys 720 ) containing Tyr 569 and the 6 tyrosine residues of the cluster, the mutated form His 6 -Wzc-(Ser 447 -Lys 720 )-Y569F containing the 6 terminal tyrosines, and the substituted form His 6 -Wzc-(Ser 447 -Lys 720 )-Y569F-L6 with no tyrosine residue were analyzed in parallel.
The corresponding autoradiograms are presented in Fig. 5  (lanes 1, 2, and 3, respectively). Extensive labeling was observed when the tyrosine cluster was present (lane 2), especially when Tyr 569 was not mutated (lane 1). By contrast, no phosphorylation occurred in His 6 -Wzc-(Ser 447 -Lys 720 )-Y569-L6 (lane 3), which provided evidence that, besides Tyr 569 , the phosphorylation sites for the protein-tyrosine kinase activity of Wzc are all located in the C-terminal tyrosine cluster. The analysis of the six different mutant peptides containing each only one terminal restored tyrosine (Fig. 5, lanes 4 -9) showed, in addition, that they all were able to undergo phosphorylation except the mutant containing tyrosine 705 (lane 4). This finding indicated that, in the C-terminal cluster of Wzc, the 5 tyrosine residues at positions 708, 710, 711, 713, and 715 are phosphorylation sites, whereas tyrosine at position 705 is not.
Wzc Is Subject to Both Intra-and Interphosphorylation-Because the phosphorylation sites of Wzc appeared to be located in two distinct parts of the Wzc molecule, one at Tyr 569 and the others in the 5 neighboring tyrosines of the C-terminal end, an attempt was made to assess the existence of a functional relationship between these sites during the phosphorylation process.
This possibility was supported primarily by the fact that the extent of phosphorylation of a complete C-terminal fragment His 6 -Wzc-(Ser 447 -Lys 720 ) (Fig. 2, lane 1) is about 700-fold higher than the phosphorylation signal produced by a fragment like His 6 -Wzc-(Ser 447 -Ala 704 ) containing Tyr 569 but missing the tyrosine cluster (Fig. 4, lane 1). The average amount of radioactivity in the relevant bands, as determined by fluid scintillation counting in three different experiments, was indeed, respectively, 380,740 and 560 cpm. Similarly, when comparing directly the extent of phosphorylation of the same complete C-terminal fragment His 6 -Wzc-(Ser 447 -Lys 720 ) (Fig. 2, lane 1) with that of fragment His 6 -Wzc-(Ser 447 -Lys 720 )-Y569F containing the tyrosine cluster but missing Tyr 569 (Fig. 5, lane 2), a 45-fold difference was emerging, as confirmed by the average scintillation counting of the corresponding bands: 380,740 cpm in the former fragment and 8580 cpm in the latter. Together, these data suggested the existence of a cooperative effect between the phosphorylation sites of Wzc, the tyrosine at position 569 and the 5 tyrosines of the cluster, leading to a substantial increase of the overall degree of phosphorylation of the Cterminal fragment. More precisely, at this point, it seemed that the presence of Tyr 569 would enhance the phosphorylation of the tyrosine cluster.
To check this hypothesis, we investigated in more detail the process of phosphorylation. First, we produced a class of Cfragments unable to sustain phosphorylation. For this, the lysine residue located at position 540 in the Walker A motif essential for ATP binding (9, 10) was substituted to methionine so as to yield Wzc-(Ser 447 -Lys 720 )-K540M or Wzc-(Ser 447 -Ala 704 )-K540M, depending on the plasmid used for site-directed mutagenesis (Table II). In addition, to mark out these peptides, the corresponding DNA fragments were cloned in the pGEX-KT vector (Table II) to obtain GST-fused peptides, GST-Wzc-(Ser 447 -Lys 720 )-K540M and GST-Wzc-(Ser 447 -Ala 704 )-K540M. A similar population of peptides, also fused to GST but containing a non-mutated K 540 , was also prepared for control assays: GST-Wzc-(Ser 447 -Lys 720 ) and GST-Wzc-(Ser 447 -Ala 704 ). The various GST peptides could thus be distinguished from His 6 -tagged peptides on the basis of a larger molecular mass. Moreover, in view of testing the possible effect of the nature of the tag on the phosphorylation reaction, a symmetrical class of  peptides bearing a mutated Lys 540 and fused with 6 histidine residues was also constructed by using pQE30-41 and pQE30-42 as templates (Table II) In control assays, the fusion peptide GST-Wzc-(Ser 447 -Lys 720 ) was shown to actively autophosphorylate (Fig. 6, lane 1) as well as the homologous peptide GST-Wzc-(Ser 447 -Ala 704 ) missing the tyrosine cluster, although to a lesser extent (lane 2). When the same two peptides were mutated at Lys 540 , no phosphorylation occurred, which confirmed the crucial role played by this lysine residue in the kinase activity of Wzc ( lanes  3 and 4).
When inactive GST-Wzc-(Ser 447 -Ala 704 )-K540M was incubated with either active His 6 -Wzc-(Ser 447 -Lys 720 ) or active His 6 -Wzc-(Ser 447 -Ala 704 ), no phosphorylation was detected at the level of the GST-fused peptide (lanes 5 and 6), whereas the active 6 His-tagged peptides were phosphorylated. The same result was obtained when, conversely, inactive His 6 -Wzc-(Ser 447 -Ala 704 )-K540M was incubated with either active GST-Wzc-(Ser 447 -Lys 720 ) or active GST-Wzc-(Ser 447 -Ala 704 ) (data not shown). This means that, in both situations, whatever the nature of the tag fused, no transfer of the radioactive moiety present on a tagged C-fragment, with or without tyrosine cluster, to the phosphorylation site Tyr 569 of another wzc-(Ser 447 -Lys 720 ) fragment can take place. It therefore seems that when Tyr 569 becomes phosphorylated, the reaction is strictly due to an intramolecular phosphorylation, with no possible intermolecular transfer of phosphoryl groups to Tyr 569 .
In contrast, when inactive GST-Wzc-(Ser 447 -Lys 720 )-K540M was incubated with either active His 6 -Wzc-(Ser 447 -Lys 720 ) or active His 6 -Wzc-(Ser 447 -Ala 704 ), the GST-fused peptide was effectively labeled (lanes 7 and 8), like the 6 His-tagged peptides. An identical result was obtained when inactive His 6 -Wzc-(Ser 447 -Lys 720 )-K540M was incubated with either active GST-Wzc-(Ser 447 -Lys 720 ) or active GST-Wzc-(Ser 447 -Ala 704 ) (data not shown). It can be concluded that, in these conditions, the phosphorylated residues belong to the tyrosine cluster and accept radioactivity through an intermolecular transfer of phosphoryl groups arising from ATP hydrolysis and catalyzed by another copy of the C-fragment of Wzc. In other words, tyrosine 569 would be specifically phosphorylated in an intramolecular phosphorylating process, whereas the tyrosine cluster would be phosphorylated in an intermolecular reaction between two distinct molecules of Wzc.
Finally, to test further the possible stimulatory effect of Tyr 569 on the phosphorylation of the tyrosine cluster, the GST-Wzc-(Ser 447 -Lys 720 )-K540M peptide, inefficient in phosphorylation, was incubated with either His 6 -Wzc-(Ser 447 -Lys 720 )-Y569F or His 6 -Wzc-(Ser 447 -Ala 704 )-Y569F, which both lack a phosphorylatable native Tyr 569 residue. The phosphorylation of GST-Wzc-(Ser 447 -Lys 720 )-K540M was drastically diminished (lanes 9 and 10) compared with the assays with peptides containing an active Tyr 569 , suggesting that Wzc needs to autophosphorylate at Tyr 569 before acquiring its maximal interphosphorylating activity (Fig. 7). DISCUSSION Protein Wzc is an inner-membrane protein from E. coli K12 able to undergo autophosphorylation on tyrosine residues and required for the production of a particular exopolysaccharide, colanic acid (14). In this study, we determined, for the first time, the precise number and location of the different phosphorylation sites within the protein molecule. Then we newly showed that phosphorylation of Wzc proceeds through a cooperative two-step mechanism that involves both intramolecular and intermolecular phosphorylation.
Based on the hydropathic profile of its amino acid sequence, it was predicted that Wzc contained two separable domains: an N-terminal domain, with two transmembrane ␣-helices, and a C-terminal cytoplasmic domain, harboring Walker A and B ATP-binding motifs. This type of molecule topology was previously demonstrated in the case of protein ExoP, a Wzc homologue in S. meliloti, and the C-terminal domain alone was shown to autophosphorylate in vitro (12). Our data confirmed this finding, because we observed that the C-terminal fragment of Wzc, Wzc-(Ser 447 -Lys 720 ), exhibits an intrinsic protein-tyrosine kinase activity. Discrepant results were, however, described for another class of E. coli strain, K30, and for the Gram-positive bacterium S. pneumoniae. Indeed, it was reported (10) that a C-terminal fragment of Wzc from E. coli K30 is unable to autophosphorylate per se but needs the presence of the N-terminal part of the protein to be phosphorylated. Similarly, it was found that in S. pneumoniae the phosphorylation at the tyrosine of protein CpsD requires the presence of another The phosphorylation of Wzc proceeds through a cooperative two-step mechanism. First, Tyr 569 is phosphorylated in an intramolecular reaction, involving the Walker A motif, which generates a substantial increase of the kinase activity of Wzc. Then the 5 residues of the tyrosine cluster are phosphorylated in an intermolecular Tyr 569 -dependent reaction. Any molecule of Wzc can thus be successively intraphosphorylated then interphosphorylated. tivity promotes extensive phosphorylation of the tyrosine cluster (interphosphorylation) in the C-terminal region; then the phosphorylation of Wzc affects, directly or indirectly, the production of the exopolysaccharide, colanic acid. Further experiments are now needed to check the plausibility of this hypothesis and to decipher the molecular mechanism that interconnects protein tyrosine phosphorylation and bacterial pathogenicity, via the production of capsular and/or extracellular polysaccharides.