C-kinase phosphorylates the epidermal growth factor receptor and reduces its epidermal growth factor-stimulated tyrosine protein kinase activity.

The Ca2+- and phospholipid-dependent protein kinase (C-kinase) binds tightly in the presence of Ca2+ to purified membranes of A431 human epidermoid carcinoma cells. The major membrane substrate for C-kinase is the epidermal growth factor (EGF) receptor. Phosphorylation of the EGF receptor is Ca2+-dependent and occurs at threonine and serine residues. After tryptic digestion of the receptor, three major phosphothreonine-containing peptides were identified. These are identical with three new phosphopeptides present in the EGF receptor isolated from A431 cells treated with either of the tumor promoters 12-O-tetradecanoylphorbol 13-acetate or teleocidin. C-kinase catalyzes phosphorylation at these same sites in purified EGF receptor protein. These results indicate that, in A431 cells exposed to tumor promoters, C-kinase catalyzes phosphorylation of a significant population of EGF receptor molecules. This phosphorylation of EGF receptors results in decreased self-phosphorylation of the EGF receptor at tyrosine residues both in vivo and in vitro and in decreased EGF-stimulated tyrosine kinase activity in vivo.

saturated diacylglycerol to increase the affinity of C-kinase for Ca2+ and phosphatidylserine in activating the enzyme ( 5 ) . The phorbol diester receptor and C-kinase from rat brain extensively co-purify, suggesting that the two activities reside in the same or closely associated molecules (6,7). C-kinase can also be irreversibly activated by a specific Ca2+-dependent protease (8).
Addition of TPA to parietal yolk sac cells increases the fraction of C-kinase associated with the plasma membrane (9), suggesting that initial substrates for phosphorylation may be located there. Because tumor promoters alter epidermal growth factor binding (10-13) and because tumor promoters and growth factors produce several similar biological effects, the EGF receptor has been examined as a substrate for Ckinase. We have found that C-kinase phosphorylates membrane-associated as well as purified EGF receptors at threonine residues at sites identical with those phosphorylated when intact cells are exposed to tumor promoters. The EGF receptor protein itself possesses protein kinase activity with specificity for tyrosine residues in substrate proteins (14). This tyrosine protein kinase activity is reduced after phosphorylation of the EGF receptor by C-kinase.
For some experiments C-kinase was further purified on an hydroxyapatite column which was eluted with a linear phosphate gradient (50-500 mM). The enzyme eluting between 120 and 250 mM was brought to 1 M NaCl and loaded on a phenyl-Sepharose column. The column was then developed with a decreasing sodium chloride gradient from 1 M to 0 M. The enzyme was eluted between 0.2 and 0 M NaCl. The specific activity of the purified enzyme was not determined because stabilizing soybean trypsin inhibitor protein was present.
For purification of the EGF receptor, monoclonal anti-EGF receptor antibody 528 was coupled to Sepharose 4B with 98% yield by the cyanogen bromide procedure (21) to provide an affinity matrix containing 10 PM antibody. A4318 cells were homogenized in buffer containing 20 mM HEPES-NaOH, pH 7.2, 2 mM EDTA, 1 mM dithiothreitol, 3 mM benzamidine HCl, 3.6% aprotinin, 14% glycerol, and 1% Triton X-100, then diluted with 10 vol of 20 mM HEPES, pH 7.2, containing 160 mM NaCl, 0.15% 2-mercaptoethanol, and 10% glycerol; the homogenate was clarified by centrifugation; and the supernatant was applied directly to the affinity matrix. After extensive washing with buffers of increasing ionic strength and 1 M urea, the EGF receptor was eluted with either 1 M acetic acid or with 6 M urea and was dialyzed against 20 mM HEPES-NaOH, pH 7.2/0.25% Triton X-100. The purified receptor preparation contained a single species of approximately 170,000 M, on SDS-polyacrylamide gel electrophoresis and a single NH2-terminal leucine by manual Edman degradation.
Immunoprecipitation of EGFReceptors-EGF receptor was isolated using TL5 monoclonal antibody (19), from 32P-labeled cell extracts, made with RIPA buffer (25) containing 2 mM EDTA. Lysate corresponding to -IO6 A431s cells was incubated with 12 pmol of TL5 for 30 min, followed by 5 p1 of goat antiserum to mouse immunoglobulin for 30 min then 1.5 mg of Pansorbin for 60 min, all at 0 "C. These immunoprecipitates were washed as described (25). I n vitro phosphorylated EGF receptor was isolated with 1 p M 528 monoclonal antibody (18) by incubation for 30 min at 0 "C followed by addition of an aliquot of Pansorbin calculated to bind a 10-fold excess of IgG. Immunoprecipitates were sequentially washed with 1 M sucrose/ RIPA/O.5 M NaCl and H20. Samples were boiled in Laemmli sample buffer and phosphoproteins were separated by electrophoresis on 10% SDS-polyacrylamide gels (10% acrylamide, 0.13% N,N'-methylenebisacrylamide). Phosphorylated EGF receptor was located by autoradiography. Molecular weight standards were run in a parallel lane and identified by Coomassie blue staining.
Two-dimensional Peptide Mapping-Phosphorylated EGF receptor was located by autoradiography of dried gels and eluted, oxidized, and digested with trypsin as described (26). Phosphopeptides were separated on cellulose thin layer plates by electrophoresis at pH 8.9 (anode at left) and ascending chromatography (27). Phosphoamino acids present in the EGF receptor or in individual phosphopeptides were released by 1 h of hydrolysis in 6 N HCl at 110 "C and separated by electrophoresis at pH 1.9 followed by electrophoresis at pH 3.5 (28).

RESULTS
When A431 cell plasma membranes which contain a large amount of EGF receptor (29) were incubated with rat brain C-kinase, [-p3'P]ATP, CaC12, and MgC12, two major membrane phosphoproteins were labeled (Fig. lA, lunes a-c). Phosphorylation was dependent on both C-kinase and Ca2+. The 170,000 M , phosphoprotein was identified as the EGF receptor based on co-migration with purified EGF receptor and on specific interaction with monoclonal anti-EGF receptor antibody (lune e ) . The 70,000 M, phosphoprotein was derived from the membranes because it was not phosphorylated when A4318 membranes or C-kinase were incubated individually with Ca2+ ( Fig. 1A and data not shown). The 70,000 M, protein, however, was not derived from the EGF receptor as determined by comparing partial proteolytic digests of 32Plabeled 170,000 and 70,000 Mr proteins (Fig. 1B). The 83,000 M , phosphoprotein was labeled when the C-kinase preparation was incubated alone and corresponds to self-phosphorylated enzyme (20). EGF receptor phosphorylated by C-kinase contained phosphothreonine, some phosphoserine, and small amounts of phosphotyrosine, whereas EGF receptor/kinase self-phosphorylated in the presence of EGF and Mn2+ contained only phosphotyrosine (Fig. 1C). The KO for Ca'+ activation of C-kinase-catalyzed phosphorylation of EGF receptor protein was less than 20 FM, identical with the K,, for Ca2+ using histone H1 as substrate. The reaction was strongly inhibited by 0.5 mM chlorpromazine but was not affected by 1 PM calmodulin (data not shown).
Two-dimensional maps of tryptic digests of 32P-labeled EGF receptor revealed that C-kinase catalyzed phosphorylation of 3 peptides ( X , Y , 2 in Fig. 2F). All contained phosphothreonine. The sites of serine phosphorylation were not identified as discrete peptides and presumably the phosphoserine is distributed among many minor sites. The additional peptide phosphorylated at tyrosine (spot 1 in Fig. 2F) corresponded to the major site of self-phosphorylation in EGF receptor ( Fig.  2E and Refs. 31 and 32). To confirm the specificity of Ckinase-catalyzed phosphorylation of EGF receptor, the latter was purified to homogeneity by affinity chrqmatography on Sepharose immobilized anti-EGF receptor IgG. Purified EGF receptor retained lZ5I-EGF and lZ5I-anti-receptor IgG binding activity but not EGF-dependent protein kinase activity. Ckinase catalyzed phosphorylation of this purified EGF receptor at phosphothreonine and phosphoserine and the same threonine-containing phosphopeptides were observed (Fig.  2H). The phosphotyrosine-containing peptide was not detected (Fig. 2H). An additional phosphorylation site was noted which was apparently less accessible in the native protein ( Fig. 2H uersus 2F).
Similar experiments were performed after the C-kinase had been submitted to two additional purification steps. After the final step, the enzyme was able to phosphorylate the membrane-bound EGF receptor in a calcium-dependent manner. In addition this phosphorylation was strongly inhibited by a specific inhibitor of C-kinase prepared from placenta termed phorbol ester-binding inhibitory factor which was kindly provided by I. Vilgrain.' The presence of this inhibitor during the reaction blocked phosphorylation of the EGF receptor on threonine residues, but allowed autophosphorylation of the EGF receptor on tyrosine residues (data not shown). Although a requirement for phospholipid and diacylglycerol was not observed with A431 membranes as s&strate, phosphorylation of purified EGF receptor protein showed a strong requirement for phospholipid and diacylglycerol (data not shown).
Addition of TPA or teleocidin to A431 cells labeled to . . " EGF receptor from control cells or cells treated with the inactive parent phorbol (Fig. 2, A, B, and C) or cells treated with carrier dimethyl sulfoxide (data not shown). These tryptic peptides, which together contained about 10% of the total radioactivity as phosphothreonine, were well resolved from the phosphotyrosine-containing tryptic peptide observed in EGF-treated cells (Fig. 2 E ) and from several other phosphopeptides whose pattern was unaffected by treatment of cells with tumor promoters. New peptides containing phosphoserine were not detected. The new sites phosphorylated in the EGF receptor isolated from tumor promoter-treated cells were identical with the sites phosphorylated in the EGF receptor by C-kinase in vitro (Fig. 2G). These results support the idea that tumor promoters activate C-kinase which catalyzes phosphorylation of EGF receptor a t 3 specific sites (Fig. 2D). The increase in phosphorylation is rapid, being detected within 3 min of tumor promoter addition and persists for a t least 1 h. When EGF is added to A431 cells, increased phosphoryla-tion of the EGF receptor occurs because of increased labeling of several serine and threonine phosphopeptides as well as appearance of a single phosphotyrosine peptide (31, 32). Under these conditions, EGF did not increase labeling of the specific phosphopeptides labeled by C-kinase (Fig. 2E), while tumor promoters did not increase labeling of the EGF-specific phosphotyrosine-containing peptide (Fig. 2, B and C).
To characterize C-kinase-catalyzed phosphorylation of EGF receptors further, A4318 membranes were incubated in the absence or presence of C-kinase in buffer containing either CaCI, or EGTA. The membranes were then washed in NaCIcontaining buffer and incubated for different times with [y-'"PIATP, MgCI?, and CaCI2. It can be seen in Fig. 3A that, in the absence of C-kinase, the EGF receptor undergoes rapid phosphorylation followed by rapid dephosphorylation which is almost complete in 5 min. This dephosphorylation is due to the presence in the membrane of an active tyrosine-specific protein phosphatase (33). When membranes were incubated supplemented with 4% complete calf serum. After 16 h, 2 pI of dimethyl sulfoxide containing 4&phorbol, teleocidin, or TPA were added, and 1 h later the cells were lysed and the EGF receptor was recovered by immunoprecipitation ("Experimental Procedures"). In a similar hut separate experiment, EGF was added to "P-labeled cells, in the absence of dimethyl sulfoxide, for 1 h before lysis and immunoprecipitation. EGF receptor was digested with trypsin and an equal fraction of each sample was analyzed ("Experimental Procedures"). Autoradiographs were exposed with the aid of a fluorescent screen. Arrowheads. points of sample application. A, 100 ng/ml of 4iJ"phorbol with C-kinase in the presence of CaCl,, a phosphoprotein of 83,000 M, corresponding to the self-phosphorylated C-kinase was found associated with the membrane (Fig. 3B, lane c ) .
The C-kinase was tightly bound to membrane and could not be removed by washing with 0.5 M NaCI. When membranes were incubated with enzyme in the absence of Ca,', C-kinase failed to bind to the membranes (Fig. 3B, lane 6). Once bound to the membrane, C-kinase was able to phosphorylate the EGF receptor rapidly; but, like the autophosphorylation process, the EGF receptor was then dephosphorylated by membrane protein phosphatases which hydrolyze phosphothreonine and phosphoserine (Fig. 3). Because Rosner et al. (34) recently reported that tumor promoters blocked phosphorylation of EGF receptors on tyrosine residues, we examined the effect of tumor promoters on EGF-induced tyrosine phosphorylation of the EGF receptor/kinase and other cell proteins in intact A431 cells. Addition of EGF increased the relative content of phosphotyrosine in total cell proteins from 0.04 to 0.21% (Table I). Treatment with TPA alone had no effect, but TPA reduced the EGF-stimulated increase in protein phosphotyrosine content by -45%. TPA also reduced the EGF-stimulated increase in phosphotyrosine content of the EGF receptor by -60% (Table   I).
EGF receptor from EGF-or TPA-treated cells contains 2.2-3.6 mol of phosphate/mol of polypeptide (data not shown). One can therefore calculate that about 20-35% of the molecules were phosphorylated at tyrosine in EGF-treated cells and at peptides X , Y, or 2 in TPA-treated cells.
T o evaluate the role of C-kinase in this effect, EGF receptor/kinase was phosphorylated by C-kinase prior to selfphosphorylation. A4318 membranes containing EGF receptor/ kinase were incubated with C-kinase, CaCI,, MgC12, and unlabeled ATP for 5 min at 22 "C to phosphorylate the EGF receptor/kinase (Fig. 3, A and B ) . EGF at 100 nM was then added for 2 min at 22 "C followed by addition of [y-"'P]ATP and MnCl,. Fig. 3C shows that :12P incorporation into alkalistable phosphoamino acids was markedly reduced in EGF receptor/kinase phosphorylated by C-kinase (Fig. 3C), indicating a reduced ability of EGF receptor/kinase to self-phos- phorylate a t tyrosine residues in uitro.

DISCUSSION
In R2P-labeled A431 cells, the EGF receptor is constitutively phosphorylated a t several serine and threonine residues, and a tyrosine residue is phosphorylated in response to EGF (31). When either of two active tumor promoters with two distinct structures are added, two new major and one minor phosphothreonine-containing tryptic peptides are observed. Iwashita and Fox have also found that active tumor promoters stimulate phosphorylation of EGF receptors at threoninecontaining peptides (38). In uitro, C-kinase catalyzes specific phosphorylation a t these same sites in the EGF receptor, indicating that the in uiuo phosphorylation, which is stimulated by tumor promoters, is most likely mediated by this enzyme. In uitro phosphorylation requires Ca"; phospholipid and diacylglycerol are not required with A431s cell membranes, presumably because of the lipid present, but phosphorylation of purified EGF receptor protein required phospholipid, and either TPA or diacylglycerol.
A431" cell membranes isolated in the presence of EGTA contain little C-kinase although the enzyme is readily detected in the cytosol. Added C-kinase binds strongly to such membranes and binding as well as activation of the enzyme is dependent on Ca'+. Tumor promoters and Ca2+ are likely to cooperate in uiuo to fix C-kinase at the membrane (9). Isolated membranes contain phosphatase activities which rapidly re-

TABLE I Effect of TPA on ECF-dependent protein kinase activity in vivo
In viuo labeling was carried out for 16 h as described in Fig. 2. TPA (100 ng/ml) was then added, followed SO min later by 60 ng/ml of EGF, and cells were harvested 10 min later. A portion of each lysate was extracted into phenol and phosphoamino acid content of hydrolyzed proteins quantitated (IO6 cpm analyzed in each case) (31). EGF receptor was isolated from a second portion of extract by immunoprecipitation and SDS-polyacrylamide gel electrophoresis, and phosphoamino acids were quantitated following hydrolysis of isolated EGF receptor protein (200-400 cpm analyzed in each case). Internal authentic standards were included, and results are expressed as the percentage of total phosphoamino acids which was released as phosphotyrosine. move not only tyrosine-linked phosphate but also threonineand serine-linked phosphate. In uiuo the state of phosphorylation of the EGF receptor therefore depends on the balance between kinase activation in response to exogenous signals such as EGF and tumor promoters and these phosphatase activities.