Mechanism of Phosphorylation of the Epidermal Growth Factor Receptor at Threonine 669”

The major site of phosphorylation of the epidermal growth factor (EGF) receptor after treatment of cells with EGF is threonine 669. Phosphorylation of this site is also associated with the transmodulation of the EGF receptor caused by platelet-derived growth factor and phorbol ester. A distinctive feature of the primary sequence surrounding threonine 669 is the proximity of 2 proline residues (-Pro-Leu-Threas-Pro-). This site is not a substrate for phosphorylation by protein kinase C. To investigate the mechanism of the increased phosphorylation of the EGF receptor at threonine 669, in vitro assays were used to measure protein kinase and protein phosphatase activities present in homogenates prepared from cells treated with and without EGF. No evidence for the regulation of protein phosphatase ac- tivity was obtained in experiments using the [S2P]phos-phate-labeled EGF receptor as a substrate. A synthetic peptide corresponding to residues 663-681 of the EGF receptor was used as a substrate for protein kinase assays. Incubation of murine 3T3 L1 pre-adipocytes and human WI-38 fibroblasts with EGF caused a rapid increase (3-10-fold) in the level of threonine protein kinase activity detected in cell homogenates. Similar results were obtained after EGF treatment of Chinese hamster ovary cells expressing wild-type (Threes) and mutated (Alaaee) human EGF receptors. Activation of the threonine protein kinase activity was also observed in cells treated with platelet-derived growth factor, serum,

The major site of phosphorylation of the epidermal growth factor (EGF) receptor after treatment of cells with EGF is threonine 669. Phosphorylation of this site is also associated with the transmodulation of the EGF receptor caused by platelet-derived growth factor and phorbol ester. A distinctive feature of the primary sequence surrounding threonine 669 is the proximity of 2 proline residues (-Pro-Leu-Threas-Pro-). This site is not a substrate for phosphorylation by protein kinase C. To investigate the mechanism of the increased phosphorylation of the EGF receptor at threonine 669, in vitro assays were used to measure protein kinase and protein phosphatase activities present in homogenates prepared from cells treated with and without EGF. No evidence for the regulation of protein phosphatase activity was obtained in experiments using the [S2P]phosphate-labeled EGF receptor as a substrate. A synthetic peptide corresponding to residues 663-681 of the EGF receptor was used as a substrate for protein kinase assays. Incubation of murine 3T3 L1 pre-adipocytes and human WI-38 fibroblasts with EGF caused a rapid increase (3-10-fold) in the level of threonine protein kinase activity detected in cell homogenates. Similar results were obtained after EGF treatment of Chinese hamster ovary cells expressing wild-type (Threes) and mutated (Alaaee) human EGF receptors. Activation of the threonine protein kinase activity was also observed in cells treated with platelet-derived growth factor, serum, and phorbol ester. Insulin-like growth factor-1 caused no significant change in protein kinase activity. Together these data indicate a role for the regulation of the activity of a threonine protein kinase in the control of the phosphorylation state of the EGF receptor at threonine 669. The significance of the identification of a growth factor-stimulated threonine protein kinase to the mechanism of signal transduction is discussed.
The cell surface receptor for epidermal growth factor (EGF)' is a 170-kDa transmembrane glycoprotein. The binding of EGF to the extracellular domain of the receptor causes * These studies were supported by Grants GM37845 and CA39240 from the National Institutes of Health. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The abbreviations used are: EGF, epidermal growth factor; CHO, Chinese hamster ovary; EGTA, [ethylenebis(oxyethylenenitrilo)]tetraacetic acid HEPES, 4-(2-hydroxyethyl)-l-piperazineethanesulfonic acid IGF-1, insulin-like growth factor-1; synthetic peptide T669, Glu-Leu-Val-Glu-Pro-Leu-Thr-Pro-Ser-Gly-Glu-Ala-Pro-Asn-Gln-Ala-Leu-Leu-Arg; PDGF, platelet-derived growth factor; PMA, phorbol 12-myristate 13-acetate; HPLC, high performance liquid chromatography. 108 an increase in the tyrosine protein kinase activity of the receptor cytoplasmic domain (1). Treatment of cells with PDGF or with phorbol ester causes rapid alterations in the apparent affinity and tyrosine protein kinase activity of the EGF receptor. This process has been termed transmodulation (1). Treatment of human fibroblasts with EGF, PDGF, or with phorbol ester causes an increase in the phosphorylation state of the EGF receptor at several sites (2). One of these sites, threonine 654 (3,4), is a substrate for protein kinase C. It has been proposed that the phosphorylation of the EGF receptor by protein kinase C at threonine 654 is mechanistically related to EGF receptor transmodulation (1,2). Recently this hypothesis has been critically tested by site-directed mutagenesis of the EGF receptor cDNA and the investigation of the properties of wild-type and mutated receptors expressed in cultured cells. Substitution of threonine 654 with an alanine (5) or a tyrosine (6) residue does not significantly affect the regulation of EGF binding caused by PMA (but see Ref. 7). These data indicate that the phosphorylation of threonine 654 by protein kinase C is not required for the inhibition of EGF binding caused by PMA and that other mechanisms must account for this action. However, it is possible that the mechanism of receptor regulation does involve the phosphorylation of the EGF receptor. This is because multiple EGF receptor serine and threonine residues are phosphorylated in phorbol ester (3,4,(8)(9)(10)(11)-, diacylglycerol(l2,13)-, and PDGF (14, 15)-treated cells under conditions in which a decreased apparent affinity of the receptor is observed. These sites are also phosphorylated in phorbol ester-treated CHO cells that express [Ala654]EGF receptors (5). It is therefore possible that the phosphorylation of one or more of these sites is responsible for the regulation of the apparent affinity of the EGF receptor.
Recently the major site of regulated phosphorylation of the EGF receptor has been identified as threonine 669 (16). EGF receptor threonine 669 is not a substrate for phosphorylation by protein kinase C. However, threonine 669 is located close to the site phosphorylated by protein kinase C, threonine 654 (3,4). The primary sequence surrounding threonine 654 contains several basic residues and is similar to the sequence of protein kinase C substrate sites on other proteins. In contrast, there are no basic residues in the primary sequence surrounding threonine 669. A distinctive feature of the location of threonine 669 is the proximity to 2 proline residues. The mechanism by which threonine 669 is phosphorylated in cells incubated with EGF, PDGF, or phorbol ester is not understood.
The purpose of the experiments reported here was to evaluate the role of the phosphorylation of the EGF receptor at threonine 669 and to determine the mechanism by which EGF, PDGF, and phorbol ester regulate the phosphorylation of the EGF receptor at this site. There are three potential mechanisms that could account for the regulation of the phosphorylation state of threonine 669: 1) activation of a protein kinase; 2) inhibition of a protein phosphatase; and 3) EGF binding and receptor transmodulation may alter receptor conformation or change the receptor subcellular distribution. The approach that we used to examine these hypotheses was to measure the activities of protein kinases and protein phosphatases present in cell homogenates that utilize threonine 669 as a substrate. We report that no marked alteration in protein phosphatase activity was detected, but that EGF, PDGF, and phorbol ester markedly stimulated the activity of a protein kinase that phosphorylated threonine 669. The effect of EGF to stimulate this protein kinase activity was also observed in cells expressing a mutated [AlaM9]EGF receptor. Together these data indicate a role for the regulation of protein kinase activity in the control of the phosphorylation state of the EGF receptor at threonine 669.

Plasmid Construction
Oligonucleotide-directed mutagenesis of T h P ' (ACA), T h P (ACC), and ThrW9 (ACA) to alanine (GCA or GCC) was carried out using 17-mer oligonucleotides according to Zoller and Smith (20) using methods described previously (21). The wild-type and mutated EGF receptor cDNAs were cloned as 4-kilobase XbaI-Hind111 fragments into the expression vector pX (obtained from Dr. G. Johnson, Jewish National Center, Denver) which contains the murine dihydrofolate reductase gene as a selectable marker and allows the expression of the EGF receptor cDNA using the SV40 early promoter and polyadenylation signals. The plasmids obtained were designated and pXER, P X E R ( A~~~~' ) , P X E R ( A~~~~~) , a n d pXER(AlaS6').
Cell Culture 3T3-Ll and WI-38 fibroblasts were obtained from the American Type Culture Collection and were maintained in modified Eagle's medium supplemented with 5% calf serum. A431 cells were obtained from Dr. G. Todaro (Oncogen) and were maintained in Dulbecco's modified Eagle's medium supplemented with 5% calf serum. CHO cells expressing the human PDGF receptor (B-type) cDNA using a pZipNeoSV(X) vector (22) were obtained from Drs. L. Claesson-Welsh and C. H. Heldin (University of Uppsala, Sweden). The cells were maintained in Ham's F-12 medium supplemented with 5% fetal bovine serum and 0.25 mg/ml '2418 (Geneticin, Gibco). CHO cells were transfected with the plasmids pXER, pXER(Ala6'j'), P X E R ( A~~~~' ) , and pXER(AlaW9) using the calcium phosphate technique. After 3 days the cells were passaged and selected using modified Eagle's medium 01 supplemented with 5% dialyzed fetal bovine serum, 0.5 p M amethopterin, and 0.25 mg/ml G418. Stable colonies were isolated using cloning rings and screened for the expression of EGF receptors by measuring the cell surface binding of lz5I-EGF at 0 "C. No specific binding of '=I-EGF was observed to the parental CHO cells.
Phosphorylation of Synthetic Peptide T669 in Vitro Cells were seeded in 35-mm wells and grown to a density of 2 x lo5 cells/well. The monolayers were washed in serum free medium and incubated for 30 min with 1 ml of 120 mM NaC1,6 mM KC1, 1.2 mM CaC12, 1 mM MgC12, 25 mM HEPES (pH 7.4), 30 p M bovine serum albumin at 37 "C. The cells were then treated without and with growth factors or phorbol ester for defined times. The medium was aspirated and the cells were collected by scraping in 0.5 ml of 25 mM HEPES (pH 7.4), 5 mM EGTA, 50 mM NaF, 10 pg/ml leupeptin (0 "C), and were homogenized by 10 passages through a 26-gauge needle. Phosphorylation reactions were performed at 22 "C using 5 p1 of the cell extract, 25 p1 of 50 mM HEPES (pH 7.41, 20 mM MgCL, and 10 pl of synthetic peptide (5 mg/ml). The phosphorylation reaction was initiated by the addition of 10 pl of 50 p M [Y-~'P]ATP (100 pCi/nmol) and was terminated by the addition of 10 p1 of 90% formic acid. Two procedures were used to isolate the phosphorylated synthetic peptide from the reaction mixture.
High Pressure Liquid Chromatography-This purification procedure was used for experiments designed to investigate the physical and chemical properties of the phosphorylated synthetic peptide (Fig.  1). The synthetic peptide T669 phosphorylated in vitro (60 pl) was diluted with 1 ml of 0.1% trifluoroacetic acid (v/v) and applied to a C,, Sep-pak cartridge (Millipore). The Sep-pak was washed with 10 ml of 0.1% trifluoroacetic acid and subsequently with 6 ml of 5% acetonitrile, 0.1% trifluoroacetic acid. The synthetic peptide was eluted from the Sep-pak with 1 ml of 99.9% acetonitrile, 0.1% trifluoroacetic acid. The eluant was lyophilized, dissolved in 0.5 ml of 0.1% trifluoroacetic acid and injected onto a Vydac Cla reverse-phase HPLC column (0.46 X 25 cm) equilibrated with 0.1% trifluoroacetic acid. The column was eluted with a linear gradient of acetonitrile (l%/min) in 0.1% trifluoroacetic acid. The eluant was monitored using in-line detectors for both absorbance at 214 nm and Cerenkov radiation. The synthetic peptide T669 eluted from the column at 28.5% acetonitrile and was detected by absorbance at 214 nm. After phosphorylation with [-p3'P]ATP two peaks of UV absorbance were eluted from the column at 27 and 28.5% acetonitrile. A peak of Cerenkov radiation was detected at 27% acetonitrile, but not at 28.5% acetonitrile. The peak of UV absorbance and Cerenkov radiation eluted from the column at 27% acetonitrile was not observed if either the ATP or the synthetic peptide were omitted from the phosphorylation reaction. Analysis by thin layer electrophoresis demonstrated the presence of a single radioactive peptide which stained with ninhydrin and was detected by autoradiography. Phosphoamino acid analysis indicated the presence of [32P]phosphothreonine. These observations indicate that the radioactive UV absorbance peak that eluted from the reverse-phase column at 27% acetonitrile was a phosphorylated form of the synthetic peptide T669.
Thin Layer Electrophoresis-This method was used for assays designed to measure the activity of protein kinases present in cell extracts using the peptide T669 as a substrate. The synthetic peptide was isolated from the phosphorylation reaction mixture by electrophoresis (4 "C) for 3 h at 500 V on a 100-pm cellulose thin layer plate using 30% (v/v) formic acid as solvent. The phosphorylated peptide was identified by autoradiography, and the incorporation of radioactivity into the peptide was quantitated by liquid scintillation counting.
Protein Phosphatase Assays A431 cells were seeded in 35-mm wells and grown to a density of 1 X lo6 cells/well. The cells were washed in phosphate-free Dulbecco's modified Eagle's medium and incubated in 1 ml of the same medium supplemented with 0.1% calf serum and 1 mCi/ml [3ZP]phosphate. The cells were incubated for 20 h at 37 "C. EGF receptors were isolated by immunoprecipitation as described (4). The immunoprecipitates were washed and incubated in 45 pl of 25 mM HEPES (pH 7.4), 5 mM MgC12,lO pg/ml leupeptin at 22 "C with 5 p1 of cell extract. The cell extracts were prepared as described for the protein kinase assays. After defined times the phosphatase reaction was terminated by the addition of 120 pl of Laemmli sample buffer containing 50 mM dithiothreitol. The phosphorylation state of the EGF receptor was analyzed by polyacrylamide gel electrophoresis and autoradiography. The results were quantitated by excising the gel slices containing the EGF receptor and measuring the associated radioactivity by liquid scintillation counting.

Threonine Protein
Kinase Activation by EGF

Analysis of '251-EGF Binding
CHO cells were seeded in 16-mm wells and grown to a density of 5 X lo4 cells/well. The cells were incubated for 24 h in medium supplemented with 0.1% calf serum. The monolayers were then washed with 120 mM NaC1, 6 mM KC1, 1.2 mM CaC12, 1 mM MgC12, 25 mM HEPES (pH 7.4), 30 yM bovine serum albumin, and incubated for 30 min at 37 "C in the same medium. The cells were treated with and without phorhol ester or PDGF at 37 "C and then rapidly cooled to 0 "C. The binding of "'1-EGF to cell surface receptors was measured by incubation of the cells at 0 "C for 3 h as described (15). Nonspecific binding was estimated in incubations containing a 500fold excess of unlabeled ligand.

Analysis of PHIThymidine Incorporation
CHO cells were seeded in 16-mm wells and grown to a density of 5 X lo4 cells/well. The cells were then incubated for 48 h in medium supplemented with 0.1% calf serum. Growth factors were added to the medium together with 5 pCi/ml [3H]thymidine. After a further 24 h of incubation, the incorporation of radioactivity into acidinsoluble material was measured as described (12).

Purification of EGF Receptor Tryptic [32P]Phosphopeptides
A431 cells were labeled with [3ZP]phosphate for 20 h by incubation with phosphate-free Dulbecco's modified Eagle's medium supplemented with 0.1% calf serum and 2 mCi/ml [32P]phosphate. EGF receptors were isolated from the cells by immunoprecipitation of detergent extracts with a polyclonal anti-EGF receptor antibody as described (4). The immunoprecipitates (100 pl) were reduced by heating to 60 "C for 15 min in the presence of 80 yl of 10% sodium dodecyl sulfate, 14 mM dithiothreitol. After cooling, the EGF receptors were alkylated by adding 40 p1 of 0.4 M iodoacetamide, 0.25 M Tris-HC1 (pH 8. 8), and incubation at room temperature for 15 min. Subsequently, 80 pl of 75% glycerol, 25% 2-mercaptoethanol was added, and the sample was heated to 60 "C for 15 min. After polyacrylamide gel electrophoresis, the gel slice containing the receptor was excised. The receptor was eluted with sodium dodecyl sulfate and precipitated with trichloroacetic acid as described (23). The sample was then digested with 1 yg of tosylphenylalanyl chloromethyl ketonetreated trypsin in 100 mM N-ethylmorpholine (pH 8.0). After 5 h, a second addition of trypsin was made, and the incubation was allowed to proceed for a further 19 hours. Phosphopeptide mapping of the trypsin-digested EGF receptor was performed by reverse-phase HPLC using a Vydac C18 column (0.46 X 25 cm) equilibrated with 0.1% trifluoroacetic acid (15). Peptides were eluted with a linear gradient of acetonitrile (I%/min) in 0.1% trifluoroacetic acid. Fractions were collected at 20-s intervals, and the [3ZP]phosphopeptides were detected by measuring the Cerenkov radiation associated with each fraction. The peptide containing the major site of EGF receptor threonine phosphorylation eluted from the column at 27% acetonitrile and has been described previously (4,5,16).

Two-dimensional Phosphopeptide Mapping
[32P]Phosphopeptides were analyzed by two-dimensional separation on 100-pm cellulose thin layer plates by electrophoresis in 30% formic acid (v/v) for 2 h at 400 V and ascending chromatography using water/butan-1-ol/pyridine/acetic acid (60:75:5015) as solvent. The mobility of the peptides was analyzed by autoradiography.

Phosphoamino Acid Analysis
Phosphoamino acid analysis was performed by partial acid hydrolysis (1 h at 110 "C in 6 M HCl) and thin layer electrophoresis by the method of Hunter and Sefton (24) as described (4).

Automated Amino-terminal Sequence Analysis
Sequence analysis of [32P]phosphate-labeled peptides was performed in the presence of 4 nmol of myoglobin using a modified Beckman 490C liquid-phase sequenator and a 0.1 M Quadrol Program (Beckman 121078). Two precycles were performed prior to the first cleavage. The anilino-thiazolinones were converted to phenylthiohydantoins by reaction in 25% trifluoroacetic acid at 56 "C and were identified and quantitated by a modification of the reverse-phase HPLC procedure described by Zimmerman et al. (25) using acetonitrile. The radioactivity associated with the phenylthiohydantoins derived from the peptide that were released at each cycle was measured by liquid scintillation counting.

RESULTS
Characterization of Synthetic Peptide T669-A431 human epidermoid carcinoma cells were labeled with [32P]phosphate, and the EGF receptors were isolated by immunoprecipitation and polyacrylamide gel electrophoresis. The receptors were eluted from the gel and digested with trypsin. The ["PI phosphopeptides obtained were separated by reverse-phase HPLC. The peptide containing the major site of EGF receptor phosphorylation in vivo was eluted from the column at 27% acetonitrile (4, 15) and collected. The radiochemical purity of the phosphopeptide was investigated by two-dimensional chromatography and electrophoresis on a 100-pm cellulose thin layer plate, Autoradiography of the thin layer plate demonstrated a single [32P]phosphopeptide (Fig. 1). This ["PI phosphopeptide was used for phosphoamino acid analysis and radiochemical sequencing. Fig. 1 shows that the peptide contained a [32P]phosphothreonine residue that was released at the seventh cycle of automated Edman degradation. Inspection of the cDNA sequence of the EGF receptor (26) indicated three predicted receptor tryptic peptides that contained a threonine residue located at position 7 from the amino terminus. These threonine residues are located at positions 669, 759, and 969 in the predicted primary sequence of the EGF receptor (26). To identify which of these threonine residues is the phosphorylation site, oligonucleotide-directed mutagenesis of the receptor cDNA was used to replace each residue with alanine. The mutated cDNAs were cloned into an expression vector and transfected into CHO cells. Stable clones expressing EGF receptors were isolated. The clones were labeled with [32P]phosphate and the in vivo phosphorylation state of the EGF receptors was investigated. Similar ["PI phosphopeptide maps were observed for wild-type, [Ala7"]-, and [Ala=']EGF receptors. In contrast, the incorporation of [32P]phosphate into the [Ala"']EGF receptor was observed to be very low (not shown). These data are consistent with the possibility that threonine 669 is the major site of EGF receptor phosphorylation in vivo. To confirm this hypothesis, a synthetic peptide was prepared that corresponds to residues 663-681 of the EGF receptor. This synthetic peptide was identical to the predicted EGF receptor tryptic peptide that contains threonine 669 and was designated T669. A phosphorylated derivative of the synthetic peptide was prepared by incubation of T669 with [r-32P]ATP as described under "Experimental Procedures." The phosphorylated peptide was then purified by reverse-phase HPLC and was eluted from the column at 27% acetonitrile. The synthetic peptide phosphorylated in vitro and the receptor tryptic peptide phosphorylated in vivo were compared by two-dimensional peptide mapping and by phosphoamino acid analysis. No significant difference between the properties of the two peptides was observed (Fig.  1). We conclude that the major site of phosphorylation of the EGF receptor in vivo is threonine 669. A similar conclusion has recently been reported by Heisermann and Gill (16).
Characterization of the [AAlaM9JEGF Receptor-Wild-type and [AlaM9]EGF receptors were expressed in CHO cells. NO specific binding of lZ5I-EGF was observed to the parental cells, but specific high affinity binding of lZ6I-EGF to the surface of the transfected cells was detected. The lz5I-EGF binding isotherm was investigated (Fig. 2) and found to be curvilinear when plotted by the method of Scatchard (27). Analysis of the binding isotherm using the computer program LIGAND (28) showed that a two-site model provided a significantly better description of the experimental data than a one-site model ( p > 0.97). The results of this analysis are summarized in Table I tors was observed. Exposure of the CHO cells to high concentrations of EGF caused the internalization and down-regulation of both the wild-type and the mutant receptors (not shown). Signal transduction by the EGF receptors was investigated by examining the effect of EGF to regulate the incorporation of ["]thymidine into DNA. Fig. 3 shows that EGF caused an increase in the [3H]thymidine incorporation by CHO cells expressing wild-type and [Ala=']EGF receptors.
As the phosphorylation of threonine 669 (16) is associated

TABLE I Analysis of the lZ5I-EGF binding isotherm
The '"I-EGF binding isotherm (Fig. 2) was analyzed by the computer program LIGAND (28). A significantly better fit of the experimental data to a two-site model than to a one-site model was obtained for control cells ( p > 0.97). In contrast, after treatment with PMA the data were fitted to a one-site model better than a two-site model. The table summarizes the best fit of the data obtained (mean f S.E.) for each condition. with the transmodulation of the EGF receptor caused by PMA (4,10,11) and PDGF (14,15), experiments were performed to investigate the regulation of the [Alafifig]EGF receptor. ., EGF.
2 shows that PMA caused the loss of the high affinity binding of lZ5I-EGF to cells expressing wild-type and mutated [Ala669] EGF receptors. The CHO cells employed for these experiments express functional human PDGF receptors (B type).

Treatment of the cells with PDGF caused an inhibition of the high affinity binding of lZ5I-EGF to wild-type cell surface EGF receptors.' No significant difference between the results obtained for cells expressing wild-type and [Ala669]EGF receptors was observed (not shown).
Measurement of Protein Phosphatase Activity-The increase in the phosphorylation state of the EGF receptor at threonine 669 (16) caused by EGF (29) could be due to the inhibition of protein phosphatase activity. To test this hypothesis, the level of protein phosphatase activity present in homogenates prepared from CHO cells treated with and without EGF was measured. The substrate for the phosphatase assay used was the [32P]phosphate-labeled EGF receptor isolated from control A431 cells by immunoprecipitation using a rabbit polyclonal antibody directed against the extracellular domain of the receptor. Quantitative studies (15) of the phosphorylation of the EGF receptor in control A431 cells indicate that approximately 70% of the total [32P]phosphate incorporated into the receptor was at threonine 699 (16). Fig. 4 shows the time course of phosphatase activity using the [32P]phosphate-labeled EGF receptor as a substrate. Rapid dephosphorylation of the receptor was observed after the addition of a homogenate prepared from CHO cells expressing wild-type EGF receptors. Treatment of the cells with EGF before homogenization caused no significant change in the observed rate of dephosphorylation of the receptor (Fig. 4).
Measurement of Protein Kinase Activity-The increased phosphorylation state of the EGF receptor at threonine 669 (16) observed in EGF-treated cells (29) could be caused by the stimulation of the activity of a protein kinase that phosphorylates the EGF receptor at threonine 669. To test this hypothesis, the protein kinase activity in homogenates prepared from cells incubated with and without EGF was measured. The experimental strategy employed was to use the synthetic peptide T669 as a protein kinase substrate. CHO cells were homogenized and incubated with the peptide T669 and [3ZP]Phosphate-labeled EGF receptors isolated from A431 cells by immunoprecipitation were used as a substrate for the assay of protein phosphatase activity. The time course of dephosphorylation of the EGF receptor by extracts prepared from CHO cells expressing the wildtype EGF receptor is presented. The CHO cells were treated with and without 100 nM EGF for 15 min at 37 "C prior to homogenization. The data presented are normalized to the amount of radioactivity associated with the EGF receptor prior to incubation with the cell extract (22,314 cpm). Similar results were obtained in two separate experiments. CL-0, control; U , EGF. isolated and characterized in detail (Fig. 1). The radioactivity incorporated into the peptide was measured by liquid scintillation counting. Phosphoamino acid analysis indicated the presence of [32P]phosphothreonine (Fig. 1). As the synthetic peptide contains only a single threonine residue we conclude that the phosphorylated residue corresponds to EGF receptor threonine 669 (Fig. 1). Control experiments demonstrated that the rate of phosphorylation of the peptide under the standard assay conditions (see "Experimental Procedures") was linear for 20 min at 22 "C.
CHO cells expressing the wild-type human EGF receptor were used to investigate the effect of EGF on the level of threonine protein kinase activity measured using the synthetic peptide T669 as a substrate. Treatment of the CHO cells with EGF caused a marked increase in the protein kinase activity detected in cell homogenates compared with control cells (Fig. 3). In five experiments the increase in protein  kinase activity caused by EGF treatment of the CHO cells was 3.5 f 0.6-fold (mean & S.D.). Similar results were obtained using CHO cells expressing mutated [Ala669]EGF receptors (Fig. 3).
Regulation of Protein Kinase Activity in Fibroblusts-EGF stimulates a threonine protein kinase activity detected in homogenates prepared from CHO cells using the synthetic peptide T669 as a substrate (Fig. 3). As CHO cells do not normally express EGF receptors it was necessary to confirm these results using fibroblasts that express functional endogenous EGF receptors. Fig. 5 shows that treatment of murine 3T3 L1 pre-adipocytes and human WI-38 fetal lung fibroblasts with EGF caused an increase in the threonine protein kinase activity measured in cell homogenates. The increase in the protein kinase activity was observed within 5 min of EGF treatment (Fig. 6), and maximal effects were observed after treatment of the cells with 5 nM EGF (Fig. 7).
In further experiments, the effects of the treatment of 3T3 L1 cells with other growth factors was investigated. Fig. 8 shows that EGF, PDGF, phorbol ester, and serum caused a marked increase in the level of threonine protein kinase activity measured in cell homogenates. However, IGF-1 was observed to cause no significant change in the level of threonine protein kinase activity detected. Similar results were obtained for CHO cells expressing endogenous IGF-1 receptors and the human receptors for EGF and PDGF (not shown).

DISCUSSION
Mechanism of Phosphorylation of the EGF Receptor at Threonine 669-There are three potential mechanisms by which EGF, PDGF, and phorbol ester could regulate the phosphorylation state of the EGF receptor at threonine 669: 1) activation of a threonine protein kinase; 2) inhibition of a protein phosphatase; and 3) alteration of the receptor conformation or subcellular distribution. The experiments reported here were designed to examine these hypotheses.
Measurement of protein phosphatase activity in cell homogenates using [32P]phosphate-labeled EGF receptors as a substrate indicated that there was no significant effect of EGF treatment on the level of phosphatase activity observed (Fig.  4). As this experiment employs an in vitro assay for protein phosphatase activity, the data obtained do not allow the conclusion that protein phosphatase activity is not regulated by EGF in intact cells. Chan et al. (30) have reported that type I protein phosphatase is activated by treatment of cells with EGF. Together these data do not support the hypothesis that EGF increases the phosphorylation of the receptor at threonine 669 by inhibiting protein phosphatase activity.
To measure the activity of protein kinases that phosphorylate the EGF receptor at threonine 669 an in vitro assay was developed using a synthetic peptide substrate that corresponds to residues 663-681 of the EGF receptor. It was observed that the synthetic peptide (T669) was phosphorylated on a unique threonine residue, threonine 669 (Fig. 1). Treatment of murine 3T3 L1 pre-adipocytes and human WI-38 fetal lung fibroblasts with EGF caused a marked increase in the rate of phosphorylation of the synthetic peptide T669 by homogenates prepared from these cells. Similar results were obtained for CHO cells expressing human EGF receptors. These data are consistent with the hypothesis that the mechanism of EGF action to increase the phosphorylation of the receptor at threonine 669 is the activation of a threonine protein kinase.
It is possible that the phosphorylation state of threonine 669 is regulated by the conformation or subcellular distribution of the receptor. No evidence was obtained from the studies reported here that excludes these mechanisms from playing a role in the regulation of the phosphorylation of threonine 669. However, the observation of a growth factorstimulated threonine protein kinase (Figs. 3, 5-8) indicates that EGF could increase threonine 669 phosphorylation in the absence of changes in the conformation or subcellular distribution of the receptor. Consistent with this proposal, it was observed that EGF treatment of CHO cells expressing the [Ala669]EGF receptor caused a stimulation of threonine protein kinase activity (Fig. 3).
Mechanism of Transmodulation of the EGF Receptor-A working hypothesis that we have used to examine the mechanism of transmodulation of the EGF receptor proposes that receptor phosphorylation mediates the regulation of the receptor (for review see Ref. 2). Treatment of cells with phorbol ester (4,10,11) or PDGF (14,15) causes the phosphorylation of the EGF receptor at several sites, including threonine 654 (3,4) and threonine 669 (16). Previous studies have demonstrated that the phosphorylation of threonine 654 by protein kinase C does not fully account for the transmodulation of the EGF receptor caused by phorbol ester (5, 6, but see Ref. 7) or PDGF.' A potential role has been demonstrated for the phosphorylation of the EGF receptor at threonine 654 in the regulation of the tyrosine protein kinase activity of the EGF receptor by protein kinase C (5, 8, 9, but see Ref. 6). Substitution of threonine 654 with an alanine ( 5 ) or a tyrosine (6) residue by site-directed mutagenesis and expression of the mutated receptor in CHO cells demonstrated that threonine 654 is dispensible for the regulation of the high affinity binding of '"I-EGF to the receptor by phorbol ester. However, the [Ala654]EGF receptor was phosphorylated at threonine 669 (16) during transmodulation ( 5 ) . Based on this result it is possible that the phosphorylation of the EGF receptor at threonine 669 may be mechanistically related to the process of transmodulation. We report here that the substitution of threonine 669 with an alanine residue does not alter the inhibition of high affinity binding of EGF caused by phorbol ester and PDGF (Fig. 2). These data demonstrate that replacement of either threonine 654 or threonine 669 with alanine does not alter the regulation of the high affinity binding of lZ51-EGF to the receptor. The effect of simultaneous substitution of both threonine 654 and threonine 669 with alanine remains to be determined.
Identity of the Growth Factor-stimulated Threonine Protein Kinase-The protein kinase activity detected using in vitro assays with the synthetic peptide T669 is stimulated by the treatment of cells with EGF, PDGF, phorbol ester, and serum. The identity of this protein kinase is not known. It has been shown for many protein kinases that the primary sequence of a peptide substrate is a critical factor in determining substrate specificity (31). The proximity of 2 proline residues in the primary sequence of the EGF receptor surrounding threonine 669 is unusual for a site of protein phosphorylation (31). However, two protein kinases have been reported that exhibit a similar substrate specificity: glycogen synthase kinase-3 and the multifunctional protein kinase. Glycogen synthase kinase-3 (32) and multifunctional protein kinase (33, 34) have been reported to phosphorylate glycogen synthase and phosphatase inhibitor I1 at sites located within a proline-rich primary sequence. It is possible that the growth factor-stimulated threonine protein kinase activity reported here may be accounted for by glycogen synthase kinase-3 or by multifunctional protein kinase, but previous studies of the properties of these enzymes do not support this hypothesis. 1) Activation of glycogen synthase kinase-3 and multifunctional protein kinase by growth factors has not been reported (31). 2) The phosphorylation state of a glycogen synthase kinase-3 substrate, the nerve growth factor receptor, is not regulated by treatment of cells with phorbol ester (35). 3) It has been proposed that the substrate specificity of glycogen synthase kinase-3 is serine-X-X-X-phosphoserine/threonine (36). The primary sequence surrounding EGF receptor threonine 669 does not conform to this proposed consensus primary sequence. 4) Treatment of adipocytes with insulin has been reported to cause a rapid inhibition of multifunctional protein kinase activity (37), but no effect of IGF-1 was observed on the threonine protein kinase activity present in homogenates of 3T3 L1 pre-adipocytes or CHO cells when the peptide T669 was used as a substrate (Fig. 8). Together, these data indicate that the growth factor-stimulated threonine protein kinase may be an enzyme that has not been previously described.
Signal Transduction by the EGF Receptor-Treatment of cultured cells with EGF (29), PDGF (14, E), cell-permeable diacylglycerols (12,13), or phorbol ester (4, 10, 11) causes the phosphorylation of the EGF receptor at threonine 669 (16). The data presented here indicate a role for the activation of a threonine protein kinase. It is likely that this growth factorstimulated threonine protein kinase is able to phosphorylate protein substrates other than the EGF receptor in intact cells. Such phosphorylation of target proteins may be physiologically relevant for signal transduction by growth factor receptors (EGF and PDGF) and by protein kinase C (diacylglycerol and phorbol ester). Potential target proteins will probably share homology with the EGF receptor at the site of phosphorylation: Pr~-Leu-Thr~~'-Pro. Recently Giugni et al. (38) reported that EGF activates a serine protein kinase in A431 human epidermoid carcinoma cells that phosphorylates the synthetic peptide Leu-Arg-Arg-Ala-Ser-Leu-Gly. The identification of the substrates for these protein kinases and the elucidation of the mechanism by which EGF increases serine/ threonine protein kinase activity is an important goal for future research.