Epidermal Growth Factor and Epidermal Growth Factor Receptor-dependent Phosphorylation of a M, = 34,000 Protein Substrate for pp6Osrc*

A M, = 34,000 protein present in the 100,000 % g supernatant fraction from A431 human epidermoid carcinoma cells is the major radiolabeled phosphate accep- tor from [y-32P]ATP in a cell-free system requiring epidermal growth factor (EGF) and EGF receptor kinase. This protein is immunoprecipitated by IgG directed against avian M, = 34,000 cellular substrate for pp60"". Phosphoamino acid analysis of the M, = 34,000 protein labeled with 32Pi from [y-'2P]ATP in a cell-free system requiring EGF and EGF receptor kinase yielded radio- labeled phosphotyrosine with no detectable radioactivity in phosphoserine or phosphothreonine.

Cell Culture-Cultures were maintained on plastic dishes and plastic or glass roller bottles in Dulbecco's modified Eagle's medium plus 10% calf serum in an atmosphere of 10% COZ.
Membrane Purification"A431 cell surface membranes were purified by Procedure II (12) modified to contain I0 m~ EDTA during all steps from scraping of cells from their growth substratum to equilibrium banding in a Ficoll density gradient. This routinely produces membranes in which EGF receptor proteins are 90% or more in the undegraded 160 K d form. Details of this procedure will be described elsewhere.' Triton Solubilization ofA431 Membranes-To solubilize A431 cell surface membranes, 700 pg of membranes in 100 pl of 10 mM Hepes at pH 7.4 were mixed with 75 pl of 2.5% Triton X-100 and incubated for 15 min at 22 "C.
Gel Electrophoresis and Autoradiography-Proteins were resolved by electrophoresis in 7.5% polyacrylamide gels using buffer systems described by Laemmli (13). Gels were stained with Coomassie brilliant blue and dried onto Whatman 3 " paper. Radiolabeled proteins were visualized by autoradiography with the aid of a DuPont Lightning-Plus screen.
Phosphoamino Acid Analysis-Gel sections containing y-">P-labeled M , = 34,000 protein band resolved by sodium dodecyl sulfategel electrophoresis (13) were mashed and the labeled protein eluted during an overnight incubation in 50 mM ammonium bicarbonate solution containing 0.1% sodium dodecyl sulfate, 5% P-mercaptoethanol, and 50 pg/ml of bovine serum albumin. This procedure achieves 7040% elution with no detectable degradation of phosphorylated 34K protein. Eluted protein was precipitated with 20% trichloroacetic acid, and precipitated proteins were washed twice with acetone and hydrolyzed in 6 N HCl at 108 "C for 2 h. Samples were lyophilized; authentic phosphoserine, phosphothreonine, and phosphotyrosine standards were added and phosphoamino acids were separated by electrophoresis on cellulose thin layer plates. Electrophoresis in the first dimension was at pH 1.9 (1.5 kV for 1 h) and in the second dimension at pH 3.5 (400 V for 2% h) (14). Standards were visualized by staining with ninhydrin. Radioactive phosphoamino acid spots were visualized by autoradiography. '12P Radioactivity in the spots was quantified by Cerenkov counting.

RESULTS
The EGF receptor kinase phosphorylates tyrosine residues of exogenous substrates (15). However, there so far is no direct cell-free system evidence for EGF receptor-mediated phosphorylation of a specific cellular substrate. In initial in vitro tests for specific EGF-dependent phosphate acceptor activity, we used a cell lysate system in which the concentrations of protein components are close to those encountered in intact cells. A431 human epidermoid carcinoma cells were employed because of their unusually high density of EGF receptors (16).
Cell lysates were incubated with [y-"'P]ATP and Mg' in the presence or absence of EGF, and phosphorylated proteins were resolved by sodium dodecyl sulfate-gel electrophoresis and analyzed by autoradiography (Fig. 1). EGF greatly enhanced phosphorylation of three bands with apparent M , = 160,000, 145,000, and 34,000. The M , = 160,000 band is EGF receptor-associated with membranes, and the M , = 145,000 band is a product of that receptor produced by proteolysis (5). EGF-dependent phosphorylation of a M , = 34,000 protein is in agreement with findings by Erikson et al. (17) that EGF increased phosphate labeling of the 34K protein substrate for pp60"" in A431 cells grown in medium containing labeled phosphate. We therefore performed additional tests to determine if M, = 34,000 protein phosphorylated under cell-free conditions has properties associated with the 34K substrate for pp60"" kinase and also to determine the subcellular localization of M , = 34,000 phosphate acceptor activity. A431 cells were resolved into 100,000 X g supernatant and pellet fractions, and these were tested for phosphate acceptor activity in a system containing detergent-solubilized cell surface membranes from A431 cells as a source of EGF receptor kinase. The pellet fractions contributed no major phosphate acceptor component in addition to those present in detergent- EGF receptor kinase-dependent phosphorylation of a 34K protein present in the 100,000 X g supernatant fraction of A431 cells. A431 cells were disrupted by homogenization in a buffer containing 10 mM Hepes, pH 7.4, and 2.5 mM dithiothreitol and centrifuged at 30,000 X g for 30 min. The 30,000 X g supernatant fraction was centrifuged at 100,000 X g for 1 h. Various amounts of the 100,000 X g supernatant fraction (SUP) ( lanes 2 and 6, 20 pg;  lanes 3, 7, and 9,40 pg; lanes 4 and 8,60 pg) were incubated for 3 min at 30 "C with 40 pg of Triton X-100-solubilized A431 cell surface membranes (hfEhfl3) (lanes 1-8) in a 5Opl system containing 2 mM MgCI2 and 20 p~ unlabeled ATP. Fifty ng of EGF were then added in 1 pl of Hepes buffer (lanes 5-9), and the samples were incubated at 30 O C for an additional 3 min. Samples were cooled to ice bath temperature and 5 p1 of 50 p~ [y-3'P]ATP were added for a 15-min incubation. Reactions were terminated by the addition of 12 p1 of 5times concentrated electrophoresis buffer and samples were analyzed by gel electrophoresis as described in the legend to Fig. 1. solubilized A431 membranes (data not shown). To detect EGF-dependent phosphorylation in reactions containing A431 cell supernatant fraction, it first was necessary to incubate the complete system with unlabeled ATP prior to addition of EGF and radiolabeled ATP. This procedure reduced EGFindependent phosphorylation greatly, permitting detection of EGF-dependent phosphorylation products. When isolated detergent-solubilized A431 membranes were tested alone in the radiophosphorylation system, M, = 160,000 EGF receptor protein was the major labeled product, and its radiophosphorylation required EGF. When supernatant fraction alone was inspected for phosphate acceptor activity, a number of labeled bands were detected, including an intensely labeled one at M, = 90,000 (Fig. 2, lune 3). EGF did not enhance labeling in these bands, nor did EGF stimulate significant EGF-dependent Phosphorylation of a pp60'" Substrate 2043 cell-free system requiring both EGF and EGF receptor kinase.

EGF -+ llhb
In order to test for homology between the M, = 34,000 protein requiring EGF and EGF receptor for phosphorylation in vitro and the 34K cellular substrate for p p 6 W (18), we incubated the high speed supernatant fraction with anti-Mr = 34,000 (anticellular substrate for avian pp60"") prior to adding it to the reconstituted system described in Fig. 2. This treatment blocked EGF-dependent phosphorylation of M, = 34,000 protein, suggesting that EGF receptor-associated kinase and pp60"" play roles in the phosphorylation of the same protein of M , = 34,000 (data not shown).
The M, = 34,000 protein phosphorylated in a reaction stimulated by EGF and EGF receptor kinase also was tested directly for immunoprecipitation by IgG directed at the avian M, = 34,000 cellular substrate for pp60". The immune complexes were isolated by double antibody precipitation and analyzed by gel electrophoresis and autoradiography. There was no detectable phosphorylation when immunoprecipitation was performed in the presence of preimmune IgG (Fig. 3,   lanes 1 and 2). However, when immune IgG was used, a M, = 34,000 band was detected in the autoradiogram (lanes 3 and  4), and its intensity was enhanced 3-fold by EGF (lane 4  uersus lane 3). This provides evidence that viral transforming gene products and EGF receptor kinase stimulated phosphorylation of the same cellular protein in cell-free systems.
Since phosphorylation of M, = 34,000 protein catalyzed by pp60"" and in vitro phosphorylation of EGF receptors by EGF receptor-associated kinase occur exclusively on tyrosine residues (6,18), it was of interest to determine which amino acids in the M , = 34,000 protein are phosphorylated in the in vitro reaction requiring EGF and EGF receptor kinase. M, = 34,000 protein was phosphorylated in a reaction identical with that described in Fig. 2 (lanes 2 and 4 ) or absence (lunes 1 and 3) of EGF using the phosphorylation reaction system protocol described in the legend to Fig. 2 Phosphorylation of the 34K protein present in an A431 cell 100, OOO X g supernatant fraction was achieved in the presence of EGF and purified detergent solubilized A431 membranes as described in the legend to Fig. 2. The 34K protein was resolved by gel electrophoresis, eluted from the gel, and hydrolyzed in 6 N HCI at 108 "C for 2 h for phosphoamino acid analysis ("Experimental Procedures").
polyacrylamide gels. M, = 34,000 protein was eluted and hydrolyzed with acid, and the hydrolysis products were separated by two-dimensional electrophoresis (Fig. 4). Radiophosphate was present in tyrosine residues; no detectable radioactivity was detected in phosphoserine or phosphothreonine (Fig. 4).

DISCUSSION
The EGF receptor kinase and pp60"" arising from Rous sarcoma virus transformation have similar substrate determinants in intact cells (17) and phosphorylate exclusively a t tyrosine residues in cell-free systems (6)(7)(8). Moreover, close correlations between increased tyrosine-specific protein kinase activity, increased phosphotyrosine content, and cell proliferation have been demonstrated in several retrovirally transformed cells (19). Increased phosphorylation of proteins at tyrosine also is a property of cultured EGF-treated cells of the A431 human epidermoid carcinoma line (20).
In this report, we present evidence that in the presence of EGF, EGF receptor kinase participates in EGF-dependent phosphorylation of a M, = 34,000 protein in A431 cell extracts. This M, = 34,000 protein has homology with the M, = 34,000 cellular substrate for pp60"" based on its immunoprecipitation by IgG directed at the avian 34K cellular substrate for pp60"" (Fig. 3). Phosphoamino acid analysis of M, = 34,000 protein phosphorylated by EGF receptor kinase in a cell-free system (Fig. 4) reveals labeled phosphate exclusively on tyrosine residues. Erikson et al. (17) recently showed that EGF stimulates phosphorylation of a M, = 34,000 protein in viable A431 cells, and that this protein is immunoprecipitated by IgG directed at avian M, = 34,000 pp60" substrate. Tryptic phosphopeptide-mapping data permitted these authors to conclude that A431 cell M, = 34,000 protein is phosphorylated in intact cells in response to EGF at sites identical with those phosphorylated in vitro by p p 6 0 protein kinase activity. EGF-stimulated EGF receptor kinase and p p 6 0 kinase play roles in the phosphorylation of similar acceptor peptides. IgG directed a t avian pp60"" is an excellent substrate for both in spite of its lack of immunoreactivity toward EGF receptors. Pike et al. (21) showed that the sequence of the region around the phosphorylated tyrosine is a determinant of EGF-stimulated receptor kinase activity. When tested as substrates for EGF-stimulated receptor kinase, synthetic peptides with sequences similar to the pp60"" phosphorylation site are superior to a tyrosine-containing peptide of totally unrelated se-

EGF-dependent Phosphorylation
of a pp6W' Substrate quence. However, both pp60"" kinase and EGF-stimulated receptor kinase have broad specificity for a wide number of peptides, including casein and histones (15). Of the proteins present in an A431 cell supernatant fraction, the only one for which significant EGF direct phosphate acceptor activity was detected is M, = 34,000 protein. In contrast, none of the major proteins that are present in an A431 cell supernatant fraction and can be detected by Coomassie blue staining accepted detectable amounts of radiophosphate. Refinements in the cell-free system described here should allow determination as to whether the high acceptor specificity of M, = 34,000 protein resides solely in short sequences surrounding the phosphorylated tyrosine or also in additional recognition sites between M , = 34,000 protein and the EGF receptor kinase system.