JOURNAL OF CELLULAR PHYSIOLOGY 146:63-72 (1991)

Epidermal Growth Factor Treatment of A431 Cells Alters the Binding Capacity and Electrophoretic Mobility of the Cvtoskeletallv Associated Epidermal Growth Factor Receptor LINDA M. ROY, CYNTHIA

K. GITTINGER, AND GARY E.

LANDRETH*

Department of Neurology, Medical University of South Carolina, Charleston, South Carolina 29425 The epidermal growth factor (EGF) receptor interacts with structural elements of A431 cells and remains associated with the cytoskeleton following extraction with nonionic detergents. Extraction of cells with 0.15% Triton X-100 resulted in detection of only approximately 40% of the EGF binding sites on the cytoskeleton. If the cells were exposed to EGF prior to extraction, approximately twofold higher levels of low-affinity EGF binding sites were detected. The difference in number of EGF binding sites was not a consequence of differences in numbers of EGF receptors associated with the cytoskeleton;equal amounts of '5S-labeled receptor were immunoprecipitated from the cytoskeletons of both control and EGF-treated cells. The effect of EGF pretreatment on binding activity was coincident with a change in the mobility of the receptor from a doublet of M, - 160-1 80 kDa to a single sharp band at 180 kDa. The alteration in receptor mobility was not a simple consequence of receptor phosphorylation in that the alteration was not reversed by alkaline phosphatasetreatment, nor was the shift produced by treatment of the cells with phorbol ester. The two EGF receptor species demonstrated differential susceptibilityto V8 proteinase digestion. The EGF-induced 180 kDa species was preferentially digested by the proteinase relative to the 160 kDa species, indicating that EGF binding results in a conformational change in the receptor. The EGF-mediated preservation of binding activity and altered conformation may be related to receptor oligomerization. Epidermal growth factor (EGF) produces its mitogenic signal through its specific cell surface receptor (Carpenter and Cohen, 1979; Cohen, 1987; Schlessinger, 1988). The receptor for EGF is a 170 kDa integral membrane glycoprotein (Carpenter et al., 1979;Cohen et al., 1980; Downward et al., 1984;Ullrich et al., 19841, possessing an extracellular ligand binding domain (Kawamoto et al., 1983; King and Cuatrecasas, 1982;Rees et al., 1984) and an intrinsic tyrosine kinase activity, which is activated upon binding of EGF (Buhrow et al., 1983; Russo et al., 1985; Ushiro and Cohen, 1980). The specific biochemical events subserving the action of EGF on cellular proliferation remain unclear; however, the tyrosine kinase activity of the receptor is essential for transmission of biological signals (Chen et al., 1987; Honegger et al., 1987a,b; Livneh et al., 1986, 1987; Moolenaar et al., 1988; Prywes et al., 1986). Our approach to studying the mechanism by which the EGF receptor transmits a biological signal has focussed on the interaction of the receptor with the underlying structural elements of the cell. We have taken advantage of an in situ system that preserves cytoskeletal structures and positional relationships following hormone treatment (Burr et al., 1980; Landreth et al., 1985). We operationally define the cytoskeleton as the detergent-insoluble fraction follow0 1991 WILEY-LISS, INC.

ing extraction of cells with a nonionic detergent, Triton X-100. Using this preparation, we previously demonstrated the association of the EGF receptor with the detergent-insoluble cytoskeleton of A431 cells (Landreth et al., 1985). The cytoskeletally associated EGF receptor retained an intact tyrosine kinase activity and a functional ligand binding domain (Landreth et al., 1985; Roy et al., 1989). The EGF receptor exists on the cell surface in both high-affinity and low-affinity states (Chatelier et al., 1986; Gullick et al., 1984; Kawamoto et al., 1983; King and Cuatrecasas, 1982). Lateral diffusion studies with rhodamine-labeled EGF show that the higher affinity form of the EGF receptor was immobile, likely as a result of preferential association of this form of the receptor with strucutal elements of the cell (Rees et al., 1984; Roy et al., 1989; Wiegant et al., 1986). Signal transduction by EGF occurs through occupancy of the Received May 23, 1990; accepted September 25, 1990. *Gary E. Landreth's present address is Alzheimer Research Laboratory E504, Case Western Reserve University School of Medicine, Cleveland, Ohio 44106. Address reprint requestskorrespondence there. Linda M. Roy's present address is Department of Pharmacology, University of Colorado Health Sciences Center, Denver, Colorado 80262.

64

ROY ET AL

high-affinity receptor (Defize et al., 1989), suggesting that the cytoskeleton may have a pivotal role in the mechanism of signal transduction. The experiments described in this paper were initiated when we noted a disparity in the number of cytoskeletally associated EGF binding sites on control and EGF-treated cells. We report here that EGF treatment of A431 cells results in an apparent increase in the number of lZ5I-EGFbinding sites present on the cytoskeleton. Surprisingly, this higher level of binding was not due to ligand-induced association of EGF receptors with the cytoskeleton but rather to modification of the EGF receptor following EGF binding as evidenced by altered electrophoretic migration of the receptor. MATERIALS AND METHODS Materials Receptor-grade EGF was purchased from Biomedical Technologies, Inc. (Stoughton, MA). Triton X-100, alkaline phosphatase (from bovine intestinal mucosa), and lactoperoxidase were purchased from Sigma Chemical Co. (St. Louis, MO). Na1251,T r a n ~ - and ~~S 32Piwere obtained from ICN Radiochemicals (Irvine, CA). Radiolabeled ATP was prepared using Gamma Prep A (Promega Biotech, Madison, WI). Sephadex G-25 was purchased from Pharmacia, Inc. (Piscataway, NJ). Anti-EGF receptor antibodies were obtained from Oncogene Sciences (Manhasset, NY). Cell culture Human A431 epidermoid carcinoma cells were the generous gift of Dr. M.A. Bothwell. The cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) containing 5% fetal calf serum (Hyclone, Logan, UT) and were maintained in an atmosphere of 5% C02 at 37°C. Iodination of EGF EGF was iodinated to a specific activity of 100-200 cpm/pg using the lactoperoxidase method (Marchalonis, 1969).Unincorporated 1251 was removed by passage of the sample through a Sephadex G-25 minicolumn (Tuszynski et al., 1980). Preparation of cytoskeletons Adherent A431 cells were grown t o near confluency in 35 or 16 mm tissue culture wells. Control and EGF-treated (see below) cells were washed twice with ice-cold PBSAlGLU (phosphate-buffered saline containing 1mg/ml each bovine serum albumin [BSA] and glucose). Cytoskeletons were prepared by extracting the cells with Triton buffer (25 mM Hepes, 2 mM MnC12, 1 mM PMSF, 10 pM Na3V04, 4 mM iodoacetic acid, 0.15% [w/vl Triton X-100, unless otherwise indicated) for 2 min at 4°C. 1251-EGFbinding studies The various methods for the incubation of A431 cells with EGF are illustrated schematically in Figure 1.To determine 1251-EGFbinding to whole cells (Fig. 1, whole), A431 cells were seeded in 16 mm tissue culture wells the previous day to yield a final density of 1 x lo5 cells per well. Cells were washed twice with ice-cold

PBSA/GLU and then were incubated with 0.2 ml of PBSA/GLU containing 0.1 ng/ml(16.9 pM) to 500 ngiml (82 nM) lZ5I-EGFfor at least 8 hr at 4°C. The incubation medium then was removed, and the cells were washed twice with PBSA/GLU. The incubation medium and washes were pooled, counted on an LKB gamma counter, and considered to be unbound ligand. The cells were solubilized with 1 ml of 1 N NaOH, and bound radioactivity measured. Nonspecific binding was measured in parallel incubations in the presence of 100500-fold excess unlabeled EGF and was less than 10% in all cases. Analysis of binding was performed by the method of Scatchard (1949) using the Simplex program (Lam, 1970; Nelder and Mead, 1965). Binding of lz5I-EGF to cytoskeletons of untreated cells (see Fig. 1,csk) was performed by first extracting A431 cells with Triton buffer. The resulting cytoskeletons were washed with Hepes/BSA buffer (25 mM Hepes, 2 mM MnCl,, 1 mM PMSF, 10 pM Na3V04, 4 mM iodoacetic acid, 1%[w/v] BSA). The; were then incubated in Hepes/BSA buffer with l2 I-EGF and processed exactly as described above. 12%EGF binding by the cytoskeletons of EGFtreated cells was determined by two different methods (see Fig. 1).In the first method (Fig. 1,whole tx), whole cells were first incubated with lZ51-EGFfor 8 hr at 4°C. The incubation medium was then removed, and the cells were washed. The incubation medium and washes were pooled together and counted. The cytoskeletons were isolated with Triton buffer for 2 min at 4°C and washed with HepeslBSA buffer. The remaining cytoskeletal pellet with the bound ligand was solubilized with 1 ml of 1 N NaOH. In the second method (Fig. 1, pretreated), whole cells were washed with ice-cold PBSAIGLU and then pretreated with 200 ng/ml unlabeled EGF for 2 hr at 4°C. The medium was removed, and the cells were washed with PBSA/GLU. Bound EGF was removed from the cells by treatment with 0.2 M acetic acid/0.5 M NaCl for 5 min at 4°C (Haigler et al., 1980). After several washes with PBSA/GLU, cytoskeletons were prepared by extraction from the cells with Triton buffer containing the indicated concentrations of detergent for 2 min at 4°C. The cytoskeletons were washed with Hepes/BSA buffer and incubated with HepeslBSA buffer containing lZ5I-EGFfor 8 hr at 4°C.Measurement of ‘“I-EGF binding was performed as described above. Metabolic labeling and immunoprecipitation A431 cells were plated into 35 mm wells 24 hr prior to the experiment in DMEM containing 5% fetal calf serum and T r a n ~ - (50 ~ ~ pCilm1). S Proteins were labeled to a specific activity of approximately 3,000 cpm/p,g protein. Protein concentrations were determined by the method of Bradford (1976). Control and EGF-treated (200 ng/ml) cells were lysed in Triton buffer as described. The EGF receptors from both the detergentsoluble and insoluble fractions were immunoprecipitated using a monoclonal antibody directed against the extracellular domain of the EGF receptor (Kawamoto et al., 1983). PBS-DT buffer (10 mM NaPO,, pH 7.25, 160 mM NaCl, 1%Triton X-100,0.5%deoxycholic acid, 0.2% sodium azide, 0.1% SDS, 10 mM NaF, 4 mM iodoacetic acid, 50 FM Na,V04, 1 mM EGTA) was

CYTOSKELETALLY ASSOCIATED EGF RECEPTOR

65

WHOLE

INTACT CELLS

INTACT CELLS WITH 125 BOUND I-EGF

WHOLE TX INTACT CELLS

CSK

@ @3

INTACT CELLS WITH 125 BOUND I-EGF

CYTOSKELETONS WITH 125 BOUND I-EGF

INTACT CELLS

PRETREATED

CYTOSKELETONS

CYTOSKELETONS WITH 125 BOUND I- EC F

ACID WASH TO REMOVE BOUND

CYTOSKELETONS

CYTOSKELETONS WITH 125 BOUND I- EGF

Fig. 1. Schematic diagram illustrating the different experimental procedures used, as described in Materials and Methods.

added to both fractions. The samples were precleared Hepes, 2 mM MnC12, 1 mM PMSF, 4 mM iodoacetic with Staphylococcus aureus (Pansorbin, Calbiochem) acid, 10 pM Na,VO,), followed by the addition of by incubation for 30 min at 4°C. The monoclonal phosphorylation buffer containing 10 KM y3'P-ATP antibody was added (1:lOO dilution) to the clarified (8 pCiinmo1). The phosphorylation reaction proceeded supernatants, and the mixture was incubated for 1 hr at 4°C for 20 min at which time 1 ml of PBS-DT was at 4°C. S. aureus and 1 Fg goat antimouse IgG antibody added. The phosphorylated EGF receptor was immunowas then added and the incubation was continued for precipitated. 30 min. The immunoprecipitate was then washed four Alkaline phosphatase treatment of times in PBS-DT, and the samples were analyzed on EGF receptors 6 1 3 %SDS-PAGE (Laemmli, 1970). The EGF receptor 35S-and 32P-labeledreceptors were immunoprecipiband was excised, and incorporated radioactivity was tated as described above. The immunoprecipitates were determined by liquid scintillation counting. washed three times in 30 mM Tris, pH 8 . 0 , l mM MgC1, and resuspended in 100 p1 of the same buffer. Ten units Phosphorylation of cytoskeletons of alkaline phosphatase was added, and the samples Cytoskeletons were prepared as described, except were incubated for 1 hr at 30°C. The reaction was that, after Triton extraction, the cytoskeletons were stopped by the addition of electrophoresis sample washed once with phosphorylation buffer (25 mM buffer, and the samples were separated by SDS-PAGE

66

ROY ET AL

on a 613% gradient. Radioactivity incorporated into on cytoskeletons isolated from cells incubated with 1251-EGF(Fig. 1, whole tx) and with unlabeled EGF the EGF receptor was determined as described. (Fig. 1, pretreated) was 5.9 x lo5 sitedcytoskeleton V8 proteinase digestion and 4.9 x lo5 sites/cytoskeleton, respectively. The of sites on cytoskeletons of untreated cells (Fig. A431 cells were metabolically labeled with T r a n ~ - ~ ~number S (50 kCi/ml) for 24 hr and then incubated in the absence 1, csk) was 1.56 x lo5 siteskytoskeleton. We did not or presence of 200 ngiml EGF for 15 min at 37°C. The reproducibly detect an alteration in the number of cells were extracted in Triton buffer for 2 min at P C , high-affinity sites, and since the number of highand the cytoskeletons were collected. The EGF receptor affinity sites present on cytoskeleton is small, they was immunoprecipitated as described above. The im- could not account for the difference in levels of 1251-EGF munoprecipitates were resuspended in 120 ~1 digestion binding. In this experiment, the cytoskeletons were buffer containing 0.125 M Tris, pH 6.0, 0.5% sodium isolated using buffer containing 0.5% Triton X-100, dodecyl sulfate (SDS), 10% glycerol, and 0-5 kg V8 since the difference between untreated and EGFproteinase as indicated (Cleveland et al., 1977). The treated cells was greater than when lower detergent samples were incubated for 30 min at 37% and then concentrations were used. Similar effects were observed using 0.15% Triton X-100. It should also be analyzed by SDS-PAGE on 8-15% gradient gels. noted that the removal of cell surface EGF by acid RESULTS treatment is only about 90% efficient, thus the number Treatment of cells with EGF decreases of binding sites measured on “pretreated” cells may be detergent-lability of 1251-EGFbinding sites an underestimate. These studies were initiated to extend our previous EGF treatment does not alter EGF receptor observation that the cytokeletons of untreated cells detergent insolubility retained fewer EGF binding sites than those of EGFWe predicted that the increase in 1251-EGFbinding treated cells. We used two different methods to examine this effect of EGF. In the first method (see Fig. 1,whole due to pretreatment of cells with EGF was a consetx), intact A431 cells were incubated in the absence quence of a greater number of EGF receptors associated (Fig. 1, csk) or presence of 20 ng/ml 1251-EGF(Fig. 1, with the detergent-insoluble cytoskeleton. To directly whole tx) for 4 hr at 4°C prior to extraction with varying measure the amount of EGF receptor associated with concentrations of Triton X-100 (Table 1).Alternatively, the cytoskeleton, A431 cells were metabolically labeled A431 cells were exposed to a saturating concentration with T r a n ~ - ~for~ S 18 hr. Cytoskeletons from control (200 ngiml) of unlabeled EGF for 2 hr at 4”C, and then and EGF-treated (200 ng/ml, 15 min, 37°C) cells were bound ligand was removed by treatment with 0.2 M prepared, and the 35S-labeledEGF receptor was immuacetic acid (Haigler et al., 1980) before extraction with noprecipitated from both the detergent-soluble and Triton buffer (Fig. 1,pretreated). Two hours of pretreat- -insoluble fractions and analyzed by SDS-PAGE. There was no difference in the amount of 35S-labeled ment with unlabeled EGF was sufficient to induce the maximum increase in binding (data not shown). The EGF receptors associated with the cytoskeletons of resulting cytoskeletons were then incubated with 1251- control and EGF-treated cells (Table 2). Extraction of EGF. EGF pretreatment resulted in increased levels of A431 cells with 0.15% or 0.5% Triton solubilized 10% EGF binding associated with the cytoskeleton relative and 25% of the EGF receptors, res ectively. These data to untreated cells (Table 1).Extraction of the cells with indicate that the differences in 2! I-EGF binding sites increasing Triton concentrations revealed greater detected following EGF treatment of cells was not a numbers of cytoskeletally associated 1251-EGFbinding result of differential solubilization of the EGF recepsites on EGF-treated cells compared with those that tors. had not been exposed to EGF. There was a progressive EGF-induced alteration of the electrophoretic loss of 1251-EGFbinding sites with increasing detergent mobility of the cytoskeletally associated concentrations; however, the detergent preferentially EGF receptor reduced the numbers of binding sites from untreated cells. The use of detergent concentrations higher than EGF-treated (200 ng/ml, 37”C, 15 min) and control were ~S 0.5% did not significantly increase the detergent labil- A431 cells metabolically labeled with T r a n ~ - ~ ity of EGF binding sites. These data suggest that EGF extracted with Triton X-100, and the EGF rece tor was treatment of intact cells results in the detection of a immunoprecipitated from the detergent-solu73le and -insoluble fractions for analysis by SDS-PAGE. Inspecgreater number of binding sites on the cytoskeleton. tion of the autoradiogram revealed the electrophoretic EGF-induced retention of binding sites on mobility of the EGF receptor band was altered in both cytoskeletons is due to a greater number of the soluble and insoluble fractions of EGF-treated cells, low-affinity sites compared with control cell fractions (Fig. 3). The 35SWe wished to determine if the increase in binding on EGF receptor derived from untreated cells appeared as EGF pretreatment was due to a change in the number a doublet of M, 160-180 kDa of equal intensity, of low- or high-affinity EGF binding sites. Scatchard whereas the EGF receptor band immunoprecipitated analysis of steady-state binding revealed that the dif- from EGF-treated cells migrated as a single sharp band ference was due principally to a greater number of low at 180 kDa. The altered mobility of the EGF receptors affinity sites (Fig. 2A,B). The total number of low from pretreated cells corresponded to a 40% increase in affinity EGF binding sites on intact cells was the amount of labeled EGF receptors present in the 1.85 X lo6 sitesicell. The number of low-affinity sites upper (180 kDa) band and a concomitant decrease in

CYTOSKELETALLYASSOCIATEDEGFRECEPTOR

67

TABLE 1. Detergent-insoluble Iz5I-EGF binding' lZ5I-EGF incubation

Intact cells

0.15% Triton

71 f 2 42 f 2 35 f 4

0.54; Triton

1.0% Triton

Detergent-insolubleEGF-receptor (4;) Cytoskeletons of Cytoskeletons of EGF-pretreated cells untreated cells

*

68 k 11 21 2 17k2

46 11 9+2 7+1

*

'Detergentinsoluble '"I-EGFbinding was measuredas described in Figure 1 in the cytoskeletons prepared from intact cells incubated with 20 ng/ml'251-EGF (whole,ts);intactcells pretreatedwith unlabeled EGF, followed b extraction with Triton X-100and exposure to 20ng/ml '%EGF (pretreated);untreated cells extracted prior to exposure of rz51-EGF (csk).The data are expressed as the mean percent (+sd) of total '"I-EGF binding measured on intact cells (whole).

a

m 0

0

1 2 3 EGF Bound (prnoles/106 cells)

4

0

0.5

1

EGF Bound (pmoles/l O6 cells)

Fig. 2. Binding of '"11-EGF to cells and cytoskeletons. Binding of lZ51-EGFwas performed exactly as described in Materials and Methods. Cytoskeletons were prepared using 0.5% Triton X-100.A: Scatchard analysis of binding to cells and cytoskeletons. B: The same data as shown in A, but the scale has been changed to better illustrate the

binding data. 0,lZ51-EGFbinding to intact cells. m, cytoskeletons of cells incubated with lZ51-EGF prior to extraction (whole tx). A , '"1-EGF binding to cytoskeletons of cells pretreated for 2 hr at 4°C with unlabeled EGF (pretreated). , IZ5I-EGFbinding to cytoskeletons prepared from untreated cells (csk).

+

TABLE 2. EGF pretreatment does not alter EGF receptor detergent solubility' Experiment 0.15%Triton 1 2

Mean 0.5% Triton 1

2 3 4 5

Mean

Soluble EGF-receptor Control Pretreated 10.1 10.4 10.25 27.3 23.2 20.0 31.2 28.2 26.0

+ 0.15

+ 4.4

Cytoskeletal EGF-receptor Control Pretreated

14.2 6.1 10.2 & 4.1

89.9 89.6 89.8 k 0.15

85.7 93.9 89.8 & 4.1

32.0 24.8 31.0 17.8 24.2 26.0 k 5.2

72.7 76.8 80.0 68.8 71.1 73.9 k 4.0

68.0 76.2 69.0 82.8 75.9 74.3 f 5.2

'Detergent solubility ofthe EGF receptor in controland pretreated cells. A431 cells weremetabolically labeled with Trans.% for 18 hr. Shown is the amount ofimrnunoprecipitableEGF receptors present in the detergent-solubleand insoluble fractions of untreated and EGF-pretreated(200 ng/ml, 15 min, 37°C) cells. The data are expressed as the mean percent (ksd) of total cellular "S-labeled EGF receptor.

the lower (160 kDa) band (Fig. 3B). The magnitude of the shift to a lower mobility species was EGF dose dependent (data not shown). The EGF-dependent alteration of EGF receptor mobility was also observed when the pretreatment of cells with EGF was carried out for 2 hr at 4°C (data not shown). Interestingly, addition of EGF to cytoskeletons of untreated cells did not produce a change in mobility, indicating that treatment of intact cells is required (data not shown). The EGF-induced alteration in receptor mobility was not an artifact of our extraction methods, since the

change in receptor migration was seen in EGF receptors immunoprecipitated from untreated and treated intact cells (data not shown). Also, we were particularly concerned with suppression of proteinase activity; the inclusion of EGTA and iodoacetic acid during the lysis step minimized the possibility that the appearance of the 160 kDa band was due to proteolysis of the 180 kDa species, a result of calcium-activated neutral proteinase activity (Gates and King, 1985; King and Gates, 1985; Stoscheck et al., 1988). An extensive series of control experiments were performed in which leupep-

ROY ET AL

68

a

Detergent

Detergent S o h ble

Insoluble

180

2

-160

B

7 I

a

T

b

tin, aprotonin, soybean trypsin inhibitor, pepstatin, bacitracin, and PMSF were included in the lysis buffer; however, there was no effect of these agents on the mobility of the EGF receptor bands (data not shown). These data demonstrate that the change in receptor mobility is a consequence of physiological events occurring rapidly following EGF binding to its receptor. Alteration of electrophoretic mobility is an EGF-specific event To determine if phosphorylation of the EGF receptor at serinekhreonine residues could also result in a shift in mobility, we examined the effect of 12-0-tetradecanoylphorbol 13-acetate (TPA) on the electrophoretic mobility of the EGF receptor. A431 cells metabolically labeled with T r a n ~ - ~were ~ S treated with EGF (200 ng/ml) or TPA (100 nglml) for 15 min at 37°C. The EGF receptors were immunoprecipitated from the detergent-insoluble fractions of control and EGF- and TPAtreated cells. Although TPA treatment induces extensive phosphorylation of the EGF receptor on serine and threonine residues (Friedman et al., 1984; Davis and Czech, 1985; Heisermann and Gill, 1988; Hunter et al., 1984) it did not result in a change in electrophoretic mobility (Fig. 4).These data demonstrate that the shift in mobility is an EGF-specific event and is not simply due to enhanced receptor phosphorylation. Alteration in mobility is not reversed by dephosphorylation If receptor phosphorylation was responsible for the shift in electrophoretic mobility of the EGF receptor,

Fig. 3. Alteration of the electrophoretic mobility of the EGF receptor. Control and EGF-treated (200 ngiml, 15 min, 37°C) A431 cells (metabolically labeled with T r a n ~ - ~ were ~ S ) extracted with Triton X-100. A:An autoradiogram of labeled EGF receptors immunoprecipitated as described in Materials and Methods and analyzed by SDS-PAGE on 6 1 3 % gels. B: The fractional percentage of the radioactivity incorporated into the upper (open bars, 180 kDa) and the lower (hatched bars, 160 kDa) bands of the EGF receptors in both control (a) and pretreated (b) samples prepared by extraction with Triton.

then dephosphorylation of the receptor should reverse the shift. Immunoprecipitated 35S-labeled and 32Plabeled EGF receptors were treated with alkaline phosphatase and then analyzed b SDS-PAGE. Treatment with alkaline phosphatase di not result in the reversal of the electrophoretic shift of the EGF receptors immunoprecipitated from EGF-treated cells (Fig. 5A). The percentages of total 35S-labeled EGF receptors in the upper band for EGF-treated and EGF plus alkaline phosphatase samples were 67% and 66%, respectively. The efficiency of receptor dephosphorylation by alkaline phosphatase was measured in parallel. The EGF receptor was phosphorylated in situ by incubation with 32P-ATP, followed by treatment with alkaline phosphatase. Alkaline phos hatase treatment resulted in the removal of 85% of”P0, from the EGF receptor (Fig. 5B). The inability to reverse the shift in mobility by dephosphorylation indicates that the altered mobility was not due primarily to phosphorylation. However, its contribution to the altered mobility cannot be ruled out. Additional ex eriments with two- to threefold higher levels of a1 aline phosphatase also failed to reverse the shift in mobility. EGF pretreatment alters the susceptibility of the EGF receptor to V8 proteinase If the EGF-induced alteration in receptor mobility reflects a conformational change in the receptor, we reasoned that this may be reflected in the susceptibility of the receptor to proteinase digestion. The EGF receptor immunoprecipitation from control and EGF-treated cells were incubated with V8 proteinase. The autora-

B

K

69

CYTOSKELETALLYASSOCIATEDEGFRECEPTOR

A

CON

EGF

35s

TPA CON

Fig. 4. Effect of phorbol ester treatment on gel shift. Cytoskeletons were isolated from metabolically labeled A431 cells using 0.15% Triton buffer. EGF receptors were immunoprecipitated from the detergent-insoluble fraction of control, EGF-treated (200 ng/ml, 15 min, 37"C), and TPA-treated (100 ngiml, 15 min, 37°C) and analyzed by SDS-PAGE.

diograph revealed that the 180 kDa species was preferentially digested to a species of approximately 160 kDa species whose mobility was slightly lower than that of the 160 kDa band present in immunoprecipitations of control cells (Fig. 6). The digests from control and EGF-treated cells were otherwise qualitatively similar. The lower molecular weight bands apparent in higher V8 concentrations were not reproductively observed. These data indicate that the 180 kDa EGF receptor species is preferentially digested by V8 proteinase relative to the 160 kDa band indicating that the EGF receptor is in a conformation rendering this receptor species more susceptible to enzymatic degradation.

DISCUSSION The present study was initiated following our observation that there was a significant disparity in the number of EGF binding sites detected on the cytoskeletons of EGF-treated A431 cells compared to untreated cells. We originally predicted that the enhanced binding exhibited by EGF-pretreated cells was a consequence of ligand-induced association of the receptor with the cytoskeleton, as has been shown for the nerve growth factor receptor (Schechter and Bothwell, 1981; Vale et al., 1985; Vale and Shooter, 1982). Recently, van Bergen en Henegouwen et al. (1989) have reported a similar EGF-dependent increase in EGF binding to cytoskeletally associated receptors. They interpreted the data as evidence for a ligand-dependent association of the receptor with the cytoskeleton. Surprisingly, we failed to detect an increase in the amount of immunoprecipitable 35S-EGFreceptors in the detergent-insoluble fractions of pretreated cells when compared to that of control cells. This result contrasted sharply with binding data, where the cytoskeletons of EGF-treated cells had twice the number of binding sites cytoskeletons of control cells. The dramatic reduction in binding activity was shown not to be due principally to solubilization of the EGF receptor but rather a consequence of the inability of the cytoskeletally associated EGF receptor to bind ligand. The data suggest that receptor occupancy preserves the capacity of the receptors to bind ligand

EGF

EGF

+

AP

B 32P

CON

EGF E G F

+

AP Fig. 5. Alkaline phosphatase treatment of EGF receptors. Immunoprecipitated EGF receptors from the detergent-insoluble fractions of control and EGF-treated cells were treated with alkaline phosphatase as described in Materials and Methods. A: EGF receptors from cells metabolically labeled with T r a n ~ - ~B: ~S . receptors from cytoskelEGF etons phosphorylated in situ with 32P-ATP. Radioactivity incorporated into the phosphorylated EGF receptor was 300 cpm for control, 530 for EGF, and 70 for EGF plus AP. AP, alkaline phosphatase.

following detergent treatment. Triton treatment apparently inactivates the binding activity of a fraction of the low-affinity EGF receptors, resulting in an apparent preferential retention of high-affinity sites. This is a paradoxical effect of Triton on the cytoskeletally associated EGF receptor, since a variety of studies have extensively characterized the EGF receptor binding activity and kinase activity using detergent-solubilized preparations. A possible mechanism by which EGF protects low-affinity EGF receptors is through the formation of receptor dimers on ligand binding (Yarden and Schlessinger, 1987a,b;Cochet et al., 1988). It seems likely that the Triton extraction of the cells prevents the subsequent interaction of the receptors due to their immobilization through their association with cytoskeletal elements. It is of particular interest that the EGF-induced change in mobility and increase in binding occurred under the same conditions, although it remains unclear

ROY ET AL.

70

180% 160'

0 EGF' 0 EGF No 1'

0

I

0 EGFl 0 0.1

EGF' 0

€OF1 0

0.2 5

1

EOF 5 P9 V8

Fig. 6. V8 Proteinase digestion of EGF receptors from control and EGF-treated cells. A431 cells were by incubation in the absence or presence of 200 ngiml EGF. The EGF labeled with T r a n ~ - ~followed ~S receptors were immunoprecipitated and then digested for 30 min with the indicated concentration of V8 proteinase. The digests were separated on 8-15% SDS-PAGE. An autoradiograph is shown. The first two lanes are from immunoprecipitates in which no EGF receptor antibody was added.

whether there is a direct relationship between these two events. The diminished mobility of the EGF receptor was not wholly due to binding of EGF in that the 180 kDa band was detected, albeit at lower levels, in control cells. The data demonstrate that treatment of intact cells was required to produce the shift as addition of EGF to cytoskeletons of untreated cells had no effect on receptor mobility. The latter result is consistent with our previous observation that EGF treatment of cytoskeletons was unable to activate the receptor tyrosine kinase activity (Roy et al., 1989). This observation suggests that receptor occupancy alone is insufficient to produce the shift in receptor mobility and must involve modification of the receptor within intact cells. We investigated the possibility of receptor autophosphorylation being responsible for altered EGF receptor mobility. Chinkers and Garbers (1986) noted that the EGF receptor of EGF-treated A431 cells migrated more slowly than the receptors from untreated cells, which they resumed to be a result of increased phosphorylation.3! hifts in electrophoretic mobility due to increased phosphorylation have also been reported for other proteins, including the protein products of two oncogenes fos and src (Barber and Verma, 1987; Gould and Hunter, 1988). However, TPA, which induces the extensive phosphorylation of serine and threonine residues on the EGF receptor (Cochet et al., 1984; Friedman et al., 1984; Hunter et al., 1984; Iwashita and Fox, 19841, did not alter the mobility of the receptor. Furthermore, the failure of alkaline phosphatase to reverse the EGF-induced change in receptor mobility suggests that receptor phosphorylation cannot entirely account for the altered electrophoretic mobility of the receptor. The shift in electrophoretic mobility of the EGF receptor may result from an EGF-dependent conformational change in the receptor. The data demonstrate that the 180 kDa species was preferentially digested by

V8 proteinase relative to the 160 kDa species, suggesting that EGF-receptor occupancy elicits a conformational change in the receptor rendering it more susceptible to proteinase digestion. It is perplexing that these changes persist under the conditions used for immunoprecipitations and SDS-PAGE. Nevertheless, recent work by Greenfield et al. (1989) showed that binding of EGF to its receptor induces a change in the conformation of the extracellular domain. It remains a formal possibility that the EGF receptor oligomers persist in the immunoprecipitates and the differential susceptibility to V8 digestion is a result of such receptor interactions. The increased susceptibility of the bound receptor to proteinase may be relevant to the intracellular processing of the hormone-receptor complex. We favor a hypothesis in which the EGF receptor is normally associated with cytoskeletal elements rendering it detergent insoluble. Upon EGF binding, the receptor undergoes a conformational change that is correlated with the formation of receptor oligomers, an altered electrophoretic mobility, and susceptibility to proteinase digestion. We suggest that the occupied receptor oligomers retain their capacity to bind ligand, whereas the binding activity of unoccupied receptor is more highly sensitive to detergent treatment. What was previously interpreted as preferential retention of high-affinity receptors with the cytoskeleton in fact reflects the differential sensitivity of the two receptor species to inactivation of their binding activity by Triton X-100. The nature of the change in the occupied EGF receptor remains unclear. The altered mobility of the receptor is not likely to be due to receptor phosphorylation, nor is it a simple consequence of receptor occupancy. The nature of the EGF-induced change in the receptor remains unclear and is likely a reflection of the complex intramolecular interactions of the receptor.

CYTOSKELETALLYASSOCIATED EGF RECEPTOR

ACKNOWLEDGMENTS We thank Joan Kingsley for assistance in the preparation of the manuscript. L.M.R. was supported by a predoctoral training fellowshiu from the Medical Uniiersity of South Caiolina. This work was supported by NIH grant GM 34908.

71

induces multisite phosphorylation of pp6OC-"" and increases its protein-tyrosine kinase activity. Mol. Cell. Biol., 83345-3356. Greenfield, C., Hiles, I., Waterfield, M.D., Federwisch, M., Wollmer, A., Blundell, T.L., and McDonald, N. (1989) Epidermal growth factor binding induces a conformational change in the external domain of its receptor. EMBO J., 8:41154123. Gullick, W.J., Downward, D.J.H., Marsden, J.J., and Waterfield, M.D. (1984) A radioimmunoassay for human epidermal growth factor receptor. Anal. Biochem., 141:253-261. Haigler, H.T., Maxfield, F.R., Willingham, M.C., and Pastan, I. (1980) LITERATURE CITED Dansylcadaverine inhibits internalization of '2sI-epidermal growth Barber, J.K., and Verma, I.M. (1987) Modification of fos proteins: factor in BALB 3T3 cells. J. Biol. Chem., 255:1239-1241. Phosphorylation of c-fos, but not v-fos, is stimulated by 12-tetrade- Heisermann, G.J., and Gill, G.N. (1988) Epidermal growth factor canoyl-phorbol-13-acetateand serum. Mol. Cell Biol., 72201-2211. receptor threonine and serine residues phosphorylated in vivo. J. Bradford, M.M. (1976) A rapid and sensitive method for the quantiBiol. Chem., 263:13152-13158. tation of microgram quantities of protein utilizing the principle of Honegger, A.M., Dull, T.J., Felder, S., Van Obberghen, E., Bellot, F., protein-dye binding. Anal. Biochem., 72:248-254. Szapary, D., Schmidt, A., Ullrich, A,, and Schlessinger, J. (1987a) Buhrow, S.A., Cohen, S., Garbers, D.L., and Staros, J.V. (1983) Point mutation at the ATP binding site of EGF receptor abolishes Characterization of the interaction of 5'-p-fluorosulfonyl-benzoyl protein-tyrosine kinase activity and alters cellular routing. Cell, adenosine with the epidermal growth factor receptoriprotein kinase 51 :199-209. in A431 cell membranes. J. Biol. Chem., 258:782&7827. Honegger, A.M., Szapary, D., Schmidt, A., Lyall, R., Van Obberghen, Burr, J.G., Dreyfuss, G., Penman, S., and Buchanan, J.M. (1980) E., Dull, T.J., Ullrich, A., and Schlessinger, J . (1987b) A mutant Association of the src gene product of Rous sarcoma virus with epidermal growth factor receptor with defective protein tyrosine kinase is unable to stimulate proto-oncogene expression and DNA cytoskeletal structures of chicken embryo fibroblasts. Proc. Natl. synthesis. Mol. Cell. Biol., 7:45684571. Acad. Sci. USA, 77:3484-3488. Carpenter, G., and Cohen, S.(1979) Epidermal growth factor. Annu. Hunter, T., Ling, N., and Cooper, J.A. (1984) Protein kinase C Rev. Biochem., 48:193-216. phosphorylation of the EGF receptor at a threonine residue close to Carpenter, G., King, L. Jr., and Cohen, S. (1979) Rapid enhancement the cytoplasmic face of the plasma membrane. Nature, 311:480-483. of protein phosphorylation in A-431 cell membrane preparations by Iwashita, S., and Fox, C.F. (1984)Epidermal growth factor and potent epidermal growth factor. J. Biol. Chem., 254:48844891. phorbol tumor promoters induce epidermal growth factor receptor phosphorylation in a similar but distinctively different manner in Chatelier, R.C., Ashcroft, R.G., Lloyd, C.J., Nice, E.C., Whitehead, R.H., Sawyer, W.H., and Burgess, A.W. (1986) Binding of fluoreshuman epidermoid carcinoma A431 cells. J. Biol. Chem.,259:25592567. ceinated epidermal growth factor to A431 cell sub-populations studied using a model-independent analysis of flow cytometric Kawamoto, T., Sato, J.D., Le, A., Polikoff, J . , Sato, G.H., and Mendelsohn, J . (1983) Growth stimulation of A431 cells by epidermal fluorescence data. EMBO J . , 5:1181-1186. growth factor: identification of high-affinity receptors for epidermal Chen, W.S., Lazar, C.S., Poenie, M., Tsien, R.Y., Gill, G.N., and growth factor by an anti-receptor monoclonal antibody. Proc. Natl. Rosenfeld, M.G. (1987) Requirement for intrinsic tyrosine kinase Acad. Sci. USA, 80:1337-1341. activity in the immediate and late actions of the EGF receptor. King, A.C., and Cuatrecasas, P. (1982) Resolution of high and low Nature, 328:820-823. affinity epidermal growth factor receptors. J. Biol. Chem.. Chinkers, M., and Garbers, D.L. (1986) Suppression of protein ty2573053-3060. rosine kinase activity of the epidermal growth factor receptor by King, L.E., and Gates, R.E. (1985)Calcium-activated neutral protease epidermal growth factor. J. Biol. Chem., 261:8295-8297. purified from beef lung: Properties and use in defining structure of Cleveland, D.N., Fischer, S.G., Kirschner, M.W., Laemmli, U.K. epidermal growth factor receptors. Arch. Biochem. Biophys., (1977) Peptide mapping by limited proteolysis in sodium dodecyl 2424:14&156. sulfate and analysis gel electrophoresis. J . Biol. Chem., 252:1102Laemmli, U.K. (1970) Cleavage of structural proteins during the 1106. assembly of the head of bacteriophage T,. Nature, 227:680-685. Cochet, C., Gill, G.N., Meisenhelder, J., Cooper, J.A., and Hunter, T. (1984) C-kinase phosphorylates the epidermal growth factor recep- Lam, C.F. (1970) Computer analysis and optimization of a class of non-linear systems. PhD Dissertation, Clemson University. tor and reduces its epidermal growth factor-stimulated tyrosine Landreth, G.E.,Williams, L.K., and Reiser, G.D. (1985)Association of protein kinase activity. J. Biol. Chem., 259:2553-2558. the epidermal growth factor receptor kinase with the detergentCochet, C., Kashles, O., Chambaz, E.M., Borello, I., King, C.R., and insoluble cytoskeleton of A431 cells. J. Cell Biol., 102:1341-1350. Schlessinger, J. (1988) Demonstration of epidermal growth factorinduced receptor dimerization in living cells using a chemical Livneh, E., Prywes, R., Kashles, O., Reiss, N., Sasson, I., Mory, Y ., Ullrich, A,, and Schlessinger, J . (1986) Reconstitution of human covalent cross-linking agent. J . Biol. Chem.. 263:3290-3295. epidermal growth factor receptors and its deletion mutants in Cohen, S.(1987) Epidermal growth factor. In Vitro, 23:239-246. cultured hamster cells. J. Biol. Chem., 261:12490-12497. Cohen, S.,Carpenter, G., and King, L. (1980) Epidermal growth factor receptor-protein kinase interactions: Copurification of receptor and Livneh, E., Reiss, N., Berent, E., Ullrich, A,, and Schlessinger, J. (1987) An insertional mutant of epidermal growth factor receptor epidermal growth factor-enhanced phosphorylation activity. J . Biol. allows dissection of diverse receptor function. EMBO J., 62669Chem., 2554834-4842. 2676. Davis, R.J., and Czech, M.P. (1985) Tumor-promoting phorbol diesters cause the phosphorylation of epidermal growth factor receptors in Marchalonis, J.J. (1969) An enzymic method for the trace iodination of immunoglobulins and other proteins. Biochem. J., 113:299-305. normal human fibroblasts a t threonine-654. Proc. Natl. Acad. Sci. Moolenaar, W.H., Bierman, A.J., Tilly, B.C., Verlaan, I., Defize, USA, 82:1974-1978. L.H.K., Honegger, A.M., Ullrich, A., and Schlessinger, J. (1988) A Defize, L.H.K., Boonstra, J.,Meisenhelder, J., Kruijer, W., Tertoolen, point mutation a t the ATP-binding site of the EGF-receptor abolL.G.J., Tilly, B.C., Hunter, T., van Bergen en Henegouwen, P.M.P., ishes signal transduction. EMBO J., 7707-710. Moolenaar, W.H., and de Laat, S.W. (1989) Signal transduction by epidermal growth factor occurs through the subclass of high affinity Nelder, J.A., and Mead, R. (1965) A simplex method for function minimization. Comput. J., 7:308-313. receptors. J . Cell Biol., 109:2495-2507. Downward, J., Yarden, Y., Mayes, E., Scrace, G., Totty, N., Stockwell, Prywes, R., Livneh, E., Ullrich, A,, and Schlessinger, J. (1986) Mutations in the cytoplasmic domain of the EGF receptor affect P., Ullrich, A,, Schlessinger, J., and Waterfield, M.D.(1984) Close EGF binding and receptor internalization. EMBO J., 5:2179-2190. similarity of epidermal growth factor receptor and v-erb-B oncogene Rees, A.R., Gregariou, M., Johnson, P., and Garland, P.B. (1984)High protein sequences. Nature, 307521-526. affinity epidermal growth factor receptors on the surface of A431 Friedman, B.A., Frackelton, A.R. Jr., Ross, A.H., Connors, J.M., cells have restricted lateral diffusion. EMBO J., 3:1843-1847. Fujiki, H., Sugimura, T., and Rosner, M.R. (1984) Tumor promoters block tyrosine-specific phosphorylation of the epidermal growth Roy, L.M., Gittinger, C.K., and Landreth, G.E. (1989) Characterization of the EGF receptor associated with the cytoskeletons of A431 factor receptor. Proc. Natl. Acad. Sci. USA, 81:3034-3038. cells. J . Cell. Physiol., 140:295-305. Gates, R.E., and King, L.E. (1985) Different forms of the epidermal growth factor receptor kinase have different autophosphorylation Russo, M.W., Lukas, T.J., Cohen, S., and Staros, J.V. (1985) Identification of residues in the nucleotide binding site of the epidermal sites. Biochemistry, 24:5209-5215. growth factor receptorikinase. J . Biol. Chem., 2605205-5208. Gould, K.L., and Hunter, T. (1988) Platelet-derived growth factor

72

ROY ET AL.

Scatchard, G. (1949) The attractions of proteins for small molecules and ions. Ann. N.Y. Acad Sci., 51:660-672. Schechter, A.L., and Bothwell, M.A. (1981) Nerve growth factor receptors on PC12 cells: Evidence for two receptor classes with differing cytoskeletal association. Cell, 24:867-874. Schlessinger, J. (1988) The epidermal growth factor receptor as a multifunctional allosteric protein. Biochemistry, 273119-3123. Stoscheck, C.M., Gates, R.E., and King, L.E. (1988) A search for EGF-elicited degradation products of the EGF receptor. J. Cell. Biochem., 38:51-63. Tuszynski, G., Knight, L., P$xrno, J., and Walsh, P. (1980) A rapid method for removal of l2 Iodide following iodination of protein solutions. Anal. Biochem., 106118-122. Ullrich, A,, Coussens, L., Hayflick, J.S., Dull, T.J., Gray, A., Tam, A.W ., Lee, J.,Yarden, Y ., Libermann, T.A., Schlessinger, J., Downward, J.,Mayes, E.L.V., Whittle, N., Waterfield, M.D., andseeburg, P.H. (1984) Human epidermal growth factor receptor cDNA sequence and aberrant expression of the amplified gene in A431 epidermoid carcinoma cells. Nature, 309:418425. Ushiro, H., and Cohen, S. (1980) Identification of phosphotyrosine as a product of epidermal growth factor-activated protein kinase in A-431 cell membranes. J. Biol. Chem., 255:8363-8365.

Vale, R.D., Ignatius, M.J., and Shooter, E.M. (1985) Association of nerve growth factor receptors with the Triton X-100 cytoskeleton of PC12 cells. J. Neurosci., 5:2762-2770. Vale, R.D.,and Shooter, E.M. (1982) Alteration of binding properties and cytoskeletal attachment of nerve growth factor receptors in PC12 cells by wheat germ agglutinin. J. Cell Biol., 94:710-717. van Bergen en Henogouwen, P., Defize, L., deKroon, J., van Damme, H., Verkleij, A., and Boonstra, J. (1989) Ligand-induced association of epidermal growth factor receptor to the cytoskeleton of A431 cells. J. Cell Biochem., 39:455-465. Weigant, F.C., Blok, F.J., Defize, L.J.K., Linnemans, W.A.M., Verkley, A.J., and Boonstra, J. (1986) Epidermal growth factor receptors associated to cytoskeletal elements of epidermoid carcinoma (A4311 cells. J. Cell Biol., 103237-94. Yarden, Y., and Schlessinger, J. (1987a) Self-phosphorylation of epidermal growth factor receptor: Evidence for a model of intermolecular allosteric activation. Biochemistry, 261434-1442. Yarden, Y., and Schlessinger, J. (1987b) Epidermal growth factor induces rapid, reversible aggregation of the purified epidermal growth factor receptor. Biochemistry, 26:1443-1451.