JOURNAL OF CELLULAR PHYSIOLOGY 144:303-312 (1990)

Cellular Distribution and Biological Activity of Epidermal Growth Factor Receptors in A431 Cells Are Influenced by Cell-Cell Contact ROSEMARIE B. LICHTNER* AND VOLKER SCHIRRMACHER Department of immunology and Genetics, German Cancer Re5earch Center, 6900 Herdelberg, Federal Republrc of Germany The potential significance of cell-cell interactions on EGF receptor (EGFR) activity was investigated in cultured adherent A431 cells seeded as single-cell suspensions with different initial cell densities In dense cultures, EGFRs were mainly localised at cell boundaries and in microvilli as shown by immunofluorescence analysis with an EGFR-specific antibody while in sparse cultures the distribution of EGFRs was more diffuse Scatchard analysis showed that as cell density decreased the number of high-affinity receptors increased considerably Upon treatment of adherent intact cells with EGF all cclls in sparse cultures contained activakd ECFKs as demonstrated by immunofluorescence analysis with a phosphotyrosine-specific antibody, while in dense cultures mainly cells at the periphery of a cluster and especially at their expanding borders exhibited activated EGFRs. EGF-induced phosphorylation in intact cells was greatly enhanced in sparse compared with dense cultures as demonstrated by immunoprecipitation with a phosphotyrosine-specific antibody. In contrast to intact cells, in cytoskeleton preparations, obtained after mild detergent treatment of adherent cells, EGFRs were able to undergo EGF-independent pho3phoryldtion Pretreatment of cells with EGF led to enhanced tyrosine phosphorylation of cytoskeletal-associated proteins Our observations suggest that cell density hds a considerable effect on the subcellular localisation a5 well as biological activity of the EGFR. Thus, in intact A431 cells growing with extensive cell-cell interactions some negative control mechanisms preventing ECFR activation may be exerted by adjacent cells.

Epidermal growth factor (EGF)mediates its mitogenic response through interaction with a specific membrane receptor. The binding of EGF to a n extracellular domain of the EGF receptor (EGFR) activates its cytoplasmic tyrosine kinase which phosphorylates various cellular proteins and the receptor itself (Hunter and Cooper, 1981; Downward et al., 1984). EGF induces morphological changes in cultured cells, such as rounding up of cells and induction of membrane ruffling and extension of filopodia (Chinkers e t al., 1979, 1981; Schlessinger and Geiger, 1981). There have been contradictory reports concerning the distribution of EGFRs on the surface of A431 cells. In a recent investigation, the highest density of EGFR was detected a t the cell periphery (van Belzen e t al., 1988). This contrasts with earlier studies which had indicated a random distribution of EGFRs on the cell surface (Schlessinger et al., 1978; Haigler e t al., 1979; Boonstra et al., 1985). Addition of EGF has been reported to cause preferential phosphorylation of EGFRs in microvilli and membrane ruffles (Carpentier et al., 1987) as well as the phosphorylation of other cytoskeleton-associated proteins (Landreth e t al., 1985; Carpentier e t al., 1987; Bretscher, 1989). Since morphology and dynamics of 0 1990 WILEY-LISS, INC.

the cell are largely maintained by a n integrated action of cytoskeletal systems (Ben Ze’ev, 1985), a direct or indirect coupling of the receptor to the cytoskeleton could be expected. Indeed, a structural association of EGFRs and especially of the high-affinity class of EGFR to the cytoskeleton of Triton X-100 extracted cultured A431 cells was reported (Wiegant et al., 1986; van Bergen en Henegouwen et al., 1989; Roy et al., 1989). Cytoskeletal structures are involved in the establishment of cell-cell contacts and are associated with the adherent class of intercellular junctions (Staehelin, 1974). Thus, the degree of cell-cell contacts could have effects on the properties of EGFRs associated with the cytoskeleton. Studies using the fluorescence photobleaching recovery method have shown that high-affinity receptors in sparse cultures of A431 cells exhibit restricted lateral diffusion a s compared to that of the low-affinity receptors expressed on confluent A431

Received October 18, 1989; accepted April 24, 1990. *To whom reprint requestskorrespondence should be addressed.

304

LICHTNER AND SCHIRRMACHER

cells (Rees et al., 1984; Hillmann and Schlessinger, 1982). This would suggest a cell-density-dependent interaction between the EGFR and the underlying structural elements. Furthermore, both growth factor binding (Rizzino et al., 1988) and mitogenic response to growth factor (Paulsson et al., 1988) were reported recently to be modulated by cell density. A431 cells provide the most extensively studied model of EGF action but most of these investigations have been performed on subconfluent or even confluent cultures. I n the present study we have investigated the effect of cell density on the distribution and kinase activity of the EGFR. Both intact growing and adherent detergent permeabilized (cytoskeletal preparation) A431 cells have been examined. Our results suggest that the density of cultured cells has a considerable influence on the subcellular distribution and biologial activity of the EGFR. MATERIALS AND METHODS Materials [y-”P] ATP (3,000 Ciimmol), lz5I-EGF (100 mCi/ mmol), and 35S-methionine (1,300 Ci/mmol) were purchased from Amersham Buchler (Braunschweig, FRG) and H,32P0, (9,000 Ciimmol) was from NEN (Dreieich, FRG). EGF, receptor grade, was supplied by Collaborative Research (Lexington, MA, USA). Prestained molecular weight standards for gel electrophoresis were from Sigma (Munich, FRG) and Protein A coupled to Sepharose was from Pharmacia (Freiburg, FRG). DME free of phosphate was obtained from Biochrom (Berlin, FRG). Antibodies IG2, a monoclonal antibody (Mab) against phosphotyrosine (aPTyr),was obtained from Oncogene Sciences (Manhasset, NY, USA), while a Mab against the ligand-binding domain of the EGF receptor (EGFR1) (Waterfield et al., 1982) was obtained from Amersham Buchler. Cells Human epidermoid carcinoma cells (A431) were a gift from Dr. G.E. Gallick (M.D. Anderson Hospital, Houston, TX, USA) and they were cultured in a 1 : l mixture of Dulbecco’s-modified Eagle’s (DME) (Flow, Meckenheim, FRG) and F12 medium (Gibco, Eggenstein, FRG), supplemented with 10% non-heat-inactivated fetal calf serum (FCS) (Gibco). Monolayers of A431 cells were grown in 100 mm culture dishes (Corning).At 6 0 4 0 % confluency, the tumor cells were harvested by using 0.125% trypsin in 2 mM EDTA in calcium-magnesium-free Dulbecco’s phosphatebuffered saline (trypsin/EDTA) and diluted with DME/ F12 containing 1%of FCS. Usually cells were plated in 100 mm dishes a t 2 x lo5 (2,550 cells/cm2) or at 2 x lo6 (25,500 cells/cm2). Binding of 1251-EGFto A431 cells A431 cells were plated a t the indicated cell numbers in 1ml of medium with 1%FCS in 16 mm well plates and allowed to attach for 18-20 h. After this time the cell numbers had increased 10-20%. Cell monolayers were rinsed once with ice-cold PBS containing 2 mg/ml

bovine serum albumin (BSA) and incubated in this medium for 30 min a t 4°C. Subsequently, cells were incubated in PBSiBSA with 1251-EGF (appr. 20,000 cpm) and increasing concentrations of unlabeled EGF (0.25150 ngiml) for 1 h at 4°C. After this time, monolayers were washed three times with PBSiBSA and solubilized in 10 mM NaOHi1% SDS. The bound radioactivity was measured in a gamma spectrometer. Nonspecific binding was determined by addition of excess unlabeled EGF (300-fold) and was always < 20% of total binding.

Fluorescent labeling of cells A431 cells (2x105 or 2 x 1 0 6 , respectively) were seeded in medium with 1%FCS in 100 mm dishes and allowed to attach on glass coverslips. After 18-20 h, cells were rinsed in PBS and fixed for 5 min in methanol and for 2 min in acetone a t 4°C. After a brief rinse in PBS, the cells were treated with the first antibody diluted in PBS/BSA (1mgiml) for 1 h at 37°C. Control cells were incubated in PBSiBSA or with a n equivalent amount of nonspecific IgG. The coverslips were then washed thoroughly with PBS and incubated for 1 h at 37°C with a 1:300 dilution of the second antibody (antimouse IgG labeled with Texas Red, Dianova). After extensive washes (1 h) in PBS, the coverslips were mounted on glass slides in glycerol/PBS (9:1, viv, pH 8, containing 0.1% KJ) and sealed with Kronig Wax. Tumor cells were examined by fluorescent microscopy by using a photomicroscope equipped with Kodak Tri-X 400 film. The same exposure and development times were used within each experiment unless stated otherwise. Metabolic labeling of cells A431 cells were seeded at the indicated cell numbers in 10 ml of DMEiFl2 with 1% FCS in 100 mm culture dishes and cells were allowed to attach for 18-20 h. After two washes with DME lacking PI or RPMI without methionine, respectively, the monolayers were incubated for 4 h with 5 ml of DME without PI and serum, supplemented with 32P1(0.1 mCi/5 ml), or for 14-18 h with 5 ml of RPMI without methionine supplemented with 1%FCS and 35S-methionine (30 pCi/ 5ml for 2 x lo5 cells and 300 pCii5ml for 2 x lo6 cells, respectively). EGF was added during the last 5 min of the 4 h incubation, after which cells were washed four times with PBS a t room temperature.

Immunoprecipitation A431 monolayers were labeled with “Pi or 35S-methionine and stimulated with EGF as described above. Cells were lysed and immunoprecipitated as described by Frackelton (Frackelton et al., 1983).Briefly, 0.5 ml of ice-cold lysis buffer (18Triton X-100, 5 mM EDTA, 50 mM NaC1, 50 mM NaF, 30 mM Na,P,O,, 0.1 mM Na3V0,, 0.1% BSA, 1 mM PMSF, 10 pg/ml leupeptin, 50 pg/ml aprotinin, 10 mM Tris, pH 7.6) was added and the cells were scraped with a rubber policeman and transferred to a microfuge tube on ice. The culture dishes were rinsed with a n additonal 0.5 ml of lysis buffer which was combined with the first extraction. Samples were incubated on ice with occasional rigorous vortexing for 20 min and centrifuged at 15,OOOg for 15 min a t 4°C to remove insoluble debris. The supernatant

EGF RECEPTOR PHOSPHORYLATTON AND CELL-CELL CONTACT

305

was incubated with 50 ~1 of Sepharose 4B beads cou- and subsequently for 15 min in 1 M KOH and incupled with human serum albumin for 1 h at 4°C with bated a t 55°C for 2 h with gentle agitation in 1M KOH. continuous mixing to remove proteins binding to Afterwards, they were washed in several changes of Sepharose or serum proteins. I n order t o correct for the acetic acid-isopropanol for at least 2 h. Gels were dried differences in the intracellular labeled ATP pool of under vacuum and exposed with intensifying screens cells incubated a t different cell densities with the same at -70°C. Exposure times varied between 1 and 3 days amount of 32Pl,the precleared samples were normal- for untreated and 5 and 10 days for KOH-treated gels, ized on the basis of cpm in cell lysates or TCA-precip- respectively. itable cpm. Incorporation of 3-P (TCA-precipitable RESULTS cpm) was always between 10 and 15% of 32P-uptake (cpm in lysates) and this relation was not affected by Effect of cell density on 12'II-EGF binding cell density or EGF treatment. The lysates were added To examine the possible influence of cell-cell contact to fresh tubes containing either 10 p1 of EGFRl or 10 pl of Sepharose 4B beads to which aPTyr had been on the biological activity of EGFR, we seeded singlecovalently attached and incubated for 2 h at 4°C with cell suspensions of A431 cells at such numbers that 24 inversion mixing. For separation of the proteins phos- h after adhering to the plastic substrate they either phorylated on tyrosine, the beads were washed three formed clusters of 10-30 or 2-5 cells with extensive times with lysis buffer and twice with lysis buffer with- (Fig. 1A) or limited cell-cell contacts (Fig. lB), respecout BSA and the antigen was eluted by incubating for tively. These clusters we will refer to hereafter as 10 min on ice with gentle mixing in 60 E J . ~of elution dense (Fig. 1A) or sparse (Fig. 1B) cultures, respecbuffer (lysis buffer with 1mM phenyl phosphate, 0.01% tively. Binding experiments with lZ5I-EGFwere perovalbumin instead of 0.1% BSA and 10 mM NaCl in- formed on A431 cells growing at the densities described stead of 50 mM). This was repeated once; 30 pl of four above, including a higher one where cells had reached times concentrated Laemmli sample buffer containing nearly confluency (2 x lo5 cellsil6 mm well) (Fig. 2). 200 mM DTT was added to the combined eluates and Scatchard analysis of l2'1-EGF binding resulted in a heated for 4 min at 95°C. For separation of the labeled best fit according to the two-affinity model with two receptor recognized by EGFR1, 20 pl of Protein A- classes of binding sites (Shoyab e t al., 1979; King and Sepharose was added and incubation continued for 1h. Cuatrecasas, 1982). At high cell density < 1% of the Subsequently, the immunoprecipitates were washed sites were in the high-affinity state (K, = 0.083 nM) with lysis buffer as described above. The pellets were and >99% in the low-affinity state (K, = 11nM) while resuspended in 120 pl of Laemmli sample buffer con- at the lowest cell density 10% of the sites were in the taining 50 mM DTT, heated for 4 rnin a t 95"C, and high (K, = 0.28 nM)- and 90% in the low-affinity state centrifuged. Samples were analyzed by one-dimen- (K, = 21 nM), respectively (Table 1). Thus, cell density sional sodium dodecyl sulphate polyacrylamide gel did not influence number or affinity of low-affinity sites, but had considerable effects on the high-affinity electrophoresis (1D PAGE). sites. With decreasing cell numbers and thus decreasPhosphorylation of detergent-insoluble ing cell-cell contact, the number of high-affinity recepcytoskeleton fractions with Y - ~ ~ P - A T P tors increased by 16.8-fold. Treatment of cells with acid A431 cells were seeded at different cell densities in prior to Scatchard analysis (pH 5.0 for 6 min at 4°C) DME/F12 medium with 1% FCS as indicated. After according to the method described by Haigler e t al. 18-20 h cells were washed twice with PBS and then (1980) did not change binding parameters of high-afonce with HMCa buffer (80 mM HEPES, pH 6.9,2 mM finity receptors (data not shown). MgCl,, 1 mM CaC1,) at room temperature. SubseEGF-induced phosphorylation in intact A431 quently, the cells were incubated for 1min with 1ml of cells: immunoprecipitation studies washing buffer (HMCa with 2 mM MnCl,, 0.1 mM Na3V04, 1 mM PMSF, 10 pg/ml leupeptin, 50 pg/ml Based on our observations that ligand binding acaprotinin) containing 0.15% TX-100. The detergent- tivity and subcellular distribution of EGFRs were soluble fraction was discarded and the detergent-insol- modulated by cell density, we next investigated autouble cytoskeleton fraction was rinsed three times with phosphorylation and kinase activity by immunoprecipiwashing buffer. Phosphorylation was initiated by the tation studies. We used the Mab against the EGFR addition of 1 ml of washing buffer containing 10 pM recognizing the ligand binding domain and a n antiATP and 10-20 KCi y3'P-ATP. After 10 min incuba- phosphotyrosine Mab (aPTyr) with high affinity and tion at room temperature with occasional gentle shak- specificity for phosphotyrosine residues (Frackelton et ing, the plates were rinsed four times with washing al., 1983). In immunoprecipitates of whole cell lysates buffer and transferred on ice, and the cytoskeletons of A431 cells prelabeled with ',PI, EGFRl recognized were scraped with a rubber policeman in lysis buffer only one phosphorylated protein with M, of 170 kd (Fig. and used for immunoprecipitation a s described above. 3A, lane 1)identifying it as the EGF receptor. Addition of EGF to the cell cultures resulted in increased phosElectrophoretic techniques phorylation of this band (lane 2). However, using For identification of phosphorylated proteins 20 p1 of aPTyr this protein was not rdiolabeled in control cell sample was used and separated by 1D-SDS PAGE on cultures (even not detectable after prolonged expo4-1592 gels. In order to increase the detection of pro- sures, Carpentier et al., 1987; Honegger et al., 1988) teins labeled at tryosine (Cooper and Hunter, 1981) while in sparse cultures proteins were precipitated gels run in parallel with 50 pl of sample were soaked in with M, of 130 and 48 kd, respectively (Fig. 3, lane 5). two 15 min washes in 10% acetic acid-10% isopropanol These proteins were also detectable in dense cultures

306

LICHTNER AND SCHIRRMACHER

Fig. 1. Phase contrast micrograph of A431 cells growing at different cell densities. A431 cells were plated in 10 ml of DMEiFl2 plus 1%FCS at 2 x lo6 (A) or 2 x lo5 (B) cells in 100 mm dishes and allowed to attach overnight. x 180.

after prolonged exposure (data not shown). In EGFstimulated sparse cells the EGFR was extensively phosphorylated in addition to several proteins (Fig. 3A, lanes 4 and 6) and alkali treatment (Fig. 3B) identified proteins with M, or 170, 94, 76, 58, and 39 kd as the major proteins phosphorylated on tyrosine in whole A431 cells in an EGF-dependent fashion. These were also detectable in dense cultures but only after prolonged exposure (data not shown). In order to establish the relative amount of EGFR precipitated by aPTyr we labeled A431 cells with 35S-methionine. Figure 3C demonstrates no phosphorylated receptor protein in unstimulated cells (lanes 3 and 5, see also Fig. 3A,B). Upon EGF treatment comparable amounts of 35S-labeled receptor protein were precipitated from dense and sparse cultures (lanes 4 and 6). This contrasts with the significantly lesser amount of 32P associated with EGFR in dense cells (compare Fig. 3A, B, and C, and see text earlier). The simplest explanation would be that a certain population of EGFRs in sparse cells is highly phosphorylated in relation to the mass of EGFR in dense cells. We also tested tryosine kinase activity in dense and sparse cultures a t limiting EGF concentrations (Fig. 4). At 1 ngiml EGF the phosphorylation in either culture was below detection level. However, 10 ngiml of EGF induced a level of phosphorylated protein in sparse cells comparable to that with 150 ngiml of EGF in dense cells, which exhibit four times less highaffinity receptors (Table 1). The possibility that A431 cells release autocrine growth factors and that this may have reduced the binding of exogenously added EGF was tested by examining the effects of medium conditioned by A431 cells a t high cell density on the EGF-dependent phosphorylation. In this experiment, A431 cells were plated at high or low density with subsequent labeling with

32Plfor 4 h in fresh medium or in medium previously conditioned by A431 cells (data not shown). These experiments indicated that conditioned medium had only minor effects on EGF-dependent phosphorylation and cannot account for the immense differences seen in sparse vs. dense cultures. Distribution of phosphorylated proteins in intact A431 cells following EGF stimulation: immunofluorescent analysis The decrease in EGF-dependent stimulation of EGFR phosphorylation correlated with increase in cell density and could, therefore, be related to the establishment of cell-cell contacts. Thus, cells in the center of a cell cluster may react differently to EGF addition than those at the periphery. In order to test this hypothesis we performed immunof luorescent staining with aPTyr on A431 cells in dense and sparse cultures. Figure 5D demonstrates that in A431 cells growing in sparse cultures a 5 min incubation with EGF induced a bright staining for phosphotyrosine in all cells, which was most pronounced in microvilli and cell-cell contact areas. However, in cells growing in dense cultures, EGF-induced phosphotyrosine formation did not occur in all cells but preferentially in cells situated atlthe periphery of a cluster and especially in expanding borders (Fig. 5C). The same cells are shown in Figure 5E, but with a different development time in order t o demonstrate the outline of the cell cluster. In unstimulated cells only dull, cytoplasmic staining was observed (Fig. 5A,B). Control experiments with an equivalent amount of nonspecific IgG only slightly increased background staining (data not shown), confirming the specificity of the antibodies used (see also Fig. 6).

307

EGF RECEPTOR PHOSPHORYLATION AND CELL-CELL CONTACT

0.5

0.4

(u

2 0.3

.

LL

5

0

U

c

:0.2

10 15 Free ( nM )

20

25

m

0.1

0.0 0.0

1 .o

0.5

1.5

2.0

Bound ( # per cell ) xl 0 - 6

0.20

0.15 01

s!

LL

5

0.10

3

0

m

0.05

0.00

'

0

1

0.2

0.4

0.6

0.8

1

1.2

1.4

1.6

Bound ( # per cell ) x l 0-6 0.05

0.04

(u

.

0.03

LL

D

Free ( nM )

c

:0.02 m 0.01

0

0

0.5

1

1.5

2

Bound ( #per cell ) x I O - ~ Fig. 2. Scatchard analysis of '251-EGFbinding to A431 cells growing at different cell densities. A431 cells (A: 2 x lo5; B: 5 x lo4,and C : 5 x lo3)were plated in 1ml of D m / F 1 2 plus 1%FCS in €6 mm well plates and allowed t o attach overnight; '251-EGFbinding was determined over a range of 0.25-150 n g h l EGF by incubation for 1h at 4°C. Inserts depict amount of ligand bound vs. the logarithm of free ligand concentration.

308

LICHTNER AND SCHIRRMACHER

TABLE 1. Effect of cell densitv on EGF-binding' Cell No.

K1 (nM)' 0.083 r 0.033 (1.014 0.15 t 0.02 (1.8) 0.28 ? 0.06 (3.4)

( x 104)

20 5 0.5

n l ( x lo6) 0.022 ? 0.01 (1.01 0.088 ? 0.03 (4.0) 0.370 i 0.08 (16.8)

K2 (nM) 11 2 111.0) 5 i 0.4 (0.5) 21 F 5 (1.9)

n2 ( x 1 0 9 2 0.3 (1.0) 2 ? 0.4 (1.0) 2.9 t 0.4 (1.5)

*

'Binding parameters were determined from Scatchard plots to obtain receptor numbers per cell ( n i and ligand dissociation constants (K). The plots were fitted by using a computer routine generating a Scatchard plot of the experimental points as well as the curve resulting from the parameters chosen. "'1-EGF binding was determined a t 4°C for 1 h a s described in Materials and Methods. 'Mean of triplicates of a t least three independent experiments.

%EM. 4Relative values

1

2

3

4

5

6

1

2

3

4

5

6

7

3

4

5

6

180116-

84584836Fig. 3. EGF-induced phosphorylation in whole cell lysates of A431 cells growing at different cell densities. A431 cells (lanes 1 4 , 7 2 x lo6 and lanes 5,6: 2 X lo5 per 100 mm dish) were prelabeled with 32P, (A, B) or 35S-methionine(C) and for the last 5 min 150 ng/ml ECF (lanes 2, 4, 6) or PBS (lanes 1,3, 5, 7) was added. Immunoprecipitations were performed with EGFRl (lanes 1, 2 ) , aPTyr (lanes 3-6) or unrelated IgG, (lane 7). Autoradiograph of untreated (A) and KOH (€3-treated gel, or impregnated with DMSOiPPO (C).

Distribution of EGFRs in intact A431 cells: immunofluorescent analysis Since only subsets of EGFRs were phosphorylated in dense cultures upon EGF treatment, we next investigated the distribution of EGFRs in A431 cells seeded a t different cell densities by immunofluorescence analysis with a n antibody specific for the receptor protein (Waterfield et al., 1982). Figure 6 shows that in dense cultures of A431 cells the receptors were localized mainly in cell extensions, microvilli and at the cell boundaries where the cells establish cell-cell contacts. In contrast, in sparse cultures of A431 cells a n overall bright staining of cells was found with intense staining in cell extensions (Fig. 6B). Thus, it appears that cell-cell interactions induce localization or recruitment of EGFRs in cell-cell contact areas and that their response to EGF is suppressed in intact dense cells (Fig. 5 ) .

EGF-induced phosphorylation of cytoskeleton-associated EGFRs: immunofluorescent studies It is well established that EGFRs isolated from A431 cells are able to undergo EGF-independent (basal) and EGF-induced autophosphorylation when tested in vitro. Therefore, we investigated in detergent-permeabilized cells whether EGFRs associated with the cytoskeleton and localized in cell-cell contact areas possess kinase activity by immunofluorescent (Fig. 7) and immunoprecipitation (Fig. 8) analysis with aPTyr. In these experiments intact, adherent A431 cells were stimulated with EGF for 5 min and lysed with mild detergent treatment, and activity of kinase associated with the cytoskeleton was initiated by the addition of ATP. All manipulations were carried out at room temperature, since some cytoskeletal structures are labile

EGF RECEPTOR PHOSPHORYLATION AND CELL-CELL CONTACT

309

1 2 3 4 5 6

18011684-

5848Fig. 4. EGF-induced phosphorylation in whole cell lysates of A431 cells growing at different cell densities: EGF variation. A431 cells (lanes 1-3: 2 x 10" and lanes 4-6: 2 x 10' per 100 mm dish) were prelabeled with %'Piand for the last 5 min EGF at 1 (lanes 1, 4), 10 (lanes 2,5), or 150 (lanes 3, 6) ng/ml was added. Immunoprecipitation was performed with aPTyr. Autoradiograph of untreated gel.

at low temperature, and we wanted to preserve positional relationships between kinases and the respective substrates. Distinct phosphorylation of cytoskeleton-associated proteins of A431 cells growing in dense and sparse cultures was demonstrated in unstimulated cells (Fig. 7A,B) which was further enhanced upon EGF treatment (Fig. 7C,D). EGF-induced phosphorylation of cytoskeleton-associatedEGFRs: immunoprecipitation studies Since detergent treatment might have removed inhibitors of tyrosine kinases not related to EGFR or exposed new, maybe unnatural, substrates, we performed immunoprecipitations with aPTyr of proteins phosphorylated in cytoskeleton preparations of A431 cells growing at different cell densities (Fig. 8). The EGFR as well as proteins with M, of 130, 96, 80, and 39 kd, respectively, were phosphorylated intensively on tyrosine in dense control cells, and EGF addition resulted in slightly augmented phosphorylation (Fig. 8, lanes 1 and 2). In A431 cells growing with very limited cell-cell contact, lower phosphorylation levels were observed of EGFR as well as other proteins in control cytoskeleton preparations while a pronounced relative increase in phosphorylation occurred following EGF stimulation (Fig. 8, lanes 3 and 4).

DISCUSSION In the present study we examined the potential significance of cell-cell contacts on EGF receptor affinity, kinase activity, and distribution in cultured A431 cells seeded a t different initial cell densities. To investigate the EGF binding characteristics we subjected A431 cells to Scatchard analysis and determined that the numbers of high-affinity receptors decreased as cell

Fig. 5. Immunofluorescent staining with aPTyr of EGF treated and unstimulated A431 cells growing at different cell densities. A431 cells (A, C, E: 2 x lo6 or B, D, F: 2 x 10' per 100 mm dish) were plated in 10 ml of medium plus 1% FCS and allowed to attach on glass coverslips overnight. Intact cells were stimulated with 150 ngiml EGF (C-F) or PBS (A, B) for 5 min prior fixation. The same cell cluster is shown in C and E, but with different development times. aPTyr was used at 50 kg/ml (A-E) and was omitted in F. x 790.

density increased in agreement with previous studies (Rees et al., 1984; Rizzino et al., 1988). Immunoprecipitation with a phosphotyrosine-specific antibody demonstrated enhanced EGF-induced phosphorylation in sparse compared with dense intact cells tested at different EGF concentrations. The concomitant increase in high-affinity receptor number and EGF concentration-dependent phosphorylation in sparse cultures of A431 cells could indicate a causal relationship between affinity to EGF and phosphotyrosine kinase activity of EGFRs in intact A431 cells. However, even at extremely low concentrations of EGF a substantial amount will bind to low-affinity sites (Defize et al., 1989). Therefore, the increased tyrosine phosphorylation in sparse cells at limiting EGF concentrations is suggestive but no direct evidence to prove t h a t the number of high-affinity receptors and tyrosine phosphorylation are linked. Furthermore, comparison between aPTyr-immunoprecipitates from EGF-stimulated 35S-methionine and 32P-labeledEGFRs indicated the presence of a population of highly phosphorylated receptors in sparse cells. This correlates with the greater number of high-affinity EGFR in these cells as

310

LICHTNER AND SCHIRRMACHER

Fig. 6 . Immunofluorescent staining of EGFRs in A431 cells growing at different cell densities. A431 cells (A,C:2 x lo6;or B:2 x lo6)were plated in 10 ml of DMEIF12 plus 1% FCS in 100 mm dishes and allowed to attach on glass coverslips overnight. EGFRl was used at 10 pgiml (A, B) and was omitted in C. x 910.

compared with dense cultures. Our conclusion is based on the assumption that one antibody molecule only binds one receptor molecule, which seems likely considering the huge antibody-Sepharose complex. Similarly, in A431 (Defize et al., 1989) and in phaeochromocytoma PC12 cells (Boonstra et al., 19871, a correlation between high-affinity EGFRs and tyrosine kinase activity of EGFRs had been demonstrated. In immunolocalisation studies with a n antibody recognizing the ligand binding domain of EGFR and thus not distinguishing between activated and nonactivated EGFRs, we demonstrated preferential localization of EGFRs in cell-cell contact areas in dense but not in sparse cultures of A431 cells. A structural coupling of a subset of EGFRs with the cytoskeleton is well documented (Wiegant et al., 1986; van Belzen et al., 1988; van Bergen en Henegouwen et al., 1989; Roy e t al., 1989)although the influence of cell density has not been investigated. The cytoskeleton-associated EGFRs retained both a functional ligand-binding domain and tyrosine kinase activity (Giugni et al., 1985; Landreth et al., 1985; Roy e t al., 1989). In addition, several cytoskeleton-associated proteins have been reported to be

Fig. 7. EGF-induced in vitro phosphorylation of proteins associated with the cytoskeleton of A431 cells growing a t different cell densities: Identification by immunofluorescence. A431 cells (A,C,E: 2 x 106 or B, D, F: 2 x lo7per 100 mm dish) were plated in 10 ml of medium plus 1% FCS and allowed to attach on glass coverslips overnight. Intact cells were stimulated with 150 ng/ml EGF (C-F) or PBS (A, B) for 5 min prior mild detergent lysis (0.15% TX-100for 1 min). Kinase reactions were initiated by the addition of 10 pM cold ATP and after 10 min incubation cytoskeletons were fixed. aPTyr was used at 50 pg/ml (A-D)and was omitted in E. F. x 830.

phosphorylated in response to EGF (Giugni et al., 1985; Landreth et al., 1985; Bretscher, 1989). However, detergent treatment of cells might expose new, unnatural substrates for tyrosine kinases. Thus, it is not surprising that the pattern of phosphoproteins isolated with aPTyr in this investigation was different in intact cells when compared with cytoskeleton preparations. For example, the EGF-induced phosphorylation of proteins with relative molecular masses of 170,130, 94, and 39 kd, respectively, was observed in intact as well as permeabilised cells. In contrast, phosphoproteins with M, of 76 and 58 kd were only detected in intact cells and with M, of 80 and 48 kd, respectively, only in cytoskeleton preparations. Thus, from the phosphorylated substrates reported in the present investigation p39 might represent lipocortin (Fava and Cohen, 1984) and p80 might represent ezrin (Bretscher, 1989) while the identity of the others remains obscure. To our surprise, EGFRs were capable of undergoing EGF-independent phosphorylation in cytoskeleton preparations in contrast to intact cells. According to the allosteric oligomerization model proposed by Yarden and Schlessinger (1987a,b) activation of EGFR

EGF RECEPTOR PHOSPHORYLATION AND CELL-CELL CONTACT

1

2

3

4

1 2 3 4

18011684-

58483626-

Fig. 8. EGF-induced in vitro phosphorylation of proteins associated with the cytoskeleton of A431 cells growing a t different cell densities: Identification by immunoprecipitation. A431 cells (lanes 1 , 2 2 x lo6 and lanes 3, 4 2 x lo5 per 100 mm dish) were plated in 10 ml of medium plus 1% FCS and allowed to attach overnight. Intact cells were stimulated for 5 min with 150 ngiml EGF (lanes 2, 4) of PBS (lanes 1, 3); cytoskeletons of adherent cells were prepared by mild detergent lysis (0.158 TX-100 for 1 m i d ; and kinase reactions were initiated by the addition of 10 yM ATP plus 20 yC Y - ~ ~ P - A TAfter P. incubation for 10 min immunoprecipitation was performed with aF'Tyr. Autoradiograph of untreated (A) and KOH (B) treated gel.

kinase involves an intermolecular process. Thus, removal of membrane lipids and proteins by detergent treatment might have facilitated interactions between neighboring receptors associated with the cytoskeleton allowing EGF-independent phosphorylation. Consequently, EGF addition resulted in augmented although not marked increase in phosphorylation in dense cultures of A431 cells. However, as density decreased, the relative EGF-induced phosphorylation increased in permeabilised cells, suggesting that a smaller amount of EGFRs might be associated with the cytoskeleton in sparse cultures. Thus, EGF treatment of intact sparse cells might either induce association of EGFRs with the cytoskeleton as suggested by van Bergen en Henegouwen et al. (1989) or increase the relative phosphorylation levels of associated receptors. In a recent report by Roy et al. (1989) only addition of EGF to intact cells but not to cytoskeleton preparations resulted in increased autophosphorylation of associated EGFRs. The authors concluded that conformational restrictions of EGFRs as a consequence of their structural association with the cytoskeleton might be responsible for this effect. They based this explanation on the observation of Yarden and Schlessinger (1987a,b), who demonstrated that autophosphorylation of EGFRs immobilized on a solid matrix was not stimulated above basal levels upon addition of EGF. This leads to the pivotal question: does the association of EGFRs with the cytoskeleton cause immobilization

311

and unresponsiveness to EGF (Roy et al., 1989) or does removal of membrane constituents by detergent treatment facilitate interactions between cytoskeleton-associated neighboring receptors allowing autophosphorylation (this report)? We feel that differences in the experimental procedures, such as adherent cells vs. cells in suspension, room temperature vs. low temperature, and presence or absence of Ca2 ' may account for the different observations. Further investigations are needed to resolve the role of cytoskeleton-associated EGFRs in the transmission of the biological signal of EGF. Since in intact cells the response to EGF was restricted to some cells mainly at the margin of cell clusters we postulate that in dense cultures negative control mechanisms preventing EGFR activation may be exerted by adjacent cells. This could be due t o either density-induced down regulation of growth factor receptors (Rizzino et al., 1988) or growth factor-specific down regulation. It has been reported that TGFa anchored to the cell surface can interact with and activate EGFRS on adjacent cells (Wong et al., 1989).Two lines of evidence argue against growth factor-specific down regulation in our experiments. First, activation of EGFRs by growth factor produced by neighboring cells would have resulted in phosphorylated EGFRs in control cell cultures, which is contrary to our observation; second, EGFRs in all A431 cells seeded at high density were equally capable of undergoing ligand-induced tyrosine phosphorylation when labelled in situ in detergent-permeabilised cells. This also suggests that the EGFR in densely cultured intact cells binds exogenous EGF, but its activation is inhibited by factors which are removed during cytoskeleton preparation. In summary, this investigation demonstrates evidence that cell-cell interactions modulate both the intracellular localisation and biological activity of EGFRs in A431 cells. A431 cells provide the most extensively studied model of EGF action but most of these investigations have been performed on subconfluent or even confluent cultures. The precise mechanism whereby EGF regulates cell proliferation and differentiation is still unknown. Based on our data presented here we propose to also conduct studies on sparse cultures of A431 cells since, for example, besides the effects on subcellular distribution and kinase activity of EGFRs the availability of tyrosine kinase substrates and thus the signalling pathway could also be affected by cell-cell interactions.

ACKNOWLEDGMENTS We gratefully acknowledge the excellent technical assistance of Ms. M. Sieber. We thank Dr. K. Khazaie for fruitful discussions, Dr. L. Erkell and C. Ertel for help with the statistical analysis, Drs. G. Brunner and V. Kinzel for critical reading of the manuscript. Ms. E. Schloms for preparation of the manuscript, and Ms. B. Engelhardt, J. Kollner, and R. Kiihnl-Bontzol for expert help with photographic work. Supported by Deutsche Forschungsgemeinschaft.

LITERATURE CITED Ben-Ze'ev, A. (1985)The cytoskeleton of cancer cells. Biochim. Biophys. Acta, 789:197-212. Boonstra, J., van Belzen, N., van Maurik, P., Hage, W.J., Blok, F.J.,

3 12

LICHTNER AND SCHlRRMACHER

Wiegant, F.A.C., and Verkleij, A.J. (1985) Immunocytochemical demonstration of cytoplasmic and cell-surface EGF-receptors in A431 cells using cryoultra microtomy surface replication, freezeetching and label fracture. J . Microsc. (Oxf), 14Ot119-129. Boonstra, J.,Mummery, G.L., Feijen, A., de Hoog, W.J., van der Saag, P.T., and de Laat, S.W. (1987) Epidermal growth factor receptor expression during morphological differentiation of pheochromocytoma cells induced by nerve growth factor or dibutyryl cyclic AMP. J . Cell. Physiol., 131:409-417. Bretscher, A. (1989) Rapid phosphorylation and reorganization of ezrin and spectrin accompany morphological changes induced in A431 cells by epidermal growth factor. J . Cell Biol., 108t921-930. Carpentier, J.-L., White, M.F., Orci, L., and Kahn, R.C. (1987) Direct visualization of the phosphorylated epidermal growth factor receptor during its internalization in A431 cells. J. Cell Biol., 105.27512762. Chinkers, M., McKanna, J.A., and Cohen, S. (1979) Rapid induction of morphological changes in human carcinoma cells A431 by epidermal growth factor. J . Cell Biol., 83:260-265. Chinkers, M., McKanna, J.A., and Cohen, S. (1981) Rapid rounding of human epidermoid carcinoma cells A431 induced by epidermal growth factor. J. Cell Biol., 88.422-429. Cooper, ,J.A., and Hunter, T. (1981) Changes in protein phosphorylation in Rous sarcoma virus-transformed chicken embryo cells. Mol. Cell. Biol., It165-178. Defize, L.H.K., Boonstra, J.,Meisenhelder, J., Kruijer, W., Tertoolen, L.G.J., Tilly, B.C., Hunter, T., van Bergen en Henegouwen, P.M.P., Moolenaar, W.H., and de Laat, S.W. (1989) Signal transduction by epidermal growth factor occurs through the subclass of high affinity receptors. J. Cell Biol., 109t2495-2507. Downward, J., Parker, P., and Waterfield, M.D. (1984) Autophosphorylation sites on the epidermal growth factor receptor. Nature (Lond.), 311 t483-485. Fava, R.A., and Cohen, S. (1984) Isolation of a calciumdependent 35kilodalton substrate for the epidermal growth factor receptorikinase from A431 cells. J. Biol. Chem., 259:2636-2645. Frackelton, A.R., Jr., Ross, A.H., and Eisen, H.N. (1983) Characterization and use of monoclonal antibodies for isolation of phosphotyrosyl proteins from retrovirus-transformed cells and growth factor-stimulated cells. Mol. Cell. Biol., 3:1343-1352. Giugni, T.D., James, L.C., and Haigler, H.T. (1985) Epidermal growth factor stimulates tyrosine phosphorylation of specific proteins in permeabilized human fibroblasts. J. Biol. Chem., 260t15081-15090. Haigler, H.T., Maxfield, F.R., Willingham, M.C., and Pastan, I. (1980) Dansylcadaverine inhibits internalization of '251-epidermal growth factor in BALB 3T3 cells. J. Biol. Chem., 255t1239-1241. Haigler, H.T., McKanna, J.A., and Cohen, C. (1979) Direct visualisation of the binding and internalization of a ferritin conjugate of epidermal growth factor in human carcinoma cells A431. J. Cell. Biol., 81:382-395. Hillmann, G.M., and Schlessinger, J. (1982) Lateral diffusion of epidermal growth factor complexed to its surface receptors does not account for the thermal sensitivity of patch formation and endocytosis. Biochemistry, 21t1667-1672. Honegger, A,, Dull, T.J., Bellot, F., van Obberghen, E., Szapary, D., Schmidt, A,, Ullrich, A,, and Schlessinger, J . (1988) Biological activities of EGF-receptor mutants with individually altered autophosphorylation sites. EMBO J . , 7r3045-3052. Hunter, T., and Cooper, J.A. (1981) Epidermal growth factor induces rapid tyrosine phosphorylation of proteins in A431 human tumor cells. Cell, 245'41-752.

King, A.C., and Cuatrecasas, P.J. (1982) Resolution of high and low affinity epidermal growth factor receptors: inhibition of high a f i n ity component by low temperature, cycloheximide and phorbol esters. J. Biol. Chem., 257t3053-3060. Landreth, G.E., Williams, L.K., and Rieser, G.D. (1985)Association of the epidermal growth factor receptor kinase with the detergentinsoluble cytoskeleton of A431 cells. J. Cell Biol., IOIr1341-1350. Paulsson, Y., Beckmann, M.P., Westermark, B., and Heldin, C.-H. (1988) Density-dependent inhibition of cell growth by transforming growth factor-pl in normal human fibroblasts. Growth Factors, I : 19-27. Rees, A.R., Gregoriou, M., Johnson, P., and Garland, P.B. (1984) High affinity epidermal growth factor receptors on the surface of A431 cells have restricted lateral diffusion. EMBO J., 3~1843-1847. Rizzino, A., Kazakoff, P., Ruff, E., Kaszynski, C., and Nebelsick, J . (1988) Regulatory effects of cell density on the binding of transforming growth factor p, epidermal growth factor, platelet-derived growth factor, and fibroblast growth factor. Cancer Res., 48r42664271. Roy, L.M., Gittinger, C.K., and Landreth, G.E. (1989) Characterization of the epidermal growth factor receptor associated with cytoskeletons of A431 cells. J. Cell. Physiol., 14Ot295-304. Schlessinger, JI, and Geiger, B. (1981) Epidermal growth factor induces redistribution of actin and a-actinin in human epidermal cells. Exp. Cell Res., 134t273-279. Schlessinger, J., Shechter, Y., Willingham, M.C., and Pastan, I. (1978) Direct visualization of binding, aggregation and internalization of insulin and epidermal growth factor on living fibroblastic cells. Proc. Natl. Acad. Sci. USA, 752659-2663. Shoyab, C.M., DeLarco, J.E., and Todaro, G.J. (1979) Biologically active phorbol esters specifically alter affinity of epidermal growth factor receptors. Nature, 279:387-391. Staehelin, L.A. (1974) Structure and function of intercellular junctions. Int. Rev. Cytol., 39:191-283. van Belzen, N., Rijken, P.J., Hage, W.J., de Laat, S.W., Verkleij, A.J., and Boonstra, J. (1988) Direct visualization and quantitative analysis of epidermal growth factor-induced receptor clustering. J . Cell. Physiol., 134t413-420. van Bergen en Henegouwen, P.M.P., Defize, L.H.K., de Kroon, J.,van Damme, H., Verkleij, A.J., and Boonstra, J . (1989) Ligand-induced association of epidermal growth factor receptor to the cytoskeleton of A431 cells. J. Cell. Biochem., 39:455-465. Waterfield, M.D., Mayes, E.L.V., Stroobant, P., Bennet, P.L.P., Young, S., Goodfellow, P.N., Banting, G.S., and Ozanne, B. (1982) A monoclonal antibody to the epidermal growth factor receptor. J. Cell. Biochem., 20:149-161. Wieeant, F.A.C., Blok, F.J.. Defize. L.H.K.. Linnemans. W.A.M.. Virkleij, A.J., and Boonstra, J . (1986) Epidermal g o w t h factor re: ceptors associated to cytoskeletal elements of epidermoid carcinoma (A431) cells. J. Cell Biol., 103t87-94. Wong, S.T., Winchell, L.F., McCune, B.K., Earp, H.S., Teixido, J., Massague, J., Herman, B., and Lee, D.C. (1989) The TGFa precursor expressed on the cell surface binds to the EGF receptor on adjacent cell, leading to signal transduction. Cell, 56~495-506. Yarden, Y., and Schlessinger, J. (1987a) Self-phosphorylation of epidermal growth factor receptor: Evidence for a model of intermolecular allosteric activation. Biochemistry, 2611434-1442. Yarden, Y., and Schlessinger, J . (1987b) Epidermal growth factor induces rapid, reversible aggregation of the purified epidermal growth factor receptor. Biochemistry, 26t1443-1451.