EXPERIMENTAL CELL RESEARCH 194, 48-55 (1991)

The Role of Microfilaments in the Capping of Epidermal Growth Factor Receptor in A431 Cells I . A . KHREBTUKOVA,**’K. KWIATKOWSKA,~D. A. GUDKOVA,* A. B.

SOROKIN,*

AND

G. P. PINAEV*

*Institute of Cytology of the Academy of Sciences of the USSR, 4 Tikhoretsky Avenue, 194064 Leningrad, USSR; and tNencki Institute of Experimental Biology, Department of Cell Biology, 3 Pasteur Street, PL-02093 Warsaw, Poland

ample of this type of receptor is the epidermal growth factor (EGF) receptor, which is primarily involved in a triggering of proliferation or some other metabolic process. EGF is an intensively studied polypeptide growth factor. The various effects of EGF in its target cells have been the subject of extensive studies [8,9]. Most of the studies on the mechanism of EGF activity were performed in human epidermoid carcinoma A431 cells, which have, on the cell surface, an unusually high number of EGF receptors (EGF-R), 2-3 X lo6 receptors/cell [lo]. An association of at least a part of the EGF-R with Triton X-lOO-insoluble cytoskeleton of A431 cells has been described [ll-141. Judging from the diameter of the filaments observed in electron micrographs it has been suggested that actin is involved in this association [12]. We examined the capping of EGF-R in suspended and adherent A431 cells in order to analyze the participation of different parts of the actin cytoskeleton in this process. We report that the multivalent ligands induce the aggregation of EGF-R. The capping of EGF-R on the cell surface showed clean redistribution of actin in both suspended and adherent cells, suggesting a connection between these two proteins. The stress-fiber-like bundles of actin in cytoplasm were not involved in the capping. However, intact cortical circular actin filaments are needed for the EGF-R cap formation.

Capping of the EGF receptor (EGF-R) on the surface of suspended and adherent epidermoid carcinoma cells, A431, is studied. It was induced at 20% after treating cells with monoclonal antibody to the EGF receptor followed by the second antibody conjugated with FITC. Accumulation of cortical actin under the caps was detected by rhodamine-phalloidin. Destruction of the actin stress-fiber-like bundles was observed during incubation of cells with the ligands at 0°C. Two processes appear to take place at 20°C: redistribution of the EGFR with cortical actin into the caps within 15-30 min and reconstruction of cytoplasmic actin bundles over 45-60 min. Dihydrocytochalasin B prevented cap formation in adherent cells, but small patches of EGF-R colocalized with actin aggregates under plasma membrane were observed. The function of different actincontaining cytoskeleton structures in the process of o 1991 Academic PEWS, IUC. capping is discussed.

INTRODUCTION Association of actin microfilaments with the plasma membrane is an important factor in various cellular phenomena, in determination of cell shape, in cell motility, and in cell-cell and cell-substratum adhesion inclusive [l-3]. Activation of these processes starts from the binding of ligands to specific surface membrane receptors. It is believed that activated receptors take a part in modifying the state of the cytoskeleton [3, 41. Transmembrane interactions of ligand-receptor complexes with actin microfilament bundles at the molecular level was described in detail for cell contact phenomena [5-71. It is likely that these interactions are essential for the mechanical force production and/or cell shape stabilization. It is interesting, however, to determine interactions between the cytoskeleton, particularly actin filaments, and receptors, whose function is not a cell shape or mechanical force regulation. An ex’ To whom correspondence dressed.

and reprint

0014-4827/91$3.00 Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

requests

should

MATERIALS AND METHODS Cell culture. Human epidermoid carcinoma cells (A431) were obtained from the USSR Cell Culture Collection (Institute of Cytology of the Academy of Sciences, Leningrad). Cells were grown on glass coverslips in 35-mm culture dishes in Eagle’s medium supplemented with glutamin and 10% bovine serum, in a humidified atmosphere of 5% CO, at 37°C. For preparation of nonadherent cells, they were grown to subconfluency. Then the cells were detached from substrata by 0.02% EDTA, washed with a culture medium, and resuspended in a culture medium. Materials. EGF purified from submaxillary glands of adult male mice was kindly supplied by Dr. A. D. Sorkin (Institute of Cytology,

be ad-

48

THE

EGF-R

CAPPING

Leningrad). The monoclonal antibody to the EGF receptor, denoted as mAb 5A9, is an IgG that recognizes a protein with the molecular mass of 170 kDa from the membrane preparation of A431 cells and localizes at the EGF-rhodamine binding sites in 6xed A431 cells. The antibody neither competes with EGF to bind the receptor nor influences protein kinase activity of the receptor (for details, see [15]). TO induce redistribution and staining of the ligand-receptor complexes, FITC-conjugated swine anti-mouse IgG (FITC-SWAM) (Institute of Sera and Vaccines, Praha) was applied as a secondary antibody. Rhodamine-phalloidin (Molecular Probe) was used for the specific fluorescent staining of actin filaments in the cells. For the destruction of microtubules and microfilaments, respectively, colcemid (Calbiothem) and dihydrocytochalasin B (Calbiochem) were applied. The latter is a derivative of cytochalasin B that is shown to induce changes in actin structures but does not affect sugar transport [16].

Fluorescent labeling of the &and-receptor complexes and actin filaments. Both suspended and adherent cells were sequentially incubated at O°C for 30 min with EGF (200 rig/ml), mAb 5A9, and FITCSWAM. For the living cells there were ligands supplemented in the culture medium. In some experiments only antibodies were used as ligands. To induce redistribution of the ligand-receptor complexes, the treated cells were incubated at 37 or 20°C. These cells were then fixed at different time periods. The adherent cells were fixed by 3% formaldehyde in PBS solution containing 1 mMMgC1, for 25 min. By contrast suspended cells were fixed in methanol (-2O”C, 5 min) since the methanol fixation produced better results as indicated from our preliminary experiments. For actin staining, fixed cells were permeabilized for 10 min in 0.1% solution of Triton X-100 in PBS, washed with PBS, and stained with rhodamine-phalloidin for 45 min. The fluorescent preparations were observed with an Opton microscope equipped with filter sets for selective observation of rhodamine and fluorescein.

Drug treatment of cells. To destroy microtubules, cells were incubated with 1 fig/ml of the colcemid for 1 h at 37°C. When cytochalasin B (CB) was used cells were incubated with CB (10 pglml, 3 h at 37°C).

RESULTS

Capping of EGF-R in Suspended A431 Cells We began our investigation of EGF-R redistribution on suspended A431 cells. When EGF-R was visualized in cells at 0°C using the monoclonal anti-EGF-R antibody (mAb 5A9) followed by FITC-SWAM, uniform cell surface labeling was observed (Fig. 1A). When labeled cells were warmed to 37°C the internalization of the ligand-receptor complexes started rapidly. Obtaining caps during this process is difficult. In order to increase the capping efficiency, the temperature was decreased to 20°C to slow the internalization. Under this condition the internalization began after 60 or 100 min, respectively, with or without the EGF. Within 10 min, patching of the receptors was observed (data not shown). In 30-40 min the EGF-R commonly gathered into a cap on the surface of about 50% cells (Fig. IC). Microscopically, EGF-R caps from A431 cells appeared similar to those induced by lectins or antibodies in lymphocytes [17, 181. Neither EGF nor mAb 5A9 alone could induce detectable capping. Thus, additional cross-linking of the re-

FIG. 1. The EGF-R capping in suspended A431 cells and colocalization of actin and ligand-receptor complexes. Cells were treated with EGF, mAb 5A9, and FITC-SWAM at 0°C and then transferred to 20°C for induction of the EGF-R capping. At different periods of redistribution cells were fixed and permeabilized by methanol and actin was labeled with rhodamine-phalloidin. The localization of both EGF-R (A) and actin (B) at 0°C is uniform over the plasma membrane. After 30 min of incubation of the cells at 20°C the ligandreceptor complexes are capped (C) and similar localization of actin is observed (D). Objective, 100X.

ceptors by the second antibody may be necessary for the appearance of visible clustering. At different stages of the ligand-receptor redistribution the cells were fixed, permeabilized, and labeled with rhodamine-phalloidin for localization of actin (Figs. 1B and 1D). When there was no redistribution of surface receptors, the uniform labeling of actin under plasma membrane was observed (Fig. 1B). But underneath each cap a large accumulation of actin was detected (Fig. lD), suggesting that microfilaments and EGF-R interacted closely during the cap formation. Capping of the EGF-R in Adherent A431 CeuS Adherent A431 cultures grew in groups of closely associated flat cells (Fig. 2B). After these cells were treated with the ligands at O”C, the EGF-R distributed randomly over the plasma membrane. The cell-cell contacts increased the EGF-R labeling slightly (Fig. 2A). By raising the temperature to 20°C the EGF-R showed redistribution which proceeded in a manner similar to that of the suspended cells following the patching, capping, and internalization stages. The EGF-R cluster formed independently within 15-30 min, whether or not EGF was added together with the antibodies, and occurred at cell margins, but never at the site of cell-cell contacts (Fig. 2C). These caps were distinct under phase microscopy (Figs. 2D and 2E). The addition of EGF influenced the initiation time of internalization, i.e., it began 45 or 90 min with or with-

50

KHREBTUKOVA

ET AL.

FIG. 2. The EGF-R capping in adherent A431 cells. Cells were treated with the ligands (the same as in Fig. 1) at 0°C. (A, B) Fluorescent and phase-contrast micrograph of the same group of cells at 0°C. The localization of the EGF-R is more or less uniform over the plasma membrane (A). After transference of the treated cells to 20°C the ligand-receptor complexes are capped. (C) Cells fixed after 20 min of incubation at 20°C. Such caps are distinctly visible under the phase microscope (compare D and E). After 60 min of incubation at 20°C the internalization of ligand-receptor complexes begins (F, G). The ligand-receptor complexes are visible by means of fluorescent microscopy as indistinct areas inside the cell. Objective, 40X.

out addition of EGF, respectively. After internalization the ligand-receptor complexes were observed in central regions inside the cell as indistinct areas (Figs. 2F and 2G).

Actin Cytoskeleton of Adherent A431 Cells under Normal and Experimental Conditions The actin cytoskeleton under normal growth conditions at 37°C contains four main structural types: (a) bundles in numerous microvilli (Figs. 3A and 3C!), (b) a belt array of circular bundles at cell margins (Figs. 3B and 3C), (c) cytoplasmic fiber bundles oriented in different directions (Figs. 3A-3C), and (d) some amorphic aggregates in the central region of cytoplasm (Figs. 3A and 3C). Sometimes polygonal arrays of actin filaments in cytoplasm were noticed (Fig. 3A).

When cells were transferred from 37 to O”C, the actin organization was altered as follows: the number of microvilli decreased, the stress-fiber-like structures disappeared almost completely, and numerous uniformly distributed small aggregates arose in the cytoplasm. However, the submembrane layer of microfilaments was preserved (Fig. 4A). After returning cells back to 37°C initial actin structures were restored, cytoplasmic bundles of actin filaments appeared, and the small aggregates disappeared (Fig. 4D). At 20°C the above process started at 45 min of incubation and ended after 2 h. When cells were treated at 0°C with EGF or mAb 5A9 a difference in the cell edge could be seen (Figs. 4B and 4C). This might represent early changes induced by ligand-receptor binding, although further organization of actin filaments and redistribution of receptors do not occur, due to the low temperatures.

51

FIG. 3. Fluorescent labeling of adherent A431 cells for F-actin with rhodamine-phalloidin under normal growth conditions (37°C). (A-C) Different examples of distribution of actin in the cells. There are four main actin structures in the cytoskeleton of A431 cells: bundles in surface microvilli (A, C); cortical circular bundles under the plasma membrane on the cell margin (B, C); cytoplasmic stress-fiber-like bundles oriented in different directions (A-C); and some amorphic aggregates in the central part of the cytoplasm (A, C). Objective, 100X.

The Redistribution of Actin During EGF-R Clusterization in Adherent A431 Cells As shown above, the temperature shift from 0 to 20°C in cells treated with the ligands led to clusterization of the surface receptors. Actin always accompanied clusters of the ligand-receptor complexes. Therefore, submembrane circular bundles of microfilaments are involved in the process of redistribution of these complexes (Figs. 5D, 5F, and 5H). Numerous microvilli with actin bundles inside were often observed in the cap area. Stress-fiber restoration at 20°C started later than the capping and more time (1-2 h) was required for a total restoration of cytoplasmic bundles. The redistribution was completed after 30-40 min. The capping and restoration of cytoplasmic bundles seemed to occur simultaneously after cells were transferred to 20°C. In conclusion, therefore, actin stress-fibers are not involved in the process of redistribution of ligand-receptor complexes. As shown in Figs. 5D and 5H, the treatment of cells with colcemid does not influence cluster formation.

Thus intactness of microtubules EGF-R redistribution. Cytochalusin

B Influence

is irrelevant

to the

on the EGF-R Capping

Treatment of cells with CB (10 rglml) for 3 h at 37°C led to the following events. As a result of arborization, cells formed a number of microspikes and filopodia. Actin still persisted in these structures and in microvilli. Neither the submembrane layer of actin bundles nor the stress-fibers occurred. Small aggregates of actin under the plasma membrane as well as large ones in the central part of the cytoplasm could be detected (Fig. 6A). After treatment of these cells with EGF and antibodies at 0°C the ligand-receptor complexes distributed more or less uniformly (not shown). Actin distribution was similar at both 0 and 37°C. At 20°C caps did not form. Some cells had numerous small aggregates of the receptors. Actin submembrane aggregates and receptors were always localized together (Figs. 6B and 6C). These results indicate that EGF-R and actin are still associated.

52

KHREBTUKOVA

FIG. 4. Reorganization temperature. The cells were actin bundles are preserved aggregates of actin occur in (D). This process is started

ET AL.

of actin cytoskeleton of A431 cells at low temperatures and restoration of initial structures after raising the incubated at 0°C for 2 h without any ligands (A), with EGF (B), or with mAb 5A9 (C). The submembrane circular in any case. But independently from ligand treatment the cytoplasmic actin bundles are destroyed and small the cytoplasm. After the transference to 20°C within 2 h the restoration of cytoplasmic fiber bundles is observed 45-60 min after the rise in temperature. Objective, 100X.

After removing CB, cell shape and actin structures were restored; they were accompanied by large aggregates of surface receptors (not shown). DISCUSSION

The present study illustrates that cross-linking of the EGF receptors by the monoclonal antibody mAb 5A9 followed by the second antibody leads to aggregation of the ligand-receptor complexes in A431 cells. The results support previously obtained data for various antiEGF receptor antibodies that show that multivalency of the receptor-recognizing agent is required for a visible macroclustering formation (for review see [22]). By decreasing temperatures from 37 to 20°C to slow the receptor internalization, a single cluster (cap) on the surface of both suspended and adherent A431 cells formed. In suspended cells, the clusterization is similar to the lectin- or antibody-induced patching and capping in lymphocytes and other rounded cells [17-211. As in the case of capping in lymphocytes, submembrane actin appears to localize beneath the EGF-R caps in both sus-

pended and adherent A431 cells. Thus our experimental data support the existence of a connection between the EGF-R and actin filaments as reported by Wiegant et al. [12]. The redistribution of EGF-R on the surface of adherent cells is not similar to either the clearing of the leading edge of crawling fibroblasts from cross-linking receptors [23-251 or the immobilization of surface receptor patches along stress-fibers in nonmotile adherent cells [26,27]. By contrast, in A431 cells, large clusters of EGF-R were formed on the cell margin. The unusual state of actin cytoskeleton after incubation at low temperature could explain the uncommon behavior of cross-linking receptors in adherent A431 cells. Stress-fiber-like bundles were absent in the cytoplasm, and only the cortical circular bundles were found surrounding cellular margins. The restoration of the cytoplasmic bundles begins after capping is completed. These results show that stress-fibers are not required for the receptor clusterization. Perhaps the existence of stress-fibers in a cell determines the direction of receptor complex transposition from the margin to the

THE

EGF-R

CAPPING

53

of actin and ligand-receptor complexes during the capping of EGF-R in adherent A431 cells. EGF-R (left) and actin G I. 6. Colocalization in the same cell. The cells were treated with EGF, mAb 5A9, and FITC-SWAM at 0°C. The distribution of EGF-R (A) and t) localization subn le mbrane actin (B) is more or less uniform. No fiber bundles of actin are seen in the cytoplasm at 0°C. After the transference of the treated cells tf o 20°C the redistribution of ligand-receptor complexes and submembrane actin is observed. (C) Formation of some small clusters bati :h -like) after 5-10 min of incubation; (E) formation of a single cap on the cell margin within 20 min. The treatment of the cells by colcemid does n ot affect the cap formation (G). The submembrane actin always accompanies the surface patches and caps (D, F, H). Objective, 100X.

54

KHREBTUKOVA

FIG. 6. The effect of CB on the redistribution of the ligand-receptor complexes. The cells were treated with CB (10 pg/ml) for 3 h at 37”C, and then the above procedure of ligand treatment at 0°C and incubation at 20°C for inducing redistribution of EGF-R were used in the presence of CB (10 cg/ml). (A) Localization of actin labeled with rhodamine-phalloidin in cells fixed at 37°C after incubation for 3 h with CB; (B, C) localization of the ligand-receptor complexes (B) and actin (C) after the incubation of CB-treated cells at 20°C for 45 min. Objective, 100X.

center. The absence of cytoplasmic bundles leads to cap formation in adherent cells, which is similar to that found in suspended cells. However, it could also be due to some properties of the EGF receptor. The phenomenon of actin bundle destruction at 0°C was unexpected. It has not been described in other cell types nor has it been found in the HeLa, CHO, HOSTE, and normal fibroblasts we examined (unpublished data). The cooling-induced reorganization of cytoskeleton is unusual. Since it is not typical for the behavior of actin structures, the destruction of microfilaments might be dependent on disassembling of microtubules at low temperatures [28, 291. However, in the colcemid-treated A431 cells, the microtubule-disrupting agent did not alter the actin structures. Colcemid also had no effect on the redistribution of EGF-R (Figs. 5G and 5H). Therefore, disintegration of actin structures at low temperatures and cluster formation are not connected with the state of microtubules.

ET AL.

In contrast to colcemid, CB does inhibit the capping of EGF-R. After the CB treatment all actin structures and inclusive submembrane bundles are destroyed (Fig. 6A) and no cap formation occurs. As the destruction of cytoplasmic bundles at 0°C does not affect the process, the capping may depend on the intact cortical circular actin bundles. Note that aggregates of EGF-R still coincide with the patches of submembrane actin (Figs. 6B and 6C). This suggests some CB-resistant connection between the surface receptor and cortical actin filaments. In their early work on the association between the EGF-R and the Triton-insoluble cytoskeleton in A431 cells Landreth et al. [ll] also showed that the connection site is CB-resistant. They concluded that actin could not be involved in the connection. Our data indicate that EGFR still remains associated with actin filaments after CB treatment. On the basis of the results described above, the following conclusions concerning the participation of different parts of actin cytoskeleton in the process of ligandreceptor complex redistribution can be made. First, the connection of receptor complexes with the intact cortical actin bundles is necessary for the capping to proceed. This is true for suspended or adherent cells without any stress-fiber-like bundles. In cells with more complicated actin cytoskeleton with stress-fibers in the cytoplasm, the association of cross-linking receptors with both submembrane and cytoplasmic bundles of actin is likely to occur. This leads to the immobilization of receptors in nonmotile cells [27]. During the cell movement, the receptor patches run backward over the cell surface [24, 251. Preexisting bundles of microfilaments can direct this transposition. We are very grateful to Dr. T. R. Chen and Mrs. A. A. Staviskaya for helpful discussion of the manuscript. We thank Dr. A. D. Sorkin for providing us with EGF.

REFERENCES 1. 2. 3. 4. 5. 6. 7. a. 9. 10. 11.

Geiger, B. (1983) Biochim. Biophys. Actu 737,305-341. Mangeat, P.-H. (1988) Btil. Cell 64,261-281. Carraway, K. L., and Carraway, C. A. C. (1989) Btichim. Biophys. Acta 988,147-171. Lassing, I., and Lindberg, U. (1988) Exp. Cell Res. 174, 1-15. Geiger, B., Volk, T., Volberg, T., and Bendori, R. (1987) J. Cell Sci. S(Suppl.), 251-272. Burridge, K., Molony, L., and Kelly, T. (1987) J. Cell Sci. B(Suppl.), 211-22s. Burridge, K., Fath, K., Kelley, T., Nuckolls, G., and Turner, C. (1988) Anna Rev. Biochem. 4,487-525. Carpenter, G., and Cohen, S. (1979) Anna Rev. Biochem. 40, 193-216. Carpenter, G. (1987) Annu. Rev. Biochem. 56,881-914. Stochek, C. M., and Carpenter, G. (1983) J. Cell. Biochm. 23, 191-202. Landreth, C. E., Willians, L. K., and Rieser, G. D. (1985) J. Cell Biol. 101, 1341-1350.

THE 12.

13. 14.

15.

16. 17. 18. 19.

EGF-R

55

CAPPING

Wiegant, F. A. C., Blok, F. J., Defize, L. H. K., Linnemans, W. A. M., Verkley, A. J., and Boonstra, J. (1986) J. Cell Biol. 103,87-94.

20.

Condeelis,

21.

Pastemak, C., Spudich, (Lmdon) 341,549-551.

Roy, L. M., Gittinger, C. K., and Landreth, G. E. (1989) J. Cell. Physiol. 140, 295-304. van Bergen en Henegouwen, P. M. P., Defize, L. H. K., de Kroon, J., van Damme, H., Verkley, A. J., and Boonstra, J. (1989) J. Cell. B&hem. 39,455-465. Sorokin, A. B., Nesterov, A. M., Sorkin, A. D., Ignatova, T. N., Galaktionov, K. I., and Kudryavtseva, N. V. (1989) Tsitologiu 3 1,549-555. Lin, D. C., and Lin, S. (1978) J. Cell Biol. 263, 1415.

22.

De&e, L. H. K., Mummery, C. L., Moolenar, W. H., and de Laat, S. W. (1987) Cell Difer. 20.87-102. Abercrombie, M., Heaysman, J. E. M., and Pegrum, S. M. (1972) Exp. Cell Res. 73, 536-539. Vasiliev, J. M., Gelfand, I. M., Domnina, L. V., Dorfman, N. A., and Pletyushkina, 0. Y. (1976) Proc. Natl. Acad. Sci. USA 73, 4085-4089. Heath, J. P. (1983) Nature (London) 302, 532-534. Ash, J. F., and Singer, S. J. (1976) Proc. Natl. Acad. Sci. USA 73, 4575-4579. Ash, J. F., Louvard, D., and Singer, S. J. (1977) Proc. Natl. Acad. Sci. USA 74,55&G5563. Weisenberg, R. C. (1972) Science 177,1104-1105. Borisy, G. G., and Olmstead, J. B. (1972) Science 177, 1196 1197.

Taylor, R. B., Duffis, W. P. H., Raff, M. C., and de Petris, S. (1971) Nature New Biol. 233,225-229. Schreiner, G., and Unanue, E. R. (1976) Adv. Immunol. 24,37165. Bourguignon, L. Y. W., and Bourguignon, Cytol. 87,195-224.

Received September 13,199O Revised version received November

30,199O

G. J. (1984) Znt. Rev.

23. 24.

25. 26. 27. 28. 29.

J. (1979) J. Cell Biol. 80, 751-758. J. A., and Elson, E. L. (1989) Nature