EXPERIMENTAL
CELL RESEARCH
196,255-263
(1991)
Actin-Binding Proteins Involved in the Capping of Epidermal Growth Factor Receptors in A431 Cells KATARZYNA KWIATKOWSKA,**’ IRINA A. KHREBTUKOVA,? DINA A. GIJDKOVA,~ GEORGE P. PINAEV,~ AND ANDRZEJ SOBOTA* *Nencki Institute of Experimental Biology, Department of Cell Biology, 3 Pasteur Street, PL-02093 Warsaw, Poland; and tlnstitute of Cytology of the Academy of Sciences, 4 Tikhretsky Avenue, 194064 Leningrad, USSR
A capping process of epidermal growth factor receptors (EGF-Rs) was used for the study of the relation between the receptors and the actin-binding proteins (spectrin, vinculin, annexin I) that may be involved in EGF-R-cytoskeleton interaction. In intact, adherent A431 cells, EGF-Rs were diffusively distributed on the cell surface. Spectrin, vinculin, and annexin I were located beneath the plasma membrane. An abundance of EGF-Rs as well as submembrane proteins was observed in regions of membrane ruffles and cell-cell contacts. Annexin I was localized also in cytoplasm being attached to Alamentous structures surrounding the nucleus and extending to the cell periphery. Under polyvalent ligand treatment, EGF-Rs of adherent cells were aggregated on one side of the cell. Spectrin, vinculin, and annexin I dislocated together with EGF-Rs and were concentrated under plasma membrane at regions where cap formation took place. In suspended A431 cells only spectrin was located under the plasma membrane whereas annexin I and vinculin were diffusively distributed through the cells. During cap formation only spectrin was colocalized with EGF-Rs. The results confirmed the major role of spectrin as a receptor-mi0 1991 Academic press. he. CrOfilament linking protein.
INTRODUCTION
Several immunocytochemical and biochemical data indicate that if cells are exposed to a polyvalent ligand the cell-surface receptors become linked to the cytoskeleton and, as a result of actin-myosin interaction, are translocated and aggregated on one side of the cell [l31. Hence, assembling receptors into cap may serve as a model for studying mechanisms of receptor-cytoskeleton interactions. There have been suggestions that in human carcinoma epidermoid A431 cells the epidermal growth factor receptor (EGF-R)-cytoskeleton interac-
’ To whom reprint
requests should be addressed.
tion may be involved in the mechanism of epidermal growth factor (EGF) action [4-61. Several biochemical data suggest that the association of EGF with EGF-Rs stimulates interactions between EGF-Rs and cytoskeleton [6]. However, only a few immunoelectron microscopy observations, which suggest a linkage between EGF-Rs and microfilaments, are available [7]. Recently, it was found that the actin bundles delineating cell margins indeed take part in capping of the EGF-Rs of A431 carcinoma cells [S]. The submembrane actin-binding proteins are assumed to be involved in the linkage of microfilaments to EGF-Rs. Spectrin is one of the proteins that plays an important role in anchoring actin filaments to membrane proteins as it has been demonstrated in human erythrocytes [g-11]. The participation of spectrin, together with actin, in capping the receptors in lymphocytes [2, 12, 131, DictyosteZium discoideum [14-161, and fibroblasts [17, 181 indicates that spectrin may be considered a linker in receptor-microfilaments interactions. Another protein that may be involved in EGF-R-microfilament interactions is annexin I (calpactin 2). This protein, as well as the closely related annexin II (calpactin l), is able to interact with the membrane and with the F-actin in a Ca’+-dependent manner [19-221. In A431 cells treated with EGF, annexin I is phosphorylated by the EGF-R tyrosine kinase [23, 241 indicating the close structural relations between EGF-R and this protein. Vinculin may also be involved in binding the microfilament bundles to plasma membrane [25,26]. This protein, similar to annexin I, is a substrate for the pp66’-” tyrosine kinase [27,28], with an activity similar to that of EGF-R kinase [29]. In this study we examine the localization and redistribution of spectrin, vinculin, and annexin I during aggregation of EGF-receptors in A431 cells. The association of these proteins with EGF-R depends on the attachment of the cells to a substratum. It seems that spectrin is a protein that is actively engaged in the capping of EGF-Rs.
255 Copyright Q 1991 by Academic Press, Inc. All rights of reproduction in any form reserved
256
KWIATKOWSKA
MATERIALS
AND METHODS
Cell culture. Human epidermoid carcinoma A431 cells, obtained from the Cell Culture Collection of the Institute of Cytology of the Academy of Sciences, Leningrad, the USSR, were cultivated in Eagle’s medium supplemented with 10% bovine serum and glutamine. A431 cells were grown on glass at 37°C under 5% CO, atmosphere. Antibodies. o-Spectrin was isolated from a “0.6 M KJ extract” of chicken erythrocyte membranes [30]. The extract was applied to a Sephacryl S-200 column equilibrated and developed with solution of 0.6 M KJ, 0.5 m&f EDTA, 0.2 mM dithiothreitol (DTT), 10 n&f TRIS, pH 8.0. The void volume fractions were loaded onto a calmodulin-Sepharose column in the presence of 5 m&f CaCl, and 2 M KJ [31]. a-Spectrin selectively bound to the calmodulin column was eluated in the presence of 10 m&f EGTA and used for immunization of rabbits [32]. The o-spectrin antibody was affinity purified from the antisera according to procedure of Talian et al. [33]. Anti-annexin I antibody was a generous gift from Dr. C. Comera, Pasteur Institute, Paris, France. Anti-vinculin antibody was kindly provided by Dr. V. E. Koteliansky from the Center of Cardiology, Moscow, the USSR. Monoclonal anti-EGF-R IgG mAb5A9 was kindly supplied by Dr. A. B. Sorokin (Institute of Cytology Acad. Sci. USSR, Leningrad). Rhodamine-labeled and fluorescein-labeled antirabbit IgG (GAR-TRITC and GAR-FITC, respectively) were purchased from Sigma. Fluorescein-labeled swine anti-mouse IgG (SWAM-FITC) was obtained from the Institute of Sera and Vaccines, Prague, Czechoslovakia. Zmmunofluorescence microscopy. Adherent intact A431 cells were fixed with 3% formaldehyde prepared in phosphate-buffered saline (PBS) containing 1 m&f MgCl, for 30 min at room temperature, according to Schlessinger and Geiger [4]. Nonspecific antibody binding was blocked by 100 mA4 glycine in PBS (10 min, room temp). EGF-R was labeled with monoclonal mAb5A9 anti-EGF-R antibody (1:lOO diluted in PBS, 30 min, room temp) followed by incubation with SWAM-FITC (1:lO diluted in PBS, 30 min, room temp). Then the samples were permeabilized with solution of 0.1% Triton X-100 in PBS (10 min on ice) and treated with anti-o-spectrin (1:lO in PBS), anti-vinculin (1:20), anti-annexin I (1:lOO) antibodies, or nonimmune rabbit serum (1:lOO) for 1 h at 37°C. The samples were carefully washed and then incubated in a solution of GAR-TRITC (1:50 in PBS, 1 h, 37’C). Actin was observed by staining the cells with rhodamine-labeled phalloidin (Sigma, St. Louis, MO) at a concentration of 0.5 *g/ml. Using FITC-labeled second antibodies together with rhodamine-labeled phalloidin allowed the annexin I and actin to be localized simultaneously in the same cell. For studies of EGF-Rs redistribution two sets of experiments were performed: (1) the adherent A431 cells were treated with polyvalent ligands, i.e., they were exposed to solution of mouse anti-EGF-R antibody followed by exposure to a solution of anti-mouse SWAM-FITC antibody (20 min, 0°C); or (2) the intact cells were pretreated with EGF (200 rig/ml) in the growth medium at 0°C for 1 h. After washing, the cells were incubated with polyvalent ligands as described above. In both experiments, the living cells were subsequently transferred into the growth medium at 20°C and fixed at different stages of EGFRs redistribution. The cytoskeletal proteins were localized in Triton X-lOO-permeabilized cells as described above. For the observation of EGF-R redistribution in suspendedcells, the adherent cells were detached from substratum by treatment of culture with 0.02% EDTA in PBS solution. The detached cells were incubated in growth medium with serum at 37°C for 2 h. The cells were then transferred into an ice bath, treated with EGF, labeled with EGF-R antibody, and then labeled with SWAM-FITC by a procedure similar to that used for adherent cells. The redistribution of EGF-Rs in suspended A431 cells in the absence of EGF was not tested. The labeled cells were subsequently incubated (lo-50 min) at room temperature, fixed and permeabilized in methanol (-ZO’C, 5 min), and
ET AL. later stained for the actin-binding proteins using GAR-TRITC as the second antibody. The cells were intensively rinsed with PBS after each incubation and finally embedded in mixture of 90% glycerol and 5% propyl gallate. The specimens were examined using an OPTON microscope with filters for FITC and TRITC fluorescence. In a series of experiments the FITC-labeled samples were checked with green as well as with red filter sets. The same checking procedure was used for some TRITCstained cells. The overlapping of both fluorescences was barely seen. Zmmunoelectron microscopy. For localization of EGF-R the A431 cells were treated as in the fluorescence microscopy procedure used for the suspended cells. Briefly, the detached cells were incubated at 37°C for 2 h, then transferred to 0°C and exposed to EGF (200 rig/ml) followed by anti-EGF-R and SWAM-FITC antibodies. Finally, the samples were incubated with protein A-gold complex (15 nm gold particle size). Thereafter, the cells were shifted to room temperature and, at different stages of cap formation, fixed in a mixture of 6% formaldehyde/0.5% glutaraldehyde containing 50 m&f KC1 and 5 mM MgCl, in 100 mA4 phosphate buffer, pH 7.4, (2 h, room temp). After being washed, the cells were postfixed in 1% 0~0, solution in the same buffer (1 h, room temp). The samples were dehydrated in a series of ethanol and embedded in Epon 812. For postembedding labeling, the gold sections of A431 cells were mounted to formvar/carbon-coated nickel grids. In order to remove the surface layer of resin the samples were exposed for 30 s to a saturated alcohol solution of sodium hydroxide diluted 1:l with 96% ethanol [34]. After being washed with ethanol, distilled water, and PBS, the grids were put into 0.5% defatted milk in PBS to reduce nonspecific binding. The o-spectrin antigen was visualized by incubation of the samples with affinity-purified antibody directed against o-spectrin from chicken erythrocytes (1 pglml, 4”C, overnight), followed by incubation with gold-labeled protein A (8 nm size of gold particles). The control samples were treated with nonimmune rabbit serum diluted with PBS (100 gg of protein/ml). The unbound antibodies and protein A-gold were removed by extensive washing (5 X 15 min each). All solutions were prepared in PBS containing 1% bovine serum albumin and 0.1% Tween 20 [35]. The sections were finally rinsed with distilled water and counterstained with lead citrate and uranyl acetate. The samples were examined with an JEM 1OOB electron microscope. Analytical procedures. Gel electrophoresis was carried out in 10% polyacrylamide gel (SDS-PAGE) as described by Laemmli [36]. Proteins were transferred from the gel to nitrocellulose sheets for 2 h at 500 mA according to the procedure of Towbin et al. [37]. Immunoblotting was performed using appropriate antibodies followed by peroxidase-conjugated protein A binding. Enzymatic activity was visualized in the presence of H202 and diaminobenzidine (DAB).
RESULTS Characterization
of the Antibodies
a-Spectrin, vinculin, and annexin I were identified and localized in A431 carcinoma cells by means of polyclonal antibodies. The monospecific anti-a-spectrin antibody recognized a 240-kDa molecular weight polypeptide in whole A431 cell homogenates (Figs. lb and lc). After low temperature storage the homogenate contained 240- and 150-kDa polypeptides that crossreacted with the antibody (not shown). The 150-kDa polypeptide seems to be a typical product of proteolytic degradation of a-spectrin [ 381. Antiserum against human vinculin cross-reacted only with a 130-kDa polypeptide in whole cell homogenate
CAPPING
IN A431 CELLS
a b FIG. 1. Immunoblotting analysis of actin-binding proteins of A431 cells. Proteins of whole cells were separated by SDS-PAGE (b), transferred to nitrocellulose sheets, and treated with anti-cu-spectrin (c), anti-vinculin (d), anti-annexin I (e). (c) Anti-or-spectrin recognizes the 240-kDa polypeptide. (d) Anti-vinculin and (e) anti-annexin I react with polypeptides of 130 and 35 kDa, respectively. The molecular weight markers indicated on the left are myosin (200 kDa), ,&palactosidase (116 kDa), phosphorylase B (97 kDa), and ovalbumin (45 kDa).
(Fig. Id), which corresponded to the molecular weight of vinculin. Anti-annexin I serum recognized a low molecular weight polypeptide of 35 kDa in A431 cell homogenate (Fig. le). Thus, the 240,130, and 35 kDa polypeptides, revealed by immunoblotting analysis in A431 carcinoma cells, can be considered cr-spectrin, vinculin, and annexin I, respectively. Immunofluorescent Localization of EGF-R, Spectrin, Vinculin, and Annexin I in Adherent Cells
In intact adherent A431 cells fixed at room temperature and labeled for EGF-R the antigen was diffusively distributed on the surface of cells (Fig. 2A). However, the regions of ruffles and cell-cell contacts displayed stronger staining of receptor (Fig. 2A, arrows). Fixed, permeabilized adherent cells exposed subsequently to anti-spectrin, anti-vinculin, or anti-annexin I displayed a delicate fluorescence that was more intensive under plasma membrane in the periphery of the cells (Figs. 2B-2D). The labeling was especially well seen in regions of ruffles and cell-cell contacts (Figs. 2B-2D, arrows), similar to the case of EGF-R. Moreover, spectrin, vinculin, and annexin I formed a delicate network structure throughout the cell (Figs. 2B-2D, asterisks). In some cells, however, annexin I was also found in the cytoplasm attached to nonidentified filamentous structures that were concentrated around nucleus and extended to the cell periphery (Fig. 2G). When the cells were double labeled with phalloidin and anti-
257
annexin, no correlation between localization of actin filaments and annexin-associated filamentous structures was observed (Figs. 2G and 2H). In control cells treated with nonimmune rabbit serum instead of the first antibody a very faint labeling was visible (Figs. 2E and 2F). The quite different patterns of fluorescence observed in the cell double labeled for annexin I and actin (compare Figs. 2G and 2H) were the positive control for the proper sets of filters applied in our examinations. The bright green FITC fluorescence attributed to annexin was not observed with the red (TRITC) filter set. On the contrary, the rhodamine fluorescence of phalloidin was seen only at the filter set proper for TRITC. These observations allowed us to exclude the possible overlapping of FITC and TRITC emissions. Redistribution of Actin-Binding Proteins during Capping of EGF-Rs in Adherent Cells
It is suggested that a set of proteins involved in the process of binding of microfilaments to ligand-receptor complex is metabolically regulated [39]. Therefore it was interesting to examine whether the presence of EGF together with polyvalent ligands would define the selection of proteins involved in linkage of EGF-Rs to microfilaments during capping. For this purpose, the adherent A431 cells were treated in two different ways: (1) the cells were exposed to EGF and then to polyvalent ligands (anti-EGF-Rs and SWAM-FITC antibodies), or (2) the incubation with EGF was omitted. The whole procedure of EGF-R staining was carried out at 0°C to prevent internalization of receptors. When the EGFpolyvalent ligand-exposed cells were transferred from 0 to 2O”C, after 30 min EGF-Rs aggregated into caps which were formed along cell margins free of cell-cell contacts (Figs. 3A, 3C, and 3E, arrows). During aggregation of EGF-Rs all three actin-binding proteins comigrated to where the accumulation of EGF-Rs took place. The colocalization of EGF-R caps and aggregates of spectrin (Figs. 3A and 3B), vinculin (Figs. 3C and 3D), and annexin I (Figs. 3E and 3F) was observed in adherent cells. When the cells were exposed only to polyvalent ligands, no differences in the protein composition of subcaps, versus the EGF-polyvalent ligand-treated cells, were observed. Spectrin (Fig. 3H) as well as vinculin and annexin I accumulated at the locations of EGF-R aggregation. However, in these cells the aggregation of EGFRs proceeded slowly, and during 30 min at 2O”C, 2-3 patches of EGF-Rs were formed (Fig. 3G, arrows). Colocalization of EGF-R and submembrane actinbinding proteins was observed during the whole process of cap formation (compare Figs. 3A and 3B with 3G and 3H). It should be noted, however, that only a part of cellular spectrin, vinculin, and annexin I was shifted to
258
KWIATKOWSKA
ET AL.
FIG. 2. Localization of EGF-R and actin-binding proteins in intact, adherent A431 cells. (A) Cells stained for EGF-R. The antigen is diffisively distributed on the cell surface. (B, C, D) Cells stained for spectrin, vinculin, and annexin I, respectively. The abundance of submembrane proteins as well as EGF-R in regions of ruffles and cell-cell contacts is shown (arrows). Asterisks denote the areas of the ce 11s where the tiny network formed by spectrin (B), vinculin (C), and annexin I (D) can be seen. (G, H) A431 cells double-labeled for annexin I alnd F-actin, respectively. (G) The anti-annexin I antibody stains the filaments concentrated in the perinuclear region from which they extel nd toward the cell periphery. There is no colocalization between these filaments and actin bundles visualized by rhodamine-conjugatedphalloid lin in (H). (E) Control sample. The cells were treated with nonimmune rabbit serum and then exposed to GAR-TRITC antibody. (F) Pha ise contrast image of(E). A-D, X520; E, F, X400, G, H, X700.
CAPPING
IN A431 CELLS
259
1PIG. 3. (A-F) Coincident redistribution of EGF-Rs and actin-binding proteins induced by anti-EGF-R and SWAM-FITC antibodies in shows the cocaps formed after 30 min at room temperature by EGF-Rs (A) and spectrin (1B), the presence of EGF. Double immunotluorescence EG F-Rs (C) and vinculin (D), EGF-Rs (E) and annexin I (F). Locations of cap formation are indicated by arrows. The bright labeling of contacts is still observed (arrowheads). (G, H) Redistribution of EGF-Rs and spectrin induced by polyvale mt SUb Imembrane proteins in cell-cell liga mds in the absence of EGF. The cells, fixed after 30 min of cap formation, display a few patches of EGF-Rs (G, arrows). (H) Correspondi ng red istribution of spectrin accumulated under plasma membrane as copatches (arrows). A-F, X740; G, H, X520.
260
KWIATKOWSKA
ET AL,
the region of the formed cap. Even in the cells containing the completely formed caps of EGF-Rs we still observed the labeling of the actin-binding proteins in the submembrane layer of the cells, especially in regions of cell-cell contacts (Figs. 3B, 3D, and 3H, arrowheads). Participation of Spectrin in Capping of EGF-Rs in Suspended A431 Cells
Examination of distribution of actin-binding proteins in detached A431 cells revealed that only spectrin did not change its cortical localization which was similar to adherent cells. The antigen was seen as a continuous submembranous layer that coincided with uniformly spread surface-labeled EGF-R (Figs. 4A and 4B) as well as with cortical F-actin 181. Annexin I and vinculin, contrary to spectrin, were diffusively distributed through the cell and seemed to no longer underlie the plasma membrane (Figs. 4F and 4H). In addition, in a part of cells labeled with anti-annexin I the nuclear envelope was also stained (Fig. 4H). The uniform labeling of the proteins in cytoplasm was much stronger than it was in the case of nonspecific binding of antibodies (compare Figs. 4F and 4H with 45). During aggregation of EGF-Rs only spectrin was accumulated at the region of cap formation (Figs. 4C and 4D). At the same time, no changes in the diffusive distribution of vinculin and annexin I were observed (Figs. 4F and 4H). Once the capping was finished the number of EGF-R aggregates associated with spectrin decreased. The results presented above were confirmed by immunoelectron microscopy. When the suspended A431 cells were fixed during the first steps of EGF-Rs capping and labeled with anti-spectrin, the aggregated gold particles marking the places rich in cu-spectrin, appeared in the submembrane layer of cells in the close vicinity of EGFRs (Fig. 5A). A prolonged (45 min) exposure of EGF-Rs to polyvalent ligands allowed the first moments of internalization of the receptors to be caught. At this step the formed cap contained numerous small vesicles with EGF-R inside them. At this moment, a spatial separation of EGF-R and spectrin aggregates was visible (Fig. 5B). The specificity of spectrin labeling was tested by substituting anti-spectrin antibody for nonimmune rabbit serum. Only a few, randomly dispersed, gold particles could be found in the control samples (Fig. 5C). DISCUSSION
In this study we examined the distribution of spectrin, vinculin, and annexin I in human carcinoma A431 cells. These proteins, being actin- and membrane-binding ones, may serve as a linking system between actin filaments and cell-surface EGF-Rs. The process of ag-
FIG. 4. Suspended A431 cells stained for both surface EGF-R (left) and actin-binding proteins (right). (B) In intact cells spectrin is localized uniformly under plasma membrane reflecting the distribution of EGF-R on the cell surface (A). During cap formation spectrin (D) accompanies the receptors (C). Vinculin (F) and annexin I (H) are dispersed diffusively through the cell and do not redistribute in capped cells. The noncoincident distribution of EGF-Rs stained with FITC, and of annexin I and vinculin visualized by TRITC-fluorescence, points to the correct sets of filters used. (J) Control cells incubated with nonimmune rabbit serum instead of the first anti-intracellular protein antibodies. x500.
gregation of EGF-Rs induced by polyvalent ligands was exploited as a model system for studying the mechanisms of cytoskeleton-receptor interaction. As it was
CAPPING
IN A431 CELLS
261
shown in adherent A431 cells, spectrin, vinculin, and annexin I were located beneath the plasma membrane and were accumulated in the regions of EGF-R cap formation. In suspended A431 cells only spectrin colocalized with cap of EGF-Rs. Vinculin and annexin I were no more clearly observed beneath the plasma membrane of suspended cells and were dispersed through the cytoplasm in resting as well as in capped cells. Changes in cellular distribution of vinculin and annexin I may reflect a phenomenon of general reconstruction of the cytoskeleton in A431 cells as 8 result of detachment of cells from the substratum. A similar phenomenon has been observed in suspended fibroblasts [40]. It is known that binding of annexins I, II (calpactin 2,l) to F-actin, spectrin, and membrane phospholipids is a Ca2+-dependent process [ 19-221. Hence it is possible that changes in ionic equilibrium in suspended A431 cells induced by EDTA, used for detachment of cells from the substratum, might evoke dissociation of the proteins from the plasma membrane. Dissociation of vinculin from the plasma membrane in suspended cells is considered as the loss of cell-substratum and cell-cell contacts. A similar process occurs in the cells transformed by oncogenic viruses. This transformation leads to a significant reduction of cell adhesion correlating with increased phosphorylation of vinculin and redistribution within the cell [26,40]. Vinculin in adherent A431 cells was located uniformly beneath the plasma membrane (Fig. 2C) although the used antibody stained selectively focal contacts in human fibroblasts (unpublished). However, in l-day-old cells no focal contacts were observed by means of interference reflection microscopy. The adhesion of these cells to the substratum probably is due to small and incomplete dot contacts in which actin filaments, invisible with fluorescence microscopy, are anchored within the plasma membrane via vinculin [41]. Later, these sites are transformed into completely formed focal contacts that were
FIG. 5. Ultrastructural analysis of the relation between the EGF-R and spectrin localization at different stages of the capping process.EGF-R of suspended cells was labeled by anti-EGF-R and SWAM-FITC antibodies, and additionally, with protein A-gold particles (15 nm in diameter) and then induced to aggregate for either 30 min (A, C) or 45 min (B). Spectrin was localized in the cells by means of the postembedding technique using anti-a-spectrin antibody followed by protein A-gold particles (8 nm). (A) EGF-R is localized on the plasma membrane surface (arrowheads) close to submembrane spectrin aggregate (arrow). (B) Internalization of EGF-Rs is beginning. EGF-Rs are found in endosome-like structures (lower part of micrograph, arrowheads) whereas spectrin aggregates remain under plasma membrane (arrows). (C) Control sample. The cells were stained for EGF-R and after embedding in resin were subsequently exposed to nonimmune serum and protein A-gold conjugates (6 nm). Arrowheads mark the location of EGF-Rs (gold particles of 15 nm). Nonspecific postembedding labeling of the sections is barely detectable. A, X60,000, B, C, ~48,000.
262
KWIATKOWSKA
well developed in 5-day-old A431 cell culture (unpublished). Reorganization of the cytoskeleton in suspended cells correlates with changes in dynamics of EGF-R. In these cells the process of internalization of EGF-Rs is twice as slow as in adherent ones [42]. Also, only 30-50% of the receptors is associated with the Triton-insoluble fraction of suspended A431 cells [43, 441. Detachment of A431 cells from the substratum probably modifies EGFR-cytoskeleton interactions. These changes may explain the observed differences in composition of actinbinding proteins involved in the association of actin filaments with EGF-R-polyvalent ligand complex in adherent and suspended cells. On the other hand, these results seem to indicate that among actin-binding proteins only spectrin is actively engaged in binding of the peripheral actin bundles with EGF-Rs. The concentration of vinculin and annexin I under plasma membrane of adherent cells in regions of EGF-Rs aggregation may be a secondary phenomenon. These proteins can be transported to the EGF-R cap region as a result of interaction with actin filaments. Linking of cytoskeletal filaments with cell-surface receptors is probably an essential function of spectrin family proteins [45]. It has been suggested that spectrin is involved in receptor-microfilament interactions during such processes as: (i) control of mammalian erythrocyte shape changes [46], (ii) regulation and exposition of receptors in neurons [47,48], (iii) control of epithelial cell polarity [49,50], and (iv) axonal transport [51]. As the process of cap formation was finishing, an increased amount of EGF-R caps not associated with spectrin aggregates was detected. The electron microscopic observations showed the spatial separation of EGF-R and spectrin started at the moment of internalization of polyvalent-EGF-R complex (Fig. 5B). It seems that dissociation of spectrin from the complex is the first step of the internalization of the cell-surfaceaggregated receptors. A similar mechanism is postulated for spectrin in exocytosis [52, 531. Dissociation of spectrin from the membrane of secretory vesicles and disintegration of the spectrin-actin network seems to be necessary for fusion of vesicles with plasma membrane [53,54]. The role of spectrin as well as other actin-binding proteins in the process of aggregation and internalization of EGF-Rs induced by EGF in A431 cells under physiological conditions is still an open question [42, 551. Proteins p35 (annexin I) and p36 (annexin II) become intensively phosphorylated in the cells exposed to growth factors [29,56-581. Some biochemical and cytochemical data indicate that submembrane-localized annexin I is the specific substrate of the EGF-R kinase [23, 24,58,59]. However, the observations of adherent cells showed that only a part of annexin I is connected to the
ET AL.
plasma membrane. A large fraction of this protein was dispersed through the cell, where it was associated with filamentous structures. The cytoplasmic pool of annexin I could be the substrate for the EGF-R kinase after its internalization induced by EGF. Cohen and Fava [60] demonstrated that annexin I is phosphorylated in vitro in a Ca’+-dependent manner by the EGF-R kinase of endosomes. In viuo, phosphorylation of annexin I reaches the maximum level after 30 min of EGF treatment, a time sufficient to internalize EGF-Rs [23]. Internalization of EGF-Rs is a necessary step to provoke the mitogenic effects of EGF [61]. Phosphorylation of the cytoplasmic fraction of annexin I could play a role in with the long-lasting effects induced by EGF in A431 cells [23]. The nature of cytoplasmic filamentous structures delineated by annexin I is unclear. These filaments were resistant to low temperature (0°C) and were not labeled by phalloidin (Figs. 2G and 2H), which suggests that neither cold labile microtubules [62, 631 nor microfilaments were the filamentous structures found. The characteristic pattern of these filaments forming a network between nucleus and cell periphery (Fig. 2G) suggests that they may represent intermediate filaments [64]. It has been demonstrated that protein p35 (annexin I) is selectively accumulated in keratin-rich Hassal bodies in human thymus [65]. Therefore, in A431 cells, the cytoplasmic filamentous structures with associated high amounts of annexin I can be considered as intermediate filaments. Further studies are necessary to confirm this suggestion. We thank Drs. A. B. Sorokin and A. D. Sorkin from the Institute of Cytology, Leningrad, for providing the mAb5A9 anti-EGF-R monoclonal antibody and EGF, Dr. V. E. Koteliansky from the Center of Cardiology, Moscow, for the anti-vinculin antibody, and Dr. C. Comera from the Pasteur Institute, Paris, for the anti-annexin I antibody. We are grateful to Dr. J. Sikora from Nencki Institute, Warsaw, for his comments on the manuscript and to Mrs. K. Mrozinska for excellent technical assistance. This work was supported by Grant NO. C.P.B.P. 04.01. from the Polish Academy of Sciences.
REFERENCES 1.
Bourguignon, L. Y. W., and Singer, S. J. (1977) Proc. Natl. Acad. Sci. USA 74,5031-5035. 2. Bourguignon, L. Y. W., and Bourguignon, G. J. (1984) Int. Rev. Cytol. 87,195-224. 3. Pasternak, C., Spudich, J. A., and Elson, E. L. (1989) Nature (London) 431,549-551. 4. Schlessinger, J., and Geiger, B. (1981) Exp. Cell Res. 134, 273-
279. 5. Bretscher, A. (1989) J. Cell Biol. 108,921-930. 6. van Bergen en Henegouwen, P. M. P., Defize, L. H. K., de Kroon, J., van Damme, H., Verkleij, (1989) J. Cell Biochem. 39,455-456.
A. J., and Boonstra,
J.
CAPPING
263
IN A431 CELLS
7.
Wiegant, F. A. C., Block, T. J., Defize, L. H. K., Linnemans, W. A. M., Verkley, A. J., and Boonstra, J. (1986) J. Cell Biol. 103,87-93.
35. Roth, J. (1986) J. Microsco~ 143,125-137. 36. Laemmli, U. K. (1970) Nature (London) 227,680~685. 37. Towbin, H., StaeheIin, T., and Gordon, J. (1979) Proc. Natl.
8.
Khrebtukova, I. A., Kwiatkowska, K., Gudkova, D. A., Sorokin, A. B., and Pinaev, G. P. (1991) Exp. Cell Res., 194,48-55. Branton, D., Cohen, C. M., and Tyler, J. M. (1981) CeU 24,2431.
38. Repasky, E. A., Granger, B. L., and Lazarides, E. (1982) Cell 29,
9. 10.
Byers, T. J., and Branton, 82,6153-6157.
11.
Elgsaeter, A., Stokke, B. T., Mikkelsen, (1986) Science 234,1217-1223.
12.
Nelson, W. J., Colaco, C. A. L. S., and Lazarides, Natl. Acad. Sci. USA 80, 1626-1630.
13.
Bourguignon, L. Y. W., Suchard, S. J., Nagpal, M. L., and Glenney, J. R., Jr. (1985) J. Cell. Biol. 101, 477-487. Condeelis, J. (1979) J. Cell Biol. 80, 751-758. Carboni, J. M., and Condeelis, J. S. (1985) J. Cell Biol. 100, 1884-1893. Bennett, H., and Condeelis, J. (1988) Cell Motil. Cytoskel. 11, 303-317. Heath, J. P. (1983) Nature (London) 302, 532-534. Levine, J., and Willard, M. (1983) Proc. Natl. Acad. Sci. USA 80, 191-195. Gerke, V., and Weber, K. (1984) EMBO J. 3, 227-233. Gerke, V., and Weber, K. (1985) J. Biol. Chem. 260,1688-1695. Glenney, J. (1986) J. Biol. Chem. 261, 7247-7252.
14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
25.
A., and Branton,
D.
K.-S., Wallner, B. P., Mattaliano, R. J., Tizard, R., C., Frey, A., Hession, C., Mcgray, P., Sinclair, L. K., E. P., Browning, J. L., Ramachandran, K. L., Tang, J., J. E., and Pepinsky, R. B. (1986) Cell 46, 191-199. B. (1979) Cell 18, 193-205.
Burridge, K., Fath, K., Kelly, T., Nuckolls, (1988) Annu. Rev. Cell Biol. 4, 487-525.
27.
Sefton, B. M., Hunter, T., Ball, E. H., and Singer, S. J. (1981) Cell 24,165-174. Werth, D. K., and Pastan, I. (1984) J. Biol. Chem. 259, 52645270. Feige, J.-J., and Chambaz, E. M. (1987) Biochemie 69,379-385.
29. 30. 31. 32. 33. 34.
G., and Turner,
C.
Howe, C. L., Sacramone, L. M., Mooseker, M. S., and Morrow, J. S. (1985) J. Cell Biol. 101, 1379-1385. Glenney, J. R., Jr., and Weber, K. (1985) Anal. B&hem. 150, 364-368. Kwiatkowska, K., and Sobota, A. (1990) Histochemistry 94,8793. Talian, J. C., Olmsted, J. B., and Goldman, R. D. (1983) J. Cell Biol. 97, 1277-1283. Litwin, J. A., and Beier, K. (1988) Histochemistry 98, 193-196.
Received February 7, 1991 Revised version received May 27, 1991
40. 41. 42.
E. (1983) Proc.
26.
28.
39.
D. (1985) Proc. N&l. Acad. Sci. USA
Glenney, J. R., Jr., Tack, B., and Powell, M. A. (1987) J. Cell Biol. 104,503-511. Fava, R. A., and Cohen, S. (1984) J. Biol. Chem. 259,2636-2645. Huang, Burne, Chow, Smart, Geiger,
Acad. Sci. USA 76,4350-4354. 821-833. Burn, P., Kupfer, A., and Singer, S. J. (1988) Proc. Natl. Acad. Sci. USA 85,497-501. Cooper, J. A., and Hunter, T. (1982) J. Cell Biol. 94,287-296. Vasiliev, J. M. (1985) B&him. Biophys. Acta 780, 2165. Zidovetzki, R., Yarden, Y., Schlessinger, J., and Jovin, (1981) Proc. Natl. Acad. Sci. USA 78,6981-6985.
43. Landreth, G. E., Williams, Biol. 101, 1341-1350. 44. 45. 46.
T. M.
K., and Rieser, G. D. (1985) J. Cell
Roy, M. L., Gittinger, C. K., and Landreth, G. E. (1989) J. Cell Physiol. 140, 295-304. Coleman, T. R., Fishkind, D. J., Mooseker, M. S., and Morrow, J. S. (1989) Cell Motil. Cytoskel. 12, 225-247. Anderson, R. A., and Lovrien, R. E. (1984) Nature (London)
307,655-658. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59.
Baudry, M., Bundman, M. C., Smith, E. K., and Lynch, G. S. (1981) Science 212,937-938. Lynch, G., and Baudry, M. (1984) Science 224,1057-1063. Nelson, W. J., and Veshnock, P. J. (1986) J. Cell Biol. 103, 1751-1766. Nelson, W. J., and Veshnock, P. J. (1987) J. Cell Biol. 104, 1527-1537. Willard, M., Baitinger, C., and Cheney, R. (1987) Brain Res. Bull. 18, 817-824. Burgoyne, R. D., and Cheek, T. R. (1987) Nature &xm’on) 326, 448. Aunis, D., and Bader, M. (1988) J. Exp. Biol. 139, 253-266. Perrin, D., and Aunis, D. (1985) Nature (London) 315,589-592. Haigler, H., Ash, J. F., Singer, S. J., and Cohen, S. (1978) Proc. Natl. Acad. Sci. USA 75, 3317-3321. Erikson, E., Shealy, D. J., and Erikson, R. L. (1981) J. Biol. Gem. 256,11,381-11,384. Hunter, T., and Cooper, J. A. (1981) Cell 24, 741-752. Sawyer, T., and Cohen, S. (1985) J. Biol. Chem. 260,8233-8236. Pepinsky, R. B., and Sinclair, R. K. (1986) Nature (London)
321,81-84. 60. 61. 62. 63. 64. 65.
Cohen, S., and Fava, R. A. (1985) J. Biol. Chem. 260, 12,35112,358. Schreibeer, A. B., Liberman, T. A., Lax, I., Yarden, Y., and Schlessinger, J. (1983) J. Biol. Chem. 258,846-853. Borisy, G. C., and Olmsted, J. B. (1972) Science 177,1196-1197. Weisenberg, R. C. (1972) Science 177,1104-1105. Goldman, R. D., Goldman, A. E., Green, K. J., Jones, J. C. R., Jones, S. M., and Yang, H.-Y. (1986) J. Cell Sci. Suppl. 5,69-97. Fava, R. A., and Piltch, A. S. (1987) J. H&o&m. Cytochem. 35, 1309-1315.
CELL RESEARCH
196,255-263
(1991)
Actin-Binding Proteins Involved in the Capping of Epidermal Growth Factor Receptors in A431 Cells KATARZYNA KWIATKOWSKA,**’ IRINA A. KHREBTUKOVA,? DINA A. GIJDKOVA,~ GEORGE P. PINAEV,~ AND ANDRZEJ SOBOTA* *Nencki Institute of Experimental Biology, Department of Cell Biology, 3 Pasteur Street, PL-02093 Warsaw, Poland; and tlnstitute of Cytology of the Academy of Sciences, 4 Tikhretsky Avenue, 194064 Leningrad, USSR
A capping process of epidermal growth factor receptors (EGF-Rs) was used for the study of the relation between the receptors and the actin-binding proteins (spectrin, vinculin, annexin I) that may be involved in EGF-R-cytoskeleton interaction. In intact, adherent A431 cells, EGF-Rs were diffusively distributed on the cell surface. Spectrin, vinculin, and annexin I were located beneath the plasma membrane. An abundance of EGF-Rs as well as submembrane proteins was observed in regions of membrane ruffles and cell-cell contacts. Annexin I was localized also in cytoplasm being attached to Alamentous structures surrounding the nucleus and extending to the cell periphery. Under polyvalent ligand treatment, EGF-Rs of adherent cells were aggregated on one side of the cell. Spectrin, vinculin, and annexin I dislocated together with EGF-Rs and were concentrated under plasma membrane at regions where cap formation took place. In suspended A431 cells only spectrin was located under the plasma membrane whereas annexin I and vinculin were diffusively distributed through the cells. During cap formation only spectrin was colocalized with EGF-Rs. The results confirmed the major role of spectrin as a receptor-mi0 1991 Academic press. he. CrOfilament linking protein.
INTRODUCTION
Several immunocytochemical and biochemical data indicate that if cells are exposed to a polyvalent ligand the cell-surface receptors become linked to the cytoskeleton and, as a result of actin-myosin interaction, are translocated and aggregated on one side of the cell [l31. Hence, assembling receptors into cap may serve as a model for studying mechanisms of receptor-cytoskeleton interactions. There have been suggestions that in human carcinoma epidermoid A431 cells the epidermal growth factor receptor (EGF-R)-cytoskeleton interac-
’ To whom reprint
requests should be addressed.
tion may be involved in the mechanism of epidermal growth factor (EGF) action [4-61. Several biochemical data suggest that the association of EGF with EGF-Rs stimulates interactions between EGF-Rs and cytoskeleton [6]. However, only a few immunoelectron microscopy observations, which suggest a linkage between EGF-Rs and microfilaments, are available [7]. Recently, it was found that the actin bundles delineating cell margins indeed take part in capping of the EGF-Rs of A431 carcinoma cells [S]. The submembrane actin-binding proteins are assumed to be involved in the linkage of microfilaments to EGF-Rs. Spectrin is one of the proteins that plays an important role in anchoring actin filaments to membrane proteins as it has been demonstrated in human erythrocytes [g-11]. The participation of spectrin, together with actin, in capping the receptors in lymphocytes [2, 12, 131, DictyosteZium discoideum [14-161, and fibroblasts [17, 181 indicates that spectrin may be considered a linker in receptor-microfilaments interactions. Another protein that may be involved in EGF-R-microfilament interactions is annexin I (calpactin 2). This protein, as well as the closely related annexin II (calpactin l), is able to interact with the membrane and with the F-actin in a Ca’+-dependent manner [19-221. In A431 cells treated with EGF, annexin I is phosphorylated by the EGF-R tyrosine kinase [23, 241 indicating the close structural relations between EGF-R and this protein. Vinculin may also be involved in binding the microfilament bundles to plasma membrane [25,26]. This protein, similar to annexin I, is a substrate for the pp66’-” tyrosine kinase [27,28], with an activity similar to that of EGF-R kinase [29]. In this study we examine the localization and redistribution of spectrin, vinculin, and annexin I during aggregation of EGF-receptors in A431 cells. The association of these proteins with EGF-R depends on the attachment of the cells to a substratum. It seems that spectrin is a protein that is actively engaged in the capping of EGF-Rs.
255 Copyright Q 1991 by Academic Press, Inc. All rights of reproduction in any form reserved
256
KWIATKOWSKA
MATERIALS
AND METHODS
Cell culture. Human epidermoid carcinoma A431 cells, obtained from the Cell Culture Collection of the Institute of Cytology of the Academy of Sciences, Leningrad, the USSR, were cultivated in Eagle’s medium supplemented with 10% bovine serum and glutamine. A431 cells were grown on glass at 37°C under 5% CO, atmosphere. Antibodies. o-Spectrin was isolated from a “0.6 M KJ extract” of chicken erythrocyte membranes [30]. The extract was applied to a Sephacryl S-200 column equilibrated and developed with solution of 0.6 M KJ, 0.5 m&f EDTA, 0.2 mM dithiothreitol (DTT), 10 n&f TRIS, pH 8.0. The void volume fractions were loaded onto a calmodulin-Sepharose column in the presence of 5 m&f CaCl, and 2 M KJ [31]. a-Spectrin selectively bound to the calmodulin column was eluated in the presence of 10 m&f EGTA and used for immunization of rabbits [32]. The o-spectrin antibody was affinity purified from the antisera according to procedure of Talian et al. [33]. Anti-annexin I antibody was a generous gift from Dr. C. Comera, Pasteur Institute, Paris, France. Anti-vinculin antibody was kindly provided by Dr. V. E. Koteliansky from the Center of Cardiology, Moscow, the USSR. Monoclonal anti-EGF-R IgG mAb5A9 was kindly supplied by Dr. A. B. Sorokin (Institute of Cytology Acad. Sci. USSR, Leningrad). Rhodamine-labeled and fluorescein-labeled antirabbit IgG (GAR-TRITC and GAR-FITC, respectively) were purchased from Sigma. Fluorescein-labeled swine anti-mouse IgG (SWAM-FITC) was obtained from the Institute of Sera and Vaccines, Prague, Czechoslovakia. Zmmunofluorescence microscopy. Adherent intact A431 cells were fixed with 3% formaldehyde prepared in phosphate-buffered saline (PBS) containing 1 m&f MgCl, for 30 min at room temperature, according to Schlessinger and Geiger [4]. Nonspecific antibody binding was blocked by 100 mA4 glycine in PBS (10 min, room temp). EGF-R was labeled with monoclonal mAb5A9 anti-EGF-R antibody (1:lOO diluted in PBS, 30 min, room temp) followed by incubation with SWAM-FITC (1:lO diluted in PBS, 30 min, room temp). Then the samples were permeabilized with solution of 0.1% Triton X-100 in PBS (10 min on ice) and treated with anti-o-spectrin (1:lO in PBS), anti-vinculin (1:20), anti-annexin I (1:lOO) antibodies, or nonimmune rabbit serum (1:lOO) for 1 h at 37°C. The samples were carefully washed and then incubated in a solution of GAR-TRITC (1:50 in PBS, 1 h, 37’C). Actin was observed by staining the cells with rhodamine-labeled phalloidin (Sigma, St. Louis, MO) at a concentration of 0.5 *g/ml. Using FITC-labeled second antibodies together with rhodamine-labeled phalloidin allowed the annexin I and actin to be localized simultaneously in the same cell. For studies of EGF-Rs redistribution two sets of experiments were performed: (1) the adherent A431 cells were treated with polyvalent ligands, i.e., they were exposed to solution of mouse anti-EGF-R antibody followed by exposure to a solution of anti-mouse SWAM-FITC antibody (20 min, 0°C); or (2) the intact cells were pretreated with EGF (200 rig/ml) in the growth medium at 0°C for 1 h. After washing, the cells were incubated with polyvalent ligands as described above. In both experiments, the living cells were subsequently transferred into the growth medium at 20°C and fixed at different stages of EGFRs redistribution. The cytoskeletal proteins were localized in Triton X-lOO-permeabilized cells as described above. For the observation of EGF-R redistribution in suspendedcells, the adherent cells were detached from substratum by treatment of culture with 0.02% EDTA in PBS solution. The detached cells were incubated in growth medium with serum at 37°C for 2 h. The cells were then transferred into an ice bath, treated with EGF, labeled with EGF-R antibody, and then labeled with SWAM-FITC by a procedure similar to that used for adherent cells. The redistribution of EGF-Rs in suspended A431 cells in the absence of EGF was not tested. The labeled cells were subsequently incubated (lo-50 min) at room temperature, fixed and permeabilized in methanol (-ZO’C, 5 min), and
ET AL. later stained for the actin-binding proteins using GAR-TRITC as the second antibody. The cells were intensively rinsed with PBS after each incubation and finally embedded in mixture of 90% glycerol and 5% propyl gallate. The specimens were examined using an OPTON microscope with filters for FITC and TRITC fluorescence. In a series of experiments the FITC-labeled samples were checked with green as well as with red filter sets. The same checking procedure was used for some TRITCstained cells. The overlapping of both fluorescences was barely seen. Zmmunoelectron microscopy. For localization of EGF-R the A431 cells were treated as in the fluorescence microscopy procedure used for the suspended cells. Briefly, the detached cells were incubated at 37°C for 2 h, then transferred to 0°C and exposed to EGF (200 rig/ml) followed by anti-EGF-R and SWAM-FITC antibodies. Finally, the samples were incubated with protein A-gold complex (15 nm gold particle size). Thereafter, the cells were shifted to room temperature and, at different stages of cap formation, fixed in a mixture of 6% formaldehyde/0.5% glutaraldehyde containing 50 m&f KC1 and 5 mM MgCl, in 100 mA4 phosphate buffer, pH 7.4, (2 h, room temp). After being washed, the cells were postfixed in 1% 0~0, solution in the same buffer (1 h, room temp). The samples were dehydrated in a series of ethanol and embedded in Epon 812. For postembedding labeling, the gold sections of A431 cells were mounted to formvar/carbon-coated nickel grids. In order to remove the surface layer of resin the samples were exposed for 30 s to a saturated alcohol solution of sodium hydroxide diluted 1:l with 96% ethanol [34]. After being washed with ethanol, distilled water, and PBS, the grids were put into 0.5% defatted milk in PBS to reduce nonspecific binding. The o-spectrin antigen was visualized by incubation of the samples with affinity-purified antibody directed against o-spectrin from chicken erythrocytes (1 pglml, 4”C, overnight), followed by incubation with gold-labeled protein A (8 nm size of gold particles). The control samples were treated with nonimmune rabbit serum diluted with PBS (100 gg of protein/ml). The unbound antibodies and protein A-gold were removed by extensive washing (5 X 15 min each). All solutions were prepared in PBS containing 1% bovine serum albumin and 0.1% Tween 20 [35]. The sections were finally rinsed with distilled water and counterstained with lead citrate and uranyl acetate. The samples were examined with an JEM 1OOB electron microscope. Analytical procedures. Gel electrophoresis was carried out in 10% polyacrylamide gel (SDS-PAGE) as described by Laemmli [36]. Proteins were transferred from the gel to nitrocellulose sheets for 2 h at 500 mA according to the procedure of Towbin et al. [37]. Immunoblotting was performed using appropriate antibodies followed by peroxidase-conjugated protein A binding. Enzymatic activity was visualized in the presence of H202 and diaminobenzidine (DAB).
RESULTS Characterization
of the Antibodies
a-Spectrin, vinculin, and annexin I were identified and localized in A431 carcinoma cells by means of polyclonal antibodies. The monospecific anti-a-spectrin antibody recognized a 240-kDa molecular weight polypeptide in whole A431 cell homogenates (Figs. lb and lc). After low temperature storage the homogenate contained 240- and 150-kDa polypeptides that crossreacted with the antibody (not shown). The 150-kDa polypeptide seems to be a typical product of proteolytic degradation of a-spectrin [ 381. Antiserum against human vinculin cross-reacted only with a 130-kDa polypeptide in whole cell homogenate
CAPPING
IN A431 CELLS
a b FIG. 1. Immunoblotting analysis of actin-binding proteins of A431 cells. Proteins of whole cells were separated by SDS-PAGE (b), transferred to nitrocellulose sheets, and treated with anti-cu-spectrin (c), anti-vinculin (d), anti-annexin I (e). (c) Anti-or-spectrin recognizes the 240-kDa polypeptide. (d) Anti-vinculin and (e) anti-annexin I react with polypeptides of 130 and 35 kDa, respectively. The molecular weight markers indicated on the left are myosin (200 kDa), ,&palactosidase (116 kDa), phosphorylase B (97 kDa), and ovalbumin (45 kDa).
(Fig. Id), which corresponded to the molecular weight of vinculin. Anti-annexin I serum recognized a low molecular weight polypeptide of 35 kDa in A431 cell homogenate (Fig. le). Thus, the 240,130, and 35 kDa polypeptides, revealed by immunoblotting analysis in A431 carcinoma cells, can be considered cr-spectrin, vinculin, and annexin I, respectively. Immunofluorescent Localization of EGF-R, Spectrin, Vinculin, and Annexin I in Adherent Cells
In intact adherent A431 cells fixed at room temperature and labeled for EGF-R the antigen was diffusively distributed on the surface of cells (Fig. 2A). However, the regions of ruffles and cell-cell contacts displayed stronger staining of receptor (Fig. 2A, arrows). Fixed, permeabilized adherent cells exposed subsequently to anti-spectrin, anti-vinculin, or anti-annexin I displayed a delicate fluorescence that was more intensive under plasma membrane in the periphery of the cells (Figs. 2B-2D). The labeling was especially well seen in regions of ruffles and cell-cell contacts (Figs. 2B-2D, arrows), similar to the case of EGF-R. Moreover, spectrin, vinculin, and annexin I formed a delicate network structure throughout the cell (Figs. 2B-2D, asterisks). In some cells, however, annexin I was also found in the cytoplasm attached to nonidentified filamentous structures that were concentrated around nucleus and extended to the cell periphery (Fig. 2G). When the cells were double labeled with phalloidin and anti-
257
annexin, no correlation between localization of actin filaments and annexin-associated filamentous structures was observed (Figs. 2G and 2H). In control cells treated with nonimmune rabbit serum instead of the first antibody a very faint labeling was visible (Figs. 2E and 2F). The quite different patterns of fluorescence observed in the cell double labeled for annexin I and actin (compare Figs. 2G and 2H) were the positive control for the proper sets of filters applied in our examinations. The bright green FITC fluorescence attributed to annexin was not observed with the red (TRITC) filter set. On the contrary, the rhodamine fluorescence of phalloidin was seen only at the filter set proper for TRITC. These observations allowed us to exclude the possible overlapping of FITC and TRITC emissions. Redistribution of Actin-Binding Proteins during Capping of EGF-Rs in Adherent Cells
It is suggested that a set of proteins involved in the process of binding of microfilaments to ligand-receptor complex is metabolically regulated [39]. Therefore it was interesting to examine whether the presence of EGF together with polyvalent ligands would define the selection of proteins involved in linkage of EGF-Rs to microfilaments during capping. For this purpose, the adherent A431 cells were treated in two different ways: (1) the cells were exposed to EGF and then to polyvalent ligands (anti-EGF-Rs and SWAM-FITC antibodies), or (2) the incubation with EGF was omitted. The whole procedure of EGF-R staining was carried out at 0°C to prevent internalization of receptors. When the EGFpolyvalent ligand-exposed cells were transferred from 0 to 2O”C, after 30 min EGF-Rs aggregated into caps which were formed along cell margins free of cell-cell contacts (Figs. 3A, 3C, and 3E, arrows). During aggregation of EGF-Rs all three actin-binding proteins comigrated to where the accumulation of EGF-Rs took place. The colocalization of EGF-R caps and aggregates of spectrin (Figs. 3A and 3B), vinculin (Figs. 3C and 3D), and annexin I (Figs. 3E and 3F) was observed in adherent cells. When the cells were exposed only to polyvalent ligands, no differences in the protein composition of subcaps, versus the EGF-polyvalent ligand-treated cells, were observed. Spectrin (Fig. 3H) as well as vinculin and annexin I accumulated at the locations of EGF-R aggregation. However, in these cells the aggregation of EGFRs proceeded slowly, and during 30 min at 2O”C, 2-3 patches of EGF-Rs were formed (Fig. 3G, arrows). Colocalization of EGF-R and submembrane actinbinding proteins was observed during the whole process of cap formation (compare Figs. 3A and 3B with 3G and 3H). It should be noted, however, that only a part of cellular spectrin, vinculin, and annexin I was shifted to
258
KWIATKOWSKA
ET AL.
FIG. 2. Localization of EGF-R and actin-binding proteins in intact, adherent A431 cells. (A) Cells stained for EGF-R. The antigen is diffisively distributed on the cell surface. (B, C, D) Cells stained for spectrin, vinculin, and annexin I, respectively. The abundance of submembrane proteins as well as EGF-R in regions of ruffles and cell-cell contacts is shown (arrows). Asterisks denote the areas of the ce 11s where the tiny network formed by spectrin (B), vinculin (C), and annexin I (D) can be seen. (G, H) A431 cells double-labeled for annexin I alnd F-actin, respectively. (G) The anti-annexin I antibody stains the filaments concentrated in the perinuclear region from which they extel nd toward the cell periphery. There is no colocalization between these filaments and actin bundles visualized by rhodamine-conjugatedphalloid lin in (H). (E) Control sample. The cells were treated with nonimmune rabbit serum and then exposed to GAR-TRITC antibody. (F) Pha ise contrast image of(E). A-D, X520; E, F, X400, G, H, X700.
CAPPING
IN A431 CELLS
259
1PIG. 3. (A-F) Coincident redistribution of EGF-Rs and actin-binding proteins induced by anti-EGF-R and SWAM-FITC antibodies in shows the cocaps formed after 30 min at room temperature by EGF-Rs (A) and spectrin (1B), the presence of EGF. Double immunotluorescence EG F-Rs (C) and vinculin (D), EGF-Rs (E) and annexin I (F). Locations of cap formation are indicated by arrows. The bright labeling of contacts is still observed (arrowheads). (G, H) Redistribution of EGF-Rs and spectrin induced by polyvale mt SUb Imembrane proteins in cell-cell liga mds in the absence of EGF. The cells, fixed after 30 min of cap formation, display a few patches of EGF-Rs (G, arrows). (H) Correspondi ng red istribution of spectrin accumulated under plasma membrane as copatches (arrows). A-F, X740; G, H, X520.
260
KWIATKOWSKA
ET AL,
the region of the formed cap. Even in the cells containing the completely formed caps of EGF-Rs we still observed the labeling of the actin-binding proteins in the submembrane layer of the cells, especially in regions of cell-cell contacts (Figs. 3B, 3D, and 3H, arrowheads). Participation of Spectrin in Capping of EGF-Rs in Suspended A431 Cells
Examination of distribution of actin-binding proteins in detached A431 cells revealed that only spectrin did not change its cortical localization which was similar to adherent cells. The antigen was seen as a continuous submembranous layer that coincided with uniformly spread surface-labeled EGF-R (Figs. 4A and 4B) as well as with cortical F-actin 181. Annexin I and vinculin, contrary to spectrin, were diffusively distributed through the cell and seemed to no longer underlie the plasma membrane (Figs. 4F and 4H). In addition, in a part of cells labeled with anti-annexin I the nuclear envelope was also stained (Fig. 4H). The uniform labeling of the proteins in cytoplasm was much stronger than it was in the case of nonspecific binding of antibodies (compare Figs. 4F and 4H with 45). During aggregation of EGF-Rs only spectrin was accumulated at the region of cap formation (Figs. 4C and 4D). At the same time, no changes in the diffusive distribution of vinculin and annexin I were observed (Figs. 4F and 4H). Once the capping was finished the number of EGF-R aggregates associated with spectrin decreased. The results presented above were confirmed by immunoelectron microscopy. When the suspended A431 cells were fixed during the first steps of EGF-Rs capping and labeled with anti-spectrin, the aggregated gold particles marking the places rich in cu-spectrin, appeared in the submembrane layer of cells in the close vicinity of EGFRs (Fig. 5A). A prolonged (45 min) exposure of EGF-Rs to polyvalent ligands allowed the first moments of internalization of the receptors to be caught. At this step the formed cap contained numerous small vesicles with EGF-R inside them. At this moment, a spatial separation of EGF-R and spectrin aggregates was visible (Fig. 5B). The specificity of spectrin labeling was tested by substituting anti-spectrin antibody for nonimmune rabbit serum. Only a few, randomly dispersed, gold particles could be found in the control samples (Fig. 5C). DISCUSSION
In this study we examined the distribution of spectrin, vinculin, and annexin I in human carcinoma A431 cells. These proteins, being actin- and membrane-binding ones, may serve as a linking system between actin filaments and cell-surface EGF-Rs. The process of ag-
FIG. 4. Suspended A431 cells stained for both surface EGF-R (left) and actin-binding proteins (right). (B) In intact cells spectrin is localized uniformly under plasma membrane reflecting the distribution of EGF-R on the cell surface (A). During cap formation spectrin (D) accompanies the receptors (C). Vinculin (F) and annexin I (H) are dispersed diffusively through the cell and do not redistribute in capped cells. The noncoincident distribution of EGF-Rs stained with FITC, and of annexin I and vinculin visualized by TRITC-fluorescence, points to the correct sets of filters used. (J) Control cells incubated with nonimmune rabbit serum instead of the first anti-intracellular protein antibodies. x500.
gregation of EGF-Rs induced by polyvalent ligands was exploited as a model system for studying the mechanisms of cytoskeleton-receptor interaction. As it was
CAPPING
IN A431 CELLS
261
shown in adherent A431 cells, spectrin, vinculin, and annexin I were located beneath the plasma membrane and were accumulated in the regions of EGF-R cap formation. In suspended A431 cells only spectrin colocalized with cap of EGF-Rs. Vinculin and annexin I were no more clearly observed beneath the plasma membrane of suspended cells and were dispersed through the cytoplasm in resting as well as in capped cells. Changes in cellular distribution of vinculin and annexin I may reflect a phenomenon of general reconstruction of the cytoskeleton in A431 cells as 8 result of detachment of cells from the substratum. A similar phenomenon has been observed in suspended fibroblasts [40]. It is known that binding of annexins I, II (calpactin 2,l) to F-actin, spectrin, and membrane phospholipids is a Ca2+-dependent process [ 19-221. Hence it is possible that changes in ionic equilibrium in suspended A431 cells induced by EDTA, used for detachment of cells from the substratum, might evoke dissociation of the proteins from the plasma membrane. Dissociation of vinculin from the plasma membrane in suspended cells is considered as the loss of cell-substratum and cell-cell contacts. A similar process occurs in the cells transformed by oncogenic viruses. This transformation leads to a significant reduction of cell adhesion correlating with increased phosphorylation of vinculin and redistribution within the cell [26,40]. Vinculin in adherent A431 cells was located uniformly beneath the plasma membrane (Fig. 2C) although the used antibody stained selectively focal contacts in human fibroblasts (unpublished). However, in l-day-old cells no focal contacts were observed by means of interference reflection microscopy. The adhesion of these cells to the substratum probably is due to small and incomplete dot contacts in which actin filaments, invisible with fluorescence microscopy, are anchored within the plasma membrane via vinculin [41]. Later, these sites are transformed into completely formed focal contacts that were
FIG. 5. Ultrastructural analysis of the relation between the EGF-R and spectrin localization at different stages of the capping process.EGF-R of suspended cells was labeled by anti-EGF-R and SWAM-FITC antibodies, and additionally, with protein A-gold particles (15 nm in diameter) and then induced to aggregate for either 30 min (A, C) or 45 min (B). Spectrin was localized in the cells by means of the postembedding technique using anti-a-spectrin antibody followed by protein A-gold particles (8 nm). (A) EGF-R is localized on the plasma membrane surface (arrowheads) close to submembrane spectrin aggregate (arrow). (B) Internalization of EGF-Rs is beginning. EGF-Rs are found in endosome-like structures (lower part of micrograph, arrowheads) whereas spectrin aggregates remain under plasma membrane (arrows). (C) Control sample. The cells were stained for EGF-R and after embedding in resin were subsequently exposed to nonimmune serum and protein A-gold conjugates (6 nm). Arrowheads mark the location of EGF-Rs (gold particles of 15 nm). Nonspecific postembedding labeling of the sections is barely detectable. A, X60,000, B, C, ~48,000.
262
KWIATKOWSKA
well developed in 5-day-old A431 cell culture (unpublished). Reorganization of the cytoskeleton in suspended cells correlates with changes in dynamics of EGF-R. In these cells the process of internalization of EGF-Rs is twice as slow as in adherent ones [42]. Also, only 30-50% of the receptors is associated with the Triton-insoluble fraction of suspended A431 cells [43, 441. Detachment of A431 cells from the substratum probably modifies EGFR-cytoskeleton interactions. These changes may explain the observed differences in composition of actinbinding proteins involved in the association of actin filaments with EGF-R-polyvalent ligand complex in adherent and suspended cells. On the other hand, these results seem to indicate that among actin-binding proteins only spectrin is actively engaged in binding of the peripheral actin bundles with EGF-Rs. The concentration of vinculin and annexin I under plasma membrane of adherent cells in regions of EGF-Rs aggregation may be a secondary phenomenon. These proteins can be transported to the EGF-R cap region as a result of interaction with actin filaments. Linking of cytoskeletal filaments with cell-surface receptors is probably an essential function of spectrin family proteins [45]. It has been suggested that spectrin is involved in receptor-microfilament interactions during such processes as: (i) control of mammalian erythrocyte shape changes [46], (ii) regulation and exposition of receptors in neurons [47,48], (iii) control of epithelial cell polarity [49,50], and (iv) axonal transport [51]. As the process of cap formation was finishing, an increased amount of EGF-R caps not associated with spectrin aggregates was detected. The electron microscopic observations showed the spatial separation of EGF-R and spectrin started at the moment of internalization of polyvalent-EGF-R complex (Fig. 5B). It seems that dissociation of spectrin from the complex is the first step of the internalization of the cell-surfaceaggregated receptors. A similar mechanism is postulated for spectrin in exocytosis [52, 531. Dissociation of spectrin from the membrane of secretory vesicles and disintegration of the spectrin-actin network seems to be necessary for fusion of vesicles with plasma membrane [53,54]. The role of spectrin as well as other actin-binding proteins in the process of aggregation and internalization of EGF-Rs induced by EGF in A431 cells under physiological conditions is still an open question [42, 551. Proteins p35 (annexin I) and p36 (annexin II) become intensively phosphorylated in the cells exposed to growth factors [29,56-581. Some biochemical and cytochemical data indicate that submembrane-localized annexin I is the specific substrate of the EGF-R kinase [23, 24,58,59]. However, the observations of adherent cells showed that only a part of annexin I is connected to the
ET AL.
plasma membrane. A large fraction of this protein was dispersed through the cell, where it was associated with filamentous structures. The cytoplasmic pool of annexin I could be the substrate for the EGF-R kinase after its internalization induced by EGF. Cohen and Fava [60] demonstrated that annexin I is phosphorylated in vitro in a Ca’+-dependent manner by the EGF-R kinase of endosomes. In viuo, phosphorylation of annexin I reaches the maximum level after 30 min of EGF treatment, a time sufficient to internalize EGF-Rs [23]. Internalization of EGF-Rs is a necessary step to provoke the mitogenic effects of EGF [61]. Phosphorylation of the cytoplasmic fraction of annexin I could play a role in with the long-lasting effects induced by EGF in A431 cells [23]. The nature of cytoplasmic filamentous structures delineated by annexin I is unclear. These filaments were resistant to low temperature (0°C) and were not labeled by phalloidin (Figs. 2G and 2H), which suggests that neither cold labile microtubules [62, 631 nor microfilaments were the filamentous structures found. The characteristic pattern of these filaments forming a network between nucleus and cell periphery (Fig. 2G) suggests that they may represent intermediate filaments [64]. It has been demonstrated that protein p35 (annexin I) is selectively accumulated in keratin-rich Hassal bodies in human thymus [65]. Therefore, in A431 cells, the cytoplasmic filamentous structures with associated high amounts of annexin I can be considered as intermediate filaments. Further studies are necessary to confirm this suggestion. We thank Drs. A. B. Sorokin and A. D. Sorkin from the Institute of Cytology, Leningrad, for providing the mAb5A9 anti-EGF-R monoclonal antibody and EGF, Dr. V. E. Koteliansky from the Center of Cardiology, Moscow, for the anti-vinculin antibody, and Dr. C. Comera from the Pasteur Institute, Paris, for the anti-annexin I antibody. We are grateful to Dr. J. Sikora from Nencki Institute, Warsaw, for his comments on the manuscript and to Mrs. K. Mrozinska for excellent technical assistance. This work was supported by Grant NO. C.P.B.P. 04.01. from the Polish Academy of Sciences.
REFERENCES 1.
Bourguignon, L. Y. W., and Singer, S. J. (1977) Proc. Natl. Acad. Sci. USA 74,5031-5035. 2. Bourguignon, L. Y. W., and Bourguignon, G. J. (1984) Int. Rev. Cytol. 87,195-224. 3. Pasternak, C., Spudich, J. A., and Elson, E. L. (1989) Nature (London) 431,549-551. 4. Schlessinger, J., and Geiger, B. (1981) Exp. Cell Res. 134, 273-
279. 5. Bretscher, A. (1989) J. Cell Biol. 108,921-930. 6. van Bergen en Henegouwen, P. M. P., Defize, L. H. K., de Kroon, J., van Damme, H., Verkleij, (1989) J. Cell Biochem. 39,455-456.
A. J., and Boonstra,
J.
CAPPING
263
IN A431 CELLS
7.
Wiegant, F. A. C., Block, T. J., Defize, L. H. K., Linnemans, W. A. M., Verkley, A. J., and Boonstra, J. (1986) J. Cell Biol. 103,87-93.
35. Roth, J. (1986) J. Microsco~ 143,125-137. 36. Laemmli, U. K. (1970) Nature (London) 227,680~685. 37. Towbin, H., StaeheIin, T., and Gordon, J. (1979) Proc. Natl.
8.
Khrebtukova, I. A., Kwiatkowska, K., Gudkova, D. A., Sorokin, A. B., and Pinaev, G. P. (1991) Exp. Cell Res., 194,48-55. Branton, D., Cohen, C. M., and Tyler, J. M. (1981) CeU 24,2431.
38. Repasky, E. A., Granger, B. L., and Lazarides, E. (1982) Cell 29,
9. 10.
Byers, T. J., and Branton, 82,6153-6157.
11.
Elgsaeter, A., Stokke, B. T., Mikkelsen, (1986) Science 234,1217-1223.
12.
Nelson, W. J., Colaco, C. A. L. S., and Lazarides, Natl. Acad. Sci. USA 80, 1626-1630.
13.
Bourguignon, L. Y. W., Suchard, S. J., Nagpal, M. L., and Glenney, J. R., Jr. (1985) J. Cell. Biol. 101, 477-487. Condeelis, J. (1979) J. Cell Biol. 80, 751-758. Carboni, J. M., and Condeelis, J. S. (1985) J. Cell Biol. 100, 1884-1893. Bennett, H., and Condeelis, J. (1988) Cell Motil. Cytoskel. 11, 303-317. Heath, J. P. (1983) Nature (London) 302, 532-534. Levine, J., and Willard, M. (1983) Proc. Natl. Acad. Sci. USA 80, 191-195. Gerke, V., and Weber, K. (1984) EMBO J. 3, 227-233. Gerke, V., and Weber, K. (1985) J. Biol. Chem. 260,1688-1695. Glenney, J. (1986) J. Biol. Chem. 261, 7247-7252.
14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.
25.
A., and Branton,
D.
K.-S., Wallner, B. P., Mattaliano, R. J., Tizard, R., C., Frey, A., Hession, C., Mcgray, P., Sinclair, L. K., E. P., Browning, J. L., Ramachandran, K. L., Tang, J., J. E., and Pepinsky, R. B. (1986) Cell 46, 191-199. B. (1979) Cell 18, 193-205.
Burridge, K., Fath, K., Kelly, T., Nuckolls, (1988) Annu. Rev. Cell Biol. 4, 487-525.
27.
Sefton, B. M., Hunter, T., Ball, E. H., and Singer, S. J. (1981) Cell 24,165-174. Werth, D. K., and Pastan, I. (1984) J. Biol. Chem. 259, 52645270. Feige, J.-J., and Chambaz, E. M. (1987) Biochemie 69,379-385.
29. 30. 31. 32. 33. 34.
G., and Turner,
C.
Howe, C. L., Sacramone, L. M., Mooseker, M. S., and Morrow, J. S. (1985) J. Cell Biol. 101, 1379-1385. Glenney, J. R., Jr., and Weber, K. (1985) Anal. B&hem. 150, 364-368. Kwiatkowska, K., and Sobota, A. (1990) Histochemistry 94,8793. Talian, J. C., Olmsted, J. B., and Goldman, R. D. (1983) J. Cell Biol. 97, 1277-1283. Litwin, J. A., and Beier, K. (1988) Histochemistry 98, 193-196.
Received February 7, 1991 Revised version received May 27, 1991
40. 41. 42.
E. (1983) Proc.
26.
28.
39.
D. (1985) Proc. N&l. Acad. Sci. USA
Glenney, J. R., Jr., Tack, B., and Powell, M. A. (1987) J. Cell Biol. 104,503-511. Fava, R. A., and Cohen, S. (1984) J. Biol. Chem. 259,2636-2645. Huang, Burne, Chow, Smart, Geiger,
Acad. Sci. USA 76,4350-4354. 821-833. Burn, P., Kupfer, A., and Singer, S. J. (1988) Proc. Natl. Acad. Sci. USA 85,497-501. Cooper, J. A., and Hunter, T. (1982) J. Cell Biol. 94,287-296. Vasiliev, J. M. (1985) B&him. Biophys. Acta 780, 2165. Zidovetzki, R., Yarden, Y., Schlessinger, J., and Jovin, (1981) Proc. Natl. Acad. Sci. USA 78,6981-6985.
43. Landreth, G. E., Williams, Biol. 101, 1341-1350. 44. 45. 46.
T. M.
K., and Rieser, G. D. (1985) J. Cell
Roy, M. L., Gittinger, C. K., and Landreth, G. E. (1989) J. Cell Physiol. 140, 295-304. Coleman, T. R., Fishkind, D. J., Mooseker, M. S., and Morrow, J. S. (1989) Cell Motil. Cytoskel. 12, 225-247. Anderson, R. A., and Lovrien, R. E. (1984) Nature (London)
307,655-658. 47. 48. 49. 50. 51. 52. 53. 54. 55. 56. 57. 58. 59.
Baudry, M., Bundman, M. C., Smith, E. K., and Lynch, G. S. (1981) Science 212,937-938. Lynch, G., and Baudry, M. (1984) Science 224,1057-1063. Nelson, W. J., and Veshnock, P. J. (1986) J. Cell Biol. 103, 1751-1766. Nelson, W. J., and Veshnock, P. J. (1987) J. Cell Biol. 104, 1527-1537. Willard, M., Baitinger, C., and Cheney, R. (1987) Brain Res. Bull. 18, 817-824. Burgoyne, R. D., and Cheek, T. R. (1987) Nature &xm’on) 326, 448. Aunis, D., and Bader, M. (1988) J. Exp. Biol. 139, 253-266. Perrin, D., and Aunis, D. (1985) Nature (London) 315,589-592. Haigler, H., Ash, J. F., Singer, S. J., and Cohen, S. (1978) Proc. Natl. Acad. Sci. USA 75, 3317-3321. Erikson, E., Shealy, D. J., and Erikson, R. L. (1981) J. Biol. Gem. 256,11,381-11,384. Hunter, T., and Cooper, J. A. (1981) Cell 24, 741-752. Sawyer, T., and Cohen, S. (1985) J. Biol. Chem. 260,8233-8236. Pepinsky, R. B., and Sinclair, R. K. (1986) Nature (London)
321,81-84. 60. 61. 62. 63. 64. 65.
Cohen, S., and Fava, R. A. (1985) J. Biol. Chem. 260, 12,35112,358. Schreibeer, A. B., Liberman, T. A., Lax, I., Yarden, Y., and Schlessinger, J. (1983) J. Biol. Chem. 258,846-853. Borisy, G. C., and Olmsted, J. B. (1972) Science 177,1196-1197. Weisenberg, R. C. (1972) Science 177,1104-1105. Goldman, R. D., Goldman, A. E., Green, K. J., Jones, J. C. R., Jones, S. M., and Yang, H.-Y. (1986) J. Cell Sci. Suppl. 5,69-97. Fava, R. A., and Piltch, A. S. (1987) J. H&o&m. Cytochem. 35, 1309-1315.
