186,138-148

EXPERIMENTALCELLRESEARCH

(1990)

Disintegration of Adhesion Plaques in Chicken Embryo Fibroblasts upon Rous Sarcoma Virus-Induced Transformation: Different Dissociation Rates for Talin and Vinculin RUUD BRANDS, Division

of Cell Biology,

ANNEMIEK

The Netherlands

Cancer

DE BOER, CONSTANCE Institute

(Antoni

uan Leeuwenhoek

The localization of talin and vinculin in chicken embryo fibroblasts (CEF) during transformation was studied by immunoelectron microscopy. CEF cells were infected with a temperature-sensitive mutant of Rous sarcoma virus. After 16 h at 42”C, transformation was induced by incubation at 37°C for different intervals up to 3 h. Cells were cleaved by “wet cleaving” as reported previously by us (R. Brands and C. A. Feltkamp, 1988, Exp. Cell Res. 176,309) and labeled with affinity-purified polyclonal antibodies to talin or vinculin, or monoclonal anti-vinculin. We observed a rapid reduction of vinculin in adhesion plaques within 15 min and a much slower dissociation of talin. This was found using single-labeling procedures and also within the same cell using double labeling. Seemingly intact microfilament bundles were observed associated with adhesion plaques that contained relatively little vinculin. These observations show that an early event in src-induced transformation is the release of vinculin from adhesion plaques. Furthermore, since adhesion plaques with attached filament bundles can exist at least transiently with very little or no vinculin present, it seems likely that vinculin is not, or not the only protein, linking actin filaments to adhesion plaques. o isso Academic press, k.

INTRODUCTION

Areas of closest contact between untransformed fibroblasts and the substrate contain structures termed “focal contacts” or “adhesion plaques.” Major constituents of these plaques are transmembrane integrins [Z], mediating adhesion to extracellular matrix molecules like fibronectin and vitronectin, and the cytoplasmic components talin and vinculin (for recent reviews, see [3,4]). The latter proteins have been suggested to form part of a chain linking integrins to actin microfilaments which terminate at these plaques [3-51. In vitro studies have shown that integrins can bind to talin [6], which in 1 To whom

reprint

requests

should

0014.4827/90 $3.00 Copyright 0 1990 by Academic Press, All rights of reproduction in any form

138

Huis),

AND ED Roosl

1066 CXAmsterdam,

The Netherlands

turn can bind to vinculin [7], which is able to bind directly or indirectly to actin [3, 4, 8-101. Adhesion plaques and terminating actin microfilament bundles are greatly reduced in size and number in fibroblasts transformed by various oncogenes, in particular those encoding tyrosine kinases [ 111. Integrins, talin, and vinculin are among the substrates for these kinases, but it is not clear which of these substrates, if any is the relevant one. Phosphorylation at tyrosine residues of adhesion plaque components might alter their mutual interaction and thereby ultimately lead to dissociation of the focal contact [12, 131. Recently, it was shown that the tyrosine phosphorylated in the chick embryo fibroblast integrin complex is contained within the peptide domain binding to talin and that phosphorylation inhibited this interaction in vitro [14]. Thus, phosphorylation of integrin is likely to be a major cause of adhesion plaque disintegration. Whether phosphorylation of talin and vinculin is important at all is less evident. In particular the role of vinculin phosphorylation has been questioned [15-191. The localization of the various components in adhesion plaques was observed mainly by immunofluorescence. Electron microscopical studies are scarce, probably due to technical difficulties. In the present study we have used our recently described “wet-cleaving” method which allows for easy access of macromolecules (such as antibodies) to the cytoplasmic surface of the ventral plasma membrane and which has turned out to be quite suitable for ultrastructural studies [l]. We show that after induction of transformation by a temperature-sensitive mutant of Rous sarcoma virus, vinculin dissociates from adhesion plaques prior to talin, and that seemingly intact adhesion plaques with terminating actin filaments can temporarily exist with very few vinculin molecules present. MATERIALS Cell culture. in Dulbecco’s

be addressed.

Inc. reserved.

A. FELTKAMP,

AND

METHODS

Chicken embryo fibroblasts (CEF) were maintained modified Eagle’s medium (DMEM) containing 10% (v/

ALTERATIONS v) fetal otics.

calf serum

(FCS),

10 m&f Hepes,

IN

ADHESION

2 m&f glutamine,

PLAQUES and antibi-

Materials. Nitrocellulose (0.45 Frn membrane filter) was from Schleicher & Schiiell Scientific (Dassel, West Germany). Biotinylated sheep anti-mouse IgG (S X M-biot), donkey anti-rabbit IgG antibodies (D X R-biot), streptavidin-peroxidase (Str-PO) and streptavidingold conjugates (&r-A@) were from Amersham (Amersham Int., Amersham, UK). Gold-conjugated antibodies were all from Janssen Pharmaceutics (Beerse, Belgium). Goat anti-mouse-fluorescein (G X M-FITC) and goat anti-rabbit-fluorescein (G X R-FITC) were from Nordic Pharmaceutical (Tilburg, The Netherlands). Rhodamine-conjugated phalloidin (Phall-TRITC) was from Molecular Probes, Inc. (Junction City, Oregon). Prestained molecular weight markers were obtained from Bethesda Research Labs (Gaithersburg, Maryland). All chemicals were reagent grade from common commercial suppliers. tsNY-68 Rous Sarcoma Virus. The temperature-sensitive mutant of the Smidt-Ruppin strain of Rous sarcoma virus subgroup-A (tsRSV), originally described by Kawai and Hanafusa [20] as tsNY-68 SRA, was kindly provided by Dr. Boschek (GieBen University, West Germany). Wet cleauing. Wet cleaving was carried out as described previously ] 11. In short, subconfluent monolayers of CEF grown on plastic petri dishes (Falcon, U.S.A.) or glass coverslips were rinsed at 37°C with TBS (Tris-buffered saline, (25 mM Tris, 135 m&f NaCl, 5 mM KCl, 0.5 mM Na2H2P04, 1.0 mM CaCl,, 0.5 mM MgClz, pH 7.2, 37°C). Next, buffer was aspirated and cells were overlayed with nitrocellulose sheets for 1 min at 20°C. Tne cells were cleaved by lifting the nitrocellulose. Using standard conditions described previously, the plasma membrane and the associated cytoskeleton with embedded organelles remained predominantly attached to the substrate, whereas the greater part of the cell was removed with the nitrocellulose. If drier conditions were applied, i.e., leaving less pericellular fluid, cells were a!most completely ripped off their substrate, leaving only small areas of plasma membrane on the substrate. Vinculin and talin purification. Vinculin was purified from chicken gizzards according to O’Halloran et al. [21] and Feramisco and Burridge [22]. The vinculin-containing fractions after the final isolation step on hydroxylapatite were pooled and used for immunization. Taiin was purified according to O’Halloran et al. [21]. Gel electrophoresis (SDS-PAGE) and Western blotting were carried out according to Laemmli [23] and Towbin et al. [24], respectively. Polyclonal and monoclonal antibodies to vinculin and talin. Rabbits (New Zealand Whites) were immunized with a total of 100 pg vinculin or talin (in Freund’s complete adjuvant) administered intramuscularly and into the popliteal lymph nodes, and given boosters of 100 pg in Freund’s incomplete adjuvant at 3-week intervals. Boosters were administered intramuscularly. Reactivity of the sera was tested on immunoblots. Antibodies were affinity-purified on nitrocellulose strips carrying Western-blotted vinculin and talin, as described previously [l]. The specificity of the purified anti-vinculin and anti-talin polyclonal antibodies (pAb-Vc and pAb-Ta) thus obtained was checked on immunoblots of CEF homogenate, as was monoclonal anti-vinculin (mAb-Vc, Bio-Yeda, Rehovot, Israel). Infection with tsRSV and induction of transformation. CEF were infected with tsRSV at 37°C for 1 h (as described by Becker et al. [25]) and then cultured for 16 h at the restrictive temperature of 42°C. Induction of transformation was achieved by transferring the cells to the permissive temperature of37”C [ll]. Control CEF were either noninfected cells or tsRSV-infected cells maintained at 42°C. The latter were rinsed in 42°C wash buffer and aldehyde-fixed prior to wetcleaving. Immunoflwrescence. CEF that had been cultured on slips were wet cleaved. Tbe cell remnants were rinsed in (40 m&f Na-Pi I pH 7.0,lOO m&f KCl, 1 mM CaCl,, 1 n&f fixed in 2% paraformaldehyde + 0.01% glutaraldehyde in

glass coverwash buffer MgClz) and this buffer.

IN

TRANSFORMING

FIBROBLASTS

139

Prior to incubation, excess aldehyde was quenched in wash buffer containing 0.02 M glycin. All immunoincubations were carried out at 20°C. First, affinity-purified anti-vinculin diluted to 10 &g/ml in buffer (containing 0.02 M glycin) was applied. Fluorescein-conjugated sheep anti-rabbit Ig (25 wg/ml in glycin-containing buffer) was used as secondary antibody. In addition, actin filaments were visualized by decoration with rhodamine-conjugated phalloidin (pha&TRITC). PhalIoidin binds irreversibly to filamentous actin. Labehng procedures were similar for pAb talin and mAb vinculin, except in the latter ease fluorescein-conjugated goat anti-mouse Ig was applied as secondary antibody. Immunoelectron microscopy, critical-point drying, and carbon shadowing. CEF (grown on glass coverslips) were wet cleaved and subsequently aldehyde-fixed (2% paraformaldehyde + 0.01% glutaraldehyde) and immunolabeled with anti-vinculin (pAb or mAb) or antitalin antibodies (pAb) at 10 pg/ml (single-labeling procedure). Either of the following double-labeling procedures was applied using mAbvinculin/pAb-talin or pAb-vinculin/pAb-talin protocols: (I) mAb-vinculin was reacted with G X M-Au” conjugates, whereas talin was labeled indirectly with D X R-biotin and streptavidin-A# conjugates (alternatively G X M-A@ and G X R-Au’ were used for mAb vinculin and pAb talin, respectively). (II) tected

pAb vinculin was detected indirectly with D X R-biotin

with G X -Au’ and pAb talin deand streptavidin-Au’” conjugates.

After immunoincubation, specimens were fixed with 2.5% glutaraldehyde. 0~0, fixation and uranyl staining, ethanol d.ehydration, and critical-point drying (CPD) were carried out essentially as described by Roos et al. [26]. In a Balzers BAF301 apparatus (2 set at 60 mA, 1.6 Kv) a thin layer of carbon was evaporated over the CPD specimens, which were subseqrently detached from the glass coverslips by floating them on 40% HF (Merck), rinsed with distilled water, and transferred to Formvar (1%). and carbon-coatedEM-grids (300 mesh). The specimens were observed with standard transmission EM, using a Philips EM 301 electron microscope at 60-80 kV. An important advantage of the wet-cleaving/CPD/HF procedure compared to cryo- or plastic sectioning was that large-areas, containing many cells, could be observed. Digitation of electron-microscopic data. icrograph negatives were projected on a transparent screen with a final magnification of 42,000X. Coordinates (x, y) of the gold particles were manually imported with the use of a Numonics digitizer (Type 1220-6060, Qptronits Veenendaal The Netherlands) into a database/analysis program (Ilja v.d. Paver& this institute) running on a Hewlett-Packard 98168. The area occupied by plasma membrane on each of the micrographs was measured and normalized to 70 pm2 as was the number of detected gold particles. Particle and cluster distributions were calculated and plotted. A cluster was defined as an area that contained at least 10 immunogold particles, each located at no greater distance than 200 nm from the next neighbor. Using this method, cell areas up to 70 pm’, starting from an electron microscopic magnification of 4200X, could be digitized. This area was sufhciently large to plot a leading lamella of a CEF containing one or several adhesion plaques. Image processing of wet-cleaved CEF, double-labeled for talin and vinculin, -was carried out using a gold probe analysis program developed at this laboratory (Arjen Pet and Jan Glorie) within TCL-image software (Technical Command Language, TPD-TNO, Institute of Applied Physics, Technical University Del& distributed by Multihouse B.V., Amsterdam, The Netherlands) running on a Macintosh II. Electron micrographs were digitized by a high-resolution square-pixel CCD video camera (Cohu, GTI, Weesp, The Netherlands) and imported into TCL-image.

140

BRANDS

..

':I i " /_ -

ET

AL.

Transformation

215 190

130

I

of tsRSV-Infected

In experiments described in this study, rounding of tsRSV-infected CEF occurred for most cells within 1 h after temperature downshift. The percentage of wellspread cells fell sharply from 80 to about 35% within 30 min, whereas the percentage of rounded cells increased. With time, increasing numbers of detached cells were observed. Upon wet cleaving, completely rounded but still attached cells were almost completely ripped off the substrate and could not easily be studied with immunoEM. They were studied with immunofluorescence on saponin-permeabilized cells. About 35% of the remaining cells maintained a spread phenotype even after prolonged incubation at the permissive temperature. However, after 24-72 h, these cells also rounded off and detached from the substrate. Immunofluorescence

7

123455

FIG. 1. Characterization of anti-vinculin and anti-talin antibodies. Immunoperoxidase reaction of normal CEF lysate (run on 7.5% PAGE and transferred to nitrocellulose). All antibodies were applied at t8 pg/ml. Molecular weights for vinculin (130 kDa) and talin (215 and 190 kDa) are indicated, standard molecular weight markers referred to under Materials and Methods served as reference. Lane 1, crude pAb vinculin (Vc); Lane 2, affinity-purifiedpAb Vc; Lane 3, mAb Vc; Lane 4, crude pAb talin (Ta); Lane 5, affinity-purified pAb antiTa-215; Lane 6, affinity-purifiedpAb anti-Ta-190; Lane 7, normal rabbit serum (NRS).

RESULTS

Characterization of Polyclonal and Monoclonal Antibodies to Vinculin and Talin The affinity-purified pAb-vinculin was found by Western blotting to be specific for a single 130-kDa protein in CEF homogenate (Fig. 1, lane 2). MAb-vinculin reacted with the same band (Fig. 1, lane 3). Affinity-purified pAb-talin-215 reacted with a 215-kDa protein and in addition with a 190-kDa protein (Fig. 1, lane 5). This 190-kDa protein has been described previously as a proteolytic fragment of talin [al]. Antibodies to this 190kDa protein, affinity-purified on Western blots, reacted with both the 190-kDa and the 215-kDa proteins (Fig. 1, lane 6). The same bands also reacted with pAb talin-215 (Fig. 1, lane 5). All antibodies reacted with aldehydefixed antigens, which were immobilized on nitrocellulose (not shown).

CEF

of Normal

and Transforming

CEF

Immunofluorescence patterns obtained with the affinity-purified polyclonal antibodies to vinculin and talin on wet-cleaved or 0.2% saponin-permeabilized cells were similar to those reported by others [4, 13, 271. All antibodies, including a monoclonal antibody to vinculin, labeled the distinct patches of adhesion plaques in nontransformed CEF cells. Actin microfilament bundles, stained with phall-TRITC, terminated at these structures. No difference in labeling pattern was seen between the two anti-talin antibodies, affinity-purified on either the 190-kDa (Fig. 2B) or the 215-kDa (Fig. 2C) protein band. In all further studies we applied pAb talin215 and we refer to this as pAb talin. Normal rabbit or mouse serum (or IgG) was used as a negative control (not shown). During transformation, the peripheral adhesion plaques disappeared or became much less prominent, and actin microfilaments detached. In cells showing extensive ruffling, more centrally located adhesion plaques with terminating stress fibers were still visible. Just before still-attached cells became spherical, vinculin, talin, and actin had disappeared from distinct structures and only diffuse label was seen. In rounded cells, circular rosette contacts [28] containing talin, vinculin, and actin were occasionally observed. Actin was located at the center of these rosette contacts, and both vinculin and talin more peripherally (not shown). Electron

Microscopy

Standard conditions of wet cleaving resulted in cleaved cells with basal plasma membranes still attached to the substrate. The overall morphology was well preserved, as previously shown using Epon-embedded material [ 11. To demonstrate that cells had been infected by the RSV virus, carbon-platinum replicas were pre-

ALTERATIONS

FIG. 2. Immunofluorescence staining (B), and pAb talin 215 (6). All antibodies

IN

ADHFSIQN

of adhesion plaques react to corresponding

PLAQUES

IN

TRANSFORMING

in nontransformed sites in respective

pared of wet-cleaved CEF cultured at permissive temperature for 6 h after RSV infection and of noninfected cells as negative controls. Only the infected CEF showed numerous virions budding from the plasma membrane (not shown). Because platinum obscured most of the immunogold label, evaporation of platinum onto the specimens was omitted and only brief carbon shadowing was applied for immunocytochemical purposes. With this adapted method, structures like coated pits and vesicles emerging from the uniformly electron dense plasma membrane were clearly visible (arrows, Fig. 4). At low magnification, areas of the cell containing adhesion plaques (depicted rectangle and inset, Fig. 3) were readily discerned. These areas corresponded both in location and in size to areas reacting with anti-vinculin and anti-talin antibodies as observed by immunofluorescence (compare Figs. 2 and 3). At high magnification, adhesion plaques (Fig. 4, areas enclosed by dotted lines) could be recognized easily at the cell periphery because of actin microfila.ments terminating at these structures or, when stress fibers were no longer present, by their relatively high electron density compared to that of the surrounding membrane. Since in immunofluorescence studies talin and vinculin have been found to be highly concentrated at adhesion plaques [3, 4, 291, we considered in immunoelectronmicroscopy these two proteins to be markers for adhesion plaques (see below). During transformation the density of microfilaments in the cortical cytoskeleton decreased. Adhesion plaques with only a few or no terminating microfilaments were often observed. Also the areas adjacent to the adhesion plaques became progressively less occupied with elements of the cortical cytoskeleton, in general prior to the dissociation of the focal contacts themselves. Wetcleaved specimens of cells in advanced stages of transformation no longer contained large continuous ventral membrane layers, but only scattered membrane remnants. This was seen most prominently at the periphery

141

FIBROBLASTS

CEF with affinity-purified cells. Bar 10 urn.

pAb

vinculin

(A), pAb talk-190

of the cell. Remaining membrane areas often contained remnants of adhesion plaques, to which only a few or no microfilaments were attached. The center of the ceil in most of these instances still contained adhesion plaques with terminating microfilaments (not shown). Due to the disappearance of part of the obscuring microfilaments during transformation, a network of interconnected parallel subareas of somewhat increased electrondensity became visible at the inner face of the plasma membrane in adbesion plaques. At each of these

FIG, 3. Wet-cleaved nontransformed CEF containing several adhesion plaques at low magnification. This specimen was indirectly immunolabeled for talin, using 15-nm gold probes. Bar, 2 pm. Inset: higher magnification of an area containing several adhesion plaques (arrows). Note a dense cortical cytoskeleton is present. Bar, 0.5 ,am.

142

BRANDS

ET

AL.

FIG. 4. Electron micrograph of part of the leading lamella of a wet-cleaved tsRSV-infected CEF, maintained at the restrictive temperature, immunoincubated for talin. The plane of cleavage is close to the ventral plasma membrane. The basal plasma membrane (star) can be discerned from extracellular areas exposing an irregular pattern of extracellular matrix (A) by its homogeneous electron density and by the presence of coated pits (small arrows) and actin filaments. The latter often terminate at underlying adhesion plaques (areas enclosed by dotted lines) where immunogold-labeled talin is concentrated. Big arrow: noncleaved area. Bar, 0.25 pm.

substructures remaining microfilaments terminated. The substructure of the adhesion plaque was also observed in nontransformed cells cleaved under drier conditions, when cleavage occurred just above the ventral plasma membrane removing most of the cortical cytoskeleton [ 11. Distribution of Talin and Vinculin at Adhesion during Transformation: Immunocytochemical Localization and Semiquantitative Analysis

Plaques

Wet-cleaved CEF were indirectly immunolabeled for vinculin and talin with immunogold probes of 5, 10, or 15 nm (Au5, Au”, and Au15) using single- and doublelabeling procedures, critical-point-dried, and carbonshadowed. Vinculin and talin were found to be localized at high density in adhesion plaques in noninfected CEF as well as in tsRSV-infected CEF maintained at the restrictive temperature of 42°C. An example of talin immunogold label in the latter cells is given in Fig. 4. Distribution of vinculin and talin corresponded to the subareas present in adhesion plaques, as described above. Other cellular structures were only labeled at background level. If cleaved specimens were not rinsed with buffer prior to initial fixation with 2% paraformaldehyde + 0.01% glutaraldehyde, substantial amounts of label were noted outside adhesion plaques, presumably representing soluble vinculin and talin, aldehyde-attached to the cy-

toskeleton. Background label after using normal rabbit or mouse serum or IgG was very low. Vinculin and talin were labeled in RSV-infected CEF at different time points after temperature downshift (Fig. 5). In noninfected CEF (Fig. 5) or in tsRSV-infected CEF maintained at the restrictive temperature (not shown), talin (Fig. 5A) and vinculin (Fig. 5B) were restricted to the adhesion plaques and microfilament bundles could be seen terminating at these areas. Already 15 min after induction of transformation, label densities for vinculin and talin differed greatly. Whereas all adhesion plaques reacted densely for talin (Fig. 5C, arrows), only very limited reaction was observed for vinculin (Fig. 5D, arrows). At subsequent time points, adhesion plaques became smaller and fewer terminating microfilament bundles were observed at peripheral plaques. Adhesion plaques became devoid of vinculin and contained progressively less talin label. Terminating microfilaments were observed only in those cases where talin label was still present. An example of a digitized image is given in Fig. 6 depicting an adhesion plaque-containing cell area labeled for talin 15 min after temperature downshift. Data from a series of digitized micrographs taken at increasing time intervals during transformation, shown in Fig. 7, demonstrate that the velocities of change were different for vinculin and talin. The amount of detected vinculin fell sharply within minutes after temperature downshift,

ALTERATIONS

FIG. areas of talin (A, is much

IN

5. Immunocytochemical staining of wet-cleaved noninfected CEF (A, B) C) or mAb-vinculin (B, D). At 15 min decreased (D, arrows) or absent. Bar,

ADHESION

PLAQUES

IN

TRANSFORMING

FIBROBLASTS

talin and vinculin during transformation. Shown are examples within one experiment and tsRSV-infected CEF 15 min after temperature downshift (6, D), immunolabeled after induction of transformation, label for talin remains dense (6, arrows) while that 0.25 pm.

whereas tahn label, although decreasing, was more persistent. For vinculin after 30 min at permissive temperature this resulted in less than 10% of the initial amount, in contrast to approximately 50% for talin (Fig. 7A). The rather large standard deviations were due to the fact that although several areas of the same size were digitized at given time points, these areas contained different num-

143

of similar with pAbfor vinculin

bers of adhesion plaques of various sizes and therefore different numbers of gold particles. To circumvent this variable, the number of immunogold clusters per 1000 digitized gold particles (see Materials and calculated (Fig. 7B). In noninfected cells (c,), large numbers of either vinculin- or talk-associated gold particles were found in relatively small numbers of (large) adhe-

144

BRANDS

FIG. 6. after

Example of a micrograph, 15 min at permissive temperature

0) ,” 0 ._.

ET

digitized and imported into a database (A) is plotted using dedicated software

AL.

program. Immunogold label-detecting (B). The arrows indicate corresponding

6000 I

80

LEE

200

Ia

308

28

30

40---e--

t i me A

FIG. 7.

talin at adhesion plaques sites. Bar, 0.25 pm.

time

(min)

60

Cmin)

B

Quantitation of single-labeled vinculin (solid lines) and talin (dashed lines) at different time intervals after induction of transformation. Vertical bars indicate the standard deviation. to, Noninfected cells. (A) Number of immunogold particles on 70-pm2 membrane areas. The number of detected vinculin epitopes decreases at a higher rate than that of talin. (B) Numbers of clusters per 1000 immunogold particles during the first 60 min after induction of transformation. A cluster contains at least 10 immunogold particles, each located less than 200 nm apart from its next neighbor. Note that the number of clusters detected for talin remains constant whereas that for vinculin rises during the first 30 min.

ALTERATIONS

IN

ADHESION

PLAQUES

sion plaque areas, identified by the gold cluster computer program as (at least an equal number of) “clusters.” Within 5-10 min, the number of vinculin clusters increased, while that of talin remained about equal. The single-labeling experiments had shown that the number of vinculin epitopes within the adhesion plaque rapidly decreased (Figs. 5B and 7A). From this we deduced that the distances between remaining particles increase and, as a consequence, larger numbers of smaller clusters were detected. At 60 min the cluster number for vinculin had decreased. This was most likely due to the fact that remaining nondetached cells retained a spread morphology with less affected adhesion plaques. Double Labeling The above data on vinculin and talin were necessarily obtained in separate experiments. To show that the results were not due to sample bias, we performed doublelabeling experiments. Similar results were obtained using either of the following indirect immunogold incubations. Vincuhn was detected with Au5 and talin with Au” (Fig. 8) or, reversily, vinculin was labeled with Au” or Au15 and talin with Au5 (not shown). Large immunogold probes have the inherent disadvantage that no high label density is obtained. For instance, at least a IO-fold higher label density was obtained with Au5 probes than with Aul’ or Ad5 probes. An example of an immuno double-labeling reaction using monoclonal antibodies for vinculin and affinity-purified polyclonal antibodies for talin is given in Fig. 8. The density of vinculin label fell sharply within an hour after temperature downshift, in contrast to the talin label. Alternatively, we also double labeled with both affinity-purified polyclonal antibodies to vinculin and talin, respectively. One antigen was detected with 6 X R-Au’, whereas the other antigen was detected with D X R-biotin and streptavidin-Au15 conjugates. The biotin’streptavidin system enables labeling with large gold particles at much higher density, albeit at the cost of higher background. Figure 9 shows an example of such an immunoincubation. Again a decrease in vinculin relative to talin label was detected. The ratio between vincuhn (immunogold probe, 15 nm) and tahn (immunogold probe, 5 nm) labels within the same adhesion plaque was determined in double-labeled specimens” We observed a sharp decline in the vinculin-talin ratio after temperature shift over the course of 1 h (Fig. 10). When mAb vinculin was detected with tahn with Au’, similar results were obtained, as was also the case when vinculin and talin were detected with two pollyclonal antibodies (Fig. 9).

Several authors embryo fibroblasts

have described changes in chicken occurring upon induction of transfor-

IN

TRANSFORMING

FIBROBLASTS

145

mation by temperature-sensitive 8 san2oma virus [3, 11, 27, 301. Al! authors rep early events like ruffling and some (e.g. [3 describe complete transfor hands, transformation was the cells: 65% of tsRS within 1 h after shift to th Using immunofluorescence, vincuhn and talin were found by several groups to be localized at the distinct patches of adhesion plaques [3,29]. Geiger et al. [31] and Nicol et ak. 1321have performed ultrastructural studies showing vinculin to be localized in adhesion plaques, but such electron microscopical data are relatively scarce, possibly due to technical difficulties. this report, our recently described wet-clieaving method [l] is quite suitable for this purpose. We show here the changes in d~strib~t~~~ of vinculin and talin in CEF after induction of trans most striking finding was that the relativ an munogold particles detecting anti-vim&n multiple single-label experiments decreased very rapidly. A decrease in immunogold particles detecting tahn was slower and less pronounced. The velocity of the changes varied per cell and per ex~e~~~~@~t~so that it was difficult to exclude that the apparent changes were due to sample bias. However, we have seen the same changes in double-labeling experiments which allowed the relative amounts of vincuhn and tahn to be assessed in the same adhesion plaque. Shortly after shift to permissive temperature, the vincuhn-tahn ratio fell sharply, showing that vinculin d~sa~~ea~~~more rapidly from the same plaques than tahn. observation is observed irrenot likely to be an artifact, since i spective of whether tahn or vincuhn was labeled with the larger gold particles. Furthermore, different procedures, first with a against vinculin and a polyclonal an and second, using two polyclona which was detected with immu~~-b~~t~r~~~~re~tavi~i~gold. Our observations are in agreement with a r vincuhn dissociation from adhesion plaques in c-3T3 cells upon PDGF stimulation, in which cells vinculin dissociated from the adhesion plaque prior to tahn with kinetics similar to that we hav In witro studies have demon&rat integrin [6, 141 and talin to vincuhn [7] and suggested indirect binding of vincuhn to actin a a-actinin [?, 10 ve been propose Thus, models of adhesion plaques suggesting that the structure of an adhesion plaque is based on talin and vinculin linking the membrane-scabning integrins to actin [3, Burridge [3] has argued likely to be the only linker molecules. vations support tbis notion, since we have phologically recognizable adhesion plaqu

146

BRANDS

ET

AL.

FIG. 8. Electron micrographs of tsRSV-infected CEF adhesion plaque areas double-labeled for talin and vinculin at 5 min (A), 15 min (B), 20 min (C), and 60 min (D) at permissive temperature. Vinculin was indirectly labeled with mAb-vinculin and 5 nm immunogold and talin with pAb-talin and 10 nm gold probes. After 20 min of transformation seemingly intact adhesion plaques with terminating actin microfilaments are densely labeled for talin only, and not for vinculin (C, D). Insets show the peripheral regions of respective CEF, where the depicted adhesion plaques locate. Arrows in each micrograph direct to corresponding areas within the cell. Bars, 0.1 pm Inset bars, 2.5 pm.

transiently with attached microfilaments while very little vinculin was present. If vinculin is essential for linking actin to the adhesion plaques, it is difficult to imagine how the observed actin filaments could remain attached. However, this does not exclude an important role for vinculin in the maintenance of the stability of the adhe-

sion plaque, as suggested by recent observations of Bendori et al. [34] indicating that vinculin interconnects talin molecules due to self-association of vinculin molecules. Vinculin is one of the substrates for the ~~60”.“” tyrosine kinase. Thus, the very rapid disappearance of vin-

ALTERATIONS

IN

ADHESION

PLAQUES

IN

TRANSFORMING

147

FIBROBLASTS

FIG. 9. Double labeling of talin and vinculin during transformation using two polyclonal antibodies. Vinculin was indirectly labeled with p4b-vinculin and biotin-streptavidin-Au’“. Talin was indirectly labeled with immunogold probes of 5 nm. Shown are parts of adhesion plaques in noninfected CEF (A) and RSV-infected CEF at, respectively, 5 min (B) and 1 b (C) at permissive temperature. Note the sharp decrease in vinculin label density during this period. Bar, 0.25 am.

culin after induction of transformation by this kinase is likely to be due to the tyrosine phosphorylation of vinculin. This notion is supported by our preliminary results using 32P-labeled cells suggesting that vinculin is very rapidly phosphorylated with the same kinetics as t,he observed disappearance from adhesion plaquesmuch more rapidly than talin (Brands et al., unpublished

results). If confirmed, this means that ~~bos~hor~Iatio~ of vinculin does contribute to arc-in of adhesion plaques, even though occur in the absence of vinculin ph phosphorylation of other substrates like integrins, talin, or others 1351. We thank Caroline N. van der Maas for excellent technical assistance. The assistance of Harry van Drimmelen in early stages of this work is also acknowledged. Special thanks are due to Ilja van de Pavert for the development of the gold cluster cmmputer proigram, to Arjen Pet and Jan Glorie for the development of the software used for analysis of immunogold double-labeled specimens, and to Nice Ong for photographic work. This work was supported by Grant 85-4 from the Dutch Cancer Foundation.

FIG. 10. Quantitation of double immunolabeling of vinculin and talin within the same adhesion plaque during transformation of tsRSV-infected CEF. Depicted is the ratio between the vinculin-associated A# label and the talin-associated Au5 label (solid line). The absolute numbers of vinculin-associated (+. +) and talin-associated (O---O) gold particles per square micrometer, as detected by automated image processing, indicate that the decline of the ratio is caused by a decrease in vinculin-associated label.

1.

Brands, 318.

R., and Feltkamp,

C. A. (1988)

Exp.

Cell lies.

2.

Tamkun, J. W., DeSimone, D. W., Fonda, D., Patel, C., Horwitz, A. F., and Hynes, R. 0. (1986) Cell4

176,309R. S., Buck,

3.

Burridge,

K. (1986)

Cancer

Reu. 4,18-78.

4.

Burridge,

K. (1988)

Annu.

Rev. CeEl Biol.

5.

Geiger, B., Avnur, Z., Rinnerthales, V. J. (1984) J. Ceil Biol. SES, 83-91.

6.

Ilorwitz, A., Duggan, K., Buck, K. (1986) Nature (London) 32

7.

Burridge, Hi., and Mangeat, P. (1984) Nature (London) X)8,744146. Burridge, K., and Feramisco, J. R. (1982) Cold Spring Harbor Symp. f&ant. Biol. 46,587-597.

8.

4,487-525.

G., Hinssen, ., Beckerle,

II., and Small,

M. C., and Barridge,

148 9. 10.

Schliwa, 21a.

M.,

and Potter,

M. (1986)

Wilkins, J. A., Risinger, 103,1483-1494.

M.

Cell

Tissue

A., and Lin,

S. (1986)

Boschek, C. B., Jockusch, B. M., Friis, R. R., Back, mann, E., and Bauer, H. (1981) Cell 24,175184.

12.

Hirst, R., Horwitz, A., Buck, C., and Rohrschneider, Proc. Natl. Acad. Sci. USA 83,6470-6474.

13.

Pasquale, E. B., Maher, P. A., and Singer, Acad. Sci. USA 83,5507-5511. Tapley, P., Horwitz, A., Buck, C., Duggan, der, L. (1989) Oncogene 4,325-333.

15.

Rohrschneider, 731-746.

L. R., and Rosok,

16.

Rohrschneider, 3097-3107.

L. R., and Reynolds,

17.

Antler, A. M., Greenberg, H. (1985) Mol. Cell. Biol.

18.

Nigg, E. A., Sefton, B. M., Virology 15 1,50-65.

19.

Kellie, Wyke,

20.

Kawai,

21.

O’Halloran, T., Molony, in Enzymology (Vallee, Press, New York.

Received Revised

S., and Hanafusa,

May 17,1989 version received

AL.

211-

22.

J. Cell Biol.

23. 24.

Feramisco, 1194-1199. Laemmli,

R., Grund25. L. (1986) 26.

S. J. (1986)

Proc. Natl.

K., and RohrschneiMol.

Cell. Biol.

3,

S. (1985)

Mol.

Cell. Biol.

5,

Singer,

G. M., and Hanafusa,

S. J., and Vogt, N. M.,

P. K. (1986)

Critchley,

H. (1971)

Virology

17, 1989

29. 30. 31. 32.

33. 46,470-479.

L., and Burridge, K. (1986) in Methods R. B., Ed.), Vol. 134, pp. 69-77, Academic

August

28.

D. R., and

165,216-228.

Towbin, Acad. Becker, (1977) Roos, F. A.

J. R., and Burridge, U. K. (1970)

Nature

K. (1980) (London)

J. Biol.

Chem.

255,

227,680-685.

H., Staehelin, T., and Gordon, J. (1979) Proc. Natl. Sci. USA 76,4350-4354. D., Kurth, R., Critchley, D., Friis, R. R., and Bauer, H. J. Virol. 21,1042-1055. E., Spiele, H., Feltkamp, C. A., Huisman, H., Wiegant, C., Traas, J., and Mesland, D. A. M. (1985) J. Cell Biol.

101,1817-1825. 27.

M. J. (1983)

M. E., Edelman, 5,263-267.

S., Patel, B., Wigglesworth, J. A. (1986) Exp. Cell Res.

ET

Res. 246,

11.

14.

BRANDS

David-Pfeuty, T., and Singer, S. J. (1980) Proc. Natl. Acad. Sci. USA 77,6687-6691. Marchisio, P. C., Cirillo, D., Teti, A., Zambonin-Zallone, A., and Tarone, G. (1987) Exp. Cell Res. 169,202-214. Geiger, B. (1979) Cell l&193-205. Edelman, G. M., and Yahara, I. (1976) Proc. Natl. Acad. Sci. USA 73,2047-2051. Geiger, B., Tokuyasu, K. T., Dutton, A. H., and Singer, S. J. (1980) Proc. Natl. Acad. Sci. USA 77,4127-4131. Nicol, A., Nermut, M. V., Doeinck, A., Robenek, H., Wiegand, C., and Jockusch, B. M. (1987) J. Histochem. Cytochem. 35,499506. Herman, B., and Pledger, W. J. (1985) J. Cell Biol. 100, 10311040.

34.

Bendori, 2383-2393.

R., Salomon,

35.

Glenney,

J. R., and Zokas,

D., and

Geiger,

L. (1989)

B. (1989)

J. Cell Biol.

J. Cell Biol.

108,2401-2408.