Exp. Eye Res. (1990) 51, 191-198
Identification
and Localization of Talin in Chick Epithelial Cells NANCY
Department
of Anatomy, (Received
J. PHILP’,
University 79 September
MI-YUNG
of Pennsylvania
YOON
AND
RICKS.
School of Medicine,
Retinal
Pigment
HOCK
Philadelphia,
PA 19104, U.S.A.
1989 and accepted in revised form 27 December
1989)
Retinal pigmented epithelial cells are adherent at their basal surface to Bruch’s membrane and at their apical surface to the neural retina. We examined the expression and distribution of two proteins that are found in regions of cell-matrix interaction, talin and integrin. Talin is a 23 S&Da cytoplasmic protein that has been localized to regions of cell-substrate adhesion. It binds to both integrin, a transmembrane glycoprotein complex, and to vinculin, a cytoskeletal protein. In the present study, we produced a polyclonal antibody to chicken gizzard talin. Using this antibody we showed by western blot analysis that talin is expressedby RPE cells and is found in the triton-soluble fraction. Talin was shown to co-localize with integrin and vinculin in the basal region of chick RPEcells isolated from 18-day-old chick embryos. Neither talin nor integrin was found in the apical processesor in the zonula adherens. Antibodies to vinculin showed staining both in the apical and basal regions of the RPEcells. The localization of integrin. talin and vinculin along the basal membrane suggeststhat this complex is important in the attachment of the RPE cells to the basement membrane. The distribution of integrin and talin was examined in primary cultures of RPEcells grown on permeable filters. In these cells, a polarized distribution of integrin and talin was not observed. This may suggest that the neural retina may be important for maintaining the differentiated state of the RPE cells. Key words: talin ; retinal pigment epithelium : cell adhesion : integrin ; vinculin. and identified by immunolocallzation
1. Introduction Like most epithelia, RPEcells are both structurally and functionally polarized. The basal surface of the RPE cells rests on a continuous basal lamina and the apical surface elaborates microvillous processesthat are in intimate association with the photoreceptor cells. The apical and basal-lateral membrane domains of the RPE cells are separated by junctional complexes that form a continuous ring around the apex of the cells. While most epithelial cells are in contact with one matrix compartment, the RPE cells are in contact with two matrix compartments. The interphotoreceptor matrix fills the spacebetween the apical processesof the RPE and the outerlimiting membrane of the neural retina. The interphotoreceptor matrix is composed of sialoglycoconjugates, chondroitin sulfate A and C, hyaluronate and heparan sulfate (Hewitt and Newsome, 1988). The basal surface of the RPE rests on a continuous basement membrane that is composed primarily of type IV collagen, heparan sulfate, laminin and fibronectin (Turksen et al., 1985). The attachment of the RPE cells to both the basement membrane and
the neural retina is critical for maintaining the nonproliferative, differentiated state of these cells (for review, Steinberg, 1986). However, little is known about the adhesion molecules expressed by the RPE cells or the linkage of adhesion molecules to the cytoskeleton. Talin, a large cytoplasmic protein (235 kDa), was originally purified from chicken gizzard smooth muscle
of talin has since been studied by immunolocalllation
in a variety of cultured cells where it has been seen to be concentrated in adhesion plaques (for review see Burridge et al., 1988). Within migratory cells, talon,
like integrin, has a diffuse distribution, while in the intestinal epithelium it is found along the basal surface of the cells adjacent
to the substratum
(Drenckhahn et al., 1988). In cultured RPE cells, integrin, talin and vinculin
have been identified to be
components of focal contacts (Turksen, Opas and Kalnins, 1987 ; Opas, 1989). The distribution of talin in RPE cells in vivo has not yet been examined. In the present study, we have produced a polyclonal antibody to chicken gizzard talm. With this antibody we showed by western blot analysis that talin is expressed by RPE cells. Using indirect immunofluorescence microscopy we examined the distribution of talin, integrin and vinculin cultured RPE cells.
in freshly isolated and
2. Materials and Methods Embryos
White Leghorn chick embryos used in these studies were procured from a local supplier and incubated in a forced draft incubator at 38°C.
* For correspondence. 00144835/90/080191+08
as a component
of adhesion plaques (Burridge and Connell, 198 3a, b). It has been shown in vitro that talin binds both to vinculin (Otto, 1983; Burridge and Mangeat, 1984) and to integrin (Horwitz et al., 1986). The distribution
$03.00/O
0 1990 Academic Press Limited
192
Cell Culture Retinal pigment epithelial cells were isolated from 9day-old chick (E9) embryos in PBS/EDTA buffer, as previously described (Philp and Nachmias, 198 5). The cells were grown on millipore filter inserts (MillicellHATM, 12.5 mm) in a medium consisting of DMEM containing 5 % (v/v) FBS and penicillin (5U ml-l) and streptomycin (5 ,ugml-l) (GIBCO). Approximately 5.0 x lo5 cells were plated onto each filter unit. Cells were grown to confluence then fixed and sectioned for the immunofluorescence studies. In some cases cells were grown in serum-free medium PC-l (Ventrex, Portland, ME). The medium was tested by immunoblot analysis and was shown to be free of fibronectin. Antibodies Polyclonal antibodies to the /3 l-subunit of the avian integrin complex were a gift of Dr Clayton Buck (The Wistar Institute, Philadelphia, PA). Antibodies to vinculin were produced against chicken gizzard vinculin and were a gift of Dr Susan Craig (Johns Hopkins University, Baltimore, MD) and antibodies to human plasma fibronectin were purchased from BRL (Bethesda Research Laboratories, Bethesda, MD). Antibodies to chicken gizzard talin were raised in New Zealand rabbits in our laboratory. The talin was purified from frozen chicken gizzards (Pel-Freeze) according to published procedures (O’Halloran, Molony and Burridge, 1986; Hock, Sanger and Sanger, 1989). Highly purified talin was coupled to cyanogen-activated Sepharose 4B (Sigma) for use in affinity purification of the antibody. SDS-PAGEand Western Blots RPE cells were isolated from 17- and l&day-old embryonic chick eyes, as previously described (Philp and Nachmias, 1985). Proteins from whole RPE cells (about 50 ,ug or one quarter of an eye per lane) and triton-insoluble cytoskeletons (75 rug or two eyes per lane) were electrophoresced on microslab gels using the buffer system of Laemmli (1970). Whole platelet lysate was used as standards on the gels since they contained large amounts of talin (23 5 kDa protein) (O’Halloran, Beckerle and Burridge, 1985) and vinculin (130 kDa) (Langer, Gonnelle and Nachmias, 1984) which cross-reacted with antibodies against chicken gizzard talin and vinculin. Gelswere fixed and stained in 0.25 % CoomassieBrilliant Blue R in 50 % methanol, 10% acetic acid, and were destained by diffusion in 5% methanol and 10% acetic acid. For Western blots the proteins separated on SDS-PAGE were electrophoretically transferred to nitrocellulose paper (Biorad Laboratories, Richmond, CA) or Immobilon-P transfer membrane (Millipore Corporation, Bedford, MA) at 100 V for i hr using a MiniTransblot apparatus (Biorad Laboratories, Richmond,
N. J. PHILP
ET AL.
CA). The blots were incubated in a buffer containing 5% BSA, 20 mM Tris, 500 mM NaCl, pH 7.5 for 1 hr to block non-specific binding sites. The Immobilon membranes were hydrated with 100% methanol for 5 set, washed with distilled water, then incubated in blocking buffer. The blots were incubated in the primary anti-sera for at least 1 hr. Optimal dilutions for the primary antisera were experimentally determined to be 1: 200 for affinity purified anti-talin (1.25 mg ml-‘), 1: 100 for anti-vinculin and antiintegrin (14 pugml-‘). Antisera were diluted in 1% BSA, 0.05 % Tween-20, 20 mM Trisma Base, 500 mM NaCl, pH 7.5. The blots were washed three times with this buffer and incubated for one hour with alkaline phosphatase conjugated goat anti-rabbit IgG at a dilution of 1: 3000 (Biorad Laboratories), followed by two more washes before developing with 0.3 mg ml-’ NBT (p-nitro blue tetrazolium chloride) and O-15 mg ml-’ BCIP (S-bromo-4-chloro-3-indolyl phosphate p-toluidine salt) in a buffer of 0.1 M NaHCO,, 1.0 mM MgCI,, pH 9.8. Immunoprecipitation RPE cells were isolated from six eyes (E18) as described above and solubilized in cytoskeletal extraction buffer (CEB) containg 100 mM KCI, 1 mM MgCl,, 1 mM EGTA, 10 mM imidazole pH 7.2, 1% Triton X-100, 1 mM benzamidine and 0.05 mg ml-’ leupeptin. After 15 min on ice, the extract was centrifuged at 10 000 g for 15 min and the supernatant collected. Affinity purified anti-talin (S-10 pg) was incubated with 100 ~1 of a 10 % solution of SDS stripped IgGsorb (The Enzyme Center, Malden, MA) which had been for 1 hr in the cold, then washed four times in cytoskeleton extraction buffer. The anti-talin IgGSorb was incubated in the cold for another hour with the RPEcell supernatant. After 1 hr, the IgGsorb was spun for 2 min at 10 000 g and the supernatant was discarded. The pellet was washed four times in CEB,boiled in Laemmli sample buffer, and the sample run on 7.5 Y0or 10 y0 SDSgels. Phosphorylation Cultured chick RPE and freshly isolated El8 RPE cells were labelled with azP1 mCi ml-’ for 1 hr in MEM without phosphates (Flow Laboratories, Mclean, VA). At the end of the incubation period, the cells were washed with cold PBS three times and were then extracted into 0.5 ml of a buffer containing 1% Nonidet P40, 100 mM sodium pyrophosphate, 250m~ NaCl, 50m~ NaF, 5 mM EGTA, 1 mM benzamidine, 10 ,ugml-’ leupeptin, 15 mM pmercaptoethanol and 20 mM Tris HCl, pH 7.9. The extract was spun in an eppendorf centrifuge for 10 min and the supernatant taken and incubated for 1 hr with anti-talin bound to IgGsorb, as described above.
TALIN
193
IN RPE CELLS
Immunocytochemistry Posterior eye cups from El8 and sheets of RPE cells isolated from El 8 were ilxed for 30 min in 2 % paraformaldehyde containing 1 mu MgCl, and 3% sucrose in PBS. The tissues were washed with three changes of PBSwith 3 % sucrose, then put in PBSwith 30% sucrose overnight in the cold. Tissues were embeddedin Tissue Tek II 0.C.T Compound (American Scientific Products, Edison, NJ) and frozen in liquid nitrogen. Six to 12 pm cryostat sections were mounted on gelatin coated coverslips. The sections were flxed for 1 min in 2% paraformaldehyde, 3% sucrose and 1 mu MgCl, in PBS, and non-specific binding sites on the tissue were blocked for 30 min in 1% goat serum in PBS.All washes and antibody dilutions were made in 1 Y0BSA and 0.05 Y0Tween-20 in PBS.The sections were incubated with the primary anti-sera for 1 hr. Talin was used at 10 mg ml-‘, fibronectin was diluted 1: 100, integrin 1: 100, laminin 1: 100, and afEnity purified vinculin 1: 15. After thorough washing, rhodamine conjugated goat anti-rabbit F(ab’), fragment (Cooper Biomedical, Westchester, PA) was applied for 45 min. Sections were washed, and the coverslips mounted with gelvatol (Aii Products and Chemicals, Inc, Allentown, PA). Slides were examined with phase or epiiluoresecence on a Zeiss Research microscope equipped with a 63 x phase objective with an N.A. of 1.4. Photographs were taken with Kodak Tri-X film. The negatives were developed for about 7 min in HC-110 developer dilution B. Fresh tissue samples were prepared each time for the cryostat sectioning. Two eye cups or sheetsof RPEcells isolated from about four eyeswere frozen in each block and the tissues were stained immediately after sectioning. The immunocytochemical studies were repeated a minimum of three times with similar results.
1
2
FIG. 1. Immunoblot analysis of whole RPE cells (1) and
purified chicken gizzardtalin (2) with affiity purified talin antibody. The antibody cross-reactswith a single band in whole RFE cells that migrateswith the samemobility as chicken gizzard talin with a molecular weight of about 235 kDa (1). The antibody also cross-reacts with the 190 kDa breakdown product of talin (2).
FIG. 2. Immunoblot analysis of whole chick RPEcells and cytoskeletons with antibodies to the p-subunit of integrin (120 kDa) (A), talin (235 kDa) (B), and vinculin (117 kDa) (C). Lane 1, whole platelet lysate, lane 2, whole RPE cells and lane 3, RPE cytoskeletons.
3. Results
Polyclonal antibodies to chicken gizzard talin were produced in rabbit and aiiinity purified on a talinsepharosecolumn. The antisera were characterized on western blots and were shown to cross-react with both purified chicken gizzard talin and with a protein in RPE cells that had a similar electrophoretic mobility (Fig. 1). The antibody cross-reacted with both the intact protein (23 5 kDa) as well as with the 190 kDa degradation product (O’Halloran and Burridge, 1986). Additionally, as shown in Fig. 2(B), lane 1, the antibody cross-reactedwith the 2 3 S-kDa protein from human platelets which has been identified as talm (Beckerle, O’Halloran and Burridge, 1986). Talin from chicken gizzards, chicken RPE cells and human platelets all ran with the same relative mobility in our gel system. To determine whether or not talin remained lied to the cytoskeleton after triton extraction, whole RPE cells and cytoskeletal proteins form l&day-old 14
em-
bryos were separated on polyacrylamide gels and transferred to nitrocellulose or ImmobilonTM paper. The blots were probed with antibodies to integrin, talin and vinculin. Lanes 2 contain RPEproteins from one third of an eye per lane, while lanes 3 are RPE cytoskeletons from two eyes per lane. Despite the fact that more cytoskeletal than whole cell proteins were loaded onto the gel little talin or integrin remained bound to the triton-insoluble cytoskeleton of RPE cells [Figs 2(A) and (B), lanes 31. Vinculin, which is found both in areas of cell-cell and cell-substratum was present in the cytoskeletons of RPE cells [Fig. 2(c)]. Some minor protein bands were detected with the vinculin antibody in whole platelet lysate, whole RPE cells and in RPE cytoskeletons. These may represent isoforms and breakdown products of vinculin or it could be non-specific binding. The afBnity-purified talin antibody was used to immunoprecipitate talin from tt-iton extracts of whole EER 51
N. J. PHILP
194
A
talinm
--
,.--
vine
a
1 2 1' 2' FIG. 3. A, Immunoprecipitationof talin from triton soluble extract of whole RPEcellsrun on a 10% polyacylamidegel. Lane 1, Coomassieblue stainedgel of whole platelet lysate, lane 2, whole RPE cells, and lane 3, talin immunoprecipitatedfrom triton solubleextract of whole RPEcells. The lower molecular weight bands seenin lane 3 are IgG heavy and light chain subunitsof the anti-talin. B, Western blot analysisof proteinsfrom wholeplateletlysate(1, 1’) and immunoprecipitated talin (2, 2’) run on 7.5y0 polyacrylamide gels and transferred to ImmobilonTMpaper. The westernblotswereprobedwith antibodiesto talin (1,2) and vinculin (l’, 2’). The talin antibody cross-reactedwith the 235-kDa protein in platelets and with the immunoprecipitatedtalin. Antibody to vinculin cross-reactedwith vinculin in plateletsbut was negativein lane 2’, showing that vinculin doesnot coprecipitatewith talin. RPE cells. The immunoprecipitated protein migrated with the same electrophoretic mobility as human platelet talin [Fig. 3(A)]. When samples from human platelet and chick RPEcells were blotted and incubated with the antibody to talin, both the 235-kDa protein in platelets and the protein immunoprecipitated from RPE cells cross-reacted with the talin antibody. The same blot was probed with antibody to vinculin to determine whether vinculin coprecipitated with talin. It was shown on at least five separate occasions that only trace amounts of vinculin coprecipitated with talin [Fig. 3(B)]. Integrin did not coprecipitate with talin (data not shown). Indirect immunofluorescence was used to localize talin, fibronectin, integrin and vinculin in El8 chick
ET AL.
RPE cells (Fig. 4). Antibodies to talin showed bright staining along the basal surface of the RPE cells with no significant staining along the apical region of the RPE cells. No staining was seen with the anti-talin in the area of the junctional complexes or in the apical processes.As previously shown (Philp and Nachmias, 198 7) integrin is localized along the basal surface of the RPEcells where the cells attach to the substratum. To insure that the staining observed in the intact eye cup was from RPE cells and not from the endothelial cells, the cells were isolated from Bruch’s membrane after incubation in PBS containing EDTA. Frozen sections of the isolated sheetsof cells were stained with antibodies to talin and integrin (Fig. 5). As previously shown in Fig. 4, talin and integrin colocalized along the basal membrane of the RPE cells. Vinculin was localized at both the basal and apical membranes and at the intercellular junctions as previously shown (Philp and Nachmias, 1985; Turksen et al., 1987). Laminin and fibronectin did not remain bound to the RPEcells after they were dissociatedfrom the basement membrane in EDTA [Figs. 5(F) and (G)]. The distribution of talin and integrin was examined in cultured RPE cells that were grown on permeable supports. Frozen sections of these cells showed that in contrast to the polarized distribution of integrin and talin seen in vivo, in cultured RPEcells, these proteins were not polarized. Integrin was localized on both the apical and basal-lateral membranes [Fig. 6(B)]. Talin staining was seen to be diffuse throughout the cells [Fig. 6(C)]. Despite the fact that these cultures were grown in media containing serum, fibronectin staining was only observed along the basal surface of the RPE cells [Fig. 6(A)]. This suggests that the fibronectin in the serum is not tightly bound to integrin in the apical membrane of the RPE cells. Culturing the cells in serum free media in the apical compartment and media containing serum in the basal compartment did not alter the staining pattern of talin, integrin or fibronectin (data not shown). Even when we cultured the cells on Bruch’s membrane we did not see a polarized distribution of integrin. When whole mounts were examined, we found that there was a concentration of integrin, talin and vinculin in adhesion plaques in agreement with the work of Opas (1989). It has been reported that talin is phosphorylated both in vivo and in vitro (Litchfield and Ball, 1986). We labeled culture RPE cells with 32Pand immunoprecipitated talin from the triton-soluble extract to determine its state of phosphorylation. Gel electrophoresis and autoradiography showed that the immunoprecipitated talin was phosphorylated in cultured RPE cells (Fig. 7). Additionally, other phosphorylated proteins coprecipitated with talin. We do not know the identity of these bands or whether their coprecipitation with talin is specific or non-specific. Using the same labeling and immunoprecipitation procedures on freshly isolated RPE cells, we were unable to demonstrate that talin was phosphorylated (data not shown).
TALIN
IN RPE CELLS
FIG. 4. Frozen sections of lb-day-old chick retina. Phase micrograph (A), and similar sections stained with antibodies to fibronectin (B). integrin (C), talin (D), and rhophalloidin (E), and secondaryantibody (F). Arrows indicate the basalregion of
the RPEceils.Bar = 10 pm.
4. Discussion In the present study we examined the expression and distribution of two proteins involved in cellsubstratum adhesion, integrin and talin. Previously we reported that integrin was localized only on the basal surface of RPEcells (Philp and Nachmias, 198 7) though subsequent reports suggested that it was on the apical surface of the RPEcells as well (Anderson et al., 1989). This raised the possibility that integrin could be involved both in cell-cell and cell-substratum adhesion. To help address this issue, we decided to
study more closely another protein that binds to integrin and links it to the cytoskeleton, talin. Therefore, we produced a polyclonal antibody to talin to use in these studies. Our results showed that in RPE cells isolated from 18-day-old chick embryos, there was discrete localization of integrin and talin along the basal surface of the RPE cells. No significant staining was seen in the apical region of the RPE cells or in the area of the junctional complexes. Two different polyclonal antibodies to talin, one produced by K. Burridge and one produced in our laboratory,
yielded similar staining 14-2
196
N. J. PHILP
ET AL
FIG. 5. Frozensectionsof isolated sheetsof 18 d RPEcells. Phasemicrograph (A), and similar sectionsstained with rhophalloidin (B),and antibodiesto vinculin (C),talin (D),integrin (E).laminin (F),fibronectin(G),andsecondaryantibodyonly (H). The small arrows indicate the basalregion of the RFEcells.Bar = 10 pm. patterns on tissue. Therefore, it is unlikely that we are not detecting talin and or integrin in other regions of the cell. Vinculin, a cytoskeletal protein that is found in all types of adhesion junctions (for review seeGeiger et al., 198 7), was localized by immunoflurescent staining along the basal surface of the RPE and in the apical region of the cell. While talin binds to vinculm, they codistribute only in regions of cell-substratum adhesion in RPE cells. It was recently reported that when fluorescently-labeled talm was microinjected into Madin Darby bovine kidney cells it was localiied only in focal adhesions. However when the 200~kDa domain was injected into these cells it was incorporated into the zonula adherens as well as into the focal contacts (Nuckolls, Turner and Burridge, 1989). This suggeststhat the 47-kDa fragment regulates the specific distribution of talin. We have preliminary evidence to support a direct role for integrin in the attachment of RPE cells to the basement membrane. The monoclonal antibody to the integrin receptor, CSAT, inhibits the attachment of freshly isolated RPE cells to both laminin and fibronectm (Yoon and Philp, unpubl. res.). This would be consistent with reports which show that avian integrin binds to both laminin and fibronectin (Horwitz et al., 1985; Bronner-Fraser, 1985). The CSAT antibody did not disrupt the attachment of RPEcells to Bruch’s membrane when explants of RPE cells were
incubated in antibody containing media. This suggests that integrin is not the only protein in the RPE cells regulating cell-substratum adhesion. The distribution of integrin and talin in normal and pathological states such as retinal detachment is of interest since it has been shown that integrin can modulate both cell adhesion and cell migration (Duband et al., 1986). Under normal conditions integrin and talin may stabilize the attachment of the RPE cells to the basement membrane while in pathological states these same proteins may mediate the migration of these cells. Studies have shown that RPEcells can migrate along a fibronectin gradient and that this migration is inhibited by the tetrapeptide RGDS(Avery and Glaser, 1986). It would be of interest to determine whether or not there is a redistribution of talin and integrin after retinal detachments. We examined the distribution of talin and integrin in RPEcells grown in tissue culture. The cultured RPE cells did not exhibit a polarized distribution of integrin or talin, even after several weeks in culture. We thought that components of the serum such as fibronectin could have an effect on the distribution of integrin and talin. To test this hypothesis we grew cells on millipore inserts with serum free media in the apical compartment and serum containing media in the basal compartment. Even under these conditions we found integrin on both the apical and basal
TALIN
197
IN RPE CELLS
talint
---
1
1'
FIG. 7. Cultured RPE cells were endogenously labelled
with 32P and detergent extracted. Tahn was immunoprecipitated from the detergent soluble extract. Lane 1 shows the Coomassieblue-stained gel of the innnunoprecipitate and 1’ is the corresponding autoradiogram.
FIG. 6. Frozen sections of chick RPE cells cultured on Millicell-HAT” inserts. Sections were stained with antibodies to fibronectin (A), integrin (B), talin (C), and rhophalloidin (D). The arrows indicate the basal surface of the RPE cells. Bar = 10 ,um.
surfaces of the RPEcells. This occurred despite the fact that these cells produce all of the components of the basement membrane (Turksen et al., 1985). We also saw a polarized accumulation of fibronectin and laminin at the ventral surface of the cultured RPE cells. In chick intestinal epithelial cells (Chen et al., 1985) integrin is localized along the basal membrane of the cells. When these cells are cultured, integrin tlrst shows a diiuse distribution over both the dorsal and ventral surfaces of the cells. However, after 1 hr in culture integrin becomes polarized along the ventral surface of the cells. The differences in the observed results between the intestinal epithelium and the RPE may be attributable to the fact that in vivo, the apical surface of the RPE cells does not face an open lumen but is in intimate association with the neural retina.
The neural retina may play a significant role in the differentiation and polarization of the RPE cells. The importance of the close apposition of the RPE and the neural retina is shown by the studies of Anderson and coworkers (1986), who found that after experimental detachment of the neural retina in several mammalian species,there is a loss of apical processesand the RPE cells become proliferative. Talin phosphorylation has been reported from both in vivo and in vitro studies. In vitro, talin has been shown to be a substrate for protein kinase C (Litchfield and Ball, 1986). Talin phosphorylation has been reported in transformed cells (Pasquale, Maher and Singer, 1986). It has also been shown that treatment of BSC-1cells with the tumor promotor TPA caused a loss of adhesion plaques and an increase in phosphorylation of talin (Turner, Pavalko and Burridge, 1989). While one could speculate that phosphorylation of talin changes its affinity for integrin or vinculin there is not yet data to support this conclusion. In the present study we have shown that talin is phosphorylated in cultured RPEcells but not in the RPE cells freshly isolated from 18-day-old chick embryos. The phosphorylation of talin in vitro may be related to the changes in protein distribution that we see between cultured and freshly isolated RPE cells. Alternatively, the change in phosphorylation could occur when the 18-day-old RPE cells are removed from the basement membrane. In conclusion we have shown that a complex of proteins known to be involved in cell-substratum adhesion, integrin and talin, are found polarized to the basal region of the RPE cells. This suggests that integrin and talin play a role in the maintenance of the attachment of the RPE cells to the basement membrane but not to the neural retina.
N. J. PHILP
198
Acknowledgments We thank Dr Vivianne Nachmias for helpful discussions and for providing the laboratory space in which this work was done, and Rajasree Golla for assistance in antibody production and for providing whole platelet lysate samples. This work was supported by the National Institutes of Health grant EY05508.
References Anderson, D. H., Guerin, C. J.. Erickson, P. A., Stern. W. H. and Fisher, S. K. (1986). Morphological recovery in the reattached retina. Invest. Ophthalmol. Vis. Sci. 27, 168-83. Anderson, D. H.. Guerin, C. J., Pfeffer, B. A. and Matsumoto, B. (1989). Representatives from two integrin families are expressed simultaneously by monkey retinal pigmented epithefial (RPE) cells. Invest. Ophthalmol. Vis. Sci. 30, 414. Avery, R. L. and Glaser, B. M. (1986). Inhibition of retinal pigment epithelial cell attachment and migration by a synthetic peptide derived from the cell binding domain of fibronectin. Arch. Ophthalmol. 104. 1220-2. Beckerle, M. C.. O’Halloran, T. and Burridge, K. (1986). Demonstration of a relationship between talin and ~235, a major substrate of the calcium-dependent protease in platelets. 1. Cell. Biochem. 30, 259-70. Bronner-Fraser. M. (1985). Alteration in neural crest migration by a monoclonal antibody that affects cell adhesion. J. Cell Biol. 101, 610-7. Burridge, K. and Connell, L. (1983a). A new protein of adhesion plaques and ruffling membranes. 1. Cell BioJ. 97, 359-67. Burridge. K. and Connell, L. (1983b). Talin: A cytoskefetal component concentrated in adhesion plaques and other sites of actin-membrane interaction. Cell Motil. 3, 405-17. Burrldge, K., Fath, K., Kelly, T., Nuckolls, G. and Turner, C. ( 1988). Focal adhesions: Transmembrane junctions between the extracellular matrix and the cytoskeleton. Annu. Rev. Cell Biol. 4, 487-525.
Burrldge. K. and Mangeat, P. (1984). An interaction between vlnculin and talin. Nature 308, 744-6. Chen. W. T., Greve, J. M., Gottlieb, D. I. and Singer, S. J. (1985). Immunocytochemical localization of 140 kD cell adhesion molecules in cultured chicken fibroblasts, and in chicken smooth muscle and intestinal epithellal tissues. 1, Histochem.Cytochem. 33, 576-86. Drenckhahn, D., Beckerle, M., Burrldge, K. and Otto, J. (1988). Identification and subcellular localization of talin in various cell types and tissues by means of [125I] vinculin overlay, immunoblotting and immunocytochemistry. Euro. J. Cell Biol. 46, 513-22. Duband, J.-L., Rocher. S., Chen. W.-T, Yamada, K. M. and Thiery, J. P. (1986). Cell adhesion and migration in the early vertebrate embryo: location and possible role of the putative fibronectin receptor complex. J. Cell Biol. 102, 160-78.
Geiger. B.. Volk, T., Volberg, T. and Bendorl, R. (1987). Molecular interactions in adherens-type contacts. 1. Ceil Sci. 8 (Suppl.), 251-72. Hewitt, A. T. and Newsome, D. A. (1988). Altered proteoglycans in cultured human retmitis pigmentosa retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci. 29, 720-6.
ET AL.
Hock, R. S., Sanger, J. M. and Sanger, J. W. (1989). Talin dynamics in living microinjected nonmuscle cells. Cell Motil. [email protected], 271-87. Horwitz. A., Duggan, K., Buck, C., Beckerle,M. and Burridge. K. (1986). Interaction of plasma membrane fibronectin receptor with talin-A transmembrane linkage. Nature 320, 531-3. Horwitz, A., Duggan, K., Greggs,R., Decker, C. and Buck, C. (1985). The cell substrate attachment (CSAT) antigen has properties of a receptor for laminin and fibronectin. 1. Cell Biol. 101, 213444.
Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-S. Langer, B. G., Gonnella, P. A. and Nachmias, V. T. (1984). Alpha-actinin and vincufin in normal and thrombasthenic platelets. BZood63, 606-14. Litchfleld, D. W. and Ball, E. H. (1986). Phosphorylation of the cytoskeletal protein talin by protein kinase C. Biochem.Biophys. Res. Commun. 134, 1276-83. Nuckolls, G., Turner, C. and Burrldge, K. (1989). Functional studies of the domains of talin. J. Cell Biol. 109, 266a. O’Halloran, T. and Burridge, K. (1986). Purification of a 190 kDa protein from smooth muscle: Relationship to talin. Biochim. Biophys. Actu 869, 33749. O’Halloran, T., Beckerle, M. C. and Burridge, K. (1985). Identification of talin as a major cytoplasmic protein implicated in platelet activation. Nature 317, 449-51. O’Hafloran, T., Molony, L. and Burridge, K. (1986). Purification and assay of vinculin. metavlnculin and talin. Methods Enzymol. 134, 69-77. Opas,M . (1989). Expression of the differentiated phenotype by epitheliaf cells in vitro is regulated by both biochemistry and mechanics of the substratum. Dev. Biol. 131, 281-93.
Otto, J. J. (1983). Detection of vincufin-binding proteins with an 12 5-I vinculin gel overlay technique. 1. CeNBiol. 97, 1283-7. Pasquale, E. B., Maher, P. A. and Singer. S. J. (1986). Talin is phosphorylated on tyrosine in chicken embryo fibroblasts transformed by Rous Sarcoma Virus. Proc. N&f. Acad. Sci. U.S.A. 83, 5507-11. Philp, N. and Nachmlas, V. (1985). Components of the cytoskeleton in the retinal pigment epithellum of the chick. J. Cell Biol. 101, 358-62. Philp, N. and Nachmias, V. (1987). Polarized distribution of integrin and fibronectin in the retinal pigment epithelium. Invest. Ophthalmol Vis. Sci. 28, 1275-80. Philp. N. J. and Yoon, M.-Y. (1990). The role of Integrins in the attachment of RPE cells to basement membrane components. Invest. Ophthalmol. Vis. Sci. 31, 372. Steinberg, R. H. (1986). Research update: Report from a workshop on Cell Biology of retinal detachment. Ex~. Eye Res. 43, 695-706.
Turksen, K., Aubin, J. E., Sodek,J. and Kalnins, V. I. (198 5). Localization of laminin. Type IV collagen, fibronectin and heparan sulfate proteoglycan in the chick retinal pigment epithelium basement membrane during embryonic development. I. Histochem. Cytochem. 33. 665-71. Turksen, K., Opas,M. and Kalnins, V. I. (198 7). Preliminary characterization of cell surface-extracellular matrix linkage complexes in cultured retinal pigmented eplthefial cells. Exp. Cell Res. 147, 259-64. Turner, C. E., Pavalko. E. M. and Burridge, K. (1989). Role of talin phosphorylation in cytoskeletal disruption of BSC-I cells caused by TPA. J. Cell Biol. 107, 255a.
Identification
and Localization of Talin in Chick Epithelial Cells NANCY
Department
of Anatomy, (Received
J. PHILP’,
University 79 September
MI-YUNG
of Pennsylvania
YOON
AND
RICKS.
School of Medicine,
Retinal
Pigment
HOCK
Philadelphia,
PA 19104, U.S.A.
1989 and accepted in revised form 27 December
1989)
Retinal pigmented epithelial cells are adherent at their basal surface to Bruch’s membrane and at their apical surface to the neural retina. We examined the expression and distribution of two proteins that are found in regions of cell-matrix interaction, talin and integrin. Talin is a 23 S&Da cytoplasmic protein that has been localized to regions of cell-substrate adhesion. It binds to both integrin, a transmembrane glycoprotein complex, and to vinculin, a cytoskeletal protein. In the present study, we produced a polyclonal antibody to chicken gizzard talin. Using this antibody we showed by western blot analysis that talin is expressedby RPE cells and is found in the triton-soluble fraction. Talin was shown to co-localize with integrin and vinculin in the basal region of chick RPEcells isolated from 18-day-old chick embryos. Neither talin nor integrin was found in the apical processesor in the zonula adherens. Antibodies to vinculin showed staining both in the apical and basal regions of the RPEcells. The localization of integrin. talin and vinculin along the basal membrane suggeststhat this complex is important in the attachment of the RPE cells to the basement membrane. The distribution of integrin and talin was examined in primary cultures of RPEcells grown on permeable filters. In these cells, a polarized distribution of integrin and talin was not observed. This may suggest that the neural retina may be important for maintaining the differentiated state of the RPE cells. Key words: talin ; retinal pigment epithelium : cell adhesion : integrin ; vinculin. and identified by immunolocallzation
1. Introduction Like most epithelia, RPEcells are both structurally and functionally polarized. The basal surface of the RPE cells rests on a continuous basal lamina and the apical surface elaborates microvillous processesthat are in intimate association with the photoreceptor cells. The apical and basal-lateral membrane domains of the RPE cells are separated by junctional complexes that form a continuous ring around the apex of the cells. While most epithelial cells are in contact with one matrix compartment, the RPE cells are in contact with two matrix compartments. The interphotoreceptor matrix fills the spacebetween the apical processesof the RPE and the outerlimiting membrane of the neural retina. The interphotoreceptor matrix is composed of sialoglycoconjugates, chondroitin sulfate A and C, hyaluronate and heparan sulfate (Hewitt and Newsome, 1988). The basal surface of the RPE rests on a continuous basement membrane that is composed primarily of type IV collagen, heparan sulfate, laminin and fibronectin (Turksen et al., 1985). The attachment of the RPE cells to both the basement membrane and
the neural retina is critical for maintaining the nonproliferative, differentiated state of these cells (for review, Steinberg, 1986). However, little is known about the adhesion molecules expressed by the RPE cells or the linkage of adhesion molecules to the cytoskeleton. Talin, a large cytoplasmic protein (235 kDa), was originally purified from chicken gizzard smooth muscle
of talin has since been studied by immunolocalllation
in a variety of cultured cells where it has been seen to be concentrated in adhesion plaques (for review see Burridge et al., 1988). Within migratory cells, talon,
like integrin, has a diffuse distribution, while in the intestinal epithelium it is found along the basal surface of the cells adjacent
to the substratum
(Drenckhahn et al., 1988). In cultured RPE cells, integrin, talin and vinculin
have been identified to be
components of focal contacts (Turksen, Opas and Kalnins, 1987 ; Opas, 1989). The distribution of talin in RPE cells in vivo has not yet been examined. In the present study, we have produced a polyclonal antibody to chicken gizzard talm. With this antibody we showed by western blot analysis that talin is expressed by RPE cells. Using indirect immunofluorescence microscopy we examined the distribution of talin, integrin and vinculin cultured RPE cells.
in freshly isolated and
2. Materials and Methods Embryos
White Leghorn chick embryos used in these studies were procured from a local supplier and incubated in a forced draft incubator at 38°C.
* For correspondence. 00144835/90/080191+08
as a component
of adhesion plaques (Burridge and Connell, 198 3a, b). It has been shown in vitro that talin binds both to vinculin (Otto, 1983; Burridge and Mangeat, 1984) and to integrin (Horwitz et al., 1986). The distribution
$03.00/O
0 1990 Academic Press Limited
192
Cell Culture Retinal pigment epithelial cells were isolated from 9day-old chick (E9) embryos in PBS/EDTA buffer, as previously described (Philp and Nachmias, 198 5). The cells were grown on millipore filter inserts (MillicellHATM, 12.5 mm) in a medium consisting of DMEM containing 5 % (v/v) FBS and penicillin (5U ml-l) and streptomycin (5 ,ugml-l) (GIBCO). Approximately 5.0 x lo5 cells were plated onto each filter unit. Cells were grown to confluence then fixed and sectioned for the immunofluorescence studies. In some cases cells were grown in serum-free medium PC-l (Ventrex, Portland, ME). The medium was tested by immunoblot analysis and was shown to be free of fibronectin. Antibodies Polyclonal antibodies to the /3 l-subunit of the avian integrin complex were a gift of Dr Clayton Buck (The Wistar Institute, Philadelphia, PA). Antibodies to vinculin were produced against chicken gizzard vinculin and were a gift of Dr Susan Craig (Johns Hopkins University, Baltimore, MD) and antibodies to human plasma fibronectin were purchased from BRL (Bethesda Research Laboratories, Bethesda, MD). Antibodies to chicken gizzard talin were raised in New Zealand rabbits in our laboratory. The talin was purified from frozen chicken gizzards (Pel-Freeze) according to published procedures (O’Halloran, Molony and Burridge, 1986; Hock, Sanger and Sanger, 1989). Highly purified talin was coupled to cyanogen-activated Sepharose 4B (Sigma) for use in affinity purification of the antibody. SDS-PAGEand Western Blots RPE cells were isolated from 17- and l&day-old embryonic chick eyes, as previously described (Philp and Nachmias, 1985). Proteins from whole RPE cells (about 50 ,ug or one quarter of an eye per lane) and triton-insoluble cytoskeletons (75 rug or two eyes per lane) were electrophoresced on microslab gels using the buffer system of Laemmli (1970). Whole platelet lysate was used as standards on the gels since they contained large amounts of talin (23 5 kDa protein) (O’Halloran, Beckerle and Burridge, 1985) and vinculin (130 kDa) (Langer, Gonnelle and Nachmias, 1984) which cross-reacted with antibodies against chicken gizzard talin and vinculin. Gelswere fixed and stained in 0.25 % CoomassieBrilliant Blue R in 50 % methanol, 10% acetic acid, and were destained by diffusion in 5% methanol and 10% acetic acid. For Western blots the proteins separated on SDS-PAGE were electrophoretically transferred to nitrocellulose paper (Biorad Laboratories, Richmond, CA) or Immobilon-P transfer membrane (Millipore Corporation, Bedford, MA) at 100 V for i hr using a MiniTransblot apparatus (Biorad Laboratories, Richmond,
N. J. PHILP
ET AL.
CA). The blots were incubated in a buffer containing 5% BSA, 20 mM Tris, 500 mM NaCl, pH 7.5 for 1 hr to block non-specific binding sites. The Immobilon membranes were hydrated with 100% methanol for 5 set, washed with distilled water, then incubated in blocking buffer. The blots were incubated in the primary anti-sera for at least 1 hr. Optimal dilutions for the primary antisera were experimentally determined to be 1: 200 for affinity purified anti-talin (1.25 mg ml-‘), 1: 100 for anti-vinculin and antiintegrin (14 pugml-‘). Antisera were diluted in 1% BSA, 0.05 % Tween-20, 20 mM Trisma Base, 500 mM NaCl, pH 7.5. The blots were washed three times with this buffer and incubated for one hour with alkaline phosphatase conjugated goat anti-rabbit IgG at a dilution of 1: 3000 (Biorad Laboratories), followed by two more washes before developing with 0.3 mg ml-’ NBT (p-nitro blue tetrazolium chloride) and O-15 mg ml-’ BCIP (S-bromo-4-chloro-3-indolyl phosphate p-toluidine salt) in a buffer of 0.1 M NaHCO,, 1.0 mM MgCI,, pH 9.8. Immunoprecipitation RPE cells were isolated from six eyes (E18) as described above and solubilized in cytoskeletal extraction buffer (CEB) containg 100 mM KCI, 1 mM MgCl,, 1 mM EGTA, 10 mM imidazole pH 7.2, 1% Triton X-100, 1 mM benzamidine and 0.05 mg ml-’ leupeptin. After 15 min on ice, the extract was centrifuged at 10 000 g for 15 min and the supernatant collected. Affinity purified anti-talin (S-10 pg) was incubated with 100 ~1 of a 10 % solution of SDS stripped IgGsorb (The Enzyme Center, Malden, MA) which had been for 1 hr in the cold, then washed four times in cytoskeleton extraction buffer. The anti-talin IgGSorb was incubated in the cold for another hour with the RPEcell supernatant. After 1 hr, the IgGsorb was spun for 2 min at 10 000 g and the supernatant was discarded. The pellet was washed four times in CEB,boiled in Laemmli sample buffer, and the sample run on 7.5 Y0or 10 y0 SDSgels. Phosphorylation Cultured chick RPE and freshly isolated El8 RPE cells were labelled with azP1 mCi ml-’ for 1 hr in MEM without phosphates (Flow Laboratories, Mclean, VA). At the end of the incubation period, the cells were washed with cold PBS three times and were then extracted into 0.5 ml of a buffer containing 1% Nonidet P40, 100 mM sodium pyrophosphate, 250m~ NaCl, 50m~ NaF, 5 mM EGTA, 1 mM benzamidine, 10 ,ugml-’ leupeptin, 15 mM pmercaptoethanol and 20 mM Tris HCl, pH 7.9. The extract was spun in an eppendorf centrifuge for 10 min and the supernatant taken and incubated for 1 hr with anti-talin bound to IgGsorb, as described above.
TALIN
193
IN RPE CELLS
Immunocytochemistry Posterior eye cups from El8 and sheets of RPE cells isolated from El 8 were ilxed for 30 min in 2 % paraformaldehyde containing 1 mu MgCl, and 3% sucrose in PBS. The tissues were washed with three changes of PBSwith 3 % sucrose, then put in PBSwith 30% sucrose overnight in the cold. Tissues were embeddedin Tissue Tek II 0.C.T Compound (American Scientific Products, Edison, NJ) and frozen in liquid nitrogen. Six to 12 pm cryostat sections were mounted on gelatin coated coverslips. The sections were flxed for 1 min in 2% paraformaldehyde, 3% sucrose and 1 mu MgCl, in PBS, and non-specific binding sites on the tissue were blocked for 30 min in 1% goat serum in PBS.All washes and antibody dilutions were made in 1 Y0BSA and 0.05 Y0Tween-20 in PBS.The sections were incubated with the primary anti-sera for 1 hr. Talin was used at 10 mg ml-‘, fibronectin was diluted 1: 100, integrin 1: 100, laminin 1: 100, and afEnity purified vinculin 1: 15. After thorough washing, rhodamine conjugated goat anti-rabbit F(ab’), fragment (Cooper Biomedical, Westchester, PA) was applied for 45 min. Sections were washed, and the coverslips mounted with gelvatol (Aii Products and Chemicals, Inc, Allentown, PA). Slides were examined with phase or epiiluoresecence on a Zeiss Research microscope equipped with a 63 x phase objective with an N.A. of 1.4. Photographs were taken with Kodak Tri-X film. The negatives were developed for about 7 min in HC-110 developer dilution B. Fresh tissue samples were prepared each time for the cryostat sectioning. Two eye cups or sheetsof RPEcells isolated from about four eyeswere frozen in each block and the tissues were stained immediately after sectioning. The immunocytochemical studies were repeated a minimum of three times with similar results.
1
2
FIG. 1. Immunoblot analysis of whole RPE cells (1) and
purified chicken gizzardtalin (2) with affiity purified talin antibody. The antibody cross-reactswith a single band in whole RFE cells that migrateswith the samemobility as chicken gizzard talin with a molecular weight of about 235 kDa (1). The antibody also cross-reacts with the 190 kDa breakdown product of talin (2).
FIG. 2. Immunoblot analysis of whole chick RPEcells and cytoskeletons with antibodies to the p-subunit of integrin (120 kDa) (A), talin (235 kDa) (B), and vinculin (117 kDa) (C). Lane 1, whole platelet lysate, lane 2, whole RPE cells and lane 3, RPE cytoskeletons.
3. Results
Polyclonal antibodies to chicken gizzard talin were produced in rabbit and aiiinity purified on a talinsepharosecolumn. The antisera were characterized on western blots and were shown to cross-react with both purified chicken gizzard talin and with a protein in RPE cells that had a similar electrophoretic mobility (Fig. 1). The antibody cross-reacted with both the intact protein (23 5 kDa) as well as with the 190 kDa degradation product (O’Halloran and Burridge, 1986). Additionally, as shown in Fig. 2(B), lane 1, the antibody cross-reactedwith the 2 3 S-kDa protein from human platelets which has been identified as talm (Beckerle, O’Halloran and Burridge, 1986). Talin from chicken gizzards, chicken RPE cells and human platelets all ran with the same relative mobility in our gel system. To determine whether or not talin remained lied to the cytoskeleton after triton extraction, whole RPE cells and cytoskeletal proteins form l&day-old 14
em-
bryos were separated on polyacrylamide gels and transferred to nitrocellulose or ImmobilonTM paper. The blots were probed with antibodies to integrin, talin and vinculin. Lanes 2 contain RPEproteins from one third of an eye per lane, while lanes 3 are RPE cytoskeletons from two eyes per lane. Despite the fact that more cytoskeletal than whole cell proteins were loaded onto the gel little talin or integrin remained bound to the triton-insoluble cytoskeleton of RPE cells [Figs 2(A) and (B), lanes 31. Vinculin, which is found both in areas of cell-cell and cell-substratum was present in the cytoskeletons of RPE cells [Fig. 2(c)]. Some minor protein bands were detected with the vinculin antibody in whole platelet lysate, whole RPE cells and in RPE cytoskeletons. These may represent isoforms and breakdown products of vinculin or it could be non-specific binding. The afBnity-purified talin antibody was used to immunoprecipitate talin from tt-iton extracts of whole EER 51
N. J. PHILP
194
A
talinm
--
,.--
vine
a
1 2 1' 2' FIG. 3. A, Immunoprecipitationof talin from triton soluble extract of whole RPEcellsrun on a 10% polyacylamidegel. Lane 1, Coomassieblue stainedgel of whole platelet lysate, lane 2, whole RPE cells, and lane 3, talin immunoprecipitatedfrom triton solubleextract of whole RPEcells. The lower molecular weight bands seenin lane 3 are IgG heavy and light chain subunitsof the anti-talin. B, Western blot analysisof proteinsfrom wholeplateletlysate(1, 1’) and immunoprecipitated talin (2, 2’) run on 7.5y0 polyacrylamide gels and transferred to ImmobilonTMpaper. The westernblotswereprobedwith antibodiesto talin (1,2) and vinculin (l’, 2’). The talin antibody cross-reactedwith the 235-kDa protein in platelets and with the immunoprecipitatedtalin. Antibody to vinculin cross-reactedwith vinculin in plateletsbut was negativein lane 2’, showing that vinculin doesnot coprecipitatewith talin. RPE cells. The immunoprecipitated protein migrated with the same electrophoretic mobility as human platelet talin [Fig. 3(A)]. When samples from human platelet and chick RPEcells were blotted and incubated with the antibody to talin, both the 235-kDa protein in platelets and the protein immunoprecipitated from RPE cells cross-reacted with the talin antibody. The same blot was probed with antibody to vinculin to determine whether vinculin coprecipitated with talin. It was shown on at least five separate occasions that only trace amounts of vinculin coprecipitated with talin [Fig. 3(B)]. Integrin did not coprecipitate with talin (data not shown). Indirect immunofluorescence was used to localize talin, fibronectin, integrin and vinculin in El8 chick
ET AL.
RPE cells (Fig. 4). Antibodies to talin showed bright staining along the basal surface of the RPE cells with no significant staining along the apical region of the RPE cells. No staining was seen with the anti-talin in the area of the junctional complexes or in the apical processes.As previously shown (Philp and Nachmias, 198 7) integrin is localized along the basal surface of the RPEcells where the cells attach to the substratum. To insure that the staining observed in the intact eye cup was from RPE cells and not from the endothelial cells, the cells were isolated from Bruch’s membrane after incubation in PBS containing EDTA. Frozen sections of the isolated sheetsof cells were stained with antibodies to talin and integrin (Fig. 5). As previously shown in Fig. 4, talin and integrin colocalized along the basal membrane of the RPE cells. Vinculin was localized at both the basal and apical membranes and at the intercellular junctions as previously shown (Philp and Nachmias, 1985; Turksen et al., 1987). Laminin and fibronectin did not remain bound to the RPEcells after they were dissociatedfrom the basement membrane in EDTA [Figs. 5(F) and (G)]. The distribution of talin and integrin was examined in cultured RPE cells that were grown on permeable supports. Frozen sections of these cells showed that in contrast to the polarized distribution of integrin and talin seen in vivo, in cultured RPEcells, these proteins were not polarized. Integrin was localized on both the apical and basal-lateral membranes [Fig. 6(B)]. Talin staining was seen to be diffuse throughout the cells [Fig. 6(C)]. Despite the fact that these cultures were grown in media containing serum, fibronectin staining was only observed along the basal surface of the RPE cells [Fig. 6(A)]. This suggests that the fibronectin in the serum is not tightly bound to integrin in the apical membrane of the RPE cells. Culturing the cells in serum free media in the apical compartment and media containing serum in the basal compartment did not alter the staining pattern of talin, integrin or fibronectin (data not shown). Even when we cultured the cells on Bruch’s membrane we did not see a polarized distribution of integrin. When whole mounts were examined, we found that there was a concentration of integrin, talin and vinculin in adhesion plaques in agreement with the work of Opas (1989). It has been reported that talin is phosphorylated both in vivo and in vitro (Litchfield and Ball, 1986). We labeled culture RPE cells with 32Pand immunoprecipitated talin from the triton-soluble extract to determine its state of phosphorylation. Gel electrophoresis and autoradiography showed that the immunoprecipitated talin was phosphorylated in cultured RPE cells (Fig. 7). Additionally, other phosphorylated proteins coprecipitated with talin. We do not know the identity of these bands or whether their coprecipitation with talin is specific or non-specific. Using the same labeling and immunoprecipitation procedures on freshly isolated RPE cells, we were unable to demonstrate that talin was phosphorylated (data not shown).
TALIN
IN RPE CELLS
FIG. 4. Frozen sections of lb-day-old chick retina. Phase micrograph (A), and similar sections stained with antibodies to fibronectin (B). integrin (C), talin (D), and rhophalloidin (E), and secondaryantibody (F). Arrows indicate the basalregion of
the RPEceils.Bar = 10 pm.
4. Discussion In the present study we examined the expression and distribution of two proteins involved in cellsubstratum adhesion, integrin and talin. Previously we reported that integrin was localized only on the basal surface of RPEcells (Philp and Nachmias, 198 7) though subsequent reports suggested that it was on the apical surface of the RPEcells as well (Anderson et al., 1989). This raised the possibility that integrin could be involved both in cell-cell and cell-substratum adhesion. To help address this issue, we decided to
study more closely another protein that binds to integrin and links it to the cytoskeleton, talin. Therefore, we produced a polyclonal antibody to talin to use in these studies. Our results showed that in RPE cells isolated from 18-day-old chick embryos, there was discrete localization of integrin and talin along the basal surface of the RPE cells. No significant staining was seen in the apical region of the RPE cells or in the area of the junctional complexes. Two different polyclonal antibodies to talin, one produced by K. Burridge and one produced in our laboratory,
yielded similar staining 14-2
196
N. J. PHILP
ET AL
FIG. 5. Frozensectionsof isolated sheetsof 18 d RPEcells. Phasemicrograph (A), and similar sectionsstained with rhophalloidin (B),and antibodiesto vinculin (C),talin (D),integrin (E).laminin (F),fibronectin(G),andsecondaryantibodyonly (H). The small arrows indicate the basalregion of the RFEcells.Bar = 10 pm. patterns on tissue. Therefore, it is unlikely that we are not detecting talin and or integrin in other regions of the cell. Vinculin, a cytoskeletal protein that is found in all types of adhesion junctions (for review seeGeiger et al., 198 7), was localized by immunoflurescent staining along the basal surface of the RPE and in the apical region of the cell. While talin binds to vinculm, they codistribute only in regions of cell-substratum adhesion in RPE cells. It was recently reported that when fluorescently-labeled talm was microinjected into Madin Darby bovine kidney cells it was localiied only in focal adhesions. However when the 200~kDa domain was injected into these cells it was incorporated into the zonula adherens as well as into the focal contacts (Nuckolls, Turner and Burridge, 1989). This suggeststhat the 47-kDa fragment regulates the specific distribution of talin. We have preliminary evidence to support a direct role for integrin in the attachment of RPE cells to the basement membrane. The monoclonal antibody to the integrin receptor, CSAT, inhibits the attachment of freshly isolated RPE cells to both laminin and fibronectm (Yoon and Philp, unpubl. res.). This would be consistent with reports which show that avian integrin binds to both laminin and fibronectin (Horwitz et al., 1985; Bronner-Fraser, 1985). The CSAT antibody did not disrupt the attachment of RPEcells to Bruch’s membrane when explants of RPE cells were
incubated in antibody containing media. This suggests that integrin is not the only protein in the RPE cells regulating cell-substratum adhesion. The distribution of integrin and talin in normal and pathological states such as retinal detachment is of interest since it has been shown that integrin can modulate both cell adhesion and cell migration (Duband et al., 1986). Under normal conditions integrin and talin may stabilize the attachment of the RPE cells to the basement membrane while in pathological states these same proteins may mediate the migration of these cells. Studies have shown that RPEcells can migrate along a fibronectin gradient and that this migration is inhibited by the tetrapeptide RGDS(Avery and Glaser, 1986). It would be of interest to determine whether or not there is a redistribution of talin and integrin after retinal detachments. We examined the distribution of talin and integrin in RPEcells grown in tissue culture. The cultured RPE cells did not exhibit a polarized distribution of integrin or talin, even after several weeks in culture. We thought that components of the serum such as fibronectin could have an effect on the distribution of integrin and talin. To test this hypothesis we grew cells on millipore inserts with serum free media in the apical compartment and serum containing media in the basal compartment. Even under these conditions we found integrin on both the apical and basal
TALIN
197
IN RPE CELLS
talint
---
1
1'
FIG. 7. Cultured RPE cells were endogenously labelled
with 32P and detergent extracted. Tahn was immunoprecipitated from the detergent soluble extract. Lane 1 shows the Coomassieblue-stained gel of the innnunoprecipitate and 1’ is the corresponding autoradiogram.
FIG. 6. Frozen sections of chick RPE cells cultured on Millicell-HAT” inserts. Sections were stained with antibodies to fibronectin (A), integrin (B), talin (C), and rhophalloidin (D). The arrows indicate the basal surface of the RPE cells. Bar = 10 ,um.
surfaces of the RPEcells. This occurred despite the fact that these cells produce all of the components of the basement membrane (Turksen et al., 1985). We also saw a polarized accumulation of fibronectin and laminin at the ventral surface of the cultured RPE cells. In chick intestinal epithelial cells (Chen et al., 1985) integrin is localized along the basal membrane of the cells. When these cells are cultured, integrin tlrst shows a diiuse distribution over both the dorsal and ventral surfaces of the cells. However, after 1 hr in culture integrin becomes polarized along the ventral surface of the cells. The differences in the observed results between the intestinal epithelium and the RPE may be attributable to the fact that in vivo, the apical surface of the RPE cells does not face an open lumen but is in intimate association with the neural retina.
The neural retina may play a significant role in the differentiation and polarization of the RPE cells. The importance of the close apposition of the RPE and the neural retina is shown by the studies of Anderson and coworkers (1986), who found that after experimental detachment of the neural retina in several mammalian species,there is a loss of apical processesand the RPE cells become proliferative. Talin phosphorylation has been reported from both in vivo and in vitro studies. In vitro, talin has been shown to be a substrate for protein kinase C (Litchfield and Ball, 1986). Talin phosphorylation has been reported in transformed cells (Pasquale, Maher and Singer, 1986). It has also been shown that treatment of BSC-1cells with the tumor promotor TPA caused a loss of adhesion plaques and an increase in phosphorylation of talin (Turner, Pavalko and Burridge, 1989). While one could speculate that phosphorylation of talin changes its affinity for integrin or vinculin there is not yet data to support this conclusion. In the present study we have shown that talin is phosphorylated in cultured RPEcells but not in the RPE cells freshly isolated from 18-day-old chick embryos. The phosphorylation of talin in vitro may be related to the changes in protein distribution that we see between cultured and freshly isolated RPE cells. Alternatively, the change in phosphorylation could occur when the 18-day-old RPE cells are removed from the basement membrane. In conclusion we have shown that a complex of proteins known to be involved in cell-substratum adhesion, integrin and talin, are found polarized to the basal region of the RPE cells. This suggests that integrin and talin play a role in the maintenance of the attachment of the RPE cells to the basement membrane but not to the neural retina.
N. J. PHILP
198
Acknowledgments We thank Dr Vivianne Nachmias for helpful discussions and for providing the laboratory space in which this work was done, and Rajasree Golla for assistance in antibody production and for providing whole platelet lysate samples. This work was supported by the National Institutes of Health grant EY05508.
References Anderson, D. H., Guerin, C. J.. Erickson, P. A., Stern. W. H. and Fisher, S. K. (1986). Morphological recovery in the reattached retina. Invest. Ophthalmol. Vis. Sci. 27, 168-83. Anderson, D. H.. Guerin, C. J., Pfeffer, B. A. and Matsumoto, B. (1989). Representatives from two integrin families are expressed simultaneously by monkey retinal pigmented epithefial (RPE) cells. Invest. Ophthalmol. Vis. Sci. 30, 414. Avery, R. L. and Glaser, B. M. (1986). Inhibition of retinal pigment epithelial cell attachment and migration by a synthetic peptide derived from the cell binding domain of fibronectin. Arch. Ophthalmol. 104. 1220-2. Beckerle, M. C.. O’Halloran, T. and Burridge, K. (1986). Demonstration of a relationship between talin and ~235, a major substrate of the calcium-dependent protease in platelets. 1. Cell. Biochem. 30, 259-70. Bronner-Fraser. M. (1985). Alteration in neural crest migration by a monoclonal antibody that affects cell adhesion. J. Cell Biol. 101, 610-7. Burridge, K. and Connell, L. (1983a). A new protein of adhesion plaques and ruffling membranes. 1. Cell BioJ. 97, 359-67. Burridge. K. and Connell, L. (1983b). Talin: A cytoskefetal component concentrated in adhesion plaques and other sites of actin-membrane interaction. Cell Motil. 3, 405-17. Burrldge, K., Fath, K., Kelly, T., Nuckolls, G. and Turner, C. ( 1988). Focal adhesions: Transmembrane junctions between the extracellular matrix and the cytoskeleton. Annu. Rev. Cell Biol. 4, 487-525.
Burrldge. K. and Mangeat, P. (1984). An interaction between vlnculin and talin. Nature 308, 744-6. Chen. W. T., Greve, J. M., Gottlieb, D. I. and Singer, S. J. (1985). Immunocytochemical localization of 140 kD cell adhesion molecules in cultured chicken fibroblasts, and in chicken smooth muscle and intestinal epithellal tissues. 1, Histochem.Cytochem. 33, 576-86. Drenckhahn, D., Beckerle, M., Burrldge, K. and Otto, J. (1988). Identification and subcellular localization of talin in various cell types and tissues by means of [125I] vinculin overlay, immunoblotting and immunocytochemistry. Euro. J. Cell Biol. 46, 513-22. Duband, J.-L., Rocher. S., Chen. W.-T, Yamada, K. M. and Thiery, J. P. (1986). Cell adhesion and migration in the early vertebrate embryo: location and possible role of the putative fibronectin receptor complex. J. Cell Biol. 102, 160-78.
Geiger. B.. Volk, T., Volberg, T. and Bendorl, R. (1987). Molecular interactions in adherens-type contacts. 1. Ceil Sci. 8 (Suppl.), 251-72. Hewitt, A. T. and Newsome, D. A. (1988). Altered proteoglycans in cultured human retmitis pigmentosa retinal pigment epithelium. Invest. Ophthalmol. Vis. Sci. 29, 720-6.
ET AL.
Hock, R. S., Sanger, J. M. and Sanger, J. W. (1989). Talin dynamics in living microinjected nonmuscle cells. Cell Motil. [email protected], 271-87. Horwitz. A., Duggan, K., Buck, C., Beckerle,M. and Burridge. K. (1986). Interaction of plasma membrane fibronectin receptor with talin-A transmembrane linkage. Nature 320, 531-3. Horwitz, A., Duggan, K., Greggs,R., Decker, C. and Buck, C. (1985). The cell substrate attachment (CSAT) antigen has properties of a receptor for laminin and fibronectin. 1. Cell Biol. 101, 213444.
Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 680-S. Langer, B. G., Gonnella, P. A. and Nachmias, V. T. (1984). Alpha-actinin and vincufin in normal and thrombasthenic platelets. BZood63, 606-14. Litchfleld, D. W. and Ball, E. H. (1986). Phosphorylation of the cytoskeletal protein talin by protein kinase C. Biochem.Biophys. Res. Commun. 134, 1276-83. Nuckolls, G., Turner, C. and Burrldge, K. (1989). Functional studies of the domains of talin. J. Cell Biol. 109, 266a. O’Halloran, T. and Burridge, K. (1986). Purification of a 190 kDa protein from smooth muscle: Relationship to talin. Biochim. Biophys. Actu 869, 33749. O’Halloran, T., Beckerle, M. C. and Burridge, K. (1985). Identification of talin as a major cytoplasmic protein implicated in platelet activation. Nature 317, 449-51. O’Hafloran, T., Molony, L. and Burridge, K. (1986). Purification and assay of vinculin. metavlnculin and talin. Methods Enzymol. 134, 69-77. Opas,M . (1989). Expression of the differentiated phenotype by epitheliaf cells in vitro is regulated by both biochemistry and mechanics of the substratum. Dev. Biol. 131, 281-93.
Otto, J. J. (1983). Detection of vincufin-binding proteins with an 12 5-I vinculin gel overlay technique. 1. CeNBiol. 97, 1283-7. Pasquale, E. B., Maher, P. A. and Singer. S. J. (1986). Talin is phosphorylated on tyrosine in chicken embryo fibroblasts transformed by Rous Sarcoma Virus. Proc. N&f. Acad. Sci. U.S.A. 83, 5507-11. Philp, N. and Nachmlas, V. (1985). Components of the cytoskeleton in the retinal pigment epithellum of the chick. J. Cell Biol. 101, 358-62. Philp, N. and Nachmias, V. (1987). Polarized distribution of integrin and fibronectin in the retinal pigment epithelium. Invest. Ophthalmol Vis. Sci. 28, 1275-80. Philp. N. J. and Yoon, M.-Y. (1990). The role of Integrins in the attachment of RPE cells to basement membrane components. Invest. Ophthalmol. Vis. Sci. 31, 372. Steinberg, R. H. (1986). Research update: Report from a workshop on Cell Biology of retinal detachment. Ex~. Eye Res. 43, 695-706.
Turksen, K., Aubin, J. E., Sodek,J. and Kalnins, V. I. (198 5). Localization of laminin. Type IV collagen, fibronectin and heparan sulfate proteoglycan in the chick retinal pigment epithelium basement membrane during embryonic development. I. Histochem. Cytochem. 33. 665-71. Turksen, K., Opas,M. and Kalnins, V. I. (198 7). Preliminary characterization of cell surface-extracellular matrix linkage complexes in cultured retinal pigmented eplthefial cells. Exp. Cell Res. 147, 259-64. Turner, C. E., Pavalko. E. M. and Burridge, K. (1989). Role of talin phosphorylation in cytoskeletal disruption of BSC-I cells caused by TPA. J. Cell Biol. 107, 255a.
