A monoclonal antibody, CSAT, which inhibits the adhesion of chick cells to substrata coated with fibronectin, laminin and vitronectin, has been used to identify a cell surface receptor required for cell—substratum adhesion. This receptor, termed integrin, is found on the ventral surface of cells in close contact adhesion sites, at the periphery of adhesion plaques and beneath stress fibres. It is a heterodimer consisting of non-covalently linked alpha and beta subunits. Integrin binds directly to laminin, fibronectin and vitronectin with dissociation constants in the micromolar range. The binding of integrin to matrix molecules is sensitive to peptides carrying the cell-binding sequence Arg-Gly-Asp and requires heteromeric integrity. Integrin also binds directly to the cytoskeleton-associated protein talin. Thus, integrin has the properties of a transmembrane molecule capable of bringing extracellular matrix and cytoskeleton-associated molecules in proper juxtaposition to form adhesion structures. The integrin beta subunit is phosphorylated following Rous sarcoma virus transformation. Phosphorylation alters the ability of the receptor to bind extracellular matrix molecules as well as talin, suggesting a mechanism for the alteration of cellular adhesive and morphological properties following malignant transformation. A major phosphorylation site is on the cytoplasmic domain of the beta subunit. Synthetic peptides homologous with this region of integrin inhibit integrin-talin binding. The gene for the beta subunit of integrin has been sequenced. Its structure is consistent with the membrane-spanning properties of the receptor. Integrin is structurally and serologically related to adhesion receptors from mammalian tumour cells, fibroblasts, platelets and lymphocytes. It appears to be a member of a supergene family of receptors involved in cellular adhesive interactions. Antibody and peptide inhibition experiments have suggested a role for integrin and integrin-like molecules in cell migration, neurite extension, •neural differentiation, histogenesis and embryonic development in Drosophila. Thus, integrin appears representative of a set of evolutionarily conserved, biologically important adhesive molecules.

The interaction of cells with their extracellular matrix is basic to such processes as histogenesis, wound healing, metastasis, neuronal organization and angiogenesis, to name but a few. To study the process of cell—matrix adhesion, most investigators have turned to model in vitro systems focusing on structures such as adhesion plaques described by Abercrombie et al. (1971). At sites of cell-matrix contact, molecules of the extracellular matrix, the surface membrane and cytoskeleton are brought into highly organized juxtaposition. Over the past 15 years, considerable effort has gone into defining the molecular basis of this organization, with various groups focusing their research efforts on one of the three regions involved in adhesive events, i.e. the cell surface, the extracellular matrix or the cytoskeleton.

Our interest has been in the molecules within the surface membrane that serve to coordinate the organization of these adhesive structures. Our working hypothesis has been that a cell surface molecule (or molecules) exists that serves as a bridge spanning the surface membrane, interacting with both the extracellular matrix and elements of the cytoskeletal complex. To search for these molecules, we first developed polyclonal, and later monoclonal, antibodies that could perturb cell-matrix adhesion in a reversible, non-toxic manner and could then be used to identify the cell surface constituents required for adhesion (Wylie et al. 1979; Knudsen et al. 1981; Neff et al. 1982). The monoclonal antibody we have used is designated CSAT (Neff et al. 1982). A second monoclonal antibody, which interfere” with chick fibroblast adhesion, was independently isolated and characterized by Greve & Gottlieb (1982). This antibody, designated JG22, has properties that appear indistinguishable from those of CSAT. The cell surface complex with which these antibodies react has been called CSAT antigen, 140K complex and integrin. For the sake of clarity, the term integrin will be used here.

The effect of the monoclonal antibody CSAT on the adhesion of chick fibroblasts to fibronectin and laminin is both cell-type- and substratum-dependent. While the antibody interferes with the adhesion of chick tendon fibroblasts to both fibronectin and laminin, it prevents adhesion of chick cardiac fibroblasts to laminin only (Fig. 1), unless the cells are briefly trypsinized prior to plating on fibronectin, at which time the antibody will also prevent adhesion of the cardiac fibroblasts to fibronectin (Decker et al. 1984 ; Horwitz et al. 1985). The simplest interpretation of these results is that cells exhibit more than one adhesion mechanism. Chick tendon fibroblasts adhere to both fibronectin and laminin by a CSAT-sensitive mechanism, whereas cardiac fibroblasts adhere to laminin primarily by a CSAT-sensitive mechanism and to fibronectin by at least two different mechanisms, one of which is trypsin-sensitive. Thus, integrin is clearly not the only cell surface molecule involved in cell-matrix adhesion. While early cell adhesion and spreading appear to take place by a CSAT- sensitive mechanism and hence involve integrin, the establishment of mature adhesion plaques requires other molecules (Oesch & Birchmeier, 1982; Rogalski & Singer, 1985) including, in some instances (depending upon the cell type), proteoglycan and proteoglycan receptors (Lark et al. 1985; Izzard et al. 1986; Rapraeger et al. 1986; Singer et al. 1987).

Fig. 1.

Effect of CSAT monoclonal antibody on adhesion of chick cardiac fibroblasts to fibronectin and laminin. Chick cardiac fibroblasts were plated onto wells previously coated with fibronectin or laminin and allowed to spread for 2h. CSAT monoclonal antibody was then added (40μgml−1) and monolayers photographed 2h later. FN, fibronectin; LM, laminin.

Fig. 1.

Effect of CSAT monoclonal antibody on adhesion of chick cardiac fibroblasts to fibronectin and laminin. Chick cardiac fibroblasts were plated onto wells previously coated with fibronectin or laminin and allowed to spread for 2h. CSAT monoclonal antibody was then added (40μgml−1) and monolayers photographed 2h later. FN, fibronectin; LM, laminin.

The antigen to which the CSAT monoclonal antibody binds is located on the ventral surface of well-spread cells beneath actin-containing stress fibres, at the periphery of adhesion plaques, and in presumptive close contact-like structures (Fig. 2). In double immunofluorescence experiments, it colocalizes with fibronectin, actin, talin and vinculin (Damsky et al. 1985; Chenet al. V)%Sa,b, 1986b). Thus, its position in adherent cells is consistent with its serving as a coordinator of cell-matrix adhesion.

Fig. 2.

Distribution of integrin on the ventral surface of chick fibroblasts. Chick tendon fibroblasts were plated on coverslips coated with fibronectin. Integrin was localized by indirect immunofluorescence using a polyclonal anti-integrin serum as the primary antibody. A. Distribution of integrin around the periphery of adhesion plaques forming needle’s-eye-like patterns. B. Distribution of integrin in close contact-like regions of initially spreading fibroblasts, note lack of needle’s eye pattern.

Fig. 2.

Distribution of integrin on the ventral surface of chick fibroblasts. Chick tendon fibroblasts were plated on coverslips coated with fibronectin. Integrin was localized by indirect immunofluorescence using a polyclonal anti-integrin serum as the primary antibody. A. Distribution of integrin around the periphery of adhesion plaques forming needle’s-eye-like patterns. B. Distribution of integrin in close contact-like regions of initially spreading fibroblasts, note lack of needle’s eye pattern.

The use of monoclonal antibodies has greatly simplified the biochemical characterization of the antigen, due to the ease with which it can be purified by monoclonal antibody affinity chromatography (Neff et al. 1982; Greve & Gottlieb, 1982). The antigen purified by this method (or identified by immunoprecipitation) consists of at least three distinct glycoproteins (Fig. 3), which can only be resolved by SDS-PAGE under non-reducing conditions (Knudsen et al. 1985 ; Hasegawa et al. 1985). At this point, several obvious questions arise. First, does this complex serve as a receptor for extracellular matrix and cytoskeleton-associated molecules? Second, is the antigen a heteromeric complex, or are these merely different glycoproteins all sharing a common epitope? Third, is heteromeric integrity required for function? In the past few years, we have addressed each of these questions.

Fig. 3.

Autoradiogram of integrin analysed by SDS-PAGE. Monoclonal antibody affinity-purified [3SS]integrin was subjected to SDS-PAGE in the presence (reduced) or absence (non-reduced) or beta-mercaptoethanol. Position of Mr standards (×10−3) shown to the right of each gel.

Fig. 3.

Autoradiogram of integrin analysed by SDS-PAGE. Monoclonal antibody affinity-purified [3SS]integrin was subjected to SDS-PAGE in the presence (reduced) or absence (non-reduced) or beta-mercaptoethanol. Position of Mr standards (×10−3) shown to the right of each gel.

The binding properties of the complex have been examined by equilibrium gel filtration (Horwitz et al. 1986). Since this method permits the measurement of receptor—ligand interactions in the constant presence of excess ligand, it is possible to detect binding that occurs at moderate to low affinities and possesses rapid equilibria. This was necessary in the case of matrix receptor—ligand interactions, as earlier work had shown that fibronectin bound to the surface of cells with an affinity in the micromolar range (Akiyama et al. 1985). The results of equilibrium gel filtration of integrin in the presence of laminin and fibronectin are shown in Figs 4 and 5. The change in the elution profile of radioactive integrin in the presence of either ligand indicates that ligand-receptor interaction occurred. The interaction of laminin with integrin could be blocked by the CSAT monoclonal antibody, suggesting that the integrin—laminin interaction monitored by equilibrium gel filtration resembled that which occurred between the cell-associated integrin and laminin (Fig. 4C). Because the complex formed between the monoclonal antibody and the receptor eluted from the gel filtration column in the same position as the fibronectin—receptor complex, it was not possible to determine if the CSAT monoclonal antibody would block fibronectin—integrin binding. However, the interaction of fibronectin with its receptor is sensitive to the fibronectin cell-binding peptide Arg-Gly-Asp (Pytela et al. 1985a). Therefore, it was possible to determine if this peptide would interfere with fibronectin—integrin binding as measured by gel filtration. Fig. 5B shows that this was indeed the case. In the presence of peptides containing the Arg-Gly-Asp sequence, no complex was formed between fibronectin and integrin. Further, this peptide also competed with laminin for binding to integrin (Fig. 4D). Control experiments showed that a related peptide containing the sequence Arg-Gly-Glu was not able to block either fibronectin or laminin binding to integrin (Horwitz et al. 1985). The Arg-Gly-Asp, but not the control peptide, also inhibited chick fibroblast adhesion to both fibronectin and laminin (Horwitz et al. 1985). Thus, the interaction between the receptor complex, integrin, and the matrix molecules, fibronectin and laminin, as measured by equilibrium gel filtration, resembled that which occurred between the membrane-bound receptor and the extracellular ligand.

Fig. 4.

Equilibrium gel filtration elution profile of integrin in the presence of laminin. ,3S-labelled integrin was subjected to equilibrium gel filtration in the presence of: A, CSAT monoclonal antibody; B, laminin; C, laminin plus CSAT monoclonal antibody; D, laminin plus Arg-Gly-Asp-containing peptide. A. (○) Antigen alone; (▪) integrin plus monoclonal antibody. B. (A) Integrin in the presence of 25μgml−1 laminin; (•) integrin plus 100μgml−1 laminin; (▪) integrin plus 200fig ml−1 laminin; (○) integrin plus 400μgml−1 laminin. C. (○) Integrin plus 400μgml−1 laminin; (▪) integrin plus CSAT monoclonal antibody and 400μgml−1 laminin. D. (•) Integrin plus 400μg ml−1 laminin and 1mgml−1 cell-binding tetrapeptide; (▫) integrin plus 400μgml−1 laminin and 1mgml−1 control peptide.

Fig. 4.

Equilibrium gel filtration elution profile of integrin in the presence of laminin. ,3S-labelled integrin was subjected to equilibrium gel filtration in the presence of: A, CSAT monoclonal antibody; B, laminin; C, laminin plus CSAT monoclonal antibody; D, laminin plus Arg-Gly-Asp-containing peptide. A. (○) Antigen alone; (▪) integrin plus monoclonal antibody. B. (A) Integrin in the presence of 25μgml−1 laminin; (•) integrin plus 100μgml−1 laminin; (▪) integrin plus 200fig ml−1 laminin; (○) integrin plus 400μgml−1 laminin. C. (○) Integrin plus 400μgml−1 laminin; (▪) integrin plus CSAT monoclonal antibody and 400μgml−1 laminin. D. (•) Integrin plus 400μg ml−1 laminin and 1mgml−1 cell-binding tetrapeptide; (▫) integrin plus 400μgml−1 laminin and 1mgml−1 control peptide.

Fig. 5.

Equilibrium gel-filtration elution profile of integrin in the presence of fibronectin. [35S]integrin was subjected to equilibrium gel filtration in the presence of: A, fibronectin; or B, fibronectin plus cell-binding tetrapeptide. A. (▫) Integrin alone; (•) integrin plus 200μgml−1 fibronectin. B. (A) Integrin plus 100μgml−1 fibronectin; (•) integrin plus 200figml fibronectin and 1mg ml−1 Arg-Gly-Asp-containing peptide.

Fig. 5.

Equilibrium gel-filtration elution profile of integrin in the presence of fibronectin. [35S]integrin was subjected to equilibrium gel filtration in the presence of: A, fibronectin; or B, fibronectin plus cell-binding tetrapeptide. A. (▫) Integrin alone; (•) integrin plus 200μgml−1 fibronectin. B. (A) Integrin plus 100μgml−1 fibronectin; (•) integrin plus 200figml fibronectin and 1mg ml−1 Arg-Gly-Asp-containing peptide.

If integrin were to serve as a transmembrane link between the extracellular matrix and the cytoskeleton, it should also bind to either cytoskeletal or cytoskeleton- associated molecules. This was also tested by equilibrium gel filtration (Horwitz et al. 1986). It was found that integrin would not bind directly to the cytoskeleton- associated molecules vinculin or alpha actinin. It would, however, bind to talin. The binding of integrin to talin resulted in the formation of a complex of larger Stokes’ radius that eluted closer to the void volume of the gel filtration column in the same manner as when integrin bound fibronectin (Fig. 6). The talin-integrin binding was, however, not sensitive to the cell-binding peptide, showing that talin reacted with integrin at a different site from fibronectin.

Fig. 6.

Equilibrium gel filtration of integrin in the presence of cytoskeleton-associated molecules. [35S]integrin was subjected to equilibrium gel filtration in the presence of 100μgml−1 vinculin (○); 100μl−1 talin (δ); 200μgml−1 talin plus 200μml− 1 vinculin (•). Inset: profile of internal catalase standard included in all experiments. Vo, void volume; Vt, inclusion volume.

Fig. 6.

Equilibrium gel filtration of integrin in the presence of cytoskeleton-associated molecules. [35S]integrin was subjected to equilibrium gel filtration in the presence of 100μgml−1 vinculin (○); 100μl−1 talin (δ); 200μgml−1 talin plus 200μml− 1 vinculin (•). Inset: profile of internal catalase standard included in all experiments. Vo, void volume; Vt, inclusion volume.

Talin and vinculin bind to one another, and in some undetermined manner, appear to mediate actin—membrane interactions (reviewed by Burridge, 1987; and see Geiger et al. this volume). Thus, while vinculin itself could not bind to integrin, it should bind to a talin-integrin complex if the configuration of the complex is physiologically appropriate. This was indeed the case (Fig. 6). The presence of vinculin in an equilibrium gel filtration column did not change the elution position of integrin. The addition of talin and integrin to the column resulted in the characteristic shift in the elution position of integrin, showing that a complex had been formed. If vinculin was added along with talin, the elution profile of integrin was further shifted towards the void volume of the column (Fig. 6), showing that an even larger molecular complex had been formed presumably due to the addition of vinculin to the talin-integrin complex. These experiments show that the cytoskeleton-associated molecule, talin, binds to integrin at a site on the integrin complex different from that which binds fibronectin. It also shows that talin binds to integrin at a site different from that required for talin—vinculin interactions. Thus, it appears that integrin has all the properties of a transmembrane receptor capable of binding both extracellular matrix molecules and cytoskeleton-associated linking molecules.

Structural information that allows us to gain further insight into the function of different regions of integrin is becoming available as the genes coding for members of the complex are being isolated and sequenced. The gene coding for band 3 of integrin has recently been isolated (Tamkun et al. 1986). Nucleotide sequence determination shows that this molecule consists of 803 amino acids. It has multiple glycosylation sites, four cysteine-rich repeats in the extracellular domain, a hydrophobic trans-membrane domain and a consensus tyrosine kinase phosphorylation site in the 47- amino-acid cytoplasmic domain. The structure of this subunit is consistent with the role of integrin as a transmembrane bridge between the extracellular matrix and the cytoplasm. The role of the cytoplasmic domain in talin binding to integrin has been examined using a synthetic peptide consisting of the amino acids Trp-Asp-Thr-Gly- Glu-Asn-Pro-Ile-Tyr-Lys. This peptide is the equivalent of a tryptic fragment from the consensus tyrosine kinase phosphorylation site. If this region of the molecule is required for integrin-talin binding, this peptide should block the interaction of talin and integrin as measured by equilibrium gel filtration. That this is the case is shown in Fig. 7. This experiment makes two important points. First, talin binds to the cytoplasmic domain of integrin; and second, the tyrosine kinase phosphorylation site is in the region of the molecule involved in talin binding.

Fig. 7.

Effect of a synthetic peptide from the cytoplasmic domain of integrin on talin-integrin binding. [3sS]integrin was subjected to equilibrium gel filtration alone, in the presence of talin, or in the presence of talin plus 1 mg ml−1 of synthetic peptide from the cytoplasmic domain of integrin corresponding to the tyrosine kinase phosphorylation site on band 3 glycoprotein (10-mer). (δ) Integrin; (•) integrin + 0·4 mg ml−r talin; (○) integrin + 0·4mgml−1 talin + 1 mg ml−1 10-mer.

Fig. 7.

Effect of a synthetic peptide from the cytoplasmic domain of integrin on talin-integrin binding. [3sS]integrin was subjected to equilibrium gel filtration alone, in the presence of talin, or in the presence of talin plus 1 mg ml−1 of synthetic peptide from the cytoplasmic domain of integrin corresponding to the tyrosine kinase phosphorylation site on band 3 glycoprotein (10-mer). (δ) Integrin; (•) integrin + 0·4 mg ml−r talin; (○) integrin + 0·4mgml−1 talin + 1 mg ml−1 10-mer.

The tyrosine kinase phosphorylation site on integrin is of further interest in the light of the fact that it may play an important role in cellular changes noted during malignant transformation. For example, viral transformation leads to alterations in the morphological and adhesive properties of cells. Immunolocalization studies have shown that the distribution of fibronectin receptors including integrin is greatly altered following viral transformation (Hirst et al. 1986; Chen et al. 19866; Marchisio et al. 1987). In addition, pp60 src has been localized to adhesion plaques, suggesting that proteins found within this complex may be the targets of src kinase. Consistent with this is the observation by Hirst et al. (1986) that band 3 glycoprotein of integrin was phosphorylated following viral transformation. In order to determine if phosphorylation results in an altered ability of integrin to bind talin or extracellular matrix molecules, Rous sarcoma virus-transformed chick cells were labelled with 32P and control non-transformed cells were labelled with [3’S]methionine. Integrin was isolated from both sets of cells. The 32P- and 3’S-labelled integrins were then mixed, and subjected to equilibrium gel filtration in the presence of either fibronectin or talin. The fractions from the column were then analysed in a scintillation counter for 32P and 35S. Fig. 8A shows that the 35S-labelled integrin retained its ability to bind fibronectin and talin. ‘2P-labelled integrin, on the other hand (Fig. 8B), bound neither fibronectin nor talin suggesting that a reduction of the receptor function of integrin following phosphorylation could, at least partially, contribute to the adhesive and morphological changes noted in transformed cells. While this effect of phosphorylation on adhesion has been observed in the pathological process of malignant transformation, it might well be a mechanism whereby a migrating cell could regulate its interaction with the extracellular matrix or its morphology during normal morphogenetic movements.

Fig. 8.

Effect of the phosphorylation of integrin on its ability to bind fibronectin and talin. Preparations of 35S- and 32P-labelled integrin were mixed and subjected to equilibrium gel filtration in the presence or absence of talin and fibronectin. The column eluates were differentially analysed for either 35S or 32P in a liquid scintillation counter. A. Elution profile of [35S]integrin. B. Elution profile of [32P]integrin. Vo, void volume; Vt, inclusion volume. A. (δ) 35S-labelled non-transformed integrin alone; (○) +0·4mgml−1 talin; (•) or + 0·8mgml−1 fibronectin.

Fig. 8.

Effect of the phosphorylation of integrin on its ability to bind fibronectin and talin. Preparations of 35S- and 32P-labelled integrin were mixed and subjected to equilibrium gel filtration in the presence or absence of talin and fibronectin. The column eluates were differentially analysed for either 35S or 32P in a liquid scintillation counter. A. Elution profile of [35S]integrin. B. Elution profile of [32P]integrin. Vo, void volume; Vt, inclusion volume. A. (δ) 35S-labelled non-transformed integrin alone; (○) +0·4mgml−1 talin; (•) or + 0·8mgml−1 fibronectin.

We have demonstrated that integrin is a complex of glycoproteins, not merely a single molecule. The question arises as to whether integrin functions as a complex of glycoproteins, or is merely a mixture of antigenically related, but functionally distinct, molecules that are co-purified with the CSAT monoclonal antibody. Biochemically, it has not been possible to separate the integrin complex except under the denaturing conditions of SDS-PAGE (Knudsen et al. 1985; Hasegawa et al. 1985). We have, however, succeeded in producing a monoclonal antibody specific for the lower molecular weight band 3. of the integrin complex, which is capable of dissociating the complex in such a way as to permit the separation of band 3 from bands 1 and 2 on an antibody affinity column (Bucket al. 1986). Using this antibody, we have been able to divide integrin into two fractions under non-denaturing conditions. One fraction enriched in band 3 glycoprotein, and the other enriched in glycoproteins making up SDS-PAGE bands 1 and 2. When each of the fractions was mixed separately with fibronectin and applied to an equilibrium gel filtration column, there was no change in their elution profile from that seen with each fraction chromatographed in the absence of fibronectin (Fig. 9A,B). However, when the two fractions were mixed and applied to the gel filtration column, two changes were noted (Fig. 9C). First, the elution profile of the mixed complex appears different from that of either integrin fraction alone. The profile is sharper than that of the band 1 plus 2 mixture and elutes earlier from the column than band 3 itself, in the precise position of the intact integrin complex. These hydrodynamic changes indicate that the macromolecular complex has been reconstituted, supporting the contention that integrin exists as a complex of glycoproteins. This contention is further strengthened by sucrose density gradient experiments showing that integrin sediments as a complex (Buck et al. 1985, 1986; Hasegawa et al. 1985). Second, upon mixing the two fractions of integrin glycoproteins, their ability to bind fibronectin was restored. Similarly, they regained their ability to bind talin, which was lost after separation into two fractions (Buck et al. 1986). Thus, the ability of integrin to act as a receptor for extracellular matrix molecules and talin requires that its integrity as a heteromer be maintained. This rules out the possibility that each glycoprotein of the integrin complex may have a separate binding function.

Fig. 9.

Fibronectin binding to reconstituted integrin. J’S-labelled integrin was separated into two fractions, one consisting of bands 1 and 2 glycoproteins, and the other consisting of the band 3 glycoprotein. A. Equilibrium gel filtration of bands 1 plus 2 in the presence and absence of fibronectin. B. Equilibrium gel filtration of band 3 in the presence and absence of fibronectin. C. Equilibrium gel filtration of reconstituted integrin (bands 1 plus 2 mixed with band 3) in the presence and absence of fibronectin. (○) Integrin fractions in the absence of fibronectin: (•) integrin fractions in the presence of fibronectin. VAg, position of antigen elution.

Fig. 9.

Fibronectin binding to reconstituted integrin. J’S-labelled integrin was separated into two fractions, one consisting of bands 1 and 2 glycoproteins, and the other consisting of the band 3 glycoprotein. A. Equilibrium gel filtration of bands 1 plus 2 in the presence and absence of fibronectin. B. Equilibrium gel filtration of band 3 in the presence and absence of fibronectin. C. Equilibrium gel filtration of reconstituted integrin (bands 1 plus 2 mixed with band 3) in the presence and absence of fibronectin. (○) Integrin fractions in the absence of fibronectin: (•) integrin fractions in the presence of fibronectin. VAg, position of antigen elution.

Integrin appears to differ from other mammalian extracellular matrix receptors in its ligand specificity (reviewed by Buck & Horwitz, 1987). Integrin has binding activity for more than one matrix molecule, whereas the mammalian receptors described to date are specific for a single matrix molecule. This suggests that either integrin is a single promiscuous receptor, or it is a mixture of receptors each possessing a common subunit. Support for the latter possibility has come from the observation that the epitope for the CSAT monoclonal antibody is located only on integrin band 3 glycoprotein (Bucket al. 1986). However, when the issue of receptor promiscuity is explored further by competition experiments, a different conclusion seems likely. For these experiments, we have taken advantage of the fact that vitronectin binds to membrane receptors more tightly than either fibronectin or laminin (Pytela et al. 19856). The interaction of vitronectin with integrin is shown in Fig. 10. When vitronectin was mixed with integrin and applied to the equilibrium gel filtration column, most of the vitronectin eluted in the same position as integrin (Fig. 10). Vitronectin-integrin binding is sensitive to both the CSAT monoclonal antibody and the Arg-Gly-Asp cell binding peptide (Horwitz & Buck, unpublished) and therefore resembles that of integrin binding to other matrix molecules. The ability of vitronectin to compete with fibronectin or laminin for integrin binding sites was tested by mixing radioactive integrin with either of these matrix molecules in the presence or absence of vitronectin, and subjecting the mixture to equilibrium gel filtration. The presence of vitronectin abolished the ability of integrin to bind either fibronectin or laminin (Fig. 11). By this criterion, integrin would appear to be a promiscuous receptor with respect to the binding of laminin, fibronectin and vitronectin.

Fig. 10.

Integrin-vitronectin binding as measured by equilibrium gel filtration. Vitronectin, integrin or a mixture of vitronectin and integrin was subjected to gel filtration. In this case, gels were not pre-equilibrated with the ligand since vitronectin-integrin binding was more stable than integrin binding to other matrix molecules. Lowry protein analysis was performed on each column fraction to determine the location of each protein or complex. Note the decrease in the amount of protein in the region of free vitronectin and the concomitant increase in protein in the region of integrin elution upon mixing the two molecules. Vo, void volume; Vt inclusion volume. (δ) Integrin; (○) vitronectin; (•) integrin + vitronectin.

Fig. 10.

Integrin-vitronectin binding as measured by equilibrium gel filtration. Vitronectin, integrin or a mixture of vitronectin and integrin was subjected to gel filtration. In this case, gels were not pre-equilibrated with the ligand since vitronectin-integrin binding was more stable than integrin binding to other matrix molecules. Lowry protein analysis was performed on each column fraction to determine the location of each protein or complex. Note the decrease in the amount of protein in the region of free vitronectin and the concomitant increase in protein in the region of integrin elution upon mixing the two molecules. Vo, void volume; Vt inclusion volume. (δ) Integrin; (○) vitronectin; (•) integrin + vitronectin.

Fig. 11.

Vitronectin competes with fibronectin and laminin for binding to integrin. 35S- labelled integrin was subjected to equilibrium gel electrophoresis in the presence of fibronectin, laminin or a mixture of each of these molecules and vitronectin. A. Integrin plus fibronectin (0 · 8 mg ml−1, (○) or a mixture of fibronectin and vitronectin (0 · 8mgml−1 + 10 μ g (•)). B. Integrin plus laminin (0 · 4mgml−1 (○) or a mixture of laminin and vitronectin (0 · 1mgml− 1 + 10 μ g (•)). Note, in each case, the vitronectin prevents the characteristic shift in the elution profile of integrin in the presence of one of the matrix molecules.

Fig. 11.

Vitronectin competes with fibronectin and laminin for binding to integrin. 35S- labelled integrin was subjected to equilibrium gel electrophoresis in the presence of fibronectin, laminin or a mixture of each of these molecules and vitronectin. A. Integrin plus fibronectin (0 · 8 mg ml−1, (○) or a mixture of fibronectin and vitronectin (0 · 8mgml−1 + 10 μ g (•)). B. Integrin plus laminin (0 · 4mgml−1 (○) or a mixture of laminin and vitronectin (0 · 1mgml− 1 + 10 μ g (•)). Note, in each case, the vitronectin prevents the characteristic shift in the elution profile of integrin in the presence of one of the matrix molecules.

To compare further the properties of integrin with those of a mammalian receptor, we have purified a fibronectin receptor from the rat myoblast cell line L6A. The receptor is, as expected, a heterodimer, and antibodies prepared against this receptor will interfere with L6A adhesion to fibronectin-coated tissue culture dishes (Fig. 12). This receptor will bind to fibronectin, but not vitronectin or laminin (Horwitz & Buck, unpublished). The ability of antibodies against integrin to bind to this receptor was determined by immunoblot analysis (Fig. 13). Antibodies against integrin or against band 3 of integrin reacted with the lower molecular weight (beta) band of the L6A fibronectin receptor. There was weak cross-reactivity with the higher molecular weight (alpha) band of the L6A receptor also. The antibody against the rat fibronectin receptor reacted only weakly with integrin and, in this case, it was with the band 2 glycoprotein. These data indicate structural similarities between fibronectin receptors from various species and are consistent with the suggestions that these receptors belong to a single family of cell-surface adhesion molecules (Ruoslahti & Pierschbacher, 1986; Leptin, 1986; Hynes, 1987; Hemler et al. 1987; Takada et al. 1987a,b).

Fig. 12.

Fibronectin receptor from rat L6A cells. Silver staining of purified receptor following SDS-PAGE under non-reducing conditions. Antiserum was prepared against this receptor and tested for its effect on L6A cell adhesion. A. Control cells in the presence of pre-immune serum. B. L6A cells 4h after addition of antisera raised against the rat fibronectin receptor. C. L6A cells exposed to anti-fibronectin receptor that had been mixed with purified receptor. The purified receptor completely blocked the antibody-induced adhesive alterations.

Fig. 12.

Fibronectin receptor from rat L6A cells. Silver staining of purified receptor following SDS-PAGE under non-reducing conditions. Antiserum was prepared against this receptor and tested for its effect on L6A cell adhesion. A. Control cells in the presence of pre-immune serum. B. L6A cells 4h after addition of antisera raised against the rat fibronectin receptor. C. L6A cells exposed to anti-fibronectin receptor that had been mixed with purified receptor. The purified receptor completely blocked the antibody-induced adhesive alterations.

Fig. 13.

Immunoblot comparisons of integrin and rat fibronectin receptor. Lanes 1 – 5, integrin; and lanes 6 – 9, rat fibronectin receptor. Lanes 1 – 4, from a different gel from lanes 5-9. Lanes 1 and 5, reacted with a monoclonal antibody specific for band 3 glycoprotein mixed with one specific for band 1 glycoprotein. These serve as internal controls and mark the position of their respective antigens on the gels. Lanes 2 and 3, reacted with a polyclonal antibody raised against integrin. Lanes 3 and 7, reacted with a polyclonal antibody raised against the rat fibronectin receptor. Lane 9, the pre-immune control. Antigen-antibody interactions were detected using antibodies conjugated with alkaline phosphatase.

Fig. 13.

Immunoblot comparisons of integrin and rat fibronectin receptor. Lanes 1 – 5, integrin; and lanes 6 – 9, rat fibronectin receptor. Lanes 1 – 4, from a different gel from lanes 5-9. Lanes 1 and 5, reacted with a monoclonal antibody specific for band 3 glycoprotein mixed with one specific for band 1 glycoprotein. These serve as internal controls and mark the position of their respective antigens on the gels. Lanes 2 and 3, reacted with a polyclonal antibody raised against integrin. Lanes 3 and 7, reacted with a polyclonal antibody raised against the rat fibronectin receptor. Lane 9, the pre-immune control. Antigen-antibody interactions were detected using antibodies conjugated with alkaline phosphatase.

Fig. 14 summarizes our current thinking about the general structure and function of integrin. Structurally, we are beginning to view integrin as a heterodimer consisting of an alpha and a beta subunit (reviewed by Buck & Horwitz, 1987). The alpha subunit is composed of a short membrane-spanning peptide disulphide linked to a larger extracellular peptide. The beta subunit is a highly disulphide cross-linked molecule that interacts non-covalently with the alpha subunit to produce a functionally active receptor. Our knowledge of the beta subunit structure comes from actual gene sequence data (Tamkun et al. 1986). Our concept of the structure of the alpha subunit is formed predominantly from analogy with the known structure of the mammalian vitronectin and fibronectin receptor (Argraves et al. 1986; Suzuki et al. 1986) and the fact that it decreases in apparent molecular weight following reduction, suggesting that a small fragment is lost (Knudsen et al. 1985; Hasegawa et al. 1985). Although as purified by antibody affinity chromatography integrin appears to contain at least three subunits, the molecular weight of the complex as determined hydrodynamically (Bucket al. 1985; Hasegawa et al. 1985) is between 215 and 230 (X 10’). This would not accommodate a trimeric structure made up of three glycoproteins with estimated molecular weights of 160, 140 and 110 (X103). The fact that several substratum molecules appear to compete for binding to the same site on integrin suggest that it functions as a promiscuous receptor capable of binding more than one matrix molecule. In this respect, it differs from the mammalian fibronectin and vitronectin receptors (Pytelaef al. 1985a,b) and is more like the platelet cytoadhesin Ilb/lIIa (Plow et al. 1985).

Fig. 14.

Schematic drawing of integrin. Top portion showing integrin as a transmembrane bridging molecule between the extracellular matrix (ECM) and the cytoskeleton- associated molecules talin and vinculin. Question marks designate undetermined linkages between vinculin and actin. Lower portion is an enlarged drawing of integrin. The subunits have been labelled α and β to conform to the convention established for other receptors and as suggested by Hynes (1987). Integrin as shown here is a heterodimeric promiscuous receptor capable of binding to more than one extracellular matrix molecule. See the text for detailed discussion of structure.

Fig. 14.

Schematic drawing of integrin. Top portion showing integrin as a transmembrane bridging molecule between the extracellular matrix (ECM) and the cytoskeleton- associated molecules talin and vinculin. Question marks designate undetermined linkages between vinculin and actin. Lower portion is an enlarged drawing of integrin. The subunits have been labelled α and β to conform to the convention established for other receptors and as suggested by Hynes (1987). Integrin as shown here is a heterodimeric promiscuous receptor capable of binding to more than one extracellular matrix molecule. See the text for detailed discussion of structure.

If integrin is, in fact, a promiscuous heterodimer, what about the other glycoproteins that co-purify with it on an antibody affinity column? As pointed out earlier, the CSAT monoclonal antibody reacts with band 3 or the beta subunit of integrin. It is possible that this subunit is a constituent of more than one receptor and that while one combination of alpha and beta subunits come together to form integrin, another alpha-like subunit could combine with the same beta subunit to form a receptor for another group of matrix molecules such as the collagens or proteoglycans. The function of alpha-like subunits remains to be determined.

The discovery of similarities in structure and sensitivity of binding to the fibronectin cell-binding peptide Arg-Gly-Asp have led to the speculation that the cell matrix receptors, lymphoid antigens and platelet cytoadhesions might all be members of a single supergene family (Ruoslahti & Pierschbacher, 1986; Hynes, 1987; Takada et al. 1987a,b; Kishimoto et al. 1987; Ginsberg et al. 1987). On the basis of careful serological studies, Hemler and his associates (Hemler et al. 1987; Takada et al. 1987a,b) have shown that the beta subunit of the VLA group of antigens expressed on mitogen-stimulated T cells (Sanchez-Madrid et al. 1982, 1983 ; Hemler et al. 1983 ; Hemler et al. 1985a,b) and the beta subunit of avian integrin and mammalian fibronectin receptors are antigenically similar. This is consistent with our comparisons of the fibronectin receptor from rat cells and integrin. Thus, one subfamily of receptors defined by a related beta subunit would consist of integrin, the mammalian fibronectin receptors and the VLA antigens. The second subfamily is defined by another common beta subunit, which is associated with LFA-1, Mac-1 and pl50,95 (Sanchez-Madrid et al. 1983; Springer et al. 1985). The alpha subunits of these receptors are different. The third subfamily the cytoadhesins has been defined by Ginsberg et al. (1987). It consists of the platelet cytoadhesion (Ilb/lIIa) and the mammalian vitronectin receptor. Again, these receptors share a common structurally and antigenically related beta subunit and distinct alpha subunits. The relationship between these receptors is summarized in Table 1, using receptor subunit designations suggested by Hynes (1987). The properties of these receptors are the following. They are heterodimers; some, but not all alpha subunits consist of disulphide-linked heavy and light chains; some, but not all of the receptor-ligand interactions are blocked by the fibronectin cell-binding peptide. All beta subunits contain disulphide cross-linked cysteine-rich repeats. The binding specificities are determined primarily by the alpha subunit. The importance of the alpha subunits is best illustrated in the case of the cytoadhesins (family 3) with gpIIb-IIIa being a promiscuous receptor that binds several ligands and the mammalian vitronectin receptor, which contains a different alpha subunit, being a specific receptor that binds a single ligand only. As the genes for more receptors are isolated and sequenced we will gain further insight into the structural and functional inter-relationships of the members of this supergene family as well as an understanding of how the genes are regulated, the proteins processed and the receptor subunits sorted out and assembled. New additions and combinations will undoubtedly be discovered, which will necessitate refinement of these relationships.

Table 1.

Supergene family of related receptors

Supergene family of related receptors
Supergene family of related receptors

Functionally, integrin is most probably involved in cell motility, morphogenetic movements and the initial establishment of cell-matrix interactions. Its role in development has been demonstrated indirectly by the ability of both CSAT and JG22 monoclonal antibodies as well as Arg-Gly-Asp-containing peptides to interfere with early development in chick embryos (Boucaut et al. 1984; Bronner-Fraser, 1985 ; Duband et al. 1986; Jaffredo et al. 1986). Also integrin-like molecules termed position-specific (PS) antigens, have been implicated in Drosophila development (Wilcox et al. 1984; Wilcox & Leptin, 1985). Here too, the Arg-Gly-Asp peptides have been shown to interfere with morphogenesis (Naidet et al. 1987). The apparent evolutionary conservation of integrin-like molecules would argue for its importance in biological systems.

In conclusion, over the past 16 or so years since the description of adhesion plaques, we have made considerable advances in our understanding of the molecules involved in the initial establishment of cell-matrix interactions and coordinating the organization of matrix and cytoskeletal elements. This work has led to the discovery of relationships between unexpected groups of molecules and has begun to lay the ground work for investigations into the mechanisms tissue morphogenesis, metastasis and cellular adhesive interactions involved in various pathological disorders.

We are grateful for the assistance of Ms Marie Lennon in the preparation of the manuscript, Ms Kimberly Duggan for her technical assistance and the preparation of figures, and Mr Jim Averbach for the preparation of drawings. This work was supported by National Institutes of Health grants CA 19144 and CA 10815 to C.A.B., and GM 23244 to A.F.H., and by the H. M. Watts Jr Neuromuscular Disease Research Center (A.F.H.).

Abercrombie
,
M.
,
Heaysman
,
E. M.
&
Pegrum
,
S. M.
(
1971
).
The locomotion of fibroblasts in culture IV. Electron microscopy of the leading lamella
.
Expl Cell Res
.
67
,
359
367
.
Axvíxma
,
S. K.
,
Hasegawa
,
E.
,
Hasegawa
,
T.
&
Yamada
,
K.
(
1985
).
The interaction of fibronectin fragments with fibroblastic cells
.
J. biol. Chem
.
260
,
13 256
13 260
.
Argraves
,
W. S.
,
Pytela
,
R.
,
Suzuki
,
S.
,
Millan
,
J. L.
,
Pierschbacher
,
M. D.
&
Ruoslahti
,
E.
(
1986
).
cDNA sequences from the alpha subunit of the fibronectin receptor predict a transmembrane domain and a short cytoplasmic peptide
.
J. biol. Chem
.
261
,
12922
12024
.
Boucaut
,
J. C.
,
Darriebere
,
T.
,
Poole
,
T. J.
,
Aoyama
,
H.
,
Yamada
,
K. M.
&
Thiery
,
J. P.
(
1984
).
Biologically active synthetic peptides as probes of embryonic development: a competitive peptide inhibition of fibronectin function inhibits gastrulation in an amphibian embryos and neural crest migration in avian embryos
.
J. Cell Biol
.
99
,
1822
1830
.
Bronner-Fraser
,
M.
(
1985
).
Alterations in neural crest migration by a monoclonal antibody that affects cell adhesion
.
J. Cell Biol
.
101
,
610
617
.
Buck
,
C. A.
&
Horwitz
,
A. F.
(
1987
).
Cell surface receptors for extracellular matrix molecules. A
.
Rev. Cell Biol
.
3
,
179
205
.
Buck
,
C.
,
Knudsen
,
K. A.
,
Damsky
,
C. H.
,
Decker
,
C.
,
Greggs
,
R. R.
,
Duggan
,
K. E.
,
Bozyczko
,
D.
&
Horwitz
,
A.
(
1985
).
Intergral membrane protein complexes in cell-matrix adhesion
.
In The Cell in Contact: Adhesions and Junctions as Morphogenic Determinants
(ed.
G. M.
Edelman
&
J. P.
Thiery
), pp.
345
364
.
New York
:
Wiley
.
Buck
,
C. A.
,
Shea
,
E.
,
Duggin
,
K.
&
Horwitz
,
A. F.
(
1986
).
Integrin (the CSAT antigen): Functionality requires oligomeric integrity
.
J. Cell Biol
.
103
,
2421
2428
.
Burridge
,
K.
(
1987
).
Substrate adhesions in normal and transformed fibroblasts: Organization and regulation of cytoskeletal, membrane and extracellular matrix components at focal contacts
.
Cancer Rev
.
4
,
18
78
.
Chen
,
W.-T.
, Chen, J.-M. &
Mueller
,
S. C.
(
1986a
).
Coupled expression and co-localization of 140 K cell adhesion molecules, fibronectin, and laminin during morphogenesis and cyto differentiation of chick lung cells
.
J. Cell Biol
.
103
,
1073
1090
.
Chen
,
W.-T.
,
Greve
,
J. M.
,
Gottlieb
,
D. I.
&
Singer
,
S. J.
(
1985a
).
Immunocytochemical localization of 140 kD cell adhesion molecules in cultured chicken fibroblasts and in chicken smooth muscle and intestinal epithelial tissues
.
J. Histochem. Cytochem
.
33
,
576
586
.
Chen
,
W.-T.
,
Hasegawa
,
E.
,
Hasegawa
,
T.
,
Weinstock
,
C.
&
Yamada
,
K. M.
(
1985b
).
Development of cell surface linkage complexes in cultured fibroblasts
.
J. Cell Biol
.
100
,
1103
1104
.
Chen
,
W.-T.
,
Wong
,
J.
,
Hasegawa
,
T.
,
Yamada
,
S.
&
Yamada
,
K. M.
(
1986b
).
Regulation of fibronectin receptor distribution by transformation, exogenous fibronectin, and synthetic peptides. J
.
Cell Biol
.
103
,
1649
1661
.
Damsky
,
C. H.
,
Knudsen
,
K. A.
,
Bradley
,
D.
,
Buck
,
C. A.
&
Horwitz
,
A. F.
(
1985
).
Distribution of the cell-substratum attachment (CSAT) antigen on myogenic and fibroblastic cells in culture
.
J. Cell Biol
.
100
,
1528
1539
.
Decker
,
C.
,
Greggs
,
R.
,
Duggan
,
K.
,
Stubbs
,
J.
&
Horwitz
,
A.
(
1984
).
Adhesive multiplicity in the interaction of embryonic fibroblasts and myoblasts with extracellular matrices. ‘
.
Cell Biol
.
99
,
1398
1404
.
Duband
,
J. L.
,
Rocker
,
S.
,
Chen
,
W. T.
,
Yamada
,
K. M.
&
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
178
.
Geiger
,
B.
,
Volk
,
T.
,
Volberg
,
T.
&
Bendori
,
R.
(
1987
).
Molecular interactions in adherens-type contacts
.
J. Cell Sci
.
Suppl
.
8
,
251
272
.
Ginsberg
,
M. H.
,
Loftun
,
J.
,
Pierschbacher
,
M. D.
,
Ruoslahti
,
E.
&
Plow
,
E. F.
(
1987
).
Immunochemical and N-terminal sequence comparison of two cytoadhesins indicates they contain similar or identical beta subunits and distinct alpha subunits
.
J, biol. Chem
. (in press).
Greve
,
J. M.
&
Gottlieb
,
D. I.
(
1982
).
Monoclonal antibodies which alter the morphology of culture chick myogome cells
.
J. Cell Biochem
.
18
,
221
230
.
Hasegawa
,
T.
,
Hasegawa
,
E.
,
Chen
,
W. T.
&
Yamada
,
K. M.
(
1985
).
Characterization of a membrane-associated glycoprotein complex implicated in cell adhesion to fibronectin
.
J. Cell Biochem
.
28
,
307
318
.
Hemler
,
M. E.
,
Huang
,
C.
&
Schwarz
,
L.
(
1987
).
The VLA protein family
.
J. biol. Chem
.
262
,
3300
3309
.
Hemler
,
M. E.
,
Jacobson
,
J. G.
,
Brenner
,
M. B.
,
Mann
,
D.
&
Strominger
,
J. L.
(
1985a
).
VLA-1: a T cell surface antigen which defines a novel late stage of human T cell activation
.
Eur.J. Immun
.
15
,
502
508
.
Hemler
,
M. E.
,
Jacobson
,
J. G.
&
Strominger
,
J. L.
(
1985b
).
Biochemical characterization of VLA-1 and VLA-2 cell surface heterodimers on activated T cells
.
J. biol. Chem
.
260
,
15246
15252
.
Hemler
,
M. E.
,
Ware
,
C. S.
&
Strominger
,
J. L.
(
1983
).
Characterization of a novel differentiation antigen complex recognized by a monoclonal antibody (A-1A5): Unique activation-specific molecular forms on stimulated T cells
.
J. Immun
.
131
,
334
340
.
Hirst
,
R.
,
Horwitz
,
A.
,
Buck
,
C.
&
Rohrschneider
,
L.
(
1986
).
Phosphorylation of the fibronectin receptor complex in cells transformed by onocogenes that encode tyrosine kinases
.
Proc. natn. Acad. Sri. U.S.A
.
83
,
6470
6474
.
Horwitz
,
A.
,
Duggan
,
K.
,
Buck
,
C.
,
Beckerle
,
K.
&
Burridge
,
K.
(
1986
).
Interaction of plasma membrane fibronectin receptor with talin-a transmembrane linkage
.
Nature, Land
.
320
,
531
533
.
Horwitz
,
A.
,
Duggan
,
K.
,
Greggs
,
R.
,
Decker
,
C.
&
Buck
,
C.
(
1985
).
Cell substrate attachment (CSAT) antigen has properties of a receptor for laminin and fibronectin
.
J. Cell Biol
.
103
,
2134
2144
.
Hynes
,
R. O.
(
1987
).
Integrins: A family of cell surface receptors
.
Cell
48
,
549
554
.
IZZARD
,
C. S.
,
Radinsky
,
R.
&
Culp
,
L. A.
(
1986
).
Substratum contacts and cytoskeletal reorganization of BALB/c 3T3 cells on a cell-binding fragment and heparin-binding fragment of plasma fibronectin
.
Expl Cell Res
.
165
,
320
336
.
Jaffredo
,
T.
,
Horwitz
,
A. F.
,
Buck
,
C. A.
,
Rong
,
P. W.
&
Dieterlen-Lievre
,
F.
(
1986
).
CSAT antibody interferes with in vivo migration of somitic myoblast precursors into the body wall
.
In The Somites in Developing Embryos
(ed.
R.
Bellaires
&
J. W.
Lash
), pp.
225
236
.
New York
:
Plenum
.
Kishimoto
,
T. K.
,
O’Connor
,
K.
,
Lee
,
A.
,
Roberts
,
T. M.
&
Springer
,
T. A.
(
1987
).
Cloning of the beta subunit of the leukocyte adhesion proteins: homology to an extracellular matrix receptor defines a novel supergene family
.
Cell
48
,
681
690
.
Knudsen
,
K. A.
,
Horwitz
,
A.
, &
Buck
,
C. A.
(
1985
).
A monoclonal antibody identifies a glycoprotein complex involved in cell–substratum adhesion
.
Expl Cell Res
.
157
,
218
226
.
Knudsen
,
K. A.
,
Rao
,
P.
,
Damsky
,
C. H.
&
Buck
,
C. A.
(
1981
).
Membrane glycoproteins involved in cell-substratum adhesion
.
Proc. natn. Acad. Sri. U.S.A
.
78
,
6071
6075
.
Lark
,
M. W.
,
Laterra
,
J.
&
Culp
,
L. A.
(
1985
).
Close and local contact adhesions of fibroblasts to a fibronectin-containing matrix
.
Fedn Proc. Fedn Am. Socs Exp. Biol
.
44
,
394
403
.
Leptin
,
M.
(
1986
).
The fibronectin receptor family
.
Nature, Land
.
321
,
728
729
.
Marchisio
,
P.
,
Cirillo
,
D.
,
Teti
,
A.
,
Zambonin-Zallone
,
A.
&
Tarone
,
G.
(
1987
).
Rous sarcoma virus transformed fibroblasts and cells of monocytic origin display a peculiar dot-like organization of cvtoskeletal proteins involved in microfilament-membrane interactions
.
Expl Cell Res
.
169
,
202
214
.
Naidet
,
C.
,
Semeriva
,
M.
,
Yamada
,
K.
&
Thiery
,
J.-P
. (
1987
).
Peptides containing the cellattachment recognition sequence Arg-Gly-Asp prevent gastrulation in Drosophila embryos
.
Nature, Lond
.
325
,
348
350
.
Neff
,
N. T.
,
Lowrey
,
C.
,
Decker
,
C.
,
Tovar
,
A.
,
Damsky
,
C.
,
Buck
,
C.
&
Horwitz
,
A.
(
1982
).
A monoclonal antibody detaches embryonic skeletal muscle from extracellular matrices
.
J. Cell Biol
.
95
,
654
666
.
Oesch
,
B.
&
Birchmeier
,
W.
(
1982
).
New surface component of fibroblast focal contacts identified by a monoclonal antibody
.
Cell
31
,
671
679
.
Plow
,
E. F.
,
Pierschbacher
,
M. D.
,
Ruoslahti
,
E.
,
Marguerie
,
G. A.
&
Ginsberg
,
M. H.
(
1985
).
The effect of Arg-Gly-Asp-containing peptides on fibrinogen and von Willerbrand factor binding to platelets
.
Proc. natn. Acad. Sci. U.S.A
.
82
,
8057
8061
.
Pytela
,
R.
,
Pierschbacher
,
M. D.
&
Ruoslahti
,
E.
(
1985a
).
Identification and isolation of a 140 kd cell surface glycoprotein with properties expected of a fibronectin receptor
.
Cell
40
,
191
198
.
Pytela
,
R.
,
Plerschbacher
,
M. D.
&
Ruoslahti
,
E.
(
1985b
).
A 125/115 kDa cell surface receptor specific for vitronectin interacts with the arginine-glycine-aspartic acid adhesion sequence derived from fibronectin
.
Proc. natn. Acad. Sci. U.SA
.
82
,
5766
5700
.
Rapraeger
,
A.
,
Jalkanen
,
M.
&
Bernfield
,
M.
(
1986
).
Cell surface proteoglycan associates with the cytoskeleton at the basolateral celt surface of mouse mammary epithelial cells
.
J. Cell Biol
.
103
,
2683
2696
.
Rogalski
,
A. A. Singer
, S. J
. (
1985
).
An integral membrane glycoprotein associated with membrane sites of actin with microfilaments
.
J. Cell Biol
.
101
,
785
801
.
Ruoslahti
,
E.
&
Plerschbacher
,
M. D.
(
1986
).
Arg-Gly-Asp: A versatile cell recognition signal
.
Cell
44
.
517
518
.
Sanchez-Madrid
,
F.
,
Krensky
,
A. M.
,
Ware
,
C. F.
,
Robbins
,
E.
,
Strominger
,
J. L.
,
Burakoff
,
S. J.
&
Springer
,
T. A.
(
1982
).
Three distinct antigens associated with human T-lymphocyte mediate cytolysis: LFA-1, LFA-2 and LFA-3
.
Proc. natn. Acad. Sci. U.S.A
.
79
,
7489
7493
.
Sanchez-Madrid
,
F.
,
Nagy
,
J. A.
,
Robbins
,
E.
,
Simon
,
P.
&
Springer
,
T. A.
(
1983
).
A human leukocyte differentiation antigen family with distinct alpha subunits and a common beta subunit : the lymphocyte function-associated antigen (LFA-1), the C3bi complement receptor (OKM1/ Mac-1), and the pl50,95 molecule
.,
J, exp. Med
.
158
,
1875
1803
.
Singer
,
I. L
,
Kawka
,
D. W.
,
Scott
,
S.
,
Mumford
,
R. A.
&
Lark
,
M. W.
(
1987
).
The fibronectin cell attachment sequence Arg-Gly-Asp-Ser promotes focal contact formation during early fibroblast attachment and spreading
.
J. Cell Biol
.
104
,
573
584
.
Springer
,
T. A.
,
Teplow
,
D. B.
&
Dreyer
,
W. J.
(
1985
).
Sequence homology of the LFA-1 and Mac-1 leukocyte adhesion glycoproteins and unexpected relation to leukocyte interferon
.
Nature, Land
.
314
,
540
542
.
Suzuki
,
S.
,
Argraves
,
W. S.
,
Pytela
,
R.
,
Arai
,
H.
,
Krusius
,
T.
,
Pierschbacher
,
M. D.
&
Ruoslahti
,
E.
(
1986
).
cDNA and amino sequences of the cell adhesive protein receptor recognizing vitronectin reveal a transmembrane domain and homologies with other adhesion protein receptors
.
Proc. natn. Acad. Sci. U.S.A
.
83
,
8614
8618
.
Takada
,
Y.
,
Strominger
,
J. L.
&
Hemler
,
M. E.
(
1987a
).
The VLA family of heterodimers are members of a superfamily of molecules involved in adhesion and embryogenesis
.
Proc. natn. Acad. Sci. U.S.A
.
84
,
3239
3243
.
Takada
,
Y.
,
Strominger
,
C. H.
&
Hemler
,
M. E.
(
1987b
).
Fibronectin receptor structures within the VLA family of heterodimers
.
Nature, Lond
.
326
,
607
609
.
Tamkum
,
J. W.
,
DeSimone
,
D. W.
,
Fonda
,
D.
,
Patel
,
R. S.
,
Buck
,
C. A.
,
Horwitz
,
A. F.
&
Hynes
,
R. O.
(
1986
).
Structure of integrin, a glycoprotein involved in the transmembrane linkage between fibronectin and actin
.
Cell
46
,
271
282
.
Wilcox
,
M.
,
Brown
,
N.
,
Piovant
,
M.
,
Smith
,
R. J.
&
White
,
R. A. H.
(
1984
).
The Drosophila position-specific antigens are a family of cell surface glycoprotein complexes
.
EMBO J
.
3
,
2307
2313
.
Wilcox
,
M.
&
Leptin
,
M.
(
1985
).
Tissue-specific modulation of a set of related cell surface antigens in Drosophila
.
Nature, Lond
.
316
,
351
354
.
Wylie
,
D. E.
,
Damsky
,
C. H.
&
Buck
,
C. A.
(
1979
).
Studies on the function of cell surface glycoproteins
.
J. Cell Biol
.
80
,
385
401
.