Specification of the Direction of Adhesive Signaling by the Integrin β Cytoplasmic Domain*

Integrin adhesion receptors can signal in two directions: first, they can regulate cellular behaviors by modulating cellular signaling enzymes (“outside-in signaling”); second, cells can regulate the affinity of integrins (“inside-out signaling”) by such pathways. Integrin β cytoplasmic domains (tails) mediate both types of signaling, and Src family kinases (SFKs) and talin, which bind to β tails, are important for integrin signaling. Here, we utilized “homology scanning” mutagenesis to identify β tail mutants selectively defective in c-Src binding and found that amino acid exchanges affecting a combination of an Arg and Thr residue in the integrin β3 tail control the binding specificity for SFKs but have no effect on talin binding. Using β tail mutants at these residues, we found that SFK binding to integrin β tails is dispensable for inside-out signaling but is obligatory for cell spreading, a marker of outside-in signaling. Conversely, we found that point mutations that disrupt talin binding abolish integrin activation, but they do not inhibit SFK binding to the β3 tail or the initiation of outside-in signaling once the integrins are in a high affinity form. Thus, we show that inside-out and outside-in integrin signaling are mediated by distinct and separable interactions of the integrin β tails. Furthermore, based on our results, it is possible to discern the relative contributions of the direction of integrin signaling on biological functions in cell culture and, ultimately, in vivo.

Integrin adhesion receptors can signal in two directions: first, they can regulate cellular behaviors by modulating cellular signaling enzymes ("outside-in signaling"); second, cells can regulate the affinity of integrins ("inside-out signaling") by such pathways. Integrin ␤ cytoplasmic domains (tails) mediate both types of signaling, and Src family kinases (SFKs) and talin, which bind to ␤ tails, are important for integrin signaling. Here, we utilized "homology scanning" mutagenesis to identify ␤ tail mutants selectively defective in c-Src binding and found that amino acid exchanges affecting a combination of an Arg and Thr residue in the integrin ␤3 tail control the binding specificity for SFKs but have no effect on talin binding. Using ␤ tail mutants at these residues, we found that SFK binding to integrin ␤ tails is dispensable for inside-out signaling but is obligatory for cell spreading, a marker of outside-in signaling. Conversely, we found that point mutations that disrupt talin binding abolish integrin activation, but they do not inhibit SFK binding to the ␤3 tail or the initiation of outside-in signaling once the integrins are in a high affinity form. Thus, we show that inside-out and outside-in integrin signaling are mediated by distinct and separable interactions of the integrin ␤ tails. Furthermore, based on our results, it is possible to discern the relative contributions of the direction of integrin signaling on biological functions in cell culture and, ultimately, in vivo.
Cell-cell and cell-matrix adhesion, mediated by integrin adhesion receptors, is essential to the development and functioning of multicellular organisms (1)(2)(3)(4). In addition to supporting adhesion, these receptors are bona fide signaling receptors that inform the cell about the physical and biochemical nature of its environment, leading to regulation of gene expression and cell proliferation, differentiation, and survival (1, 4 -7). This process is frequently referred to as "outside-in signaling." The ability of many integrins to bind ligands is regulated by cellular signaling mechanisms (8 -11). This process, often called integrin "activation" or "inside-out signaling," can arise through conformational changes in the integrin extracellular domain and/or changes in the physical distribution of these receptors on the cell surface (12). Thus, integrins can transmit information in both directions across the plasma membrane. A central question is to what extent each form of integrin signaling contributes to particular biological functions.
Integrins are heterodimers of non-covalently associated ␣ and ␤ subunits. The N terminus of each subunit is extracellular, and a single transmembrane domain separates the ϳ700 -1200-residue extracellular domain from a usually short (5-65residue) cytoplasmic domain, or tail (13). The cytoplasmic domains of integrins play an essential role in bi-directional signaling processes, and intensive work has sought to identify cellular proteins that interact directly with these domains (14). Recent studies have focused attention on two groups of proteins that bind to integrin ␤ cytoplasmic domains and regulate integrin signaling: Src family kinases (SFKs) 1 (15) and talin (16).
SFKs are constitutively associated with integrins, and their activation is a proximal and early consequence of integrin clustering (17,18). The presence of SFKs is obligatory for the majority of tyrosine phosphorylation events that follow integrin-mediated adhesion (19). Cell spreading, a measure of integrin signaling into cells, depends on the presence of SFKs. We recently found that SFKs can bind directly to integrin ␤ cytoplasmic domains via their SH3 domains (15). This interaction requires the C terminus of the integrin ␤ tail and promotes the activation of the SFKs (15) and downstream transcriptional factors such as nuclear factor-B (20). Thus, the binding of SFKs to integrin ␤ tails is important for outside-in signaling; however, its role in integrin activation remains to be clarified.
Talin 1 is required for many integrin functions (21)(22)(23), in part because it is a pivotal connection between integrins and the actin cytoskeleton. However, recent studies have also implicated talin as a key regulator of inside-out integrin signaling. Overexpression of the N terminus of talin leads to integrin activation (24,25), and short hairpin RNA knockdown of talin expression in cultured cells (16) inhibits the activation of ␤1 and ␤3 integrins. Furthermore, in platelets and their megakaryocyte precursors, intracellular signals initiated by agonist binding to G-protein-coupled receptors result in integrin ␣IIb␤3 activation. Expression of talin short hairpin RNA, but not mismatched short hairpin RNA, blocked agonist-stimulated fibrinogen binding to megakaryocytes. Thus, talin is a downstream target of cellular signaling pathways that activate integrins. In addition, x-ray crystallography and NMR spectroscopy have indicated that a 96-residue F3 domain of talin interacts with multiple sites on the integrin ␤3 tail (26 -28). Thus, binding of the talin F3 domain to the ␤3 cytoplasmic tail is a final common step in integrin activation; however, the role * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. of talin binding to integrin ␤ cytoplasmic tails in outside-in signaling is less clear.
Relationships between inside-out and outside-in integrin signaling remain uncertain. Structural and immunochemical studies have suggested that the similar conformational rearrangements in the extracellular domain might lead to both forms of signaling (29,30). In contrast, truncation of the last 3 residues of the ␤3 cytoplasmic domain can inhibit cell spreading, with less of an effect on integrin activation (31), suggesting that the two processes might be separable, at least at the cytoplasmic face of the integrin. Furthermore, talin 1 deficiency impacts the force-dependent reinforcement of initial integrincytoskeleton bonds with little effect on cell spreading or SFK activation (32). These results suggested that the interaction of SFKs and talin with integrin ␤ tails might have separable roles in integrin signaling in the two directions.
Here we report that amino acid exchanges affecting a combination of an Arg and Thr residue in the integrin ␤3 tail control the binding specificity for SFKs but have no effect on talin binding. Using mutants at these residues and cells engineered to express only c-Src among SFKs, we find that SFK binding to integrin ␤ tails is not required for inside-out integrin signaling but is obligatory for cell spreading. Conversely, we show that point mutations that disrupt talin binding abolish integrin activation, but they do not inhibit SFK binding to the ␤3 tail or the initiation of outside-in signaling by SFKs. Thus, we provide decisive evidence that inside-out and outside-in integrin signaling are biochemically distinct.

MATERIALS AND METHODS
Antibodies and Reagents-Antibodies against the integrin ␤3 subunit (SSA6 and 8053) were described previously (15). Monoclonal antibody 327 is specific for the c-Src SH3 domain. Monoclonal antibody 7E2 to the ␤1 integrin subunit was from Rudolph Juliano (33). Polyclonal antibody 828 is specific for the ␤2 integrin subunit (34). Human ␣IIb␤3specific monoclonal antibody, D57, and activating anti-␤3 antibody, anti-LIBS6, have been described previously (35)(36)(37). Monoclonal antibody 8d4 to talin was from Sigma. Antibodies to His 6 , GST (B14), Fyn, and Lyn were from Santa Cruz Biotechnology. Antibody to c-Yes was from Transduction Laboratories. Horseradish peroxidase-conjugated secondary antibodies were from Bio-Rad. Monoclonal antibody, PY20, to phosphotyrosine was from BD Biosciences. Purified human fibrinogen was from Enzyme Research Laboratories, Inc. Rhodamine-phalloidin was from Molecular Probes. Protein A-Sepharose was from Amersham Biosciences. All other reagents were obtained from Sigma.
Cell Lines, Plasmids, and Transfections-Chinese hamster ovary cell lines that stably express human wild-type ␣IIb␤3 have been described previously (38). SYF cells (mouse embryonic fibroblasts deficient in c-Src, Fyn, and c-Yes) and SYFϩc-Src cells (reconstituted with c-Src) (19) were obtained from the American Type Culture Collection. Cells were maintained in Dulbecco's modified Eagle's medium supplemented with 10% fetal bovine serum, L-glutamine, and antibiotics at 37°C with 6% CO 2 . Transient transfections were carried out with Lipofectamine Plus (Invitrogen) as described by the manufacturer. After 24 h, cells were serum-starved in 0.5% fetal bovine serum and cultured for an additional 24 h before use in functional assays.
Immunoprecipitation-Transfected cells were lysed in Nonidet P-40 buffer containing 1% Nonidet P-40, 150 mM NaCl, 50 mM Tris (pH 7.4), 1 mM sodium vanadate, 0.5 mM sodium fluoride, and complete protease inhibitor mixture (Roche Applied Science). Lysates were clarified by centrifugation at 13,000 ϫ g for 10 min at 4°C, and 500 -800 g of protein from the soluble fraction of cell lysates were immunoprecipitated with the indicated antibodies to the ␤1, ␤2, or ␤3 integrin sub-units and protein A-Sepharose beads. Immunoprecipitates were subjected to SDS-PAGE and Western blotting with the indicated antibodies. Immunoreactive bands were detected by chemiluminescence (Pierce).
Affinity Chromatography-Recombinant integrin cytoplasmic tail model proteins containing both a His 6 and an in vivo biotinylation peptide at the N terminus were expressed and purified from Escherichia coli extracts as previously described (15). A scrambled sequence preserving the same amino acid composition of the integrin ␤3 tail sequence (KLA-TNEPTTAFIELGKHRTREKITNSYRWFTAEREFKILDATNKYA) was produced in the same manner and used as control for nonspecific binding. Affinity chromatography of cell lysates was performed using integrin tail model proteins immobilized on neutravidin-agarose (Pierce) as previously described (15).
Enzyme-linked Immunosorbent Assay-Biotinylated recombinant integrin tails (200 ng) were bound at saturating concentrations (20 g/ml) to wells of neutravidin-coated microtiter plates (Pierce). GST-SH3 domains of c-Src, Lyn, Fyn, and c-Yes were expressed in E. coli strain BL21(DE3)pLys (Novagen, Inc.). Increasing concentrations of GST-c-Src SH3 (in 100 l) were added to wells in triplicate and incubated for 1 h at room temperature. After three washes with phosphate-buffered saline-0.05% Tween 20 (PBS-T), mouse anti-GST antibody (0.4 g/ml) was added and incubated for an additional hour, followed by another three washes with PBS-T and a final 1-h incubation with horseradish peroxidase-conjugated anti-mouse immunoglobulin before wash and color development with -phenylenediamene substrate. Absorbance was measured at 490 nm. For competition assays, increasing concentrations of GST-c-src-SH3 were incubated with immobilized ␤3 tail in the presence or absence of 1 M GST-talin F23 followed by incubation with mouse monoclonal antibody 327 (anti-c-Src-SH3). Loading density of the various cytoplasmic tail constructs onto neutravidin-coated plates was determined using mouse monoclonal anti-His 6 . Binding of GST-c-Src-SH3 to the ␤3 tail was determined using GraphPad Prism Software. Non-saturable binding was estimated by performing parallel binding curves with randomized ␤3 or by competition with an excess (150 M) of peptide KGGRSLRPLPPLPPPG previously found to abrogate the binding of c-Src-SH3 to ␤3 (15). After subtraction of the non-saturable binding component, binding data were fit to a single site binding model in which Y ϭ B max (X)/(K d ϩ X) by non-linear regression. Results were normalized to the percentage of maximal binding to account for fluctuations in maximal absorbance signal at saturation between experiments.
Enzyme Activity of Purified c-Src-Purified active human c-Src (5 units/reaction; Upstate Biotechnology) was incubated with ␤ cytoplasmic tail model proteins (20 g) bound to neutravidin beads (Pierce) in the presence of 150 M Src-specific peptide substrate. Reactions were carried out using a Src kinase assay kit (Upstate Biotechnology). Briefly, tyrosine phosphorylation of substrate peptide KVEKIGEG-TYGVVYK by recombinant c-Src was measured by incorporation of [␥-32 P]ATP in the presence or absence of beads. In the cases where the effect of talin F23 or c-Src-SH3 was measured, beads containing ␤3 tail were pre-incubated with 1 M GST-talin F23 or 20 M GST-c-Src-SH3 for 20 min prior to the addition of c-Src. Activity was calculated as the percentage increase over basal activity of c-Src in the absence of beads. SDS-PAGE gels of 5 l of beads were run to monitor the tail content of the beads.
Integrin Activation-SYFϩc-Src cells were transiently cotransfected with plasmids encoding ␣IIb␣5, native or mutant ␤3, and enhanced GFP (as a marker of transfection). After 72 h, binding of the activationspecific monoclonal IgM antibody PAC1 was assessed by flow cytometry as previously described (41). PAC1 binding was only analyzed on a gated subset of live (propidium iodide-negative) and transfected (GFPpositive) cells. Integrin surface expression levels of each transfection were analyzed with ␣IIb␤3-specific monoclonal antibody D57.
Samples were incubated in presence or absence of 200 g/ml activating anti-␤3 antibody anti-LIBS6 or 10 M ␣IIb␤3 antagonist Ro44-9883 (42). PAC1 binding in the presence of Ro44-9883 was used to estimate nonspecific binding, and treatment with anti-LIBS6 was used to calculate maximal PAC1 binding. To obtain numerical estimates of integrin activation, we calculated an activation index defined as follows: 100 ϫ (F Ϫ F 0 )/(F max Ϫ F 0 ), where F is the mean fluorescence intensity of PAC1 binding, F 0 is the mean fluorescence intensity of PAC1 binding in the presence of Ro44-9883, and F max is the mean fluorescence intensity of PAC1 binding in the presence of anti-LIBS6.
Cell Spreading-Coverslips were coated with 5 g/ml fibrinogen overnight at 4°C, washed twice with phosphate-buffered saline, and blocked with heat-denatured bovine serum albumin (5 mg/ml) for 1 h at room temperature. For the spreading experiments, transfected cells FIG. 1. Mapping of the Src-binding sequences in the ␤3 cytoplasmic tail. A, amino acid sequence of the ␤3 and ␤1A cytoplasmic tails, including the scrambled ␤3 sequence used as control for nonspecific binding (random ␤3), and representation of the recombinant exchanges made between ␤3 and ␤1A cytoplasmic tails. Residues from ␤3 are indicated by gray boxes, and residues from ␤1 are indicated by white boxes. B, pull-down analysis from platelet lysates using native and chimeric ␤ cytoplasmic tails immobilized on neutravidin-agarose. Binding of c-Src was assayed by Western blot. Monoclonal anti-His 6 was used to check for tail loading. C, binding curves of GST-c-Src-SH3 to neutravidin-bound biotinylated ␤3 and ␤1A cytoplasmic tail and their respective C-terminal exchange chimeras (20 g/ml) on microtiter plates, developed with anti-GST antibodies. Results were normalized to the percentage of maximal binding after subtraction of nonspecific binding to randomized ␤3 cytoplasmic tail. D, effect of the ␤ cytoplasmic tail type on relative Src activity. Results were expressed as the percentage increase in activity relative to basal Src activity in the absence of ␤3 to permit pooling of data from three independent experiments.
were resuspended at 10 6 cells/ml in Tyrode's buffer and incubated on coated coverslips at 37°C for 15-60 min. After removal of unbound cells by washing with phosphate-buffered saline, adherent cells were fixed with 3.7% formaldehyde and permeabilized with 0.2% Triton X-100. Platelets were stained with rhodamine-phalloidin and antibody PY20 followed by a fluorescein isothiocyanate-conjugated anti-mouse IgG secondary antibody. Fluorescence images were acquired with a laser scanning confocal microscope (MRC 1024; Bio-Rad). The surface areas of transfected cells (GFP-positive) were measured using Image Pro Plus software (Media Cybernetics, Inc.).

c-Src and Talin Bind to Distinct Sites in the ␤3
Cytoplasmic Domain-We sought to discern the relative biological roles of c-Src and talin binding to integrin ␤ cytoplasmic domains. c-Src binding to integrin ␤3 involves the last 14 residues of ␤3 (15), a region partially overlapping with the talin binding site (26,27). To identify ␤ tail mutants selectively defective in c-Src binding, we utilized a previously described (43) homology scanning approach by creating chimeras between the ␤3 and ␤1A cytoplasmic tails (Fig. 1A). We had previously found that c-Src within platelet lysates binds to the ␤3 cytoplasmic tail, but not to the ␤1A or ␤2 tails, and that binding to the ␤3 tail is abrogated upon elimination of the 4 C-terminal amino acid residues (15). Upon exchanging the C-terminal 3 residues between ␤1A and ␤3 (Fig.  1B), we observed that whereas the ␤3 tail with the ␤1A substitution at positions 760 -762 (␤3␤1(C3R)) lost its ability to bind c-Src, ␤1A containing the C-terminal 3 residues of ␤3 (RGT) (␤1␤3(C3R)) exhibited a gain of binding toward c-Src (Fig. 1B). We previously found that the related SFKs Yes, Hck, and Lyn bound to ␤1 and ␤3 (15); each of these kinases bound to all of the chimeras shown in Fig. 1B. In contrast, Fyn, which is ␤3-specific (15), exhibited an identical binding pattern to c-Src (data not shown). Individual ␤3 mutations, R760E or T762K, were incapable of suppressing c-Src binding to the ␤3 tail, and individual ␤1A mutations (E760R or K762T) were unable to promote c-Src binding to the ␤1A tail (data not shown). To routinely control for nonspecific binding, a randomized sequence version of the ␤3 cytoplasmic tail (r␤3) failed to interact with c-Src (Fig. 2B) or any of the other SFKs (data not shown). Taken together, these results indicate that the binding of c-Src is dependent on the combined presence of the Arg and Thr residues of the C-terminal ␤3 RGT motif.
Because c-Src interacts with ␤3 via its SH3 domain (15), we examined the effect of integrin ␤ tail exchange on the binding of recombinant c-Src-SH3 to immobilized native or chimeric ␤1A and ␤3 tails (Fig. 1C). The ␤1␤3(C3R) chimera bound c-Src SH3 with an EC 50 similar to the native ␤3 tail (ϳ5-10 M). Conversely, the ␤3␤1(C3R) chimera showed complete loss of saturable binding to c-Src-SH3. Interaction of fulllength c-Src with ␤3 activates c-Src (15). Similarly, the ␤1␤3(C3R) chimera gained the capacity to activate c-Src in FIG. 2. Talin and c-Src bind to different regions of the ␤3 tail. A, enzyme-linked immunosorbent assay binding curve for GST-talin-F23 and GST-c-Src-SH3 using neutravidin-bound biotinylated native or mutant ␤3 cytoplasmic tails (20 g/ml). Binding was detected using mouse anti-GST antibody, followed by horseradish peroxidase-conjugated anti-mouse immunoglobulin. B, pull-down analysis from platelet lysates using native and mutant ␤3 cytoplasmic tails. Western blot was performed using monoclonal antibodies against c-Src and talin. Monoclonal anti-His 6 was used to check for tail loading.
vitro, whereas the ␤3␤1(C3R) chimera lost this capacity (Fig.  1D). Thus, the last 3 residues of ␤3 and ␤1 control binding to c-Src and the capacity of the integrin tail to regulate c-Src activity.
Having established that an exchange of 2 amino acid residues could regulate c-Src binding to and activation by integrins, we assessed the effects of this exchange on the binding of talin, which interacts with integrin ␤ tails principally through its N-terminal F23 subdomain (44). Whereas exchange of the last 3 residues of ␤1A with those of ␤3 (␤3␤1(C3R)) markedly reduced c-Src binding, it had no effect on the binding of the F23 fragment of talin ( Fig. 2A) or full-length talin (Fig. 2B) to the ␤3 tail. Conversely, ␤3 mutations known to abolish talin binding (␤3(W739A), ␤3(L746A/K748A)) (16, 26) had no effect on the binding of c-Src SH3 ( Fig. 2A) or full-length c-Src (Fig. 2B) to the integrin ␤3 tail. Moreover, saturating concentrations of the talin F23 fragment (26,44) failed to block the binding of the c-Src SH3 domain (Fig. 3A) to ␤3 or the capacity of ␤3 to activate c-Src (Fig. 3B). Thus, the interaction sites in ␤3 integrins for talin and c-Src are distinct, and talin does not compete with c-Src for binding to the integrin.
Certain Src Family Kinases Bind to Integrin ␤1 Cytoplasmic Tails-The foregoing results establish that exchange of the last 3 residues of ␤3 and ␤1 can regulate the binding and activation of c-Src. Most cells express multiple Src family kinases, and we previously found in pull-down studies with platelet lysates (15) that Lyn and c-Yes kinases are capable of binding to integrin ␤1A, ␤2, and ␤3 tails. We now find that the purified SH3 domains of c-Yes and Lyn can bind to ␤1A, ␤2, and ␤3 tails (Fig.  4A), whereas Fyn binds only to ␤3. Furthermore, native Lyn and c-Yes associated with heterodimeric integrins bearing the ␤1 and ␤2 tails in cells (Fig. 4B). These results suggest that although exchange of the last 3 C-terminal residues between ␤1A and ␤3 can regulate the binding of c-Src, it would not change the binding of c-Yes or Lyn. Indeed, this proved to be the case (Supplementary Fig. 1, A and B).
Biological Role of the c-Src-Integrin Interaction-To analyze the functional importance of the c-Src-integrin interaction, we sought a cellular system in which c-Src was the only SFK expressed. Klinghoffer et al. (19) have derived fibroblasts from c-Src, Fyn, and c-Yes triple knock-out embryos and showed that integrin signaling and cell spreading are dramatically depressed in these cells (SYF cells). They have also reconstituted these cells with c-Src (SYFϩc-Src cells). Immunoblotting confirmed that all three kinases were absent from SYF cells and that c-Src was the only one present in the SYFϩc-Src cells (data not shown). The SYFϩc-Src cells were then transiently transfected with the ␣IIb subunit in combination with integrin ␤3 or with the talin binding-defective ␤3(L746A/K748A) mutant. Alternatively, the cells were transfected with ␣IIb and chimeras of ␤3 containing the ␤3 extracellular and transmembrane domains joined to either the ␤1 cytoplasmic domain (34,45) or one of the chimeric cytoplasmic domains described above, ␤3␤1(C3R) or ␤1␤3(C3R). c-Src was present in ␤3 immunoprecipitates when either the wild-type ␤3, ␤3(L746A/K748A), or ␤1␤3(C3R) chimeric tail was present (Fig. 4C). In contrast, no c-Src was associated with integrins bearing either the ␤1A tail or the ␤3␤1(C3R) tail. Thus, these various SYFϩc-Src cells provide us with a tool to examine the biological role of the c-Src-integrin interaction.
The ␤3 Integrin-Talin Interaction, but Not the ␤3-c-Src Interaction, Is Required for Integrin Activation-Integrin activation is regulated by the interaction of talin with the integrin ␤ tail. To examine the relative roles of talin and c-Src binding to the integrin ␤ tail on integrin activation, we transiently transfected SYFϩc-Src cells with plasmids encoding the integrin ␤3 subunit, a talin binding-deficient mutant (␤3(L746A/K748A)), or a c-Src binding-deficient mutant (␤3␤1(C3R)). To produce an activated integrin, we cotransfected a chimeric integrin ␣ subunit in which the extracellular and transmembrane domains of ␣IIb were joined to the cytoplasmic domain of ␣5 (41). Integrin activation was assessed with the ␣IIb␤3 activation-specific monoclonal antibody, PAC1 (46). The c-Src binding-deficient mutant supported specific PAC1 binding to the same extent as wild-type ␣IIb␤3 (activation index, 12.2 and 12.3, respectively) (Fig. 5). In sharp contrast and as expected, the talin binding-deficient ␣IIb␤3(L746A/K746A) mutant failed to support specific PAC1 binding. Importantly, all three integrins were equally well expressed, as judged by the binding of an ␣IIb␤3-specific antibody, D57 (data not shown). Thus, whereas talin binding is required, c-Src binding is dispensable for integrin activation.
The ␤3 Integrin-c-Src Interaction, but Not the ␤3-Talin Interaction, Is Required for Integrin-mediated Cell Spreading-Integrin-mediated adhesion initiates the SFK-dependent spreading of cells. We exploited the fact that ␤3 integrins, but not the widely expressed ␤1 integrins, mediate cell spreading on fibrinogen-coated surfaces (47). SYFϩc-Src cells failed to spread on fibrinogen but did spread when transiently transfected with cDNAs encoding the ␤3 integrin subunit (Fig. 6). Importantly, ␤3 transfection failed to induce spreading of the SYF cells (data not shown), confirming that the spreading was both ␤3and c-Src-dependent. Transfection of SYFϩc-Src cells with the ␤3␤1 chimera or the ␤3␤1(C3R) chimera failed to induce cell spreading beyond that seen in untransfected cells, indicating that c-Src binding to the integrin tail was required for spreading. Conversely, the gain of c-Src binding function in the ␤1␤3(C3R) chimera enabled that integrin subunit to mediate cell spreading (Fig. 6).
To quantify the results depicted in Fig. 6, the cell areas for 50 -100 randomly sampled positively transfected cells were measured using Image Pro Plus software. The ␤3 integrins that promoted visual spreading (␤3, ␤1␤3(C3R)) on average resulted in a 2-fold increase in cell area relative to the chimeras that failed to promote visual spreading (␤1, ␤3␤1(C3R)) (Fig. 7). Thus, SFK binding to the integrin ␤ tail is required for and mediates cell spreading.
To analyze the role of talin binding to the ␤3 tail in cell spreading, we examined the behavior of cells transfected with ␤3(L746A/K748A) upon attachment to fibrinogen. As expected from the fact that deletion of the ␤3 tail only modestly inhibits adhesion to fibrinogen (47), cells bearing this mutant adhered; FIG. 5. Interaction between c-Src and integrin ␤3 tail is not required for integrin activation in vivo. c-Src-expressing SYF cells were cotransfected with plasmids encoding GFP, ␣IIb␣5, and native or mutant ␤3 subunit, as indicated. Binding of the activation-specific anti-␣IIb␤3 monoclonal antibody PAC1 was assessed by flow cytometry 48 h later. Dot plots represent the degree of transfection (GFP fluorescence; vertical) and PAC1 binding (horizontal). PAC1 binding was determined in the presence and absence of the ligand mimetic Ro44-9883 (10 M) to demonstrate specificity. The depicted activation indices were calculated as described under "Materials and Methods." The subpopulation of PAC1-negative cells expressing ␣IIb␣5 in combination with ␤3 probably represents cells expressing the GFP transfection marker, but not the integrin.
B, Chinese hamster ovary cells stably expressing ␣IIb␤3 were cotransfected with ␣M, ␤2, and the indicated Src family kinase. Immunoprecipitates of ␤ integrins were probed on Western blot with antibodies to c-Src, Fyn, c-Yes, and Lyn. Immunoprecipitation with normal IgG was used as negative immunoblotting control. C, SYF cells or c-Src-expressing SYF cells (SYFϩc-Src) were transfected with ␣IIb and wild-type ␤3, mutant ␤3(L746A/K748A), or C-terminal exchange ␤3␤1 chimeric constructs. These constructs consist of the indicated recombinant ␤ cytoplasmic domains (as described in Fig. 1A) fused to the extracellular and transmembrane domain of human integrin ␤3. Cell lysates were immunoprecipitated with a polyclonal antibody to ␤3 (antibody 8053) and immunoblotted as indicated. however, they failed to spread as judged visually and by quantitative image analysis (Fig. 7A). We hypothesized that the lack of activation of the ␤3 integrin could lead to reduced engagement of ␣IIb␤3, resulting in a secondary inhibition of cell spreading. Indeed, addition of the activating monoclonal antibody, anti-anti-LIBS6 (35,37), restored cell spreading in ␤3(L746A/K748A)expressing cells (Fig. 7B). By contrast, even in the presence of anti-anti-LIBS6, ␤3␤1(C3R)-expressing cells spread poorly on fibrinogen, indicating that the reduced spreading of these cells was not due to a defect in integrin activation. These results establish that talin contributes to cell spreading by activating the integrin, but once the integrin is activated, talin binding to the integrin ␤ tail is no longer required.
Perspective and Conclusions-The studies described here provide unambiguous evidence that inside-out and outside-in integrin signaling are mediated by distinct and separable interactions of the integrin cytoplasmic domains. Talin binding is required for integrin activation, and the binding of c-Src kinases is required for the signaling events that lead to cell spreading. Thus, it is possible to exert precise mutational control over the vectorial direction of integrin signaling. Furthermore, previous studies have suggested that integrin activation could precede and regulate the ability of these receptors to generate intracellular signals (48,49). The present studies directly establish this point by showing that the spreading defect caused by loss of talin binding (and hence, loss of activation) can be rescued by antibody-mediated increase in ligand binding affinity. This result also provides additional evidence that cells can undergo spreading even when talin binding to the integrin tail is blocked by mutation (present studies) or by lack of talin (32). Many studies have examined integrin signaling in development, the immune response, hemostasis, and a wide variety of pathological states. The present studies will provide a paradigm to dissect the relative contributions of the direction of integrin signaling on these processes in cell culture and, ultimately, in vivo. FIG. 6. Interaction between c-Src and integrin ␤3 tail is required for cell spreading on fibrinogen. SYF cells or c-Src-expressing SYF cells (ϩ c-Src) were cotransfected with GFP (as a marker for transfection; arrows) and native ␤3 or C-terminal exchange chimeras, as indicated. Cells were plated on fibrinogen-coated coverslips (5 g/ml) at 37°C. After 30 min, adherent cells were fixed, permeabilized, and stained with rhodamine-phalloidin (F-Actin; red) and monoclonal antibody PY20 (P-Tyr; blue). Cell images were analyzed by confocal microscopy. The figure contains representative images of four independent experiments. Bars, 50 m.  Ն 60). B, c-Src-expressing SYF cells were transfected with native or mutant ␤3, as indicated. Cells were plated on fibrinogencoated coverslips at 37°C for 30 min in the presence or absence of the activating anti-␤3 antibody anti-LIBS6 (200 g/ml). Cell area was analyzed as described in A.