The role of conserved amino acid motifs within the integrin beta3 cytoplasmic domain in triggering focal adhesion kinase phosphorylation.

Integrin-mediated adhesion of cells to extracellular matrix proteins triggers a variety of intracellular signaling pathways including a cascade of tyrosine phosphorylations. In many cell types, the cytoplasmic focal adhesion tyrosine kinase, FAK, appears to be the initial protein that becomes tyrosine-phosphorylated in response to adhesion; however, the molecular mechanisms regulating integrin-triggered FAK phosphorylation are not understood. Previous studies have shown that the integrin β1, β3, and β5 subunit cytoplasmic domains all contain sufficient information to trigger FAK phosphorylation when expressed in single-subunit chimeric receptors connected to an extracellular reporter. In the present study, β3 cytoplasmic domain deletion and substitution mutants were constructed to identify amino acids within the integrin β3 cytoplasmic domain that regulate its ability to trigger FAK phosphorylation. Cells transiently expressing chimeric receptors containing these mutant cytoplasmic domains were magnetically sorted and assayed for the tyrosine phosphorylation of FAK. Analysis of these mutants indicated that structural information in both the membrane-proximal and C-terminal segments of the β3 cytoplasmic domain is important for triggering FAK phosphorylation. In the C-terminal segment of the β3 cytoplasmic domain, the highly conserved NPXY motif was found to be required for the β3 cytoplasmic domain to trigger FAK phosphorylation. However, the putative FAK binding domain within the N-terminal segment of the β3 cytoplasmic domain was found to be neither required nor sufficient for this signaling event. We also demonstrate that the serine 752 to proline mutation, known to cause a variant of Glanzmann's thrombasthenia, inhibits the ability of the β3 cytoplasmic domain to signal FAK phosphorylation, suggesting that a single mutation in the β3 cytoplasmic domain can inhibit both “inside-out” and “outside-in” integrin signaling.

Integrins are a family of ␣/␤ heterodimeric adhesion receptors used by cells to interact with their extracellular matrix (1). In addition to mediating cell adhesion, integrins also function as signaling receptors, integrating information from the extra-cellular environment (2)(3)(4)(5). Integrin engagement triggers intracellular signals that direct cell adhesion and regulate other aspects of cell behavior including cell proliferation, differentiation, and survival. A major integrin-triggered signal is a cascade of tyrosine phosphorylations. In many cell types, an early integrin-triggered event is the tyrosine phosphorylation and activation of the focal adhesion kinase, FAK 1 (2)(3)(4)(5), which is followed by the tyrosine phosphorylation of additional cytoskeletal and signaling proteins, including paxillin, tensin, MAP kinase, cortactin, and p130 CAS (6 -12). These phosphorylation events are likely to be central to the assembly of adhesion-dependent signaling complexes that regulate cell adhesion and other aspects of cell behavior (2)(3)(4)(5).
The integrin-triggered tyrosine phosphorylation of FAK induces the formation of specific adhesion-dependent signaling complexes by creating binding sites for SH2 domains of other signaling molecules, thus providing pathways linking integrin engagement with downstream signaling events. Integrin engagement triggers the autophosphorylation of FAK, which in turn generates a binding site for the SH2 domain of Src family kinases (13,14). These interactions result in the further phosphorylation and activation of FAK (15) and have recently been shown to be required for the function of FAK in modulating the migratory behavior of cells (16), as well as the ability of constitutively activated FAK to inhibit apoptosis in some circumstances and function as a transforming agent in others (17). The phosphorylation of FAK at other tyrosine residues provides binding sites for additional SH2-containing proteins, such as phosphatidylinositol 3-kinase (18) and the adapter protein, GRB2 (19), both of which are important in growth factor signaling pathways. These phosphorylation events may therefore serve to link integrin engagement to downstream signaling pathways including the Ras and MAP kinase cascades known to become activated by integrin-mediated adhesion (8,20,(22)(23)(24).
Since integrins have no intrinsic enzymatic activity, they must depend on their association with other cellular proteins to initiate intracellular signals. Integrin ␤ subunit cytoplasmic domains are likely to serve as the major link between integrins and the cytoskeleton and signal transduction apparatus of the cell (25). ␤ subunit cytoplasmic domains have been demonstrated to be required for many integrin-mediated processes, including cell adhesion, cell spreading, cell migration, and fibronectin matrix assembly, as well as to signal FAK phosphorylation (26 -30). Additionally, cytoplasmic domains of specific integrin ␤ subunits expressed in the context of single-subunit chimeric receptors are sufficient to both mimic and inhibit various functions of ligand-occupied integrins (31)(32)(33)(34)(35)(36), further supporting the notion that ␤ cytoplasmic domains interact with specific cytoplasmic proteins required for endogenous integrin function.
Chimeric receptors containing the integrin ␤ 1 , ␤ 3 , or ␤ 5 cytoplasmic domain connected to a reporter consisting of either the IL-2 or CD4 receptor extracellular domain were previously utilized to demonstrate that ␤ subunit cytoplasmic domains contain sufficient information to activate the FAK signaling pathway (34,35). Furthermore, this was found to be dependent upon their primary structure, since the integrin ␤ 1B and ␤ 3B subunit cytoplasmic domains, which are modified by alternative splicing, did not signal FAK phosphorylation (34,37). In the present study, we have used single-subunit chimeric receptors to analyze the role of specific amino acids within the ␤ 3 cytoplasmic domain in signaling FAK phosphorylation. To do this, we constructed chimeric receptors containing the extracellular and transmembrane domains of the IL-2 receptor connected to mutant ␤ 3 cytoplasmic domains with specific amino acid deletions or substitutions (Fig. 1). We found that amino acids within both the N-terminal and the C-terminal segment of the ␤ 3 cytoplasmic domain were necessary to signal increases in the tyrosine phosphorylation of FAK, and that the putative FAK binding domain within the N-terminal segment of the ␤ 3 cytoplasmic domain was neither sufficient nor necessary to signal FAK phosphorylation. We have further defined an important role for the conserved NPXY motif in this signal transduction event.
Tyrosine Phosphorylation Assay-Human fibroblasts, transiently expressing the chimeric receptors, were harvested 16 -18 h after transfection. Positively transfected cells were sorted with magnetic beads conjugated with goat anti-mouse IgG (Advanced Magnets, Cambridge, MA) and mAb 7G7B6 to the IL-2 receptor (Upstate Biotechnology Inc., Lake Placid, NY) as described previously (39). Magnetically sorted cells were lysed with a buffer containing 50 mM Tris (pH 7.4), 150 mM NaC1, 1% Nonidet P-40, 0.25% sodium deoxycholate, 1 mM EDTA, 1 mM sodium vanadate, 1 mM sodium fluoride, 10 g/ml aprotinin, 10 g/ml leupeptin, and 1 mM phenylmethylsulfonyl fluoride. Protein concentrations were determined using the Micro-BCA assay (Pierce). Cell lysates (15 g) were resolved under reducing conditions by 10% SDS-PAGE. Tyrosine phosphorylation was analyzed by Western blotting using the mAb 4G10 to phosphotyrosine (Upstate Biotechnology Inc.) and visualized by enhanced chemiluminescence (Amersham). The filters were then stripped by incubation at 75°C for 40 min in stripping buffer containing 2% SDS, 62.5 mM Tris-HCl, pH 6.8, and 100 mM ␤-mercaptoethanol, and the level of FAK present in each sample was assayed by reprobing the filter with mAb to FAK (Transduction Laboratories, Lexington, KY). Relative amounts of FAK and phosphorylated FAK in each lysate were determined by densitometry using a Millipore Bioimage 60S video densitometer utilizing Visage software. For this purpose, exposures were chosen so that the signal from each lysate was in the linear range of the film. Expression levels of the chimeric receptors were analyzed in each experiment by flow cytometry using a Becton Dickinson FACScan flow cytometer as described previously (32). In some experiments, magnetically sorted cells were lysed with a 1% Nonidet P-40 buffer (40), and equal amounts of lysates (300 g) were immunoprecipitated by incubation with polyclonal antibodies to FAK. Immune complexes were collected with Protein A-Sepharose and then analyzed under reducing conditions by 10% SDS-PAGE. Tyrosine phosphorylation was then visualized by Western blotting as described above.

RESULTS
To identify the amino acids within the ␤ 3 cytoplasmic domain required to signal increases in FAK phosphorylation, we made a series of amino acid deletions and substitutions in the ␤ 3 cytoplasmic domain ( Fig. 1) based on three criteria: (a) the identity of amino acid residues important for FAK binding in vitro (41), (b) the identity of the amino acids replaced by alternative splicing of ␤ 1B and ␤ 3B (25), and (c) the regions of homology in the C-terminal regions of the ␤ 1 , ␤ 3 , and ␤ 5 cytoplasmic domains known to contain sufficient information to trigger FAK phosphorylation (25,34). These ␤ 3 cytoplasmic mutants were expressed in the context of single-subunit chimeric receptors as described previously (31,32). Their ability to signal FAK phosphorylation was assayed using a magnetic sorting procedure that not only enriched for positively express- ing cells but also triggered FAK phosphorylation in a manner dependent upon the presence of the integrin ␤ cytoplasmic domain (34).
Elements in Both the N-terminal and C-terminal Segments of the ␤ 3 Cytoplasmic Domain Are Required to Trigger FAK Phosphorylation-Differences in the abilities of the ␤ 3 and ␤ 3B cytoplasmic domains to trigger FAK phosphorylation suggested either that the C-terminal amino acids of the ␤ 3 cytoplasmic domain are required for signaling FAK phosphorylation or that the C-terminal amino acids of the ␤ 3B cytoplasmic domain inhibit this signaling event. To distinguish between these possibilities, three deletion mutants were constructed: ␤ 3 -d742-762, ␤ 3 -d728 -762, and ␤ 3 -d718 -741 ( Fig. 1). Chimeric receptors containing these mutant cytoplasmic domains or the wildtype ␤ 3 cytoplasmic domain were transiently expressed in normal human fibroblasts and then assayed for their ability to trigger FAK phosphorylation (Fig. 2). Increases in tyrosine phosphorylation of FAK were observed in lysates from cells expressing chimeric receptors containing the wild-type ␤ 3 cytoplasmic domain, whereas cells expressing a chimeric receptor lacking the entire ␤ 3 cytoplasmic domain did not show this increase in phosphorylation in agreement with previously published results (34). The ␤ 3 deletion mutant, ␤ 3 -d742-762, which lacked the C-terminal amino acids that differ between the ␤ 3 and ␤ 3B cytoplasmic domains, consistently did not signal FAK phosphorylation above background levels ( Figs. 2 and 6). This result suggests that the C-terminal region of ␤ 3B does not play a negative role, but rather that amino acids in the C-terminal region of ␤ 3 are required to signal FAK phosphorylation. The deletion mutant, ␤ 3 -d728 -762, containing the first 12 membrane-proximal amino acids of the ␤ 3 cytoplasmic domain reported to bind to FAK in vitro (41), also was not sufficient to signal increases in tyrosine phosphorylation of FAK above background levels (Figs. 2 and 6), further suggesting that additional C-terminal amino acids are required for this signaling event. Similarly, chimeric receptors containing the deletion mutant, ␤ 3 -d718 -741, which contained only the 21 C-terminal amino acids of ␤ 3 that are replaced in ␤ 3B , also did not signal FAK phosphorylation, indicating that these amino acids are not sufficient to signal FAK phosphorylation (Figs. 2 and 6). These results suggest that the sequence information in both the N-terminal and C-terminal regions is required for the ␤ 3 cytoplasmic domain to trigger FAK phosphorylation. Although the level of expression of each chimeric receptor varied from each other and from one experiment to the next, differences in the ability of various chimeric receptors to trigger FAK phosphorylation were not found to be due to these differences in expression levels.
The Level of Expression of Chimeric Receptors Versus Their Ability to Signal FAK Phosphorylation-We also directly tested whether differences in the level of expression of individual chimeric receptors affected their ability to signal FAK phosphorylation. Since the amount of DNA used for transfection directly influenced the average level of transfected gene expression in individual cells, as well as the total number of positively expressing cells observed after transfection, we transfected cells with either 15 or 40 g of plasmid DNA encoding the chimeric receptors containing either the wild-type ␤ 3 cytoplasmic domain or the deletion mutant, ␤ 3 -d728 -762. We analyzed chimeric receptor expression levels in these different transfected cell populations by flow cytometry. Transfection with either 15 or 40 g of DNA encoding the chimeric receptor containing the wild-type ␤ 3 cytoplasmic domain generated populations of cells with mean fluorescence intensities of 42.9 and 135.99, respectively. Similarly, transfection with 15 or 40 g of DNA encoding the ␤ 3 -d728 -762 deletion mutant generated populations of cells with mean fluorescence intensities of 124.59 and 352.99, respectively. When we assayed these different populations of cells for FAK phosphorylation, we consistently found that the wild-type ␤ 3 cytoplasmic domain signaled FAK phosphorylation when expressed at either higher or lower levels, whereas the deletion mutant ␤ 3 -d728 -762 was consistently unable to signal tyrosine phosphorylation of FAK, even at higher levels of expression (Fig. 3).
The Conserved Acidic Residues Asp and Glu in the N-terminal Segment of the ␤ 3 Cytoplasmic Domain Are Not Required to Signal FAK Phosphorylation-The chimeric receptor ␤ 3 -d728 -762, which contains the putative FAK binding domain, was unable to signal FAK phosphorylation, indicating that in some circumstances the putative FAK binding is not sufficient to trigger FAK phosphorylation (Fig. 2). To determine whether the putative FAK binding domain was required for the ␤ 3 cytoplasmic domain to signal FAK phosphorylation, we also generated a chimeric receptor, ␤ 3 -723,726(*/A), that contained a mutant ␤ 3 cytoplasmic domain with alanine substitutions at the acidic residues, aspartic acid (Asp-723) and glutamic acid (Glu-726). Previous studies had shown that these specific mutations inhibit FAK binding in vitro to peptides modeled after the membrane-proximal region of ␤ cytoplasmic domains (41). Interestingly, chimeric receptors containing these alanine substitutions consistently triggered FAK phosphorylation to levels similar to those induced by chimeras containing the wild-type ␤ 3 cytoplasmic domain (Fig. 4, A and B, and Fig. 6). These results indicate that the putative FAK binding domain is nei- ther sufficient nor required for the ␤ 3 cytoplasmic domain to trigger FAK phosphorylation.
The ability of the substitution mutant, ␤ 3 -723,726(*/A), to signal FAK phosphorylation and the inability of the deletion mutant, ␤ 3 -d728 -762, containing only the putative FAK binding site to signal FAK phosphorylation was also examined by immunoprecipitation. Again, chimeric receptors containing either the wild-type ␤ 3 cytoplasmic domain or the substitution mutant, ␤ 3 -723,726(*/A), were able to signal FAK phosphorylation, whereas the deletion mutant containing only the putative FAK binding domain was not (Fig. 4C).
To ensure that our lack of signal with the chimeric receptor containing the putative FAK binding domain was not due to the loss of phosphorylated FAK with the magnetic beads in our original lysates, we also compared the level of FAK phosphorylation in lysates obtained by boiling the cell-magnetic bead complexes under reducing conditions in a buffer containing 2% SDS (Fig. 4D). These results further confirm our findings that the putative FAK binding domain is not sufficient to trigger FAK phosphorylation above background levels.
Conserved Amino Acids in the C-terminal Segment of the ␤ 3 Cytoplasmic Domain Are Required to Signal FAK Phosphorylation-To identify the amino acids within the C-terminal portion of the ␤ 3 cytoplasmic domain that are required to signal FAK phosphorylation, amino acid substitution mutants were generated (Fig. 1). Initial substitutions were targeted to the NPXY motif which is conserved in the ␤ 1 , ␤ 3 , and ␤ 5 cytoplasmic domains (25). Three amino acid substitutions were made in this conserved motif (Fig. 1) and assayed as above. Chimeric receptors containing alanine substitutions for either asparagine 744, the ␤ 3 -744(N/A) mutant, or tyrosine 747, the ␤ 3 -747(Y/A) mutant, consistently did not signal FAK phosphorylation (Figs. 5 and 6), even when expressed at much higher levels (data not shown). These results indicate that the NPXY motif is important for integrin intracellular function in this signal transduction event. However, chimeric receptors containing the conservative substitution of phenylalanine for tyrosine 747 were able to signal FAK phosphorylation (Figs. 5 and 6), in agreement with published reports that a phenylalanine substitution for tyrosine in the NPXY motif of the ␤ 1 cytoplasmic domain does not inhibit FAK phosphorylation (30). These results indicate that the NPXY motif is important for integrin-triggered FAK phosphorylation and that although an aromatic amino acid is required at residue 747, it need not be tyrosine, indicating that phosphorylation at this residue is not required for integrin-triggered FAK phosphorylation.
Similar to the NPXY motif, the NXXY motif is also conserved in the ␤ 1 , ␤ 3 , and ␤ 5 cytoplasmic domains (25). To determine the contribution of this motif to the ability of the ␤ 3 cytoplasmic domain to signal FAK phosphorylation, we constructed two additional chimeric receptors, ␤ 3 -756(N/A) and ␤ 3 -759(Y/A), containing alanine substitutions for either asparagine 756 or tyrosine 759, respectively (Fig. 1). Chimeric receptors containing these mutant ␤ 3 cytoplasmic domains were able to signal FAK phosphorylation, although to a consistently lower level compared with the wild-type ␤ 3 cytoplasmic domain (Figs. 5  and 6). These results indicate that the NXXY motif is not strictly required for the ␤ 3 cytoplasmic domain to signal the phosphorylation of FAK, but may regulate the level of FAK phosphorylation induced by integrin engagement.
In addition to targeting the conserved NPXY and NXXY motifs, we also tested the consequences of the naturally occurring serine 752 to proline mutation (Fig. 1). This mutation is responsible for a variant of Glanzmann's thrombasthenia, in which the absence of platelet aggregation is due to the inability of the ␣ IIb ␤ 3 integrin to become activated in response to platelet agonists (42). Thus, the serine 752 to proline mutation makes ␣ IIb ␤ 3 defective in "inside-out" signal transduction (42)(43)(44). Chimeric receptors containing this mutant ␤ 3 cytoplasmic domain, ␤ 3 -752(S/P), consistently could not trigger FAK phosphorylation (Figs. 5 and 6), indicating that it was also defective in "outside-in" signal transduction. However, chimeric receptors containing the mutant ␤ 3 cytoplasmic domain, ␤ 3 -751-753(*/ A), containing alanine substitutions at residues threonine 751, serine 752, and threonine 753 (Fig. 1) signaled an increase in tyrosine phosphorylation of FAK (Fig. 5), although at a consistently lower level than the wild-type ␤ 3 cytoplasmic domain (Fig. 6). These results demonstrate that, although the substitution of proline at residue 752 has a negative effect on this signal transduction event, there is not a strict requirement for a serine residue at this position. DISCUSSION The importance of FAK activity is evident by the observation that FAK deficient embryos fail to develop normally, with an overall phenotype similar to fibronectin deficient embryos, supporting the notion that FAK activity is an important downstream component of signals generated by integrins (57). FAK activity has been implicated in regulating cell survival, proliferation, and migration (16,17,58). Although integrin ␤ subunit cytoplasmic domains are known to be an important link in the pathway by which cell adhesion activates FAK (30,34,35), the molecular mechanisms regulating this signal transduction event are not fully understood. Several different experimental approaches are likely to be required to delineate the molecular mechanisms leading from integrin engagement to FAK phosphorylation. The approach we have taken, expressing wild-type and mutant ␤ 3 cytoplasmic domains as separate domains connected to an extracellular reporter, allows examination of the role of specific amino acid motifs within the ␤ 3 cytoplasmic domain in triggering FAK phosphorylation, independent of their role in upstream events such as ligand binding and cell adhesion. This is important, since mutations have been identified in the ␤ 3 cytoplasmic domain that inhibit both these processes (27,52,53). Using this approach, we have demonstrated 1) that elements in both the membrane-proximal and C-terminal segments of the ␤ 3 cytoplasmic domain are necessary to trigger FAK phosphorylation; 2) that the conserved acidic residues in the membrane proximal region, proposed to  3). B, the same filter stripped and reprobed with antibodies to FAK. C, FAK was immunoprecipitated from cells expressing chimeric receptors containing the wild-type ␤ 3 cytoplasmic domain, the ␤ 3 -d728 -762, or the ␤ 3 -723,726(*/A) mutant (lanes 2-4), respectively. FAK phosphorylation was then assayed by Western blotting with mAb 4G10 to phosphotyrosine. Lane 1 shows control lysate from cells expressing chimeric receptors containing the wild-type ␤ 3 cytoplasmic domain. D, FAK phosphorylation assayed in cell lysates generated by lysing the positively transfected cells from the magnetic beads by boiling in a buffer containing 2% SDS: lysates from cells expressing chimeric receptors containing the wild-type ␤ 3 cytoplasmic domain (lane 1) or cells containing the ␤ 3 -d728 -762 mutant (lane 2).
be important for FAK binding to the integrin, are not required for this signaling event; 3) that the conserved NPXY motif is required for integrin ␤ cytoplasmic domains to trigger FAK phosphorylation, although a conservative tyrosine to phenylalanine mutation is tolerated; and 4) that the naturally occurring serine 752 to proline mutation that inhibits ␣ IIb ␤ 3 activation also inhibits the ability of the ␤ 3 cytoplasmic domain to trigger FAK phosphorylation.
Although FAK can bind to peptides modeled after the membrane-proximal region of ␤ cytoplasmic domains, FAK has not been co-precipitated with integrins. Therefore, it is not known whether FAK directly interacts with the ␤ cytoplasmic domains of endogenous heterodimeric integrins and whether the linkage of FAK and integrins differs for different integrin heterodimers and in different cell types. Nonetheless, clustering unoccupied integrins on the cell surface was shown to result in the coclustering of only FAK and tensin, as well as the tyrosine phosphorylation of FAK (49), suggesting that additional protein interactions are not required for FAK phosphorylation at least when signaled by the clustering of some unoccupied integrins. However, clustering ligand-occupied integrins or clustering integrins with antibodies that mimic the clustering of ligand-occupied integrins was shown to result in the coclustering of additional cytoskeletal and signaling proteins including actin, talin, ␣-actinin, paxillin, Src, and FAK (49 -51), as well as the phosphorylation of FAK (49). This suggests that ligand occupancy may either direct the association of additional proteins with FAK and tensin, or may dictate alternative interactions between integrins and these and other cytoplasmic components. Interestingly, clustering chimeric receptors containing a ␤ cytoplasmic domain resulted in the colocalization of the same cytoplasmic proteins as was observed for the clustering of ligand-occupied integrins, 2 as well as triggering FAK phosphorylation (34). These results suggest that the intracellular protein interactions mediated by the transfected chimeric receptors are similar to those mediated by ligand-occupied integrins.
Interestingly, the ␤ 1B and ␤ 3B cytoplasmic domains contain the putative FAK binding domain (41); however, they do not signal FAK phosphorylation (34,37), suggesting that either the amino acids inserted by alternative splicing interfered with the ability of FAK to bind to ␤ cytoplasmic domains within the cell, or that these alternative amino acids inhibited additional or alternative protein interactions required for this signaling event. Our results with mutant cytoplasmic domains lacking either the membrane-proximal segment or the C-terminal segment indicated that structural information in both regions of the ␤ 3 cytoplasmic domain is important for the regulation of integrin-triggered FAK phosphorylation and that the putative FAK binding domain alone is not sufficient to trigger FAK phosphorylation. Consistent with this notion, ␤ 1 cytoplasmic domain deletion mutants lacking the C-terminal segment also did not signal FAK phosphorylation (30). The ability of the chimeric receptors containing substitutions in the putative FAK binding domain to signal FAK phosphorylation further suggested that FAK binding to this region is not required for 2 S. Miyamoto, S. LaFlamme, and K. Yamada, unpublished results. the integrin ␤ 3 cytoplasmic domain to trigger this signaling event. These results suggest that perhaps integrin-cytoplasmic protein interactions other than FAK binding to its putative binding site on the ␤ 3 cytoplasmic domain may couple integrins to the FAK signaling pathway, and that additional structural information for this interaction must be encoded in the Cterminal region of ␤ cytoplasmic domains.
When we further analyzed the requirements in the C-terminal segment of the ␤ 3 cytoplasmic domain for this signaling event, we found that the NPXY motif present in the C-terminal segment of many integrin ␤ cytoplasmic domains is important in the regulation of FAK phosphorylation. The NPXY motif may function by regulating the overall conformation of the ␤ cytoplasmic domain, thereby influencing many protein interactions involving the ␤ cytoplasmic domain involved in several integrin-mediated processes. Alternatively, the NPXY motif may itself interact with specific cytoplasmic components necessary to trigger FAK phosphorylation. One of these proteins may be the cytoskeletal protein, talin, since peptides spanning this motif can inhibit the association of talin and integrins (56). Consistent with this notion, the C-terminal segment of the ␤ 1 cytoplasmic domain has been shown to be required for the association of both FAK and talin with integrins (40,51), as well as FAK phosphorylation (30). This suggests a correlation between the association of talin and integrins with FAK phosphorylation.
In contrast to these findings, others have recently demonstrated that recombinant heterodimeric ␣ IIb ␤ 3 integrin receptors containing only the putative FAK binding site of the ␤ 3 cytoplasmic domains were able to signal FAK phosphorylation (45). These results differ from ours and those of Guan et al. (30), which suggest that amino acids C-terminal to the putative FAK binding site are also required to signal FAK phosphorylation. Although the reasons for these differences are not known, it seems likely that signaling pathways leading to FAK phosphorylation may differ for different integrin heterodimers and different cell types (46). Recently, it has been reported that integrins can associate with transmembrane 4 proteins and that these associations occur with only specific integrin heterodimers (47), including ␣ IIb ␤ 3 (48). Since these associations may promote integrin clustering by extracellular interactions (47), they may serve in some cases to bypass a requirement for cytoskeletal interactions mediated by the C-terminal segment of integrin ␤ cytoplasmic domains.
Several laboratories have mutagenized the conserved motifs within the ␤ 3 cytoplasmic domain and have assayed several different integrin functions. The inhibitory phenotypes reported are strikingly similar to those reported here for the function of the ␤ 3 cytoplasmic domain in triggering FAK phosphorylation. Mutations in the NPXY motif have recently been shown to inhibit the regulation of the affinity state of the platelet integrin, ␣ IIb ␤ 3 (52), as well as other integrin-mediated processes such as cell adhesion, cell spreading, cell migration, and integrin localization to adhesion sites (53)(54)(55). However, our results are the first demonstration to our knowledge of a role for this motif in a specific post-ligand binding integrin signaling event. Mutation of serine 752 to proline had similar inhibitory effects (52,54). In addition, mutations in the NXXY motif, as well as alanine substitutions for serine 752, had intermediate phenotypes, not only on FAK phosphorylation described in this report, but also on cell spreading and the regulation of integrin affinity (52,54), suggesting that the NXXY motif may modulate the binding of a protein required for integrin function in these processes.
The similar inhibitory phenotypes described in this study and those discussed above suggest that conserved motifs tar-geted in these studies may function to regulate the overall conformation of the ␤ cytoplasmic domain. Therefore, mutations in these motifs may have global effects on many different protein interactions. However, it is also possible that since ␤ intracellular domains are small in size, yet function in many different integrin-mediated processes, they may encode overlapping binding sites for several cytoplasmic proteins. Additional mutagenesis, functional assays, and protein binding assays will be necessary to dissect the specific interactions required to mediate specific integrin functions.