Cooperative Role of the Membrane-proximal and -distal Residues of the Integrin β3 Cytoplasmic Domain in Regulation of Talin-mediated αIIbβ3 Activation*

Integrin cytoplasmic tails regulate integrin activation that is required for high affinity binding with ligands. The interaction of the integrin β subunit tail with a cytoplasmic protein, talin, largely contributes to integrin activation. Here we report the cooperative interaction of the β3 membrane-proximal and -distal residues in regulation of talin-mediated αIIbβ3 activation. Because a chimeric integrin, αIIbβ3/β1, in which the β3 tail was replaced with the β1 tail was constitutively active, we searched for the residues responsible for integrin activation among the residues that differed between the β3 and β1 tails. Single amino acid substitutions of Ile-719 and Glu-749 in the β3 membrane-proximal and -distal regions, respectively, with the corresponding β1 residues or alanine rendered αIIbβ3 constitutively active. The I719M/E749S double mutant had the same ligand binding activity as αIIbβ3/β1. These β3 mutations also induced αVβ3 activation. Conversely, substitution of Met-719 or Ser-749 in the β1 tail with the corresponding β3 tail residue (M719I or S749E) inhibited αIIbβ3/β1 activation, and the M719I/S749E double mutant inhibited ligand binding to a level comparable with that of the wild-type αIIbβ3. Knock down of talin by short hairpin RNA inhibited the I719M- and E749S-induced αIIbβ3 activation. These results suggest that the β3 membrane-proximal and -distal residues cooperatively regulate talin-mediated αIIbβ3 activation.

Platelet integrin ␣IIb␤3 is a transmembrane receptor that mediates platelet adhesion and aggregation (1,2). ␣IIb␤3 exists in a low affinity state in resting platelets and requires activation for high affinity binding with soluble ligands (1,2). Activation of ␣IIb␤3 is tightly linked to structural rearrangements of the ␣IIb␤3 molecule. Recent crystal and NMR structure studies have revealed the precise conformational changes that occur during these rearrangements, including separation of the cytoplasmic, transmembrane, and extracellular leg domains of the ␣IIb and ␤3 subunits, a switchblade-like extension from a bent to an extended conformation of the extracellular domains, outward swing of the hybrid domain, and structural rearrange-ments of the head domain containing a ligand binding pocket (3)(4)(5)(6)(7)(8). This long range conformational transduction is initiated from the cytoplasmic tails of the ␣IIb and ␤3 subunits in the integrin activation process referred to as inside-out signaling.
Several NMR analyses have identified multiple hydrophobic and electrostatic contacts within the membrane-proximal helices of the ␣IIb and ␤3 cytoplasmic tails (3,9). Deletion or mutation of the membrane-proximal ␣IIb GFFKR or ␤3 LLITIHD sequence renders ␣IIb␤3 constitutively active (10,11). Therefore, it is considered that the membrane-proximal regions of the ␣IIb and ␤3 cytoplasmic tails associate to form a clasp, maintaining ␣IIb␤3 in a low affinity state. On the other hand, the structural basis of the membrane-distal regions of the ␣IIb and ␤3 tails is less obvious.
The structure of the ␣IIb membrane-distal region has been explored by NMR. Analysis of the myristoylated peptide containing the entire ␣IIb tail sequence revealed that the membrane-distal loop folded back toward the membrane-proximal helix, while this conformation was not observed in recent analysis of the ␣IIb tail in a membrane-embedding environment (5,12). A lipid-modified peptide containing the membrane-distal sequence (RPPLEED) inhibits agonist-induced ␣IIb␤3 activation (13). Mutation of the ␣IIb RPP distal sequence activates ␣IIb␤3 in a metabolic energy-dependent manner (14,15). These findings suggest that the membrane-distal region of the ␣IIb tail suppresses energy-dependent ␣IIb␤3 activation, although the mechanism responsible for activation of the ␣IIb␤3 mutants remains unclear.
In view of the link between integrin activation and allosteric structural rearrangements of the extracellular segments of integrins (23), one would expect that structural changes in the ␤3 membrane-distal region containing binding sites for intracellular proteins would be relayed to the membrane-proximal region, leading to ␣IIb␤3 activation. However, there has been no evidence that structural rearrangement of the ␤3 membrane-distal region is directly linked to integrin activation. Except for the talin binding site, the structure of the ␤3 membrane-distal region is ill-defined, and no activating mutation has so far been reported in the ␤3 membrane-distal region. In this context, we considered that a previously reported ␣IIb␤3 mutant in which the ␤3 tail was replaced by the ␤1 tail was noteworthy (24). This chimeric integrin, ␣IIb␤3/␤1, was constitutively active. Because the ␤1 and ␤3 subunits have relatively high sequence homology, we reasoned that the residues differing between the ␤1 and ␤3 tails are responsible for ␣IIb␤3 activation. The ␤3 and ␤1 tails have 20 residues that differ, and 18 of them are located in the membrane-distal region, raising the possibility that the membrane-distal residue(s) may contribute to ␣IIb␤3 activation.
In this study, we attempted to identify the critical residue(s) for regulation of ␣IIb␤3 activation using a model of activated ␣IIb␤3, ␣IIb␤3/␤1. We obtained evidence that structural perturbation of specific residues in the ␤3 membrane-distal region can induce ␣IIb␤3 activation and that the membrane-proximal and -distal residues in the ␤3 tail cooperatively regulate talinmediated ␣IIb␤3 activation.
Site-directed Mutagenesis-Site-directed mutagenesis was carried out by overlap extension polymerase chain reaction (PCR) or by using a QuikChange site-directed mutagenesis kit (Stratagene) as described previously (15,25). Two recombinant ␤3 fragments were generated from a full-length ␤3 cDNA template by PCR with Pfu polymerase and complementary primers containing the desired mutations. These two fragments were combined and subjected to PCR using primers containing Hin-dIII and XbaI restriction sites at the 5Ј-and 3Ј-ends of ␤3, respectively. The amplified PCR product was digested with HindIII/XbaI and cloned into a mammalian expression vector, pCDNA3 (Invitrogen). The nucleotide sequences of the inserts were confirmed by sequence analysis.
Expression of Recombinant Integrins-The mutated ␤3 and wild-type ␣IIb cDNAs were cotransfected into Chinese hamster ovary (CHO) cells with Lipofectamine (Invitrogen) in accordance with the manufacturer's instructions. Human 293T cells were also used for transfection. The cells were then cul-tured in Dulbecco's modified Eagle's medium containing 10% fetal calf serum for 48 h to achieve cell surface expression of the recombinant integrins.
Ligand Binding Assay-PAC1 and fibrinogen binding assays were performed as described previously (15,25,26). CHO cells were suspended in Tyrode's buffer (137 mM NaCl, 12 mM NaHCO 3 , 2.6 mM KCl, 2 mM CaCl 2 , 2 mM MgCl 2 , 5 mM Hepes, 5.5 mM glucose, and 1 mg/ml bovine serum albumin, pH 7.4) and incubated for 30 min at room temperature with 10 g/ml PAC1 and 5 g/ml biotin-labeled 4B10 in the presence of 2 mM RGDS peptide or buffer. The cells were then washed and resuspended in a mixture of a 1:20 dilution of fluorescein isothiocyanate-conjugated goat anti-mouse IgM antibody (BIOSOURCE, Camarillo, CA) and a 1:50 dilution of phycoerythrin-streptavidin (Molecular Probes, Eugene, OR) for 25 min on ice. Then, 1 l of 1 mg/ml propidium iodide (Sigma) was added. After 5 min the cells were washed and resuspended in 500 l of ice-cold Tyrode's buffer and analyzed on a FACSCalibur flow cytometer (BD Biosciences). After electronic compensation, PAC1 binding (FL1) was analyzed on the gated subset of live cells (propidium iodide-negative, FL3) that were positive for ␣IIb␤3 expression (FL2). In the assays for fibrinogen binding to ␣IIb␤3, fluorescein isothiocyanate-conjugated fibrinogen was used instead of PAC1. The binding of fibrinogen to ␣V␤3 in CHO cells was measured as described previously (25).
Talin Knock Down-Knock down of talin expression in CHO cells was performed by the small interfering RNA technique according to the method of Tadokoro et al. (18). CHO cells were transfected with plasmids encoding ␣IIb cDNA, ␤3 cDNA, and shRNAs targeting talin-1. Scrambled shRNA was used as a control. A preparation without shRNA transfection was included as another control. The cells were cultured for 72 h to allow the shRNA to exert its maximal inhibitory effect (18). The transfected cells were then subjected to ligand binding assay as described above. To assess talin expression, CHO cells were transfected with shRNA and pEGFP-C1 (BD Biosciences Clontech). The cells were then fixed with 0.5% paraformaldehyde, permeabilized with 0.05% saponin, and incubated for 30 min at room temperature with anti-talin-1 and -2 monoclonal antibody 8d4 (Sigma). After washing, the cells were incubated for 30 min on ice with phycoerythrin-conjugated goat anti-mouse IgG (BIOSOURCE). They were then washed and resuspended in phosphate-buffered saline containing 0.025% saponin, 0.25% bovine serum albumin, and 1 g/ml propidium iodide. Binding of 8d4 antibody to talin was analyzed using flow cytometry by gating propium iodide-negative live cells that were positive for green fluorescent protein expression.

Mutation of the Ile-719 and Glu-749 Residues in the ␤3 Tail
Activates ␤3 Integrins-To identify the residues critical for regulation of ␣IIb␤3 activation in the ␤3 tail, we focused on amino acids in the ␤3 tail that differed from those in the ␤1 tail as the ␣IIb␤3/␤1 chimeric integrin, in which the ␤3 tail is replaced with the ␤1 tail, has been reported to be constitutively active (24). We produced 13 ␣IIb␤3 mutants in which the individual or group residues in the ␤3 tail were substituted with the corresponding ␤1 tail residues (Fig. 1). The ␣IIb␤3 mutants were expressed on the surface of CHO cells by cotransfection of mutant ␤3 and wild-type ␣IIb cDNAs, and surface expression of ␣IIb␤3 was determined by flow cytometry using a non-function-blocking anti-␣IIb␤3 monoclonal antibody, 4B10 (15,25).
The expression of ␣IIb␤3 mutants on the cell surface was 93-136% that of the wild-type ␣IIb␤3, indicating almost equivalent expression of the mutated integrins ( Fig. 2A). Among these mutants, ␣IIb␤3 bearing the ␤3I719M or ␤3E749S mutation bound significantly higher levels of PAC1 than wild-type ␣IIb␤3 without any stimulation and the RGDS peptide abolished PAC1 binding to the mutants, indicating a constitutively active state (Fig. 2B). Other mutants, as well as wild-type ␣IIb␤3, were in a low affinity state. To determine whether activation of ␣IIb␤3 by I719M and E749S mutation is due to loss of Ile-719 and Glu-749 residues or insertion of methionine and serine into the ␤3 tail, we examined the effects on PAC1 binding of I719A and E749A mutants in which the residues concerned were replaced with alanine. I719A and E749A bound the same level of PAC1 as I719M and E749S did, suggesting that loss of the side chains of Ile-719 and Glu-749 residues is responsible for ␣IIb␤3 activation (Fig. 2B).
Although the I719M and E749S mutants were constitutively active, each exhibited less PAC1 binding than the ␣IIb␤3/␤1-activating integrin in which the entire ␤3 tail was replaced by the ␤1 tail (Fig. 3). Because we screened all the ␤3 tail residues that were different from the ␤1 residues, it is unlikely that tail residues other than Ile-719 and Glu-749 are involved in ␣IIb␤3/␤1 activation. To clarify whether the combination of I719M and E749S mutations was able to reproduce an activating state comparable with ␣IIb␤3/␤1, we produced the I719M/E749S double mutant. This mutant was expressed normally on the surface of CHO cells and showed more PAC1 binding than the single mutants, reaching the same level of PAC1 binding as ␣IIb␤3/␤1 (Fig. 3). Similar effects were observed when soluble fibrinogen, a physiological ligand for ␣IIb␤3, was used instead of PAC1 (Fig. 3). Similar results were also obtained when human 293T cells were used instead of CHO cells (data not shown). These results suggest that mutation of the membrane-proximal Ile-719 and membrane-distal Glu-749 residues in the ␤3 tail has an additive effect on ␣IIb␤3 activation and that activation of the chimeric integrin ␣IIb␤3/␤1 can be recapitulated by double mutations of Ile-719 and Glu-749 residues in the ␤3 tail. The ␤3 subunit can associate with ␣IIb and ␣V subunits, forming ␣IIb␤3 and ␣V␤3 integrins. To determine whether the mutations of the ␤3 tail responsible for ␣IIb␤3 activation also give rise to ␣V␤3 activation, recombinant ␣V␤3 was expressed on the surface of CHO cells by transfection with wild-type ␣V and mutant ␤3 cDNAs. The surface expression of mutant ␣V␤3 was comparable with that of wildtype ␣V␤3. The ␣V␤3I719M and ␣V␤3E749S mutants exhibited significant fibrinogen binding without any stimulation, and the I719M/E749S double mutant bound more fibrinogen, reaching the same level of fibrinogen as ␣V␤3/␤1 (Fig. 4). These results suggest that ␤3I719M and ␣V␤3E749S mutations activate ␣V␤3 as well as ␣IIb␤3.

Introduction of the ␤3I719 and ␤3E749 Residues into the ␤1 Tail Suppresses Integrin Activation-The foregoing results
showed that the I719M and E749S mutations were responsible for activation of the ␣IIb␤3/␤1 chimeric integrin. We next examined whether the converse mutations, M719I and S749E, prevented ␣IIb␤3/␤1 activation. The M719I and S749E ␣IIb␤3/␤1 mutants reduced PAC1 binding significantly, even though expression of these mutants was sufficient (Fig. 5). The M719I/S749E double mutant reduced PAC1 binding to a level comparable with wild-type ␣IIb␤3. These results are consistent with the foregoing results obtained with ␣IIb␤3I719M and ␣IIb␤3E749S mutations and further suggest the cooperative role of the ␤3I719 and ␤3E749 residues in regulation of ␣IIb␤3 activation.
The ␤3I719 and ␤3E749 Residues Suppress Talin-dependent ␣IIb␤3 Activation-To explore the mechanism of ␣IIb␤3 activation by the Ile-719 and Glu-749 mutations, we examined whether activation of these mutant integrins is mediated by talin, a final common element of signaling pathways to integrin activation (18). Knock down of talin was carried out by the use of shRNA and was confirmed by inhibition of talin expression in CHO cells by 66% (Fig. 6A). This inhibitory effect was the same as that shown in the previous study (18). Talin knock down significantly reduced PAC1 binding to the CHO cells expressing the mutant ␣IIb␤3, whereas control shRNA had no effect (Fig. 6B). This result indicates that both the I719M and E749S mutations induce ␣IIb␤3 activation in a talin-dependent manner, suggesting that the ␤3I719 and ␤3E749 residues suppress talin-mediated ␣IIb␤3 activation to retain ␣IIb␤3 in a default low affinity state.

DISCUSSION
In this study, we have identified two residues in the ␤3 tail, Ile-719 and Glu-749, that are critical for maintaining ␣IIb␤3 in a low affinity state. Ile-719 is located in the membrane-proximal region of the ␤3 tail and binds to the ␣IIb tail in NMR  Activation of ␣IIb␤3 by ␤3 Tail Mutations FEBRUARY 29, 2008 • VOLUME 283 • NUMBER 9 structure models (3,9). Because the membrane-proximal interface between the ␣ and ␤ tails forms a clasp to maintain integrins at a low affinity state (23), one would expect that mutations within the membrane-proximal clasping region would activate integrins. A simple explanation for ␣IIb␤3 activation by Ile-719 mutation is that loss of the bond between the ␤3I719 and the ␣IIb tail directly disrupts the membrane-proximal clasp. Although the Ile-719-mediated contact is only part of the multiple hydrophobic and electrostatic interactions within the clasp, disturbing this specific contact may weaken the interaction and facilitate ␣IIb␤3 activation. However, the Ile-719 mutation-induced ␣IIb␤3 activation is talin-dependent, as demonstrated in this study. A similar mutational effect on ␣IIb␤3 activation has been demonstrated for the Asp-723 residue in the ␤3 membrane-proximal region (18,27). Talin binds to not only the ␤3 membrane-distal region but also the ␤3 membraneproximal region, and a recent study has shown that binding of talin to the latter is critical for ␣IIb␤3 activation (17). Therefore, it is possible that structural changes around the ␤3 membrane-proximal Ile-719 and Asp-723 residues give access to the talin protein, thereby activating ␣IIb␤3.
Another activating mutation of the ␤3 tail residue, Glu-749, is intriguing. Unlike Ile-719, Glu-749 is located in the ␤3 membrane-distal region. This region directly interacts with many signaling and cytoskeletal molecules, including tyrosine kinases and talin, that are involved in the inside-out signaling pathway (18,28,29). Deletion or mutation of the ␤3 membrane-distal region has been reported to result in failure of ␣IIb␤3 activation by inside-out signaling (19 -22). These studies clearly demonstrate that the ␤3 membrane-distal region functions as a receiver of intracellular signals activating ␣IIb␤3. Because integrin activation is brought about by allosteric structural rearrangement of each extracellular segment of integrins (23), it is likely that the structural changes in the ␤3 membrane-distal region following interaction with intracellular molecules are relayed to the membrane-proximal region, leading to ␣IIb␤3 activation. However, there is no direct evidence that the structural rearrangement of the ␤3 membranedistal region induces integrin activation. Our present study shows that a single amino acid mutation of the ␤3 membrane-distal residue Glu-749 activates ␣IIb␤3 in a talin-dependent manner. No such activating mutation has been found in the ␤3 membrane-distal region despite numerous reports of loss-of-function mutations in this region (19 -22). Our result provides experimental evidence that structural perturbation of the ␤3 membrane-distal region is linked to integrin activation through talin binding. Furthermore, we have found that the ␤3E749S mutation activates ␣V␤3 and that ␣V␤3/␤1 is constitutively active although ␣V can associate with ␤1, forming integrin ␣V␤1. This finding suggests that E749S-induced integrin activation is not ascribed to simple dissociation of the ␤3 tail from the ␣ tail but to the structural change in the ␤3 membrane-distal region. Talin binds to the 739 WDTANNPLYK sequence that is outside Glu-749 (16). However, Glu-749 is adjacent to this talin binding site and NMR perturbation of Glu-749 by interaction with talin has been observed (16). It is therefore possible that the Glu-749 mutation has a direct effect on talin binding and thus affects ␣IIb␤3 activation through talin-mediated mechanism. To demonstrate this hypothesis, a direct binding assay for interaction of talin with recombinant ␤3 tails is necessary. Unfortunately, we were unable to perform such an assay because detection of low affinity integrin-talin interactions requires integrin tails connected with specially designed model proteins bearing coiled-coil and glycine spacer structures that were not commercially available (30).
Talin was recently reported to compete with filamin for binding to the ␤3 tail, and this competition may regulate integrin activation (31). Filamin binds to the 745 PLYKEATSTFTN sequence encompassing Glu-749 (31). It is therefore possible that reduced binding of filamin to the E749S mutant results in more talin binding, leading to ␣IIb␤3 activation. However, this mechanism is unlikely because the ␤1 tail bound much less filamin than the ␤7 tail although both tails have serine at the ␤3E749 position, indicating that E749S is not a determinant of filamin binding affinity (24).
A recent NMR analysis has demonstrated that the ␤3 membrane-distal NPLY motif is anchored to the membrane surface and that this membrane anchoring may restrict binding of talin to the ␤3 tail to constrain ␣IIb␤3 in a default low affinity state (5). Because Glu-749 lies within the helix flanking the NPLY loop (5), Glu-749 mutation may release the membrane anchoring of the NPLY, thus facilitating talin binding to the ␤3 tail. This hypothesis could account for talin-dependent activation of ␣IIb␤3 by Glu-749 mutation. Structural analysis of the talin/ Glu-749 mutant complex at the atomic level will be required to verify this hypothesis.
Because the ␤3 membrane-distal region binds intracellular signaling molecules, it is possible that Glu-749 mutation affects the interaction of these proteins with the membrane-distal tail that regulates ␣IIb␤3 activation. The ␤3E749 residue is critical for binding of some phosphotyrosine-binding proteins, including Dok1 and Numb (32). Dok1 competes with talin for binding to the ␤3 tail and has the ability to inhibit ␣IIb␤3 activation (17). It is therefore possible that reduced binding of Dok1 to the ␤3E749S tail results in increased binding of talin, thereby activating ␣IIb␤3, although overexpression of Dok1 is required for partial inhibition of ␣IIb␤3 activation (17). The effect of Numb on integrin signaling is unknown. Another phosphotyrosine-binding protein, Dab2, binds to the ␤3 tail and functions as a negative regulator of ␣IIb␤3 activation (33). However, it is unlikely that Dab2 is involved in the mechanism of E749S-induced ␣IIb␤3 activation because the E749S mutation had no effect on Dab2 binding (32). Because the Glu-749 residue is specific for ␤3 among all the integrin ␤ subunits, it may interact with some ␤3-specific partner to restrain ␣IIb␤3 activation. Based on the present study, screening of intracellular proteins competent to bind the wild-type but not E749S mutant ␤3 tail may be useful for detecting novel regulators of ␣IIb␤3 activation.
It should be noted that the mutational effects of the membrane-proximal Ile-719 and the membrane-distal Glu-749 res-idues were additive and talin-dependent. This result suggests that the conformational changes in both the membrane-proximal and -distal regions play a cooperative role in talin function. Very recently, Wegener et al. (17) demonstrated the cooperative roles of the membrane-proximal and -distal regions in the process of talin-mediated ␣IIb␤3 activation. They proposed a model of talin-induced integrin activation in which interaction of talin with the membrane-distal region anchors talin to the ␤ subunit tail tightly and subsequent interaction of talin with the membrane-proximal region induces separation or reorientation of the integrin tails, leading to integrin activation. Disruption of either interaction with the membrane-proximal or -distal region by ␤3 tail mutations resulted in failure of ␣IIb␤3 activation (17). Our study showed that either Ile-719 or Glu-749 mutation induced talin-dependent ␣IIb␤3 activation, thus supporting this recent model of talin-induced integrin activation in terms of gain-of-function mutations. Wegener et al. (17) employed F2/F3 fragments of talin to activate ␣IIb␤3. In contrast, we observed the effects of the endogenous intact talin molecule on integrin activation. Therefore, our findings may indicate the ␤3 tail site critical for interaction with the intact talin molecule. Although we could not define the mechanism underlying the activating effects of the ␤3 tail mutations, identification of the ␤3 tail residues as regulators of talin-mediated ␣IIb␤3 activation will help to unravel the molecular basis for inside-out signaling of integrins.