HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

J. Biol. Chem., Vol. 280, Issue 26, 24775-24783, July 1, 2005

This Article Abstract Full Text (PDF) All Versions of this Article: 280/26/24775    most recent M409548200v1 Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Kamata, T. Articles by Aiso, S. Search for Related Content PubMed PubMed Citation Articles by Kamata, T. Articles by Aiso, S. Social Bookmarking                      What's this?

Membrane-proximal / Stalk Interactions Differentially Regulate Integrin Activation^*

Tetsuji Kamata, Makoto Handa¶, Yukiko Sato, Yasuo Ikeda||, and Sadakazu Aiso

From the Departments of Anatomy, ^¶Transfusion Medicine and Cell Therapy, and ^||Internal Medicine, Keio University School of Medicine, Tokyo 160-8582, Japan

Received for publication, August 19, 2004 , and in revised form, April 21, 2005.

	ABSTRACT

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
The affinity of integrin-ligand interaction is regulated extracellularly by divalent cations and intracellularly by inside-out signaling. We report here that the extracellular, membrane-proximal / stalk interactions not only regulate cation-induced integrin activation but also play critical roles in propagating inside-out signaling. Two closely related integrins, IIb3 and V3, share high structural homology and bind to similar ligands in an RGD-dependent manner. Despite these structural and functional similarities, they exhibit distinct responses to Mn²⁺. Although V3 showed robust ligand binding in the presence of Mn²⁺, IIb3 showed a limited increase but failed to achieve full activation. Swapping stalk regions between IIb and V revealed that the stalk, but not the ligand-binding head region, was responsible for the difference. A series of IIb/V domain-swapping chimeras were constructed to identify the responsible domain. Surprisingly, the minimum component required to render IIb3 susceptible to Mn²⁺ activation was the V calf-2 domain, which does not contain any divalent cation-binding sites. The calf-2 domain makes interface with epidermal growth factor 4 and tail domain in three-dimensional structure. The effect of calf-2 domain swapping was partially reproduced by mutating the specific amino acid residues in the calf-2/epidermal growth factor 4- tail domain interface. When this interface was constrained by an artificially introduced disulfide bridge, the Mn²⁺-induced V3-fibrinogen interaction was significantly impaired. Notably, a similar disulfide bridge completely abrogated fibrinogen binding to IIb3 when IIb3 was activated by cytoplasmic tail truncation to mimic inside-out signaling. Thus, disruption/formation of the membrane-proximal / stalk interface may act as an on/off switch that triggers integrin-mediated bidirectional signaling.

	INTRODUCTION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Integrins are a family of / heterodimeric transmembrane cell surface receptors that mediate cell-extracellular matrix and cell-cell interactions. The hallmark of integrin-dependent adhesive interaction is its regulation by intracellular signaling events (inside-out signaling) and by divalent cations. In addition to mediating adhesive interactions, liganded integrins initiate signals inside the cell to modify cell behavior (outside-in signaling) and thus play fundamental roles in numerous biological processes such as differentiation, cell survival, apoptosis, and cell motility (1). Integrin-mediated bidirectional signaling is accompanied by conformational change of the integrin structure. The crystal structure of V3 extracellular domains revealed an unexpected bent conformer distinct from the extended conformer observed under electron microscope (2, 3). High-resolution electron microscopic observation on truncated recombinant V3 has confirmed the presence of both conformers, suggesting that the transition from one conformer to another might take place under physiological conditions (4). However, integrin extension per se is not required for the activation, but the swing-out of the hybrid domain (the transition from "closed" headpiece to "open" headpiece) that accompanies the extension is the critical event (5). Thus, integrin extracellular domains undergo extensive structural rearrangement (so-called "switchblade-like" movement) upon Mn²⁺/ligand binding. In support, fluorescence resonance energy transfer measurements between 41-bound peptide ligand and plasma membrane revealed that such movement actually takes place in living cells (6, 7).

Divalent cations differentially regulate integrin-ligand interaction. Whereas Mn²⁺ and, to a lesser extent, Mg²⁺ have an enhancing effect, Ca²⁺ typically has an inhibitory effect on ligand binding. Biochemical studies have suggested the existence of three classes of cation-binding sites with distinct function and preference for cations (8). The V3 extracellular domains turned out to contain eight cation-binding sites. Five of them are located in the -propeller and at the junction between the thigh and calf-1 domains of the subunit. The other three are located in the A domain of the subunit. Besides the metal ion-dependent adhesion site (MIDAS), which is essential for ligand binding (9, 10), the A domain possesses two additional sites designated ADMIDAS (adjacent to MIDAS) and LIMBS (ligand-associated metal binding site), respectively (11). Whereas ADMIDAS is occupied by a cation regardless of the presence of bound ligand, MIDAS and LIMBS have been shown to bind Mn²⁺ only in the presence of bound ligand (3, 11). A recent report by Chen et al. (12) suggests that the ADMIDAS is the negative regulatory site for Ca²⁺, whereas LIMBS is the positive regulatory site for Ca²⁺. These reports implicate that the cation-binding sites in the A domain represent the three classes of cation-binding sites described by Mould et al. (8), thus they are primarily responsible for integrin affinity regulation by divalent cations. In contrast, the role of the subunit, particularly the five cation-binding sites, has not been clearly defined.

Two 3 integrins, IIb3 and V3,share high structural homology (13). Both integrins share a common 3 subunit and bind fibrinogen (Fbg),¹ von Willebrand factor, fibronectin, and vitronectin in an RGD-dependent manner. Despite these structural and functional similarities, divalent cations regulate ligand binding differently in these integrins. Kinetic studies have shown that Mn²⁺ supports Fbg binding to both IIb3 and V3, albeit with a relatively slow association rate, whereas Ca²⁺ supports Fbg binding only to IIb3, but not to V3 (14). In agreement, real-time measurements of Fbg binding interaction have shown that Mn²⁺ increased the affinity for Fbg in both integrins by increasing the association rate (4, 15). Consistent with these observations, V3-mediated cell attachment to immobilized Fbg was greatly enhanced by Mn²⁺ but not by Ca²⁺ (14). In contrast, intact IIb3 expressed on the cell surface exhibits different characteristics. First of all, cations including Mn²⁺ are unable to initiate platelet aggregation or induce Fbg binding to platelets, unless platelets are stimulated by agonists. The agonist-stimulated platelet aggregation is also poorly supported by Mn²⁺ (14, 16, 17). These lines of evidence suggest that cation-binding sites in the A domain do not account for the distinct response to Mn²⁺ in 3 integrins, but rather that the structural difference in the subunit may be responsible. In this study, we sought to determine the mechanism that regulates Mn²⁺-induced activation in 3 integrins. We provide evidence that the calf-2 domain, but not the cation-binding sites, in the subunit plays a critical role in regulating the Mn²⁺-induced integrin activation. Furthermore, the results suggest that disruption/formation of the membrane-proximal calf-2/epidermal growth factor (EGF)4- tail domain (TD) interface may act as an on/off switch that propagates the conformational signals in integrin-mediated bidirectional signaling.

	EXPERIMENTAL PROCEDURES

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Antibodies and Reagents—Anti-IIb monoclonal antibody (mAb) PL98DF6 (18) was a generous gift from Drs. J. Ylänne (University of Uppsala, Uppsala, Sweden) and I. Virtanen (University of Helsinki, Helsinki, Finland). Conformation-dependent anti-3 mAbs anti-LIBS1 and anti-LIBS2 (19) were generous gifts from Dr. Mark H. Ginsberg (University of California, San Diego, CA). Anti-IIb3 complex-specific ligand-mimetic mAb OP-G2 (20) was a kind gift from Dr. Yoshiaki Tomiyama (University of Osaka, Osaka, Japan). The IIb3 complex-specific anti-IIb mAb P2, anti-3 mAb SZ21, and anti-V mAb AMF-7 were purchased from Beckman Coulter (Fullerton, CA). Anti-IIb3 complex-specific activating mAb PT25-2 and non-functional anti-3 mAb VNR5-2 have been characterized previously (21). Fluorescein isothiocyanate (FITC)- and R-phycoerythrin-conjugated goat anti-mouse polyclonal antibodies were purchased from BIOSOURCE. Synthetic peptide Gly-Arg-Gly-Asp-Ser (GRGDS) was purchased from Peptide Research Institute (Osaka, Japan). FITC was purchased from Sigma-Aldrich. Human Fbg was purchased from Experimental Cell Research (South Bend, IN).

Construction of Mutant IIb, V, and 3 cDNA Clones—The full-length cDNAs for integrin IIb, V, and 3 subunits (generous gifts from Dr. Joseph C. Loftus, Mayor Clinic, AZ) were cloned into mammalian expression vector pBJ-1 (kindly provided by Dr. Mark Davis, University of California, San Francisco, CA). A SacI site was engineered into the V cDNA at site homologous to the endogenous SacI site in the IIb cDNA by site-directed mutagenesis using Transformer Site-Directed Mutagenesis Kit (BD Biosciences). The cDNAs for B/V and V/B chimeras were created by replacing the SacI-BamHI fragment between IIb and V. The cDNAs for IIb/V domain-swapping chimeras TC1C2, TC1, T, C1, C2, D, C1C2D, TC2D, and TC1D were created using the overlap extension PCR method. The domain boundaries for each chimera were set as shown in Fig. 1A. The cDNAs for B/V 753–755, B/V 760–764, B/V 774, B/V 781–783, B/V 787, B/V 899–900, B/V 902–904, and B/V 958–960 mutants were created by site-directed mutagenesis. In these IIb mutants, the amino acid sequences in each of the indicated regions were replaced by the corresponding V sequences (⁷⁵³NSF to VSS, ⁷⁶⁰VVAAE to FLPIP, Asp⁷⁷⁴ to Glu, ⁷⁸¹EHT to QHI, His⁷⁸⁷ to Arg, ⁸⁹⁹QR to KS, ⁹⁰²MTV to ILY, and ⁹⁵⁸ALE to GIQ, respectively). Likewise, B/V 760, B/V 761, B/V 762, B/V 763, B/V 764, B/V 899, and B/V 900 mutants represent V760F, V761L, A762P, A763I, E764P, Q899K, and R900S mutations, respectively. The cDNAs for B/V 787/899–900 and B/V 787/900 were created from cDNAs for B/V 787 and B/V 899–900 and for B/V 787 and B/V 900, respectively, by replacing the BamHI fragment. The cDNA for B/V 760–764/787/899–900 was created from cDNAs for B/V 760–764 and B/V 787/899–900 by replacing the SacI fragment. The V D599A, V E636A, V S749C, IIb F755C, 3 D606C, 3 L717tr (3tr), and IIb G991tr (IIbtr) mutants were created by site-directed mutagenesis. The 3 D606C/L717tr double mutant (606tr) was created from 3 D606C and 3 L717tr by replacing the ApaI fragment. The IIb F755C/G991tr double mutant (755tr) was created from IIb F755C and IIb G991tr by replacing the BamHI fragment. The authenticity of the constructs was confirmed by DNA sequencing.

Cell Culture and Transfection—Chinese hamster ovary (CHO)-K1 cells were cultured in Dulbecco's modified Eagle's medium (Invitrogen) supplemented with 10% fetal calf serum (BIOSOURCE), 1% penicillin and streptomycin (Invitrogen), and 1% non-essential amino acids (Sigma-Aldrich) and maintained at 37 °C in a humidified incubator supplemented with 5% CO₂. Fifty µg of V or IIb cDNA construct was co-transfected with 50 µg of 3 cDNA construct into CHO-K1 cells by electroporation. After 48 h, the cells were detached and used for assays.

Flow Cytometry—Cells were detached with phosphate-buffered saline containing 3.5 mM EDTA. After washing, cells were incubated with mAbs in modified HEPES-Tyrode buffer (5 mM HEPES, 5 mM glucose, 0.2 mg/ml bovine serum albumin, and 1x Tyrode's solution) supplemented with 1 mM CaCl₂ and 1 mM MgCl₂ for 30 min at 4 °C. For some experiments, 1 mM GRGDS peptide was included together with mAbs. After washing, cells were incubated with R-phycoerythrin-conjugated F(ab')₂ fragment of goat anti-mouse IgG for 30 min at 4 °C. After washing, cells were resuspended in HEPES-buffered saline (10 mM HEPES, 150 mM NaCl, pH 7.4) containing 1 mM CaCl₂ and 1 mM MgCl₂, and fluorescence was measured on FACSCalibur (BD Biosciences). To compare the binding of conformation-dependent mAbs among cells expressing different V3 mutants, each mAb binding was normalized by the expression of V3. This relative mAb binding was calculated by dividing the mean fluorescent intensity obtained with each anti-LIBS mAb by the mean fluorescent intensity with non-conformation-dependent anti-3 mAb SZ21.

Fibrinogen Binding Assay—FITC labeling of human Fbg was performed as previously described (22). Briefly, after adjusting the pH of human Fbg at 1 mg/ml in phosphate-buffered saline to 8.5 with 5% Na₂CO₃, 0.01 volume of 10 mg/ml FITC in Me₂SO was added and incubated at room temperature for 10 min. FITC-labeled Fbg was separated from free FITC on a PD-10 column (Amersham Biosciences) equilibrated with HEPES-buffered saline. The concentration and fluorescence/protein ratio of FITC-labeled Fbg was calculated as previously described. Forty-eight h after transfection, cells were detached and incubated with non-functional anti-3 mAb VNR5-2 followed by a R-phycoerythrin-conjugated F(ab')₂ fragment of goat anti-mouse IgG. In some experiments, cells were treated with 10 mM dithiothreitol (DTT) prior to incubation with mAbs as described previously (21). After washing, cells were incubated with FITC-labeled Fbg (200 µg/ml) with or without 1 mM GRGDS peptide in modified HEPES-Tyrode buffer containing 1 mM CaCl₂ and 1 mM MgCl₂ or 1 mM MnCl₂ for 1 h at 4 °C. In some experiments, mAb PT25-2 was included at 10 µg/ml to activate IIb3. After washing, fluorescence was measured on FACSCalibur. The mean Fbg binding (mean fluorescence intensity in FL1 channel) to cell populations expressing high 3 (fluorescence intensity in FL2 channel > 500) was calculated. Background binding in the presence of 1 mM GRGDS peptide was subtracted to obtain specific binding.

	RESULTS

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
Swapping Stalk Region Switches the Enhancing Effect of Mn²⁺ on Ligand Binding—To locate the structure in the subunit responsible for the difference in Mn²⁺ activation, we first swapped the C-terminal stalk region consisting of the thigh, calf-1, calf-2, transmembrane (TM), and cytoplasmic domains between IIb and V. The B/V chimera has IIb -propeller joined with V stalk, whereas the V/B chimera has V-propeller joined with IIb stalk. In terms of cation-binding site, only the site at the junction between the thigh and the calf-1 domains is replaced in those chimeras, and the other four sites in the -propeller domain remain unchanged (Fig. 1A). The chimeras were transiently expressed in CHO cells. Surface expression of each chimeric molecule was monitored using mAbs PL98DF6, P2, SZ21, PT25-2, OP-G2, and AMF-7. All chimeras showed mAb binding comparable with that of wild-type IIb3 or V3 (data not shown). In the presence of 1 mM Ca²⁺/Mg²⁺ without any activators, none of the cells expressing chimeras including wild-type IIb3 or V3 significantly bound Fbg (Fig. 1B). In the presence of mAb PT25-2, which binds to IIb -propeller and activates IIb3 (21–23), Fbg bound to wild-type IIb3 and to B/V chimera, but not to wild-type V3 or to V/B chimera. This is because this particular antibody is unable to bind to V3 or V/B. In the presence of 1 mM Mn²⁺, although IIb3 did show some binding, it was not as avid as that seen in the presence of PT25-2. By contrast, Mn²⁺ induced robust Fbg binding to wild-type V3. These results are in agreement with previous reports. However, when stalk was swapped, the response to Mn²⁺ changed completely from the wild-type. The B/V bound Fbg just as avidly as wild-type V3. By contrast, V/B binding significantly decreased to a level similar to that of wild-type IIb3. These results clearly indicate that the integrin stalk, but not the head, contains the critical component that regulates Mn²⁺-induced integrin activation.

View larger version (32K): [in this window] [in a new window]   FIG. 1. The calf-2 domain in integrin subunit regulates Mn2+-induced integrin activation. A, schematic representation of the IIb/V chimeras. Positions of the domain boundaries are indicated. The specific binding of FITC-Fbg is presented in mean fluorescent intensity. B, FITC-Fbg binding to cells expressing stalk-swapping chimeras (B/V and V/B) in the presence of 1 mM Ca2+/Mg2+ (open bar), 1 mM Ca2+/Mg2+ + 10 µg/ml PT25-2 (cross-hatched bar), and 1 mM Mn2+ (filled bar) is shown. C, FITC-Fbg binding to cells expressing IIb/V domain-swapping chimeras (TC1C2, TC1, and T) in the presence of 1 mM Ca2+/Mg2+ (open bar) and 1 mM Mn2+ (filled bar) and cells pretreated with DTT in the presence of 1 mM Ca2+/Mg2+ (hatched bar) is shown. D and E, FITC-Fbg binding to cells expressing IIb/V domain-swapping chimeras (T, C1, C2, D, C1C2D, TC2D, TC1D, and TC1C2) in the presence of 1 mM Mn2+ (filled bar) and cells pretreated with DTT in the presence of 1 mM Ca2+/Mg2+ (hatched bar) is shown.

The V Calf-2 Domain, Not the IIb Calf-2 Domain, Facilitates Mn²

Mutation in the Cation-binding Site at Genu Does Not Affect Mn²⁺-induced Ligand Binding—In the V3 crystal structure, the only cation-binding site in the stalk is located at the junction between the thigh and the calf-1 domains, but not in the calf-2 domain. To examine the effect of this cation-binding site on ligand binding, Asp⁵⁹⁹ and Glu⁶³⁶, which coordinate a cation at V genu, were mutated to Ala. These two acidic residues are conserved in most integrins. Mutating individual amino acid residues to Ala did not inhibit V3-Fbg interaction induced by Mn²⁺ (Fig. 2). It did not induce constitutive activation in the presence of 1 mM Ca²⁺/Mg²⁺, either. Although these results do not rule out the possibility that this cation-binding site may play some role in ligand binding, they are consistent with the results obtained using domain-swapping chimeras.

View larger version (12K): [in this window] [in a new window]   FIG. 2. The effect of mutation in the cation-coordinating amino acid residues on the V3-ligand interaction. Asp599 and Glu636, which coordinate a cation at V genu, were individually mutated to Ala. FITC-Fbg binding to CHO cells expressing mutant V3 in the presence of 1 mM Ca2+/Mg2+ (hatched bar) and 1 mM Mn2+ (filled bar) is shown.

EGF4-TD Interface Regulate Mn²

Constraining the Calf-2/EGF4-TD Interface in V3 Inhibits Mn²⁺-induced Activation—If calf-2/EGF4-TD interface interaction is involved in the regulation of Mn²⁺-induced ligand binding, constraint in the interface should have a profound impact on integrin activation and conformation. To examine this hypothesis, we introduced an artificial disulfide bridge to put constraint in this interface. As seen in Fig. 4A, V Ser⁷⁴⁹ and 3 Asp⁶⁰⁶ are physically close, and if these residues are mutated to Cys at the same time, we will be able to introduce a disulfide bridge. Although individual V S749C or 3 D606C did not have any significant effect, double cysteine mutation 749/606 reduced Fbg binding to about one-third (Fig. 4B). This suppression was neutralized when the cells were treated with DTT to disrupt the disulfide bridge. These results suggest that restraining the possible movements in this membrane-proximal domain interface has a significant inhibitory effect on Mn²⁺-induced integrin activation.

Next we examined the effect of the constraint on the conformational change of V3. To monitor the conformational change, we examined the binding of anti-LIBS antibodies. The anti-LIBS mAbs bind preferentially to ligand-occupied form of integrins. Thus, they are believed to detect conformational changes associated with ligand binding (19). In this assay, the binding of non-functional anti-3 mAb VNR5-2 did not change among those mutants. The presence of 1 mM GRGDS peptide did not significantly affect the binding, either (Fig. 4C). The binding of anti-LIBS1 was significantly weaker than that of VNR5-2. However, the binding of anti-LIBS1 showed a robust increase in the presence of 1 mM GRGDS peptide. No significant difference was observed among wild-type and cysteine mutants (Fig. 4D). By contrast, anti-LIBS2 binding in the presence of GRGDS peptide was suppressed in the double mutant (Fig. 4E). Although the localization of the epitope for anti-LIBS1 is not known, the epitope for anti-LIBS2 has been located within amino acid residues 602–690 in the TD (25). These results suggest that the constraint on the membrane-proximal / interface restricts the conformational change within the membrane-proximal domain upon ligand binding. However, it may not prevent the conformational change in other parts of the molecule.

View larger version (50K): [in this window] [in a new window]   FIG. 3. The effect of IIb to V mutation in amino acid residues located at the calf-2/EGF4-TD interface on the IIb3-ligand interaction. A, amino acid residues 753–755, 760–764, 774, 781–783, 787, 899–900, 902–904, and 958–960 located at the calf-2/EGF4-TD interface in IIb were mutated to the corresponding residues in V. The resulting IIb3 mutants are designated B/V 753–755, B/V 760–764, B/V 774, B/V 781–783, B/V 787, B/V 899–900, B/V 902–904, and B/V 958–960, respectively. B, amino acid residues in the IIb 760–764 and 899–900 regions were individually mutated to the corresponding residues in V. C, B/V 760–764, B/V 787, and B/V 899–900 or B/V 900 mutations were combined. These IIb3 mutants were expressed in CHO cells. FITC-Fbg binding to cells expressing IIb3 in the presence of 1 mM Mn2+ (filled bar) and cells pretreated with DTT in the presence of 1 mM Ca2+/Mg2+ (hatched bar) is shown. As a reference, Fbg binding to cells expressing wild-type IIb3 and C2 chimera are also shown. D, the spacefill model of V3 crystal structure. The V chain is shown in green, and the 3 chain is shown in red. The membrane-proximal domains are magnified on the left. The calf-2 domain is shown in light green, and the EGF4 and TD are shown in transparent red. Positions of amino acid residues corresponding to 753–755, 760–764, 774, 781–783, 787, 900, 902–904, and 958–960 in IIb are shown and labeled. Amino acid residues 787/900 and 760–764, which promoted Mn2+-induced Fbg binding when mutated to V residue, are shown in yellow and orange, respectively. Amino acid residues 753–755, 774, and 781–783, which did not affect Fbg binding, are in cyan. Amino acid residues 902–904, which jeopardized surface expression and Fbg binding, are shown in blue. The models were created using RasMol.

EGF4-TD Interface in IIb3 Completely Abrogates Inside-out Signaling—

The effect of the constraint is far more dramatic in inside-out signaling. To mimic inside-out signaling, the cytoplasmic domain of 3 was truncated at the membrane-proximal Leu⁷¹⁷ (26). The truncated IIb3 (IIb/3tr) bound Fbg in the presence of Ca²⁺/Mg²⁺ without any activators, as reported previously (Fig. 5B). When individual cysteine mutation IIb F755C or 3 D606C was introduced in truncated IIb3 (designated 755/3tr and IIb/606tr, respectively), it did not significantly block Fbg binding. In contrast, Fbg binding was completely abrogated in double cysteine mutant 755/606tr. Activating antibody did not override the inhibition completely. Fbg binding was completely restored only when the disulfide bridge was disrupted with DTT treatment. The cysteine mutations had similar effects when IIb3 was activated by IIb cytoplasmic domain truncation (Fig. 5C). Although individual IIb F755C or 3 D606C mutation (775tr/3 and IIbtr/606, respectively) did not have a significant impact as compared with wild-type (IIbtr/3), double cysteine mutation 755tr/606 completely abrogated Fbg binding. This suppression was completely neutralized with DTT treatment. Thus, disruption of the membrane-proximal domain interface is essential for propagating inside-out signaling. These results suggest that constraining the membrane-proximal / interface not only restricts integrin activation by divalent cations but also prevents activation by inside-out signaling.

	DISCUSSION

TOP ABSTRACT INTRODUCTION EXPERIMENTAL PROCEDURES RESULTS DISCUSSION REFERENCES

 
In this study, we provide evidence that interdomain interactions in the integrin stalk regions, particularly in the membrane-proximal domains, are the key regulator of integrin activation. By using Mn²⁺-induced Fbg binding to 3 integrins as a model system, we demonstrated the following: 1) that switching stalk region resulted in switching the Mn²⁺ requirement for ligand binding in IIb3 and V3, 2) that the calf-2 domain in the stalk region is the sole determinant for the distinct response to Mn²⁺, 3) that this response is mediated in part by the specific amino acid residues in the calf-2 domain that create the calf-2/EGF4-TD interface, and 4) that constraining the calf-2/EGF4-TD interface not only inhibited Mn²⁺-induced integrin activation but also completely abrogated inside-out signaling. These results suggest that the extracellular, membrane-proximal / interaction plays important roles in regulating the cation-dependent integrin activation and critically controls the propagation of inside-out signaling.

View larger version (72K): [in this window] [in a new window]   FIG. 4. The effect of constraint in the membrane-proximal / interface in V3 on ligand binding and conformation. A, ribbon models depicting the crystal structure of the V3 extracellular domain. The and subunits are shown in cyan and red, respectively. The amino acid residues Ser749 in V and Asp606 in 3 that were mutated to Cys to facilitate disulfide bridge formation in the study are shown in yellow and orange spacefills, respectively. The model was created using RasMol. B, FITC-Fbg binding to cells expressing individual Cys mutant VS749C, 3D606C, and double Cys mutant 749/606 in the presence of 1 mM Ca2+/Mg2+ (open bar) and 1 mM Mn2+ (filled bar) and cells pretreated with DTT in the presence of 1 mM Mn2+ (hatched bar) is shown. C-E, mAb binding to cells expressing individual Cys mutant VS749C, 3D606C, and double Cys mutant 749/606 in the presence 1 mM Ca2+/Mg2+ (filled bar) and 1 mM Ca2+/Mg2+ + 1 mM GRGDS peptide (hatched bar) is shown. Relative mAb binding was calculated as described under "Experimental Procedures." The binding of VNR5-2, anti-LIBS1, anti-LIBS2 is shown in C, D, and E, respectively.

Several lines of evidence implicate the role of the calf-2 domain in integrin activation. Some integrin subunit is proteolytically cleaved into covalently linked heavy and light chains during post-translational processing. Prevention of this endoproteolytic cleavage by mutagenesis abrogated 61 activation by phorbol myristate acetate (31). Neutrophil elastase cleaves IIb between Val⁸³⁷ and Asp⁸³⁸, close to the endoproteolytic cleavage site in the calf-2 domain (32). Although the elastase does not, by itself, induce platelet aggregation, it greatly potentiates IIb3 activation by cathepsin G and Mn²⁺ (17, 33). Thus, structural alterations in the calf-2 domain affect integrin activation. Conversely, ligand binding provokes structural change in the calf-2 domain. Upon ligand binding, the integrin stalk regions express neo-epitopes that are recognized by a group of anti-LIBS mAbs. These epitopes are mostly located in the stalk region (25, 34). Notably, one such mAb, PMI-1, recognizes an epitope within residues 842–856 of the IIb calf-2 domain (35). Thus, structural change in the calf-2 domain is closely associated with integrin activation. How does the calf-2 domain affect Mn²⁺-induced Fbg binding, despite the fact that it does not contain actual cation-binding sites? Previous studies suggest that the membrane-proximal / interaction modulates integrin affinity regulation by divalent cations. Kinetic studies using recombinant truncated 51 and V3 with an artificial C-terminal clasp suggest that / interaction in the TM and cytoplasmic domains decreases integrin affinity by increasing the dissociation rate. Electron microscopic observation revealed that the two stalks are widely separated at their ends in unclasped 51 (36), whereas they are still connected at their ends in a majority of molecules in unclasped recombinant V3, even in the presence of Mn²⁺ (4). These reports suggest that the extracellular, membrane-proximal / interface formation is significant in some integrins under physiological conditions and may account for the distinct response to Mn²⁺ in 3 integrins. Indeed, IIb to V mutation in selected amino acid residues that make up the calf-2/EGF4 and calf-2/TD interface greatly facilitated Mn²⁺-induced Fbg binding in IIb3. Stabilizing this interface with a disulfide bridge (749/606 mutant) prevented Mn²⁺-induced Fbg binding to V3. It is possible that constraining the membrane-proximal / interface may somehow constrain integrin in a bent state, thereby keeping integrin in a low affinity state. However, the 749/606 and 755/606 double mutants did undergo conformational change upon ligand binding, as detected by LIBS epitope expression. These mutants still showed some Fbg binding when stimulated by Mn²⁺ or activating mAb. In contrast, constraining the integrin in a bent state by introducing a disulfide bridge between A and TD resulted in failure to induce LIBS epitope expression upon ligand binding and complete loss of ligand binding, even in the presence of activating mAb.² Thus, constraining the / stalks and head/ stalk has a distinct conformational effect on integrin structure. In addition, the fact that extended, clasped conformer can be observed under electron microscopy (4) indicates that constraining the / stalks does not necessarily constrain integrin in a bent state. On the contrary, disruption of the membrane-proximal / interface may lead to integrin activation. Anti-LIBS2 that binds to TD has been shown to activate IIb3 (25). Likewise, two mAbs that bind to IIb calf-2 domain weakly induced Fbg binding in the presence of Ca²⁺/Mg²⁺. Notably, these mAbs potentiated Mn²⁺-induced Fbg binding.³ Disruption of the Cys⁶⁰⁸-Cys⁶⁵⁵ and Cys⁶⁶³-Cys⁶⁸⁷ disulfide bridges in TD that face the calf-2/EGF4-TD interface in the crystal structure constitutively activated IIb3, whereas disruption of the other two disulfide bridges did not (37). These results implicate that the membrane-proximal calf-2/EGF4-TD interface may act as a physiological clasp and thus regulate integrin activation.

View larger version (28K): [in this window] [in a new window]   FIG. 5. The effect of constraint in the membrane-proximal / interface on ligand binding to IIb3. A, amino acid residue Phe755 in IIb, which is homologous to Ser749 in V, was mutated to Cys (F755C) to facilitate disulfide bridge formation between IIb and 3. FITC-Fbg binding to cells expressing IIb3 (IIb/3) carrying individual IIb F755C mutation (755/3), 3 D606C mutation (IIb/606), and double IIb F755C and 3 D606C mutation (755/606) in the presence of 1 mM Ca2+/Mg2+ (open bar) and 1 mM Ca2+/Mg2+ plus 10 µg/ml PT25-2 (hatched bar) and cells pretreated with DTT in the presence of 1 mM Ca2+/Mg2+ (filled bar) is shown. B, to mimic inside-out signaling, 3 was truncated at Leu717 by introducing a stop codon in 3 cDNA (3tr). The D606C mutation was introduced in 3tr (606tr). Wild-type IIb or IIb carrying F755C mutation was expressed together with 3tr or 606tr in CHO cells. FITC-Fbg binding to cells expressing truncated IIb3 (IIb/3tr) carrying IIb F755C mutation (755/3tr), 3 D606C mutation (IIb/606tr), and IIb F755C and 3 D606C mutations (755/606tr) in the presence of 1 mM Ca2+/Mg2+ (open bar) and 1 mM Ca2+/Mg2+ plus 10 µg/ml PT25-2 (hatched bar) and cells pretreated with DTT in the presence of 1 mM Ca2+/Mg2+ (filled bar) is shown. C, to mimic inside-out signaling, IIb was truncated at Gly991 by introducing a stop codon in IIb cDNA (IIbtr). The F755C mutation was introduced in IIbtr (755tr). Wild-type 3 or 3 carrying the D606C mutation was expressed together with IIbtr or 755tr in CHO cells. FITC-Fbg binding to cells expressing truncated IIb3 (IIbtr/3) carrying IIb F755C mutation (755tr/3), 3 D606C mutation (IIbtr/606), and IIb F755C and 3 D606C mutations (755tr/606) in the presence of 1 mM Ca2+/Mg2+ (open bar) and 1 mM Ca2+/Mg2+ plus 10 µg/ml PT25-2 (hatched bar) and cells pretreated with DTT in the presence of 1 mM Ca2+/Mg2+ (filled bar) is shown.

In inside-out signaling, how the interactions within the short cytoplasmic tails trigger structural rearrangements in the large extracellular domain remains a mystery. The cytoplasmic, membrane-proximal interaction constrains integrin in an inactive state (40). Recently, the talin head domain has been shown to activate IIb3 by disrupting the membrane-proximal IIb/3 tail interface in vitro (43). In living cells, the L/2 cytoplasmic tails undergo separation upon agonist stimulation (44). Conversely, in outside-in signaling, ligand binding induces a structural rearrangement in the cytoplasmic tails and induces separation of the tails (44) and exposure of the LIBS epitope in tail (45). Our results unequivocally indicate that separation of the cytoplasmic tails is unable to induce integrin activation unless the extracellular stalks separate from each other. In agreement, previous studies have shown that IIb3 stalks undergo structural rearrangement and change their relative distance or orientation on the cell surface upon agonist stimulation of platelets (46). Thus, disruption of the calf-2/EGF4-TD interface that follows the / cytoplasmic tail separation triggers a cascade of structural rearrangement of integrin extracellular domains in inside-out signaling. Most likely, disruption of the calf-2/EGF4-TD interface would be required for the propagation of outside-in signaling as well.

View larger version (30K): [in this window] [in a new window]   FIG. 6. A relay switch model of integrin activation. The and subunits are illustrated in blue and green, respectively. The -propeller, thigh, calf-1, and calf-2 domains in are labeled as P, T, C1, and C2, respectively, and the A, hybrid, EGF3, EGF4, and TD domains in are labeled as A, H, E3, E4, and TD, respectively. The PSI, EGF1, and EGF2 domains in are not shown for simplicity. The putative interactions in the head/ stalk, stalk/ stalk, and cytoplasmic tail/ cytoplasmic tail interfaces are shown as a red bar. The eight cation-binding sites are shown as white circles. Manganese ions are depicted as yellow circles. Destabilization and disruption of the interface are shown as a black and a white break in the bar, respectively. The low affinity form is depicted as bent conformer. The high affinity form is depicted as extended, open, unclasped conformer. The hypothetical intermediate affinity forms in 3 integrin are depicted as the bent, unclasped form. Alternatively, the extended, closed, clasped form (in outside-in signaling) and the extended, closed, unclasped form (in inside-out signaling) shown within the broken rectangle may be the predominant intermediates in other integrins. All conformers are in structural equilibrium with each other. The presence of ligands shifts the equilibrium toward the formation of the high affinity conformer.

Our results suggest that any endogenous molecules or reagents that potentially affect / interface formation may modulate integrin activation. Indeed, CD151, platelet-derived growth factor receptor , vascular endothelial growth factor receptor 2, and CD47 have been shown to associate with integrin extracellular domains (51–54). These molecules may affect integrin activation by interfering with or stabilizing intersubunit or intrasubunit interface formation by binding to integrin stalks. Likewise, chemicals or reagents that cross-link and stabilize the calf-2/EGF4-TD interface will be able to selectively inhibit inside-out signaling (and probably outside-in signaling as well) without inhibiting low affinity ligand binding or inducing LIBS epitope expression. Thus, membrane-proximal / interaction can be a novel target for effective anti-integrin therapies.

	FOOTNOTES

 
^* This work was supported by a health and labor science research grant for research on regulatory science of pharmaceuticals and medical devices from the Ministry of Health, Labor and Welfare; a grant for leading project for biosimulation from the Ministry of Education, Culture, Sports, Science and Technology; a grant from Keio Gijuku Fukuzawa Memorial Fund for the advancement of education an