Introduction

Summary Introduction Procedures Results Discussion References

The integrin family of receptors mediate many of the cell-cell and cell-substratum interactions that are central to cell adhesion, migration, growth, and differentiation. Integrins are noncovalent / heterodimers. Each subunit contains a large extracellular region, a transmembrane domain, and a short cytoplasmic tail (1). Integrins bind to a wide variety of ligands, including extracellular matrix proteins, counter-receptors on other cells, and circulating plasma proteins (2). The affinity and specificity of an integrin binding site are defined by the specific pairing of the  and  subunits (3). In addition, previous studies support a model in which amino acid sequences in both subunits coordinate ligand and cations in close proximity to form a "reactive" center for ligand binding (4). The cytoplasmic tails of integrins interact with intracellular proteins, including cytoskeletal proteins such as talin and -actinin (5) and a number of regulatory proteins such as focal adhesion kinase (6), integrin-linked kinase (7), endonexin (8), and cytohesin-1 (9). Upon ligand binding, integrin-mediated signaling events, which include rearrangement of the cytoskeleton, gene regulation, and cellular differentiation, are induced by a process called "outside-in" signaling (10). Alternatively, intracellular signaling events can modulate the affinities of integrins for extracellular ligands (11). These pathways involve phospholipids, protein kinases (5, 6), intracellular calcium fluxes, and low-molecular weight G proteins (12, 13). The process whereby cytoplasmic signals result in changes in receptor conformation and ligand binding affinity is termed "inside-out" signaling. The modulation of outside-in and inside-out signals is important for the regulation of integrin function (2).

Platelet adhesive interactions are of primary importance in normal hemostasis as well as thrombotic disorders. IIb3 is the major integrin involved in attachment, spreading, and aggregation of platelets. On resting platelets, IIb3 is in a "latent" or basal state that does not bind fibrinogen, one ligand present in abundance in the circulation (14). A wide variety of agonists, such as thrombin, ADP, or collagen, can stimulate platelets, which results in the "activation" of IIb3 and the binding of soluble fibrinogen or other ligands (including von Willebrand factor, vitronectin, and thrombospondin) that are important for thrombus formation (15). This inside-out signaling is mediated by conformational changes in IIb3 and is modulated by intracellular events (2, 5). In addition to these agonists, synthetic peptides containing the RGD sequence, which is present within the fibrinogen molecule and serves as a recognition site for binding to IIb3, activate IIb3 (16). RGD peptides bind to resting IIb3, leading to conformational changes in its extracellular domain that enable it to bind soluble fibrinogen, after the removal of the RGD peptide. IIb3 activated by RGD peptides expresses novel sites on its extracellular domain (termed ligand-induced binding sites), whereas IIb3 activated by agonist does not express ligand-induced binding sites unless it binds fibrinogen (17, 18). These findings indicate that IIb3 has at least two distinct conformational states that can bind soluble fibrinogen, i.e., agonist- and RGD-activated states (19).

To analyze the conformational states of IIb3, the ligand-binding properties of recombinant human platelet IIb3 expressed in HEK 293 cells were examined by using synthetic peptidic and nonpeptidic RGD mimics. These molecules have differential specificity for agonist-activated and resting platelets and thus can be used as markers to identify activation states of IIb3. Here, we identify IIb3 in HEK 293 cells to be in a transitional activation state that is distinct from the fully activated receptor in platelets. This is confirmed by using the activation-specific, anti-IIb3 mAb PAC1. Furthermore, Glanzmann's thrombasthenic mutations, which disrupt either ligand binding [3(D119Y)] or receptor signaling [3(S752P)] (20), do not permit binding to PAC1 and abrogate the physiological ligand-binding functions of IIb3. However, deletion of the cytoplasmic domain of the  subunit is sufficient to convert IIb3 in HEK 293 cells to a fully competent receptor, as in activated platelets. The full-length receptor, which is in a transitional activation state, is also shown to mediate agonist-independent aggregation of cells in the presence of fibrinogen. This fibrinogen-mediated aggregation requires an intact ligand binding domain in IIb3 and is dependent on cytoskeletal rearrangement. These results suggest that IIb3 can exist in multiple transitional conformational states, which may modulate specific physiological functions of the integrin in platelets.

	Experimental Procedures

Summary Introduction Procedures Results Discussion References

Antibodies. Mouse mAbs to IIb3 (mAb CA3) and 51 (mAb JBS5) were purchased from Chemicon (Temecula, CA). mAb PAC1 (21), against the activated form of IIb3, was purchased from the Cell Center, University of Pennsylvania (Philadelphia, PA). Polyclonal antibodies against purified human platelet IIb3 were generated in rabbits and affinity-purified. These antibodies specifically bind to IIb3 in platelets and human erythroleukemia cells.¹ FITC-conjugated donkey anti-mouse IgM and FITC-conjugated goat anti-mouse IgG were purchased from Jackson Immunoresearch Laboratories (West Grove, PA).

Synthetic ligands. The IIb3 antagonists MK-852, L-692,884, and L-734,217 were synthesized by the Medicinal Chemistry Department, Merck Research Laboratories (West Point, PA). L-692,884 [cyclo-4-iodo-benzoyl-(Cys-Asn-Pro-Arg-Gly-Asp-Cys)-OH], a synthetic cyclic RGD peptide, binds preferentially to activated IIb3 (22). MK-852 [cyclo-N-acetyl-[Cys-Asn-(5,5-dimethyl-4-thiazolidinecarbonyl)-(4-aminomethyl-Phe)-Gly-Asp-Cys]-OH] and L-734,217 [N-[3(R)-(2-(piperidin-4-yl)ethyl)-2-piperidon-1-yl]acetyl-3(R)-methyl--Ala] bind specifically to activated IIb3 in platelets (23, 24). IIb3-specific antagonists Ro 43-5054 and Ro 44-9883 were also synthesized by the Medicinal Chemistry Department and characterized as previously described (25). The detailed characterization and binding properties of these fibrinogen receptor antagonists with purified human platelet IIb3 are described elsewhere² and are summarized in Table 1.

View this table: [in this window] [in a new window]   TABLE 1 Dissociation binding constants of RGD mimics with purified human IIb3 The ratio of differentiation is defined as the ratio of the Kd values for the unactivated and activated forms. It is a qualitative measure of the ability of the compounds to distinguish between the two forms of the receptor.

cDNA constructs. The full-length human cDNA encoding IIb in the expression vector pM2ADA was a kind gift from Dr. Joel Bennett (University of Pennsylvania, Philadelphia, PA) (26). The cDNA insert was subcloned into the eukaryotic expression vector pR135, which was under the transcriptional control of the cytomegalovirus promoter and contained the hygromycin-selectable marker. The isolation and construction of cDNA encoding full-length human platelet 3, 3(717), 3(D119Y), and 3(S752P) have been previously described (27). The mutation 3(693) was introduced by polymerase chain reaction using a 5 primer (CAGCTCGAGCTATTAGTCAGGGCCCTTAGGGACACTCTGG) that contained a PstI restriction site and the 3 oligonucleotide (TGCCATTGGGCCTCATA) that contained an XhoI site. The resulting polymerase chain reaction fragment was digested with PstI and XhoI. The full-length 3 cDNA was digested with HindIII and PstI. The expression vector pCDNA3 (Invitrogen, CA) was digested with HindIII and XhoI. Ligation of the three resulting fragments generated a stop codon before the transmembrane domain of the 3 cDNA. The 3 cDNA constructs were subcloned into the expression vector pCDNA3, containing the neomycin-selectable marker. All constructs were characterized by restriction digestion, purified by CsCl centrifugation, and verified by DNA sequence analysis before transfection.

Cell culture and transfection. HEK 293 cells were obtained from the American Type Culture Collection (Rockville, MD). HEK 293 cells were grown in minimal essential medium with Earle's salt supplemented with 10% fetal calf serum, 1% kanamycin, and 2 mM glutamine (GIBCO-BRL Life Technologies, Gaithersburg, MD).

Flow cytometry. Surface expression levels of integrins were analyzed by single-color flow cytometry. Cells (2 × 10⁵) were harvested with trypsin/EDTA (GIBCO-BRL), washed once with 5 volumes of minimal essential medium with Earle's salt containing 10% fetal calf serum and twice with DPBS, and incubated with either 20 µg/ml mAb CA3 (anti-IIb3) or 15 µg/ml mAb JBS5 (anti-51), in DPBS containing 1 mM CaCl₂ and 1% BSA, for 45 min at 4°, in a total volume of 100 µl. The cells were pelleted, washed once with DPBS, and incubated with FITC-conjugated goat anti-mouse IgG. After a 45-min incubation at 4°, the cells were washed once with DPBS, resuspended in 350 µl of flow cytometric buffer (100 mM HEPES, pH 7.5, 150 mM NaCl, 3 mM KCl, 1 mM CaCl₂), and analyzed by flow cytometry. The light scatter and fluorescence intensity of 10,000 cells were collected using logarithmic gain.

Cell attachment. The cell attachment assay was performed as previously described (27). Essentially, 96-well plates were coated with fibrinogen (5 µg/ml), vitronectin (1.5 µg/ml), or fibronectin (4 µg/ml) in DPBS. Cells were harvested, washed three times with serum-free minimal essential medium with Earle's salt, and then resuspended in attachment solution (calcium- and magnesium-free Hanks' balanced salt solution, 20 mM HEPES, 1 mg/ml heat-inactivated BSA, 1 mM CaCl₂, 1 mM MgCl₂). Cells (1 × 10⁴) were added to each well and allowed to attach for 1-2 hr at 37° in a humidified 5% CO₂ incubator. Unattached cells were washed with Hanks' balanced salt solution. Attached cells were determined by colorimetric development of glucosaminidase activity. The number of attached cells was quantitated spectrophotometrically at 405 nm in triplicate, according to a standard curve.

Surface labeling and immunoprecipitation. Transfectants (2 × 10⁶ cells) were surface-labeled with 2 mM Immunopure Sulfo-NHS-LC-Biotin (Pierce, Rockford, IL) and then solubilized in RIPA buffer (50 mM Tris, pH 7.5, 150 mM NaCl, 1 mM CaCl₂, 1% Nonidet P-40, 0.5% deoxycholate, 0.1% SDS) containing 1 mM phenylmethylsulfonyl fluoride, 10 µg/ml aprotinin, and 100 µg/ml leupeptin. Cell extracts were immunoprecipitated overnight with rabbit polyclonal anti-IIb3 antibodies, followed by protein A-Sepharose for 2 hr at 4°. The protein A-Sepharose beads were pelleted, washed in RIPA buffer, resuspended in sample buffer (50 mM Tris·HCl, pH 6.8, 2% SDS, 0.002% bromphenol blue, 10% glycerol), and boiled for 5 min. After centrifugation, immunoreactive proteins were resolved by reducing 8% SDS-polyacrylamide gel electrophoresis (Novex, San Diego, CA). The proteins were transferred to nitrocellulose, stained with horseradish peroxidase-conjugated streptavidin (Amersham, Arlington Heights, IL), and developed with the enhanced chemiluminescence system (NEN-DuPont, Boston, MA).

Ligand binding assays. Saturation binding studies were performed using 1 × 10⁴ cells/tube and increasing concentrations of ¹²⁵I-L-692,884 in the presence or absence of unlabeled L-692,884 (1 µM), in a total volume of 200 µl, as described below. Using the LIGAND program, IIb3 cells were shown to express, on average, 4 × 10⁵ receptors/cell (K_d = 12 × 10⁹ M), whereas IIb3(717) cells express 7.5 × 10⁵ receptors/cell (K_d = 2 × 10⁹ M) (data not shown).

Aggregation of recombinant cells. Agonist-independent aggregation of cells was performed as previously described (18). Typically, 100 µl of cells (1 × 10⁷/ml in DPBS containing 1 mM CaCl₂) were added to wells of a 24-well tissue culture plate in the presence or absence of inhibitor (1 µM Ro 43-4054 or L-734,217), in a total volume of 200 µl, and were allowed to remain at room temperature for 30 min. Fibrinogen (1 µM; Sigma Chemical Co., St. Louis, MO) was added, the contents of the wells were mixed by hand-swirling, and the plates were then subjected to gyrorotation at 100 rpm for 30 min. Aggregation was stopped after 20 min by addition of 2% paraformaldehyde (150 µl). The plate was allowed to remain at room temperature for 15 min before analysis by light microscopy. Aggregation was monitored as the formation of aggregates during the rotary agitation. In some cases, cells were pretreated with 0.5 µM cytochalasin D (Sigma) or 0.1% dimethylsulfoxide for 30 min at 4° before addition of fibrinogen. All experiments were performed in triplicate.

	Results

Summary Introduction Procedures Results Discussion References

Surface expression of IIb3 and mutants in HEK 293 cells. Full-length IIb3 and its mutants were expressed in HEK 293 cells (Fig. 1). Parental HEK 293 cells do not express IIb or 3, as shown in Fig. 1A. To obtain high levels of integrin expression, we first transfected IIb cDNA into HEK 293 cells and selected clones that expressed high IIb mRNA levels (data not shown). The full-length 3 subunit and its mutants were then transfected into these cell lines expressing IIb. The 3 mutants included those with deletion of the cytoplasmic domain [3(717)] or deletion of the transmembrane and cytoplasmic domains [3(693)] and the two variants of Glanzmann's mutants [3(D119Y) and 3(S752P)].

View larger version (23K): [in this window] [in a new window]   Fig. 1.   Surface expression of recombinant IIb3 in HEK 293 cells. A, The expression pattern of IIb3 and 51 in the presence or absence of 2 mM DTT was determined by flow cytometry. Cells (2 × 104) transfected with cDNAs coding for the  and  subunits noted were stained with either anti-IIb3 (mAb CA3) or anti-51 (mAb JBS5) and then incubated with FITC-conjugated goat anti-mouse IgG. The cells were then subjected to flow cytometry as described in Experimental Procedures. Filled histograms, cells labeled with anti-IIb3; open histograms, cells labeled with anti-51. Note that all IIb3 heterodimers were expressed at similar levels. B, Cell extracts of the biotin-surface-labeled HEK 293 cells were immunoprecipitated with polyclonal anti-IIb3 antibody. The immunoprecipitated proteins were resolved by 8% SDS-polyacrylamide gel electrophoresis under reducing conditions, transferred, and visualized as described in Experimental Procedures. Deletion of the cytoplasmic domain of the 3 subunit leads to a shift in the mobility of the 3 subunit bands on the gels. Compared with the level of receptor expression determined by flow cytometry, the decrease in the level of 3 subunit in the point mutants (D119Y and S752P) may be the result of the decrease in the efficiency of biotinylation, rather than the level of expression.

View larger version (31K): [in this window] [in a new window]   Fig. 2.   Cell attachment assay of cells expressing IIb3 constructs. Cells (1 × 104/well) were plated on 96-well plates that had been precoated with fibrinogen (A), vitronectin (B), or fibronectin (C), as described in Experimental Procedures. The numbers of adherent cells after a 2-hr incubation at 37° are reported as the means of triplicate determinations.

Discrimination, by binding affinities, of the transitional state of IIb3 in HEK 293 cells. The cyclic RGD peptide L-692,884, cyclo-4-iodo-benzoyl-(Cys-Asn-Pro-Arg-Gly-Asp-Cys)-OH, binds differentially to activated (K_d = 1.4 nM) and unactivated (K_d = 75 nM) IIb3 in human platelets (22).² To evaluate the affinity state of the receptor, we used a number of IIb3 antagonists (RGD mimics) (Table 1) to displace the binding of radioiodinated L-692,884 from recombinant cells. These antagonists fall into two categories, i.e., ones that selectively bind to the activated IIb3 (MK-852, L-734,217, L-692,884, and Ro 43-5054) and others that do not discriminate between the activated and unactivated IIb3 (echistatin and Ro 44-9883). Echistatin and L-692,884 bind to both IIb3 and V3¹; all of the other compounds listed in Table 1 are specific for IIb3.²

View larger version (19K): [in this window] [in a new window]   Fig. 3.   Binding of 125I-L-692,884 to IIb3 transfectants. A, Specific binding of 125I-L-692,884 to cell lines, determined as described in Experimental Procedures. Cells (5 × 104) were incubated with 20 pM 125I-L-692,884 in the absence or presence (1 µM) of inhibitor (echistatin or L-692,884). The radiolabeled ligand bound specifically only to cells expressing IIb3, IIb3(717), or IIb3(693). B, Displacement of 125I-L-692,884, in HEK 293 cells expressing IIb3, by echistatin, L-692,884, MK-852, Ro 44-9883, Ro 43-5054, or L-734,217.

View this table: [in this window] [in a new window]   TABLE 2 Competitive binding of RGD mimics to HEK 293 cells expressing IIb3 and IIb3(717) Inhibition of 125I L-692,884 binding to HEK 293 cells expressing IIb3 or IIb3(717) was determined at various concentrations of the RGD mimics. Cells (2.5 × 104) were incubated with 20 pM 125I L-692,884 in the presence of increasing concentrations of RGD mimics, as described in Experimental Procedures. The concentrations of the compounds inhibiting binding of 125I L-692,884 by 50% (IC50) were determined by a four-parameter nonlinear analysis and are represented as mean ± standard error. Under the experimental conditions, IC50 is equal to the dissociation binding constant (Kd). The ratio of differentiation is defined as the ratio of the IC50 (Kd) values for IIb3 and IIb3(717).

PAC1 binding to IIb3 in HEK 293 cells. PAC1 is a murine IgM antibody specific for the high affinity conformation of IIb3 (21). PAC1 mimics the ligand binding characteristics of the natural ligand fibrinogen (30) and fails to bind to ligand binding-defective mutants of IIb3 (20). Consequently, PAC1 binding to the receptor mimics the binding of the physiological soluble ligand fibrinogen and serves as a marker for the activated state of IIb3.

View larger version (19K): [in this window] [in a new window]   Fig. 4.   PAC1 binding to activated platelets and HEK 293 cells expressing IIb3. A, PAC1 binding to platelets in the absence (filled histograms) or presence (open histograms) of Ro 43-5054 (1 µM). Resting (1 and 2) or thrombin-activated (3 and 4) platelets were treated with 2 mM DTT (2 and 4) or not treated (1 and 3) before incubation with PAC1. PAC1 does not bind to resting platelets and binds specifically to activated platelets or DTT-treated resting platelets. B, Flow cytometric histograms illustrating PAC1 binding to HEK 293 cells expressing wild-type IIb3 and the mutants, in the presence (open histograms) or absence (filled histograms) of the competitive inhibitor Ro 43-5054.

View larger version (58K): [in this window] [in a new window]   Fig. 5.   AI values of IIb3. To obtain a numerical estimate of integrin activation, an AI was calculated for each of the IIb3 mutants, as described in Experimental Procedures. PAC1 binding was measured in HEK 293 cells expressing wild-type IIb3, IIb3(717), or IIb3(693). Depicted are the mean AI ± standard error of three independent experiments.

Aggregation of cells. Aggregation of platelets requires the presence of IIb3 in the activation state capable of binding to fibrinogen or other adhesive macromolecules with high affinity (14). In CHO cells, recombinant IIb3 exhibits fibrinogen-dependent aggregation only after the addition of activating antibodies. This fibrinogen-dependent aggregation has also been observed for recombinant cells transfected with an IIb3 construct lacking the entire cytoplasmic domain of the  subunit (19). In the present study, HEK 293 cells expressing the full-length IIb3 were capable of promoting fibrinogen-dependent aggregation in the absence of activating antibodies (Fig. 6A). This IIb3-mediated, fibrinogen-dependent aggregation of HEK 293 cells was completely blocked by L-734,217, an IIb3-specific inhibitor (Fig. 6B). Furthermore, an intact ligand binding site was required for this fibrinogen-dependent aggregation, because cells expressing the Glanzmann's mutation IIb3(D119Y) failed to aggregate (Fig. 6A). Aggregation of cells was completely inhibited by the presence of cytochalasin D (0.5 µM) (Fig. 6B), which indicates that cytoskeletal reorganization is required for aggregation. These results suggest that in HEK 293 cells the full-length IIb3 is capable of mediating agonist-independent cell aggregation.

View larger version (104K): [in this window] [in a new window]   Fig. 6.   Agonist-independent aggregation of cells expressing IIb3. A, Cells expressing IIb3 and parental HEK 293 cells were subjected to aggregation in the presence or absence of 1 mM fibrinogen. The cells were subjected to gyrorotation for 30 min at room temperature, and the aggregation was stopped by the addition of (2%) paraformaldehyde. B, Aggregation of recombinant cells was also performed in the absence (A) or presence (B) of 1 µM L-734,217 or in the absence (C) or presence (D) of cytochalasin D (Cyt D), as described in Experimental Procedures.

	Discussion

Summary Introduction Procedures Results Discussion References

Adhesion is controlled by the kinetics of integrin binding to extracellular matrix. Cells can rapidly change integrin function by altering the binding affinity as well as the avidity of integrins for ligands (3). This modulation of the binding characteristics of integrins is cell type specific and depends on the cytoplasmic domains of widely divergent structures (5).

In this study, we have examined the expression of recombinant IIb3 and its mutants in HEK 293 cells. Cells expressing IIb3, IIb3(717), or IIb3(693) preferentially adhered to immobilized fibrinogen. In contrast, parental HEK 293 cells and cells expressing the Glanzmann's mutants, with mutations in either the ligand binding site [IIb3(D119Y)] or the receptor signaling region [IIb3(S752P)], did not adhere to fibrinogen. As a control, the adhesion to fibronectin, which is presumably mediated by the endogenous integrin 51, was similar in parental HEK 293 cells and all recombinant cell lines.

The activation state of the full-length IIb3 receptor in HEK 293 cells was examined by the binding of an activation-dependent antibody (PAC1), peptidic RGD mimics (echistatin, L-692,884, and MK-852), and nonpeptidic RGD mimics (L-734,217, Ro 43-5054, and Ro 44-9883). These RGD mimics include those that discriminate between resting and activated IIb3, as well as those that do not. For example, echistatin, an RGD-containing peptide initially isolated from snake venom, shows no selectivity in binding to unactivated or activated IIb3 (31). Similarly, Ro 44-9883 does not discriminate in its binding to activated or unactivated IIb3. In contrast, L-692,884, MK-852, L-734,217, and Ro 43-5054 bind preferentially to activated IIb3 (Table 1) (22-25).²

The cyclic RGD mimic L-692,884 was shown to bind specifically to IIb3 expressed in HEK 293 cells. This binding was sensitive to the presence of inhibitors of IIb3. As summarized in Table 2, the binding specificities of the compounds for the full-length IIb3 in HEK 293 cells are intermediate between those observed for unactivated and activated platelets. However, all of these compounds bind to cells expressing the -cytoplasmic deletion mutant IIb3(717) with affinities that resemble those in activated platelets. These results indicate that the full-length IIb3 expressed in HEK 293 cells is indeed in a conformational state different from those in resting and activated platelets. We suggest that the receptor may exist in a transitional state.

To further explore the aforementioned hypothesis, the activation state of the recombinant IIb3 expressed in HEK 293 cells was evaluated using PAC1, an antibody specific for activated IIb3. PAC1 bound to cells expressing IIb3, IIb3(717), and IIb3(693). Binding of PAC1 to the full-length IIb3 receptor, although weak, was sensitive to DTT. Pretreatment of the cells with DTT before the addition of the antibody PAC1 increased PAC1 binding to the full-length receptor. In contrast, DTT treatment of cells expressing IIb3(717) and IIb3(693) did not alter PAC1 binding. An AI was defined as a measure of the activation state of the receptor. The full-length receptor, with an AI of 52, was identified to be in an activation state intermediate between that of resting platelets (AI = 8) and that of activated platelets (AI = 93). The AIs for the 3 subunit cytoplasmic and transmembrane deletion mutants IIb3(717) (AI = 88) and IIb3(693) (AI = 86) were indicative of fully activated receptors. These results suggest that the full-length IIb3 expressed in HEK 293 cells is in a "transitional" activation state, which is converted to a fully activated form by pretreatment with DTT. Recent studies¹ have shown that FITC-labeled fibrinogen binds to HEK 293 cells expressing IIb3 and IIb3(717) but does not bind to parental cells or the cells expressing the Glanzmann's mutants. The binding of FITC-labeled fibrinogen to IIb3 was energy dependent and could be abolished by the presence of sodium azide.¹

The functional state of the full-length IIb3 receptor expressed in HEK 293 cells was assessed by its ability to mediate cell aggregation in the presence of fibrinogen. This aggregation required an intact ligand binding site on the receptor; thus, the transitional activation state of IIb3 in HEK 293 cells was sufficient to support aggregation.

Ginsberg et al. (5) examined IIb3 expression in CHO cells and found that in this cell system the receptor exists in an unactivated state. Deletion of the membrane-proximal region of either of the cytoplasmic domains of the integrin subunit is sufficient to convert the receptor into an activated form (5, 19). Recently, those authors showed that a charge-reversal mutation in the membrane-proximal region can activate the receptor, presumably through the disruption of a potential salt bridge between the membrane-proximal portions of the  and  subunit cytoplasmic domains (32). Our results are consistent with those studies, in that deletion of the cytoplasmic domain of the  subunit leads to the activation of the receptor. These cytoplasmic deletions render the receptor competent to bind PAC1 and RGD mimics with affinities that resemble those of activated IIb3 in platelets. In contrast to IIb3 expressed in CHO or K562 cells, IIb3 expressed in HEK 293 cells appears to be in a transitional activation state, as determined by the intermediate affinities for RGD peptide and RGD mimics and the ability to bind to PAC1 and mediate fibrinogen-dependent, cell-cell aggregation. We have not been able to isolate a form of IIb3 in HEK 293 cells equivalent to that found in resting platelets.

It is conceivable that several conformations of IIb3 can exist in platelets. In fact, comparative binding studies of soluble fibrinogen and fibronectin suggest that in platelets conformational changes in IIb3 are not "all or none." Intermediate states in the conformational range of the receptor may further modulate the selectivity of soluble versus surface-bound conformations of any one ligand (33). Nakatani et al. (34) have established that, when platelets are stimulated with different agonists (and RGD-containing peptides), they bind to fibrinogen with different affinities. Thus, varied conformational states of activated IIb3 may exist in platelets. Kunicki et al. (35) have suggested the existence of subpopulations of IIb3 by direct binding of Fab fragments of the antibodies AP7 and PAC1 to IIb3 purified from human platelets, and it has been postulated that these subpopulations of IIb3 result in distinctive phenotypes of human platelets (36). These observations support the notion that recombinant IIb3 may exist in multiple conformational states in HEK 293 cells, with the most predominant being the "transitional" activation state, which we have identified here.

Although the reason for recombinant IIb3 being in a unique conformation in HEK 293 cells is unclear, previous studies have demonstrated that other integrins have cell-specific differences in ligand binding activity. For example, 21 purified from platelets does not bind to laminin; however, that purified from endothelial cells binds readily to laminin, which suggests a role for cell-specific factors in the modulation of integrin affinity (37). Cell-specific intracellular factors that are unique to HEK 293 cells could also account for this activation state of IIb3. Recently, a novel cytosolic regulatory molecule, cytohesin-1, was shown to interact with the cytoplasmic domain of 2 and to increase the strength of L2 interactions with its ligand, intercellular adhesion molecule-1 (9). Alternatively, proteins associated with the 3 integrin, such as integrin-associated protein (38), cell adhesion regulator (39), and endonexin (8), do not appear to influence the affinity state of IIb3 in CHO cells. The regulatory influence of these proteins in modulating IIb3 function in HEK 293 cells requires further investigation.

In conclusion, our results demonstrate that, with the use of novel small peptidic and nonpeptidic RGD mimics, the activation state of IIb3 can be precisely determined. We show that, in HEK 293 cells, IIb3 exists in a transitional activation state that is functionally competent to mediate aggregation of cells in the presence of fibrinogen. Cell-specific conformations of IIb3 in vivo could indicate an alternative ligand binding specificity for IIb3 in platelets or megakaryocytes. Such subpopulations of transitionally active IIb3 may serve a physiological role in such events as attachment to specific extracellular matrices or transportation of these proteins into the cell for storage. To understand the conformation-specific integrin-matrix interactions that may be found in different cells, the functional role of the transitionally active IIb3 receptor must be further explored.