A naturally occurring Tyr143Hisα_IIb mutation abolishes α_IIbβ₃ function for soluble ligands but retains its ability for mediating cell adhesion and clot retraction: comparison with other mutations causing ligand-binding defects

Integrins are a family of noncovalently associated αβ heterodimeric adhesion receptors that mediate cellular attachment to the extracellular matrix and cell–cell cohesion.1,2 Integrins are involved in a variety of physiological processes, including development, immune response, wound healing, and hemostasis.1 They are also involved in pathologic processes, such as tumor metastasis, thrombosis, and athelosclerosis.1 

Although its expression is restricted to the megakaryocyte/platelet lineage, α_IIbβ₃ (GPIIb-IIIa) is a prototypic integrin that functions as a physiologic receptor for fibrinogen and von Willebrand factor and that plays a crucial role in platelet aggregation, normal hemostasis, and pathologic thrombosis.3 Indeed, Glanzmann thrombasthenia (GT) is an autosomal recessive bleeding disorder caused by a defect in the expression or the function of integrin α_IIbβ₃.4 Recent clinical studies have shown the beneficial effects of α_IIbβ₃ antagonists in patients undergoing coronary angioplasty and in those with unstable angina.5,6Integrin α subunits are grouped into 2 classes based on the presence or absence of an inserted domain of approximately 200 amino acid residues (I- or A-domain); α_IIb and α_vsubunits do not have the I-domain.7-9 Recently, the crystal structures of the extracellular segment of the other β₃ integrin, α_vβ₃—in the presence or absence of its small ligand—have been described.10,11 As predicted, the putative ligand-binding head is primarily formed by a 7-bladed β-propeller domain from α_v and a β I-domain from β₃.10-12 The β-propeller domain contains 7 4-stranded β-sheets (W1-W7) arranged in a torus around a pseudosymmetry axis. However, the ligand-binding sites in the β-propeller domain of the α_IIb subunit remain to be determined.

The characterization of molecular defects in GT from dysfunctional α_IIbβ₃ (variant GT) has succeeded in pinpointing ligand-binding sites and functionally important domains.13-15 We have demonstrated that 2-amino acid insertion (Arg-Thr) between amino acid residues 160 and 161 in α_IIb is responsible for the ligand-binding defect in a Japanese variant GT known as KO.16 The insertion is located within the small loop (Cys146-Cys167) between W2 and W3 (W3 4-1 loop) located on the upper face of the β-propeller. Alanine substitutions further indicate that Asp163 within the Cys146-Cys167 loop is one of the critical residues for ligand binding.16In this context, 2 other naturally occurring missense mutations, Pro145Ala and Leu183Pro, which impair α_IIbβ₃ expression and its ligand-binding function, have been identified in the W3 4-1 loop and the W3 2-3 loop of the β-propeller, respectively.17,18 

In this study we demonstrate that a novel naturally occurring missense mutation (Tyr143His) within the W3 4-1 loop in α_IIb is responsible for a binding defect in α_IIbβ₃for soluble ligands. Because Tyr143His, KO, and Asp163Ala mutations are located within the same loop in α_IIb, we further compared functions of these mutant α_IIbβ₃ by using cell adhesion and clot retraction assays. Compared with the KO mutation and Asp163Alaα_IIbβ₃, Tyr143Hisα_IIbβ₃-expressing cells still had some ability for cell adhesion and clot retraction. Our results indicate that the KO mutant and Asp163Alaα_IIbβ₃ impair ligand-binding function in α_IIbβ₃ more severely than Tyr143Hisα_IIbβ₃.

Patient Osaka-12, a product of nonconsanguineous parents, was a 21-year-old Japanese woman who had a history of moderate mucocutaneous bleeding in her childhood. Hematologic examinations revealed a prolonged bleeding time (more than 15 minutes) and an absence of platelet aggregation in response to adenosine diphosphate (ADP), epinephrine, and collagen but a normal response to ristocetin. Clot retraction using the MacFarlane method was 46% (normal range, 40%-70%).19 Informed consent for analyzing their molecular genetic abnormalities was obtained from Osaka- 12 and her parents.

AP1 (GPIb-specific monoclonal antibody [mAb]), AP2 (α_IIbβ₃-specific mAb), and AP5 (β₃-specific mAb) were generously provided by Dr Thomas J. Kunicki (Scripps Research Institute, La Jolla, CA).20,21 AP3 (β₃-specific mAb) was a generous gift from Dr Peter J. Newman (Blood Center of Southeastern Wisconsin, Milwaukee).22 OP-G2 is a ligand-mimetic α_IIbβ₃-specific mAb that binds to nonactivated and activated α_IIbβ₃.23 PAC-1, a ligand-mimetic α_IIbβ₃-specific mAb that binds specifically to activated α_IIbβ₃, was kindly provided by Dr Sanford J. Shattil (Scripps Research Institute).24 PT25-2 (α_IIbβ₃-specific mAb) activates α_IIbβ₃ and was a kind gift from Drs Makoto Handa and Yasuo Ikeda (Keio University, Tokyo, Japan).25PMI-1 (α_IIb-specific mAb), anti-LIBS1 (β₃-specific mAb), and anti-LIBS6 (β₃-specific mAb) were generously provided by Dr Mark H. Ginsberg (Scripps Research Institute).26,27 PMI-1, AP5, anti-LIBS1, and anti-LIBS6 recognize ligand-induced conformational changes on α_IIbβ₃, termed ligand-induced binding sites (LIBS).21,26,27 TP80 (α_IIb-specific mAb) and MOPC21 (mouse myeloma immunoglobulin G₁ [IgG₁]) were purchased from Nichirei (Tokyo, Japan) and Sigma Chemical (St Louis, MO), respectively. FK633 (α_IIbβ₃-specific peptidomimetic antagonist) was generously provided by Dr Jiro Seki (Fujisawa Pharmaceutical, Osaka Japan),28 and cyclo(RGDfV) [cyclo(-Arg-D-Gly D-Asp-D-Phe-L-Val-D-)] peptide (α_vβ₃-specific antagonist) was a generous gift from Merck KGaA (Darmstadt, Germany).29 Fibrinogen was purchased from Calbiochem-Novabiochem (San Diego, CA). MOPC21, AP1, AP2, TP80, AP3, PAC-1, and fibrinogen were labeled with fluorescein isothiocyanate (FITC), as previously described.28 

Flow cytometric analysis using various mAbs was performed as previously described.30 FITC-labeled mAbs were used to quantify the expression levels of α_IIbβ₃ on 293 cells and on platelets.

Total cellular RNA of platelets was isolated from 30 mL whole blood, and α_IIb or β₃ messenger RNA (mRNA) was specifically amplified by reverse transcription–polymerase chain reaction (RT-PCR), as previously described.30 Primers for the amplification of α_IIb or β₃ mRNA and conditions for RT-PCR were described elsewhere.30,31Nucleotide sequences of PCR products were determined by using Taq DyeDeoxy Terminator Cycle Sequencing kit and an ABI373A DNA sequencer (Applied Biosystems, Foster City, CA).

Amplification of the region around exon 4 of the α_IIb gene was performed by using primers IIbE3, 5′-GTCGGTCGTCAGCTGGAGC-3′ (sense, nucleotide [nt] 3940-3958 in the α_IIb gene), and IIbE4, 5′-CAGGTCGTAGCTGGCGCTTAC-3′ (antisense, nt 4192-4172) and by using 250 ng DNA as a template. First-round PCR products were reamplified using primers IIbg3947S, 5′-GTCAGCTGGAGCGACGTCATTGTG-3′ (sense, nt 3947-3970) and IIbE4 and then were digested with restriction enzyme ScaI. Resultant fragments were electrophoresed in a 1.5% agarose gel.

Wild-type α_IIb and β₃ complementary DNAs (cDNAs) cloned into a mammalian expression vector pcDNA3 (Invitrogen, San Diego, CA) were generously provided by Dr Peter J. Newman. To construct the expression vector containing the⁵²¹C (His143) form of α_IIb, PCR-based cartridge mutagenesis was performed as previously described.16 Platelet α_IIb cDNA from Osaka-12 was amplified by RT-PCR using primers IIb1 and IIb3αAS, 5′-CCCACGATCAGGTCTGGGTATCCG-3′ (antisense, nt 1407-1384). Second-round amplification was then performed using 1 μL first-round PCR products as a template with nested primers IIb187S, 5′-GAGAGTGGCCATCGTGGTGG-3′ (sense, nt 187-206), and IIb3αAS2C, 5′-CCGTTGTCATCGATGTCTACGGC-3′ (antisense, nt 1364-1386). To construct the expression vector for Tyr143Ala α_IIb substitution, we carried out the overlapping extension PCR as previously described.16Mismatched sense primer IIb143Ala-S, 5′-GAGCGGCCGCCGCGCCGAGGCCTCCCCCTG-3′ (sense, nt 502-531, 2 bp mismatched), and antisense primer IIb143Ala-AS, 5′-CAGGGGGAGGCCTCGGCGCGGCGGCCGCTC-3′ (antisense, nt 531-502, 2 bp mismatched) were synthesized. PCR was performed by using α_IIb cDNA as a template and primers IIb1 and IIb143Ala-AS, or primers IIb143Ala-S and IIb3αAS. The 2 individually amplified PCR products were mixed and used as a template of PCR using primers IIb1 and IIb3αAS. Amplified fragments were digested with restriction enzyme SacII and ClaI, and the resultant fragments were extracted using GeneClean II kit (Bio 101, La Jolla, CA). Fragments were introduced into the pcDNA3 that had been digested with SacII and ClaI. Inserted fragments were characterized by sequence analysis to verify the absence of any other substitutions and the proper insertion of the PCR cartridge into the vector. Expression vectors containing the KO mutant (2-amino acid [Arg-Thr] insertion between residues 160 and 161) α_IIbcDNA and Asp163AlaαIIb cDNA were constructed as previously described.16 

The wild-type or mutant α_IIb construct was cotransfected into 293 cells with the wild-type β₃ construct by the calcium phosphate method, as previously described.16 The 293 cells transiently expressing mutant α_IIbβ₃ were obtained and analyzed 2 days after transfection. In selected experiments, CD36 expression vector was cotransfected with α_IIb and β₃ constructs to monitor transfection efficiency.32 In addition, stable transfectants expressing wild-type, Tyr143Hisα_IIbβ₃, or KO mutant α_IIbβ₃ were selected for G418 resistance (Gibco BRL, Grand Island, NY) and were cultured in Dulbecco modified Eagle medium (DMEM) with 10% heat-inactivated fetal calf serum (Life Technologies, Gaithersburg, MD).

Adhesion assays were performed as described by Faull et al.33 Wells from 96-well microtiter plates were coated with up to 0.5 μg fibrinogen per well in 100 μL phosphate-buffered saline (PBS) and were incubated at 4°C overnight. After washing with PBS, wells were blocked with PBS containing 1% bovine serum albumin (BSA) (Sigma) for 90 minutes at 22°C. To determine background adhesion, control wells were coated with 1% BSA. Then 293 cells transiently expressing wild-type or mutant α_IIbβ₃ were washed twice with PBS and resuspended in DMEM containing 0.1% BSA at a concentration of 1 × 10⁶ cells/mL, and 100 μL aliquots cell suspension were added to wells in triplicate. The plate was incubated in a humidified 37°C incubator for 60 minutes. After washing with PBS 3 times, the adherent cells were quantified by measuring endogenous cellular acid phosphate activity in an enzyme-linked immunosorbent assay.34 

For morphologic analysis, 293 cells stably expressing wild-type or mutant α_IIbβ₃ were added on fibrinogen-coated glass coverslips for 60 minutes at 37°C. After they were washed with PBS, adherent cells were fixed in 3.7% formaldehyde for 10 minutes, permeabilized in 0.5% Triton X-100 in PBS for 5 minutes at room temperature, and washed twice with PBS. Cells on coverslips were stained with rhodamine-phalloidin (Sigma) and were analyzed under a fluorescence microscope (Olympus, Tokyo, Japan).35 

Adherent cells on fibrinogen were lysed in Triton X-100 buffer (1% Triton X-100, 25 mM Tris-HCl, 100 mM NaCl, pH 7.4, 0.1 mg/mL leupeptin, 4 μg/mL pepstatin A, 1 mM phenylmethylsulfonyl fluoride, and 10 mM benzamide) containing sodium vanadate and were scraped into microcentrifuge tubes. Lysates were incubated on ice for 30 minutes, and clarified supernatants were processed for immunoprecipitation. Focal adhesion kinase (pp125^FAK) was immunoprecipitated with 1 μg rabbit polyclonal antibody specific for FAK (Santa Cruz Biotechnology, CA), and protein-G Sepharose (Pharmacia, Uppsala, Sweden). Precipitates were separated on 7.5% sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE) and were transferred to a polyvinylidene difluoride (PVDF) membrane. Phosphotyrosine was detected with mAb 4G10. To monitor the loading of gel lanes, the blots were stripped (2% SDS, 62.5 mM Tris, pH 6.7, 100 mM 2-mercaptoethanol for 30 minutes at 70°C) and reprobed with anti-FAK.36 

Clot retraction of stable cell lines was performed based on the method described by Katagiri et al.37 In brief, 293 cells stably expressing α_IIbβ₃ in Tyrode/HEPES (N-2-hydroxyethylpiperazine-N′-2-ethanesulfonic acid) buffer (2 × 10⁶ cells/mL) were incubated with 10 mM tranexamic acid, 250 μg/mL fibrinogen, and 2 mM CaCl₂ at 37°C in a siliconized glass tube. To block the effect of α_vβ₃ composed of endogenous α_v and transfected β₃, these stable transfectants were preincubated with 50 μM cyclo(RGDfV) (α_vβ₃-specific antagonist) at room temperature for 30 minutes. Then 1 U thrombin was added to 1 mL cell suspension. Fibrin gels began to form immediately after the addition of thrombin, and the tubes were kept at 37°C. Clot retraction was monitored by taking photographs every 30 minutes using a digital camera. Quantification of retraction was performed by an assessment of the clot area using the NIH Image 1.67e software (Bethesda, MD). Data were expressed as follows: % clot retraction = [(areat₀ − area t)/areat₀] × 100.

We first examined the surface expression of α_IIbβ₃ on platelets from patient Osaka-12 and 3 control subjects using flow cytometry. Although FITC-AP1 (GPIb-specific mAb) bound to Osaka-12 platelets slightly higher than to control platelets (113% of control value [mean value obtained from 3 control subjects]), FITC-AP3 (β₃-specific mAb) and FITC-TP80 (α_IIb-specific mAb) revealed a significant reduction in the expression of α_IIbβ₃ on Osaka-12 platelets (AP3 binding, 41% of control value; TP80 binding, 36% of control value; n = 2) (Figure1). Compared with AP3 and TP80 binding to Osaka-12 α_IIbβ₃, AP2 binding was markedly impaired (13% of control value; n = 2) probably because of a disturbance of AP2 epitope formation. Because Osaka-12 platelets expressed 36% to 41% of normal amounts of α_IIbβ₃ on their surfaces, we examined their ligand-binding function. OP-G2 and PAC-1 mAbs are activation-independent and activation-dependent, ligand-mimetic mAbs specific for α_IIbβ₃, respectively. Neither OP-G2 nor PAC-1 in the presence of an activating PT25-2 mAb bound to Osaka-12 platelets (Figure 1). Thus, Osaka-12 α_IIbβ₃ seems to have a ligand-binding defect and a quantitative defect.

PMI-1, AP5, anti-LIBS1, and anti-LIBS6 mAbs preferentially recognize LIBS on α_IIbβ₃, which are exposed following occupancy of the receptor by ligands. To further investigate the ligand-binding function of Osaka-12 α_IIbβ₃, we examined the effect of FK633, α_IIbβ₃-specific peptidomimetic antagonist on the expression of these LIBS. Compared with Arg-Gly-Asp-Trp peptide, FK633 has approximately 100-fold higher potency for LIBS induction.28 Although we could not simply compare the binding of mAbs recognizing LIBS between Osaka-12 and control platelets because of the different expression levels of α_IIbβ₃, even 10 μM FK633 showed a negligible effect on the expression of LIBS1 of Osaka-12 α_IIbβ₃ (Table1). On the other hand, Table 1 shows that Osaka-12 α_IIbβ₃ aberrantly expressed LIBS recognized by PMI-1, AP5, and anti-LIBS6. These data further suggested that the ligand-binding sites of Osaka-12 α_IIbβ₃ were markedly impaired.

To identify the molecular genetic defect responsible for Osaka-12, all coding regions of α_IIb and β₃ cDNA amplified by RT-PCR from Osaka-12 and control platelet mRNA were analyzed. Direct sequence of the PCR fragments showed a single T>C substitution at nt 521 in α_IIb cDNA that would lead to a Tyr143His substitution in exon 4 of α_IIb (Figure2A). No other nucleotide substitutions were detected in the coding regions of either α_IIb or β₃ cDNAs. The T>C substitution abolished a restriction site for ScaI, and allele-specific restriction enzyme analysis for the amplified α_IIb cDNA around exon 4 suggested that Osaka-12 might be homozygous for the substitution. However, allele-specific restriction enzyme analysis for the PCR fragments from genomic DNA revealed that Osaka-12 was heterozygous for the T>C substitution. The substitution was inherited from her father. The molecular genetic defect inherited from her mother remains determined. However, only the allele having ⁵²¹C from her father was amplified by RT-PCR from Osaka-12 α_IIb mRNA, suggesting that the expression levels of maternal α_IIbmRNA seem to be extremely low in Osaka-12 platelets (Figure 2B-C).

To confirm that the ⁵²¹T>C substitution leading to Tyr143His substitution in α_IIb is responsible for the functional defect, we constructed an expression vector that contained the wild-type or mutant His143 form of α_IIb and cotransfected it with wild-type β₃ vector into 293 cells. Effects of a Tyr143Ala substitution in α_IIb on α_IIbβ₃ expression and function were also examined. Because Tyr143 is located within the W3 4-1 loop of the β-propeller domain, close to the KO mutation (the Arg-Thr insertion between 160 and 161 residues of α_IIb) and the Asp163Ala α_IIb mutation, we compared the expression and function of these mutant α_IIbβ₃ in parallel.

Transfection efficiency monitored by the cotransfected CD36 expression vector was essentially the same between wild-type and mutant α_IIbβ₃ (Figure3). Flow cytometric analysis using FITC-AP3 or FITC-TP80 mAb showed that the expression levels of Tyr143Hisα_IIbβ₃ were essentially the same as wild-type α_IIbβ₃ (AP3 binding, 91% ± 15% of wild-type; TP80 binding, 102% ± 2% of wild-type; mean ± SD; n = 3). In sharp contrast, Tyr143Hisα_IIbβ₃ failed to bind OP-G2 mAb, FITC-PAC-1 mAb, or FITC-fibrinogen in the presence of PT25-2 mAb (Figure 3). We also confirmed that the binding failure of these activation-dependent ligands was not caused by the binding failure of the activating mAb, PT25-2. In addition, FITC-AP2 binding was significantly impaired compared with TP80 binding (AP2 binding; 15% ± 3% of wild-type; mean ± SD; n = 3), which is consistent with the data obtained from Osaka-12 platelets (Figures 1and 3). Given that the expression levels of abnormal α_IIbmRNA derived from her mother were so low that they were not detected in Osaka-12 platelets by RT-PCR, this maternal abnormality would reduce the expression of α_IIbβ₃ by 50%. As expected, the expression level of Tyr143Hisα_IIbβ₃ on 293 cells was roughly twice as much as that on Osaka-12 platelets. These data indicate that the Tyr143His mutation is responsible for the variant GT phenotype in Osaka-12. The Tyr143Ala substitution induced essentially the same effects on α_IIbβ₃ as the Tyr143His mutation (AP3 binding, 117% ± 14% of wild-type; TP80 binding, 113% ± 8% of wild-type; AP2 binding, 14% ± 1% of wild-type; mean ± SD; n = 3). Because Tyr143Ala abolished the ligand binding, it is likely that the presence of Tyr143 is critical for α_IIbβ₃ function. Interestingly, the phenotype of Asp163Alaα_IIbβ₃ appears the same as that of Tyr143Hisα_IIbβ₃. The phenotype of KO α_IIbβ₃ was essentially the same as wild-type α_IIbβ₃ except for the failure in OP-G2 and PAC-1 binding, as previously described.15 Thus, none of these mutant α_IIbβ₃ showed a binding ability for soluble ligands (Figure 3).

To further examine the functions of mutant α_IIbβ₃, we performed cell adhesion assay to immobilized fibrinogen. In contrast to parent cells, 293 cells transiently transfected with wild-type α_IIbβ₃ or mutant α_IIbβ₃ became adhesive to the immobilized fibrinogen. Given that endogenous α_v in 293 cells can form α_vβ₃ with the exogenous β₃ and may contribute to the adhesion, we preincubated the transfected cells with 50 μM cyclo(RGDfV) to block the α_vβ₃ function. Under these conditions, 293 cells transfected with wild-type β₃ alone (data not shown), KO variant α_IIbβ₃, or Asp163Alaα_IIbβ₃ failed to adhere to immobilized fibrinogen (Figure 4). However, 293 cells expressing Tyr143Hisα_IIbβ₃ showed a significant adhesion to immobilized fibrinogen, and the amounts of adherent cells were approximately 50% compared with 293 cells expressing wild-type α_IIbβ₃. The adhesion of 293 cells expressing Tyr143Hisα_IIbβ₃ in the presence of 50 μM cyclo(RGDfV) was mediated solely by α_IIbβ₃ because 10 μM FK633 completely blocked the cell adhesion (data not shown).

We then obtained the cells stably expressing wild-type, Tyr143His, and KO mutant α_IIbβ₃ with G418 selection for morphologic analysis of the adherent cells. Adhesive functions of 293 cells stably expressing Tyr143His α_IIbβ₃ or KO α_IIbβ₃ were the same as those transiently expressing these mutant α_IIbβ₃(data not shown). Although 293 cells stably expressing wild-type α_IIbβ₃ adhered firmly and showed complete spreading, 293 cells stably expressing KO mutant α_IIbβ₃ failed to adhere. In contrast to KO α_IIbβ₃, 293 cells stably expressing Tyr143Hisα_IIbβ₃ moderately impaired cell spreading and cell adhesion (Figure 5). Because FAK, a 125-kDa cytoplasmic tyrosine kinase, is a component of focal adhesions and is a well-established component of integrin signaling pathways,38 the tyrosine phosphorylation of pp125^FAK in these adherent 293 cells stably expressing wild-type α_IIbβ₃ or Tyr143Hisα_IIbβ₃ was examined. As shown in Figure 6, pp125^FAKphosphorylation was observed but impaired in cells expressing Tyr143Hisα_IIbβ₃ compared with wild-type α_IIbβ₃ (Figure 6).

We examined the capacity of Tyr143Hisα_IIbβ₃ and KO mutant α_IIbβ₃ to mediate fibrin clot retraction by using stable transfectants. This process reflects the capacity of α_IIbβ₃ to transduce mechanical force between the intracellular cytoskeleton and the extracellular fibrin strands. Parent 293 cells failed to retract fibrin clots. Given that α_vβ₃ is involved in mediating fibrin clot retraction,37 we preincubated all stable transfectants with 50 μM cyclo(RGDfV) to block the α_vβ₃function. Under these conditions, 293 cells stably transfected with wild-type β₃ alone failed to mediate fibrin clot retraction, whereas those with wild-type α_IIb and β₃ showed a marked retraction. Although 293 cells stably expressing Tyr143Hisα_IIbβ₃ showed some ability for fibrin clot retraction (Figure7), 293 cells stably expressing KO mutant α_IIbβ₃ failed to mediate fibrin clot retraction.

The molecular basis for the interaction between α_IIbβ₃ and its ligands remains to be determined. In this study, we have described a novel missense mutation (Tyr143His) in α_IIb associated with a ligand-binding defect in this receptor. Because the Tyr143His mutation is located in the W3 4-1 loop of the β-propeller domain of α_IIb, we compared the effects of Tyr143Hisα_IIb on the expression and function of α_IIbβ₃ with those of previously described KO and Asp163Ala mutations located in the same loop.16 Although GT patient analysis demonstrated that the surface expression level of Tyr143Hisα_IIbβ₃on platelets from heterozygous Osaka-12 was 36% to 41% of control, this reduction in the expression resulted mainly from the null expression of α_IIb mRNA from the maternal allele. Indeed, Tyr143Hisα_IIb mutation resulted in almost normal α_IIbβ₃ expression on 293 cells (91%-102% of wild-type α_IIbβ₃). In sharp contrast, this mutation abolished the interaction between α_IIbβ₃ and its macromolecular ligands OP-G2 mAb, PAC-1 mAb, and fibrinogen. The failure of LIBS expression by FK633 further suggested that Tyr143Hisα_IIbβ₃ has a binding defect for high-affinity α_IIbβ₃-specific small ligand. The effects of Tyr143Alaα_IIb on the expression and function of α_IIbβ₃ were essentially the same as Tyr143Hisα_IIb, suggesting that the presence of Tyr143, rather than the presence of a mutated residue such as His143, is critical for α_IIbβ₃ function. Interestingly, the defects for soluble ligands are essentially the same as those induced by KO or Asp163Ala mutation. However, cell adhesion and fibrin clot retraction assays revealed distinctive features among Tyr143Hisα_IIbβ₃, KO α_IIbβ₃, and Asp163Alaα_IIbβ₃. Thus, KO α_IIbβ₃ and Asp163Alaα_IIbβ₃ show a more profound binding defect for immobilized fibrinogen and fibrin strands than Tyr143Hisα_IIbβ₃, suggesting that Asp163 is located at or close to the ligand-binding sites in the β-propeller domain of α_IIb.

Recently described crystal structure of the extracellular segment of α_vβ₃ shows that the ligand-binding head is formed by a 7-bladed β-propeller domain from α_v and a β I-domain from β₃.10 Moreover, there is mounting evidence that the W3 4-1 loop, W3 2-3 loop, and W4 4-1 loop of the β-propeller domain are important for ligand binding in non–I-domain α subunits.39-41 In terms of the W3 4-1 loop of α_IIb, several residues critical for ligand binding have been suggested by alanine-scanning mutagenesis.16,41 In addition, KO, Pro145Ala, and Tyr143His (this study) are naturally occurring mutations responsible for the ligand-binding defect in patients with variant GT. Compared with soluble ligands, immobilized fibrinogen and fibrin are activation-independent ligands for α_IIbβ₃. Immobilized fibrinogen binding to α_IIbβ₃ initiates outside-in signaling that leads to cellular responses such as cell spreading. In addition to outside-in signaling, fibrin clot retraction requires inside-out transmission of the contractile forces from intracellular cytoskeleton to fibrin strand. Much higher concentrations of α_IIbβ₃-specific antagonists are needed to block cell adhesion to fibrinogen and fibrin clot retraction.42-44 Thus, these ligands show higher affinity/avidity for α_IIbβ₃ than soluble fibrinogen. Because some mutations in the W3 4-1 loop, such as Pro145Ala, impaired α_IIbβ₃ biosynthesis and its function,17 such mutations are likely to affect the conformation of this receptor, including the W3 4-1 loop. Therefore, we used these higher affinity/avidity ligands to further examine the residual function of α_IIbβ₃ mutants. Neither the KO mutant nor Asp163Alaα_IIbβ₃-expressing 293 cells adhered to immobilized fibrinogen or showed fibrin clot retraction. In contrast, Tyr143Hisα_IIbβ₃-expressing 293 cells showed some ability for fibrin clot retraction and for cell adhesion to fibrinogen with a reduction in spreading. Consistent with these findings, the impaired tyrosine phosphorylation of pp125^FAK mediated through α_IIbβ₃ was observed in Tyr143Hisα_IIbβ₃-expressing 293 cells compared with wild-type α_IIbβ₃. These findings indicate that Tyr143Hisα_IIbβ₃still has some ligand-binding function. Although a differential engagement of α_IIbβ₃ in fibrin clot retraction versus aggregation has been suggested,45,46 KO mutant and Asp163Alaα_IIb abolished the interaction with soluble fibrinogen and immobilized fibrin(ogen), as shown in this study. These findings suggest that the integrin residues involved in these phenomena, at least in part, overlap. The close location of Tyr143 with Asp163 suggests that the defect in soluble ligand binding induced by the Tyr143Hisα_IIb mutation is likely the result of its allosteric effect rather than of a defect in the ligand-binding site itself.

More recently, the crystal structure of α_vβ₃ with cyclo(RGDfV) has been described.11 The Arg and Asp side chains exclusively contact the propeller and β I-domain, respectively, confirming the critical roles of these domains in ligand binding. The Arg side chain inserts into a narrow groove at the top of the propeller domain, formed primarily by the W3 4-1 loop and the W4 4-1 loop, and the Arg guanidinium group is held in place by a salt bridge to Asp218 in the W4 4-1 loop and to Asp150 in the W3 4-1 loop.11 However, our previous data showed that alanine-scanning mutagenesis in the W3 4-1 loop, including Asp150Ala, did not inhibit the ligand-binding function of α_vβ₃; this is not consistent with the crystal structure data.34 The role of the W3 4-1 loop appears to be different between non–I-domain α subunits. The W3 4-1 loops in α_IIb and in α₃ (Thr162 and Gly163) appear critical for ligand binding, whereas those in α₄and in α_v do not.16,34,39,40 Although alanine-scanning mutagenesis suggested several residues within the W3 4-1 and W3 2-3 loops of α_IIb may be critical for ligand binding, the same procedure showed only Tyr178 in the W3 2-3 loop of α_v as a critical residue.34,41 In addition, our recent study suggests that the expression and function of α_IIbβ₃ are more strictly regulated than α_vβ₃.47 Among non–I-domain α subunits, α_IIb has the longest W3 4-1 loop, and it is possible that differences in the structure between α_IIband α_v may account for the distinctive role of the W3 4-1 loop in ligand binding between these β₃ integrins. Unique features of the W3 4-1 loop of α_IIbβ₃ are also likely to account for the discrepancy between the ability for soluble ligand binding and cell adhesion. Detailed structure–function analyses of β₃ integrins and of their crystal structure would provide a better understanding of integrin function and a new antagonist design for these integrins to prevent thrombosis.

We thank Dr Thomas Kunicki for mAbs AP1, AP2, and AP5; Dr Peter Newman for mAb AP3 and the α_IIb and β₃ cDNA cloned into a mammalian expression vector pcDNA3; Dr Sanford Shattil for mAb PAC-1; Drs. Makoto Handa and Yasuo Ikeda for mAb PT25-2; Dr Mark Ginsberg for mAbs PMI-1, anti-LIBS1, and anti-LIBS 6; Dr Jiro Seki for FK633; and Dr P. Raddatz for cyclo(RGDfV).