Crystal Structure of the Wild-type von Willebrand Factor A1-Glycoprotein Ibα Complex Reveals Conformation Differences with a Complex Bearing von Willebrand Disease Mutations
2004; Elsevier BV; Volume: 279; Issue: 22 Linguagem: Inglês
10.1074/jbc.m401659200
ISSN1083-351X
AutoresJohn J. Dumas, R. Anand Kumar, Thomas McDonagh, Francis Sullivan, Mark Stahl, W.S. Somers, Lidia Mosyak,
Tópico(s)Blood disorders and treatments
ResumoThe adhesion of platelets to the subendothelium of blood vessels at sites of vascular injury under high shear conditions is mediated by a direct interaction between the platelet receptor glycoprotein Ibα (GpIbα) and the A1 domain of the von Willebrand factor (VWF). Here we report the 2.6-Å crystal structure of a complex comprised of the extracellular domain of GpIbα and the wild-type A1 domain of VWF. A direct comparison of this structure to a GpIbα-A1 complex containing "gain-of-function" mutations, A1-R543Q and GpIbα-M239V, reveals specific structural differences between these complexes at sites near the two GpIbα-A1 binding interfaces. At the smaller interface, differences in interaction show that the α1-β2 loop of A1 serves as a conformational switch, alternating between an open α1-β2 isomer that allows faster dissociation of GpIbα-A1, as observed in the wild-type complex, and an extended isomer that favors tight association as seen in the complex containing A1 with a type 2B von Willebrand Disease (VWD) mutation associated with spontaneous binding to GpIbα. At the larger interface, differences in interaction associated with the GpIbα-M239V platelet-type VWD mutation are minor and localized but feature discrete γ-turn conformers at the loop end of the β-hairpin structure. The β-hairpin, stabilized by a strong classic γ-turn as seen in the mutant complex, relates to the increased affinity of A1 binding, and the β-hairpin with a weak inverse γ-turn observed in the wild-type complex corresponds to the lower affinity state of GpIbα. These findings provide important details that add to our understanding of how both type 2B and platelet-type VWD mutations affect GpIbα-A1 binding affinity. The adhesion of platelets to the subendothelium of blood vessels at sites of vascular injury under high shear conditions is mediated by a direct interaction between the platelet receptor glycoprotein Ibα (GpIbα) and the A1 domain of the von Willebrand factor (VWF). Here we report the 2.6-Å crystal structure of a complex comprised of the extracellular domain of GpIbα and the wild-type A1 domain of VWF. A direct comparison of this structure to a GpIbα-A1 complex containing "gain-of-function" mutations, A1-R543Q and GpIbα-M239V, reveals specific structural differences between these complexes at sites near the two GpIbα-A1 binding interfaces. At the smaller interface, differences in interaction show that the α1-β2 loop of A1 serves as a conformational switch, alternating between an open α1-β2 isomer that allows faster dissociation of GpIbα-A1, as observed in the wild-type complex, and an extended isomer that favors tight association as seen in the complex containing A1 with a type 2B von Willebrand Disease (VWD) mutation associated with spontaneous binding to GpIbα. At the larger interface, differences in interaction associated with the GpIbα-M239V platelet-type VWD mutation are minor and localized but feature discrete γ-turn conformers at the loop end of the β-hairpin structure. The β-hairpin, stabilized by a strong classic γ-turn as seen in the mutant complex, relates to the increased affinity of A1 binding, and the β-hairpin with a weak inverse γ-turn observed in the wild-type complex corresponds to the lower affinity state of GpIbα. These findings provide important details that add to our understanding of how both type 2B and platelet-type VWD mutations affect GpIbα-A1 binding affinity. The adhesion of blood platelets to sites of vascular injury is mediated by von Willebrand factor (VWF), 1The abbreviations used are: VWF, von Willebrand factor; Gp, platelet glycoprotein receptor; VWD, von Willebrand disease.1The abbreviations used are: VWF, von Willebrand factor; Gp, platelet glycoprotein receptor; VWD, von Willebrand disease. a large multimeric plasma glycoprotein that binds to both exposed connective tissue and platelet surface receptors (1Sadler J.E. Annu. Rev. Biochem. 1998; 67: 395-424Crossref PubMed Scopus (1109) Google Scholar, 2Ruggeri Z.M. Curr. Opin. Hematol. 2003; 10: 142-149Crossref PubMed Scopus (106) Google Scholar). VWF is localized to the site of injury via attachment of its A3 domain to exposed collagen in the subendothelium (3Lankhof H. vanHoeij M. Shiphorst M.E. Bracke M. Wu Y.P. Ijsseldijk M.J. Vink T. deGroot P.G. Sixma J.J. Thromb. Haemostasis. 1996; 75: 950-958Crossref PubMed Scopus (130) Google Scholar). Subsequently, platelets are recruited through a direct interaction between the multimeric glycoprotein receptor complex GpIb-IX-V on the platelet surface, and the A1 domain of immobilized VWF (4Berndt M.C. Shen Y. Dopheide S.M. Gardiner E.E. Andrews R.K. Thromb. Haemostasis. 2001; 86: 178-188Crossref PubMed Scopus (238) Google Scholar). GpIb-IX-V is comprised of four transmembrane subunits, GpIbα, GpIbβ, GpIX, and GpV, and the binding site for A1 is localized to the extracellular domain of GpIbα (5Andrews R.K. Lopez J.A. Berndt M.C. Int. J. Biochem. Cell Biol. 1997; 29: 91-105Crossref PubMed Scopus (174) Google Scholar). There is no measurable binding of normal VWF to platelets in circulating blood, and the binding of VWF to GpIbα requires high shear conditions generated by rapidly flowing blood (6Savage B. Saldivar E. Ruggeri Z.M. Cell. 1996; 84: 289-297Abstract Full Text Full Text PDF PubMed Scopus (994) Google Scholar). At lower flow rates, platelet adhesion is independent of GpIbα and VWF and involves other adhesive interactions, including those between collagen and GpIa-IIa (integrin α2β1) (7Nieuwenhuis H.K. Akkerman J.W. Houdijk W.P. Sixma J.J. Nature. 1985; 318: 470-472Crossref PubMed Scopus (388) Google Scholar, 8Watson S.P. Gibbons J. Immunol. Today. 1998; 19: 260-264Abstract Full Text Full Text PDF PubMed Scopus (166) Google Scholar) and those between fibrinogen and GpIIb-IIIa (integrin αIIbβ3) (6Savage B. Saldivar E. Ruggeri Z.M. Cell. 1996; 84: 289-297Abstract Full Text Full Text PDF PubMed Scopus (994) Google Scholar). In addition to its role in platelet adhesion, data from confocal videomicroscopic studies suggest that GpIbα-VWF association contributes to platelet aggregation and thrombus growth (9Savage B. Almus-Jacobs F. Ruggeri Z.M. Cell. 1998; 94: 657-666Abstract Full Text Full Text PDF PubMed Scopus (672) Google Scholar, 10Ruggeri Z.M. Dent J.A. Saldivar E. Blood. 1999; 94: 172-178Crossref PubMed Google Scholar). Von Willebrand disease (VWD) is the one of the most common congenital bleeding disorders with a prevalence of at least 100/million individuals (11Rodeghiero F. Castaman G. Dini E. Blood. 1987; 69: 454-459Crossref PubMed Google Scholar, 12Sadler J.E. Mannucci P.M. Berntorp E. Bochkov N. Boulyjenkov V. Ginsburg D. Meyer D. Peake I. Rodeghiero F. Srivastava A. Thromb. Haemostasis. 2000; 84: 160-174Crossref PubMed Scopus (442) Google Scholar). Type 2B VWD is caused by a qualitative abnormality of VWF in which normal sized multimers of VWF are secreted, and platelet-VWF interaction is augmented by increased affinity of VWF for GpIbα that does not require any mediating substance (13Ruggeri Z.M. Pareti F.I. Mannucci P.M. Ciavarella N. Zimmerman T.S. N. Engl. J. Med. 1980; 302: 1047-1051Crossref PubMed Scopus (267) Google Scholar, 14DeMarco L. Mazzucato M. DeRoia D. Casonato A. Fererici A.B. Girolami A. Ruggeri Z.M. J. Clin. Investig. 1990; 86: 785-792Crossref PubMed Scopus (33) Google Scholar). Paradoxically, this gain-of-function is associated with bleeding, perhaps because the largest multimers spontaneously associate with platelets, leaving the circulation deficient in large potent forms of VWF (13Ruggeri Z.M. Pareti F.I. Mannucci P.M. Ciavarella N. Zimmerman T.S. N. Engl. J. Med. 1980; 302: 1047-1051Crossref PubMed Scopus (267) Google Scholar, 15Ruggeri Z.M. Lombardi R. Gatti L. Bader R. Valsecchi C. Zimmerman T.S. Blood. 1982; 60: 1453-1456Crossref PubMed Google Scholar, 16DeMarco L. Girolami A. Zimmerman T.S. Ruggeri Z.M. Proc. Natl. Acad. Sci. U. S. A. 1985; 82: 7424-7428Crossref PubMed Scopus (96) Google Scholar). At least 14 distinct mutations are associated with type 2B VWD (1Sadler J.E. Annu. Rev. Biochem. 1998; 67: 395-424Crossref PubMed Scopus (1109) Google Scholar, 17Ginsberg D. Sadler J.E. Thromb. Haemostasis. 1993; 69: 177-184Crossref PubMed Scopus (193) Google Scholar, 18Budde U. Schneppenheim R. Rev. Clin. Exp. Hematol. 2001; 5: 335-368Crossref PubMed Scopus (54) Google Scholar), and most of these mutations cluster in a single disulfide loop of the VWF A1 domain between residues Cys509 and Cys695 (17Ginsberg D. Sadler J.E. Thromb. Haemostasis. 1993; 69: 177-184Crossref PubMed Scopus (193) Google Scholar, 19Fujimura Y. Titani K. Holland L.Z. Russell S.R. Roberts J.R. Elder J.H. Ruggeri Z.M. Zimmerman T.S. J. Biol. Chem. 1986; 261: 381-385Abstract Full Text PDF PubMed Google Scholar, 20Randi A.M. Rabinowitz I. Mancuso D.J. Mannucci P.M. Sadler J.E. J. Clin. Investig. 1991; 87: 1220-1226Crossref PubMed Scopus (89) Google Scholar, 21Sadler J.E. J. Biol. Chem. 1991; 266: 22777-22780Abstract Full Text PDF PubMed Google Scholar). In a viscosimeter at low shear, VWF with type 2B mutations bind to platelets constitutively, however binding of wild-type VWF requires the presence of biological agents like ristocetin (22Murata M. Fukuyama M. Satoh K. Fujimura K. Yoshioka A. Takahashi H. Handa M. Kawai Y. Watanabe K. Ikeda Y. J. Clin. Investig. 1993; 92: 1555-1558Crossref PubMed Scopus (28) Google Scholar). Biochemical studies (23Berndt M.C. Ward C.M. Booth W.J. Castaldi P.A. Mazurov A.V. Andrews R.K. Biochemistry. 1992; 31: 11144-11151Crossref PubMed Scopus (116) Google Scholar, 24Sobel M. Soler D.F. Kermode J.C. Harris R.B. J. Biol. Chem. 1992; 267: 8857-8862Abstract Full Text PDF PubMed Google Scholar, 25Christophe O. Obert B. Meyer D. Girma J.P. Blood. 1991; 78: 2310-2317Crossref PubMed Google Scholar) show that this disulfide loop encompasses the binding sites for GpIbα, heparin, and sulfatides. Platelet-type or "pseudo" VWD is associated with mutations in GpIbα that enhance affinity for VWF in the absence of injury and elevated shear stress, including Met239 to Val substitution (26Russell S.D. Roth G.J. Blood. 1993; 81: 1787-1791Crossref PubMed Google Scholar, 27Miller J.L. Thromb. Haemostasis. 1996; 758: 865-869Google Scholar, 28Moriki T. Murata M. Kitaguchi T. Anbo H. Handa M. Watanabe K. Takahashi H. Ikeda Y. Blood. 1997; 90: 698-705Crossref PubMed Google Scholar). Patients with either type 2B or platelet-type VWD show similar abnormalities, including loss of high molecular weight multimers of VWF in plasma, increased ristocetin-induced platelet aggregation, prolonged bleeding time, and intermittent thrombocytopenia (13Ruggeri Z.M. Pareti F.I. Mannucci P.M. Ciavarella N. Zimmerman T.S. N. Engl. J. Med. 1980; 302: 1047-1051Crossref PubMed Scopus (267) Google Scholar, 27Miller J.L. Thromb. Haemostasis. 1996; 758: 865-869Google Scholar, 29Ruggeri Z.M. Zimmerman T.S. Blood. 1987; 70: 895-904Crossref PubMed Google Scholar, 30Weiss H.J. Meyer D. Rabinowitz R. Pietu G. Girma J.P. Vicic W.J. Rogers J. N. Engl. J. Med. 1982; 306: 326-333Crossref PubMed Scopus (176) Google Scholar). Recent binding studies (31Doggett T.A. Girdhar G. Lawshe A. Schmidtke D.W. Laurenzi I.J. Diamond S.L. Diacovo T.G. Biophys. J. 2002; 83: 194-205Abstract Full Text Full Text PDF PubMed Scopus (160) Google Scholar, 32Doggett T.A. Girdhar G. Lawshe A. Miller J.L. Laurenzi I.J. Diamond S.L. Diacovo T.G. Blood. 2003; 102: 152-160Crossref PubMed Scopus (59) Google Scholar) suggest that both disorders have similar alterations in GpIbα-VWF bond formation and dissociation. Crystal structures of the unliganded VWF A1 domain (A1), the unliganded extracellular domain of GpIbα, and a complex comprised of gain-of-function mutants of GpIbα and A1 have been determined (33Emsley J. Cruz M. Handin R. Liddington R. J. Biol. Chem. 1998; 273: 10396-10401Abstract Full Text Full Text PDF PubMed Scopus (235) Google Scholar, 34Uff S. Clemetson J.M. Harrison T. Clemetson K.J. Emsley J. J. Biol. Chem. 2002; 38: 35657-35663Abstract Full Text Full Text PDF Scopus (158) Google Scholar, 35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar). The overall fold of GpIbα consists of a central domain of eight leucine-rich repeats that is flanked by a β-hairpin at the N-terminal end (residues Cys4-Cys17) and a conserved disulfide loop and the anionic region at the C-terminal end (34Uff S. Clemetson J.M. Harrison T. Clemetson K.J. Emsley J. J. Biol. Chem. 2002; 38: 35657-35663Abstract Full Text Full Text PDF Scopus (158) Google Scholar, 35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar). Structural, biochemical, and mutagenesis data identify two sites on the concave face of GpIbα that interact with A1, the N-terminal β-hairpin and a flexible loop at the C-terminal end termed the regulatory "β-switch" region, which undergoes conformational rearrangement upon binding of VWF (35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar, 36Matsushita T. Sadler J.E. J. Biol. Chem. 1995; 270: 13406-13414Abstract Full Text Full Text PDF PubMed Scopus (130) Google Scholar, 37Cauwenberghs N. Vanhoorelbeke K. Vauterin S. Westra D.F. Romo G. Huizinga E.G. Lopez J.A. Berndt M.C. Harsfalvi J. Deckmyn H. Blood. 2001; 98: 652-660Crossref PubMed Scopus (78) Google Scholar). Based on the 3.1-Å co-crystal structure and binding data, it has been proposed that type 2B VWD mutations increase the affinity of GpIbα-A1 interaction by displacing the N and C termini of A1 that could otherwise block GpIbα binding (35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar). An alternate mechanism, based on the structure of A1 mutant I546V, has been proposed suggesting that this mutation causes positional changes in the β2-β3-hairpin of A1 (27 Å away) that result in a higher affinity for GpIbα (38Celikel R. Ruggeri Z.M. Varughese K.I. Nat. Struct. Biol. 2000; 7: 881-884Crossref PubMed Scopus (66) Google Scholar). In addition, binding studies of VWD mutants have indicated that both type 2B and platelet-type VWD mutations enhance platelet adhesion, in part, by reducing the GpIbα-A1 dissociation rate (31Doggett T.A. Girdhar G. Lawshe A. Schmidtke D.W. Laurenzi I.J. Diamond S.L. Diacovo T.G. Biophys. J. 2002; 83: 194-205Abstract Full Text Full Text PDF PubMed Scopus (160) Google Scholar, 32Doggett T.A. Girdhar G. Lawshe A. Miller J.L. Laurenzi I.J. Diamond S.L. Diacovo T.G. Blood. 2003; 102: 152-160Crossref PubMed Scopus (59) Google Scholar). To gain insight into the structural basis for the enhancement of adhesion observed with VWD mutations and to identify specific changes associated with these mutations, we have crystallized a complex comprised of the extracellular domain of GpIbα and the wild-type A1 domain of VWF and determined the structure at a 2.6-Å resolution. Comparison of this GpIbα-A1 structure with unliganded GpIbα, the A1 domain, and the co-structure of the complex containing gain-of-function proteins GpIbα-M239V and A1-R543Q reveals a conformational rearrangement within the regions of GpIbα and A1 that form the binding surfaces of these proteins. The structural changes accompanying both mutations provide a rationale for the observed enhanced binding in VWD. Expression and Purification—Constructs encoding residues 1–305 of the mutated GpIbα domain lacking sites of glycosylation (N21D and N159D) were fused to the CH2-CH3 region of human IgG1 via an intervening enterokinase cleavage sequence (Asp-Asp-Asp-Asp-Lys) and expressed in Chinese hamster ovary cells. The sites of N-linked glycosylation are remote from the A1 binding site. GpIbα-Fc was recovered from conditioned medium by protein A-Sepharose (Amersham Biosciences) chromatography. Monomeric GpIbα domains (residues 1–288, note secondary cleavage site) were produced by digestion of the dimeric GpIbα-Fc with enterokinase (39LaVallie E.R. Rehemtulla A. Racie L.A. DiBlasio E.A. Ferenz C. Grant K.L. Light A. McCoy J.M. J. Biol. Chem. 1993; 268: 23311-23317Abstract Full Text PDF PubMed Google Scholar). The liberated, heterogeneous GpIbα proteins were further purified by gel filtration chromatography on a HiTrap Q-Sepharose HP column (Amersham Biosciences). Amino acid composition analyses and matrix-assisted laser desorption ionization-time of flight and electrospray ionization tandem mass spectrometry (in the negative ion modes) after proteolytic digestion were used to determine the GpIbα primary structure. Two peaks from the Q column corresponded to the predicted sizes of GpIbα with either two or three sulfated residues, and only the latter was used for crystal screening. The wild-type A1 domain (residues 496–709, numbering derived from the mature VWF subunit) was expressed as insoluble protein (inclusion bodies) in Escherichia coli cells grown at 37 °C. To express the selenomethionine-substituted form of A1, cells were grown in defined LeMaster medium supplemented with 75 μg/ml selenomethionine, induced at an A600 of 0.47 by the addition of 500 μm isopropyl-1-thio-β-d-galactopyranoside and harvested after 4 h at 37 °C. Cells were lysed at 4 °C in 50 mm Tris-HCl (pH 8.0), 10 mm NaCl, 10% glycerol, 1 mm EDTA, 1 mm phenylmethylsulfonyl fluoride, 1 mm dithiothreitol. The lysate was centrifuged at 9000 × g at 4 °C for 60 min. The pellet containing insoluble VWF-A1 was resuspended with 8 m urea, 20 mm HEPES, 20 mm dithiothreitol, 100 mm NaCl (pH 8.0), 20 mm dithiothreitol and centrifuged at 21,000 × g at 4 °C for 60 min. The urea-solubilized A1 domain was subjected to cation exchange chromatography using a Toyopearl SP 550C column at room temperature. The column was developed using a linear gradient from 8 m urea, 20 mm HEPES, 100 mm NaCl (pH 8.0) to 8 m urea, 20 mm HEPES, 400 mm NaCl (pH 8.0). The resulting A1-containing fraction was dialyzed against 6 m GdnHCl, 20 mm HEPES (pH 8.0) and diluted 20-fold to 2 m GdnHCl, 20 mm HEPES (pH 8.0). Oxidized glutathione was added to a final concentration of 0.1 mm, and the mixture was incubated for 16 h at room temperature. Refolded A1 was dialyzed against 20 mm HEPES, 100 mm NaCl (pH 8.0) before application to a Toyopearl DEAE 550C column. The unbound fraction from the DEAE column was applied to a Toyopearl CM 550S column. Refolded A1 was recovered using a linear gradient from 20 mm HEPES, 100 mm NaCl (pH 8.0) to 20 mm HEPES, 500 mm NaCl (pH 8.0) at room temperature. The A1-containing fraction was further purified on a Bio-Rad hydroxyapatite column. Purified A1 was isolated using a linear gradient from 20 mm HEPES (pH 8.0) to 20 mm HEPES, 0.5 m sodium phosphate monobasic, 0.5 m sodium phosphate dibasic (pH 6.6) at room temperature. Purified A1 was judged to be pure (>95%) by SDS-PAGE. A complex containing GpIbα and A1 complex was purified to homogeneity by gel filtration chromatography over Superdex-200 (Amersham Biosciences). The formation of 1:1 complexes of GpIbα and A1 was confirmed by native polyacrylamide gel electrophoresis and size exclusion chromatography prior to crystal screening. Crystallization and Structure Determination—Co-crystals of a complex comprised of GpIbα and selenomethionine-A1 were grown at 18 °C in microseeded hanging drops containing 10 mg/ml protein complex, 13% polyethylene glycol 3350, 50 mm Tris-HCl (pH 8.0), 200 mm potassium thiocyanate. Drops were microseeded using crushed crystals of the native GpIbα-A1 complex obtained under similar conditions (12% polyethylene glycol 3350, 50 mm Tris-HCl (pH 8.0), 200 mm potassium thiocyanate). The crystals have the symmetry of space group P21212 with unit cell dimensions of a = 72.5 Å, b = 116.1 Å, and c = 68.0 Å. Prior to data collection, crystals were briefly transferred to 25% ethylene glycol plus mother liquor and flash cooled in liquid nitrogen. Throughout data collection, the crystal was maintained at 100 K. Data were collected at the 5.0.2 beam line at the Advanced Light Source, Berkeley, CA. Intensities were integrated and scaled using DENZO (40Otwinowski Z. Minor W. Methods Enzymol. 1997; 276: 307-326Crossref Scopus (38253) Google Scholar) and SCALA (41CCP4Acta Crystallogr. Sect. D. 1994; 50: 760-763Crossref PubMed Scopus (19668) Google Scholar). The structure was solved with experimental phases derived from multiwavelength anomalous diffraction at the Se2+ edge (see Table I). Four Se2+ sites were identified by Patterson methods (SHELXS) based on the Bijvoet differences from a three-wavelength anomalous diffraction data set. The "heavy atom" model was then refined against a maximum likelihood target function using SHARP (42de La Fortelle E. Brigogne G. Methods Enzymol. 1997; 276: 472-494Crossref PubMed Scopus (1796) Google Scholar). Initial phases were improved by solvent flattening using Solomon as implemented in CCP4 (41CCP4Acta Crystallogr. Sect. D. 1994; 50: 760-763Crossref PubMed Scopus (19668) Google Scholar).Table IStatistics for data collection and refinementData collectionMultiwavelength anomalous diffractionPeakInflectionRemoteWavelength (Å)λ2 = 0.97942λ1 = 0.97959λ3 = 0.9686SourceaALS, Advanced Light SourceALS 5.0.2ALS 5.0.2ALS 5.0.2Resolution range (Å)30–2.730–2.730–2.7RsymbRsym = Σ|Ih — 〈Ih〉|/ΣIh, where 〈Ih〉 is the average intensity over symmetry equivalents. Number in parentheses reflects statistics for the last resolution shell (2.8–2.7 Å)(%)9.0 (29.5)9.8 (33.8)10.7 (38.2)I/σ(I)17.3 (3.5)15.3 (3.1)12.2 (2.4)Completeness (%)94.9 (74.5)93.0 (68.1)91.3 (61.3)Redundancy444Accentric phasing power1.25 (0.43)0.64 (0.23)0.35 (0.12)Figure of meritAcentric0.39 (0.12)Centric0.25 (0.06)Space groupP21212Unit cell dimensionsa = 72.5 Å; b = 116.1 Å; c = 68.0 ÅModel refinementNativeMaximum resolution (Å)2.6Number of reflections (free)17150 (1685)Rwork/RfreecRwork = Σ‖Fobs| — |Fcalc‖/Σ|Fobs|, Rfree is equivalent to Rwork, but is calculated for a randomly chosen 6.4% of reflections omitted from the refinement process (%)19.0/23.9No. of protein atoms3666No. of waters318Root mean square deviation from ideal geometryBonds (Å)0.006Angles (°)1.26a ALS, Advanced Light Sourceb Rsym = Σ|Ih — 〈Ih〉|/ΣIh, where 〈Ih〉 is the average intensity over symmetry equivalents. Number in parentheses reflects statistics for the last resolution shell (2.8–2.7 Å)c Rwork = Σ‖Fobs| — |Fcalc‖/Σ|Fobs|, Rfree is equivalent to Rwork, but is calculated for a randomly chosen 6.4% of reflections omitted from the refinement process Open table in a new tab Experimental maps with continuous density were obtained, and an initial model was constructed using QUANTA (Molecular Simulations, Inc.) and refined against data from 30 to 2.6 Å with CNS (43Brunger A.T. Adams P.D. Clore G.M. De Lano W.L. Gros P. Grosse-Kunstleve R.W. Jiang J.S. Kuszewski J. Nilges M. Pannu N.S. Read R.J. Rice L.M. Simonson T. Warren G.L. Acta Crystallogr. Sect. D. 1998; 54: 905-921Crossref PubMed Scopus (16919) Google Scholar). The final refined model, which includes polypeptide chains of GpIbα (residues 1–265) and of the VWF A1 domain (residues 506–702), as well as 318 water molecules, has a working R value of 0.190 and a free R value of 0.239 (Table I). The stereochemistry is excellent, and there are no backbone torsion angles outside of the allowed regions of the Ramachandran plot. Structural figures were generated using Ribbons (44Carson M. J. Appl. Crystallogr. 1991; 24: 958-961Crossref Scopus (783) Google Scholar) and PyMOL (45De Lano W.L. The PyMOL Molecular Graphics System. DeLano Scientific, San Carlos, CA2002Google Scholar). Overall Structure of GpIbα-A1 and Comparison to Unliganded GpIbα and A1—The x-ray crystal structure presented herein (Fig. 1A) shows that the interaction between the wild-type A1 and GpIbα involves the same binding interfaces as in the GpIbα-A1 complex comprised of mutant proteins, described previously (35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar). As seen in the mutant structure, changes in the conformation between unliganded GpIbα (34Uff S. Clemetson J.M. Harrison T. Clemetson K.J. Emsley J. J. Biol. Chem. 2002; 38: 35657-35663Abstract Full Text Full Text PDF Scopus (158) Google Scholar, 35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar) and GpIbα bound to A1 map to three regions: the N-terminal β-hairpin (residues Cys4-Cys17), the β-switch region (residues Val227-Ser241), and the C-terminal sulfated anionic region (Asp269-Tyr279). Early studies (46Dong J.F. Li C.Q. Lopez J.A. Biochemistry. 1994; 33: 13946-13953Crossref PubMed Scopus (106) Google Scholar, 47Marchese P. Murata M. Mazzucato M. Pradella P. DeMarco L. Ware J. Ruggeri Z.M. J. Biol. Chem. 1995; 270: 9571-9578Abstract Full Text Full Text PDF PubMed Scopus (113) Google Scholar, 48Ward C.M. Andrews R.K. Smith A.I. Berndt M.C. Biochemistry. 1996; 35: 4929-4938Crossref PubMed Scopus (180) Google Scholar) suggested that the sulfated anionic region at the C-terminal end of GpIbα (residues Asp269-Asp287) is important for binding to VWF and presumably represents an additional site of interaction. However, this region is disordered in the wild-type GpIbα-A1 co-structure, as well as in the mutant co-structure, and more recent biochemical evidence shows that deletion of this region does not effect binding affinity to the A1 domain itself (35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar). The A1 domain bound to GpIbα remains mostly unchanged, except for the conformational divergence that is seen in three adjacent regions (Fig. 1, B and C): the N-terminal end (residues beyond Ser510, inclusively), the C-terminal end (residues beyond Glu700), and the α1-β2 loop (residues Arg543-Arg552). In our structure the Cys509-Cys695 disulfide bridge that links the A1 termini is well ordered, but residues N-terminal to Asp506 and C-terminal to Pro702 are disordered. In the unliganded A1 structure an additional stretch of residues at the N-terminal end (Asp498-Asp506) and at the C-terminal end (Glu700-Thr705) extend toward the site of GpIbα interaction, but the significance of these extensions is not clear because of the presence of water-mediated crystal contacts involving several residues within these termini in the unliganded A1 (33Emsley J. Cruz M. Handin R. Liddington R. J. Biol. Chem. 1998; 273: 10396-10401Abstract Full Text Full Text PDF PubMed Scopus (235) Google Scholar). The location of the N-terminal (Asp506-Ser510) and C-terminal (Glu700-Pro703) regions with well defined electron density is quite different in our structure compared with unliganded A1. Differences in these regions and structural disorder in the A1 termini extensions have also been noted previously in comparison of unliganded A1 and the mutant GpIbα-A1 complex (35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar). The substantial rearrangement of the α1-β2 loop in the complex containing wild-type A1 represents a novel observation, which was not apparent in the mutant GpIbα-A1 complex, as will be described in greater detail below. In wild-type A1, the α1-β2 loop is shifted by more than 6 Å and is rotated away from GpIbα upon complex formation (Fig. 1B), whereas the α1-β2 loop of the unliganded A1 adopts a position that would create a steric clash with GpIbα and therefore would be incompatible with receptor binding (Fig. 1C). Comparison of Wild-type and Mutant GpIbα-A1 Co-structures—The overall GpIbα-A1 complex structure is in general similar to that of the GpIbα-M239V-A1-R543Q complex with a root mean square deviation of 1.1 Å after a superposition of 459 matched C-α atoms (Fig. 2). However, a thorough comparison of these co-structures reveals notable changes that have implications with regard to the different binding affinities of these proteins (14DeMarco L. Mazzucato M. DeRoia D. Casonato A. Fererici A.B. Girolami A. Ruggeri Z.M. J. Clin. Investig. 1990; 86: 785-792Crossref PubMed Scopus (33) Google Scholar, 32Doggett T.A. Girdhar G. Lawshe A. Miller J.L. Laurenzi I.J. Diamond S.L. Diacovo T.G. Blood. 2003; 102: 152-160Crossref PubMed Scopus (59) Google Scholar, 35Huizinga E.G. Tsuji S. Romijn R.A. Shiphorst M.E. de Groot P.G. Sixma J.J. Gros P. Science. 2002; 297: 1176-1179Crossref PubMed Scopus (492) Google Scholar). Structural Differences in the α1-β2 Loop of A1—The most prominent differences between wild-type and mutant A1 domains are observed more than 15 Å away from the R543Q mutation site. These differences include reconfiguration of the α1-β2 loop and changes involving residues that are in direct contact with GpIbα (Arg571, Glu613). The single amino acid Arg → Gln substitution affects the position of equivalent atoms without extensive disruption of structure in nearby regions. In both structures, with either Arg or Gln present at position 543, main chains of these residues adopt similar conformations and participate in intrahelical hydrogen bonding at the C-terminal end of the α1-helix. Structural deviations associated with main chain rearrangement begin at Leu544 where the orientation of its carbonyl oxygen is flipped relative to its orientation in the mutant complex, thus extending the hydrogen bond capping at the C-terminal end of the α1-helix (Fig. 3). As a result, structural neighbors Arg545 and Ile546 adjust their position
Referência(s)