Crystal structure of the A3 domain of human von Willebrand factor: implications for collagen binding

Abstract

The structure of the A3 domain suggests that adhesion to collagen is primarily achieved through interactions between negatively charged residues on A3 and positively charged residues on collagen. The absence of a pronounced binding groove precludes a large van der Waals surface interaction between A3 and collagen and is consistent with the low affinity for collagen of a single A3 domain and the requirement for multimeric vWF for tight association with collagen. The absence of bound metal ions upon soaking the crystal in MgCl2 and vWF A3's conformational incompatibility for metal binding is consistent with the absence of a functional role for metal ion binding in A3, which contrasts the metal ion activation required for ligand binding by the homologous integrin I type domains.

Structural and functional features of a collagen-binding matrix protein from the mussel byssus

“…2c), with predominating negative clusters, corresponding to the theoretical isoelectric point of PTMP1 (pI ¼ 5.9). The edges of each domain are interconnected by a disulphide bond, which is often observed in VWA domains [22][23][24] . The two PTMP1 domains are arranged 'back-to-back' with a small contact area involving only five potential hydrogen bonds (Q212-E413; K217-S411; K217-I407/N408/A410; H187-E414; T215-S412/Y415).…”

“…2e; Supplementary Table 1), demonstrating the stability of the relative domain arrangement. The two structures only differ in the length of a single a-helix (a6, numbering of the secondary structure elements according to published VWA domain structures 19,20,23 , see also Supplementary Fig. 2), which is three amino acids longer in crystal form 1 (residues 226-241) than in crystal form 2 (residues 229-241).…”

“…3b,c). Furthermore, in contrast to their unique tandem arrangement, the individual domains show a high structural homology to other VWA domains, like the A3 domain of human von Willebrand factor 23 or the I-domains of Integrin a1 and a2 19,20 (Fig. 3d-f), despite an only moderate sequence identity ( Supplementary Fig.…”

“…A structural overlay of the collagen-binding MIDAS of integrin a2 (ref. 19) and the respective motif of the human von Willebrand factor A3 domain 23 , which is incapable of binding metal ions, with the A1 domain of PTMP1 reveals that PTMP1 is structurally more related to the integrins than to vWFA3 (Fig. 5f), since A1 lacks an aspartate residue present in the MIDAS motif of vWFA3 that is thought to inhibit metal binding 23 .…”

“…19) and the respective motif of the human von Willebrand factor A3 domain 23 , which is incapable of binding metal ions, with the A1 domain of PTMP1 reveals that PTMP1 is structurally more related to the integrins than to vWFA3 (Fig. 5f), since A1 lacks an aspartate residue present in the MIDAS motif of vWFA3 that is thought to inhibit metal binding 23 . The A2 MIDAS motif, in contrast, did not bind any metal ions tested despite a perfectly conserved MIDAS sequence, and it appears to harbour subtle structural differences to the A1 MIDAS motif (see Fig.…”