When collagen meets VWF

Comment on Lisman et al, page 3753

In this issue of Blood, Lisman and colleagues report on the identification of the VWF-binding site within collagen III: of 57 triple-helical peptides spanning the whole triple-helical domain, one single 27-residue collagen III sequence, with an essential 9-residue stretch, sustained binding of the VWF A3 domain.

Von Willebrand factor (VWF) forms a bridge between collagens within the connective tissue exposed upon damage of the blood vessel and the platelet membrane glycoprotein Ib (GPIb), and therefore plays a pivotal role in the formation of an arterial platelet-rich thrombus. Several types of fibrillar collagen bind VWF, the major collagen-binding site of which is located within the A3 domain,¹  with a second site in the A1 domain.²  Collagen types I and III, in particular, provide strong binding sites for VWF. Although the binding site for collagen within the A3 domain has been resolved by mutagenesis and cocrystallization experiments with inhibitory antibodies, the interacting features within the collagens until now remained elusive.

In this issue, Lisman and colleagues used the Collagen III Toolkit,³  which contains 57 overlapping triple-helical peptides spanning the whole triple-helical domain of human collagen III, to identify one single 27-residue binding sequence within collagen. This was further narrowed down by analyzing truncation and Ala-substitution mutants to a final single stretch of 9 amino acids within the collagen sequence, provided it is presented in a triple-helical conformation. As the peptide binding was absent in the presence of inhibitory anti–A3-domain antibodies, or when using either an A3-domain deletion or dysfunctional mutant, it is clear that indeed the VWF A3 domain is involved.FIG1 

With these breakthrough materials at hand, the road is open to finally resolve the precise interaction mechanisms by computer modeling (see figure) and crystallography, providing information that ultimately may be used to design new compounds with specific antiadhesive and antithrombotic activity.

However, in the present selection and analysis, no experiments were performed under high shear flow conditions, where, for example, the A1-domain binding to collagen becomes apparent, but where also the binding interactions of the A3 domain may be modulated. Hence, it is likely that additional VWF binding sequences are present in collagen III.

Furthermore, as collagen III is not the only adhesive constituent within the exposed connective tissue, additional interacting proteins such as, for example, collagen I need to be considered. Since collagen I, however, is a heterotrimer, in contrast to the homotrimeric collagen III, it is much more difficult to produce corresponding synthetic triple-helical peptides. Based on the information obtained with collagen III, the authors offer a binding model for collagen I that awaits experimental validation.

Whatever the answers to these questions are, it is clear that the data presented here already represent a significant step forward in the understanding of arterial thrombus initiation.

The authors declare no competing financial interests. ▪