Von Willebrand disease is a bleeding disorder with reduced or abnormal function of VWF. While VWF has at least three functions (binding to FVIII, platelet GPIb, and collagen), the function of VWF has been quantified primarily using techniques in which ristocetin promotes the binding of VWF to platelet GPIb. It is recognized that the physiologic agonist for VWF-platelet interactions is shear but ristocetin facilitates a change in VWF structure to induce binding to platelet GPIb. For most individuals, this assessment of VWF function is valid, but over the past several years our group, the Zimmerman Program for the Molecular and Clinical Biology of VWD, has identified “polymorphisms” that do not affect VWF function but interfere with ristocetin binding to VWF in those laboratory tests of VWF function that utilize ristocetin. The two “polymorphisms” are P1467S and D1472H. P1467S is a rare sequence variant not seen in 12,900 alleles in the 1000-genome project. In contrast, the D1472H is a polymorphism seen in 63% of African Americans and in 17% of Caucasians. Since these two sequence variations are physically located in close proximity just C-terminal from the A1 loop, we initiated a systematic study of the adjacent N- and C-terminal sides of the A1 domain by alanine scanning mutagenesis.

We concentrated on the sequence between p.C1237 – p.C1272 and p.C1458 – p.K1491 and replaced each amino acid with an alanine. These sequences were then transferred into a full-length expression construct. The individual altered VWF proteins were studied by GPIb-binding in two solid phase assays. The first assay, VWF:RCo, is an ELISA using an antibody-captured recombinant GPIb-alpha in which ristocetin-induced bound VWF is detected using an anti-VWF monoclonal antibody. The second assay, VWF:IbCo, is a similar assay, but the expressed GPIb contained two mutations, D235Y and M239V, that confer spontaneous binding to VWF without the addition of ristocetin as our laboratory has previously published.

Where possible all assays were done at similar concentrations of VWF antigen and expressed as a ratio of VWF:IbCo/VWF:RCo compared to VWF:Ag concentration. An increased ratio indicates a reduction in ristocetin-induced binding. A pronounced difference was identified at several amino acids between p.D1459 and p.P1471. Interestingly, the D1472H polymorphism was not abnormal by this approach (D1472A) – thereby indicating that alanine substitution for the aspartic acid does not confer the same conformation that the histidine does in African Americans. Although earlier studies by Mohri and coworkers (JBC 1988) suggest a second binding site, only a minimal signal was identified for the region between p.C1237 –p.P1251. Hot spots were identified with VWF:IbCo/VWF:RCo scores 7- to 30-fold higher than background. To date, natural sequence variations have not yet been identified in these positions. Studies by Scott and coworkers (JBC 1991), suggested that ristocetin binding to proteins was proportional to the presence of X-P-G-X’ beta-turns. The region of VWF that we are studying contains one of these sequences and is associated with a major effect on the VWF:IbCo/VWF:RCo signal. Since ristocetin dimerizes at a concentration of 1 mg/ml and the dose used in VWF:RCo is usually 1 mg/ml, these studies support the concept that alterations in ristocetin binding brought about by sequence variations may give the false impression that VWF function is profoundly affected when it is only the in vitro assay that is abnormal.

Alanine scanning mutagenesis provides further substantiation of the role of the VWF-A1-loop in the binding of ristocetin to VWF. “Potential” Type 2M VWD subjects with sequence variations in this region may not have VWF functional abnormalities. Therefore, such individuals may require additional studies – particularly if bleeding symptoms are not identified in the presence of significantly abnormal laboratory testing.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.