Structural evidence proposes that the VWFpp mutations affect the N-terminal assembly and/or VWF recruitment for anterograde transport
Our VWF N-terminal model is organized into N terminal globular domains D1 and D2 (VWFpp) connected to the C terminal globular D3 domain by a filamentous like D´ domain. The D1 and D2 domains are characterized by several regions of disorder connecting three helix-bundles to beta sheeted regions in each domain. The disordered regions are most likely instrumental in the conformational changes occurring in the course of disulfide exchange within the ER. In the present model apart from 9 cysteines in the D1-D2 domains, all others were observed in an oxidized form. The D3:D3 domain interaction regions in the model suggest a clear symmetry within the monomer that serves the purposes of uni-directionality for the appropriate alignment of the dimeric pro-VWF subunits and during the multimerization process (Figure 4B). Therefore the D1, D2 domain surfaces most likely serve one of the two purposes: 1) as dimeric assembly interfaces or as 2) binding region for recruitment proteins (like ARF and Sec24) during anterograde transport of VWF from ER to Golgi. The mutations on the D1, D2 domains could, therefore, affect either (N-terminal assembly/recruitment process) or both functional aspects of these domains. The docking of Sec24 and ARF1 protein on the VWF N terminal model strongly supports this idea. The docking poses/clusters of both Sec24 and ARF1 show strong similarity in spatial positioning. The majority of the high scoring docking poses for both proteins occupy the D1 domain surfaces while select few were concentrated on the D2 domain, the docking surfaces on both domains facing away from the D3:D3 domain interface (Figures 4B, 4D, and 4F). The structural alignment of the active and inactive forms of ARF-1 i.e. GTP and GDP bound structures shows a specific ARF-1 region that undergoes a conformational change (also called switch region) and therefore is the part most likely to acts as the protein recruitment surface/interface for ARF-1 (Supplemental Figure 1). In most of our observed docks, this region forms a major part of the interface between ARF 1 and VWF (Supplemental Figure 2). Interestingly, the majority of our naturally occurring mutations (3 out of 4), are located exactly on the ARF 1/ or Sec24-VWF-D1 domain interaction interface. One D2 domain mutation i.e. p.Cys608Trp does not occur directly at the interface but close to it. Since the Cys608 participates in a disulfide bond, the mutation will result in an unpaired cysteine which could significantly influence disulfide bond combinations and also disulfide exchange (especially since there exist other unpaired Cysteines in close spatial proximity). Effectively this mutation, therefore, could have potential allosteric effects on the noncovalent D1-D2 dimer interfaces or affect interaction with ARF 1 or Sec24.