Type 2B von Willebrand disease mutations differentially perturb autoinhibition of the A1 domain

Abstract

Type 2B von Willebrand disease (VWD) is an inherited bleeding disorder in which a subset of point mutations in the von Willebrand factor (VWF) A1 domain and recently identified autoinhibitory module (AIM) cause spontaneous binding to glycoprotein Ibα (GPIbα) on the platelet surface. All reported type 2B VWD mutations share this enhanced binding; however, type 2B VWD manifests as variable bleeding complications and platelet levels in patients, depending on the underlying mutation. Understanding how these mutations localizing to a similar region can result in such disparate patient outcomes is essential for detailing our understanding of VWF regulatory and activation mechanisms. In this study, we produced recombinant glycosylated AIM-A1 fragments bearing type 2B VWD mutations and examined how each mutation affects the A1 domain's thermodynamic stability, conformational dynamics, and biomechanical regulation of the AIM. We found that the A1 domain with mutations associated with severe bleeding occupy a higher affinity state correlating with enhanced flexibility in the secondary GPIbα-binding sites. Conversely, mutation P1266L, associated with normal platelet levels, has similar proportions of high-affinity molecules to wild-type (WT) but shares regions of solvent accessibility with both WT and other type 2B VWD mutations. V1316M exhibited exceptional instability and solvent exposure compared with all variants. Lastly, examination of the mechanical stability of each variant revealed variable AIM unfolding. Together, these studies illustrate that the heterogeneity among type 2B VWD mutations is evident in AIM-A1 fragments.

Graphical abstract

Figure 1.

Type 2B mutations destabilize purified AIM-A1. Locations of represented type 2B mutations on the 3-dimensional crystal structure (PDB:7A6O aligned with PDB:1SQ0 GPIb chain) (A) and linearized schematic on AIM-A1 (B). (C) Coomassie-stained sodium dodecyl sulfate– polyacrylamide gel electrophoresis gels of purified AIM-A1 variants under nonreducing (left) and reducing (right) conditions. (D) Unfolding temperatures (Tm₅₀) of AIM-A1 variants as determined by the first derivative of thermal shift data (n = 4 independent replicates). Significance was determined via 1-way analysis of variance with post hoc Tukey test correction compared with WT, where ∗∗∗∗P < .0001.

Figure 2.

Fast- and slow-dissociating AIM-A1 molecules describe a 2-state binding model to GPIbα. (A) Representative sensorgrams of indicated AIM-A1 variants at concentrations from 1 μM to 15.625 nM show that type 2B VWD mutants yield greater binding response to immobilized PT-LBD (W230L). (B) Steady-state binding data of sensorgrams displaying means ± standard deviations of n = 3 independent replicates.

Figure 3.

HDX data show differences in protection among AIM-A1 variants. Heatmaps of relative deuterium uptake produced in DynamX at indicated time points between 10 and 10 000 seconds. The color spectrum represents fractional deuterium uptake at near-residual resolution by subtracting exchange differences of overlapping peptides. A linear schematic of secondary structure is aligned to heatmaps and displayed above. WT AIM-A1 was denatured as described in “Materials” before HDX to compare maximal exchange. Values are produced from n = 3 technical replicates.

Figure 4.

AIM-A1 and type 2B mutants reveals conformational differences. (A-G) Volcano plots showing differences in deuterium uptake among shared peptides after 10 seconds of exchange. The threshold for significance (data points above the dotted lines) requires both statistical significance of P ≤ .01 and a minimum difference of 0.5 dalton per peptide. Plots were generated using HD-eXplosion version 1.2. Peptides with significant HDX differences are displayed in color in the quadrant of the AIM-A1 variant with greater deuterium uptake. (H) Relative deuterium uptake over time for a representative peptide in the α3β4 loop. All type 2B VWD mutations tested have increased solvent accessibility in this region compared with WT. (I) Representative uptake of peptide in β3α2 loop showing P1266L has protection comparable with WT in this region at the fastest time point tested but loses this protection over time similarly to other type 2B VWD mutants. The crystal structure of AIM-A1 (PDB:7A6O) in gray and the indicated peptide mapped in black are displayed on each graph.

Figure 4.

AIM-A1 and type 2B mutants reveals conformational differences. (A-G) Volcano plots showing differences in deuterium uptake among shared peptides after 10 seconds of exchange. The threshold for significance (data points above the dotted lines) requires both statistical significance of P ≤ .01 and a minimum difference of 0.5 dalton per peptide. Plots were generated using HD-eXplosion version 1.2. Peptides with significant HDX differences are displayed in color in the quadrant of the AIM-A1 variant with greater deuterium uptake. (H) Relative deuterium uptake over time for a representative peptide in the α3β4 loop. All type 2B VWD mutations tested have increased solvent accessibility in this region compared with WT. (I) Representative uptake of peptide in β3α2 loop showing P1266L has protection comparable with WT in this region at the fastest time point tested but loses this protection over time similarly to other type 2B VWD mutants. The crystal structure of AIM-A1 (PDB:7A6O) in gray and the indicated peptide mapped in black are displayed on each graph.

Figure 5.

Single-molecule force spectroscopy data reveal the force required to unfold AIM. (A) The optical tweezers were configured with BioSpy-AIM-A1 fixed between 2 streptavidin-functionalized polystyrene beads. The biotinylated N-terminal AIM was bound directly to the fixed bead, whereas the C-terminal AIM appended with a SpyTag connected to the laser-trapped bead with a DNA handle functionalized with biotin and a SpyCatcher molecule at either end. (B) Plots of the worm-like chain model fitting results of unfolding force vs unfolding extension for each mutant compared with the WTAIM-A1. Force data are presented as mean values ± standard deviation, and extension data are presented as the peak of the Gaussian fit ± the full width at half maximum of Gaussian fit divided by the square root of counts. (C) Plots of Bell-Evans model fitting results of unfolding force vs loading rate for each AIM-A1 mutant compared with the WT AIM-A1. Unfolding force data are presented as the center of the tallest bin of the histogram ± one-half of the bin width. Numerical values of associated fittings are summarized in Table 2.

Figure 6.

Proposed model illustrating type 2B VWD mutational effect on AIM-A1 thermodynamic cycling. Under normal blood flow, AIM-A1 exists in the bloodstream in an autoinhibited state because of cooperative protection from the N-terminal AIM (blue) and C-terminal AIM (orange), where both the α3β4 and β3α2 loops (light gray) have enhanced protection. Shear forces of the blood or other endogenous or exogenous factors lead to activation of the A1 domain, which equilibrates between low- and high-affinity states, in which the structure of the AIM is likely more flexible but is yet to be determined. The low-affinity A1 domain may transiently interact with GPIbα LBD as platelets roll over VWF with a higher dissociation rate constant. Based on this study, a potential molecular signature of an activated A1 domain is enhanced solvent exposure of the α3β4 loop (dark gray). The high-affinity state of the A1 domain dissociates from the LBD more slowly and can more effectively capture platelets. In this state, the A1 domain has enhanced exposure of both the α3β4 and β3α2 loops (dark gray). Severity of type 2B VWD may correlate with relative abundance of these high-affinity molecules.