Fig 1.
Linear VWF multimers are built from dimers as smallest repeating subunits, which are linked in a head-to-head fashion via the disulfide bonds Cys1099-Cys1099’ and Cys1142-Cys1142’ between their N-terminal D`D3 assemblies. Dimers consist of two monomers–comprising the domains D`D3, A1, A2, A3, D4, C1-C6, and CK–that are linked in a tail-to-tail fashion via the disulfide bonds Cys2771-Cys2773’, Cys2773-Cys2771’, and Cys2811-Cys2811’ between their C-terminal CK domains.
Fig 2.
Multimer analysis of wildtype and mutant VWF by AFM-based imaging of individual molecules.
(A, B) Representative AFM images of individual native (A) and non-native (B) VWF molecules. Numbers in images indicate the multimer size i (i = 1 corresponds to a dimer). White arrowheads mark paired, red arrowheads unpaired CK domains. For more details on the identification of dimeric and monomeric building blocks within VWF molecules, see Materials and Methods and S1 Fig. Scale bars represent 20 nm, range of color scale is 2.4 nm. (C) Size distributions of wtVWF (i) and mutant p.Cys1099Tyr (C1099Y, ii), and schematic of step-growth multimerization (iii). Insets in subpanels i and ii show linear fits to the data represented in logarithmic space, yielding values for the extent p of multimerization of 0.43 and 0.21, respectively. (D) Size distributions of VWF mutants p.Cys2771Arg (C2771A, i) and p.Cys2773Arg (C2773R, ii), and schematic of the underlying multimerization process (iii). Native and non-native molecules are depicted in blue and red, respectively. Non-native molecules are characterized by ending on a C-terminal CK domain (small, closed circle) at one or both termini, while native molecules end on N-terminal D’D3 assemblies (open circle). Monomers are “non-native” because they are never secreted after expression of wtVWF. From the observed size distributions, values for the dimerization abolishment of 87% and 73% were determined for p.Cys2771Arg and p.Cys2773Arg, respectively. (Ei) Size distribution of VWF mutant p.Cys2811Ala, for which non-native molecules had been hypothesized to result from reopening of disulfide-linked CK domains. The overall ratio of non-native molecules was found to be 51%. The (apparent) reopening probability was determined to be 23%. (Eii) Size distribution (shown for i ≤ 5) obtained from a simulation that assumed multimers to initially follow an exponential size distribution–with p = 0.43 as observed for wtVWF–and to afterwards reopen partially at their CK domains with the experimentally determined probability for p.Cys2811Ala. Simulations yielded, similarly to p.Cys2811Ala, a ratio of 53% non-native molecules, and very low fractions of even-numbered non-native molecules. (Eiii) Schematic representation of the hypothesized scenario of initial wildtype-like step-growth multimerization and subsequent reopening of C-terminal disulfide bonds within constituent dimers. The data presented in this figure are listed in S1 Table.
Fig 3.
Quantitative electrophoretic multimer analysis of wildtype and mutant VWF.
(A) High-resolution agarose gel of multimer samples. Numbers beneath bands indicate the multimer size i (i = 1 corresponds to a dimer). (B-F) Frequencies of molecules of size i, as determined from the gel (for more details, see Materials and Methods and S4 Fig), for wtVWF (B) and mutants p.Cys1099Tyr (C), p.Cys2771Arg (D), p.Cys2773Arg (E), and p.Cys2811Ala (F). Insets in panels B and C show linear fits to the frequency data represented in logarithmic space, yielding values for the extent p of multimerization of 0.47 and 0.15 for wtVWF and mutant p.Cys1099Tyr, respectively.
Fig 4.
Stability of pseudo-WPBs in HEK293 cells expressing wtVWF or mutant p.Cys2811Ala.
HEK293 cells stably overexpressing wtVWF (middle lane) or mutant p.Cys2811Ala (bottom lane) were lysed using M-PER lysis buffer containing 1% (w/v) octylglycoside. The lysates were analyzed by electrophoretic multimer analysis. As size reference, wtVWF secreted into the medium is shown (top lane).