Force-induced cleavage of single VWFA1A2A3 tridomains by ADAMTS-13

Abstract

A disintegrin and metalloprotease with a thrombospondin type 1 motifs 13 (ADAMTS-13) regulates hemostasis by cleaving the folded A2 domain of von Willebrand factor (VWF). The cleavage is regulated by forces as it occurs in flowing blood. We tested the hypothesis that force-induced A2 domain unfolding facilitates cleavage using atomic force microscopy to pull single VWF A1A2A3 tridomain polypeptides by platelet glycoprotein Ibalpha or antibodies to measure time, distance, and force. Structural destabilization of A1A2A3 was induced by 5- to 80-pN forces, manifesting as an abrupt molecular length increase distributed around 20 and 50 nm, probably because of uncoupling A1A2A3 (or partially unfolding A2) and fully unfolding A2, respectively. Time required to destabilize A1A2A3 first increased (catch), reaching a maximum of 0.2 seconds at 20pN, then decreased (slip) with increasing force, independent of ADAMTS-13. The time required to rupture A1A2A3 exhibited a similar catch-slip behavior when pulled by glycoprotein Ibalpha but only slip behavior when pulled by antibody, which was progressively shortened by increasing concentration of ADAMTS-13 after (but not before) structural destabilization, indicating that cleavage of A2 requires the force-induced A2 unfolding. Analysis with a model for single-substrate trimolecular enzymatic kinetics estimated a cleavage rate k(cat) of 2.9 (+/- 59) seconds and a K(d) of 5.6 (+/- 3.4) nM for ADAMTS-13/A1A2A3 binding. These findings quantify the mechanical regulation of VWF cleavage by ADAMTS-13 at the level of single A1A2A3 tridomain.

Figure 1

Atomic force microscopy (AFM) setup. (A) AFM schematic. (B) Functionalization of AFM. Molecules depicted represent a composite of 2 sets adsorbed or captured on the AFM tip or the polystyrene dish. A1A2A3 tridomain was directly adsorbed or captured by anti-His mAb preadsorbed on the surface. GPIbα or CR1 was adsorbed on the cantilever tip. (C) Binding specificity. Immobilized A1A2A3 bound GPIbα- or CR1-coated cantilever tips but not BSA-coated tips. ND indicates not done. GPIbα did not bind captured A1A2A3 probably because of its different conformation from adsorbed A1A2A3.

Figure 2

Force curves. (A-B) Schematic illustrations. (C-F) Example data. The PZT retracted the cantilever base linearly to a preset position (B,C-F, red lines and right ordinate) and then held it there to stretch the molecules with a constant force (A,C-F, blue curves and left ordinate) to measure various times characteristic of molecular structural destabilization or rupture induced by this force. (C) A curve of 2 force drops (indicated by 2 in the first subscript of t_ij) and 2 periods at constant force (indicated by 1 and 2 in the second subscript of t_ij) with respective durations of time-to-destabilization t₂₁ and time-to-rupture t₂₂. Force drop triggered via feedback control further retraction of the PZT attempting to bring the force back to the preset level, which was successful after the first force drop by retracting a distance l_d but was unsuccessful after the second force drop even by long-distance retraction, indicating rupture. (D) A curve of 2 force drops but only 1 period at constant force with duration of t₂₁ but no t₂₂ resulting from premature rupture occurred before force was resumed to the preset level. (E) A curve of 2 force drops but only 1 period at constant force with duration of t₂₂ but no t₂₁ resulting from premature destabilization occurred before arriving at the preset force. (F) A curve of 1 force drop (indicated by 1 in the first subscript of t_ij, i, j = 1 or 2) and 1 period at constant force (indicated by 1 in the second subscript of t_ij) with duration of time-to-rupture t₁₁. Data shown were acquired by GPIbα-pulling, but those obtained by CR1-pulling were similar.

Figure 3

Analysis of the length increase resulted from loss of stability. (A) Histograms of molecular length increment, l_d, obtained using adsorbed A1A2A3 pulled by GPIbα (blue bars) or anti-His mAb captured A1A2A3 pulled by CR1 (red bars). (B-C) Force-time and displacement-time curves with 2 force drops were converted to force–molecular extension curves: 16 and 10 curves, respectively, obtained by GPIbα (blue) and CR1 (red) pulling with short (l_d ∼ 10-20 nm) length increase (B) as well as 10 and 8 curves (obtained by the 2 pulling methods with the same color codes) with long (l_d ∼ 45-50 nm) length increase (C) are overlaid by aligning the first ascending segment. (D) Possible modes of structural destabilization. A model for the native A1A2A3 structure is shown in the far left panel. Force may disrupt the interdomain interactions to uncouple the quaternary structure of the A1A2A3 tridomain (second panel from left) or disrupt the intradomain interactions to partially (third panel from left) or fully (far right panel) unfold the tertiary structure of the A2 domain.

Figure 4

Impact of mode of destabilization. (A-B) Comparison of destabilization times (A) or rupture times (B) for molecules with structural destabilization that yielded short (l_d < 35 nm, open bars) and long (l_d > 35 nm, closed bars) length increases in the absence or presence of 5 μg/mL ADAMTS-13 without or with 5mM EDTA. (C-D) Comparison of time-to-destabilization t₂₁ (C) or time-to-rupture t₂₂ (D) measured in the absence (open bars) or presence of ADAMTS-13 without (closed bars) or with (hatched bars) EDTA for molecules with structural destabilization that yielded short or long length increases. Data were acquired by GPIbα-pulling and presented as mean ± SEM of several tens of measurements. P values of Student t test are shown to indicate the statistical significance (or lack thereof) of the differences.

Figure 5

Effects of ADAMTS-13 concentration. Time-to-rupture t₁₁ (A), time-to-destabilization t₂₁ (B), and time-to-rupture t₂₂ of the short (C) and long (D) groups were plotted versus ADAMTS-13 concentration. With the exception of t₂₂, which was shortened by ADAMTS-13 in a dose-dependent manner, other time parameters were indifferent to the changing ADAMTS-13 concentration. DisC is a control construct without VWF-cleaving activity. Data were acquired by GPIbα-pulling and presented as mean ± SEM of several tens of measurements.

Figure 6

Enzymatic kinetics. (A) Data from Figure 5D (points) were replotted using molar concentration and fitted by Equation 1 (curve). (B) Lineweaver-Burk plot of reciprocal initial rate of substrate reduction versus reciprocal substrate concentration. The data (points) were fitted by a straight line. The best-fit parameters are indicated. The goodness of fit is indicated by R² (B) or adjusted R² (A).

Figure 7

Force-dependent kinetics of destabilization and rupture. Time-to-rupture t₁₁ (A,E) or t₂₂ (C,G) or time-to-destabilization t₂₁ (B,F) is plotted versus force measured in the absence (□) or presence of 5 μg/mL ADAMTS-13 without (◊) or with (○) 5mM EDTA for A1A2A3 adsorbed directly on polystyrene dish and pulled by GPIbα (A-C) or captured by preadsorbed anti-His mAb and pulled by CR1 (E-G). Data are mean ± SEM of several tens of measurements for each point. The rate of cleavage as a function of force was calculated by k_cat = 1/t₂₂ − 1/t₂₂⁰ denotes using the respective data in panel C (D) or panel G (H).