Integrin alphaIIbbeta3:ligand interactions are linked to binding-site remodeling

Abstract

This study tested the hypothesis that high-affinity binding of macromolecular ligands to the alphaIIbbeta3 integrin is tightly coupled to binding-site remodeling, an induced-fit process that shifts a conformational equilibrium from a resting toward an open receptor. Interactions between alphaIIbbeta3 and two model ligands-echistatin, a 6-kDa recombinant protein with an RGD integrin-targeting sequence, and fibrinogen's gamma-module, a 30-kDa recombinant protein with a KQAGDV integrin binding site-were measured by sedimentation velocity, fluorescence anisotropy, and a solid-phase binding assay, and modeled by molecular graphics. Studying echistatin variants (R24A, R24K, D26A, D26E, D27W, D27F), we found that electrostatic contacts with charged residues at the alphaIIb/beta3 interface, rather than nonpolar contacts, perturb the conformation of the resting integrin. Aspartate 26, which interacts with the nearby MIDAS cation, was essential for binding, as D26A and D26E were inactive. In contrast, R24K was fully and R24A partly active, indicating that the positively charged arginine 24 contributes to, but is not required for, integrin recognition. Moreover, we demonstrated that priming--i.e., ectodomain conformational changes and oligomerization induced by incubation at 35 degrees C with the ligand-mimetic peptide cHarGD--promotes complex formation with fibrinogen's gamma-module. We also observed that the gamma-module's flexible carboxy terminus was not required for alphaIIbbeta3 integrin binding. Our studies differentiate priming ligands, which bind to the resting receptor and perturb its conformation, from regulated ligands, where binding-site remodeling must first occur. Echistatin's binding energy is sufficient to rearrange the subunit interface, but regulated ligands like fibrinogen must rely on priming to overcome conformational barriers.

Figure 1.

Integrin:echistatin interactions measured with a solid-phase binding assay. Data are expressed as the fraction of maximum change in the absorbance at 405 nm (Signal) following incubation with increasing concentrations of biotinylated echistatin variant ([BT-rEch] nM) to αIIbβ3 immobilized in the wells of a microtiter plate. The dashed and black lines were obtained by fitting the data to a single-site saturable binding model to determine the dissociation constant, K_d, as indicated in the text. The following symbols apply: rEch (1–49) M28L D27W (black hexagons, dashed line), primed rEch (1–49) M28L (concentric circles, black line); resting (open circles), rEch (1–49) M28L R24A (gray diamonds), rEch (1–49) D26A (black triangles, dotted line obtained by linear regression).

Figure 2.

Effects of recombinant echistatin variants on the distribution of αIIbβ3 sedimenting species, g(S) versus S (time-derivative software; DCDT+; Philo 1997). Each panel compares the distribution obtained with ligand-free integrin (open circles) to a paired sample containing excess echistatin variant. Continuous lines were obtained by fitting the resultant distribution functions to a one- or two-species model, as required. (A) rEch (1–49) M28L D26A (black triangles). (B) rEch (1–49) M28L R24A (gray diamonds) (C) rEch (1–49) M28L D27W (black hexagons).

Figure 3.

Concentration-dependent perturbation of αIIbβ3 solution conformation as measured by the fractional change in sedimentation coefficient, S/S₀, versus the molar excess of echistatin variant, rEch:αIIbβ3. The dotted line denotes S/S₀ = 1.0. Symbols as follows: rEch (1–49) M28L D26A (black triangles), rEch (1–49) M28L D26E (open triangles), rEch (1–49) M28L R24A (gray diamonds), rEch (1–49) M28L R24K (black diamonds), rEch (1–49) M28L (open circles), rEch (1–49) M28L D27W (black hexagons), rEch (1–49) M28L D27F (gray hexagons). The black and gray lines were obtained by fitting the data with full-length echistatin and the R24A mutant, respectively, to a hyperbolic inhibition model (Hantgan et al. 1999).

Figure 4.

Effects of recombinant 30-kDa γ-module (γ 148–411) on the distribution of αIIbβ3 sedimenting species, g(S) versus S, as determined with time-derivative software (DCDT+; Philo1997). Each panel depicts the distributions obtained with ligand-free integrin (open circles; left axis), αIIbβ3 + γ 148–411 (black squares; right axis), and γ 148–411 alone (gray triangles; right axis). Continuous lines were obtained by fitting the resultant distribution functions to a one- or two-species model, as required. (A) Resting αIIbβ3 + γ 148–411; (B) primed αIIbβ3^* + γ 148–411.

Figure 5.

Integrin:echistatin interactions measured in solution by changes in fluorescence anisotropy with Oregon Green-labeled echistatin. (A) Data obtained in a receptor:ligand titration are expressed as the fractional change in anisotropy, A − A₀/A₀ versus the mole ratio of αIIbβ3 to fluorophore-labeled rEch (1–49) M28L D27W (black hexagons). The black line was obtained by fitting the data to a single-site saturable binding model. (B) Size exclusion chromatography profile (G-75 Sephadex) of a mixture of αIIbβ3 and excess Oregon Green-rEch (1–49) M28L D27W with detection by fluorescence intensity (open triangles, continuous line) and fluorescence anisotropy (black hexagons, dashed line). Complex formation is detected in the early-eluting fraction by the increase in anisotropy compared to the later eluting free ligand.

Figure 6.

Integrin:γ-module interactions measured in solution by changes in fluorescence anisotropy with Oregon Green-labeled ligands. (A) Data obtained in a receptor:ligand titration with primed αIIbβ3* are expressed as fractional change in anisotropy, A − A₀/A₀ versus the mole ratio of αIIbβ3 to fluorophore-labeled γ148–411 in buffer containing 1 mM Ca⁺⁺/1 mM Mg⁺⁺ (black circles) or 1 mM Ca⁺⁺ (gray squares) and γ148–392 in the presence of 1 mM Ca⁺⁺/1 mM Mg⁺⁺ (dark gray triangles). The black line was obtained by fitting the data to a single-site saturable binding model. Data obtained with these same γ-module ligands and resting αIIbβ3 are denoted by open symbols and the dashed line. (B) Size exclusion chromatography profile (G-75 Sephadex) of a mixture of primed αIIbβ3* and excess Oregon Green-γ148–411 with detection by fluorescence intensity (open diamonds, continuous line) and fluorescence anisotropy (black circles, dashed line). Complex formation is detected in the early-eluting fraction by the increase in anisotropy compared to the later eluting free ligand.

Figure 7.

Modeling eptifibatide/echistatin:αIIbβ3 interactions. (A) This figure is based on the crystallographically determined structure of the eptifibatide:αIIbβ3 ectodomain complex, 1TY6.pdb (Xiao et al. 2004). The αIIb subunit is shown in light blue, the β3 subunit in red, and eptifibatide in yellow with its functional groups colored as follows: nitrogen, blue; oxygen, red; sulfur, orange. Selected αIIb residues that interact with eptifibatide are highlighted including Asp 224, located 3.2 Å distant from the ligand's positively charged guanidinium moiety, and Phe 160, Tyr 190, and Phe 231, which pack around the homoarginine's aliphatic segment. Likewise, eptifibatide aspartate, which interacts with the β3 subunit's MIDAS Mg⁺⁺ cation (2.0 Å away), is emphasized. (B) This figure compares the structures of eptifibatide (as in panel A) and echistatin, based on the NMR determined structure, 1R03.pdb (Monleon et al. 2005). The distances between key electrostatic residues on each ligand, eptifbatide's homoarginine (C_ζ) and its aspartate (C_δ), echistatin's Arg 24 (C_ζ) and Asp 26 (C_δ), are indicated by dashed lines. Echistatin's polypeptide backbone is shown as a green tube with selected residues highlighted.