Abstract
Integrin αIIbβ3 exists in a low-affinity state and requires activation for high-affinity binding with soluble ligands. αIIbβ3 activation is linked to rearrangements of the β3 I-like domain structure and the largest movement in the β3 I-like domain following ligand binding occurs in the β6-α7 loop. Although this loop does not comprise a ligand binding site, a recent mutational study has shown that introduction of disulfide bonds into the β6/α7 region to lock the β3 I-like domain in the open or closed conformation renders αIIbβ3 constitutively active or inactive, suggesting the regulatory role of this region in integrin activation. However, it remains to be determined which residues of the β6-α7 loop in the β I-like domain are critical for integrin activation. We therefore conducted alanine-scanning mutagenesis of the β6-α7 loop residues in the β3 I-like domain and tested for ligand binding to mutant β3 integrins. The β6-α7 loop is composed of residues from S334 to N339, and the region between L333 and V340 was targeted for mutagenesis. The mutant β3 cDNA was transfected into CHO cells together with wild-type αIIb cDNA. The expression of αIIbβ3 mutants on the cell surface was 56–116% that of the wild-type αIIbβ3. Then binding of an activation-dependent antibody. PAC1, to the αIIbβ3 mutants was examined in the presence or absence of the αIIbβ3-activating antibody PT25-2. As expected, wild-type αIIbβ3 showed PAC1 binding only after activation of αIIbβ3 by PT25-2. The L333A, S334A, M335A and V340A mutations had no effect on PAC1 binding. In contrast, the S337A and N339A mutations induced significant PAC1 binding in the absence of PT25-2, indicating a constitutively active state. Although the D336A and S338A mutations retained αIIbβ3 in an inactive state in the absence of PT25-2, they induced 3- and 4.5-fold as much PAC1 binding as wild-type αIIbβ3 in response to PT25-2, respectively. The S337A and N339A mutations, which rendered αIIbβ3 constitutively active, induced further PAC1 binding in the presence of PT25-2 (2.7~5.1-fold as much PAC1 binding as wild-type αIIbβ3). When soluble fibrinogen was used instead of PAC1, similar results were obtained. None of the mutations tested had any effect on PT25-2 binding. We next quantified adhesion of CHO cells stably expressing mutant or wild-type αIIbβ3 to immobilized fibrinogen. The S337A and N339A mutations enhanced cell adhesion to fibrinogen and the extent of adhesion of the mutants was comparable to that of wild-type αIIbβ3 activated by Mn2+. To determine whether the mutations of the β3 subunit responsible for αIIbβ3 activation also induce constitutive activation of αVβ3, CHO cells expressing αVβ3 mutants were tested for soluble fibrinogen binding. The surface expression of mutant αVβ3 was comparable to that of wild-type αVβ3. The αVβ3S337A and N339A mutants bound significantly more fibrinogen than wild-type αVβ3 without any stimulation, indicating that the S337A and N339A mutations also rendered another β3 integrin, αVβ3, constitutively active. These results suggest that the S337 and N339 residues in the β3 I-like domain are essential for constraining β3 integrins in a default low-affinity state and that structural rearrangement in the 336DSSN sequence of the β6-α7 loop alters the affinity state of αIIbβ3. Since the DSSN sequence is highly conserved among integrin β subunits, this motif may be a common regulatory component of integrin activation.
Computational redesign of the Escherichia coli ribose-binding protein ligand binding pocket for 1,3-cyclohexanediol and cyclohexanol