The integrins αIIbβ3 and αvβ3 on circulating platelets are maintained in inactive states by an intra-molecular clasp involving the cytoplasmic, transmembrane, and extracellular stalk domains of their α and β subunits. Following platelet stimulation, αIIbβ3 and αvβ3 rapidly undergo global rearrangements in which their clasps are disrupted and their headpieces open to expose a ligand binding site. Protein-protein interfaces, such as those of the αIIbβ3 and αvβ3 clasps, are usually large complementary surfaces in which specific amino acids, termed energetic "hot spots", contribute disproportionately to the binding free energy. Previously, we scanned the β3 EGF-3, EGF-4, and βTD domains computationally with alanine residues and identified a limited number of interface residues that stabilized the extracellular clasps of αIIbβ3 and αvβ3. Here, we have used computational alanine scanning to identify corresponding "hot spots" in the Calf-1 and Calf-2 domains of the αIIb and αv components of the αIIbβ3 and αvβ3 clasps. The identified alanine substitutions were predicted to destabilize the respective stalk heterodimers. We identified 10 αIIb substitutions and 11 αv substitutions that would destabilize the stalk interfaces by a change in binding free energy (ΔΔG) of >0.3 kcal/mol. To confirm that a ΔΔG as small as 0.3 kcal/mol is sufficient to destabilize the β3 integrin stalk interfaces, we introduced 8 of the 10 αIIb substitutions into full-length αIIb, co-expressed the mutated αIIb with wild type β3 in CHO cells, and measured spontaneous binding of Alexa Fluor 488-labeled fibrinogen to the cells by flow cytometry. αIIbR751Aβ3 did not express, implying that R751 is important for correct αIIb folding and/or αIIbβ3 assembly. Each of the other mutations expressed and caused constitutive αIIbβ3 activation. Importantly, αIIbβ3 containing alanine substitutions whose effect on ΔΔG was <0.3 kcal/mol was inactive. Nonetheless, while the alanine scanning algorithm correctly identified destabilizing alanine substitutions, it did not correctly rank their functional effects. Hot spots could occur because they bury specific side chains in the protein-protein interface. While we found a modest correlation between ΔΔG and the buried surface area of specific αIIb side chains (R2=0.7), there was no correlation with αv (R2=0.1), a result consistent with an analysis of a compiled database of 2325 similar alanine mutants (Bogan and Thorn, J Mol Biol 280:1-9,1998). This suggests that other mechanisms, such as inter-subunit hydrogen bonds, are involved. Previously, in studies of the growth hormone-growth hormone receptor complex, Clackson and Wells (Science 267:383-86,1995) found that functionally important residues in the interface of the complex were in direct contact. To determine if this is the case for αIIbβ3 and αvβ3, we mapped the mutation-sensitive positions onto the αIIbβ3 and αvβ3 crystal structures to localize the αIIb, αv, and β3 hot spots in models of the αIIbβ3 and αvβ3 stalks. The energetic hot spots we identified in αIIb and αv were located in a strip along one face of the Calf-1 and Calf-2 domains. Similarly, the β3 hot spots were located in a strip along on one face of the EGF-3, EGF-4, and βTD domains. However, the hot spots were not contiguous and there was only partial overlap of the β3 side chains that interacted with αIIb versus αv. On the other hand, in models of the assembled stalk heterodimers, there was remarkable complementarity of the computationally identified hot spots in the αIIb and αv stalks with those in β3, indicating that only a limited number of specific inter-subunit interactions actually make significant contributions to stalk assembly. These results confirm the utility of computational alanine scanning for identifying and localizing functionally important interaction sites in αIIbβ3 and αvβ3. They also demonstrate that although the same surface of β3 interacts with the αIIb and αv stalks, unique, as well as overlapping, side chain interactions are responsible for maintaining the inactive state of the respective integrins. They also suggest that stabilizing the extracellular stalk heterodimers may be a way to attenuate β3 integrin activation.
No relevant conflicts of interest to declare.
Asterisk with author names denotes non-ASH members.