Structural and Functional Evidence on the Role of Sulfated Tyrosine Residues In Glycoprotein Ib Binding to α-Thrombin Exosites

Abstract 326

In spite of two known crystal structures, the mechanisms supporting the interaction between the amino terminal domain of glycoprotein (GP) Ibα (GPIb-N) and α-thrombin (FIIa) are still debated, and a controversial issue concerns the involvement of FIIa exosites I and II in binding. Competition for exosite I could influence processes important for hemostasis and thrombosis. Both known crystal structures show two independent contact interfaces between GPIb-N and bound FIIa in a conformation that involves each exosite interacting with a different GPIb-N molecule. This notwithstanding, a majority of investigators in the field has concluded that exosite II is solely required for FIIa binding to GPIb-N, suggesting that the interface with exosite I is the consequence of crystal packing. The goal of this work was to probe experimentally the role of FIIa exosites in GPIb-N binding. Human GPIb-N contains three Tyr residues (at positions 276, 278 and 279) that can undergo post-translational sulfation (sulfated Tyr = Tys), although this was not the case for Tyr278 in the known crystal structures. To address this discrepancy, we expressed GPIb-N in Drosophila cells - which endogenously contain a single tyrosylprotein sulfotransferase (TPST) gene - co-transfected with human TPST-2, and showed that we could obtain different GPIb-N species with 0 to 3 sulfate moles/protein moles. Using these different GPIb-N forms immobilized onto a surface plasmon resonance (SPR) chip, we determined that the kD of human FIIa binding decreased from 1290 to 89 nM going from 0 to 3 sulfate moles/protein moles. We crystallized the fully sulfated GPIb-N complexed with FIIa and found that Tys278 established contacts not previously seen with exosite II (residues Arg35 and Lys236), thus explaining the contribution of full sulfation to maximal binding efficiency. To establish the effect of TPST-2 on the process of sulfation, we compared the affinity of FIIa binding to distinct wild type GPIb-N species of known sulfate content with that to GPIb-N mutants containing distinct single, double or triple Tyr “Phe substitutions (Phe differs from Tyr for the lack of an OH group and cannot be sulfated) in which the identity of Tys residues could be established. We found that TPST-2 favors Tyr sulfation in the order 276–278-279, which is more efficient for a complete process than the order 276–279 predominant in the absence of TPST-2. We then used different Tyr "Phe (Y to F) mutants to evaluate the effects of the substitution preventing sulfation on FIIa binding to GPIb-N in solution or immobilized onto a SPR chip. We fount that the Y276F mutant had no capacity to form a soluble complex with FIIa, while Y279F could complex about as much FIIa as fully sulfated wild type GPIb-N (82 vs 99% FIIa incorporation). Of note, both Y276F and Y279F mutants had a complete to nearly complete loss of FIIa binding activity in the SPR system. In the crystal structure, the sulfate group on Tys279 establishes three close contacts (3.1, 3.3, 2.8 □) with Trp148 in a FIIa loop neighboring exosite I. On the other hand, Tys276 has closer contacts than Tys279 with residues in exosite II, suggesting that the latter may be sufficient for FIIa binding when GPIb-N is in solution but not immobilized onto a surface. To confirm this hypothesis, we used specific aptamer inhibitors of FIIa exosite II (HD22) or I (HD1) and found that the latter, similar to the Y279F substitution, indeed had no effect on FIIa-GPIb-N soluble complex formation (thus ruling out possible allosteric effects on exosite II influencing GPIb-N binding) but completely prevented FIIa binding to immobilized GPIb-N. Of note, as shown by crystallographic evidence, bound HD1 alters the orientation of Trp148 in manner that would oppose the interaction with Tys279 in GPIb-N, providing a structural explanation for the similar functional effects of the mutation and the inhibitor. Finally, we expressed transgenically wild type or Y279F mutant human GPIbα to replace the homologous chain in the GPIb-IX-V complex of murine platelets and showed that the mutation almost completely impairs FIIa binding to platelets, which is also prevented by inhibition of exosite I with HD1. These results provide functional evidence and a structural explanation for a key role of exosite I, concurrently with exosite II, for FIIa binding to GPIbα. Additional studies are now demonstrating that interfering with this interaction modifies responses to vascular injury in vivo.