Abstract
Glycoprotein (GP) Ib in the GPIb-IX-V receptor complex is the most abundant binding site for thrombin on the platelet surface. Virtually the entire thrombin binding capacity of GPIb has been shown to reside on the N-terminal region of the GPIba subunit of GPIb. Recently, Celikel et al and Dumas et al independently solved the structure of the thrombin-GPIba complex. Although comparable N-terminal fragments, comprising residues 1–290 of GPIba, were used for crystallization in both studies, significant differences existed between the two structures. Thus, it is still unclear how GPIb interacts with thrombin. In this study we have examined the interaction of GPIba with thrombin in solution phase. Human a-thrombin was labeled active site-specifically with either dansyl (D) dye via a Glu-Gly-Arg (EGR) linker to yield DEGR-thrombin or with a fluorescein or 5-((((2-iodoacetyl)amino)ethyl)amino)naphthalene-1-sulfonic acid dye (IAEDANS) via a Phe-Pro-Arg tether to yield Fluorescein-thrombin and AEDANS-thrombin, respectively. When DEGR-thrombin (initially 100 nM) was titrated with human glycocalicin, the N-terminal fragment of GPIbα compring ~400 residues, the steady state anisotropy of DEGR-thrombin decreased by ~ 22% before reaching a plateau at ~ 100 nM protein suggesting an interaction between Glycocalicin and DEGR-thrombin. A ~ 10% increase in anisotropy of the dansyl moiety was observed when a recombinant wild-type fragment of GPIba (residues 1–290) containing the three sulfated-tyrosines at positions 276, 277 and 279 was titrated into DEGR-thrombin. However, this change in anisotropy was not observed when either a mutant with tyrosine 276 mutated to phenylalanine (Y276F) or a Y279F mutant (named analogously) were titrated into DEGR-thrombin. To examine if dimerization of GPIba was important for thrombin interaction, a construct was made such that residues 1–288 of GPIba were covalently linked through a C-terminal extended sequence containing 4 Cys residues, and expressed as dimer (C65 +). When C65+ was titrated into DEGR-thrombin, the anisotropy of the dansyl probe increased by ~ 29% before reaching a plateau at 130 nM C65+, suggesting that thrombin can bind dimeric GPIba. To elucidate the stoicheometry of the thrombin-GPIba complex, resonance energy transfer (RET) experiments were performed between AEDANS donor-labeled thrombin and Fluorescein acceptor labeled thrombin. The AEDANS-thrombin and Fl-thrombin were mixed in equimolar ratios and then titrated with increasing amounts of GPIba. No change in donor intensity was observed suggestive of the absence of a AEDANS-thrombin- GPIba- Fl-thrombin complex. In conclusion, our data suggests that the GPIba interaction with thrombin can be observed in solution phase using steady state fluorescence by appropriately active site-labeled thrombin. Tyrosine sulfation at positions 276 and 279 are critical for this interaction. This observation is in complete agreement with both crystal structures where the contact site with exosite II of thrombin seems to be mediated by residues 275–279 of GPIba. However, using RET, we could not find a thrombin-GPIba-thrombin complex in solution under the present experimental conditions.
Author notes
Corresponding author