Identification of a Region within the Cytoplasmic Domain of Sema4D That Binds Calmodulin and Regulates Shedding of the Sema4D Exodomain in Platelets,

Abstract 3249

We have previously shown that the 150 kDa semaphorin family member, Sema4D (or CD100), is expressed on the surface of human and mouse platelets, where it is able to selectively reinforce collagen-initiated platelet activation by engaging receptors in trans on adjoining platelets in a contact-dependent manner. Key to this effect is the Sema4D extracellular domain, which in addition to being a ligand for Sema4D receptors, is a substrate for the metalloprotease, ADAM17 in platelets. Our previous studies suggest that ADAM17 cleaves Sema4D and other platelet surface proteins close to the platelet plasma membrane gradually producing, in the case of Sema4D, a single large (≈120 kDa) exodomain fragment and a smaller (≈28 kDa) fragment that includes the transmembrane domain and the cytoplasmic domain and remains associated with the platelet (Zhu, et al., PNAS 2007). Exodomain shedding in platelets can be triggered by the phorbol ester, PMA, and by physiologic agonists such as thrombin that raise the cytoplasmic Ca⁺⁺ concentration, but the mechanisms that regulate the shedding of Sema4D have not been defined. Here we have studied the potential role of an interaction between calmodulin and the Sema4D cytoplasmic domain. Using a public resource (http://calcium.uhnres.utoronto.ca/ctdb/ctdb/home.html), we identified a potential calmodulin-binding sequence (GYLPRQCLKFRSALLIGKKKPKS-COOH, Gly⁷⁵⁸–Ser⁷⁸⁰) within the membrane-proximal region of the Sema4D cytoplasmic domain. To test whether this region binds calmodulin, a 23 amino acid peptide corresponding to the predicted Sema4D calmodulin binding sequence (SCBP) was synthesized, as was a scrambled control peptide (RLIKACRQPKPKYKLLGFGSSKL or scrambled SCBP), which is not predicted to bind calmodulin. The results show that SCBP, but not scrambled SCBP, was able to bind to calmodulin-agarose and retrieve calmodulin from platelet lysates. As constitutive association of calmodulin with glycoprotein (GP) Ib has been shown prevent ADAM17-dependent GPIb alpha shedding in platelets, we incubated human platelets with the calmodulin inhibitor, W7. The inhibitor induced gradual Sema4D shedding that was detectable after 5 min and reached a maximum at 60 min, kinetics that are similar to those we have observed with platelet agonists. However, in contrast to platelet agonists, W7-induced Sema4D shedding generated a smaller retained fragment (≈24 kDa Vs. 28 kDa) suggesting that there is either a second or different site of cleavage. Despite their polybasic sequences, flow cytometry and confocal microscopy showed that FITC-conjugated SCBP and scrambled-SCBP are able to cross the plasma membrane. Addition of SCBP, but not scrambled-SCBP, to platelet caused cleavage of Sema4D, producing the same 28 kDa fragment observed with thrombin and PMA. In all cases cleavage of Sema4D was blocked by the metalloprotease inhibitor, TAPI-2. Combined with our earlier observations, these results suggest that 1) Sema4D is a calmodulin binding protein with a site of interaction in the membrane-proximal cytoplasmic domain and a site of cleavage by ADAM17 in the membrane-proximal exodomain, 2) the detachment of calmodulin from Sema4D may be the trigger for Sema4D cleavage in response to platelet agonists, and 3) in contrast to W7, decoying calmodulin from binding sites on Sema4D and other metalloprotease substrates on the platelet surface, as we have done here with a Sema4D cytoplasmic domain peptide, may trigger the same events seen in activated platelets and provide a tool to understand the underlying mechanisms.