The structural basis of the EPCR-APC complex induced biased PAR1 signaling

Abstract

Activated Protein C (APC) is an effector enzyme of the natural anticoagulant pathway. In addition to its anticoagulant function, endothelial protein C receptor (EPCR)-bound APC induces biased protease-activated receptor type 1 (PAR1)-mediated signaling. Despite intensive investigation, the mechanism of biased signaling is not completely clear. To gain new insights into APC-induced PAR1-biased signaling we reviewed the published data and created three-dimensional models of the proteins and their complexes involved in the early stages of PAR1 signaling. A comparative study of models related to canonical and biased signaling demonstrated that interactions between APC, EPCR, PAR1, and Caveolin-1 (Cav1) can provide plausible explanations for the differences between the two types of PAR1 signaling. The model suggests that the interaction of the PAR1 peptide 22-ARTRARRPESK-32 with 162-helix of APC positions the PAR1 N-terminus for the preferential cleavage at R46. By contrast, the hirudin-like sequence of PAR1 is involved in the positioning of the N-terminus of PAR1 for cleavage at R41 by thrombin in canonical signaling. The model and molecular dynamics (MD) simulations of the tethered ligand (TL) interaction with APC suggest that the TL facilitates direct interaction of the EPCR transmembrane (TM) domain with the PAR1 TM helices 6 and 7 by transient binding to the light chain of APC and keeping EPCR-APC in close proximity to PAR1. The biased signaling paradigm considers the ligand-induced conformational changes in PAR1 as solely being responsible for the biased signaling. Our models suggest that Cav1, EPCR, and PAR1 interactions can provide a selective advantage to biased signaling over canonical signaling. First, the complex comprised of caveolin-1 oligomer-EPCR-APC-PAR1 positions EPCR-APC and PAR1 at a distance favorable for PAR1 activation. Second, the Cav1 presence favors selectivity for the PAR1 bound β-arrestin-2, not the PAR1-bound G protein alpha (Gα) subunit. The potential reason for β-arrestin-2 selectivity includes Gα binding to the Cav1 and its immobilization resulting in the inability of PAR1-bound Gα to periodically interact with the plasma membrane required for its function. MD simulations of the PAR1-EPCR-β-arrestin-2 complex demonstrated that one of the mechanisms of the APC-induced PAR1-biased signaling is the interaction of the EPCR TM domain with the PAR1-bound β-arrestin-2, leading to the stabilization of the PAR1-β-arrestin-2 complex and activation of β-arrestin-2. Thus, models suggest that Cav1 and EPCR-APC mediated interactions provide a selective advantage for the β-arrestin-2 dependent biased signaling, not the G proteins mediated canonical signaling by the PAR1 receptor.