EPCR occupancy regulates β-arrestin-2 biased PAR1 signaling by APC and thrombin. Confluent EA.hy926 endothelial cells were transfected with the control siRNA (A-B) or siRNA specific for either β-arrestin-1 (C-D) or β-arrestin-2 (E-F) before treating cells with the PAR1 agonists (2 nM thrombin, 20 nM APC, or 2 nM thrombin + 50 nM PC-S195A for 3 hours) and monitoring cell permeability in response to either thrombin (10 nM for 10 minutes in panels A, C, E) or LPS (10 ng/mL for 4 hours in panels B, D, F) as described in “Materials and methods.” (G-H) The same as previous panels except that cell permeability in response to thrombin (10 nM) was carried in the presence of function-blocking anti-EPCR or anti-PAR1 antibodies. (I-J) sodium dodecyl sulfate–polyacrylamide gel electrophoresis (SDS-PAGE), immunoprecipitation, and immunoblotting of lysates derived from cells treated with different agonists. Total cellular proteins, derived from nontreated cells or cells treated with thrombin (2 nM), APC (20 nM), and thrombin (2 nM) + PC-S195A (50 nM), were immunoprecipitated with anti-β-arrestin-1 (I) or anti-β-arrestin-2 (J) and separated on SDS-PAGE followed by immunoblotting with anti-PAR1, anti-EPCR, anti-caveolin-1, or different pairs of the same anti-β-arrestin antibodies. (K) The same as panel A, except that permeability in response to thrombin (10 nM) was monitored in endothelial cells treated with MβCD (10 mM for 1 hour) before incubation with the agonists. (L) The same as previous panel, except that permeability in response to thrombin (10 nM) was monitored in endothelial cells transfected with specific siRNA for S1P₁ before incubation with the agonists. All results are shown as means ± standard deviation of 3 different experiments. ***P < .001. IB, immunoblotting; IP, immunoprecipitation; MβCD, methyl-β-cyclodextrin; PC195, PC-S195A; Th, thrombin; UT, untreated control.