Biased agonism of protease-activated receptor 1 by activated protein C caused by noncanonical cleavage at Arg46

Abstract

Activated protein C (APC) exerts endothelial cytoprotective actions that require protease-activated receptor 1 (PAR1), whereas thrombin acting via PAR1 causes endothelial disruptive, proinflammatory actions. APC's activities, but not thrombin's, require PAR1 located in caveolae. PAR1 is a biased 7-transmembrane receptor because G proteins mediate thrombin's signaling, whereas β-arrestin 2 mediates APC's signaling. Here we elucidate novel mechanisms for APC's initiation of signaling. Biochemical studies of APC's protease specificity showed that APC cleaved PAR1 sequences at both Arg41 and Arg46. That PAR1 cleavage at Arg46 can occur on cells was supported by APC's cleavage of N-terminal-SEAP-tagged R41Q-PAR1 but not R41Q/R46Q-PAR1 mutants transfected into cells and by anti-PAR1 epitope mapping of APC-treated endothelial cells. A synthetic peptide composing PAR1 residues 47-66, TR47, stimulated protective signaling in endothelial cells as reflected in Akt and glycogen synthase kinase 3β phosphorylation, Ras-related C3 botulinum toxin substrate 1 activation, and barrier stabilization effects. In mice, the TR47 peptide reduced VEGF-induced vascular leakage. These in vitro and in vivo data imply that the novel PAR1 N-terminus beginning at residue Asn47, which is generated by APC cleavage at Arg46, mediates APC's cytoprotective signaling and that this unique APC-generated N-terminal peptide tail is a novel biased agonist for PAR1.

Figure 1

APC cleaves a synthetic PAR1 N-terminal peptide at Arg41 and Arg46. Cleavage of a PAR1 N-terminal tail peptide (TR33-62) by APC (500nM) or thrombin (10nM) in the presence of CaCl₂ (4mM) and MgCl₂ (0.6mM) and resolution of cleavage fragments by HPLC using C18 reverse phase chromatography. (A) Chromatograms of APC-mediated TR33-62 cleavage over time. (B) Time course of TR33-62 cleavage by APC (□) or thrombin (■). (C) Generation of the N-terminal cleavage fragment derived from cleavage at Arg41 (TR33-41) by APC (▿) or thrombin (▾). (D) Generation of the C-terminal cleavage fragment derived from cleavage at Arg41 (TR42-62) by APC (Δ) or thrombin (▴). (E) Generation of the N-terminal TR42-46 (♢, ♦) and C-terminal TR47-62 fragments (○, ●) derived from cleavage at Arg46 by APC (open symbols) or by thrombin (closed symbols). (F) Time courses for cleavage of peptide TR42-66 (□, ■) and generation of TR42-46 (♢, ♦) and TR47-66 (○, ●) cleavage fragments by APC (open symbols) or thrombin (closed symbols). Representative chromatograms (A) and (B-F) data points represent the mean ± SD (N ≥ 3).

Figure 2

Noncanonical cleavage of PAR1 at Arg46 by APC on cells. Cleavage of wt-SEAP-PAR1 and SEAP-PAR1 cleavage site mutants by (A) APC and (B) thrombin in the presence of wt-EPCR and in the absence of EPCR (C). (□, ■) represent wt-SEAP-PAR1; (Δ, ▴), R41Q-SEAP-PAR1; (○, ●), R46Q-SEAP-PAR1; and (♢, ♦), R41Q/R46Q-SEAP-PAR1. (C) Closed symbols indicate thrombin; and open symbols, APC. (D) Mapping of the canonical and noncanonical cleavage site sensitive epitopes of anti-PAR1 antibodies SPAN12, ATAP2, and WEDE15 using synthetic peptides coated onto microtiter plate wells. Peptides represent the N-terminal (TR24-41) or C-terminal fragment (TR42-66) of PAR1 after cleavage at Arg41 or the N-terminal (TR24-47) or C-terminal fragment (TR47-66) of PAR1 after cleavage at Arg46 or the entire cleavage site region of the PAR1 N-terminal tail (TR33-62). (E) OCW of anti-PAR1 antibodies WEDE15 and ATAP2 on EA.hy.926 endothelial cells after incubation with control buffer (NONE), APC (100nM), or thrombin (0.25nM) for 3 hours. Cell-bound anti-PAR1 antibodies were detected with biotinylated goat anti–mouse secondary antibodies and IRDye 800CW streptavidin in the 800-nm channel of the Odyssey Imager (green). Draq5 was used for cell number normalization and detected in the 700-nm channel (red). Overlay of the Draq5 700 channel and anti-PAR1 800-nm channel is indicated in yellow (merge). The presence and loss of PAR1 epitopes on endothelial cell surfaces were determined using OCW quantification of cleavage site sensitive and insensitive PAR1 antibodies on EA.hy926 endothelial cells versus time for incubation of cells with (F) APC (100nM) or (G) thrombin (0.25nM). The cell surface was probed for PAR-1 with (○) WEDE15 (detecting all PAR1 regardless of cleavage at either Arg41 or Arg46), (□) ATAP2 (detecting uncleaved and PAR1 cleaved only at Arg41), or (♢) SPAN12 (detecting only uncleaved PAR1). (D-E) Representative experiment in triplicates. (A-C,F-G) Data points represent the mean ± SEM (N ≥ 3). (F-G) *P < .001, compared with WEDE15. ns indicates not significant.

Figure 3

The novel tethered ligand TR47 peptide induces PAR1-dependent signaling in endothelial cells. Phosphorylation of pSer473-Akt and pSer9-GSK3β by TR47 peptide was monitored in endothelial cells. (A) Time course of Akt phosphorylation at Ser473 by TR47 (50μM). (B) Time course of GSK3β phosphorylation at Ser9 by TR47 (50μM). (C) Phosphorylation of Akt at Ser473 by TR47 (50μM) or a control scrTR47 peptide (50μM) over time. (D) Phosphorylation of GSK3β at Ser9 by TR47 (50μM) and a control scrTR47 peptide (50μM) over time. (E) TR47 peptide dose dependence for phosphorylation of Akt at Ser473 at 60 minutes. (F) Time course for phosphorylation of Akt at Ser473 by TR47 (50μM) in the presence of the PAR1 inhibitor, SCH79797, or vehicle control (DMSO). (G) Time course for phosphorylation of GSK3β at Ser9 by TR47 (50μM) in the presence of the PAR1 inhibitor, SCH79797, or vehicle control (DMSO). (A-B) Representative experiments. (C-G) Data points represent the mean ± SEM (N ≥ 3).

Figure 4

The TR47 peptide and TRAP are PAR1 biased agonists. Differential phosphorylation was determined for pThr202/Tyr204-ERK1/2, pSer473-Akt, and pSer9-GSK3β that was induced by TRAP, TR47 peptide, control scrTR47 peptide, APC, and thrombin. (A) Time course of ERK1/2 phosphorylation at Thr202/Tyr204 by APC (70nM) and TR47 (50μM). (B) Phosphorylation of ERK1/2 at Thr202/Tyr204 by TRAP (50μM), TR47 (50μM), scrTR47 (50μM), and APC (70nM) over time. (C) Time course of Akt phosphorylation at Ser473 by APC (70nM) and thrombin (54nM). (D) Phosphorylation of Akt at Ser473 by APC (70nM), thrombin (54nM), and TRAP (50μM) over time. (E) Time course of GSK3β phosphorylation at Ser9 by APC (70nM) and TRAP (50μM). (F) Phosphorylation of GSK3β at Ser9 by APC (70nM) and TRAP (50μM) over time. (A,C,E) Representative experiments. (B,D,F) Data points represent the mean ± SEM (N ≥ 3).

Figure 5

The TR47 peptide inhibits endothelial permeability in vitro and in vivo vascular leakage in mice. (A) Activation of Rac1 by TR47 peptide, control peptide, and APC was determined and is shown as pull-down of active Rac1 (GTP-Rac1) with PAK1 (top left), total Rac1 (top right panel), and quantification of Rac1 activation at 180 minutes (bottom panel) using APC (70nM), TR47 (50μM), and scrTR47 (50μM). *P < .05. ns indicates not significant. (B) In vitro protection against thrombin-induced endothelial permeability by APC (25nM), TR47 (50μM), and scrTR47 (50μM). (C) Experimental time line for the in vivo VEGF-induced vascular leakage model. (D) Photographs are seen for Evans blue extravasation in the skin of TR47-treated mice (right) or of control PBS-treated mice (left) injected subcutaneously with VEGF (3 left arrows) or with BSA control (2 right arrows). (E) Heat map display of Evans blue extravasation quantified by the Odyssey near-infrared imager at 700 nm. TR47-treated mice (right) or control PBS-treated (left) mice were injected subcutaneously with VEGF (3 left arrows) or with BSA control (2 right arrows). (F) In vivo VEGF-induced vascular leakage in mice is shown for mice treated with PBS control, TR47 peptide (125 μg), or control scrTR47 peptide (125 μg). Data are also shown for BSA-treated control mice that received PBS, TR47, or scrTR47. (A top panels, D-E) Representative experiments. (A bottom panel, B,F) Data points represent the mean ± SEM (N ≥ 3). *P < .05.

Figure 6

Biased agonism of PAR1 resulting from canonical cleavage at Arg41 by thrombin or by noncanonical cleavage at Arg46 by APC. PAR1 is a biased receptor with biased agonists. Four different PAR1 agonists are shown along the top bar that can vary in their ability and efficacy to activate different signaling pathways in cells, thus depicting a signaling phenomenon that has been labeled “functional selectivity” or “biased agonism.” The receptor's bias is directly based on the spectrum of signaling that is mediated either via 1 or more G proteins or via β-arrestin-2, whereas the agonist bias is directly related to the sites of cleavage in the extracellular N-terminus (ie, resulting from cleavage either at Arg41 or at Arg46). The TRAP peptide(s) represent the N-terminal sequence of PAR1 that exists after cleavage at Arg41, whereas TR47 represents the N-terminal sequence of PAR1 that exists after cleavage at Arg46, as described in this report. Each agonist can have distinctly different properties because each one can stabilize a different subset of the dynamic conformational ensembles of PAR1. The conformer subsets that are stabilized by TRAP and thrombin promote signaling via different G proteins, whereas conformer subsets stabilized by APC's action, and presumably TR47, promote signaling via β-arrestins, especially β-arrestin-2, and dishevelled-2. However, one must note that TRAP is similar but not functionally equivalent to thrombin; and it is probable, but not yet shown, that the TR47 peptide does not have all the same activities and efficacy as APC. Not depicted are the effects of MMP1 cleavage at Asp39 in PAR1 or potential allosteric PAR1 modulatory factors, including binding of thrombin to the hirudin-like sequence of PAR1, localization of PAR1 in caveolae with caveolin-1, association with EPCR containing bound protein C or APC, dimerization with other PARs, or small molecular allosteric modulators.^,