Molecular Basis for G-protein-Coupled Receptor (GPCR) Activation and Biased Signalling at the Platelet Thrombin Receptor Proteinase Activated Receptor-4 (PAR4)

Proteinase Activated Receptor-4 (PAR4) is a member of the proteolytically-activated PAR family of G-Protein-coupled Receptors (GPCRs). PARs are activated following proteolytic cleavage of the receptor N-terminus by enzymes such as thrombin, trypsin, and cathepsin-G to reveal the receptor-activating motif termed the tethered ligand. The tethered ligand binds intramolecularly to the receptor and triggers receptor signalling and cellular responses. In spite of this unusual mechanism of activation, PARs are fundamentally peptide receptors and can also be activated by exogenous application of short synthetic peptides derived from the tethered ligand sequence. In order to gain a better understanding of the molecular basis for PAR4-dependent signalling, we examined signalling responses to a library of peptides derived from the canonical PAR4 activating peptide (PAR4-AP), AYPGKF-NH2. We examined peptide residues involved in activation of the Gαq/11-coupled calcium signalling pathway, β-arrestin recruitment, and mitogen-activated protein kinase pathway activation. The peptide N-methyl-alanine-YPGKF-NH2 was identified as a compound that is a poor activator of PAR4-dependent calcium signalling but was fully competent in recruiting β-arrestin-1 and -2. In order to gain a better understanding of the ligand-binding pocket, we used in silico docking to identify key residues involved in PAR4 interaction with AYPGKF-NH2. The predicted interactions were verified by site-directed mutagenesis and analysis of calcium signalling and β-arrestin-1/-2 recruitment following proteolytic activation (with thrombin) or activation with the synthetic agonist peptide (AYPGKF-NH2). We determined that a key extracellular loop-2 aspartic acid residue (Asp230) is critical for signalling following both proteolytic and peptide activation of PAR4. Finally, we investigated platelet aggregation in response to AyPGKF-NH2 (a peptide with D-tyrosine in position two) which is unable to activate calcium signalling, and AYPGRF-NH2 a peptide that is equipotent to the parental peptide AYPGKF-NH2 for calcium signalling but is more potent at recruiting β-arrestins. We found that AyPGKF-NH2 fails to activate platelets while AYPGRF-NH2 causes a platelet aggregation response that is greater than that seen with the parental peptide and is comparable to that seen with thrombin stimulation. Overall, these studies uncover molecular determinants for agonist binding and signalling through a non-canonically activated GPCR and provide a template for development of small molecule modulators of PAR4.


242
We were excited to observe evidence of β-arrestin-biased peptides within this group, all of 243 which showed decreased calcium signalling and/or increased b-arrestin recruitment. Moreover, 244 the leftward shifts in their associated b-arrestin concentration-effect profiles imply that 245 substitution of N-methylated alanine 1 , tyrosine 2 , lysine 5 , or phenylalanine 6 residues increases 246 peptide potency with respect to G-protein-independent signalling. Only N-methylation of 247 glycine 4 (sarcosine) resulted in a reduced overall ability to activate PAR4, with relative calcium 248 and b-arrestin signals being reduced by approximately 90% and 60%, respectively.  Table 1).

253
Previously it was demonstrated that protonation of the N-terminus of the PAR1 tethered ligand 254 (SFLLRN) is critical for agonist function (Scarborough et al., 1992). Additionally, substitution 255 of glycine for mercaptoproprionic acid (Mpr), which lacks an amino-terminal protonated amine 256 found in the native tethered-ligand-mimicking peptide GYPGQV, resulted in complete loss of 257 agonist activity . Given the evidence for N-terminal protonation in PAR 258 agonism, we sought to determine the impact of N-terminal acetylation on PAR4-AP activity.

301
Effect of additional substitutions in positions 2-4 on PAR4-AP-mediated signalling. 302 Previously, structure activity studies with PAR4 activating peptides revealed an apparent 303 requirement for an aromatic residue in position 2 (tyrosine) and found that alterations to 304 position 3 (proline) and position 4 (glycine) were not well tolerated . In the 305 present study, we observed that substitution of either tyrosine, proline, or glycine with alanine 306 reduced calcium signalling potency and b-arrestin recruitment, however none of these 307 substitutions completely abolished the activity of the peptide (Fig. 1). Interestingly, substitution 308 of either L-tyrosine or L-proline with the respective D-isomer significantly reduced the activity 309 of the PAR4-AP (Fig. 2, Table 1). Together, these data suggest that while no individual residue 310 is strictly required for agonist activity, the side chains of these residues and their positioning 311 are important for agonist function. To further probe the contribution of these residues to the 312 PAR4-AP, we generated peptides with alterations in the aromatic side chain of position 2, the 313 backbone conformation of position 3, or position 4 (Fig. 5, Table 1).

335
Elongation of the backbone in position four (glycine) by substitution with beta-alanine (AYP-336 betaAla-KF-NH2) was also significantly detrimental to both calcium signalling (EC50 = n.d.) 337 and b-arrestin recruitment (β-arr-1 42.5 ± 1.8%, β-arr-2 24.9 ± 1.6%). Together these data   substituted with a charged residue such as arginine or ornithine . In the 348 present study, we have identified that substitution of position 5 (lysine) of the PAR4-AP 349 (AYPGKF-NH2) with alanine does not significantly affect the agonist capacity of the peptide 350 ( Fig. 1). Further, we identified that substitution of L-lysine with D-lysine significantly 351 decreases agonist stimulation of calcium signalling and β-arrestin recruitment (Fig. 2). To 352 further probe the effect of positively charged residues in position 5, we generated several 353 PAR4-AP analogues with lysine 5 substitutions by either arginine or ornithine (Fig. 6, Table 1).

354
Substitution of lysine 5 with arginine resulted in a partial agonist with respect to calcium 355 signalling but increased β-arrestin recruitment compared to the parental peptide (AYPGRF-356 NH2 calcium EC50 = 4.0 ± 1.1 µM, β-arr-1 154.2 ± 10.3%, β-arr-2 140.1 ± 9.4%). This for calcium signalling (EC50 = 6.7 ± 1.9 µM) and a modestly more efficacious agonist for the 366 recruitment of β-arrestins (β-arr-1 122.5 ± 12.4%; β-arr-2 116.7 ± 11.2%). All of these   and there appears to be modest freedom in the side-chain of this position. This makes the side 380 chain a good candidate for further investigations of structure-activity relationships as well as a 381 good region to label with an imagine modality. for an aromatic residue, as substitution with non-aromatic residues such as lysine or ornithine 389 significantly decreased or abolished the activity of the peptide . In addition 390 to being an aromatic residue, phenylalanine is a hydrophobic residue so we investigated the 391 impact of altering the hydrophobicity/hydrophilicity and stericity while maintaining an 392 aromatic residue (Fig. 7, Table 1).

394
We investigated the effect of polar changes that decrease hydrophobicity. Substitution of 395 phenylalanine 6 with tyrosine results in the addition of a hydroxyl group to position 6 396 (AYPGKY-NH2). We observed partial agonism of calcium signalling with no significant 397 change in potency compared to the PAR4-AP (AYPGKY-NH2, EC50 = 5.2 ± 1.0 µM).

437
PAR4-dependent calcium signalling is Gαq/11-dependent. In order to verify that PAR4-438 dependent calcium signalling was Gaq/11-dependent, HEK-293 cells stably expressing PAR4-439 YFP were stimulated with 100 µM AYPGKF-NH2 following pre-incubation with selective 440 Gαq/11-inhibitor YM254890. YM254890 inhibited AYPGKF-NH2-stimulated calcium 441 signalling in a concentration-dependent manner, with complete inhibition observed at 100 nM 442 YM254890. The control vehicle (0.001% DMSO) had no effect on calcium signalling (Fig. 9).  Interestingly, we find that most peptides that were unable to stimulate calcium signalling were  We generated a homology model of wild-type PAR2 on the PAR2 structure, which 479 contained modifications necessary for crystallization. Twenty human PAR4 homology models, 480 using the wild-type human PAR2 model as a template, were generated in MODELLER (Webb 481 and Sali, 2014). The PAR4 model with the lowest discrete optimized protein energy (DOPE) 482 score was utilized for in silico analyses.

509
To investigate whether these sites are important for tethered ligand activation of the receptor,  the key requirements that a PAR4 agonist peptide must meet. Previous structure-activity work 557 starting with the proteolytically-revealed sequences of human (GYPGQV) and murine 558 (GYPGKF) PAR4 identified the hexapeptide AYPGKF-NH2 or SYPGKF-NH2 to be potent 559 and selective agonists for PAR4 . In this study we started with the Previously, it was demonstrated that substitution of alanine or serine in position 1 of the 566 GYPGKF-NH2 activating peptide retained the ability to trigger tritiated inositol 1,4,5-567 triphosphate release to a level comparable to that observed with thrombin activation and much 568 higher than release caused by the GYPGKF-NH2 . Interestingly, 569 substitution of position 1 with threonine was reported to significantly reduce the potency of the 570 activating peptide . In keeping with results from previous studies, we find  Table 1). Additionally, we observed that both stereochemical inversion of L-alanine to D-   Table 1). Our findings, together with previously reported data , point 593 to the necessity of an aromatic residue in position 2 to maintain agonist potency. Further, our 594 data reveal that backbone orientation in position 2 is a key determinant of agonist potency.

595
Future experiments fine-tuning the aromatic substituents may be beneficial for peptide 596 analogue development.

598
Our data reveal that alterations affecting proline backbone conformation or glycine residue 599 compactness are not well-tolerated, as these decreased calcium signalling, p44/42 MAP kinase 600 signalling, and b-arrestin recruitment to PAR4. Calcium signalling has previously been 601 reported to be negatively impacted by alterations to either of these two residues (Faruqi et al.,602 2000). Both proline and glycine residues are frequently found in b-turns, in which proline 603 provides a characteristic kink, and the available data support the notion that the conformation(s) 604 (proline 3 and glycine 4 ) and rigidity (proline 3 ) of PAR4 agonists permitted by these residues 605 is/are important for receptor activation. To test this hypothesis further, we made several 606 substitutions in position 3 (proline) aimed at evaluating favoured backbone conformations. We 607 observed that a return to a predominantly trans-amide conformation by substitution of proline 608 with alanine ( Fig. 1), results in decreased potency in all pathways studied. Interestingly, this 609 substitution did not abolish signalling but did significantly decrease agonist potency, which 610 suggests that the backbone conformation of proline may be necessary for receptor binding but 611 not receptor activation. Substitution of proline to pipecolic acid (Pip., Fig. 5), which has a 612 similar peptide backbone orientation to proline with a six-membered aromatic ring, decreased 613 b-arrestin recruitment with relatively modest effects for calcium signalling activation.   (Fig. 1). Substitution of L-lysine 5 with D-lysine, however, significantly 642 reduced potency and efficacy in calcium signalling and b-arrestin recruitment assays and 643 furthermore resulted in a loss of stimulation of MAP kinase signalling (Fig. 2). We continued We next turned our attention to analogues that resemble the side chain of lysine 5 , with a specific 655 interest on structure-activity in addition to ascertaining the importance of charge in this residue.  Fig. 6). Interestingly, however, we observed significantly increased 662 p44/42 phosphorylation in response to AYPGOF-NH2 stimulation (Fig. 10). This is somewhat 663 puzzling since we found p44/42 phosphorylation to be Gaq/11-dependent (Fig. 11); however, it To determine if any of the calcium signalling-null peptides were able to act as antagonists, cells 714 were pre-treated with calcium null peptides and subsequently stimulated with AYPGKF-NH2 715 to see if the parental peptide was able to still elicit a signal. We observed that both AY-Nip-716 GKF-NH2 and AYP-Sar-KF-NH2 were both able to significantly decrease the calcium signal 717 elicited by PAR4-AP (Fig. 8). Interestingly, each of these substitutions resulted in ablation of 718 both calcium and p44/42 signalling as well as significantly decreasing b-arrestin recruitment.

719
These results may indicate that targeting PAR4 with small molecules mimicking position 3 and 720 4 dipeptides may provide a starting place for the design of novel PAR4 antagonists. 721 Additionally, peptides that further explore modifications to position 3 and 4 could be 722 investigated for potent PAR4 antagonists.

724
In platelets, it has previously been demonstrated that PAR4-mediated calcium mobilization µM) did not stimulate platelet aggregation (Fig. 14). Further, stimulation of washed platelets 731 with the agonist peptide, AYPGRF-NH2 (Fig. 6), enhanced platelet aggregation to levels 732 comparable to those observed with thrombin activation of PAR4 (Fig. 14). AYPGRF-NH2 733 stimulates calcium signalling to levels comparable to those seen with PAR4-AP but triggers 734 more b-arrestin-1 recruitment than the parental peptide. These data suggest that calcium 735 signalling is essential for triggering platelet activation, but arrestin recruitment may also play 736 a role.

738
Having ascertained some of the governing peptide residue characteristics enabling PAR4 739 activation and validated their signalling consequences in an ex vivo system, we turned our

805
Finally, in the context of platelet activation we find that our modified peptides that are able to 806 cause a greater level of calcium signalling and b-arrestin recruitment than PAR4-AP could also 807 trigger greater platelet aggregation. In our hands, PAR4-AP was un able to elicit greater than 808 50% of the aggregation observed in response to 1 unit/mL of thrombin, while the peptide 809 AYPGRF-NH2 was able to trigger aggregation comparable to that seen with thrombin. A role  In summary, the present studies identify key mechanisms involved in PAR4 activation and 819 signalling. We also present and characterize a novel toolkit of PAR4 agonist peptides that could 820 be used to study biased signalling through PAR4 in platelets as well as other physiological 821 systems. We further show that platelet activation by PAR4 critically depends on the Gaq/11 822 pathway, and selective targeting of this pathway might yield a useful anti-platelet agent. 823 Overall, our work advances our understanding of agonist binding to PAR4 and will support 824 future efforts aimed at defining signalling contribution in homeostatic signalling and the 825 discovery of novel therapeutics targeting PAR4.   (Table S1). All peptides had purity >95% as determined by analytical 902 HPLC (see supplemental information, Figure S1).    or PAR4 mutant receptors (described above) using the calcium phosphate transfection protocol 990 as described above. Calcium signalling was investigated as described above for the wild type   using one-way ANOVA of response compared to PAR4-AP (*p < 0.05). control; *p > 0.05 AYPGKF-NH2 compared to other peptide agonists; one-way ANOVA).

1291
Peptides that were unable to stimulate calcium signalling were also generally unable to 1292 stimulate phosphorylation of p44/42. Interestingly, several peptides were able to significantly   Figure S1: Analytical RP-HPLC UV detection chromatograms of synthesized peptides.