Transient ligand contacts of the intrinsically disordered N-terminus of neuropeptide Y2 receptor regulate arrestin-3 recruitment

Abstract

Previous efforts in delineating molecular mechanisms of G protein-coupled receptor (GPCR) activation have focused on transmembrane regions and ligand-receptor contacts of the extracellular loops. The role of the highly flexible N-termini of rhodopsin-like GPCRs have not been well characterized to date. We hypothesize that transient contacts between the peptide ligand and the intrinsically disordered N-terminus (NT) of the neuropeptide Y (NPY) receptor Y₂ (Y₂R) will affect receptor signaling. We employ cross-linking mass spectrometry to capture ligand-receptor contacts including transient binding modes. A photo-reactive NPY analogue allows mapping the interaction between NPY and Y₂R NT resulting in a total number of 40 cross-links. The cross-links provide distance constraints for deriving structural models of the interaction. Molecular dynamics simulations highlight the structural flexibility and rapid interconversion of ligand-receptor contacts. Mutagenesis of Y₂R and functional characterization suggest that the cross-linking hotspots in the NT electrostatically control its conformational ensemble. The NT engages in transient contacts to the peptide and prolongs ligand residence time, which is required for efficient interaction of Y₂R with arrestin-3, but not G_i. We delineate structure-function relationships for the intrinsically disordered Y₂R NT and propose a functional role for transient binding modes involving the NT of a peptide-binding receptor.

Fig. 1. XL-MS between NPY and Y₂R.

a Scheme of photoaffinity labeling. NPY containing photo-leucines at positions 17, 24, and 30 was cross-linked to bicelle-reconstituted Y₂R preparations by UV-A light. The cross-linked complex is enzymatically digested, peptides are analyzed by liquid chromatography and trapped ion mobility spectroscopy tandem mass spectrometry (LC-TIMS-MS/MS) using the MeroX software. b Overview of XL-MS results. Each cross-link is represented by a green line. Inset: The majority of cross-links were found between L24^NPY and L30^NPY to Y₂R NT. Cross-linking experiments were conducted at least three times independently. Related to Supplementary Data 1, listing all observed cross-links.

Fig. 2. Structural flexibility of the Y₂R NT in microsecond MD simulations.

a Violin plot of the backbone-RMSD values for each residue of the NT during the MD simulations, relative to the initial frame, grouped in sets of five. b Example conformations sampled during MD simulations. Residue coloring matches the groups in the violin plot with a color gradient from orange to pink from distal to membrane-proximal regions and opacity increasing over simulation time.

Fig. 3. Residue Contacts between Y₂R NT and NPY.

a Flare plot showing the contact frequency between residues of Y₂R NT and NPY from microsecond MD simulations. Contacts are defined as residues within a distance of 4.5 Å (excluding direct neighbors) and are represented by black lines. Line opacity indicates contact frequencies, with higher opacity reflecting more frequent contacts. The acidic patches E15-E20 and D35-E39 of the NT are highlighted. b Visualization of overall contact frequency between residues of NPY and Y₂R NT. Contact frequencies are shown by red hue intensity, with higher frequencies indicated by darker shades. Residues with high contact frequencies are labeled for clarity. c Distance plots for selected residue pairs M17^Y2R-R25^NPY, D35^Y2R-R25^NPY, and L40^Y2R-R19^NPY. Different colors represent different simulation runs. Runs were smoothed by averaging over ten frames per point. The original, unsmoothed time trace is shown in the same color with higher transparency. Corresponding structural snapshots at 0.2 µs are shown below the plots using the same color scheme.

Fig. 4. Y₂R NT mutations differentially modulate G protein activation and recruitment of arrestin-3.

a Close-up view to membrane proximal Y₂R NT from cryo-EM (PDB 7X9B) highlighting interactions of D42-S43-T44^NT. b Live-cell fluorescence microscopy shows that Y₂R NT variants are transported to the plasma membrane like the wild-type. Y₂R-eYFP variants are shown in yellow, cell nuclei are stained by H33342 and shown in blue, scale bar equals 10 µm. All pictures were acquired with identical light exposure and picture processing. c Quantification of cellular receptor expression is based on eYFP fluorescence in a plate reader, normalized to wild-type Y₂R. d Activity of Y₂R variants towards G_i1 proteins as measured by a direct BRET assay (G_i1-CASE) 10 min after ligand stimulation. All variants show full activation; D42A-S43A-T44A has the strongest shift in potency, while all others show mild or no effects. e Kinetics of G_i1 activation after stimulation with 1 nM NPY is overall similar for most variants except Δ2–41 and D42A-S43A-T44A, which have a tendency for faster apparent rate constants (k_obs). 95% CI of wild-type Y₂R k_obs is given as a gray rectangle for comparison. f Recruitment of arrestin-3 to Y₂R variants as measured by BRET 10 min after NPY stimulation. D42A-S43A-T44A and Δ2–41 variants reduce recruitment of arrestin-3 to the receptor. Neutral (in green) and charge-inverted (in blue) variants of the acidic patch in Y₂R NT lead to distinct behavior, with only charge reversal impairing arrestin-3 recruitment. g Kinetic analysis of arrestin-3 recruitment after stimulation with 100 nM/ 1 µM NPY. Charge reversal in acidic patches 1 and 2 leads to faster initial recruitment, but also early signal decay. The bar plot shows quantification of the initial rate of arrestin-3 recruitment to receptor variants. Apparent rate constants (k_obs) are plotted in comparison to wild-type Y₂R-eYFP and its 95% CI (gray rectangle). * P < 0.05, ** P < 0.01 in one-way ANOVA, with Dunnett’s post-hoc test corrected for multiple comparisons against wild-type Y₂R. For the color legend, please see panel (e). Data in (c–g) are the mean ± SEM of n = 5 (c), n = 4-9 (d), n = 5-6 (e), n = 3 (f), n = 3-4 (g) independent experiments, each performed in technical triplicate.

Fig. 5. Ligand binding at Y₂R NT variants.

a Equilibrium binding between K18-Tamra-NPY and Nluc-Y₂R-eYFP variants. Charge reversal in the acidic patches retains wild-type-like equilibrium binding properties, while D42A-S43A-T44A displays a significantly smaller BRET window and about three-fold reduced low-affinity binding compared to wild-type Y₂R. b N-terminal deletion in Δ2–41 Y₂R increases overall BRET window as expected from the reduced distance, and the low-affinity state has about two-fold reduced affinity compared to wild-type Y₂R. Please note that the Y-axis scaling is different between panels (a and b). c Ligand dissociation from the low-affinity state (bound to 300 nM K18-Tamra-NPY; re-binding blocked by 50 µM antagonist) is faster for Δ2–41 and acidic patch charge reversal mutants. Data are the mean ± SEM of n = 3-4 (a, b) or n = 4 (c) independent experiments, performed in technical duplicate or triplicate. * P < 0.05 in one-way ANOVA, with Dunnett’s post-hoc test corrected for multiple comparisons against wild type Y₂R. Related to Supplementary Fig. 11: Expression and activity of Nluc-Y₂R-eYFP variants used for NanoBRET binding.