Helix-bundle and C-terminal GPCR domains differentially influence GRK-specific functions and β-arrestin-mediated regulation

Abstract

G protein-coupled receptors (GPCRs) orchestrate diverse physiological responses via signaling through G proteins, GPCR kinases (GRKs), and arrestins. While most G protein functions are well-established, the contributions of GRKs and arrestins remain incompletely understood. Here, we investigate the influence of β-arrestin-interacting GPCR domains (helix-bundle/C-terminus) on β-arrestin conformations and functions using refined biosensors and advanced cellular knockout systems. Focusing on prototypical class A (b2AR) and B (V2R) receptors and their chimeras (b2V2/V2b2), we show that most N-domain β-arrestin conformational changes are mediated by receptor C-terminus-interactions, while C-domain conformations respond to the helix-bundle or an individual combination of interaction interfaces. Moreover, we demonstrate that ERK1/2 signaling responses are governed by the GPCR helix-bundle, while β-arrestin co-internalization depends on the receptor C-terminus. However, receptor internalization is controlled via the overall GPCR configuration. Our findings elucidate how individual GPCR domains dictate downstream signaling events, shedding light on the structural basis of receptor-specific signaling.

Fig. 1. Model class A and B GPCRs, β2 adrenergic receptor (b2AR) and Vasopressin receptor 2 (V2R), induce distinct patterns of β-arrestin conformational patterns.

a Schematic representation of NanoBRET-based assay to measure β-arrestin2-NanoLuciferase (Nluc) recruitment to the GPCR-Halo-Tag in Control cells. For the b2AR measurements, the BRET pair was switched. b, c β-arrestin2 recruitment to b2AR (b) and V2R (c), as published in Drube et al.. Here, the Δ net BRET fold change was analyzed over time when cells were stimulated with 10 µM Isoprenaline (Iso) (b) or 3 µM [Arg8]-Vasopressin (AVP) (c). Timepoint 0 is defined as midpoint of the average time needed for ligand addition, considering all analyzed biological and technical replicates. All data are shown ±SEM of n = 3 independent experiments. d Schematic visualization of the utilized intramolecular NanoBRET-based β-arrestin2 sensors, as described in detail in Haider et al.. e, f Conformational change of β-arrestin2-FlAsH (F) 5-Nluc in Control cells, co-transfected with untagged b2AR (e) or V2R (f). Data were analyzed over time as in (b, c) and are shown as Δ net BRET change in percent. g–j Fingerprint of β-arrestin2 conformational change sensors measured in Control cells in presence of untagged b2AR (g, i) or V2R (h, j). The Δ net BRET change at 10 µM Iso (g) or 3 µM AVP (h) are shown as bar graphs. All data are shown ±SEM of n = 3 independent experiments. Sensor conditions, which did not fulfill the set pharmacological parameters (Hill slope and EC₅₀ analysis as described in methods) were classified as non-responding conditions and assigned zero. These data were visualized onto the surface of the inactive β-arrestin2 crystal structure (PDB: 3P2D) by coloring the respective loop (-fragments) of the labeled FlAsH positions (i, j) ranging from blue to red. For each receptor, the Δ net BRET change was normalized to the maximally reacting sensor (red). Non-responding conditions are shown in gray. Source data are provided as a Source Data file.

Fig. 2. Different combinations of GPCR helix bundles and C-termini orchestrate distinct, GRK-dependent β-arrestin2 interactions.

a Schematic representation of the utilized receptor constructs and indication of the exchanged amino acids to create the chimeric constructs (b2V2 with the receptor transmembrane helix bundle of the b2AR and C-terminus of the V2R and vice versa for the V2b2) after Oakley et al.. b (Partial) amino acid sequence of the b2AR and V2R C-terminus, including a schematic representation of the phospho-sites, targeted with specific antibodies. Antibody shape created in BioRender. Klement, L. (2025) https://BioRender.com/721vno0. c–j Measurements of proximal and distal C-terminal phosphorylation of b2AR (c, g), V2b2 (d, h), b2V2 (e, i) and V2R (f, j) in quadruple GRK2/3/5/6 knockout cells (ΔQ-GRK) or Control cells, utilizing a bead-based GPCR phosphorylation immunoassay. The stably expressed b2AR (c, g) and b2V2 (e, i) were stimulated with different concentrations of Iso and the V2b2 (d, h) and V2R (f, j) with AVP, as indicated. Data are shown as optical density (OD) at 405 nm ±SEM of n = 5 independent experiments, normalized to the maximum ligand concentration for each receptor in Control cells, respectively. k–n β-arrestin2 recruitment to the b2AR (k), V2b2 (l), b2V2 (m) and V2R (n) in ΔQ-GRK or Control cells stimulated with Iso (k, m) or AVP (l, n) as indicated. Data for b2AR, b2V2 and V2R (k, m, n) were published in Drube et al. and are shown again to allow a direct comparison. For each receptor, Δ net BRET changes ±SEM of n = 3 independent experiments measured in ΔQ-GRK cells are shown in percent, normalized to the respective maximal change in Control cells. o–r β-arrestin2-F5 conformational changes in presence (Control) or absence of endogenous GRK2/3/5/6 (ΔQ-GRK) when coupling to the b2AR (o), V2b2 (p), b2V2 (q) or V2R (r). Data of n = 3−5 independent experiments are shown as Δ net BRET change (%) ± SEM (exact n numbers for each receptor-, GRK- and sensor-specific condition can be accessed in the source data). s–v Complete β-arrestin2 conformational change fingerprints are shown for each receptor in Control and ΔQ-GRK cells, stimulated with 10 µM Iso (s, u) or 3 µM AVP (t, v), as Δ net BRET change (%) in radar plots. Sensor conditions, which did not fulfill the pharmacological parameters (Hill slope and EC₅₀ analysis as described in methods) were classified as non-responding conditions and assigned zero. Source data are provided as a Source Data file.

Fig. 3. The kinetic profile of ERK1/2 phosphorylation is governed by the GPCR transmembrane helix bundle.

a–d Representative Western blots of phosphorylated ERK (pERK) and total ERK levels over time in Control or ΔQ-GRK cells, stably expressing the b2AR (a), V2b2 (b), b2V2 (c) or V2R (d). Cells were stimulated with 1 µM Iso (a, c) or 100 nM AVP (b, d) for the time indicated. e–h Data of n = 3 independent experiments shown in (a–d) were quantified and are shown as ERK1/2 activation over time ±SEM (pERK1/2 divided by total ERK1/2). To compare the pERK changes over time, the area under the curve (AUC) was quantified for each condition (Supplementary Fig. 4) and complete results of the statistical analysis can be accessed in Supplementary Table. 1. Source data are provided as a Source Data file.

Fig. 4. Differential GPCR phosphorylation by cytosolic and membrane-bound GRK isoforms induces distinct β-arrestin interactions with (chimeric) class A and B GPCRs.

a–h Measurements of proximal and distal C-terminal phosphorylation in ΔQ-GRK stably expressing b2AR (a, e), V2b2 (b, f), b2V2 (c, g) or V2R (d, h) and GRK2-YFP or GRK6-YFP, analogously to Fig. 2c–j. Data are shown as optical density (OD) at 405 nm ±SEM of n = 5 independent experiments, normalized to the maximum ligand concentration for each receptor in Control cells, respectively. Statistical differences between measurements in ΔQ-GRK + GRK2, ΔQ-GRK + GRK6 or Control cells were compared using a two-way analysis of variance (ANOVA), followed by a Tukey’s test for vehicle or highest stimulating concentrations (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001; ns not significant). Complete results of the statistical analysis can be accessed in Supplementary Table 2. i–l Fingerprint of β-arrestin2 conformational change sensors measured in ΔQ-GRK cells, individually overexpressing GRK2 or GRK6, in presence of untagged b2AR (i), V2b2 (j), b2V2 (k) or V2R (l). The Δ net BRET changes at 10 µM Iso (i, k) or 3 µM AVP (j, l) are shown as Δ net BRET change (%) in radar plots, analogously to Fig. 2s–v. Sensor conditions, which did not fulfill the set pharmacological parameters (Hill slope and EC₅₀ analysis as described in methods) were classified as non-responding conditions and assigned zero. m, n The Δ net BRET changes (%) at highest stimulating ligand concentrations for each FlAsH position were clustered for all receptor–β-arrestin pairs in presence of GRK2 (m) or GRK6 (n) according to Manhattan distance. GPCR transmembrane helix bundle or C-terminus identity is indicated by colored bar (b2AR in orange, V2R in blue). Source data are provided as a Source Data file.

Fig. 5. β-arrestin translocation to early endosomes is controlled by the receptor C-terminus.

Control and ΔQ-GRK cells, overexpressing GRK2, GRK6 or no GRKs (empty vector (EV)-transfected) were transfected with the indicated receptor-CFP, β-arrestin-YFP and early endosome marker Rab5-mCherry. Confocal images were taken before (basal) and after 15 min of specified ligand. a Representative images for all receptors, expressed in Control cells. b The co-localization of Rab5 with β-arrestin2 in Control cells was quantified in over 28 images for each condition using Squassh and SquasshAnalyst^,. Data are presented as mean fold change in co-localization (signal) + SEM, normalized to the respective unstimulated (basal) ΔQ-GRK + EV condition. Statistical comparison between basal and stimulated values was performed using a two-way ANOVA, followed by a Sidak’s test (ns not significant; ****p < 0.0001). Detailed results, also for the two-way ANOVA, followed by a Tukey’s test to compare basal and stimulated values between different conditions, can be accessed in Supplementary Tables 3–5. c–f Analogous to the data shown in (b), the fold change in co-localization (signal) was analyzed for the receptor or β-arrestin2 with Rab5, as fold change over the respective basal ΔQ-GRK + EV condition for each receptor and GRK condition (Control, ΔQ-GRK + EV, +GRK2, +GRK6). The stimulated values for receptor–Rab5 co-localization are show on the x-axis and for β-arrestin2–Rab5 co-localization on the y-axis. For both dimension, the data points were normalized to the respective maximum (for receptor–Rab5 co-localization: V2b2 in Control cells, for β-arrestin2–Rab5 co-localization: b2V2 in Control cells) and are shown in percent. Source data are provided as a Source Data file.

Fig. 6. Distinct aspects of receptor–β-arrestin2 interactions are differentially influenced by GPCR transmembrane helix-bundles or C-termini.

a Complex configurations for β-arrestins with an adrenergic receptor (β1 adrenergic receptor (b1AR), here exchanged by the aligned structure of active b2AR, PDB: 3SN6) and the V2R (PDB: 7R0C) are shown. Both complex structures feature β-arrestin2 aligned in place of β-arrestin1, while calculated transferability coefficients (see Supplementary Fig. 11 and methods) of β-arrestin2 biosensors in Control cells are projected onto its surface structure (PDB: 3P2D) to indicate the influence of individual GPCR domains on conformational changes (blue indicates helix-bundle transferability of conformational changes, while shades of red indicate C-terminus transferability and biosensors that do not show any transferability are colored in gray). b Calculated transferability coefficients of all β-arrestin2 biosensors in Control cells are shown as bubble plots, also indicating the initial difference in signal between WT GPCRs, shown as the size of the plotted symbols. c, d β-arrestin configurational positioning and tilt angles (with respect to the GPCR alignment axis and plasma membrane) are compared between complexes formed with an adrenergic receptor (orange outline) and the V2R (blue outline). e, f Calculated transferability coefficients are depicted for functional assay data, featuring ERK1/2 activity, β-arrestin2 recruitment, GPCR and β-arrestin2 endosomal delivery, as well as proximal and distal, C-terminal GPCR phosphorylation, in all GRK-specific conditions. These coefficients are shown schematically in (e), and as bubble plots in (f), while the size of the plotted symbols indicates the initial difference in signal between WT GPCRs. Source data are provided as a Source Data file.