Live-cell quantitative monitoring reveals distinct, high-affinity Gβγ regulations of GIRK2 and GIRK1/2 channels

Handklo-Jamal, Reem; Raifman, Tal Keren; Shalomov, Boris; Hofer, Patrick; Kahanovitch, Uri; Friesacher, Theres; Tabak, Galit; Tsemakhovich, Vladimir; Reddy, Haritha P.; Chomsky-Hecht, Orna; Tripathy, Debi Ranjan; Zuhlke, Kerstin; Dessauer, Carmen W.; Klussmann, Enno; Haitin, Yoni; Hirsch, Joel A.; Stary-Weinzinger, Anna; Yakubovich, Daniel; Dascal, Nathan

doi:10.1038/s41467-025-66730-8

Download PDF

Article
Open access
Published: 24 November 2025

Live-cell quantitative monitoring reveals distinct, high-affinity Gβγ regulations of GIRK2 and GIRK1/2 channels

Nature Communications volume 16, Article number: 11607 (2025) Cite this article

4071 Accesses
1 Altmetric
Metrics details

Subjects

Abstract

G_i/o protein-coupled receptors (GPCRs) inhibit cardiac and neuronal excitability via G protein-activated K⁺ channels (GIRK), assembled by combinations of GIRK1 - GIRK4 subunits. GIRKs are activated by direct binding of the Gβγ dimer of inhibitory G_i/o proteins. However, key aspects of this textbook signaling pathway remain debated. Recent studies suggested no G_i/o-GIRK pre-coupling and low (>250 µM) Gβγ-GIRK interaction affinity, contradicting earlier sub-µM estimates and implying low signaling efficiency. We show that Gγ prenylation, which mediates Gβγ membrane attachment required for GIRK activation, also contributes to the Gβγ-GIRK interaction, explaining the poor affinity obtained with non-prenylated Gβγ. Using quantitative protein titration and electrophysiology in live Xenopus oocytes, Gβγ affinity for homotetrameric GIRK2 ranges from 4-30 µM. Heterotetrameric GIRK1/2 shows a higher Gβγ apparent affinity due to the Gβγ-docking site (anchor) in GIRK1, which enriches Gβγ at the channel. Biochemical approaches and molecular dynamic simulations reveal that the Gβγ anchor is formed by interacting N-terminal and distal C-terminal domains of the GIRK1 subunits, distinct from the Gβγ-binding “activation” site(s) underlying channel opening. Thus, the affinity of Gβγ-GIRK interaction is within the expected physiological range, while dynamic pre-coupling of Gβγ to GIRK1-containing channels through high-affinity interactions further enhances the GPCR-G_i/o-GIRK signaling efficiency.

A selectivity filter mutation provides insights into gating regulation of a K⁺ channel

Article Open access 11 April 2022

Computationally designed GPCR quaternary structures bias signaling pathway activation

Article Open access 11 November 2022

GRK specificity and Gβγ dependency determines the potential of a GPCR for arrestin-biased agonism

Article Open access 03 July 2024

Introduction

G protein-activated inwardly rectifying K⁺ channels (GIRK; Kir3) mediate inhibitory effects of G_i/o protein-coupled receptors (GPCRs), controlling neuronal and cardiac excitability; GIRK malfunction is linked to neurological, cardiac and endocrine disorders^1,2,3,4. GIRKs form homotetramers (GIRK2, GIRK4) or heterotetramers (GIRK1/2, 1/4, 1/3, 2/3), differing in tissue distribution and gating properties. Homotetrameric GIRK2 is best characterized quantitatively, including a crystal structure of GIRK2-Gβγ complex⁵. GIRKs are activated by direct, cooperative binding of up to 4 molecules of Gβγ^6,7,8,9 (Fig. 1a). This membrane-delimited process requires posttranslational Gγ prenylation, essential for Gβγ accumulation at the plasma membrane (PM)¹⁰ and GIRK activation^11,12.

**Fig. 1: Lipid modification of Gγ is essential for GIRK activation and important for GIRK-Gβγ interaction.**

The coupling between GPCR and Gα_i/oβγ in the GPCR-Gα_i/oβγ-GIRK cascade varies by receptor, G protein and cell type, ranging from collision-coupling (e.g., muscarinic m2 receptor (m2R)^13,14,15,16) to precoupling within dynamic multiprotein complexes (e.g., GABAB receptor with G_i/o and GIRK¹⁷), or a combination of both modes within protein-enriched membrane “hot spots”¹⁸, organized by specific scaffolds¹⁹ or driven by low-affinity protein interactions²⁰.

Controversies linger regarding the affinity, specificity, and efficiency of Gα_i/oβγ-GIRK coupling. Early in vitro measurements of GIRK interaction with prenylated Gβγ yielded dissociation constants (K_d) between 50-800 nM^8,21, comparable to other Gβγ interactors (3 nM-3 µM; Supplementary Table 1). Contrastingly, an NMR study reported a K_d of 250 µM for the interaction of non-prenylated Gβγ with GIRK1’s truncated cytosolic domain²². Wang, Touhara, MacKinnon and colleagues analyzed Gβγ activation of purified recombinant GIRK2 while controlling Gβγ’s surface density by titrating a non-prenylated His-Gγ into GIRK2- and NTA lipid-containing bilayers. Their studies revealed high cooperativity of Gβγ binding and its allosteric enhancement by Na⁺ and PIP₂^15,23,24,25. The resulting model, termed here WTM model, postulated sequential Gβγ binding to GIRK2, with channel opening when all four Gβγ sites are occupied^15,24 (Fig. 1a). Unexpectedly, binding of the first Gβγ showed an exceptionally low affinity, with K_d ~ 1.9 mM at [Na⁺]=0 and ~300 µM at saturating [Na⁺]²⁴. (Due to cooperativity, the affinity increases for subsequent Gβγ bindings; Supplementary Table 2).

Low affinity entails inefficient signaling. With a K_d > 250 µM, GIRK activation (10-80%, depending on intracellular Na⁺ concentration, [Na⁺]_in) would require free surface Gβγ exceeding 1200 µm⁻² (molecules/µm²)²⁴, hundredfold higher than the 2-10 µm⁻² GIRK channel density in PM of neurons or atrial myocytes^16,26. While there is no evidence for such massive accumulation of Gαβγ around GIRKs, it could theoretically occur in membrane “hot spots”. Alternatively, higher affinity or GIRK-G protein preassociation could enable fast and efficient signaling²⁷. Several studies suggest preassociation of GIRKs with Gβγ or Gαβγ heterotrimers^{17,28,29,30,31,32}, while others support pure collision coupling^16,23,24. Subunit-specific differences may play a role. GIRK1, but not GIRK2, recruits Gβγ to the PM; the cytosolic distal C terminal segment of GIRK1 (G1-dCT) is essential for Gβγ recruitment³³. We previously proposed that G1-dCT is part of a Gβγ-docking site (Gβγ anchor) that facilitates high-affinity, dynamic (reversible) pre-association of GIRK1/2 with Gβγ^{4,14,33,34,35}. However, the exact composition and interaction mode of the Gβγ-anchor remain unclear. Here we show that, besides driving Gβγ’s attachment to the PM, Gγ’s prenylation directly contributes to GIRK-Gβγ interaction. We quantitate interactions between Gβγ and GIRK channels in living cells by titrated protein expression and PM level monitoring, combined with biochemical assays and computational approaches. We demonstrate efficient, low-micromolar affinity, subunit-specific GIRK regulation by Gβγ and determine the composition of GIRK1’s Gβγ-docking site. Our protein titration methodology can facilitate quantitative studies of additional membrane-delimited signaling cascades in living cells.

Results

Lipid modification of Gγ is essential for GIRK activation and important for GIRK-Gβγ interaction

All high-affinity estimates of Gβγ-GIRK binding were obtained using prenylated Gβγ. We hypothesized that Gγ’s prenylation enhances Gβγ-GIRK interaction, similarly to Gβγ interactions with GPCRs, Gα, adenylyl cyclase and phospholipase Cβ^{36,37,38,39,40,41,42}.

In cells, the prenyl (geranylgeranyl in Gγ₂) moiety, Gγ_prenyl, is attached to Cys68 within the C-terminal CAAX motif, while the remaining residues are cleaved¹⁰. To assess the role of prenylation we used the non-prenylated mutant Gγ_C68S that associates with Gβ³⁹; however, Gβγ_C68S fails to activate GIRK channels in excised PM patches^11,12. We expressed GIRK2 channels with m2R (adjusted to maximize I_evoked¹⁴) and Gβ₁γ₂ (Gβγ) or Gβγ_C68S in Xenopus oocytes and measured whole-cell basal (I_basal), agonist (acetylcholine; ACh)-evoked (I_evoked), and Gβγ-induced (I_βγ) GIRK currents. GIRK2 had a small I_basal³⁴, which was enhanced 4-8-fold by ACh (by activating the endogenous Gα_i/oβγ) and 30-60 fold by coexpressing nearly-saturating doses of Gβγ. In contrast, the non-prenylated Gβγ_C68S neither activated GIRK nor affected I_evoked (Fig. 1b–d, and Supplementary Fig. 1). We verified that N-terminally labeled YFP-Gγ and YFP-Gγ_C68S were comparably expressed in whole oocytes and supported the expression of Gβ (Supplementary Fig. 1b, c). To assess PM localization, we immunostained Gβ in excised giant membrane patches (GMP)^32,43 using wild-type (WT) Gβ or an N-terminally myristoylated Gβ (myr-Gβ). Only WT Gγ (Gγ_WT), but not Gγ_C68S, supported GIRK2 activation and, correspondingly, PM enrichment of Gβ_WT and myr-Gβ (Supplementary Fig. 1d–f).

These results confirm proper prenylation of Gγ in oocytes, which is essential for PM attachment of Gβγ and GIRK2 activation; but is it also involved in Gβγ interaction with GIRKs? We examined the interaction of purified, His-tagged Gβγ and Gβγ_C68S with in vitro translated (ivt) Gβγ-binding proteins: Gα_i3; phosducin; cytosolic domains of GIRK1 and GIRK2 (G1NC and G2NC, respectively); and their truncated versions, G1NC_ΔdCT and G2NC_trunc (Fig. 1e). G1NC is a fusion protein of N- and C-terminal domains of GIRK1 (G1-NT and G1-CT). G1NC_ΔdCT lacks the G1-dCT and binds Gβγ much weaker than G1NC³³ (Supplementary Fig. 2). G2NC_trunc lacks the distal segments of the N- and C-terminal domains (G2-NT and G2-CT, respectively), as in structural and bilayer studies^23,44,45. All ivt proteins bound Gβγ. Remarkably, lack of prenylation dramatically reduced Gβγ interaction with Gα_i3 and phosducin, corroborating previous reports^36,37,38, and with all GIRK constructs (Fig. 1f, g), suggesting that Gγ prenylation directly contributes to Gβγ-GIRK interaction.

Estimating Gβγ density in PM using calibrated fluorescence and quantitative Western blotting

We aspired to quantitatively analyze the membrane-delimited GIRK-Gβγ interaction in intact cells. To accurately calibrate protein surface density, we extended our previously developed calibration methods in Xenopus oocytes³⁵, which use two independent approaches.

The calibrated fluorescence (CF) approach measures the surface density of yellow, cyan or Split-Venus fluorescent proteins (YFP, CFP or SpV; collectively xFP), using xFP-labeled channels as molecular calipers. We used Gβγ-activated xFP-GIRK1/2³⁵, and additionally the constitutively active homotetrameric IRK1-YFP (Fig. 2a, b). Calibration involved expressing these channels at varying RNA doses, measuring whole-cell currents, and calculating the surface density of functional channels based on open probability (P_o), single-channel current (i_single) and cell’s surface area⁴⁶ (Eq. 1 in Methods, Supplementary Fig. 3, Supplementary Table 3). YFP surface density was calculated assuming two or four YFP molecules per YFP-GIRK1/2 or IRK1-YFP channel, respectively. To avoid artifacts arising from any non-functional channels, we used channels’ RNA doses in the 0.01–1 ng range, ensuring a linear relationship between fluorescence and whole-cell current and, accordingly, the calculated YFP surface density (Fig. 2a). Deviations were observed only at high levels of YFP-GIRK1/2 (5 ng RNA; Supplementary Fig. 6a). Additionally, comparing calibration with both YFP-GIRK1/2 and IRK1-YFP in the same experiment gave almost identical estimates of YFP surface density (Fig. 2a). Concomitantly, we expressed Gβ·_YFPGγ (Gβ and YFP-Gγ) in separate groups of oocytes, measured YFP fluorescence at the oocyte’s perimeter, and converted it to YFP-Gγ surface density with each caliper. The estimates of YFP-Gγ with both calipers showed strong linear correlation with a slope of 0.9 (Fig. 2b), validating the calibration protocol.

**Fig. 2: Estimating Gβγ density in PM using calibrated fluorescence (CF) and quantitative Western blotting (qWB).**

The CF procedure with Gβ·_YFPGγ monitors YFP-Gγ rather than Gβ. We directly assessed the surface density of Gβ using the independent approach³⁵, quantitative Western blotting (qWB) of manually separated oocyte plasma membranes. We measured PM-associated Gβ with a Gβ antibody, using purified recombinant Gβγ for calibration (Fig. 2c, d). The PM density of the endogenous oocyte’s Gβ was 30 ± 13 µm⁻², consistent with previous estimates³⁵ and comparable to ~40 µm⁻² in HEK cells⁴⁷. Expressed Gβ surface levels were similar with either coexpressed Gγ or YFP-Gγ (Fig. 2e, and Supplementary Table 4). Overall, expressed surface Gβ (with 5 ng Gβ RNA) measured by qWB was 35 ± 9 µm⁻² (n = 6), about 2.5-fold lower than surface YFP-Gγ estimated by CF (91 ± 19 µm⁻², n = 7, Fig. 2f). The difference is probably not related to methodology, because previously both CF and qWB gave similar estimates of 22-28 µm⁻² for a YFP-labeled Gβ³⁵. Thus, evaluating YFP-Gγ may overestimate the coexpressed Gβ’s surface density, possibly because YFP-Gγ associates with endogenous Gβ, or exists as a separate protein^48,49. Therefore, we tested a variety of C- or N-terminally xFP-fused Gβ constructs (Supplementary Fig. 4). However, they yielded partial or no GIRK2 activation, and usually poorly activated GIRK1/2. SpV-Gβγ activated both GIRK1/2 and GIRK2 but induced smaller currents than Gβγ_WT. Only Gβ·_YFPGγ activated GIRK channels like the Gβγ_WT³³.

We next varied expression levels of Gβ·_YFPGγ and examined changes in surface densities of YFP-Gγ in intact oocytes and Gβ in GMPs (Fig. 2g). Reassuringly, surface levels of Gβ and YFP-Gγ were linearly correlated, with either GIRK2 or GIRK1/2 coexpressed (Fig. 2h, i). Thus, RNA dose-dependent changes in surface YFP-Gγ reflect corresponding changes in surface Gβ. Consequently, we routinely used Gβ·_YFPGγ in the following experiments.

Affinity of Gβγ-GIRK2 interaction is in the low µM range

We investigated the dose-dependent activation of GIRK2 by Gβ·_YFPGγ using the CF approach. We expressed GIRK2 with a range of Gβ·_YFPGγ RNA doses. Following calibration (Fig. 3a), we quantified surface Gβ·_YFPGγ density in individual oocytes and then measured single-channel P_o in cell-attached patches of the same cells (Fig. 3b–d). The activation of GIRK2 was steeply Gβ·_YFPGγ dose-dependent, with an initial slope of almost 3 on log-log coordinates (Fig. 3e). This indicates the requirement for ≥3 Gβγ molecules to open the channel, corroborating the WTM model²⁴ (Figs. 1a, 3e). Therefore, we analyzed the dose-response data using the WTM model version adjusted for real-cell conditions¹⁵ (Fig. 3e, and Supplementary Fig. 8a #2, Methods Eq. 5) and, for comparison, the familiar but mechanistically less informative Hill equation (Eq. 4). We added to the equations a constant component (c) corresponding to I_basal. To convert the two-dimensional surface density to concentration, we used a standard procedure^24,35,50 assuming a submembrane 10 nm thick interaction volume.

**Fig. 3: Coexpressed Gβ·_YFPGγ activates single GIRK2 channels with low-µM apparent affinity.**

Fitting the data with the WTM model (Fig. 3e) yielded cooperativity factor for each successive Gβγ binding (µ) of 0.44 and dissociation constant (K_d) of 44 Gβγ µm⁻² (7.4 µM). Fixing µ=0.3 as in Touhara et al.¹⁵, yielded a K_d of 17.3 µM, and Hill equation fit yielded a K_d of ~4 µM (Fig. 3f). This is much lower than the 300 µM measured in bilayers even at saturating [Na⁺] of >20 mM²⁴, as highlighted with simulated dose-response curves in Fig. 3g.

Similar K_d values were obtained for whole-cell currents of GIRK2 or HA-tagged GIRK2_HA (which is activated by Gβγ like GIRK2^33,34, Supplementary Fig. 1a). Fitting with WTM model (with fixed µ, reducing the number of free parameters) yielded K_d of ~11 µM with µ=0.44 and ~31 µM with µ=0.3 (Fig. 4f; and Supplementary Table 6). GIRK2_WT and truncated GIRK2 (as used in lipid bilayers) showed similar Gβγ sensitivity (one experiment; Supplementary Fig. 5).

**Fig. 4: GIRK2 and dCT-truncated GIRK1 show lower apparent affinity to Gβγ than GIRK1/2.**

GIRK1/2 vs. GIRK2: higher apparent affinity to Gβγ and the role of Gβγ docking to GIRK1

Heterologously expressed GIRK1/2 has a high, Gβγ-dependent I_basal, contrasting the smaller, Gβγ-independent I_basal of homotetrameric GIRK2^34,51,52. Gβγ recruitment³³ and high I_basal of GIRK1/2 and GIRK1/4 require an intact G1-dCT^34,52,53, suggesting that Gβγ docking at GIRK1 increases the local concentration of Gβγ around GIRK⁴. We hypothesized that this may also render higher apparent Gβγ affinity for GIRK1/2 compared to GIRK2.

We previously observed GIRK1/2 activation by expressing Gβγ at relatively low densities (5-50 µm⁻²)³⁵. Here, we compared activation of GIRK2 and GIRK1/2 by Gβ·_YFPGγ in the same experiment (Fig. 4a–c). Surface levels of YFP-Gγ and, subsequently, GIRK currents were measured in individual intact oocytes. Fitting these data with the WTM model revealed a significant difference between K_d of GIRK2 and GIRK1/2 (45 and 9 µM, respectively, with µ=0.3, p = < 0.0001; Fig. 4b, and Supplementary Fig. 7a). Similar K_d values were obtained for data grouped according to the RNA dosage (Fig. 4c). On average, the K_d of GIRK1/2 was about 6-fold lower than GIRK2 ( ~ 5.5 µM vs. ~31 µM, p = 0.027, Fig. 4e, f, Supplementary Table 6). We also observed an ~8-fold difference in K_d of GIRK2-CFP and GIRK1/2-CFP (Supplementary Fig. 6). GIRK1/2 also exhibited the expected higher I_basal than GIRK2. The basal current fraction (c) was higher in GIRK1/2 ( ~ 0.26) than in GIRK2 (0.02-0.03; Fig. 4e, f).

To investigate the role of Gβγ-anchor, we compared the Gβγ dose-dependence of GIRK1/2 to GIRK1_∆dCT/2. GIRK1_∆dCT/2 lacks the Gβγ-anchor, does not recruit Gβγ and has a reduced I_basal³³. Remarkably, the K_d of GIRK1_∆dCT/2 was 9-fold higher compared to GIRK1/2 (Fig. 4d; p = 0.0003) and 3.8-fold higher in another experiment (Supplementary Fig. 7c; p = 0.0009). Thus, GIRK1’s Gβγ-anchor contributes to the high apparent Gβγ affinity of GIRK1/2.

We added the c parameter to the original WTM model to account for I_basal. Instead of fitting c, I_basal can be mechanistically explained and calculated using algorithms utilizing I_basal, I_evoked and I_βγ to estimate basal Gβγ and Gα in GIRK1/2 microenvironment^14,35. We compared the modified WTM (concerted cooperative), the graded contribution (channel opens with one Gβγ and sequential Gβγ binding progressively increases P_o^9,54), and two non-cooperative models (Fig. 4g, and Supplementary Methods, Supplementary Fig. 8). With each model, we calculated basal Gα, Gβγ and I_basal for a range of K_d values, and subsequently simulated dose-response curves for expressed Gβγ with µ=0.3. Both cooperative models matched the experimental data with K_d between 1-10 µM (Fig. 4g). Expectedly, the non-cooperative models predicted lower K_d. The cooperative models also provided stable estimates of basal Gα and Gβγ across a wide K_d range, 0.1-30 µM (Supplementary Fig. 8).

The interactions of Gβγ with the PM and the channel are reversible. Therefore, we expected that removing the cytosolic Gβγ reserve by excising a membrane patch into a Gβγ-free solution would reduce the PM- and GIRK-associated Gβγ, deactivating GIRK channels. We anticipated slower deactivation in channels with a high-affinity Gβγ-anchor.

To test this hypothesis, we recorded Gβγ-activated channels in cell-attached patches and then excised them into an ATP and Na⁺-containing bath solution (Fig. 5a–d). GIRK1/2 activity decayed (deactivated) slowly, with 30-50% persisting after 5 minutes (Fig. 5a, d). The decay followed a single exponent with a time constant (τ) of >2 min and a non-deactivating fraction (C) of 0.34. In contrast, GIRK2 and GIRK1_ΔdCT/2 exhibited faster and more complete decay (Fig. 5b–d, f). Excising patches into an ATP-free solution, which could deplete PIP₂ in the PM⁵⁵, had a minimal impact on GIRK2 and GIRK1_ΔdCT/2 decay, and slightly affected GIRK1/2 (Fig. 5e, f). This suggests that GIRK deactivation is mainly governed by the depletion of Gβγ associated with or surrounding the channel, rather than PIP₂ depletion.

**Fig. 5: Different patterns of deactivation of GIRK2 and GIRK1/2 after patch excision and the role of G1-dCT.**

G1-NT and G1-dCT form a Gβγ-binding site and contribute to channel’s interaction with Gγ’s prenylation tail, Gγ_prenyl

Although deleting G1-dCT thwarts Gβγ binding, G1-dCT alone does not strongly bind Gβγ⁵⁶, indicating that the Gβγ-anchor includes additional Gβγ-binding segments³³. To identify them, we scanned arrays of overlapping peptides covering the cytosolic domains of GIRK1 and GIRK2 (G1NC, G2NC) for His-Gβγ binding (Figs. 1e, 6a–c, and Supplementary Fig. 9). Scanning revealed three Gβγ-binding segments mainly overlapping the C1 and C3 segments from previous biochemical studies^28,56. Two segments fully (in GIRK2) or partially (in GIRK1) overlapped the Gβγ-binding amino acid (a.a.) clusters from the crystallized GIRK2/Gβγ complex⁵ (Fig. 6d). Additionally, Gβγ bound to segments in G1-NT (a.a. ~20-50), parts of G1-dCT (a.a. ~390-440 and ~485-501), and G2-NT and G2-dCT.

**Fig. 6: Peptide array scanning for Gβγ binding sites in the cytosolic domains of GIRKs.**

If a separate GIRK1’s Gβγ-binding segment combines with G1-dCT to form the Gβγ-anchor, deleting it from G1NC should reduce Gβγ binding. We used prenylated His-Gβγ to pull-down the full-length ivt G1NC or G1NC with specific segment deletions, and a fusion protein of G1-NT and G1-dCT, G1NdCT (Fig. 7). Gβγ binding was unaffected by the deletion of internal segments C1-C3 and tended to decrease after the deletion of G1-NT (G1CT construct). G1-dCT and G1-NT showed weak and negligible Gβγ binding, respectively. However, both G1NdCT and the fusion of N-terminal a.a. 40-84 with G1-dCT, G1N(40-84)dCT, strongly bound Gβγ, suggesting that the GIRK1’s Gβγ-anchor comprises G1-dCT and part(s) of G1-NT.

**Fig. 7: Fused G1-NT and G1-dCT of GIRK1 form a high-affinity Gβγ-binding site.**

We conducted coarse-grain (CG) and all-atom molecular dynamics (MD) simulations to further investigate the involvement of G1-NT, G1-dCT and Gγ_prenyl in GIRK-Gβγ interactions. These elements are missing from the available high-resolution structures. We modeled full-length and truncated G1NC and G2NC tetramers complexed with Gβγ using AlphaFold3 and manually added the prenylation tails (Fig. 8, and Supplementary Fig. 10a, Supplementary Table 11). The initial CG system included four Gβγ bound to a G1NC or G2NC tetramer without the PM and bulk Gβγ in the cytosol. MD simulations accurately captured the two Gβγ-interacting surfaces from the GIRK2-Gβγ crystal structure⁵ and predicted additional Gβγ-binding segments, most of which showed excellent (in G2NC) or considerable (in G1NC) agreement with peptide arrays (Figs. 6d, 8b, and Supplementary Fig. 10b), lending credibility to the combined analysis. Further analysis revealed that Gγ_prenyl spent 100% of the simulation time interacting with G1NC, mainly with the beginning of G1-NT, as compared to only 6.4% ± 1.3 with G2NC (p = 0.039; Fig. 8c, d; Supplementary Tables 12, 14). Gγ_prenyl likely accounts for most of the Gβγ binding to the first G1-NT segment predicted by the MD (compare Fig. 8b, c), explaining the poor Gβγ labeling of a.a. 1-25 in peptide array overlays, where solid support-spotted peptides may be less accessible to Gγ lipid moiety. Additionally, Gγ_prenyl also interacted with hydrophobic a.a. in the C-terminus of Gβ (Supplementary Table 13). Backmapped atomistic simulations of G1NC yielded results consistent with the CG simulations, further supporting the robustness of the approach (Supplementary Fig. 10c). In simulations where PM was included, Gγ_prenyl interactions remained stable, supporting the robustness of the binding site predictions (Fig. 8d). Our simulations started with Gβγ pre-bound to the channel; fully ab initio simulations would require significantly longer sampling but could potentially reveal additional membrane interactions of Gγ_prenyl. However, such analyses are beyond the scope of the current study.

**Fig. 8: MD simulations corroborate the role of G1-NT and G1-dCT in interactions with Gβγ and the prenylation tail, Gγ_prenyl.**

Deleting G1-dCT abolished Gγ_prenyl interaction with the remainder of G1NC, G1NC_ΔdCT (p = 0.0003), whereas truncated G2NC retained Gγ_prenyl interaction (Fig. 8c, and Supplementary Table 12). Remarkably, Gγ_prenyl interaction with the first segment of G1-NT was lost upon G1-dCT deletion (Fig. 8c), reinforcing the idea that G1-NT and G1-dCT form a Gβγ-binding unit. MD simulations also revealed details of the GIRK1’s NT-dCT structural unit, with segments of a.a. 27-31 (NT) and ~450-460 (dCT) interacting 99% of the simulation time (Fig. 8e, f, Supplementary Data 1). Notably, this NT-dCT unit is not predicted in the G1NC-Gβγ model by AlphaFold but assembles dynamically during the simulation. In support of the important role of the NT-dCT unit for GIRK1 interaction with Gγ_prenyl, we observed a complete loss of G1NdCT-Gβγ binding with the non-prenylated Gγ_C68S (Supplementary Fig. 11).

Discussion

In this study we address two key issues in the GPCR-Gαβγ-GIRK signaling cascade: the Gβγ-GIRK interaction affinity and the subunit-dependent GIRK-Gβγ preassociation. We hypothesized that Gγ prenylation contributes to Gβγ-GIRK interactions and demonstrated that elimination of prenylation thwarts Gβγ interaction with cytosolic domains of GIRK1 and GIRK2 (G1NC and G2NC; Fig. 1). Expectedly, PM targeting was also abolished (Supplementary Fig. 1). However, membrane targeting was not involved in our Gβγ binding assays, performed in membrane-free detergent solutions. The importance of Gγ_prenyl in full channel context in PM is supported by higher GIRK2-Gβγ affinity in intact oocytes (Figs. 3, 4) compared to non-prenylated Gβγ in bilayers²⁴. We conclude that, besides its well-established role in membrane attachment of Gβγ, Gγ prenylation enhances Gβγ-GIRK interaction, as in many other Gβγ interactors^{36,37,38,39,40,41,42}. The mechanism could involve transient interactions of Gγ_prenyl with hydrophobic sites in Gβγ’s partner⁵⁷ or Gβ itself, stabilizing the conformation favoring Gβγ function^38,40,42,58. In support, MD simulations reveal interactions of Gγ_prenyl with both, specific sites in GIRK1 and GIRK2, and C-terminal hydrophobic residues of Gβ (Fig. 8, and Supplementary Fig. 10, Supplementary Tables 12-14).

The dual role of Gγ prenylation complicates the interpretation of in vitro affinity measurements. Measuring GIRK’s K_d in excised PM patches with prenylated Gβγ in bath solution grossly overestimates affinity (K_d = 2-11 nM, Supplementary Table 8) due to Gβγ’s preferential partitioning to the PM. We addressed the challenge of quantitating GIRK activation by Gβγ in intact cells utilizing Xenopus oocytes, which are highly suitable for accurate titration and monitoring of expression and function of membrane proteins. We constructed Gβγ-GIRK dose-response relationships by varying Gβγ expression and measuring surface densities of Gβγ and GIRK responses. Our results support the WTM model^15,24 of collision-coupled, cooperative activation of GIRK2 by four Gβγ molecules. However, our affinity estimates are substantially higher.

K_d estimates rely on accurate calibrations used to measure Gβγ surface levels. We validated our CF calibrations using two molecular calipers, YFP-GIRK1/2 and IRK1-YFP (Fig. 2). These results, along with previous compatibility tests between CF and qWB methods³⁵, enhance confidence in both calibration procedures. Importantly, only prenylated Gβγ dimer reaches the PM and is captured in our measurements of surface Gβγ, irrespective of total prenylated/non-prenylated cellular Gβγ content. The CF approach is advantageous for measuring expression and function of fluorescently labeled proteins in individual, intact cells. Disappointingly, xFP-Gβ constructs poorly activated GIRKs, especially GIRK2, calling for caution in using xFP-labeled Gβ in functional studies. Consequently, in most dose-response experiments we used Gβ·_YFPGγ, which activated GIRKs like Gβγ_WT. When expressing Gβ·_YFPGγ, the surface densities of Gβ and YFP-Gγ were linearly related, but measuring YFP-Gγ might overestimate coexpressed Gβ, and accordingly the K_d, by up to 2.5-fold (Fig. 2). To avoid overinterpretation, we did not apply the YFP-Gγ correction (for measuring YFP-Gγ as a proxy for Gβγ) in our tables and figures.

Even before formal curve fitting, the Gβγ-GIRK2 dose-responses clearly show that only 10 to 150 µm⁻² of free Gβγ is needed for 10% to 80-90% GIRK2 activation in intact oocytes (Figs. 3, 4), much less than the > 1200 µm⁻² predicted by bilayer results²⁴. Applying the 2.5-fold YFP-Gγ correction shifts the activation range to 4–60 µm⁻². We propose that the higher affinity that we find is mainly due to Gγ prenylation. Interestingly, GIRK2’s I_evoked (via m2R) is only 10% of Gβγ-evoked (Supplementary Fig. 1a). Thus, activation of endogenous G_i/o (Gα_i/oβγ) releases 10-15 molecules/µm⁻² of free Gβγ, corresponding to 30–50% of total endogenous Gβγ in oocyte’s PM, ~30 µm⁻² (Fig. 2). Importantly, coexpressing Gα_i3 and Gβγ with m2R yields I_evoked matching I_βγ⁵⁹. Clearly, endogenous G_i/o is insufficient to activate all GIRKs; but m2R can activate all channels when enough Gα_i/oβγ is present.

Comparing K_d for a multistep cooperative reaction is complex, even with the same kinetic model. The K_d derived from dose-response data is interdependent with the Gβγ cooperativity factor µ: higher µ gives a lower K_d. µ is Na⁺-dependent²⁴ but can be considered constant at stable cytosolic [Na⁺]¹⁵. (We consider [Na⁺]_in in oocytes, 10-20 mM, as close to saturating for GIRK2).

Our average K_d estimates for GIRK2 are 11 µM with µ=0.44 (from Fig. 3) and 31 µM with µ=0.3¹⁵. These are likely overestimates, for two reasons. First, Hill and WTM models assume ligand excess over receptors. This is uncommon in cellular protein-protein interactions, leading to ligand depletion and K_d overestimation: more receptors (GIRK) mean less free ligand (Gβγ) per receptor⁶⁰. This is relevant to our whole-cell experiments, where GIRK2 surface density was 17 ± 5 µm⁻² (Supplementary Table 6), comparable to the functional Gβγ range. Second, applying the 2.5-fold YFP-Gγ correction would shift K_d to 4-12 µM, quite close to the most accurate in vitro measurement available for prenylated Gβγ, 0.8 µM (interaction with CT of GIRK4, by surface plasmon resonance)²¹.

Notably, less Gβγ is needed for GIRK1/2; 50 µm⁻² yields full activation (Fig. 4), confirming previous results³⁵. The 10-15 Gβγ molecules/µm² released by GPCR activation would yield I_evoked of about 50% of I_βγ (Fig. 4), consistent with experiments³⁵. Accordingly, GIRK1/2’s apparent K_d from WTM fits is 5-6-fold lower than GIRK2’s. We further analyzed the GIRK1/2 dose-response data by including explicit calculations of Gα and Gβγ needed to produce the observed I_basal and I_evoked^14,35. Across a broad K_d range (0.1 to 10 µM), the two cooperative models (Fig. 4g) predicted that both I_basal and I_evoked could be generated by physiologically relevant amounts of 1-2 Gα and 3-4 Gβγ per channel (Supplementary Fig. 8). This corresponds to less than 40 µm⁻² of Gβγ assuming physiological densities of GIRKs (2-10 µm⁻²)^16,26.

GIRK1’s Gβγ docking site (anchor) emerges as the major factor determining the higher affinity of GIRK1/2. This is suggested by (i) the 4-9-fold affinity drop in GIRK1_ΔdCT/2, which lacks the main part of the anchor, G1-dCT³³ (Fig. 4); (ii) the fast deactivation after patch excision of GIRK1_ΔdCT/2, mirroring GIRK2, indicating faster Gβγ dissociation (Fig. 5). These results, along with the preservation in GIRK1_ΔdCT of Asn-217 that renders GIRK1 Na⁺-insensitive⁶¹, imply a minor role for the differences in Na⁺-dependence of Gβγ affinity in GIRK1 and GIRK2²⁵ in our experiments. The anchor probably increases the apparent affinity through local enrichment of Gβγ (see below).

We proposed that Gβγ-anchors are distinct from the Gβγ-binding “activation” sites, which induce channel opening⁴ and are located at the interface between core-CTs of two adjacent GIRK subunits^5,62. Removal of G1-dCT preserves maximal Gβγ activation and P_o but eliminates Gβγ recruitment and high I_basal³³, suggesting functional separation of docking and activation. Structural separation is suggested by strong Gβγ binding to G1NC that persists after removing major components of the activation site (C1-C3) and even the whole core-CT, leaving only the fused NT and dCT (G1NdCT) (Fig. 7). Thus, the anchor dominates the overall Gβγ affinity of GIRK1’s cytosolic domain and does not include elements from core-CT. Both G1-NT and G1-dCT bind Gβγ^56,63 (Fig. 6) but much weaker than their fusion protein, G1NdCT (Fig. 7). These results suggest that the Gβγ-anchor is formed jointly by G1-NT and G1-dCT. Interestingly, truncation of G2NC did not significantly reduce Gβγ binding, and the functional impact was minor (Fig. 8c, and Supplementary Figs. 2, 5). However, adding G1-dCT to GIRK2 increased I_basal and conferred Gβγ recruitment³³, suggesting that G2-NT may form Gβγ anchors with G1-dCT.

Peptide array scan and MD simulations provide additional insights. Both approaches identify known Gβγ-binding sites in core-CT, and NT and dCT Gβγ-binding sites in GIRK1 and GIRK2. Our MD analysis used AlphaFold-models including unstructured but essential elements absent from crystal structures: Gγ_prenyl and GIRKs’ NT and dCT. Despite the low-confidence of AlphaFold predictions for some of these elements, MD calculates interactions based on physical parameters and can capture dynamic interactions even if the initial structure is uncertain. Importantly, congruent with experimental results (Fig. 7, and Supplementary Fig. 11), the simulations reveal a dynamically arising structural unit formed by G1-NT and G1-dCT, and extensive interactions of Gγ_prenyl with G1NC, particularly G1-NT, and some with G2NC (Fig. 8). Remarkably, Gγ_prenyl–G1-NT interaction is lost, and Gβγ–G1-NT interaction is reduced after deleting G1-dCT, although G1-dCT itself barely interacts with Gγ_prenyl. These results corroborate the idea that the Gβγ-anchor is a standalone structural and functional unit formed by G1-NT and G1-dCT, with G1-dCT essential for its integrity. Notably, Gγ assists Gβ in GIRK activation^49,58. Gγ_prenyl-anchor interaction may also be involved, since removing G1-dCT or Gγ’s C-terminal region, which includes the prenylation site, eliminates Gγ’s enhancing effect^49,58.

Figure 9 summarizes our view of Gβγ-GIRK2 vs. Gβγ-GIRK1/2 interactions, gating, and the anchor’s role. The dynamic equilibrium between channel-bound, membrane-associated and cytosolic Gβγ determines the local Gβγ concentration in channel’s microdomain. Free Gβγ can reversibly partition from the cytosolic reserve to the PM, activating GIRKs. Comparing K_d for GIRK1/2 activation by Gβγ in whole oocytes (Fig. 4f) and excised patches⁴³ yields a Gβγ PM/cytosol partition coefficient between 140 and 425 (Supplementary Fig. 12), close to earlier estimates of ~300⁶⁴.

**Fig. 9: Differences between GIRK2 and GIRK1/2 in their interaction and gating by Gβγ.**

Our findings confirm that GIRK2 is gated through collision-coupling with Gβγ, cooperative Gβγ binding, and concerted activation by four Gβγ occupying all activation sites (Fig. 9a), consistent with the WTM model^15,16,24. However, in intact Xenopus oocytes (with [Na⁺_in] between 10-20 mM), the Gβγ-GIRK2 affinity is significantly higher than the bilayer estimates, primarily due to Gγ prenylation, which enhances Gβγ functionality and interaction with GIRKs. The high affinity guarantees efficient G_i/o-GIRK2 signaling without the need for obligatory hotspots to account for physiological response (although we cannot exclude hotspots or crowding in oocyte PM, which could upshift our K_d estimates).

In distinction, GIRK1/2 operates within a more complex dynamic system featuring two kinds of binding sites, docking (Gβγ-anchors) and activating. The anchor, formed jointly by G1-NT and G1-dCT, is functionally and topologically separate from the activation sites. The similarity of K_d and deactivation rates in GIRK1_ΔdCT/2 and GIRK2 indicates that the activation sites in GIRK2 and GIRK1/2 have similar Gβγ affinities. If the anchor does not participate in channel opening, how does it increase the apparent affinity? We propose that this occurs by local enrichment of Gβγ around GIRK1/2 due to Gβγ recruitment⁴, through kinetic scaffolding-like mechanisms^65,66,67, functionally equivalent to dynamic preassociation. The increased local Gβγ concentration, in excess over Gα, leads to partial activation sites’ occupation and high I_basal^33,35. Moreover, the added Gβγ will bind to the subsequent (unoccupied) sites with higher affinity due to cooperativity, explaining the leftward shift in GIRK1/2’s Gβγ dose-response curve. Added efficiency could arise if Gβγ’s binding surfaces for docking and activation sites are non-overlapping, allowing the docked Gβγ to repeatedly contact the nearby activation site before Gβγ dissociation from the anchor. Mapping the anchor-Gβγ interface is a challenge for the future.

What is the role of Gα? Gα_i/o interacts with GIRKs and was hypothesized to dock the Gα_i/oβγ heterotrimer to GIRKs^43,68,69. However, the affinity of Gα to GIRK1 is lower than Gβγ⁷⁰. Importantly, binding of Gα_i to G1NC is enhanced by added Gβγ, suggesting that the heterotrimer is docked via Gβγ^4,34,71. Both Gβγ-dependent Gα_i3-GIRK1 interactions and the speed and amplitude of I_evoked are maximized when both G1-NT and G1-dCT are present^29,52,53,71, indicating that the NT-dCT anchor is involved in docking the heterotrimer (Fig. 9b). The stoichiometry of anchor-associated Gαβγ and Gβγ in cells likely varies with GIRK1/x density, constitutive GPCR activity, and other factors⁴.

Method

Ethical approval and Xenopus oocytes handling

Experiments were approved by Tel Aviv University Institutional Animal Care and Use Committee (permits #01-20-083 and TAU-MD-IL-2411-174-3). Maintenance and surgery of female frogs were done as described⁴³. Female frogs, aged 1.5-5 years, were purchased from Xenopus 1 Corp. (Dexter, MI, USA) and kept in dark colored plastic tanks at 20 ± 2 °C at 10/14-hour light-dark cycle. During surgeries, frogs were anesthetized with a 0.25% Tricaine methanesulfonate (MS-222, Sigma-Aldrich #886-86-2) solution, and parts of ovary were removed through a small abdominal incision. Oocytes were defolliculated with collagenase in Ca²⁺ free ND96 solution (in mM: 96 NaCl, 2 KCl, 1 MgCl₂, 5 HEPES, pH 7.5). 2 hours later oocytes were washed with NDE solution (in mM: 96 NaCl, 2 KCl, 1 MgCl₂, 1 CaCl₂, 5 HEPES, 2.5 mM sodium pyruvate, 50 mg/ml gentamycin, pH 7.5) and left in NDE for 2-24 hours before injection. Oocytes were injected with 50 nl RNA using microinjection pipette (Drummond Scientific, Broomall, PA, USA) and incubated at 20 °C for 72 hours for two-electrode voltage clamp, or 48-72 hours for single-channel patch clamp experiments.

DNA constructs, RNA, antibodies

DNA constructs encoding the proteins used are summarized in Supplementary Table 9. Antibodies are described in relevant sections of the Methods and summarized in Supplementary Table 10. All antibodies were commercially available and validated by vendors. In each experiment involving detection of a protein in a sample derived from living cells, the specificity was validated by the absence of signal in cells not expressing the protein under study, and by comparison with purified recombinant proteins such as Gβ.

Gβγ stands for Gβ₁γ₂ throughout the paper. All DNA constructs used to produce RNA were inserted in vectors containing 5’ and 3’ untranslated sequences of Xenopus β-globin (pGEM-HE, pGEM-HJ or pBS-MXT)⁷¹. New constructs were prepared using standard PCR-based procedures (see Supplementary Data 2 for list of primers) and fully sequenced. We used the mouse isoform GIRK2A, which is 11 a.a. shorter than the longer isoform (mouse and human) not studied here, which includes a PDZ-binding consensus sequence at the dCT⁷². The truncated GIRK2 construct (GIRK2_trunk) was prepared by deleting a.a. 1–51 and 381–414 from the GIRK2A construct by PCR. G2NC_trunc was prepared by deleting, from G2NC, of the same regions. YFP-Gγ_C68S was prepared through single-nucleotide mutation of YFP-Gγ. Myristoylated Gβ₁ (myr-Gβ) was created by adding the myristoylation signal (the first 15 aa of Src added to the N terminus of Gβ₁)³². GIRK2-CFP was created by fusing CFP_A207K to the CT of GIRK2 via a Ser-Arg linker. IRK1-YFP and IRK1-CFP were created by fusing YFP_A207K and CFP_A207K, respectively, to the CT of IRK1 via a Lys-Leu linker⁷³. N-terminally Split Venus labeled Gβ₁ (SpV-Gβ) and N-terminally Split Venus labeled Gγ₂ (SpV-Gγ)⁵⁹ were subcloned into pGEM-HJ. G1NdCT (the fused cytosolic G1-NT and G1-dCT), G1N(1-40)dCT (the first 40 a.a. of G1-NT fused to G1-dCT), G1N(40-84)dCT (the last 44 a.a. of G1-NT fused to G1-dCT), Sumo-G1NT and Sumo-G1dCT (Sumo fused to G1-NT or G1-dCT). In all cases the fusion was via the 8-a.a. linker, QSTASQST. The Sumo construct used here was a truncated version of human Sumo 2 protein (a.a. 3-95; PDB: 5ELU_B).

RNAs were transcribed in vitro as described⁴³. The amounts of injected RNAs varied according to the experimental design. For whole-cell electrophysiology experiments we used, in ng/oocyte: 0.01-1 of GIRK1 or YFP-GIRK1, 0.2-10 GIRK2, 0.2-10 Gβ, 0.04-2.5 Gγ, 0.08-5 YFP-Gγ. Equal amounts of GIRK1 and GIRK2 RNAs were injected to express GIRK1/2 channels. In all experiments where several Gβγ expression levels were tested, the ratio of Gβ:Gγ RNA was kept constant: for Gβ:Gγ, the RNA ratio was 5:1 or 2.5:1, and for Gβ:YFP-Gγ the ratio was 2:1 or 2.5:1. For single channel patch clamp, the injected RNAs (in ng/oocyte) were: 0.005-0.01 IRK1-CFP; for GIRK2 alone, 0.02-0.05; for GIRK1/2_HA, GIRK1 0.01-0.02 of GIRK1 with 0.01-0.02 GIRK2_HA. In the experiments of Fig. 5, we injected, in ng/oocyte: GIRK2 alone, 0.2-0.5; GIRK1/2, 0.02-0.05 of GIRK1 and 0.01-0.025 of GIRK2; for GIRK1ΔdCT/2, 0.02-0.05 of GIRK1ΔdCT and 0.01-0.025 of GIRK2. In all patch clamp experiments with Gβγ-activated GIRKs, we injected 5 ng Gβ₁ and 1-2 ng Gγ₂ RNA, and 25-50 ng/oocyte of the GIRK5 antisense oligonucleotide³⁵ to prevent the formation of GIRK1/5 channels.

Gβγ expression and purification

His₆-Gβγ and His₆-Gβγ_C68S were purified essentially as described⁷⁴. For full details see Supplementary Methods. Gβ₁ and Gγ₂ were expressed in Trichoplusia ni (T.ni) cells. The non-prenylated His₆-Gβγ_C68S was extracted from the soluble fraction of the cells’ homogenate and His₆-Gβγ_WT from the membrane fraction, which ensures that the final purified protein is >95% prenylated³⁹. Protein purity was analyzed using SDS-PAGE and by Western blot using anti-Gβ₁ and anti-His tag antibodies (Supplementary Table 10).

Electrophysiology

Whole-cell GIRK currents were measured using standard two-electrode voltage clamp at 20–22 °C using GeneClamp 500B amplifier (Molecular Devices, Sunnyvale, CA, USA) and digitized using Axon Digidata 1440a using pCLAMP software (Molecular Devices). Agarose cushion microelectrodes were filled with 3 M KCl, with resistances of 0.1–1 MΩ³⁴. GIRK currents were measured in either low-[K⁺] solution ND96 (same as Ca²⁺-free but with 1 mM CaCl₂) or high-K solution with 24 mM [K]_out (in mM: 24 KCl, 72 NaCl, 1 CaCl₂, 1 MgCl₂ and 5 Hepes). In experiments of Fig. 1, to maximize GIRK2’s I_basal, we used a 96 mM high-[K]_out solution (in mM: 96 KCl, 2 NaCl, 1 CaCl₂, 1 MgCl₂ and 5 Hepes). Net GIRK currents (I_basal and I_βγ) were determined by subtraction of currents recorded in presence of 1-2.5 mM Ba²⁺ that blocked GIRK currents. The pH of all solutions was 7.5–7.6. Cell-attached patch clamp recordings were performed at 20–23 °C, using borosilicate glass pipettes with resistances of 1.5–3.5 MΩ. The electrode solution contained (in mM): 146 KCl, 2 NaCl, 1 CaCl₂, 1 MgCl₂, 10 Hepes and 1 GdCl₃ (pH 7.6). Bath solution contained (in mM): 146 KCl, 2 MgCl₂, 6 NaCl, 10 Hepes and 1 EGTA (pH 7.6). Block of stretch-activated channels by GdCl₃ was confirmed by recording currents at +80 mV. Single channel currents were recorded at −80 mV in cell-attached patches with the Axopatch 200B amplifier (Molecular Devices) at −80 mV, filtered at 2 or 5 kHz and sampled at 10 or 25 kHz.

Giant membrane patches (GMPs)

GMPs were prepared and imaged as described⁵⁹. Oocytes were devitellinized using tweezers in hypertonic solution (in mM: 6 NaCl, 150 KCl, 4 MgCl2, 10 Hepes, pH 7.6). The devitellinized oocytes were transferred onto a Thermanox^TM coverslip (Nunc, Roskilde, Denmark) and immersed in Ca²⁺-free ND96 solution with their black hemisphere facing the coverslip, for 30–45 min. The oocytes were then suctioned using a Pasteur pipette, leaving a GMP attached to the coverslip, with the cytosolic part facing the medium. The coverslip was washed thoroughly with fresh ND96 solution, and fixated using 4% formaldehyde for 30 min. Fixated GMPs were immunostained in 5% milk in PBS and non-specific binding was blocked with Donkey IgG 1:200 (Jackson ImmunoResearch, West Grove, PA, USA). Primary rabbit anti-Gβ (1:200; Santa Cruz, SC-378 or GeneTex, GTX114442) was applied for 45 min at 37 °C either alone or with blocking peptide supplied with the antibody. Then DyLight549 or DyLight® 650-conjugated anti-rabbit secondary antibodies (KPL) were applied at 1:300 dilution for 30 min at 37 °C, washed with PBS and mounted on a slide for visualization. Immunostained slides were kept at 4 °C in the dark.

Preparation of whole oocyte lysates and separated plasma membranes for pull-down and WB

Lysates from whole nucleus-free oocytes were prepared as described⁴³. 6–10 oocytes were homogenized on ice (20 mM Tris, pH 7.4, 5 mM EGTA, 5 mM EDTA, 100 mM NaCl) containing Roche Complete Protease Inhibitors Cocktail (Merck 11697498001, 1 tablet/100 ml, pH=7.5), 6 µl buffer/oocyte. Nucleus and yolk were removed by centrifugation (1000×g, 15 minutes, 4 ^oC). Supernatant was stored in aliquots corresponding to two oocytes at −80 ^oC. Manually separated oocytes’ PMs for qWB have been prepared as described³⁵. PMs together with the vitelline membranes (extracellular collagen-like matrix) were manually separated from the rest of the oocyte (“cytosol”) with fine forceps, after a 5–15 min incubation in a low osmolarity solution (5 mM NaCl, 5 mM HEPES, and protease inhibitor as above. PMs of ~20 oocytes were pooled for each sample (lane on gel).

Pull-down assays, autoradiograms and WB

Pull-down binding experiments were performed as described³³. For full description see Supplementary Methods. Briefly, pull-down was done with in vitro translated (ivt) [³⁵S]methionine-labelled proteins prepared in rabbit reticulocyte lysate, or unlabeled proteins from whole-oocyte lysates, with ~2 µg of either purified His-Gβγ_WT or purified His-Gβγ_C68S, in 300 μl of the incubation buffer (in mM: 150 KCl, 50 Tris, 0.6 MgCl₂, 1 EDTA, 0.1% Lubrol or 0.5% CHAPS; pH 7.4), followed by 60 min incubation and then addition of Ni-NTA Resin affinity beads and imidazole for 30 min. After repetitive washing, His-Gβγ and bound material were eluted with 250 mM imidazole and subjected to SDS-PAGE. [³⁵S]Methionine-labeled proteins were detected by autoradiography and unlabeled proteins by WB with the appropriate antibodies, and quantified with ImageJ/Fiji (https://imagej.net/software/fiji/). For expressed Gβ, endogenous Gβ signal from oocytes expressing the channel alone was subtracted from the total signal.

Confocal imaging

Confocal imaging and analysis were performed as described⁷³. See Supplementary Methods for details. Live oocytes were imaged at their animal hemisphere. Giant membrane patches were imaged at their edges to show both the membrane and the background. Net signals were calculated by subtracting the average net signal from uninjected (native) oocytes of the same experiment.

Peptide spot array

Peptide arrays were generated by automatic SPOT synthesis and blotted on a Whatman membrane⁷⁵. N-terminal and C-terminal parts of GIRK1 and GIRK2 were spot-synthesized as 25-mer peptides overlapping sequences, shifted by 5 a.a. along the sequence, using AutoSpot Robot ASS 222 (Intavis Bioanalytical Instruments, Cologne, Germany). The peptides were designed according to human GIRK2 (NCBI: NM_002240.5) (NT: a.a. 1-93, CT: a.a. 193-423) and rat GIRK1 (NCBI: NP_113798.1) (NT: a.a. 1-84, CT: a.a. 183-501). The interaction with spot-synthesized peptides was investigated by an overlay assay. Following blocking of 1 hour at room temperature with 5% BSA in 20 mM Tris and 150 mM NaCl with 0.1% Tween-20 (TBST), 0.016–0.16 μM purified His-Gβγ were incubated with the immobilized peptide-dots, overnight at 4 °C. His-Gβγ was detected by anti-GNB1 antibody (GTX114442) at 1:500 or 1:1000 dilution, and anti-rabbit HRP-coupled secondary antibody (1:40000) incubated with 5% BSA/TBST, and the membrane was imaged using Fusion FX7, as for Western blotting.

Electrophysiological data analysis and surface density calibration

Whole-cell and single-channel data were analyzed using Clampex and Clampfit (pCLAMP suite, Molecular Devices, Sunnyvale, CA, USA). In oocytes expressing the m2 receptor, the fold activation by agonist, R_a, was measured in each cell and defined as

$${R}_{{{\rm{a}}}}={I}_{{{\rm{total}}}}/{I}_{{{\rm{basal}}}}$$

(1)

where I_total=I_basal+I_evoked. R_a = 1 when there is no response to agonist.

The fold activation by Gβγ, R_βγ, was defined as

$${R}_{\upbeta \upgamma }={I}_{\upbeta \upgamma }/{\bar{{\mbox{I}}}}_{{{\rm{basal}}}}$$

(2)

where I_βγ is the net GIRK current in a Gβγ-expressing oocyte, and Ῑ_basal is the average GIRK current in oocytes of control group, that express only the channel, from the same experiment³⁴.

Single channel amplitudes were calculated from Gaussian fits of all-points histograms of 30–90 s segments of the record. The open channel probability (P_o) was estimated from 1–5 min segments of 4–20 min recordings from patches containing one to three channels using a standard 50% idealization criterion³⁵.

The PM density of functional channels was determined from the whole-cell current, I, using the classical equation⁴⁶

$$I={{{\rm{N}}}}_{{ch}}\cdot {i}_{{{\rm{single}}}}\cdot P{{{\rm{o}}}}$$

(3)

where N_ch is the total number of channels in the cell, i_single is the single-channel current and P_o is the open probability. P_o and i_single for Gβγ-activated GIRK1/2 are known, and for GIRK2, GIRK1/2_HA and IRK1-xFP we determined them here (Supplementary Fig. 3, Supplementary Table 3). The surface density, in channels/µm² (µm⁻²) was calculated by dividing N_ch by the membrane surface area of the oocyte⁷⁶, 2 · 10⁷ µm². Protein surface densities were converted to concentrations using the standard procedure based on a submembrane interaction space 10 nm deep. i_single was measured in cell-attached patches in 146 mM [K⁺]_out; whole-cell currents were measured in 24 mM [K⁺]_out. The amplitude translation factor for these solutions was 4.63. The conversion factor from surface density to sub-PM space concentration was 1 µm⁻² = 0.166 µM³⁵. In calculating the surface density of channel-attached YFP (two for YFP-GIRK1/2 and four for IRK1-YFP), we assumed similar levels of fluorescence maturation of channel- and Gβ-attached YFP molecules, therefore no correction for such maturation was made. For CF calibrations with YFP-GIRK1/2 or IRK1-YFP, the linear fit included the zero-fluorescence point (with no expressed channels).

In the analysis of Gβγ dose-response data in intact oocytes, we assumed that the PM level of the GIRK2 channels was not significantly altered by Gβγ, as shown previously^34,59 and confirmed for CFP-GIRK2 (Supplementary Fig. 6). In one experiment we monitored GIRK2HA and observed changes at different doses of Gβγ, and corrected the currents accordingly (Supplementary Table 6). Similarly, coexpression of Gβγ causes no significant changes in PM levels of GIRK1/2 up to 2 ng RNA of Gβ⁵⁹. In most experiments, the maximal GIRK1/2 current was observed already with 1 or 2 ng Gβ RNA. With 5 ng Gβ RNA, a 20–30% decrease in channel expression is occasionally seen⁵⁹. No correction for this potential change has been made.

Modeling, simulation and curve fitting for Gβγ dose-response data

Standard fitting for Gβγ-GIRK dose-response curves with Hill or modified WTM models was done assuming that, in the absence of GPCR simulation, the endogenous G proteins are in the form of heterotrimers. Data were fitted to Hill equation in the following form:

$${I}_{{{\rm{GIRK}}}}=(1-c){I}_{\max }{x}^{{{\rm{nH}}}}/({x}^{{{\rm{nH}}}}+{Kd}^{{{\rm{nH}}}})+{{cI}}_{\max }$$

(4)

where x is the concentration of coexpressed Gβγ ([Gβγ]), I_GIRK is GIRK current, I_max is the maximal GIRK current at saturating concentrations of coexpressed Gβγ, n_H (or nH) is the Hill coefficient, c is a constant component corresponding to I_basal;

or a modified WTM model¹⁵ with the addition of a constant component c:

$${I}_{{{\rm{GIRK}}}}= ((1-c){I}_{\max }{x}^{4}/({{K}_{{{\rm{d}}}}}^{4}{{{{\rm{\mu }}}}}^{6}+{{4K}_{{{\rm{d}}}}}^{3}{{{{\rm{\mu }}}}}^{6}x+{{6K}_{{{\rm{d}}}}}^{2}{{{{\rm{\mu }}}}}^{5}{x}^{2}+{4K}_{{{\rm{d}}}}{{{{\rm{\mu }}}}}^{3}{x}^{3}+{x}^{4})) \\ +{{cI}}_{\max }$$

(5)

where x, c and I_max have the same meaning as in Eq. 4, K_d is the dissociation constant of the first Gβγ binding to the one of the four sites in GIRK molecule, µ is the cooperativity factor for each successive Gβγ binding²⁴ for the specific case of a constant Na⁺ concentration¹⁵. In whole-cell of cell-attached recordings from intact Xenopus oocytes, both intracellular Na⁺ and the membrane PIP₂ can be assumed constant during the experiment. Therefore, in most WTM model fits, we utilized a constant cooperativity factor µ=0.3¹⁵ or µ=0.44 (from Fig. 3). In two experiments with GIRK2 we were able to obtain independent estimates of µ from fit, which were 0.44 and 0.62 (Supplementary Table 6, “free µ”).

To simulate Gβγ activation of GIRK1/2, we tested four kinetic schemes (models) (Supplementary Fig. 8a). First, we calculated the basal available Gβγ and Gα from the experimentally observed I_basal^14,35. For simulation, we constructed systems of differential equations based on these schemes and solved them numerically. See Supplementary Methods for details.

Molecular dynamics simulations

All MD analyses were performed with publicly available software. Systems were built using CHARMM-GUI (accessed July 2024). Simulations were run in GROMACS 2022.3. The Martini Elnedyn22p coarse-grained force field was applied, and atomistic refinements used the Amber14SB force field. Simulations were run using GROMACS 2022.3. Molecular graphics and analyses were performed with VMD 1.9.4a12 (December 2017); trajectories were analyzed with MDAnalysis 0.20.1 implemented in Python 3.7.4 and VMD 1.9.4a12. We also used Python 3.7.4 routines NumPy 1.21.6, pandas 1.3.5, matplotlib 3.1.3, seaborn 0.11.1. Full details and references related to MD simulation are in Supplementary Methods and Supplementary Tables 11–14.

Primary structures of G1NC and G2NC were generated by fusing the NT and CT of human GIRK1 and human GIRK2, respectively (Fig. 8). The heatmaps in Fig. 8 show G412 as the last a.a., which corresponds to G414 of mGIRK2 (Supplementary Fig. 9). Additionally, Gβγ units were incorporated into the sequences. Truncated constructs were the same as G1NC_ΔdCT and G2NC_trunc used in biochemical experiments (Fig. 1e).

Statistical analysis

Statistical analysis was performed using GraphPad Prism (GraphPad, La Jolla, CA, USA). For normally distributed data (by Shapiro-Wilk test), pairwise comparison was done by t-test and multiple comparisons by one-way ANOVA, and data were presented as bar graphs with individual data points and mean ± SEM (except if non-normally distributed data were presented on the same panel, in which case box plots were shown). If the data did not pass the normal distribution test, they were analyzed using Mann-Whitney (pairwise) and Kruskal-Wallis non-parametric ANOVA tests, and data were presented as box plots and individual data points. The boxes represent the 25th and 75th percentiles, the whiskers show the smallest and maximal values, and the horizontal line represents the median. Statistical analysis for differences between dose-response curves for two different GIRK compositions was done on WTM model fits of normalized dose-response data from individual oocytes for two fits (as in Fig. 4b, d), as well on three fits (details in Supplementary Fig. 7).

Graphics

Structures of GIRK2, Gα and Gβγ were drawn with PyMOL (Schrodinger LLC). All final figures were produced with Inkscape (inkscape.org). Molecular graphics and analyses were performed with VMD 1.9.4a12 (December 2017).

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.

Data availability

All data are presented in figures and tables in the main paper and in Supplementary Material. Source Data are provided with this paper. The molecular dynamics (MD) simulation data generated in this study have been deposited in Zenodo (https://zenodo.org/records/17117723)⁷⁷ and are publicly available without restrictions. All materials created in this paper, such as DNA constructs, are fully available upon request. The source data underlying Fig. 1b–d, f, g, Fig. 2a, b, c–f, h, i, Fig. 3, Fig. 4a–d, g, Fig. 5d, e, Fig. 7, Supplementary Figs. 1b, c, e, f, 3c, 4e, f, 6a–d, 7c, 11, and Supplementary Tables 3, 4 are provided as Source Data file. Previously published structures referred to in this study are available from the Protein Data Bank under accession codes 5ELU and 1GP2. Source data are provided with this paper.

Code availability

No custom code was used in this study. All analyses were performed with publicly available software as described in the Methods and Supplementary Methods.

References

Luscher, C. & Slesinger, P. A. Emerging roles for G protein-gated inwardly rectifying potassium (GIRK) channels in health and disease. Nat. Rev. Neurosci. 11, 301–315 (2010).
Article PubMed PubMed Central Google Scholar
Luo, H., Marron Fernandez de Velasco, E. & Wickman, K. Neuronal G protein-gated K⁺ channels. Am. J. Physiol. Cell Physiol. 323, C439–C460 (2022).
Article PubMed PubMed Central Google Scholar
Chiamvimonvat, N. et al. Potassium currents in the heart: functional roles in repolarization, arrhythmia and therapeutics. J. Physiol. 595, 2229–2252 (2017).
Article PubMed PubMed Central Google Scholar
Dascal, N. & Kahanovitch, U. The roles of Gβγ and Gα in gating and regulation of GIRK channels. Int Rev. Neurobiol. 123, 27–85 (2015).
Article PubMed Google Scholar
Whorton, M. R. & MacKinnon, R. X-ray structure of the mammalian GIRK2-βγ G-protein complex. Nature 498, 190–197 (2013).
Article ADS PubMed PubMed Central Google Scholar
Logothetis, D. E., Kurachi, Y., Galper, J., Neer, E. J. & Clapham, D. E. The βγ subunits of GTP-binding proteins activate the muscarinic K⁺ channel in heart. Nature 325, 321–326 (1987).
Article ADS PubMed Google Scholar
Yamada, M. et al. * and brain Gβγ activate muscarinic K⁺ channel through the same mechanism. J. Biol. Chem. 268, 24551–24554 (1993).
Article PubMed Google Scholar
Krapivinsky, G., Krapivinsky, L., Wickman, K. & Clapham, D. E. Gβγ binds directly to the G protein-gated K⁺ channel, I_KACh. J. Biol. Chem. 270, 29059–29062 (1995).
Article PubMed Google Scholar
Ivanova-Nikolova, T. T., Nikolov, E. N., Hansen, C. & Robishaw, J. D. Muscarinic K⁺ channel in the heart. Modal regulation by G protein βγ subunits. J. Gen. Physiol. 112, 199–210 (1998).
Article PubMed PubMed Central Google Scholar
Escriba, P. V., Wedegaertner, P. B., Goni, F. M. & Vogler, O. Lipid-protein interactions in GPCR-associated signaling. Biochim Biophys. Acta 1768, 836–852 (2007).
Article PubMed Google Scholar
Schreibmayer, W. et al. Inhibition of an inwardly rectifying K channel by G-protein α-subunits. Nature 380, 624–627 (1996).
Article ADS PubMed Google Scholar
Nakajima, Y., Nakajima, S. & Kozasa, T. Activation of G protein-coupled inward rectifier K⁺ channels in brain neurons requires association of G protein βγ subunits with cell membrane. FEBS Lett. 390, 217–220 (1996).
Article PubMed Google Scholar
Vorobiov, D., Bera, A. K., Keren-Raifman, T., Barzilai, R. & Dascal, N. Coupling of the muscarinic m2 receptor to G protein-activated K⁺ channels via Gα_z and a receptor-Gα_z fusion protein. Fusion between the receptor and Gαz eliminates catalytic (collision) coupling. J. Biol. Chem. 275, 4166–4170 (2000).
Article PubMed Google Scholar
Berlin, S. et al. A collision coupling model governs the activation of neuronal GIRK1/2 channels by muscarinic-2 receptors. Front Pharm. 11, 1216 (2020).
Article Google Scholar
Touhara K. K. et al. MacKinnon R. Molecular basis of signaling specificity between GIRK channels and GPCRs. Elife 7, e42908 (2018).
Zhang, Y. & MacKinnon, R. Higher-order transient structures and the principle of dynamic connectivity in membrane signaling. Proc. Natl. Acad. Sci. USA 122, e2421280121 (2025).
Article PubMed Google Scholar
Fowler, C. E., Aryal, P., Suen, K. F. & Slesinger, P. A. Evidence for association of GABA_B receptors with Kir3 channels and regulators of G protein signalling (RGS4) proteins. J. Physiol. 580, 51–65 (2007).
Article PubMed Google Scholar
Hein, P. & Bunemann, M. Coupling mode of receptors and G proteins. Naunyn Schmiedebergs Arch. Pharm. 379, 435–443 (2009).
Article Google Scholar
Sungkaworn, T. et al. Single-molecule imaging reveals receptor–G protein interactions at cell surface hot spots. Nature 550, 543–547 (2017).
Article ADS PubMed Google Scholar
Zhang, Y. et al. Higher-order transient membrane protein structures. Proc. Natl. Acad. Sci. USA 122, e2421275121 (2025).
Article PubMed Google Scholar
Doupnik, C. A. et al. Time resolved kinetics of direct Gb₁g₂ interactions with the carboxyl terminus of Kir3.4 inward rectifier K⁺ channel subunits. Neuropharmacol 35, 923–931 (1996).
Article Google Scholar
Yokogawa, M., Osawa, M., Takeuchi, K., Mase, Y. & Shimada, I. NMR analyses of the Gβγ binding and conformational rearrangements of the cytoplasmic pore of G protein-activated inwardly rectifying potassium channel 1 (GIRK1). J. Biol. Chem. 286, 2215–2223 (2011).
Article PubMed Google Scholar
Wang, W., Whorton, M. R. & MacKinnon, R. Quantitative analysis of mammalian GIRK2 channel regulation by G proteins, the signaling lipid PIP₂ and Na⁺ in a reconstituted system. Elife 3, e03671 (2014).
Article PubMed PubMed Central Google Scholar
Wang W., Touhara K. K., Weir K., Bean B. P., MacKinnon R. Cooperative regulation by G proteins and Na⁺ of neuronal GIRK2 K⁺ channels. Elife 5, e15751 (2016).
Touhara K. K., Wang W., MacKinnon R. The GIRK1 subunit potentiates G protein activation of cardiac GIRK1/4 hetero-tetramers. Elife 5, e15750 (2016).
Fernandez-Alacid, L., Watanabe, M., Molnar, E., Wickman, K. & Lujan, R. Developmental regulation of G protein-gated inwardly-rectifying K⁺ (GIRK/Kir3) channel subunits in the brain. Eur. J. Neurosci. 34, 1724–1736 (2011).
Article PubMed PubMed Central Google Scholar
Shea, L. D., Neubig, R. R. & Linderman, J. J. Timing is everything - The role of kinetics in G protein activation. Life Sci. 68, 647–658 (2000).
Article PubMed Google Scholar
Huang, C. L., Slesinger, P. A., Casey, P. J., Jan, Y. N. & Jan, L. Y. Evidence that direct binding of Gβγ to the GIRK1 G protein- gated inwardly rectifying K⁺ channel is important for channel activation. Neuron 15, 1133–1143 (1995).
Article PubMed Google Scholar
Slesinger, P. A., Reuveny, E., Jan, Y. N. & Jan, L. Y. Identification of structural elements involved in G protein gating of the GIRK1 potassium channel. Neuron 15, 1145–1156 (1995).
Article PubMed Google Scholar
Rebois, R. V. et al. Heterotrimeric G proteins form stable complexes with adenylyl cyclase and Kir3.1 channels in living cells. J. Cell Sci. 119, 2807–2818 (2006).
Article PubMed Google Scholar
Riven, I., Iwanir, S. & Reuveny, E. GIRK channel activation involves a local rearrangement of a preformed G protein channel complex. Neuron 51, 561–573 (2006).
Article PubMed Google Scholar
Rishal, I., Porozov, Y., Yakubovich, D., Varon, D. & Dascal, N. Gβγ-dependent and Gβγ-independent basal activity of G protein-activated K⁺ channels. J. Biol. Chem. 280, 16685–16694 (2005).
Article PubMed Google Scholar
Kahanovitch, U. et al. Recruitment of Gβγ controls the basal activity of G-protein coupled inwardly rectifying potassium (GIRK) channels: crucial role of distal C terminus of GIRK1. J. Physiol. 592, 5373–5390 (2014).
Article PubMed PubMed Central Google Scholar
Rubinstein, M. et al. Divergent regulation of GIRK1 and GIRK2 subunits of the neuronal G protein gated K⁺ channel by Gα_iGDP and Gβγ. J. Physiol. 587, 3473–3491 (2009).
Article PubMed PubMed Central Google Scholar
Yakubovich, D. et al. A quantitative model of the GIRK1/2 channel reveals that its basal and evoked activities are controlled by unequal stoichiometry of Gα and Gβγ. PLoS Comput Biol. 11, e1004598 (2015).
Article PubMed PubMed Central Google Scholar
Higgins, J. B. & Casey, P. J. In vitro processing of recombinant G protein γ subunits. Requirements for assembly of an active βγ complex. J. Biol. Chem. 269, 9067–9073 (1994).
Article PubMed Google Scholar
Jian, X. et al. Gβγ affinity for bovine rhodopsin is determined by the carboxyl-terminal sequences of the γ subunit. J. Biol. Chem. 276, 48518–48525 (2001).
Article PubMed Google Scholar
Lukov, G. L. et al. Role of the isoprenyl pocket of the G protein βγ subunit complex in the binding of phosducin and phosducin-like protein. Biochemistry 43, 5651–5660 (2004).
Article PubMed Google Scholar
Iñiguez-Lluhi, J. A., Simon, M. I., Robishaw, J. D. & Gilman, A. G. G. protein βγ subunits synthesized in Sf9 cells. Functional characterization and the significance of prenylation of γ. J. Biol. Chem. 267, 23409–23417 (1992).
Article PubMed Google Scholar
Myung, C. S. & Garrison, J. C. Role of C-terminal domains of the G protein β subunit in the activation of effectors. Proc. Natl. Acad. Sci. USA 97, 9311–9316 (2000).
Article ADS PubMed PubMed Central Google Scholar
Fogg, V. C. et al. Role of the γ subunit prenyl moiety in G protein βγ complex interaction with phospholipase Cβ. J. Biol. Chem. 276, 41797–41802 (2001).
Article PubMed Google Scholar
Akgoz, M., Azpiazu, I., Kalyanaraman, V. & Gautam, N. Role of the G protein γ subunit in βγ complex modulation of phospholipase Cβ function. J. Biol. Chem. 277, 19573–19578 (2002).
Article PubMed Google Scholar
Peleg, S., Varon, D., Ivanina, T., Dessauer, C. W. & Dascal, N. Gα_i controls the gating of the G-protein-activated K⁺ channel, GIRK. Neuron 33, 87–99 (2002).
Article PubMed Google Scholar
Niu Y., Tao X., Touhara K. K., MacKinnon R. Cryo-EM analysis of PIP₂ regulation in mammalian GIRK channels. Elife 9, e60552 (2020).
Mathiharan, Y. K. et al. Structural insights into GIRK2 channel modulation by cholesterol and PIP2. Cell Rep. 36, 109619 (2021).
Article PubMed PubMed Central Google Scholar
Hille B. et al. Ion Channels of Excitable Membranes. Sinauer (2002).
Falkenburger, B. H., Jensen, J. B. & Hille, B. Kinetics of M1 muscarinic receptor and G protein signaling to phospholipase C in living cells. J. Gen. Physiol. 135, 81–97 (2010).
Article PubMed PubMed Central Google Scholar
Knol, J. C., Engel, R., Blaauw, M., Visser, A. J. & van Haastert, P. J. The phosducin-like protein PhLP1 is essential for Gβγ dimer formation in Dictyostelium discoideum. Mol. Cell Biol. 25, 8393–8400 (2005).
Article PubMed PubMed Central Google Scholar
Tabak, G., Keren-Raifman, T., Kahanovitch, U. & Dascal, N. Mutual action by Gγ and Gβ for optimal activation of GIRK channels in a channel subunit-specific manner. Sci. Rep. 9, 508 (2019).
Article ADS PubMed PubMed Central Google Scholar
Runnels, L. W. & Scarlata, S. F. Determination of the affinities between heterotrimeric G protein subunits and their phospholipase C-β effectors. Biochemistry 38, 1488–1496 (1999).
Article PubMed Google Scholar
Leaney, J. L. & Tinker, A. The role of members of the pertussis toxin-sensitive family of G proteins in coupling receptors to the activation of the G protein-gated inwardly rectifying potassium channel. Proc. Natl. Acad. Sci. USA 97, 5651–5656 (2000).
Article ADS PubMed PubMed Central Google Scholar
Wydeven, N., Young, D., Mirkovic, K. & Wickman, K. Structural elements in the Girk1 subunit that potentiate G protein-gated potassium channel activity. Proc. Natl. Acad. Sci. USA 109, 21492–21497 (2012).
Article ADS PubMed PubMed Central Google Scholar
Chan, K. W., Sui, J. L., Vivaudou, M. & Logothetis, D. E. Specific regions of heteromeric subunits involved in enhancement of G protein-gated K⁺ channel activity. J. Biol. Chem. 272, 6548–6555 (1997).
Article PubMed Google Scholar
Sadja, R., Alagem, N. & Reuveny, E. Graded contribution of the Gβγ binding domains to GIRK channel activation. Proc. Natl. Acad. Sci. USA 99, 10783–10788 (2002).
Article ADS PubMed PubMed Central Google Scholar
Sui, J. L., Petit-Jacques, J. & Logothetis, D. E. Activation of the atrial KACh channel by the βγ subunits of G proteins or intracellular Na⁺ ions depends on the presence of phosphatidylinositol phosphates. Proc. Natl. Acad. Sci. USA 95, 1307–1312 (1998).
Article ADS PubMed PubMed Central Google Scholar
Ivanina, T. et al. Mapping the Gβγ-binding sites in GIRK1 and GIRK2 subunits of the G protein-activated K⁺ channel. J. Biol. Chem. 278, 29174–29183 (2003).
Article PubMed Google Scholar
McIntire W. E. et al. A model for how Gβγ couples Gα to GPCR. J. Gen. Physiol. 154, e202112982 (2022).
Peng, L., Mirshahi, T., Zhang, H., Hirsch, J. P. & Logothetis, D. E. Critical determinants of the G protein γ subunits in the Gβγ stimulation of G protein-activated inwardly rectifying potassium (GIRK) channel activity. J. Biol. Chem. 278, 50203–50211 (2003).
Article PubMed Google Scholar
Reddy, H. P. et al. Encephalopathy-causing mutations in Gβ1 (GNB1) alter regulation of neuronal GIRK channels. iScience 24, 103018 (2021).
Article ADS PubMed PubMed Central Google Scholar
Edelstein, S. J., Stefan, M. I. & Le Novère, N. Ligand depletion in vivo modulates the dynamic range and cooperativity of signal transduction. PLOS ONE 5, e8449 (2010).
Article ADS PubMed PubMed Central Google Scholar
Ho, I. H. & Murrell-Lagnado, R. D. Molecular determinants for sodium-dependent activation of G protein-gated K⁺ channels. J. Biol. Chem. 274, 8639–8648 (1999).
Article PubMed Google Scholar
Li, D., Jin, T., Gazgalis, D., Cui, M. & Logothetis, D. E. On the mechanism of the GIRK2 channel gating by phosphatidylinositol bisphosphate (PIP2), sodium, and the Gβγ dimer. J. Biol. Chem. 294, 18934–18948 (2019).
Article PubMed PubMed Central Google Scholar
Huang, C. L., Jan, Y. N. & Jan, L. Y. Binding of the G protein βγ subunit to multiple regions of G protein-gated inward-rectifying K⁺ channels. FEBS Lett. 405, 291–298 (1997).
Article PubMed Google Scholar
Runnels, L. W. & Scarlata, S. F. Regulation of the rate and extent of phospholipase Cβ2 effector activation by the βγ subunits of heterotrimeric G proteins. Biochemistry 37, 15563–15574 (1998).
Article PubMed Google Scholar
Mori, M. X., Erickson, M. G. & Yue, D. T. Functional stoichiometry and local enrichment of calmodulin interacting with Ca²⁺ channels. Science 304, 432–435 (2004).
Article ADS PubMed Google Scholar
Zhong, H. et al. A spatial focusing model for G protein signals. regulator of G protein signaling (RGS) protien-mediated kinetic scaffolding. J. Biol. Chem. 278, 7278–7284 (2003).
Article PubMed Google Scholar
Ross, E. M. Coordinating speed and amplitude in G-protein signaling. Curr. Biol. 18, R777–R783 (2008).
Article PubMed PubMed Central Google Scholar
Kano, H. et al. Structural mechanism underlying G protein family-specific regulation of G protein-gated inwardly rectifying potassium channel. Nat. Commun. 10, 2008 (2019).
Article ADS PubMed PubMed Central Google Scholar
Clancy, S. M. et al. Pertussis-toxin-sensitive Gα subunits selectively bind to C-terminal domain of neuronal GIRK channels: evidence for a heterotrimeric G-protein-channel complex. Mol. Cell Neurosci. 28, 375–389 (2005).
Article PubMed Google Scholar
Mase, Y., Yokogawa, M., Osawa, M. & Shimada, I. Structural basis for modulation of gating property of G protein-gated inwardly rectifying potassium ion channel (GIRK) by i/o-family G protein α subunit (Galphai/o). J. Biol. Chem. 287, 19537–19549 (2012).
Article PubMed PubMed Central Google Scholar
Berlin, S. et al. Two distinct aspects of coupling between Gαi protein and G protein-activated K⁺ channel (GIRK) revealed by fluorescently labeled Gα_i3 protein subunits. J. Biol. Chem. 286, 33223–33235 (2011).
Article PubMed PubMed Central Google Scholar
Marron Fernandez de Velasco, E. et al. GIRK2 splice variants and neuronal G protein-gated K⁺ channels: implications for channel function and behavior. Sci. Rep. 7, 1639 (2017).
Article ADS PubMed PubMed Central Google Scholar
Berlin, S. et al. Gα_i and Gβγ jointly regulate the conformations of a Gβγ effector, the neuronal G protein-activated K⁺ channel (GIRK). J. Biol. Chem. 285, 6179–6185 (2010).
Article PubMed Google Scholar
Dessauer, C. W. & Gilman, A. G. Purification and characterization of a soluble form of mammalian adenylyl cyclase. J. Biol. Chem. 271, 16967–16974 (1996).
Article PubMed Google Scholar
Dema, A. et al. The A-kinase anchoring protein (AKAP) glycogen synthase kinase 3β interaction protein (GSKIP) regulates β-catenin through its interactions with both protein kinase A (PKA) and GSK3β. J. Biol. Chem. 291, 19618–19630 (2016).
Article PubMed PubMed Central Google Scholar
Dascal, N. The use of Xenopus oocytes for the study of ion channels. CRC. Crit. Rev. Biochem 22, 317–387 (1987).
Article Google Scholar
Stary-Weinzinger A., Hofer P. MD simulations of G1NC and G2NC G-beta and G-gamma and prenylation tail interactions. Zenodo, https://doi.org/10.5281/zenodo.17117723 (2025).

Download references

Acknowledgements

This work was supported by grants 1282/18 and 581/22 from Israel Science Foundation (ND), KL1415/14-1 from the Deutsche Forschungsgemeinschaft (EK), R35GM145921 from the National Institute of General Medical Science, National Institutes of Health, W1232 from the Austrian Science Fund (A.S.W., T.F.). The computational results presented have been achieved in part using the Vienna Scientific Cluster (V.S.C.).

Author information

Haritha P. Reddy
Present address: Department of Neuroscience, Faculty of Health and Medical Sciences, University of Copenhagen, Blegdamsvej 3B, 2200, Copenhagen, Denmark

Authors and Affiliations

Department of Physiology and Pharmacology, Faculty of Health and Medical Sciences, Tel Aviv University, Tel Aviv, Israel
Reem Handklo-Jamal, Tal Keren Raifman, Boris Shalomov, Uri Kahanovitch, Galit Tabak, Vladimir Tsemakhovich, Haritha P. Reddy, Debi Ranjan Tripathy, Yoni Haitin & Nathan Dascal
Department of Pharmaceutical Sciences, Division of Pharmacology and Toxicology, University of Vienna, Josef-Holaubek-Platz 2, Vienna, Austria
Patrick Hofer, Theres Friesacher & Anna Stary-Weinzinger
Department of Biochemistry & Molecular Biology, School of Neurobiology, Biochemistry and Biophysics, George S. Weiss Faculty of Life Sciences, Tel Aviv University, Tel Aviv, Israel
Orna Chomsky-Hecht, Debi Ranjan Tripathy & Joel A. Hirsch
Max-Delbrück-Center for Molecular Medicine in the Helmholtz Association (MDC), Berlin, Germany
Kerstin Zuhlke & Enno Klussmann
Department of Integrative Biology and Pharmacology, University of Texas Health Science, Center, Houston, Texas, USA
Carmen W. Dessauer
DZHK (German Centre for Cardiovascular Research), partner site Berlin, Berlin, Germany
Enno Klussmann
Sagol School of Neuroscience, Tel Aviv University, Tel Aviv, Israel
Yoni Haitin, Joel A. Hirsch & Nathan Dascal
The Adelson School of Medicine, Ariel University, Ariel, Israel
Daniel Yakubovich

Authors

Reem Handklo-Jamal
View author publications
Search author on:PubMed Google Scholar
Tal Keren Raifman
View author publications
Search author on:PubMed Google Scholar
Boris Shalomov
View author publications
Search author on:PubMed Google Scholar
Patrick Hofer
View author publications
Search author on:PubMed Google Scholar
Uri Kahanovitch
View author publications
Search author on:PubMed Google Scholar
Theres Friesacher
View author publications
Search author on:PubMed Google Scholar
Galit Tabak
View author publications
Search author on:PubMed Google Scholar
Vladimir Tsemakhovich
View author publications
Search author on:PubMed Google Scholar
Haritha P. Reddy
View author publications
Search author on:PubMed Google Scholar
Orna Chomsky-Hecht
View author publications
Search author on:PubMed Google Scholar
Debi Ranjan Tripathy
View author publications
Search author on:PubMed Google Scholar
Kerstin Zuhlke
View author publications
Search author on:PubMed Google Scholar
Carmen W. Dessauer
View author publications
Search author on:PubMed Google Scholar
Enno Klussmann
View author publications
Search author on:PubMed Google Scholar
Yoni Haitin
View author publications
Search author on:PubMed Google Scholar
Joel A. Hirsch
View author publications
Search author on:PubMed Google Scholar
Anna Stary-Weinzinger
View author publications
Search author on:PubMed Google Scholar
Daniel Yakubovich
View author publications
Search author on:PubMed Google Scholar
Nathan Dascal
View author publications
Search author on:PubMed Google Scholar

Contributions

N.D., D.Y., A.S.W.–conceptualization. R.H., T.K.R., B.S., U.K., T.F., C.W.D., Y.H., J.A.H., A.S.W., D.Y., N.D.–experimental design. R.H., T.K.R., B.S., P.H., U.K., T.F., G.T., V.T., H.R.P., A.S.W., D.Y., N.D.–investigation. R.H., G.T., O.C.H., D.R.T., K.Z., C.W.D., E.K., J.A.H.–resources. R.H., T.K.R., P.H., U.K., T.F., A.S.W., N.D.–data curation. E.K., Y.H., J.A.H., A.S.W., D.Y., N.D.–supervision. C.W.D., E.K., Y.H., J.A.H., A.S.W., N.D.–funding acquisition. R.H., A.S.W., D.Y., N.D.–writing the initial version of the paper. R.H., H.R.P., U.K., E.K., C.W.D., Y.H., J.A.H., A.S.W., D.Y., N.D.—review and editing. All authors read and confirmed the paper.

Corresponding authors

Correspondence to Anna Stary-Weinzinger, Daniel Yakubovich or Nathan Dascal.

Ethics declarations

Competing interests

The authors declare no competing interests.

Peer review

Peer review information

Nature Communications thanks Heidi Hamm who co-reviewed with Montana Young, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available.

Additional information

Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary information

Supplementary Information

Description of Additional Supplementary Files

Supplementary Data 1

Supplementary Data 2

Reporting Summary

Transparent Peer Review file

Source data

Source Data

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and permissions

About this article

Cite this article

Handklo-Jamal, R., Raifman, T.K., Shalomov, B. et al. Live-cell quantitative monitoring reveals distinct, high-affinity Gβγ regulations of GIRK2 and GIRK1/2 channels. Nat Commun 16, 11607 (2025). https://doi.org/10.1038/s41467-025-66730-8

Download citation

Received: 30 March 2025
Accepted: 12 November 2025
Published: 24 November 2025
Version of record: 29 December 2025
DOI: https://doi.org/10.1038/s41467-025-66730-8