Agonist-induced formation of unproductive receptor-G₁₂ complexes

Significance

G protein-coupled receptors (GPCRs) are targeted by a large fraction of approved drugs and regulate many important cellular processes. Association of GPCRs with heterotrimeric G proteins in response to agonist activation is thought to invariably lead to G protein activation. We find instead that G₁₂ heterotrimers can associate with agonist-bound receptors in a manner that does not lead to activation. These unproductive agonist–receptor-G protein ternary complexes sequester G₁₂ heterotrimers and thus inhibit rather than support G₁₂ signaling. These findings reveal a mechanism whereby agonist activation of GPCRs can inhibit as well as promote G protein signaling.

Abstract

G proteins are activated when they associate with G protein-coupled receptors (GPCRs), often in response to agonist-mediated receptor activation. It is generally thought that agonist-induced receptor-G protein association necessarily promotes G protein activation and, conversely, that activated GPCRs do not interact with G proteins that they do not activate. Here we show that GPCRs can form agonist-dependent complexes with G proteins that they do not activate. Using cell-based bioluminescence resonance energy transfer (BRET) and luminescence assays we find that vasopressin V₂ receptors (V₂R) associate with both G_s and G₁₂ heterotrimers when stimulated with the agonist arginine vasopressin (AVP). However, unlike V₂R-G_s complexes, V₂R-G₁₂ complexes are not destabilized by guanine nucleotides and do not promote G₁₂ activation. Activating V₂R does not lead to signaling responses downstream of G₁₂ activation, but instead inhibits basal G₁₂-mediated signaling, presumably by sequestering G₁₂ heterotrimers. Overexpressing G₁₂ inhibits G protein receptor kinase (GRK) and arrestin recruitment to V₂R and receptor internalization. Formyl peptide (FPR1 and FPR2) and Smoothened (Smo) receptors also form complexes with G₁₂ that are insensitive to nucleotides, suggesting that unproductive GPCR-G₁₂ complexes are not unique to V₂R. These results indicate that agonist-dependent receptor-G protein association does not always lead to G protein activation and may in fact inhibit G protein activation.

G protein-coupled receptors (GPCRs) mediate important physiological activities and exert most of their effects through activation of G proteins. In the conventional model of coupling, unliganded receptors are poor recruiters and activators of G proteins, whereas agonist-bound GPCRs take on more active conformations that effectively recruit G protein heterotrimers (1, 2). Productive receptor-G protein association promotes GDP release by stabilizing the nucleotide-free state of the Gα subunit, which in turn allows GTP binding, G protein activation, and downstream signaling (3, 4). According to this model, agonist-dependent GPCR-G protein complex formation is essentially synonymous with G protein activation. Four families of G proteins (G_s, G_i/o, G_q/11, and G_12/13) can be activated, and each leads to a distinct set of downstream signaling outcomes. It is generally thought that selection of G protein subtypes by GPCRs occurs at the receptor-G protein association step, such that receptors interact with and activate cognate G protein subtypes and do not interact with noncognate G protein subtypes. Here we find that agonist-dependent GPCR-G protein association can occur without promoting subsequent G protein activation, thus, whether a G protein subtype is activated can be determined after initial receptor-G protein engagement. Moreover, noncognate G proteins can impede downstream events, perhaps by competing with other intracellular transducers for access to activated receptors. These findings revise the standard model of G protein coupling by indicating that agonist-induced GPCR-G protein association does not always promote G protein activation and may in some circumstances inhibit downstream signaling.

Results

V₂R Interacts with G₁₂ Heterotrimers.

Conventional GPCR-G protein coupling is understood as an allosteric interaction where an agonist-bound active receptor mediates GDP release by stabilizing the nucleotide-free state of an associated Gα subunit (3–5). Receptor complexes with nucleotide-free G proteins are quite transient when guanine nucleotides are present at concentrations similar to those found in cells, but are stabilized when guanine nucleotides are absent. In order to observe allosteric coupling we monitored receptor-G protein association under conditions that allowed us to control both ligand binding to the receptor and nucleotide binding to the G protein. We used bioluminescence resonance energy transfer (BRET) between GPCRs fused to Renilla luciferase (Rluc8) and Gβ₁ and Gγ₂ subunits fused to complementary fragments of Venus fluorescent protein (6–8) to monitor receptor-G protein association. These components and unlabeled Gα subunits were transfected into HEK 293 cells in which most of the endogenous G proteins had been deleted using CRISPR/Cas9-mediated gene editing (9, 10). In order to control nucleotide binding to G proteins, cells were permeabilized and either supplemented with nucleotides or treated with apyrase to remove residual nucleotides.

Because receptor-G protein complexes are transient it can be difficult to detect agonist-induced BRET signals between receptors and G proteins when guanine nucleotides are present (6). For example, arginine vasopressin (AVP) did not detectably increase BRET between vasopressin V₂ receptors (V₂R) and G_s heterotrimers in the presence of GDP (Fig. 1A). In contrast, AVP produced large BRET increases in the absence of nucleotides (Fig. 1A). Stabilization of agonist–receptor-G protein complexes when Gα subunits are nucleotide-free indicates conventional “productive” allosteric coupling and predicts that agonist-bound receptors will promote GDP release and G protein activation under physiological conditions. Nucleotide-free conditions also enhanced AVP-induced BRET between V₂R and G_i1 or G_q heterotrimers (SI Appendix, Fig. S1). These results are consistent with cognate V₂R activation of G_s and G_q heterotrimers (11, 12) and predict some ability to activate G_i1 heterotrimers. In contrast, we observed surprisingly robust agonist-induced BRET between V₂R and G₁₂ heterotrimers in the presence of GDP that was not enhanced by nucleotide depletion (Fig. 1A). These results suggest that AVP-bound V₂R can form complexes with GDP-bound G₁₂ heterotrimers that do not progress to the nucleotide-free state and therefore are not stabilized when GDP is removed (Fig. 1B). As an index of allosteric coupling we divide the increase in BRET produced by agonist in the presence of GDP (ΔBRET_ag) by the increase in BRET produced by the combined effect of agonist and nucleotide depletion (ΔBRET_ag+apy), and refer to this index as the GDP-resistance ratio, or R_GDP. R_GDP values that are less than 1 indicate conventional productive coupling, whereas an R_GDP value of 1 indicates nucleotide-resistant or unproductive coupling. R_GDP for V₂R and G_s was 0.08 ± 0.12 (mean ± SD; n = 6), whereas R_GDP for V₂R and G₁₂ was 1.01 ± 0.06 (n = 6). In contrast to V₂R, we observed more conventional productive coupling of both endothelin A (ET_A) and thromboxane A₂ (TP) receptors with G₁₂ heterotrimers, with R_GDP values of 0.66 ± 0.09 (n = 3) and 0.55 ± 0.09 (n = 6), respectively. Both of these receptors also coupled productively with G_q heterotrimers, with R_GDP values of 0.29 ± 0.08 (n = 3) and 0.11 ± 0.07 (n = 6), respectively (SI Appendix, Fig. S2).

Fig. 1.

V₂R forms GDP-resistant agonist-induced complexes with G₁₂ heterotrimers. (A) BRET between V₂R-Rluc8 and Gαβγ-Venus in the presence or absence of AVP (1 μM), and the presence or absence of GDP. When GDP was absent, apyrase (apy) was added to remove residual nucleotides. AVP-induced BRET to G_s (Left) but not G₁₂ (Right) heterotrimers was enhanced when GDP was absent; **P < 0.005; n.s., not significant (P = 0.58); one-way ANOVA (Sidak’s test) compared to GDP+AVP; n = 6. (B) Cartoon representation of two steps of V₂R-G protein coupling: agonist-induced formation of receptor-G protein complexes, and GDP release. (C) Time course of BRET between V₂R-Rluc8 and Gαβγ-Venus in response to injection of 1 μM AVP, followed by injection of 100 μM GTP in permeabilized cells treated with apyrase (mean ± SEM; n = 4–6). (D) BRET between V₂R-Rluc8 and Gαβγ-Venus as a function of AVP concentration in permeabilized cells expressing either G_s (Left) or G₁₂ (Right) heterotrimers treated with apyrase, GDP (100 μM), or GTP (100 μM). The logEC₅₀ for association with G_s was −6.8 ± 0.5 in apyrase-treated cells, and the logEC₅₀s for association with G₁₂ were −7.5 ± 0.3, −7.4 ± 0.4, and −7.5 ± 0.5 in the presence of apyrase, GDP, and GTP, respectively. Data points represent the change in BRET (ΔBRET) in response to AVP (mean ± SEM; n = 3).

Receptors that couple to one member of a Gα subunit family can usually couple to other members of the same family. Therefore, we examined V₂R coupling to G₁₃ heterotrimers, the other member of the G_12/13 family (13). We found that stimulation with AVP increased BRET between V₂R and G₁₃ heterotrimers in the presence of GDP (SI Appendix, Fig. S3). However, unlike what we observed with G₁₂, these responses were enhanced by nucleotide depletion (R_GDP = 0.67 ± 0.08; n = 4), indicating productive coupling, consistent with weak V₂R-mediated activation of G₁₃ (14). Similar results were obtained with G₁₃ heterotrimers and ET_A (R_GDP = 0.21 ± 0.03; n = 4) and TP (R_GDP = 0.18 ± 0.01; n = 4) receptors (SI Appendix, Fig. S3).

Because nucleotide-resistant V₂R-G₁₂ association was unexpected we performed additional experiments to rule out the possibility that our standard BRET assay was simply detecting an agonist-induced change in V₂R-Rluc8 conformation. We reasoned that if both ET_A and V₂R receptors were able to associate with G₁₂ the two receptors should compete for a common pool of heterotrimers. Indeed, we found that stimulation of unlabeled ET_A receptors inhibited AVP-induced BRET between V₂R-Rluc8 and G₁₂ heterotrimers in intact cells (SI Appendix, Fig. S4A). Conversely, stimulation of unlabeled V₂R receptors inhibited endothelin-1–induced BRET between ET_A-Rluc8 and G₁₂ heterotrimers (SI Appendix, Fig. S4B). Second, we found that stimulation of unlabeled V₂R receptors increased BRET between Gα₁₂-Rluc8 and Gβγ-Venus in intact cells (SI Appendix, Fig. S5 A and B). This increase persisted in permeabilized cells that were treated with apyrase and supplemented with GDPβS to prevent the possibility of heterotrimer activation by residual GTP (SI Appendix, Fig. S5B). This suggests that active V₂R receptors may impose a conformational change in G₁₂ heterotrimers that does not require GTP binding or G₁₂ activation. In contrast, stimulation of unlabeled ET_A and TP receptors decreased BRET between Gα₁₂-Rluc8 and Gβγ-Venus, and these decreases were largely blocked in permeabilized cells when only GDPβS was present (SI Appendix, Fig. S5C). Finally, we found that AVP increased luciferase complementation when a small fragment (SmBit) of Nanoluc was fused to V₂R, and a large fragment of Nanoluc was fused to Gγ₂, and these proteins were coexpressed with unlabeled Gα₁₂ and Gβ₁ (SI Appendix, Fig. S6). These results are consistent with AVP-induced association of V₂R receptors and G₁₂ heterotrimers.

Additional experiments revealed that V₂R recruited G_s and G₁₂ heterotrimers at similar rates (Fig. 1C) and that agonist-induced V₂R-G₁₂ complexes were equally stable in the presence of GDP or GTP (Fig. 1D). Stimulation of V₂R with the agonist oxytocin produced similar responses to AVP, indicating that nucleotide-insensitive V₂R-G₁₂ interactions are not restricted to AVP (SI Appendix, Fig. S7A), and AVP-induced responses were inhibited by the antagonist mozavaptan (SI Appendix, Fig. S7B).

V₂R Does Not Activate G₁₂ Heterotrimers.

The above results suggested that AVP-stimulated V₂R should not activate G₁₂ heterotrimers. To test this prediction we turned to sensitive assays that monitor signaling downstream of G₁₂ activation. We first monitored translocation of full-length p115-RhoGEF and a fragment (amino acids 281–483) of PDZ-RhoGEF from the cytosol to the plasma membrane using bystander BRET assays (15, 16). These proteins bind to activated Gα₁₂ subunits at the plasma membrane to regulate Rho GTPase activity and actin fiber formation (17, 18). ET_A and TP receptors robustly recruited p115-RhoGEF to the plasma membrane in a G₁₂-dependent manner (Fig. 2A). In contrast, stimulation of V₂R receptors failed to recruit p115-RhoGEF and instead decreased the baseline abundance of this reporter at the plasma membrane (Fig. 2A). Similar results were obtained with V₂R and TP receptors and PDZ-RhoGEF recruitment to the plasma membrane (Fig. 2B) and direct recruitment of the RGS homology domain (amino acids 1–246) of p115-RhoGEF, p115-RGS-GFP, to Gα₁₂-RlucII (Fig. 2C). The AVP-induced decrease in p115-and PDZ-RhoGEF at the plasma membrane suggests that active V₂R may sequester G₁₂ heterotrimers, preventing activation by endogenous receptors. Consistent with this suggestion, we found that activation of V₂R could significantly reduce p115-RhoGEF recruitment mediated by activation of ET_A receptors (Fig. 2D). A second sensitive assay of G₁₂ activity is gene transcription driven by activation of the serum response element (SRE) (19). Stimulation of V₂R receptors failed to activate SRE-dependent gene transcription, whereas stimulation of both ET_A and TP receptors could activate SRE in a G₁₂-dependent manner (Fig. 2E). A similar trend was observed with transcription driven by nuclear factor kappa-light-chain enhancer of activated B cells (NFκB; Fig. 2F). These results demonstrate that active V₂R receptors do not detectably activate G₁₂ heterotrimers, even though the two proteins interact in an agonist-dependent manner.

Fig. 2.

V₂R does not activate G₁₂ heterotrimers. (A) Activation of V₂R decreases bystander BRET between p115RhoGEF-Rluc8 and the plasma membrane marker Venus-Kras when G₁₂ is expressed, whereas activation of ET_A and TP receptors increases this signal, indicating association of p115RhoGEF-Rluc8 and active Gα₁₂ at the plasma membrane; *P < 0.05; paired t test compared to mock-transfected control (−); n = 3–4. (B) V₂R activation decreases BRET between PDZ-RhoGEF-RlucII and the plasma membrane marker rGFP-CAAX, whereas TP activation increases this signal (mean ± SEM; n = 3). (C) TP activation increases BRET between two different Gα₁₂-RlucII constructs and p115-RGS-GFP, whereas V₂R activation has no effect on this signal; ***P < 0.0005; n.s., not significant; one-sample t test compared to zero; n = 6. RlucII was fused to Gα₁₂ after amino acids 84 and 136 in the two different probes. Data points in A–C represent the change in BRET (ΔBRET) in response to agonist, and the broken gray line represents zero. (D) Activation of V₂R reduces p115-RhoGEF recruitment mediated by activation of ET_A. Activation of ET_A (ET1; 100 nM) increases BRET between p115RhoGEF-Rluc8 and Venus-Kras, and this response is significantly inhibited when V₂R receptors are coexpressed and activated (AVP; 1 μM); n.s., not significant (P = 0.43); **P < 0.005; one-way ANOVA (Sidak’s test); n = 4. Data points represent the change in bystander BRET (ΔBRET) in response to ET1, and the broken gray line represents zero. (E) Activation of V₂R fails to activate the SRE when G₁₂ is expressed, whereas activation of ET_A and TP receptors increases SRE-driven gene expression; *P < 0.05; n.s., not significant; paired t test compared to (−); n = 5. (F) V₂R receptors fail to activate NFκB-driven gene expression when G₁₂ is expressed; n.s., not significant; paired t test compared to (−); n = 3. Data points in E and F represent luminescence normalized to vehicle-treated controls, and the broken gray line represents one (no change). Agonists were AVP (1 μM), ET1 (100 nM), and U-46619 (10 μM).

V₂R Activates G₁₂ Chimeras and Mutants.

Canonical GPCR-mediated activation of G proteins involves extension of the Gα subunit C terminus (helix 5; H5) into the active receptor core (20, 21). This region of Gα is necessary for productive coupling and is also a key determinant of receptor-G protein selectivity. Therefore, we hypothesized that exchanging the Gα₁₂ C terminus with C-terminal peptides from other Gα subunits might allow productive coupling with V₂R. Indeed, we found that Gα₁₂ chimeras bearing the last 10 amino acids of either Gα_s or Gα_q (Fig. 3A) interacted with AVP-activated V₂R in a GDP-sensitive manner; R_GDP values were significantly less than 1 for G_12s and G_12q heterotrimers and were similar to R_GDP for G_s and G_q heterotrimers (Fig. 3B). G_12s and G_12q chimeras also supported V₂R-mediated translocation of p115-RhoGEF to the plasma membrane, consistent with productive coupling to these heterotrimers and activation of G₁₂ signaling pathways (Fig. 3C). Conversely, we found that replacing the C-terminal peptides of either Gα_s or Gα_q with that of Gα₁₂ (Fig. 3A) dramatically increased R_GDP, indicating much less productive coupling to G_s12 and unproductive coupling to G_q12 (Fig. 3B). We next made point mutations in the Gα₁₂ C terminus to introduce residues with properties shared by the corresponding residues in Gα_s and Gα_q (Fig. 3A). We found that Gα₁₂ mutants with a hydrophobic residue in the −1 position (Q381L and Q381V) still coupled unproductively with V₂R (Fig. 3 D and E). In contrast, Gα₁₂ mutants with a tyrosine in the −4 position (I378Y) coupled productively with V₂R; R_GDP was less than 1 (Fig. 3D), and I378Y supported V₂R-mediated translocation of p115-RhoGEF (Fig. 3E). Similar weak but productive coupling to V₂R was observed when Gα₁₂ was simply truncated by a single amino acid (Δ1; Fig. 3 D and E). These results indicate that the Gα₁₂ C terminus is required for unproductive coupling to active V₂R. Together with the observation that subtle modifications of the Gα₁₂ C terminus overcome the barrier to productive coupling, this result suggests that G₁₂ heterotrimers are likely to interact with active V₂R in a manner that is structurally similar to canonical GPCR-G protein complexes.

Fig. 3.

Role of the Gα₁₂ C terminus in unproductive coupling with V₂R. (A) Amino acid sequences of the distal C terminus of Gα₁₂ (green), Gα_s (blue), Gα_q (black), and the chimeras and mutants used in B–E. Boxes indicate residues whose properties are shared between Gα_s and Gα_q but not Gα₁₂. (B) GDP resistance (R_GDP) of interactions between V₂R-Rluc8 and heterotrimers incorporating the indicated Gα subunits; n.s., not significant; all other groups significantly different from one; P < 0.05; one-sample t test; n = 4–10. (C) Activation of V₂R decreases bystander BRET between p115RhoGEF-Rluc8 and Venus-Kras when G₁₂ is expressed, but increases bystander BRET when G_12s or G_12q are expressed, indicating receptor-mediated activation of these chimeras; all groups were significantly different from mock-transfected controls (−); P < 0.05; one-way ANOVA (Dunnett’s test); n = 7. (D) GDP resistance of interactions between V₂R-Rluc8 and G₁₂ heterotrimers bearing the indicated point mutations; n.s., not significant; all other groups significantly different from one; P < 0.05; one-sample t test; n = 3–14. (E) Activation of V₂R decreases p115RhoGEF-Rluc8 translocation when G₁₂ wt, Q381L, or Q381V are expressed, but increases translocation when G₁₂ I378Y or Δ1 are expressed; all groups were significantly different from mock-transfected controls (−); P < 0.05; one-way ANOVA (Dunnett’s test); n = 3–13.

The V₂R-G₁₂ Interaction Interferes with Other Transducers.

The robust agonist-induced BRET signal between V₂R receptors and G₁₂ heterotrimers in the presence of nucleotides suggested that this interaction might be stable enough to interfere with recruitment of other intracellular transducer molecules to V₂R. As V₂R receptors canonically activate G_s heterotrimers (11), we first asked how overexpressing G₁₂ would influence activation of adenylyl cyclase and cAMP accumulation. We found that overexpressing G₁₂ resulted in modest inhibition of G_s activation, as indicated by a G_s biosensor (SI Appendix, Fig. S8A). Surprisingly, this did not lead to detectable inhibition of V₂R-mediated cAMP accumulation, as indicated by two different cAMP sensors (SI Appendix, Fig. S8 B and C). Active V₂R receptors are phosphorylated by G protein receptor kinases (GRKs), and phosphorylated V₂R bind tightly to β-arrestins (22). Remarkably, we found that overexpressing G₁₂ significantly reduced AVP-induced BRET between V₂R-Rluc8 and β-arrestin2–Venus (Fig. 4 A and B). A much smaller but still significant reduction was observed after overexpressing G_s heterotrimers (Fig. 4 A and B). In contrast, overexpressing G₁₂ did not significantly reduce β-arrestin2 recruitment to ET_A, β₂-adrenergic, or angiotensin AT₁ receptors (SI Appendix, Fig. S9). Because V₂R–β-arrestin interactions are very stable and because phosphorylated V₂R can accommodate G protein and arrestin binding simultaneously (23), we suspected that G₁₂ overexpression was acting upstream of arrestin binding to inhibit V₂R interactions with GRKs. Consistent with this hypothesis, we found that G₁₂ overexpression greatly reduced the AVP-induced interaction of V₂R and GRK2 (SI Appendix, Fig. S10). Because arrestin binding is critical for agonist-dependent V₂R internalization (22) we then asked if G₁₂ overexpression would inhibit receptor endocytosis. Indeed, overexpression of G₁₂ but not G_s heterotrimers inhibited V₂R trafficking from the plasma membrane to the endosomal compartment as assessed by enhanced bystander BRET (ebBRET; Fig. 4 C and D). Conversely, there was a small but significant enhancement of V₂R internalization in cells lacking Gα₁₂ and Gα₁₃ subunits (SI Appendix, Fig. S11).

Fig. 4.

Overexpression of G₁₂ inhibits arrestin recruitment to V₂R and receptor internalization. (A and B) Time course and concentration-dependence of BRET between V₂R-Rluc8 and β-arrestin2–Venus in response to AVP (1 μM in A; mean ± SEM; n = 3). Overexpression of G₁₂ but not G_s heterotrimers inhibits arrestin recruitment. (C and D) AVP-induced changes in BRET between V₂R-RlucII and the plasma membrane marker rGFP-CAAX (C) and the endosome marker rGFP-FYVE (D), indicating trafficking of V₂R-RlucII from the plasma membrane to endosomes (mean ± SEM; n = 6–8). Overexpression of G₁₂ but not G_s heterotrimers inhibits AVP-induced internalization of V₂R-RlucII.

Other Receptors Also Form Unproductive Complexes with G₁₂.

In the course of experiments examining coupling of multiple different GPCRs to G proteins we encountered three additional examples of receptors that interact with G₁₂ heterotrimers in a nucleotide-resistant, unproductive manner. Smoothened (Smo) displays constitutive activity when the sterol transporter Patched is inhibited by Hedgehog or is not present, as is the case in HEK 293 cells. Smo is known to couple to and activate G_i heterotrimers (24). We found that unliganded Smo-Rluc8 did indeed interact with G_i heterotrimers in BRET assays, and this interaction was inhibited by either the inverse agonist cyclopamine or GDP, indicative of productive coupling (R_GDP = 0.19 ± 0.10; n = 3). In contrast, BRET between Smo-Rluc8 and G₁₂ heterotrimers was inhibited by cyclopamine but not GDP, indicative of unproductive coupling (R_GDP = 1.01 ± 0.22; n = 3; Fig. 5A). Similarly, activation of formyl peptide 2 receptors (FPR2) with the agonist peptide WKYMVm (WKY) promoted productive coupling with G_i heterotrimers (R_GDP = 0.30 ± 0.07; n = 4), but unproductive coupling with G₁₂ heterotrimers (R_GDP = 0.92 ± 0.04; n = 4; Fig. 5B). Although neither of these two receptors is known to activate G₁₂ we directly assessed activation of downstream G₁₂ signaling pathways by FPR2. As was the case with V₂R, activation of FPR2 failed to recruit p115-RhoGEF to the plasma membrane and failed to activate SRE-dependent gene transcription (SI Appendix, Fig. S12). Formyl peptide 1 receptors (FPR1) are highly homologous with FPR2 (68% identical), and we found that FPR1 also coupled productively with G_i heterotrimers (R_GDP = 0.26 ± 0.01; n = 3), but unproductively with G₁₂ heterotrimers (R_GDP = 0.96 ± 0.02; n = 3).

Fig. 5.

Smo and FPR2 form GDP-resistant complexes with G₁₂ heterotrimers. (A) BRET between Gαβγ-Venus heterotrimers and Smo-Rluc8 in the presence and absence of the inverse agonist cyclopamine (CPA; 10 μM), in the presence and absence of GDP. Smo is constitutively active in the absence of CPA. (B) BRET between Gαβγ-Venus heterotrimers and FPR2-Rluc8 in the presence and absence of the agonist WKYMVm (WKY, 0.5 μM), in the presence and absence of GDP. Both receptors coupled productively to G_i1 heterotrimers, but unproductively to G₁₂ heterotrimers; **P < 0.005; n.s., not significant (P > 0.75); one-way ANOVA (Sidak’s test) compared to GDP-CPA (A) and GDP+WKY (B); n = 3–12.

When we examined association of GPCRs with G proteins from all four Gα subtype families, we found that highly GDP-resistant interactions (R_GDP > 0.5) were restricted to G₁₂ heterotrimers (Fig. 6). For 19 of the 20 interactions that were studied with G_s, G_i, and G_q heterotrimers, R_GDP was <0.3, whereas this was the case for only two of the nine interactions we studied with G₁₂ heterotrimers. These results suggest that receptor-G₁₂ complexes may generally be more stable than other receptor-G protein complexes when G proteins are bound to GDP.

Fig. 6.

Receptor interactions with G₁₂ heterotrimers demonstrate a wide range of sensitivity to GDP. GDP resistance (R_GDP) of interactions between a panel of GPCRs fused to Rluc8 and the indicated Gαβγ-Venus heterotrimers (mean; n = 3–6). Receptors that interact in an agonist-dependent manner with G₁₂ heterotrimers are individually identified. Additional receptors are indicated by closed circles, and include: β₂AR, H₂R, D₁R (with G_s), α₂AR, M₄R, A₁R, D₂R, MOR (with G_i), AT₁R, M₃R, H₁R (with G_q). A full listing of receptor genes and ligands is provided in SI Appendix, Table S1.

Discussion

Taken together, our results suggest that several GPCRs bind to G₁₂ heterotrimers in an activation-dependent manner, but the resulting GPCR-G₁₂ complexes are insensitive to guanine nucleotides. These interactions do not activate G₁₂ signaling, but may instead have a negative effect on RhoGEF recruitment and signaling by sequestering G₁₂, thus preventing activation by other receptors. These interactions may also interfere with recruitment of other intracellular transducers and thus change signaling or trafficking of receptors that recruit but fail to activate G₁₂ heterotrimers. Whether or not these inhibitory effects occur under physiological conditions will depend on several factors, most notably the local abundance of G₁₂ heterotrimers and the stoichiometry of receptors and intracellular transducers. The normal physiological role of the V₂R is to enhance water reabsorption in the kidney by stimulating G_s, which ultimately leads to incorporation of aquaporin-2 water channels to the luminal surface of collecting duct cells (25). An inhibitory effect of V₂R activation on G₁₂ signaling could conceivably contribute to the physiological activity of this receptor, as Rho activity has been reported to act as an inhibitor of aquaporin transport (26). An inhibitory effect of the V₂R-G₁₂ interaction on arrestin recruitment could also play a regulatory role to limit receptor internalization. Although we found that G₁₂ overexpression weakly inhibited V₂R-mediated G_s activation when assessed using a direct G_s activation assay, we were surprised to find that this did not lead to a detectable decrease in cAMP accumulation. It is possible that GRK and arrestin recruitment are more sensitive to competition with G₁₂ than cAMP accumulation because cAMP signals are amplified downstream of G_s. Similar observations have been made after expression of some intrabodies that recognize the active state of β₂-adrenergic receptors (27). Further studies with native systems will be required to determine if unproductive GPCR-G₁₂ association has physiological significance.

At present, our findings significantly change the current model of GPCR coupling by demonstrating robust agonist-induced receptor-G protein interactions that do not lead to nucleotide exchange and G protein activation. GPCRs are thought to have access to all G protein subtypes expressed in a given cell, but possible interactions with noncognate heterotrimers (defined as G proteins that cannot be activated by a given GPCR) have, with a few exceptions (28), been overlooked. It is commonly assumed that stable agonist-induced GPCR-G protein interactions are restricted to cognate G proteins and are associated with G protein activation. One implication of this idea is that the conventional selection process whereby receptors reject noncognate G proteins occurs at an early stage of receptor-G protein association, such that complexes with noncognate G proteins do not progress past weak and transient encounter complexes. This seems to be true in the majority of cases, as several previous studies using sensitive methods have shown that interactions between GPCRs and noncognate G proteins are usually undetectable (7, 29). In contrast, our results suggest that some receptors functionally reject G₁₂ heterotrimers despite forming relatively stable GPCR-G₁₂ complexes. It is thought that GPCR-G protein complexes evolve through multiple intermediate conformations prior to receptor-stimulated nucleotide release (30–34). It is possible that receptors such as V₂R and FPR2 form similar intermediate complexes with G₁₂ heterotrimers that are unusually stable (Fig. 1B) and are unable to promote the changes in G₁₂ that lead to GDP release. Spontaneous GDP release from G₁₂ heterotrimers is particularly slow (35), and it may be that receptor-mediated GDP release requires relatively stable complexes with G₁₂-GDP, even for receptors that do activate G₁₂. G protein chimeras and mutants revealed that the Gα₁₂ C terminus is necessary for unproductive complexes with V₂R, implying that these complexes share some structural features with conventional productive ternary complexes. Our results with intramolecular G₁₂ BRET sensors suggest that V₂R and FPR2 may promote conformational changes in G₁₂ that do not lead to activation. However, the resolution of such probes is insufficient to determine if the heterotrimer itself changes conformation or, alternatively, if only the attached BRET donor and acceptor labels are rearranged. In either case, these results suggest that changes in G protein conformation reported by sensors similar to those used here do not necessarily indicate G protein activation.

In summary, our results reveal a mode of GPCR-G protein interaction wherein agonist-activated receptors bind to G₁₂ heterotrimers but do not promote nucleotide exchange and activation. These findings show that receptors can inhibit as well as activate G proteins, adding to the complexity of GPCR-mediated signaling.

Materials and Methods

Materials.

Trypsin, DPBS, PBS, HBSS, FBS, MEM, DMEM, penicillin/streptomycin, and l-glutamine were from Gibco (ThermoFisher Scientific). Polyethyleneimine MAX (PEI MAX) was purchased from Polysciences, Inc. Some receptor ligands, luciferin-d, and forskolin were purchased from Cayman Chemical. The remaining receptor ligands, digitonin, apyrase, GDP, GTP, GDPβS, and GTPγS were purchased from MilliporeSigma. Coelenterazine h and coelenterazine 400a were purchased from Nanolight Technologies. NanoGlo luciferase substrate was purchased from Promega.

Plasmid DNA Constructs.

GPCR plasmids were purchased from cdna.org (Bloomsburg University) or were provided by Bryan Roth (PRESTO-Tango Kit - #1000000068, Addgene). The V₂R-Rluc8 plasmid was received as a gift from Kevin Pfleger (Harry Perkins Institute of Medical Research, Nedlands, Western Australia). A plasmid encoding β-arrestin2–Venus was a gift from Vsevolod Gurevich (Vanderbilt University, Nashville, TN). V₂R-SmBit and ETAR-SmBit were generated by replacing the GPCR coding sequence in β₂AR-SmBit digested with EcoRI and NotI, which appended the SmBit peptide to the C terminus of each receptor behind a GGRGGGGSG linker. Plasmids encoding Gα subunits, Gβ₁, and Gγ₂ were purchased from cdna.org. Gα₁₂-Rluc8 was generated by inserting Rluc8 (flanked by GGSG linkers) between a residues N136 and K137 of Gα₁₂ using Quikchange mutagenesis. GRK2-Venus-Kras and GRK2-Venus-Kras R587Q were generated by appending Venus fused to the last 25 amino acids of Kras to the C terminus of bovine GRK2 or GRK2 R587Q using Quikchange mutagenesis. Plasmids encoding the S1 subunit of pertussis toxin (PTX-S1) and LgBit-Gγ₂ were kindly provided by Stephen R. Ikeda (NIAAA, Rockville, MD), and the Nluc-EPAC-VV plasmid was provided by Kirill Martemyanov (Scripps Research Institute, Jupiter, FL). The Glosensor-22F cAMP plasmid (E2301) was obtained from Promega. Plasmids encoding Gα_sΔ10, Venus-Kras, Venus-1–155-Gγ₁, and Venus-155–239-Gβ₁ GPCR-luciferase constructs, and p115RhoGEF-Rluc8 have been described previously (6, 15, 36). Plasmids encoding rGFP-CAAX, rGFP-FYVE, and V₂R-RlucII have been described previously (16). PDZ-RhoGEF-RlucII was generated by amplifying the cytosolic G12/13 interacting domain of PDZ-RhoGEF (aa 281–483) with linkerD (GIRLREALKLPAT) on its C terminus which was then subcloned onto the N terminus of RlucII in pcDNA3.1/Zeo(+) by Gibson assembly. GRK2-RlucII D110A was generated by digesting hGRK2-GFP10 D110A and pcDNA3.1/Hygro(+) GFP10-RlucII db v.2 with NheI and HindIII to excise hGRK2 from the former and GFP10 from the latter. hGRK2 was subsequently ligated in frame with pcDNA3.1/Hygro(+) RlucII db v.2 to produce a C-terminal RlucII construct. All plasmid constructs were verified by Sanger sequencing.

Cell Culture and Transfection.

HEK 293 cells (ATCC) were propagated in plastic flasks and on 6-well plates according to the supplier’s protocol. HEK 293 cells with targeted deletion of GNAS and GNAL (G_s knockouts; GSKO), targeted deletion of GNAS, GNAL, GNAQ, GNA11, GNA12, and GNA13 (G protein three family knockouts; 3GKO), and HEK 293 cells with additional targeted deletions to the 3GKO cells of GNAI1, GNAI2, GNAI3, GNAT1, GNAT2, GNAZ, and GNAO1 (G protein four family knockouts; 4GKO) were derived, authenticated, and propagated as previously described (9, 10). HEK 293 cells with additional targeted deletion of ARRB1 and ARRB2 (beta-arrestin knockouts; ARRBKO) were derived, authenticated, and propagated as previously described (37, 38). Cells were transfected in growth medium using linear PEI MAX (MW 40,000) at a nitrogen/phosphate ratio of 20 and were used for experiments 12–48 h later. Up to 3.0 μg of plasmid DNA was transfected in each well of a 6-well plate. For ebBRET experiments, up to 1.0 μg of plasmid DNA was transfected in suspension to a cell density of 350,000 cells/mL in white 96-well plates.

BRET and Luminescence Assays.

Measurement of coupling between receptor and G protein in nucleotide-depleted cells.

Cells were transfected with a GPCR-Rluc8 and Gα subunit pair, Venus-1–155-Gγ₂, Venus-155–239-Gβ₁, and pcDNA3.1(+) or PTX-S1 in a (1:3:1:1:1) ratio. Experiments with Gα_i were conducted in 4GKO cells for Gα_i cognate receptors and in 3GKO cells for all other receptors. Experiments with Gα_i were conducted without PTX-S1; all other Gα subunits were cotransfected with PTX-S1. After a 48-h incubation, cells were washed twice with permeabilization buffer (KPS) containing 140 mM KCl, 10 mM NaCl, 1 mM MgCl₂, 0.1 mM KEGTA, 20 mM NaHEPES (pH 7.2); harvested by trituration; permeabilized in KPS buffer containing 10 μg mL⁻¹ high-purity digitonin; and transferred to opaque black 96-well plate. Measurements were made from permeabilized cells supplemented either with 100 μM GDP or 2U mL⁻¹ apyrase, in both cases with or without agonist (SI Appendix, Table S1).

Luciferase complementation.

Cells were transfected with a GPCR-SmBit, Gα, LgBit-Gγ₂, Gβ₁, and pcDNA3.1(+) or PTX-S1 in a (1.5:4:1:2.5:3) ratio. After a 24-h incubation, cells were washed twice with DPBS, harvested by trituration, and transferred to opaque white 96-well plates.

GPCR competition assays.

Cells were transfected with an untagged GPCR or pcDNA3.1(+), GPCR-Rluc8, Gα, Venus-1–155-Gγ₂, Venus-155–239-Gβ₁, and PTX-S1 in a (10:1:2:2:4) ratio. After a 48-h incubation, cells were washed twice with DPBS, harvested by trituration, and transferred to opaque black 96-well plates.

G protein BRET conformational biosensor.

HEK 293 cells were transfected with an untagged GPCR, Gα₁₂-Rluc8, Venus-1–155-Gγ₂ and Venus-155–239-Gβ₁ in a (15:1:4:4) ratio. After a 48-h incubation, cells were washed twice with DPBS, harvested by trituration, and transferred to opaque black 96-well plates.

G_s activation nanoBiT sensor.

NanoBiT-G_s protein (39) consisting of Gα_s subunit fused with a large fragment (LgBiT) at the alpha-helical domain and an N-terminally small fragment (SmBiT)-fused Gγ₂ subunit along with untagged Gβ₁ subunit was expressed in the presence or absence of Gα₁₂ subunit. HEK 293 cells were seeded in a 6-well culture plate at a concentration of 2 × 10⁵ cells mL⁻¹ (2 mL per well in DMEM; Nissui) supplemented with 10% fetal bovine serum (Gibco), glutamine, penicillin, and streptomycin) 1 d before transfection. Transfection solution was prepared by combining 5 μL (per well hereafter) of PEI MAX solution (1 mg mL⁻¹), 200 μL of Opti-MEM (ThermoFisher Scientific), and a plasmid mixture consisting of 100 ng LgBiT-containing Gα_s subunit, 500 ng Gβ₁, 500 ng SmBiT-fused Gγ₂, 100 ng RIC8b, 200 ng untagged V₂R (pCAGGS plasmid) with or without 20 ng Gα₁₂ subunit (pCAGGS plasmid; gene-synthesized with codon optimization). After incubation for 1 d, transfected cells were harvested with 0.5 mM EDTA-containing DPBS, centrifuged and suspended in 4 mL of HBSS containing 0.01% BSA (fatty acid–free grade; SERVA) and 5 mM Hepes (pH 7.4) (assay buffer). The cell suspension was dispensed in a white 96-well plate at a volume of 80 µL per well and loaded with 20 µL of 50 µM coelenterazine (Carbosynth) diluted in the assay buffer. After 2 h incubation at room temperature, the plate was measured for baseline luminescence (Spectramax L, Molecular Devices), and 20 µL of titrated ligand (AVP) were manually added. The plate was immediately read at room temperature for the following 10 min at a measurement interval of 20 s with an accumulation time of 0.17 s per read. The luminescence counts over 5–10 min after ligand addition were averaged and normalized to the initial count. The fold-change values were further normalized to that of vehicle-treated samples.

Translocation of p115RhoGEF.

Cells were transfected with an untagged GPCR, Gα, Gγ₂, Gβ₁, p115RhoGEF-Rluc8, Venus-Kras, and PTX-S1 in a (2:12:4:4:1:6:2) ratio. After a 48-h incubation, cells were washed twice with DPBS, harvested by trituration, and transferred to opaque black 96-well plates.

Translocation of PDZ-RhoGEF.

ΔβARR1/2 HEK 293 cells were transfected with either FLAG-V₂R or HA-TPα, Gα₁₂, PDZ-RhoGEF-RlucII and rGFP-CAAX in a (8:4:1:12) ratio. After a 48-h incubation, cells were washed once with Tyrode’s buffer (140 mM NaCl, 2.7 mM KCl, 1 mM CaCl₂, 12 mM NaHCO₃, 5.6 mM d-glucose, 0.5 mM MgCl₂, 0.37 mM NaH₂PO₄, 25 mM Hepes [pH 7.4]) and maintained in the same buffer. Cells were stimulated for 5 min with agonist before BRET measurements.

p115-RGS-GFP biosensor to monitor Gα₁₂ activity.

A BRET-based biosensor composed of RGS homology (RH) domain (amino acids 1–246) of p115RhoGEF fused to GFP10 (p115-RGS-GFP) and one of two Gα₁₂-RlucII fusions (RlucII inserted after amino acid 84 or 136) was used to measure Gα₁₂ activity (40). HEK 293 cells were transfected with 40 ng of Gα₁₂-RlucII, 500 ng of p115-RGS-GFP, and 300 ng of receptor per row of a 96-well plate. BRET was monitored 2 min after agonist addition.

SRE transcriptional reporter assay.

Cells were transfected with a GPCR, Gα subunit, SRE-Luc, and PTX-S1 in a (10:1:100:25) ratio. Medium was exchanged to serum-free 2 h after transfection. After a 24-h incubation, cells were treated with or without agonist for 5 h. Cells were washed twice with DPBS, harvested by trituration, centrifuged at 500 × g for 3 min, and resuspended in equilibration buffer (1× HBSS, 20 mM NaHEPES; pH 7.5) supplemented with 10% FBS by volume, and 2 mM d‐luciferin. Cells equilibrated in this solution at room temperature for 30 min and were transferred to opaque white 96-well plates.

NFκB transcriptional reporter assays.

Cells were transfected with a GPCR, Gα subunit, NFκB-Luc, and empty vector in a (300:1:300:199) ratio. After a 24-h incubation, cells were treated with or without agonist for 5 h. Cells were washed twice with DPBS, harvested by enzyme-free, centrifuged at 500 × g for 3 min, and resuspended in equilibration buffer (1× HBSS, 20 mM NaHEPES; 0.1% wt/vol BSA, pH 7.5) and transferred into 96-well black/white Isoplates (Perkin-Elmer). Cells were incubated with 2 mM D‐luciferin for 30 min before reading luminescence emission at 525 nm after 30 of incubation using a PHERAstar FS (BMG LABTECH).

Nluc-EPAC-VV cAMP assay.

Cells were transfected with a pcDNA3.1(+), a GPCR, Gα subunit or pcDNA3.1(+), Gγ₂, Gβ₁, and Nluc-EPAC-VV in a (59:15:5:15:10:10:1) ratio. After a 24-h incubation, cells were washed twice with DPBS, harvested by trituration, and transferred to opaque black 96-well plates.

Glosensor cAMP assay.

GSKO cells were transfected with a GPCR, Gα subunit, Gγ₂, Gβ₁, Glosensor 22F, and either pcDNA3.1(+) or PTX‐S1 in a (1:1:1:1:4:1) ratio. After a 24‐h incubation, cells were washed twice with DPBS and treated with trypsin‐EDTA (0.05%). Detached cells were harvested and centrifuged at 250 × g for 5 min, and the cell pellet was resuspended in equilibration buffer supplemented with 10% FBS by volume and 2 mM d‐luciferin. Cells were incubated at room temperature for 1 h and then distributed to opaque white 96‐well plates. Luminescence measurements were made from cells treated with vehicle, agonist, or 100 μM forskolin.

Arrestin recruitment.

HEK 293 cells were transfected with a GPCR-Rluc8, Gα, Gγ₂, Gβ₁, and β-arrestin2–Venus in a (1:2:1:1:1) ratio. After a 24-h incubation, cells were washed twice with DPBS, harvested by trituration, and transferred to opaque black 96-well plates.

Bystander BRET V₂R trafficking.

HEK 293 cells were transfected with V₂R-RlucII, Gα, and either rGFP-CAAX or rGFP-FYVE in a (1:20:60) ratio. After a 48-h incubation, cells were washed once with Tyrode’s buffer and maintained in the same buffer. Cells were stimulated for 30 min with agonist before BRET measurements.

HiBiT-based V₂R internalization.

Parental HEK 293 cells, G_12/13-deficient HEK 293 cells (39), or β-arrestin1/2-deficient HEK 293 cells (37) in growth phase were seeded in a 6-well culture plate at a concentration of 2 × 10⁵ cells mL⁻¹. Cells were transfected with 100 ng of HiBiT-V₂R, which contained an Interleukin 6-derived signal sequence followed by a HiBiT sequence and a linker at the N terminus (MNSFSTSAFGPVAFSLGLLLVLPAAFPAPVSGWRLFKKISGGSGGGGSG; gene synthesized with codon optimization) and an unintended SmBiT tag at the C terminus. After 1 d, cells were harvested, suspended in 1 mL of assay buffer, dispensed in a white 96-well half-area plate at a volume of 25 µL per well, and mixed with 25 µL of 2× substrate buffer consisting of 1:200 of a LgBiT stock solution (Promega) and 20 µM furimazine in the assay buffer. After 40 min at room temperature, the plate was measured for baseline luminescence, and a titrated ligand (10 µL) diluted in the 1× substrate buffer was manually added. The plate was immediately read at room temperature for the following 30 min at a measurement interval of 30 s with an accumulation time of 0.4 s per read. The luminescence counts over 27–30 min after ligand addition were averaged and normalized to the initial count.

GRK2 recruitment.

For the experiments shown in SI Appendix, Fig. S10A, ΔβARR1/2 HEK 293 cells were transfected with FLAG-V₂R, GRK2-RlucII D110A, Gα, and GFP-CAAX in a (2:1:2:6) ratio in suspension and distributed into white 96-well plates. After a 48-h incubation, cells were washed once with Tyrode’s buffer and maintained in the same buffer. Cells were stimulated with agonist immediately after addition of coelenterazine 400a. For the experiments shown in SI Appendix, Fig. S10B, ARRBKO cells were transfected with a GPCR-Rluc8, Gα subunit, Gγ₂, Gβ₁, and GRK2-Venus in a (1:3:1:1:3) ratio. After a 48-h incubation, cells were washed twice with DPBS, harvested by trituration, and transferred to opaque black 96-well plates.

BRET, luminescence measurements.

Steady-state BRET and luminescence measurements were made using a Mithras LB940 photon-counting plate reader (Berthold Technologies GmbH). Kinetic BRET and luminescence time course measurements were made using a Polarstar Optima plate reader (BMG Labtech). Coelenterazine h (5 μM; Nanolight) or furimazine (NanoGlo; 1:1,000, Promega) were added to all wells immediately prior to making measurements with Rluc8 and Nluc, respectively. Raw BRET signals were calculated as the emission intensity at 520–545 nm divided by the emission intensity at 475–495 nm. Net BRET signals were calculated as the raw BRET signal minus the raw BRET signal measured from cells expressing only the Rluc8 donor.

Data Availability.

All study data are included in the article and SI Appendix.