Crosstalk between Gα_i- and Gα_q-coupled receptors is mediated by Gβγ exchange

Abstract

Activation of Gα_i-coupled receptors often causes enhancement of the inositol phosphate signal triggered by Gα_q-coupled receptors. To investigate the mechanism of this synergistic receptor crosstalk, we studied the Gα_i-coupled adenosine A₁ and α_2C adrenergic receptors and the Gα_q-coupled bradykinin B₂ and a UTP-preferring P2Y receptor. Stimulation of either Gα_i-coupled receptor expressed in COS cells increased the potency and the efficacy of inositol phosphate production by bradykinin or UTP. Likewise, overexpression of Gβ₁γ₂ resulted in a similar increase in potency and efficacy of bradykinin or UTP. In contrast, these stimuli did not affect the potency of direct activators of Gα_q; a truncated Gβ₃ mutant had no effect on the receptor-generated signals whereas signals generated at the G-protein level were still enhanced. This suggests that the Gβγ-mediated signal enhancement occurs at the receptor level. Almost all possible combinations of Gβ_1–3 with Gγ_2–7 were equally effective in enhancing the signals of the B₂ and a UTP-preferring P2Y receptor, indicating a very broad specificity of this synergism. The enhancement of the bradykinin signal by (i) Gα_i-activating receptor ligands or (ii) cotransfection of Gβγ was suppressed when the B₂ receptor was replaced by a B₂Gβ₂ fusion protein. Gβγ enhanced the B₂ receptor-stimulated activation of G-proteins as determined by GTPγS-induced decrease in high affinity agonist binding and by B₂ receptor-enhanced [³⁵S]GTPγS binding. These findings support the concept that Gβγ exchange between Gα_i- and Gα_q-coupled receptors mediates this type of receptor crosstalk.

Seven transmembrane domain receptors signal via heterotrimeric G-proteins. So far, 21 different gene products encoding Gα subunits, 5 Gβ, and 11 Gγ subunits have been distinguished. With these subunits, the formation of >1,000 different G-protein heterotrimers is conceivable. This panoply of combinations is theoretically sufficient to provide each individual G-protein-coupled receptor with a specific G-protein heterotrimer. For the various Gα subunits, specific interactions with receptors have been defined by various approaches. In addition, injection of antisense oligonucleotides or the transfection of a ribozyme specific for defined Gβ or Gγ differentially inhibited receptor-mediated signals, suggesting that in intact cells receptors are specifically coupled to defined Gβγ-heterodimers (1–5). These findings argue for a receptor-specific G-protein heterotrimer interaction. In contrast, transfection experiments revealed that a single combination, Gβ₁γ_2, is equally effective in enhancing signals of several G-protein coupled receptors (6), and most biochemical assays investigating Gβγ interactions with Gα and with effector molecules have shown little specificity (7–10). Thus, it is unclear whether the direct interaction of defined Gβγ subunits with a receptor is selective. This question is fundamental because a lack of biochemical specificity between different Gβγ combinations would allow different receptors to share a common pool of Gβγ subunits.

Under physiological conditions, such nonselective interactions might be important for the crosstalk between different receptors. It has been known for several years that stimulation of Gα_i-coupled receptors can enhance the inositol phosphate signals triggered by activation of Gα_q-coupled receptors, even though stimulation of the Gα_i-coupled receptors alone often has no effect on inositol phosphates (11–13). It is thought that Gβγ subunits are responsible for this effect (12–16) because relatively high concentrations of agonists are necessary. However, the mechanism of such an enhancement by Gβγ subunits has remained unclear. Direct stimulation of phospholipase Cβ by Gβγ has been reported (17), and this stimulation can occur in addition to that by Gα_q (18). However, this stimulation is not synergistic (18–19), and this leaves the question unanswered of whether direct stimulation of phospholipase Cβ by Gβγ subunits is indeed the mechanism for the enhancing effects caused by Gα_i-coupled receptors.

To analyze the mechanism of this synergistic receptor crosstalk and the role of Gβγ in this process, we chose the Gα_q-coupled bradykinin B₂ and a UTP-preferring P2Y receptor (20–21), and the Gα_i-coupled adenosine A₁ and the α_2C adrenergic receptors. These receptors also are expressed together in native cells and are known to enhance each other’s signaling (15, 22). We report that the receptor crosstalk can occur at the level of the receptors themselves and propose that exchange of Gβγ subunits by the Gα_i- and the Gα_q-coupled receptors mediates this type of receptor crosstalk.

MATERIALS AND METHODS

Construction of Expression Vectors.

All of the cDNAs used in these studies were ligated into an expression vector that is driven by the cytomegalovirus promoter. Identity of the subcloned cDNAs coding for bovine Gγ_1,2,3,5,7, for mouse Gγ₄, and for human Gβ_1–3 was determined by DNA sequencing. The cDNA coding for a B₂ receptor-Gβ2 fusion protein was constructed by overlap extension using PCR as described (23). The following primers were used for the PCRs: sense-B₂, 5′-TCCCTCTAGAAGCTTATGTTCTCTCCCTGGAG-3′; antisense-B₂, 5′-TGCTCCAGCTCACTCATCTGTCTGCTCCCTGCCCA-3′; sense-Gβ₂, 5′-TGGGCAGGGAGCAGACAGATGAGTGAGCTGGAGCA-3′; antisense-Gβ₂, 5′-ATCCGAGCTCGAATTCTTAGTTCCAGATCTTGA-3′. The resulting PCR-product was ligated into the HindIII/EcoRI sites of the pcDNA3 vector (Invitrogen). An N-terminally truncated mutant of the Gβ₃ subunit (amino acids 45–340) was created by PCR. Identity of the constructs was confirmed by DNA sequencing.

Cell Transfection and Determination of Inositol Phosphate Levels.

COS-1 cells (American Type Culture Collection) were grown in DMEM supplemented with 10% (vol/vol) fetal calf serum and were kept in a 92.5% air/7.5% CO₂ atmosphere. Cells were transfected by using Lipofectamine (GIBCO/BRL), or the DEAE-dextran method with 5 μg of DNA/ml of the DEAE-dextran solution (6). The transfection efficiency was usually 20–30% as determined by staining of the cells with X-Gal (5-bromo-4-chloro-3-indolyl-β-d-galactopyranoside) after transfection with a plasmid for β-galactosidase (6). For determination of inositol phosphate levels, cells were seeded on 12- or 6-well plates coated with 0.1% (wt/vol) gelatine in PBS and were transfected with 5 μg of DNA/ml of transfection mixture. In cotransfection experiments, the total amount of DNA was held constant with plasmid pcDNA3 (Invitrogen). Forty-eight hours after transfection, cells were assayed for inositol phosphate levels or for binding of [2,3-prolyl-3,4-³H]bradykinin (specific activity 98 Ci/mmol) (Amersham) or were scraped for membrane preparation as described (24, 25).

Measurement of Changes in the Intracellular Free Ca²⁺ Concentration.

Intracellular free Ca²⁺ concentration, [Ca²⁺]_i, of HEK-293 cells transfected with cDNAs coding for the B₂ or the B₂Gβ₂ receptor was determined on detached cells by fura-2/AM as described (24). Fluorescence at 510 nm was recorded. Excitation wavelength alternated between 340 and 380 nm in intervals of 40 ms. Changes in [Ca²⁺]_i are given as the ratio between 340 and 380 nm.

Immunoblotting.

Proteins of transfected COS-1 cells were resolved by SDS/PAGE on 10% acrylamide gels. Proteins were transferred to polyvinylidenedifluoride membranes by using semidry blotting. Proteins (Gβ subunits, B₂ receptor) were detected on immunoblots as described (24) by anti-Gβ common (Transduction Research, Framingham, MA) and anti-B₂ receptor antibodies (24), respectively.

RESULTS

Enhancement of Gα_q-Coupled Receptor Signals by Stimulation of Gα_i-Coupled Receptors.

To analyze the mode of receptor crosstalk between Gα_i and Gα_q-coupled receptors, we chose the Gα_q-coupled bradykinin B₂ and a UTP-preferring P2Y receptor, and the Gα_i-coupled adenosine A₁ and α_2C adrenergic receptors. Expression of α_2C-adrenergic or A₁ adenosine receptors in COS cells and stimulation with 1 μM medetomidine (α_2C-adrenergic receptor) or cyclohexyladenosine (A₁ adenosine receptor) did not per se result in the generation of inositol phosphates. However, their stimulation increased the potency and the efficacy of UTP and bradykinin. Fig. 1 shows this for the α_2C-adrenergic receptor; its stimulation increased the potency of UTP from 3.5 ± 0.4 μM to 1.5 ± 0.4 μM (Fig. 1A) and that of bradykinin from 2.0 ± 0.2 nM to 0.9 ± 0.1 nM (Fig. 1B) (n = 3). In addition, the efficacy of both hormones was increased by the costimulation of the α_2C-adrenergic receptor to 158 ± 7% and 137 ± 6%, respectively. Very similar results were obtained for the coexpression and costimulation of the A₁ adenosine receptor (data not shown).

Figure 1.

Gβγ-dependence of the inositol phosphate responses to UTP (A), bradykinin (B), AlF₄⁻ (C), and GTPγS (D) in mock-transfected COS cells (○) or in COS cells transfected with plasmids encoding for Gβ₁γ₂ (●), or for the α_2C-adrenergic receptor (■). Cells expressing α_2C-receptors were costimulated with 1 μM medetomidine. For GTPγS stimulation (D), cells were permeabilized with 10 μM digitonin in a nominally Ca²⁺-free buffer containing 138 mM KCl instead of NaCl. The ligand-stimulated increase of the inositol phosphate levels is given as percent of the maximum increase in mock-transfected control cells (100%), which was 2,430 ± 180, 2,100 ± 130, 650 ± 70, and 380 ± 40 cpm for AlF₄⁻, GTPγS, UTP, and bradykinin, respectively. Expression of α_2C-adrenergic receptors and stimulation by medetomidine did not significantly alter basal inositol phosphate levels. Basal inositol phosphate levels determined in the absence (320 ± 25 cpm) or presence (540 ± 30 cpm) of Gβ₁γ₂ were subtracted from the respective ligand-stimulated values to determine the pure “ligand-stimulated increase.” Data are from a representative experiment that was reproduced three times with similar results.

Effects of Gβγ on Receptor- and G_α-Triggered Stimulation of PLC Activity. Do Gβγ subunits released from Gα_i mediate the enhancement of Gα_q-stimulated phospholipase C (PLC) activity? We expressed Gβγ subunits and compared their effects with those of Gα_i-coupled receptor stimulation. Expression of Gβγ subunits also caused an increase in the potency and efficacy of UTP and bradykinin (Fig. 1A and B), and this enhancement was virtually indistinguishable from that caused by stimulation of the α₂-adrenergic receptor. The EC₅₀ values were increased from 3.5 ± 0.4 μM to 1.5 ± 0.5 μM for UTP and from 2.0 ± 0.2 nM to 0.9 ± 0.2 nM for bradykinin. These results are indeed compatible with the view that stimulation of Gα_i-coupled receptors enhances PLC activation by causing an increase in free Gβγ subunits at the cell membrane.

Two different mechanisms are conceivable for how Gβγ subunits could enhance a receptor-stimulated signal: (i) Free Gβγ subunits may somehow activate PLC together with Gα_q, or (ii) expression of Gβγ subunits may specifically increase receptor-mediated signals. To discriminate between both mechanisms, we stimulated Gα proteins directly with AlF₄⁻ or GTPγS (Fig. 1 C and D). Gβγ subunits increased the maximum AlF₄⁻ and GTPγS-stimulated signals (Fig. 1 C and D). However, the potencies of AlF₄⁻ and GTPγS were not enhanced by Gβγ but were actually slightly decreased from 110 ± 20 μM to 190 ± 20 μM and from 50 ± 5 μM to 72 ± 15 μM (n = 3), respectively. Concomitant Gα_i-coupled α_2C receptor stimulation had similar effects on the AlF₄⁻-stimulated signal (Fig. 1C).

Different Effects of a Gβ₃ Mutant on AlF₄⁻ Versus Receptor-Stimulated Signals.

To further discriminate between the two different modes of Gβγ-mediated signal enhancement, we created an N-terminally truncated Gβ₃ mutant (amino acids 45–340) with reduced affinity for the Gγ subunit. The Gβγ-receptor interaction depends on the Gγ subunit (26). Therefore, a Gβ mutant with reduced affinity for the Gγ subunit should have a reduced capacity to interact with receptors. The Gβ₃ mutant expressed in COS cells exhibited a slightly decreased apparent molecular weight compared with the wild-type protein in immunoblot (Fig. 2A). The mutant was still capable of significantly increasing basal inositol phosphate levels (Fig. 2B) and of enhancing an AlF₄⁻-stimulated signal (Fig. 2C). The stimulatory effects mediated by the truncated Gβ₃ depended on the cotransfection of Gγ₇, suggesting that the truncated Gβ₃ still interacted weakly with Gγ subunits (Fig. 2 B and C). However, expression of the truncated Gβ₃ mutant did not significantly enhance the UTP- or bradykinin-stimulated signals (Fig. 2 D and E). Thus, the mode of Gβγ-mediated signal enhancement at the receptor level seems to be different from the stimulatory effect of Gβγ at the G-protein level.

Figure 2.

Effect of a Gβ₃ mutant (amino acids 45–340) on basal, AlF₄⁻-, UTP-, and bradykinin-stimulated inositol phosphate levels. (A) Immunoblot of COS cells transfected with the Gβ₃ mutant (β₃mut) or with wild-type Gβ₃ (WT). Gγ₁ and Gγ₇ were cotransfected as indicated. As a control, mock-transfected COS cells were used (-). (B–E) Inositol phosphate levels of COS cells expressing the Gβ₃ mutant (β₃mut) together with Gγ_1,7 or wild-type Gβ₃ (WT) with Gγ₇. In B, basal inositol phosphate levels were determined. The increase in inositol phosphate levels over the respective basal levels after stimulation by AlF₄⁻ (300 μM AlCl₃/10 mM NaF), by UTP (100 μM), and by bradykinin (100 nM) are given in C, D, and E, respectively. Data are from a representative experiment (triplicates ± SD) that was reproduced three times with similar results.

Enhancement of B₂ and UTP-Preferring P2Y Receptor Signals by Defined Gβγ Subunits. The above-described findings raise the possibility that the synergism might be exerted at the receptor level: for example, by exchange of Gβγ subunits between Gα_i and Gα_q-coupled receptors. A prerequisite for such a mechanism is the ability of a given receptor to interact equally well with a broad spectrum of Gβγ combinations because, initially, Gα_q- and Gα_i-coupled receptors might well be coupled to different Gβγ dimers. We therefore determined the interaction of defined Gβγ subunits with the B₂ and UTP-preferring P2Y receptor. For this analysis, we chose Gβ_1–3 and Gγ_1–7 because the specific G_βγ-receptor complexes delineated from antisense studies (1–5) are all composed of these subunits. These Gβγ combinations were expressed in COS cells. Equal protein expression of the various Gβγ combinations was verified by immunoblotting with anti-Gβ_common antibodies (Fig. 3A). Compared with mock-transfected cells, the Gβ signal of overexpressing cells was increased ≈2-fold. Considering a transfection efficiency of 20–30% (as determined by staining of cells transfected with a β-galactosidase expression plasmid) (data not shown), a 2-fold increase in intensity suggests a 4- to 6-fold overexpression of Gβ in the transfected cells. All combinations of Gβ_1–3 with Gγ_2–7 were equally effective in stimulating basal PLC activity (Fig. 3B). The effect of Gβ₁γ₁ (=transducin Gβγ) was only ≈25% of maximum (Fig. 3B), and Gβ₂γ₁ and Gβ₃γ₁ did not significantly increase basal inositol phosphate levels. These findings confirm earlier data (27–28) that Gβ₂ and Gβ₃ associate weakly with Gγ₁.

Figure 3.

Gβγ-mediated enhancement of basal, bradykinin-, and UTP-stimulated inositol phosphate levels. (A) Immunoblot of COS cells transfected with combinations of cDNAs coding for Gβ_1–3 and Gγ_1–7. Note that Gγ₆ is lacking because it is identical with Gγ₂. Gβ subunit expression was detected by a monoclonal anti-Gβ_common antibody. (B) Increase of basal inositol phosphate levels of COS cells after coexpression of Gβ_1–3 and Gγ_1–7. The values represent the increase of basal inositol phosphate levels after Gβγ expression compared with mock-transfected control cells (0%) and are given as percent of maximum (100%) determined with Gβ₁γ₂. Expression of Gβ₁γ₂ increased basal inositol phosphate levels (240–380 cpm) 1.7- to 1.9-fold. (C and D) Gβγ-mediated enhancement of the bradykinin- [100 nM (C)] and UTP- [100 μM (D)] stimulated signals. The Gβγ-mediated signal enhancement over the signal of the mock-transfected control cells (0%) is given as percent of maximum stimulation (100%) after Gβ₁γ₂ expression. In mock-transfected control cells, the UTP- and bradykinin-stimulated increase over basal was 580–880 and 210–430 cpm, respectively. In Gβ₁γ₂-expressing cells, the UTP and the bradykinin-stimulated increase over basal was 1,370–1,850 cpm and 540–1,300 cpm, respectively. Data presented are the means (±SEM) of 3–7 different experiments.

Subsequently, the effects of Gβγ expression on the two different Gα_q-coupled receptors were determined. Combinations of Gβ_1–3/Gγ_2,3,4,7 enhanced the bradykinin and UTP-stimulated increase in inositol phosphate levels equally well (Fig. 3 C and D). Complexes with Gγ₅ also enhanced the hormone-stimulated signals, though to a slightly lesser extent. Taken together, the interaction of the various Gβγ subunits with the signaling cascades of the B₂ and the UTP-preferring P2Y receptor is characterized by a remarkable lack of coupling specificity, which allows combinations formed of Gβ_1–3 and Gγ_2–7 to enhance efficiently the bradykinin- or UTP-mediated signals. This lack of specificity should enable these receptors to accept different Gβγ dimers released from various differentially coupled receptors.

Signaling of a B₂ Receptor-Gβ₂ Fusion Protein. To test the hypothesis that the enhancement of Gα_q-coupled receptor signaling by Gα_i-coupled receptors reflects exchange of Gβγ subunits between two different receptors, we constructed a fusion protein between the B₂ receptor and Gβ₂. Transient transfection of COS cells resulted in the expression of the fusion protein as determined by immunoblot stained with anti-Gβ antibodies, which showed the appearance of an ≈95-kDa protein in addition to the endogenous Gβ subunits of ≈35 kDa (Fig. 4A). The apparent molecular weight of ≈95 kDa corresponds to the expected size of B₂ receptor plus Gβ subunit. When the immunoblot was stained with anti-B₂ receptor antibodies (24), again a protein of ≈95 kDa was detected in B₂Gβ₂-expressing cells whereas a protein of ≈60 kDa was apparent in wild-type B₂ receptor-expressing cells (Fig. 4B). These findings confirm that the expressed protein consists of a Gβ₂ subunit fused to the B₂ receptor.

Figure 4.

Characterization of a fusion protein between the B₂ receptor and the Gβ₂ subunit. (A and B) Expression of the B₂Gβ₂-fusion protein in COS cells (lane 1). Cells transfected with the wild-type B₂ receptor served as a control (lane 2). The immunoblots were analyzed with anti-Gβ_common (A) or anti-B₂ receptor (B) antibodies. (C and D) Bradykinin-stimulated rise in [Ca²⁺]_i in fura-2-loaded HEK-293 cells transfected with the wild-type B₂ receptor (C) or with the B₂Gβ₂ fusion protein (D). At the time point indicated by an arrow, 100 nM bradykinin (Bk) was added. (E) Bradykinin-induced rise in inositol phosphate levels in COS cells transfected with the wild-type B₂ receptor or with the B₂Gβ₂-fusion protein. Cells expressed 6.1 pmol/mg protein of wild-type and 3.9 pmol/mg protein of B₂Gβ₂ receptor. Data are from a representative experiment (triplicates ± SD) that was reproduced three times with similar results. (F) Inositol phosphate levels of COS cells expressing different levels of wild-type B₂ receptors or of B₂Gβ₂ fusion protein after stimulation by 100 nM bradykinin. The amount of expressed receptors is given as pmol/mg protein. Data are from a representative experiment (triplicates ± SD) that was reproduced three times with similar results.

The affinity of the fusion protein for the agonist bradykinin as determined by radioligand binding was 0.4 ± 0.2 nM (data not shown); this value was similar to the affinity of the wild-type receptor (0.8 ± 0.3 nM). To determine whether the fused Gβ₂ subunit altered the mode of receptor activation, we determined the B₂-receptor-mediated rise in [Ca²⁺]_i. The bradykinin-stimulated rise in [Ca²⁺]_i in cells expressing the fusion protein was virtually identical to cells expressing the wild-type receptor, suggesting that the fused Gβ subunit did not alter the kinetics of agonist-stimulated receptor activation (Fig. 4 C and D). Also, the EC₅₀ value for the bradykinin-induced rise in inositol phosphate levels of the B₂ receptor-fusion protein was similar to the wild-type B₂ receptor (Fig. 4E). Thus, B₂ receptors with C-terminally fused Gβ subunit interact efficiently with Gα proteins of COS cells.

Does the fused Gβ₂ subunit enhance the B₂Gβ₂ receptor signal? To address this question, we compared the maximum bradykinin-stimulated signals of the wild-type and the B₂Gβ₂ receptor at different receptor expression levels. With expression levels exceeding 2 pmol/mg protein, the signal of the B₂Gβ₂ receptor at a given receptor number was significantly increased compared with the wild-type receptor (Fig. 4F). This finding suggests that the C-terminally fused Gβ₂ subunit is functionally active and enhances the agonist-stimulated signal of the B₂Gβ₂ receptor similarly as seen with separately expressed Gβ(γ) subunits and the wild-type B₂ receptor.

Inhibition of Gβγ Exchange in the B₂Gβ₂ Receptor Fusion Protein.

We next asked whether the fused Gβ₂ subunit could suppress the enhancement of the B₂ receptor signal by Gβγ subunits and determined the effect of Gβ₁γ₂ coexpression on the bradykinin signal. Cotransfection with Gβ₁γ₂ increased the inositol phosphate production triggered by the wild-type B₂ receptor by up to 60% (Fig. 5A). In contrast, the signal of the B₂Gβ₂ receptor was not significantly enhanced by cotransfection of Gβ₁γ₂ (Fig. 5A). Thus, we conclude that, under the conditions applied (i) the overexpressed Gβγ did not enhance B₂ receptor-mediated signaling by direct synergistic activation of PLC, (ii) Gβγ needs access to the B₂ receptor, and (iii) this access can be blocked by fusion of Gβ₂ to the receptor’s C terminus. A small potentiating effect of expressed Gβ₁γ₂ could be observed when the fusion protein was expressed at lower levels (<1.5 pmol/mg protein), suggesting true competition between “free Gβγ” and the Gβ subunit fused to the B₂ receptor (data not shown).

Figure 5.

(A) Effect of Gβγ subunit expression on signaling by wild-type B₂ receptors and the B₂Gβ₂ fusion protein. The Gβ₁γ₂-mediated enhancement of the bradykinin-stimulated signal (100 nM) was determined on COS cells expressing the B₂ receptor (■) or the B₂Gβ₂ fusion protein (░⃞) together with Gβ₁γ₂. The enhancement is given as percent above control cells (0%): i.e., cells transfected with wild-type B₂ receptor or with B₂Gβ₂ fusion protein alone. (B and C) Suppression of crosstalk with Gα_i-coupled receptors in a B₂Gβ₂ fusion protein. COS cells expressing the B₂ receptor (■) or the B₂Gβ₂ fusion protein (░⃞) together with the A₁ adenosine (B) or the α_2C-adrenergic (C) receptor were stimulated with 100 nM bradykinin plus 1 μM N⁶-cyclohexyladenosine in the presence of 0.5 units/ml adenosine deaminase (B) or plus 1 μM medetomidine (C). Expression levels of the transfected B₂ receptors were 3.3 ± 0.4 pmol/mg protein. Expression levels of the transfected A₁ and α_2C receptor were 2.1 ± 0.4 and 1.8 ± 0.3 pmol/mg protein, respectively. Data are given as percent above control (0%): i.e., stimulation with bradykinin alone. All data are the means (±SEM) of three different experiments.

Costimulation of the B₂Gβ₂ Fusion Protein and Gα_i-Coupled Receptors.

Finally, we triggered the release of Gβγ subunits by stimulation of the two different Gα_i-coupled receptors used above, i.e., the A₁ adenosine and the α_2C adrenergic receptor, and determined again the effects on the signals generated by wild-type B₂ and B₂Gβ₂ receptors. By using wild-type B₂ receptors, the costimulation of either A₁ adenosine receptors (Fig. 5B) or of α_2C-adrenergic receptors (Fig. 5C) enhanced the maximal signal of the B₂ receptor. However, stimulation of either of the two Gα_i-coupled receptors did not alter the inositol phosphate signal of the B₂Gβ₂-fusion (Fig. 5 B and C). Thus, the Gβ₂ subunit fused to the C terminus of the B₂ receptor was capable of suppressing the A₁- or α_2C-receptor-mediated enhancement of the B₂ receptor’s signal.

Effect of Gβγ on B₂ Receptor-Stimulated G-Protein Activation.

The experiments suggest that Gβγ enhances a receptor-stimulated signal by direct interaction with the receptors. Two mechanisms are conceivable: (i) Gβγ may increase the number of functional holotrimeric G-proteins that can be activated by a given receptor, or (ii) the Gβγ-receptor interaction may enhance the receptor-stimulated Gα activation: i. e., guanine nucleotide exchange. Expression of Gβ₁γ₇ did not significantly increase the number of high affinity binding sites for [³H]bradykinin (Fig. 6A; t = 0). Because the number of high affinity binding sites reflects the interaction of functional holotrimeric G-proteins with an activated receptor, we conclude that an increase in the number of functional Gαβγ heterotrimers was not responsible for the Gβγ-mediated signal enhancement at the receptor level. By contrast, Gβ1γ7 expression accelerated the GTPγS-induced decrease in the number of high affinity [³H]bradykinin binding sites (Fig. 6A), and Gβγ subunits purified from bovine brain enhanced the B₂ receptor-stimulated GTPγS binding (Fig. 6B). Thus, Gβγ subunits accelerated the agonist-stimulated guanine nucleotide exchange: i.e., activation of the Gα subunit.

Figure 6.

(A) Effect of Gβγ subunit expression on binding of [³H]bradykinin. Membranes of COS cells transfected with B₂ receptor alone or with B₂ receptor together with Gβ₁γ₇ were incubated for 1 h at 4°C with 5 nM [³H]bradykinin. At t = 0, GTPγS (1 μM) was added, and the cells were placed at 30°C. Specific high affinity binding of [³H]bradykinin was determined at the time points indicated. Data are from a representative experiment (triplicates ± SD) that was reproduced three times with similar results. (B) B₂ receptor-stimulated GTPγS binding. Membranes of B₂ receptor expressing HEK-293 cells (5 μg protein/50 μl) were incubated in the absence (B₂) or presence (B₂+Gβγ) of 50 nM Gβγ purified from bovine brain, and bradykinin-stimulated [³⁵S]GTPγS binding was determined at the time points indicated.

DISCUSSION

Signal adaptation to new inputs is the basis for efficient signaling networks. Mechanisms to sense and to adapt to stimulatory or inhibitory inputs are generally summarized under the term “crosstalk.” An increasing number of different crosstalk mechanisms is emerging between G-protein-coupled receptors involving all levels of the signaling cascade(s). A crosstalk mechanism on the initial step of signal generation, the receptor-G-protein interface has so far not been demonstrated. In fact, antisense experiments demonstrating preferential coupling of specific receptors to specific G-protein heterotrimers in intact cells (1–4) might appear to preclude such a cross-talk at the receptor/G-protein level.

Stimulation of two Gα_i-coupled receptors, the α_2C-adrenergic and the A₁ adenosine receptor, enhanced the potency and efficacy of the signal generated by two Gα_q-coupled receptors, a UTP-preferring P2Y and the B₂ receptor. This signal enhancement was very similar to that caused by overexpression of Gβγ and is most likely caused by release of Gβγ from Gα_i. It seems to be exerted proximal to Gα_q, i.e., at the receptor level, because (i) an N-terminally truncated Gβ₃ mutant was still capable of increasing basal and AlF₄⁻-stimulated signals but had no effect on the receptor-mediated signals, and (ii) a Gβ₂ subunit fused to the B₂ receptor substituted for the signal enhancement caused by stimulation of Gα_i-coupled receptors or by Gβγ. Thus, access of Gβγ to the receptors appears to be essential for this mechanism, and this access is (partially) precluded by the fused Gβ₂-moiety. The Gα_i-coupled receptors appear to provide the Gα_q-coupled receptors with additional Gβγ subunits (Fig. 7). Gβγ subunits seem to promote Gα activation by accelerating the rate of receptor-stimulated GTP-binding (Fig. 7). A similar effect of Gβγ subunits has been observed in vitro for the rhodopsin-stimulated Gα_t activation (29). Because Gβγ subunits suppress the GDP/GTP exchange in the absence of an activated receptor by direct Gα interaction (30), the Gβγ-receptor interaction seems to be a prerequisite for the Gβγ-mediated enhancement of Gα activation. Although Gβγ subunits may enhance a receptor-stimulated signal at several steps of the signaling cascade (cf. introduction), a major component may be contributed by the newly discovered mode of receptor crosstalk. This crosstalk relies on Gβγ exchange between Gα_i-coupled and Gα_q-coupled receptors, thereby accelerating receptor-stimulated GTP-binding of Gα_q (Fig. 7).

Figure 7.

Model of Gβγ-mediated crosstalk between Gα_i and Gα_q-coupled receptors. Gβγ transfer from an activated Gα_i coupled receptor (R_i*) to an activated Gα_q-coupled receptor (R_q*) enhances the receptor-stimulated GDP/GTP exchange of Gα_q.

Although the mode of receptor crosstalk between Gα_i and Gα_q-coupled receptors was analyzed in COS cells, the observed mechanism seems to be common to a variety of different Gα_i- and Gα_q-coupled receptors in recombinant and primary cell systems (15, 16, 31). It did not depend on the overexpression of Gβγ subunits because the endogenous levels of Gβγ subunits in COS cells were sufficient to promote receptor crosstalk. In intact nonstimulated cells, different receptors appear to be preferentially coupled to different heterotrimeric G-proteins (1–5). Therefore, efficient Gβγ-mediated crosstalk must rely on a lack of biochemical coupling specificity between receptors and different Gβγ dimers. Indeed, combinations of Gβ_1–3 with Gγ_2–7 were all about equally efficient in enhancing the UTP- and bradykinin-stimulated signals. Complexes with Gγ₅ were only slightly less efficient (70–90% of maximum), an effect that is most likely attributable to a less efficient interaction with phospholipase C (U.Q., unpublished data) and not to a less efficient interaction with the receptors. One might argue that overexpression of Gβγ subunits may cover small differences in potency between several combinations, but the low efficiency of Gβ₁γ₁ is in good agreement with earlier results in which Gβ₁γ₁ was 10-fold less potent than β₁γ₂ in stimulating PLCβ₃ activity (9), the main component of Gβγ-stimulated PLC activity in COS cells (32). Thus, the receptor-Gβγ interaction lacks coupling specificity and thus fulfills the proposed prerequisite for efficient Gβγ-based crosstalk.

Gβγ subunits do not seem to exchange per se; they must be mobilized by receptor activation. Receptor signaling via Gα_i and Gα_q often leads to opposing effects in the same cell: e.g., stimulation of A₁ adenosine receptors induces vasoconstrictor responses in pulmonary vascular beads whereas B₂ receptors mediate vasodilation (33); stimulation of α_2C-adrenergic or A₁ adenosine receptors suppresses norepinephrine release from sympathetic nerve endings, which is stimulated by activation of B₂ receptors (22, 34). Therefore, the concomitant release of Gβγ subunits after Gα_i-receptor stimulation may dampen the generated signal by lowering the threshold for the antagonizing Gα_q-coupled receptor via Gβγ transfer. Thus, the Gβγ-mediated mode of receptor crosstalk may act as a feedback loop to balance inhibitory and stimulatory inputs.

ABBREVIATION

PLC

phospholipase C