Macromolecular Composition Dictates Receptor and G Protein Selectivity of Regulator of G Protein Signaling (RGS) 7 and 9-2 Protein Complexes in Living Cells*

Background: RGS7 and RGS9-2 regulate G protein signaling in the striatum, but the selectivity of their action is largely unknown. Results: RGS protein complexes show distinct patterns of receptor and G protein selectivity. Conclusion: Macromolecular composition dictates receptor and G protein selectivity of the RGS7 and RGS9-2 protein complexes. Significance: These data demonstrate novel mechanisms contributing to the regulation of striatal G protein signaling. Regulator of G protein signaling (RGS) proteins play essential roles in the regulation of signaling via G protein-coupled receptors (GPCRs). With hundreds of GPCRs and dozens of G proteins, it is important to understand how RGS regulates selective GPCR-G protein signaling. In neurons of the striatum, two RGS proteins, RGS7 and RGS9-2, regulate signaling by μ-opioid receptor (MOR) and dopamine D2 receptor (D2R) and are implicated in drug addiction, movement disorders, and nociception. Both proteins form trimeric complexes with the atypical G protein β subunit Gβ5 and a membrane anchor, R7BP. In this study, we examined GTPase-accelerating protein (GAP) activity as well as Gα and GPCR selectivity of RGS7 and RGS9-2 complexes in live cells using a bioluminescence resonance energy transfer-based assay that monitors dissociation of G protein subunits. We showed that RGS9-2/Gβ5 regulated both Gi and Go with a bias toward Go, but RGS7/Gβ5 could serve as a GAP only for Go. Interestingly, R7BP enhanced GAP activity of RGS7 and RGS9-2 toward Go and Gi and enabled RGS7 to regulate Gi signaling. Neither RGS7 nor RGS9-2 had any activity toward Gz, Gs, or Gq in the absence or presence of R7BP. We also observed no effect of GPCRs (MOR and D2R) on the G protein bias of R7 RGS proteins. However, the GAP activity of RGS9-2 showed a strong receptor preference for D2R over MOR. Finally, RGS7 displayed an four times greater GAP activity relative to RGS9-2. These findings illustrate the principles involved in establishing G protein and GPCR selectivity of striatal RGS proteins.

Signal transduction through G protein-coupled receptors (GPCRs) 2 regulates fundamental processes in the nervous sys-tem, including neuronal excitability and neurotransmitter release (1). GPCRs activate heterotrimeric G proteins, which in turn engage a wide range of the intracellular effectors to produce a cellular response. Activation of G proteins entails their binding to GTP and resulting dissociation into G␣-GTP and G␤␥ subunits. The extent and duration of signaling in GPCR pathways is critically controlled by the regulator of G protein signaling (RGS) proteins that limit G protein activity (2,3). RGS proteins bind directly to activated G␣ and facilitate the GTP hydrolysis, thus serving as GTPase-accelerating proteins (GAPs).
In humans, 17 G␣, about 865 GPCR, and ϳ30 RGS genes have been identified (4 -6). With most of the components expressed in the nervous system, this forms a formidable array of possible combinations. However, activation of individual GPCR pathways often produces unique cellular and behavioral responses. Understanding the mechanisms of this signaling selectivity is one of the biggest challenges in studying neuronal GPCR pathways.
Neurons of the striatum, a nucleus that plays a major role in reward behavior and motor control, express a number of GPCRs that respond to many neurotransmitters, including dopamine, opioids, serotonin, and acetylcholine (7,8). Several studies have demonstrated that the long splice isoform of RGS9 (RGS9-2) serves as a critical GAP in these neurons (9,10). In particular, RGS9-2 has been shown to regulate signaling downstream from D2 dopamine (D2R) and -opioid (MOR) receptors and has been implicated in drug addiction and movement disorders (11)(12)(13)(14)(15). However, no studies directly examined the impact of RGS9-2 on G protein dynamics activated by D2R and MOR.
Recent behavioral studies implicated another RGS protein in the striatum, RGS7, in controlling the effects of addictive drugs and suggested that it may be differentially involved in controlling MOR and D2R signaling (16). Both RGS9-2 and RGS7 share extensive homology in their macromolecular organization. In addition to the catalytic RGS domain, they possess the N-terminal Dishevelled, EGL-10, Pleckstrin/Dishevelled, EGL-10, Pleckstrin helical extension module and a G protein ␥ sub-unit-like domain (10). The G protein ␥ subunit-like domain forms a constitutive complex with the atypical G protein ␤ subunit, G␤5 (17,18), and both RGS7 and RGS9-2 exist as obligatory dimers with G␤5 (19). In the striatum, RGS/G␤5 dimers associate with membrane anchor R7BP (20) that recruits them to the plasma membrane and potentiates the GAP activity (21,22). What remains completely unexplored is the relative activity and selectivity of the RGS7 and RGS9-2 complexes, as well as the role of R7BP in this process.
In this study, we examined the ability of the RGS7/G␤5 and RGS9-2/G␤5 complexes to regulate G protein signaling by MOR and D2R under the native environment of living cells using a bioluminescence resonance energy transfer (BRET)based assay. We report a marked differences in the catalytic activity of complexes as well as their G protein and GPCR selectivity, depending on their macromolecular composition. Our findings illustrate mechanisms for establishing the selective regulation of striatal GPCR signaling pathways.
Fast Kinetic BRET Assay-Agonist-dependent cellular measurements of BRET between masGRK3ct-Rluc8 and G␤1␥2-Venus were performed to visualize the action of G protein signaling in living cells, as described previously, with slight modifications (25). 16 to 24 h post-transfection, HEK293T/17 cells were washed once with PBS containing 5 mM EDTA (EDTA/PBS) and detached by incubation in EDTA/PBS at room temperature for 10 min. Cells were harvested with centrifugation at 500 g for 5 min and resuspended in PBS containing 0.5 mM MgCl 2 and 0.1% glucose (BRET buffer). Approximately 50,000 -100,000 cells/well were distributed in 96-well flat-bottomed white microplates (Greiner Bio-One). The Rluc substrate, coelenterazine-h (Nanolight Technologies), was dissolved in acidified alcohol at a final concentration of 5 mM and stored at Ϫ20°C. Acidified alcohol was prepared by adding 200 l of 3N HCl to 10 ml of ethanol. Aliquots were dissolved in BRET buffer immediately before use and added to cell suspension at a final concentration of 5 M. BRET measurements were made using a microplate reader (POLARstar Omega, BMG Labtech) equipped with two emission photomultiplier tubes, allowing us to detect two emissions simultaneously with the highest possible resolution of 50 milliseconds for every data point. All measurements were performed at room temperature. The BRET signal is determined by calculating the ration of the light emitted by G␤1␥2-Venus (535 nm) over the light emitted by masGRK3ct-Rluc8 (475 nm). The average base-line value (basal R) recorded prior to agonist stimulation was subtracted from BRET signal values, and the resulting difference (⌬R) was normalized against the maximal ⌬R value (Rmax) recorded upon agonist stimulation. The rate constants (1/) of the activation and deactivation phases were obtained by fitting a single exponential curve to the traces. k GAP rate constants were determined by subtracting the basal deactivation rate (k app ) from the deactivation rate measured in the presence of exogenous RGS protein. Obtained k GAP rate constants were used to quantify GAP activity.
Western Blotting-For each sample, ϳ5,000,000 cells were lysed in 500 l of sample buffer (125 mM Tris (pH 6.8), 4 M urea, 4% SDS, 10% 2-mercaptoethanol, 20% glycerol, 0.16 mg/ml bromphenol blue). Western blot analysis of proteins was performed following SDS-PAGE. Blots were blocked with 5% skim milk in PBS containing 0.1% Tween 20 (PBST) for 30 min at room temperature, followed by a 90-min incubation with specific antibodies diluted in PBST containing 1% skim milk. Blots were washed in PBST and incubated for 45 min with a 1:10,000 dilution of secondary antibodies conjugated with horseradish peroxidase in PBST containing 1% skim milk. Proteins were visualized on x-ray film by SuperSignal West Femto substrate (Pierce). Band densities were quantified using ImageJ software by measuring the integrated intensity. The relative expression level of RGS proteins was determined by subtracting the background densities in the absence of exogenous RGS proteins and normalizing the resulting value as a fraction of the brightest band intensity expressing the maximal amount of RGS protein.
CRE-Luciferase Reporter Gene Assays-HEK293T/17 cells were transfected with CRE-luc2P reporter (Promega), MOR, G␣i1, G␤1 and G␥2, or G␤5 with or without R7 RGS at a 1:1:1: 1:1:1:1 ratio between cDNA constructs using Lipofectamine LTX reagent in 96-well plate. 16 h after transfection, cells were treated with 50 nM isoproterenol together with serial doses of morphine for 5 h. The level of expressed luciferase were determined using a Bright-Glo luciferase assay kit (Promega) according to the instructions of the manufacturer.
Statistical Analysis-Linear regression was used to relate the k GAP value to the expression level of RGS7 or RGS9-2 proteins. To compare the activities of the RGS proteins on Go versus Gi (Fig. 5) and their activities in the absence versus presence of R7BP (Figs. 6 and 7), the differences in the slopes of the regression lines were evaluated by calculating a p value (two-tailed) testing the null hypothesis that the slopes are all identical using GraphPad Prism 5. Kruskal-Wallis one-way analysis of variance on ranks followed by Tukey's post hoc test was performed to compare k GAP values in Fig. 8E using SigmaPlot 11. IC 50 values in Fig. 9C were compared by one-way analysis of variance followed by Tukey's post hoc test using SigmaPlot 11.

Live Cell Receptor-based Assays Allow Examination of the GAP Activity of R7 RGS Proteins in a Physiological Context-To
visualize RGS action in living cells, we reconstituted HEK293T cells with GPCRs, G␣ subunits, and BRET sensors (G␤␥-Venus and masGRK3ct-Rluc8, recently developed by Hollins et al. (25) (Fig. 1A)). In this assay system, activation of a GPCR by an agonist promotes the interaction of the Venus-tagged G␤␥ subunits with the Rluc8-tagged masGRK3ct reporter producing the BRET signal. Conversely, application of an antagonist quenches GPCR-driven G protein activation and results in BRET signal decay. Indeed, activation of the heterologously expressed D2R by dopamine resulted in generation of the robust BRET response and the addition of the haloperidol deactivated the signal in a time-dependent manner (Fig. 1B). The rise and decay of the BRET signal are well described by single exponential fitting, allowing us to obtain rate constants for the activation and deactivation phases, respectively. No significant agonist-elicited response was observed when D2R or G␣o was not transfected (Fig. 1C), indicating that the BRET signal recorded in the system originates from the functional coupling of the exogenously expressed receptors and G proteins and not from the endogenous proteins contained in HEK293T cells.
We next performed a series of control experiments to demonstrate the utility of the assay in studying the GAP activity of the R7 RGS proteins. Because the GAP action of RGS proteins promotes heterotrimer reformation, they are expected to influence the kinetics of the deactivation phase of the response. However, the concurrent presence of GPCR in the system makes it necessary to ensure that the kinetics of the deactivation phase are determined by the RGS action and are not influenced by changes in GPCR activity. First, we transfected cells with increasing amounts of D2R construct and stimulated cells by application of a saturating concentration of dopamine ( Fig.  1, D-F). An increase in the amount of D2R did not affect the amplitude of the response (Fig. 1D), suggesting an equal extent of the total G protein mobilization in the cells, but accelerated the onset kinetics consistent with the enhanced speed of G protein activation (Fig. 1E). Importantly, an increase in functional D2R had no effect on deactivation kinetics (Fig. 1F), indicating that the GPCR concentration does not influence the antagonist-induced termination of the response. Secondly, we applied different concentrations of dopamine to cells expressing a fixed amount of D2R (Fig. 1, G-I). We observed a typical sigmoidal dose-response curve of the response amplitude (EC 50 ϭ 1.91 ϫ 10 Ϫ7 Ϯ 1.67 ϫ 10 Ϫ8 M), indicating that an increase in the agonist concentration results in an increase in the pool of the activated G proteins. As expected, enhanced GPCR activation also resulted in the increase in the rate of G protein activation (Fig.  1H). However no changes in the deactivation phase of the response were noted (Fig. 1I). These data indicate that the amount of active G protein also does not change the deactivation rates. Finally, we tested the effect of increasing the RGS concentration on the deactivation kinetics of D2R-Go signaling ( Fig. 1, J-L). In contrast to the manipulations with GPCR and G protein concentration, an increase in the RGS7 concentration substantially accelerated the deactivation rates of the response (Fig. 1, J and K). In fact, within the concentration range tested, the calculated GAP activity of RGS7 showed a clear linear relationship with its expression level (Fig. 1L). The final set of control experiments ensured that the expression of components of the RGS complex did not affect the levels of the receptor, G␣ subunits, or BRET sensors in cells (Fig. 2). We thus conclude that, under the assay conditions, GTP hydrolysis by G proteins is the rate-limiting step that dictates the deactivation phase of the response. The function of the R7 RGS proteins can thus be quantitatively analyzed by measuring the deactivation kinetics that show no sensitivity to fluctuations in GPCR and/or G protein concentration.
RGS7 and RGS9-2 Complexes Selectively Regulate the G␣i/o Subfamily in the absence or presence of R7BP-Members of the R7 RGS family have been shown previously to be selective GAPs for the G␣i/o proteins in the in vitro biochemical assays (26). However, their selectivity was never examined in the physiological context of living cells. Furthermore, on the basis of the genetic evidence from Caenorhabditis elegans, it was proposed recently that R7BP might unlock the GAP activity of the R7 RGS proteins toward other G␣ subunits, e.g. G␣q (27). We therefore used the in-cell BRET assay to re-examine the regulation of the G␣ GTPase by R7 RGS complexes in living cells. In these experiments, we chose a panel of representative GPCRs well known to activate each of the G␣ subunits examined, e.g. D1R for G␣s, M1R for G␣q, D2R for G␣o and MOR for G␣z. Consistent with the in vitro studies, we found that, in cells, both RGS7/G␤5 and RGS9-2/G␤5 potently terminated a D2Rdriven response via G␣o, a representative member of the G␣i/o family (Fig. 3, A, D, E, and H). However, we observed no regulation of the M1R-G␣q or D1R-G␣s response termination by either RGS7/G␤5 (Fig. 3, B, C, and D) or RGS9-2/G␤5 (F, G, and H). These data indicate that the RGS7/G␤5 and RGS9-2/G␤5 dimers are Gi/o subfamily-selective regulators in living cells.

G Protein and GPCR Selectivity of RGS Complexes
Next, we examined the possibility that R7BP may enable G␣q or G␣s regulation by RGS7 or RGS9-2 complexes. Again, we found that trimeric RGS7/G␤5/R7BP and RGS9-2/G␤5/R7BP complexes potently regulated G␣o deactivation but had no effect on the deactivation rates of G␣q and G␣s (Fig. 3, I-P). Because expression levels of RGS7 or RGS9-2 were found to be similar across cells reconstituted with various G protein subunits (data not shown), the lack of GAP activity toward G␣q and G␣s cannot be attributed to a lower expression of RGS7 and RGS9-2. Thus, we conclude that complex formation with R7BP does not influence subfamily selectivity and that the RGS7/G␤5/R7BP and RGS9-2/G␤5/R7BP trimers are also Gi/o subfamily-selective GAPs.
Varying G Protein Selectivity of the RGS7 and RGS9-2 Complexes within the G␣i/o Family-In addition to G␣o, the G␣i/o family contains the highly homologous proteins G␣i1-3 and the atypical subunit G␣z, which is characterized by an extremely slow intrinsic GTP hydrolysis rate (28). Because no previous studies examined the regulation of G␣z GTPase by R7 RGS proteins, we addressed this question using cell-based BRET assays. Morphine application induced a robust BRET signal from the cells reconstituted with MOR and G␣z. Consistent with biochemical measurements, G␣z responses showed very slow deactivation kinetics upon termination of signaling at MOR, with a rate constant of 3.7 Ϯ 0.1 ϫ 10 Ϫ3 s Ϫ1 . These slow deactivation kinetics were completely unaffected by the addition of either the RGS9-2/G␤5 or RGS7/G␤5 complex both in the presence or absence of R7BP (Fig. 4). These observations suggest that R7 RGS complexes do not regulate G␣z signaling.
We next examined the impact of RGS9-2 and RGS7 on G␣i deactivation in comparison to G␣o. In these experiments, we titrated the amount of RGS proteins and normalized their expression levels by post hoc quantitative Western blotting. We used the k GAP versus RGS concentration slope as a measure of RGS catalytic efficiency. As evident from Fig. 5A, RGS9-2/G␤5 can regulate both G␣o and G␣i1 in a system containing D2R. Comparison of the two slopes indicates that RGS9-2/G␤5 shows an ϳ3-fold greater preference for G␣o over G␣i1. In contrast, although RGS7/G␤5 also effectively terminated G␣omediated responses, it had no detectable activity toward G␣i1 in a D2R-based system (Fig. 5B). We confirmed that the expression levels of G␣o and G␣i1 are similar (Fig. 5, E and F), indicating that the G protein selectivity of these two RGS proteins is not due to different levels of G protein expression. Because in  (n ϭ 4). K, the expression levels of RGS7 in cells used in J were examined by Western blotting with an RGS7-specific antibody. L, the deactivation rate constant measured in the absence of RGS7 (0.1427 Ϯ 0.0028 s) was subtracted from values measured in the presence of RGS7, and the resulting k GAP parameters were plotted against the expression level of RGS7 measured in K (n ϭ 4).

FIGURE 2. Overexpression of the R7 RGS complex does not influence the expression levels of signaling components reconstituted in HEK293T cells.
HEK293T cells were transfected with D2R, G␣o, Venus-G␤␥, masGRK3ct-Rluc8, and G␤5 with or without R7 RGS and R7BP. The expression levels of signaling components were analyzed by Western blotting with the indicated specific antibodies. Anti-Rluc (Renilla luciferase) antibody was used to detect masGRK3ct-Rluc8. Because transfection of G␤5 without R7 RGS does not change the parameters, Rmax, and activation and deactivation rates (data not shown), G␤5 was transfected in all conditions. vivo R7 RGS proteins also regulate MOR signaling, we next tested the possible effect of GPCR on G␣ selectivity of RGS proteins. As with D2R signaling, RGS9-2/G␤5 exerted a more efficacious GAP activity toward G␣o relative to G␣i1 (3-fold) (Fig. 5C). Again, RGS7/G␤5 selectively regulated G␣o but not G␣i signaling when MOR was used instead of D2R to drive the responses (Fig. 5D). These data indicate that the RGS9-2/G␤5 dimer is a GAP for both G␣o and G␣i, with a prefer-ence toward G␣o, that the RGS7/G␤5 dimer is strictly selective for G␣o and does not regulate G␣i, and that G protein selectivity of these RGS proteins does not depend on the GPCR.
R7BP Potentiates the GAP Activity of RGS7 and RGS9-2 and Enables the Regulation of G␣i by RGS7 with Both D2R and MOR-We examined the effects of R7BP on the activity of RGS7 and RGS9-2 by comparing their GAP activity toward G␣o and G␣i1 in the D2R receptor system. Coexpression with R7BP potentiated the GAP activity of both RGS7 and RGS9-2 for G␣o. The extent of the regulation was similar and reached ϳ2-fold (Fig. 6, A and B). This increase in activity was independent from the regulation of the RGS9-2 expression levels by R7BP because calculated k GAP values were normalized to the relative expression levels of the proteins. A similar stimulatory effect of R7BP (2.5-fold) was observed for RGS9-2 when G␣i was used in the assay instead of G␣o (Fig. 6C). Strikingly, coexpression with R7BP dramatically affected the ability of the RGS7/G␤5 complex to regulate G␣i, essentially resulting in an all-or-nothing effect (Fig. 6D). Although, in the absence of R7BP, we detected no GAP activity of RGS7/G␤5 toward G␣i1, the RGS7/G␤5/R7BP trimer was capable of efficiently regulating G␣i inactivation, with the k GAP values approaching those for the RGS9-2/G␤5/R7BP complex.
We next tested the possible effect of GPCRs on the function of R7BP. For this purpose, we switched receptors from D2R to MOR (Fig. 7). With MOR, we observed similar effects of R7BP on GAP activity as with the cells transfected with D2R. R7BP potentiated RGS9-2 GAP activity toward G␣o and G␣i1 ϳ2-fold (Fig. 7, A and C). Likewise, R7BP enhanced RGS7 activity toward G␣o 2.5-fold (Fig. 7B) and played a permissive role in enabling the regulation of G␣i1 GTPase activity (D). Thus, we conclude that GPCRs do not have a significant effect on the ability of R7BP to augment the GAP activity of both RGS7 and RGS9-2.
RGS9-2 Is Less Efficacious but More Selective GAP Relative to RGS7-Next, we addressed questions pertaining to the relative efficiencies of RGS7 versus RGS9-2 in G protein deactivation and the possible GPCR selectivity of their effects (Fig. 8). To enable such comparisons, we obtained absolute quantitative values for the RGS activity. Direct comparison of RGS7/G␤5 and RGS9-2/G␤5 indicate that they accelerate deactivation rates in D2R-and MOR-based systems to a different extent (Fig.  8, A and B). The expression levels of RGS7 and RGS9-2 were quantified by Western blot analysis with purified RGS proteins as standards (Fig. 8, C and D). Given the linear relationship between RGS concentration and G protein deactivation rates, these values were used to normalize k GAP values to derive the specific activity for each RGS protein in each GPCR system. The results allow direct comparisons of RGS efficiencies between MOR and D2R. As evident from such an analysis, RGS7 produced a 3-and 4.5-fold higher activity than RGS9-2 on D2R and MOR signaling, respectively (Fig. 8E). A more efficacious GAP activity of RGS7 is likely explained by its higher affinity toward the transition state of the G␣o, as evidenced by the pull-down assay between recombinant R7 RGS/G␤5 complexes and native brain lysates (Fig. 8F). Thus, we conclude that RGS7 is a more potent GAP than RGS9-2, irrespective of the GPCR used in the assay. However, although RGS7 exerted a similar GAP activity in both the D2R and MOR systems, RGS9-2 was about 2-fold more effective in deactivating D2R relative to MOR signaling (Fig.  8E). Thus, RGS9-2 exhibits GPCR selectivity and preferentially regulates D2R signaling.
RGS7 and RGS9-2 Deferentially Control the G␣i-mediated Inhibition of Adenylate Cyclase Activity-To examine how the selectivity of RGS7 and RGS9-2 on G protein inhibition propagates the regulation of downstream effector signaling, we chose an adenylyl cyclase (AC) system for its central role in cellular signaling. AC is stimulated by G␣s and inhibited by G␣i, thus integrating G protein inputs (Fig. 9A). Using this system, we analyzed the effects of RGS9-2 and RGS7 on the ability of MOR to suppress cAMP production using a CRE-luciferase reporter construct. Stimulation of MOR with morphine caused a dosedependent inhibition of ␤2AR-agonist isoproterenol-mediated CRE-luciferase induction (Fig. 9B). Cotransfection of RGS9-2 resulted in the rightward shift of the dose-response curve, increasing the IC 50 values ϳ3-fold (8.38 Ϯ 0.68 nM to 25.82 Ϯ 2.39 nM) (Fig. 9C). This indicates that RGS9-2 reduces the potency of MOR-G␣i-AC signaling and is consistent with the action of RGS9-2 as a negative regulator of G␣i. In contrast, cotransfection of RGS7 did not significantly affect MOR signaling to AC (Fig. 9, B and C), consistent with the lack of the RGS7 activity on G␣i revealed in the BRET assays. Thus, these data, together with data from the BRET assays, indicate that RGS proteins differentially control GPCR-mediated signaling to downstream effectors, consistent with their G protein selectivity profile.

DISCUSSION
The main result of this study is the establishment of the selectivity for two major striatal RGS proteins in their ability to regulate physiologically relevant GPCRs in the native environment of a living cell. The GAP activity of RGS proteins is usually assayed in in vitro systems with purified components where RGS proteins, and often only their catalytic domains, are studied in isolation from protein-protein interactions, receptors, and the membrane environment. Under those conditions, RGS proteins display very few differences in their substrate selectivity and specific activity. However, it is becoming increasingly appreciated that, in vivo, RGS proteins function in a tight association with other components of the GPCR signaling cascades and exist in larger macromolecular complexes (29). For example, both RGS7 and RGS9 form complexes with a range of partners that include G␤5; the membrane anchors R7BP and R9AP (30); and the GPCRs mGlur6 (31), D2R (13,32), MOR (33)(34)(35), m3 muscarinic (36), and GPR158/179 (37). Nevertheless, how these interactions shape RGS action in cells is poorly understood. We used a cell-based BRET assay system to study the influence of multisubunit RGS complexes on the kinetics of G protein subunit reassociation following termination of GPCR activity. We developed quantitative measures of RGS protein GAP activity in this system and applied it to investigate the FIGURE 6. R7BP augments the GAP activity of RGS7 and RGS9-2 toward G␣i1 and G␣o in the D2R-based system. G␣i1 and G␣o were reconstituted in HEK293T cells with D2R, and the GAP activity of RGS7 and RGS9-2 was examined in the absence (E) or presence (•) of R7BP. Changes in the k GAP values are plotted as a function of RGS concentration. Representative of two to three independent experiments yielding similar results are shown. Slope values obtained from linear regression analysis are 4.8 ϫ 10 Ϫ3 Ϯ 3.1 ϫ 10 Ϫ4 for D2R-Go-RGS9-2-R7BP, 1.8 ϫ 10 Ϫ3 Ϯ 1.1 ϫ 10 Ϫ4 for D2R-Go-RGS9-2, 3.3 ϫ 10 Ϫ3 Ϯ 5.9 ϫ 10 Ϫ4 for D2R-Go-RGS7-R7BP, 1.5 ϫ 10 Ϫ3 Ϯ 8.7 ϫ 10 Ϫ5 for D2R-Go-RGS7, 6.7 ϫ 10 Ϫ4 Ϯ 4.2 ϫ 10 Ϫ5 for D2R-Gi-RGS9-2-R7BP, 3.7 ϫ 10 Ϫ4 Ϯ 3.7 ϫ 10 Ϫ5 for D2R-Gi-RGS9-2, 9.8 ϫ 10 Ϫ4 Ϯ 8.8 ϫ 10 Ϫ5 for D2R-Gi-RGS7-R7BP, and 3.1 ϫ 10 Ϫ5 Ϯ 8.5 ϫ 10 Ϫ5 for D2R-Gi-RGS7. Four to six replicate samples were used for obtaining each data point. activity and selectivity of the striatal RGS proteins RGS7 and RGS9-2.
The following are the key conclusions of our study (Fig. 10). First, we show that in the cellular environment RGS9-2/G␤5 and RGS7/G␤5 display a strong preference for G␣o over G␣i. These results are in overall agreement with published in vitro data (26,38). The RGS7 complex showed the greatest selectivity for G␣o and was completely unable to inactivate G␣i in the absence of R7BP. Second, we demonstrate that the GAP activity of the RGS9-2/G␤5 complex shows a receptor preference for D2R over MOR in regulation of the G␣o deactivation. To our knowledge, this is the first clear example of GPCR selectivity of the R7 RGS action. Although the mechanisms behind this receptor preference of the RGS9-2 complex need to be established, we speculate that they are likely determined by selective interactions of RGS9-2 with the receptors, as suggested from the studies on RGS4, that selectively interacted with the ␦-opi-oid receptor over the -opioid receptor (39). Interestingly, no receptor preference was revealed for the RGS7/G␤5 action. Third, we found that R7BP acted universally to potentiate the action of both RGS7 and RGS9-2 on both G␣i and G␣o and with both D2R and MOR. It had the most pronounced all-ornothing effect on the ability of RGS7 to regulate G␣i, essentially switching it on. Although the design of our study did not allow distinguishing between allosteric effects and the general effects of positioning RGS complexes on the plasma membrane, we think that both mechanisms are likely involved in the action of R7BP, as exemplified in the studies on related membrane anchor R9AP (24,41). Finally, we report that in living cells, RGS7 shows a much more potent activity relative to the RGS9-2 complex. In contrast, previous in vitro observations with purified proteins reported approximately equal catalytic activities of these two proteins (26). This illustrates the importance of considering a native, physiologically relevant FIGURE 7. R7BP augments the GAP activity of RGS7 and RGS9-2 toward G␣i1 and G␣o in the MOR-based system. Gi1 and Go signaling were reconstituted in HEK293T cells with MOR, and the GAP activity of RGS7 and RGS9-2 was examined in the absence (E) or presence (•) of R7BP. Changes in the k GAP values are plotted as a function of RGS concentration. Representative of two to three independent experiments yielding similar results are shown. The slope values obtained from linear regression analysis are 1.7 ϫ 10 Ϫ3 Ϯ 3.4 ϫ 10 Ϫ4 for MOR-Go-RGS9-2-R7BP, 6.0 ϫ 10 Ϫ4 Ϯ 1.0 ϫ 10 Ϫ4 for MOR-Go-RGS9-2, 3.6 ϫ 10 Ϫ3 Ϯ 6.5 ϫ 10 Ϫ4 for MOR-Go-RGS7-R7BP, 1.8 ϫ 10 Ϫ3 Ϯ 2.9 ϫ 10 Ϫ4 for MOR-Go-RGS7, 6.7 ϫ 10 Ϫ4 Ϯ 4.2 ϫ 10 Ϫ5 for MOR-Gi-RGS9-2-R7BP, 3.7 ϫ 10 Ϫ4 Ϯ 3.7 ϫ 10 Ϫ5 for MOR-Gi-RGS9-2, 3.0 ϫ 10 Ϫ4 Ϯ 5.9 ϫ 10 Ϫ5 for MOR-Gi-RGS7-R7BP, and 1.2 ϫ 10 Ϫ5 Ϯ 6.8 ϫ 10 Ϫ5 for MOR-Gi-RGS7. Four to six replicate samples were used for obtaining each data point. cellular environment when comparing the activities of complex RGS proteins.
It is interesting to consider the observed differences in selectivity and activity of RGS complexes in the context of striatal G protein signaling regulation. We have reported recently that changes in neuronal excitability and oxygenation trigger a remodeling of RGS complexes in the striatum (40). During this remodeling, RGS9-2 undergoes degradation, and vacated R7BP FIGURE 8. Comparison of catalytic activities and receptor preferences of RGS7 and RGS9-2 complexes. A and B, D2R (A) and MOR (B) signaling was reconstituted in HEK293T cells with Go, and the GAP activity of the RGS7/G␤5 and RGS9-2/G␤5 dimers were examined. Trace lines represent the exponential fit of the deactivation phase. C and D, the protein levels of RGS7 (C) and RGS9-2 (D) in the transfected cells used for the BRET assay were determined by Western blotting. Band intensities (E) obtained from recombinant proteins were used to generate a standard curve. Band intensities (1.49 Ϯ 0.05 for RGS7 and 1.97 Ϯ 0.08 for RGS9-2) from the triplicate samples in question were plotted on the calibration curve (E) and used to determine the amount of RGS7 and RGS9-2. E, the k GAP values obtained from the BRET assay were divided by the amount of RGS proteins determined by Western blot analysis to compare the activity of RGS7 and RGS9-2. Six replicate samples were used for each experiment. The experiment was performed independently three times, and all showed similar results. Error bars represent mean Ϯ S.E. *, p Ͻ 0.05, Kruskal-Wallis one-way analysis of variance on ranks followed by Tukey's post hoc test; ns, not significant. F, purified His-tagged RGS7/G␤5 and RGS9-2/G␤5 were incubated with mouse brain membrane fractions treated with GDP and AlF 4 Ϫ . After detergent treatment, complexes containing His-tagged proteins were purified with nickel-nitrilotriacetic acid chromatography and analyzed by Western blot analysis with G␣o and G␤5 antibodies. recruits RGS7 to the plasma membrane compartments. Furthermore, multiple studies demonstrated that exposure to addictive drugs (e.g. cocaine, morphine, amphetamine) that influence D2R and MOR signaling also changes RGS9-2 expression (9) and, thus, likely influences the composition of RGS complexes in striatal neurons. Taken together with the results of this study, these observations suggest a model where remodeling of RGS complexes is used to adjust the strength and selectivity of striatal G protein signaling. For example, an increase in dopamine and opioid signaling may be counteracted by tweaking the GAP complex (41) to substitute less efficient RGS9-2 for the stronger RGS7 catalytic subunit. Substituting a more selective GAP for a less selective one will likely also affect the relative balance of D2R versus MOR signaling in the striatum. It thus appears that RGS protein complexes are more than just blunt indiscriminate tools and rather contribute to the homeostatic scaling of G protein signaling in a GPCR-and G protein-selective fashion. FIGURE 9. Differential effects of RGS7 and RGS9-2 on G␣i-mediated inhibition of adenylate cyclase activity. A, schematic cross-talk signaling pathway of endogenous ␤2-adrenergic receptor (␤2AR) and MOR to the firefly luciferase gene regulated by the CRE response element. B, HEK293T cells were transfected with CRE-luc2P reporter, MOR, G␣i1, G␤1, G␥2, and G␤5 with or without R7 RGS. After 5 h of treatment with 50 nM isoproterenol (ISO) together with serial doses of morphine, luciferase activity was measured. The highest dose of morphine treatment of cells transfected without RGS or with RGS7 or RGS9-2 inhibited ISO-induced luminescence by 68. 5 Ϯ 1.3%, 62.9 Ϯ 2.5%, and 61.2 Ϯ 4.7% (S.E., n ϭ 4), respectively. The average luminance at the highest dose of morphine treatment was subtracted as background, and the resulting difference (⌬R) was normalized against the maximal value upon stimulation by ISO only (Rmax). C, IC 50 values were obtained by fitting a four-parameter logistic curve to the inhibition data using GraphPad Prism 5. **, p Ͻ 0.01, one-way analysis of variance followed by Tukey's post hoc test.
FIGURE 10. A model for GPCR and G protein selectivity of striatal R7 RGS complexes. RGS7 and RGS9-2 differentially regulate D2R-and MOR-mediated signaling to G␣i and G␣o in the absence (A) or presence (B) of R7BP. The thickness of the T-shaped arrows indicates the relative strength of GAP activity observed in this study. A thicker line represents stronger activity. RGS7 is a stronger GAP than RGS9-2. RGS9-2, but not RGS7, is capable of regulating G␣i in the absence of R7BP. Both RGS9-2 and RGS7 preferentially regulate G␣o in the absence or presence of R7BP. Although RGS7 does not show a GPCR preference, RGS9-2 complexes selectively regulate D2R over MOR.