RGS6, RGS7, RGS9, and RGS11 stimulate GTPase activity of Gi family G-proteins with differential selectivity and maximal activity

Regulator of G-protein signaling (RGS) proteins are GTPase activating proteins (GAPs) of heterotrimeric G-proteins that alter the amplitude and kinetics of receptor-promoted signaling. In this study we defined the G-protein alpha-subunit selectivity of purified Sf9 cell-derived R7 proteins, a subfamily of RGS proteins (RGS6, -7, -9, and -11) containing a Ggamma-like (GGL) domain that mediates dimeric interaction with Gbeta(5). Gbeta(5)/R7 dimers stimulated steady state GTPase activity of Galpha-subunits of the G(i) family, but not of Galpha(q) or Galpha(11), when added to proteoliposomes containing M2 or M1 muscarinic receptor-coupled G-protein heterotrimers. Concentration effect curves of the Gbeta(5)/R7 proteins revealed differences in potencies and efficacies toward Galpha-subunits of the G(i) family. Although all four Gbeta(5)/R7 proteins exhibited similar potencies toward Galpha(o), Gbeta(5)/RGS9 and Gbeta(5)/RGS11 were more potent GAPs of Galpha(i1), Galpha(i2), and Galpha(i3) than were Gbeta(5)/RGS6 and Gbeta(5)/RGS7. The maximal GAP activity exhibited by Gbeta(5)/RGS11 was 2- to 4-fold higher than that of Gbeta(5)/RGS7 and Gbeta(5)/RGS9, with Gbeta(5)/RGS6 exhibiting an intermediate maximal GAP activity. Moreover, the less efficacious Gbeta(5)/RGS7 and Gbeta(5)/RGS9 inhibited Gbeta(5)/RGS11-stimulated GTPase activity of Galpha(o). Therefore, R7 family RGS proteins are G(i) family-selective GAPs with potentially important differences in activities.


Introduction
R7 family RGS protein GAP specificity Centricon centrifugal filter device (Millipore, Bedford, CT). The concentration of purified Gβ5/R7 dimers was determined by coomassie staining purified product and a standard curve of protein standards resolved by SDS-PAGE. Yield was approximately 1 mg of Gβ5/R7 dimer per 4 liters except for Gβ5/RGS11, whose purification yielded approximately 250 µg per 4 liters. Gα and Gβγ subunits (24) and muscarinic receptors (25) were purified after expression from bacculoviruses in Sf9 insect cells as described. were filtered over GF/F filters, which collect vesicles but not free protein, to quantitate incorporation of Gα subunits into vesicles, and C 12 E 10 -containing samples were filtered over nitrocellulose to quantitate total Gα.

Results
The specificity of RGS proteins for G-protein substrates determines in part their physiological effects on signaling. Previous in vitro studies with Gα-subunits in solution have illustrated specificity of R7-RGS proteins for Gαo, whereas in vivo observations have suggested broader activities. To more specifically address the selectivity of individual Gβ5/R7 heterodimers for Gα-subunits, we purified Gβ5/RGS6, Gβ5/RGS7, Gβ5/RGS9, and Gβ5/RGS11 to near homogeneity and directly measured their selectivity in steady state GTPase assays with proteoliposomes reconstituted with M1 or M2 muscarinic receptors and various heterotrimeric G proteins of the Gq and Gi families, respectively.

Gβ5/R7 Protein Purification
Full length RGS6, 7, 9, and 11 were co-expressed with hexa-histidine-tagged Gβ5 in Sf9 insect cells using the baculovirus expression system. Dimers were purified from the soluble fraction using Ni-NTA agarose and ion exchange chromatography. Twenty-five to seventy-five percent of the total cellular immunoreactive R7 protein was recovered in the soluble fraction, 25-75% of the soluble protein was recovered following Ni-NTA chromatography, and nearly 100% of the Ni-NTA eluate was recovered following the final ion exchange purification. The results of a typical purification (Gβ5/RGS7) are shown in Figure 1A. Purified Gβ5/RGS6, Gβ5/RGS9, and Gβ5/RGS11 dimers are illustrated in Figure 1B.

Vesicle Reconstitution
G-protein α-subunits (Gαo, Gαi1, Gαi2, Gαi3, Gαq, Gα11) were reconstituted in phospholipid vesicles with Gβ1γ2 and either M1 (Gq family G-proteins) or M2 (Gi family G-proteins) muscarinic receptors under conditions similar to those described by Ross and coworkers (26;27). Recovery of Gα subunits and M1 or M2 muscarinic receptors in the various vesicle preparations was quantitated as described in Methods. Essentially 100% of added Gαo, Gαi1, Gαi2, or Gαi3 was incorporated into R7 family RGS protein GAP specificity vesicles, and receptor recovery in the proteoliposomes was approximately 50%. We also prepared and resolved Gαo containing vesicles using the higher exclusion limit Sephacryl S-300 gel filtration resin, which separates vesicles from free Gαo, and observed nearly all of the Gαo immunoreactivity comigrating with vesicles in the void volume (data not shown). The four varieties of M2•Gαi/o vesicles contained similar Gα protein levels (~100 fmol/µl) and receptor: Gα ratios (1:6) (data not shown).
Quantitation of Gαq and Gα11 is difficult due to the low rates of guanine nucleotide turnover by these Gα-subunits. Therefore, calculations of GTPase activity reported below were made assuming that incorporation of Gαq/11 into vesicles was equal to that of Gi family Gα-subunits.

Steady state GTPase assays
The GAP activity of Gβ5/R7 proteins toward Gα subunits of the Gi and Gq families was assessed in steady state GTPase assays, which measure multiple rounds of GTP hydrolysis and, as such, reflect both guanine nucleotide exchange and GTPase activity. RGS4, an effective GAP against Gq and Gi family Gα-subunits (28), was used as a reference RGS protein in all experiments. Only minor increases were observed in the rate of GTP hydrolysis in the presence of either agonist (100 µM carbachol) or GAP (200 nM RGS4) alone in proteoliposomes formed by reconstitution of M2 muscarinic receptor, Gαo, and Gβ1γ2 ( Figure 2). In contrast, the combined presence of carbachol and RGS4 resulted in a markedly synergistic increase in GTPase activity, and the rate of hydrolysis of GTP was linear for at least 15 minutes.
In the presence of a maximally effective concentration of RGS protein, guanine nucleotide exchange is rate limiting, and therefore stimulation of GTPase activity by carbachol was observed with a concentration dependence of agonist that approximated its occupancy curve for binding to the M2 muscarinic receptor (data not shown). Similarly, in the presence of a maximally effective concentration of carbachol, guanine nucleotide exchange was no longer rate limiting, and marked concentration R7 family RGS protein GAP specificity dependent stimulation of GTPase was observed with RGS4 (data not shown) and Gβ5/R7 RGS proteins (see below).

Gβ5/R7 proteins stimulate steady state GTPase activity of Gi family Gα-subunits
To compare the capacity of Gβ5/R7 proteins to accelerate GTPase rates of Gi family Gα subunits, steady state GTPase activities were determined in the presence and absence of 100 µM carbachol and in the presence and absence of 1 µM RGS protein (either RGS4 or each Gβ5/R7 dimer). RGS4 markedly increased GTPase activity for Gαo, Gαi1, Gαi2, and Gαi3 in the presence of 100 µM carbachol ( Figure   3). Each of the Gβ5/R7 dimers also stimulated to varying degrees GTP hydrolysis by Gαo, Gαi1, Gαi2, and Gαi3 in the presence of agonist ( Figure 3 and Table 1). The rate observed with Gβ5/RGS11 was as high or higher than the rate with RGS4 with all four Gi family Gα-subunits, while the GTPase rates in the presence of 1 µM Gβ5/RGS6, Gβ5/RGS7, and Gβ5/RGS9 were significantly lower. Vesicles containing Gαo achieved the highest maximal GTPase rates irrespective of the RGS protein. However, the basal rate of GTP hydrolysis by Gαo in the absence of RGS protein was also higher than that observed in Gαi containing vesicles. Thus, the fold increase in activity (GTPase rate in the presence of RGS and agonist divided by the GTPase rate with agonist alone) of Gαi3 was as high or higher than that of Gαo in response to Gβ5/RGS6,7,9, and 11 stimulation. Further, the effects of R7 proteins on GTPase activity of Gαo subunits reconstituted with purified P2Y12 receptors was also determined (in the presence of the agonist 2-methylthio ADP). Similar to the results observed with M2 receptor-coupled G-proteins, each of the Gβ5/RGS11 dimers stimulated steady state GTPase activity of Gαo, and Gβ5/RGS11 stimulated much higher GTPase rates than the other R7 proteins (data not shown).

Gβ5/R7 proteins do not stimulate steady state GTPase activity of Gq family Gα-subunits
Regulation of GTPase activities of Gαq and Gα11 was examined in vesicles reconstituted with M1 muscarinic receptor and heterotrimeric G-protein ( Figure 4). RGS4 increased steady state GTPase activity of Gαq and Gα11 in the presence of agonist by nearly five fold, to final GTPase rates of approximately 200 fmol GTP/min/pmol Gα. Consistent with previous observations of guanine nucleotide exchange/GTPase kinetics of Gq (29), these rates are lower than those observed for Gi family α subunits.
In contrast to the activity of RGS4 , none of the Gβ5/R7 dimers significantly increased steady state GTPase activity of Gαq or Gα11 in the presence of carbachol ( Figure 4A). Likewise, Gβ5/R7 dimers did not stimulate GTPase activity of Gαq or Gα11 subunits reconstituted with purified P2Y1 receptors (in the presence of the agonist 2-methylthio ADP) (data not shown). Further, 1 µM Gβ5/R7 dimers had no effect on the GTPase activity of M1 receptor-coupled Gαq and Gα11stimulated by RGS4 and carbachol ( Figure   4B). The partial inhibition observed with Gβ5/RGS11 was non-specific, as demonstrated by equivalent inhibitory activity observed with boiled Gβ5/RGS11. Therefore, under the conditions of these assays, R7 proteins neither stimulate GTPase activity of Gαq or Gα11, nor affect GTPase activity stimulated by agonist and RGS4.

Gβ5/R7 proteins exhibit differences in maximal activity and potency toward Gi family Gα-subunits
To more fully elucidate any selectivity of Gβ5/R7 proteins as GAPs for Gi family subunits, full concentration effect curves of each Gβ5/R7 protein were generated in the presence of a maximally effective concentration of carbachol. Consistent with the data in Figure 3, maximally effective concentrations of Gβ5/RGS11 produced larger effects than Gβ5/RGS7, Gβ5/RGS6 and Gβ5/RGS9 on the GTPase activity of each Gα-subunit ( Figure 5 and data not shown), and the highest maximal rate observed with each RGS protein was observed with Gαo as substrate (not shown). Each of the Gβ5/RGS dimers produced a near maximal effect at a concentration of 1 µM, and therefore each activation curve was normalized to 100% of maximal activity for comparison of EC50 values ( Figure 6). All four Gβ5/R7 dimers exhibited similar potency for Gαo (EC 50 = 16-47 nM), while Gβ5/RGS9 and Gβ5/RGS11 were more potent (EC 50 = 25-80 nM) than Gβ5/RGS6 and Gβ5/RGS7 (EC 50 = 150-350 nM) for Gαi1, Gαi2, and Gαi3 (Table 2).

Gβ5/RGS7 and Gβ5/RGS9 inhibit Gβ5/RGS11-stimulated Gαo GTPase activity
Marked differences in the maximal GTPase rate of Gα-subunits were observed across the Gβ5/R7 protein family. For example, Gβ5/RGS11-stimulated GTPase activity of Gαo was twice that achieved in the presence of Gβ5/RGS9 or Gβ5/RGS7. These results suggest that Gβ5/RGS7 and Gβ5/RGS9 interaction with G-proteins results in a less active Gα conformation with respect to GTPase activity than that promoted by Gβ5/RGS11 interaction. To test this hypothesis, steady state GTPase activity of M2•Gαo•β1γ2 vesicles was measured in the presence of 100 nM Gβ5/RGS11 or 1 µM Gβ5/RGS7 alone, or with 100 nM Gβ5/RGS11 plus 1 µM Gβ5/RGS7. As illustrated in Figure 7A, GTPase activity in the presence of Gβ5/RGS11 was nearly twice that observed with a ten fold higher concentration of

Discussion
The results of this study demonstrate Gαi/o specificity of Gβ5/R7 proteins, and found no evidence of regulation of Gαq GTPase activity by these proteins. Further, we demonstrated that differences exist in the potencies and relative efficacies of the Gβ5/R7 proteins for their Gαi/o substrates. Finally, we illustrated that Gβ5/RGS7 and Gβ5/RGS9, which are less effective promoters of maximal GTPase activity than is Gβ5/RGS11, inhibit Gβ5/RGS11-stimulated GTPase activity of Gαo.
The drug selectivity of G protein coupled receptors has been widely exploited therapeutically to manipulate specific cellular processes (30)(31)(32)(33)(34). Similarly, selectivities of G proteins for effector activation and potential selectivities of RGS proteins for deactivation of G proteins may provide equally rich targets for pharmacological modulation of G protein-regulated signaling (35;36). Whereas the role of receptor activity and selectivity in regulating various classes of G-proteins has been studied extensively in the past decades, the roles of proteins exhibiting Gα GAP activity and selectivities within this class of proteins remain undefined.
Members of the Gβ5/R7 family previously were reported to be specific for Gαo in single turnover assays of soluble Gα subunits (11;20). These assays may not accurately represent physiological interactions between G-proteins and RGS proteins for several reasons, including the lack of a lipid bilayer with which G-proteins and RGS proteins may associate, the lack of a GPCR which may form a stable complex with RGS proteins (37-39) and facilitate interaction with Gα subunits, and the necessity of using GTPase-deficient mutants of Gq family G-proteins given their low rates of exchange. These limitations may explain discrepancies between selectivities for G-protein α-subunits observed in single turnover versus cell-based assays. Indeed, RGS2 behaves as a Gαq specific GAP in single turnover assays, but exhibits GAP activity towards Gαi as well as Gαq in steady state GTPase assays of proteoliposomes reconstituted with GPCR and heterotrimeric G-proteins (28).

R7 family RGS protein GAP specificity
In this study, we examined the GAP activity of Gβ5/R7 proteins using steady state GTPase assays of receptor-coupled G-proteins reconstituted in phospholipid bilayers. Because these assays measure multiple rounds of hydrolysis in the presence of receptor-stimulated guanine nucleotide exchange, wild type Gαq or Gα11 may be used, and the contributions made by agonist, receptor, Gβγ subunits, and the phospholipid bilayer to GTPase activity are likely more representative of a cellular environment. Our results differ from those from single turnover assays (11;20) with respect to the selectivity within the Gαi/o family since all four Gα subunits of this family are substrates for Gβ5/R7 proteins.
Our results also differ from published reports that indirectly suggest that Gβ5/R7 proteins stimulate GTPase activity of Gαq or Gα11 (14;21;40) in that we did not observe stimulation of Gαq or Gα11 GTPase activity in response to R7 proteins. A trivial explanation for our observation is that R7 proteins inhibit agonist promoted exchange and thereby mask GAP activity in steady state GTPase assays.
However, the lack of an effect of R7 proteins on the steady state GTPase activity achieved in the presence of carbachol and RGS4 demonstrates that Gβ5/R7 heterodimers do not significantly affect agonistpromoted exchange of guanine nucleotides under the conditions of these assays ( Figure 4B). We also observed minimal to no effects of Gβ5/R7 proteins on agonist-stimulated guanine nucleotide exchange measured directly in GTPγS binding assays (not shown). Therefore, the inability of Gβ5/R7 proteins to stimulate steady state GTPase activity of Gαq and Gα11 indicates that they do not function as Gq family GAPs under the conditions of our assay. In the absence of GAP activity, the reported effects of RGS7 on Gq family G-protein signaling could reflect direct inhibition of phospholipase enzymes, as observed by Posner et al. (20). Although we have observed some inhibition of receptor stimulated inositol phosphate accumulation in cells cotransfected with R7 RGS proteins and Gβ5, this inhibition is less pronounced and requires expression to much higher levels than does the marked inhibition of phosphoplipase C response observed in cells overexpressing RGS2 or RGS4 (data not shown). Thus, the reported effects of R7 R7 family RGS protein GAP specificity proteins on cellular Gq pathways may reflect either loss of GAP selectivity due to protein overexpression or a more complex interaction of Gβ5/R7 dimers with the G-protein signaling cycle.
Gβ5/R7 proteins exhibited differences in the potency and efficacy of their GAP activity against the Gαi/o family. Gβ5/RGS6 and Gβ5/RGS7 each exhibited ten fold lower potency for Gαi α-subunits than for Gαo. In contrast, Gβ5/RGS9 and Gβ5/RGS11 exhibited similar potency for all four Gi family Gα subunits. This pattern mirrors the grouping of R7 proteins by sequence similarity; that is, RGS6 and RGS7 have higher sequence identity to each other than to RGS9 and RGS11, and vice versa (13). The R7 proteins group differently with respect to their apparent efficacies for stimulation of GTPase activity.
Gβ5/RGS11 exhibited the highest maximal effect, while the maximal effects of Gβ5/RGS7 and Gβ5/RGS9 were much less and Gβ5/RGS6 exhibited an intermediate maximal effect. These differences in activity inversely correlate with the expression of R7 transcripts in rat brain, where RGS11 is expressed at much lower levels than RGS7 and 9, and again RGS6 is intermediate (42). We speculate that expression of the robustly active RGS11 may be regulated differently than the less active proteins.
Gαi1, Gαi2, and Gαi3 share high sequence homology, and these signaling proteins are essentially interchangeable in many signaling processes. However, although these proteins often are expressed in the same cell, they may not be entirely functionally redundant. For example, selectivity of coupling of certain G protein-coupled receptors among these three Gαi proteins has been illustrated (43)(44)(45)(46), and several reports suggest selective coupling of receptors to ion channels through specific Gαi-subunits In summary, we have demonstrated that R7 family RGS proteins selectively stimulate GTPase activity of Gi family Gα subunits. We have shown differences in potency and efficacy of Gβ5/R7 dimers as GAPs among the Gi family Gα-subunits. Further, lower efficacy GAPs were shown to inhibit GTPase activity achieved in the presence of a more efficacious GAP, indicating that RGS proteins apparently interacting with the same activating surface of a Gα-subunit promote different maximal rates of catalysis by the Gα GTPase.