Membrane Anchor R9AP Potentiates GTPase-accelerating Protein Activity of RGS11·Gβ5 Complex and Accelerates Inactivation of the mGluR6-Go Signaling*

The R7 subfamily of RGS proteins critically regulates neuronal G protein-signaling pathways that are essential for vision, nociception, motor coordination, and reward processing. A member of the R7 RGS family, RGS11, is a GTPase-accelerating protein specifically expressed in retinal ON-bipolar cells where it forms complexes with the atypical G protein β subunit, Gβ5, and transmembrane protein R9AP. Association with R9AP has been shown to be critical for the proteolytic stability of the complex in the retina. In this study we report that R9AP can in addition stimulate the GTPase-accelerating protein activity of the RGS11·Gβ5 complex at Gαo. Single turnover GTPase assays reveal that R9AP co-localizes RGS11·Gβ5 and Gαo on the membrane and allosterically potentiates the GTPase-accelerating function of RGS11·Gβ5. Reconstitution of mGluR6-Gαo signaling in Xenopus oocytes indicates that RGS11·Gβ5-mediated GTPase acceleration in this system requires co-expression of R9AP. The results provide new insight into the regulation of mGluR6-Gαo signaling by the RGS11·Gβ5·R9AP complex and establish R9AP as a general GTPase-accelerating protein activity regulator of R7 RGS complexes.

Regulators of G protein signaling (RGS) 3 are ubiquitous signaling molecules that critically shape cellular responses mediated by G protein-coupled receptor (GPCR) pathways (1). Most RGS proteins act by stimulating the rate of the GTP hydrolysis of the ␣ subunits of G proteins (G␣), thus speeding up their inactivation and limiting the duration of GPCR signaling (2). R7 subfamily of RGS proteins plays a central role in regulating fundamental neuronal functions including vision, nociception, reward processing, and motor control (3)(4)(5)(6). The subfamily consists of four homologous proteins, RGS6, RGS7, RGS9, and RGS11, that are enriched in the nervous system and share a common multidomain architecture that, in addition to the catalytic RGS domain, includes an N-terminal DEP/DHEX (Dishevelled, EGL-10, Pleckstrin/DEP helical extension) module and a GGL (G protein ␥ subunit-like) domain (7).
In vivo, R7 RGS proteins are found in complexes with two types of proteins that are becoming increasingly accepted as their subunits. The GGL forms a complex with the atypical G protein ␤ subunit, G␤ 5 (8). The DEP/DHEX module binds to either R7BP (R7 family binding protein) or R9AP (RGS9 anchor protein) (9, 10), a two-member family of small SNARE (soluble N-ethylmaleimide factor attachment protein receptor)-like proteins. Association with both G␤ 5 (11,12) and R7BP/R9AP (13)(14)(15) protects complexes from proteolytic degradation and, thus, plays an important role in regulating the expression levels of the R7 RGS proteins. In addition, membrane proteins R9AP and R7BP can localize R7 RGS proteins to cell membranes in transfected cells and native neurons (16 -18). Finally, R9AP has been shown to enhance the GTPase-accelerating protein (GAP) activity of RGS9-1 toward transducin, suggesting that membrane anchors also play a role in regulation of R7 RGS activity (10, 19 -21). However, what remains unclear is (i) if the potentiation of the GAP activity of R7 RGS members is a universal mode of action for the R9AP or R7BP membrane anchors, (ii) whether R9AP can stimulate GAP activity of R7 RGS proteins at G proteins other than transducin, and (iii) whether the stimulation of GAP function results simply from the effect of concentrating R7 protein at the cell membrane or from a more complicated mechanism, requiring additional allosteric contributions.
Previously, R9AP has been shown to be expressed in retinal photoreceptor cells where it regulates the photoreceptorspecific R7 RGS protein complex, RGS9-1⅐G␤ 5 (4,14,18). Recently, we have discovered R9AP in a novel heterotrimeric complex with another R7 RGS protein, RGS11 (23). The R9AP⅐RGS11⅐G␤ 5 complex is exclusively expressed in retinal ON-bipolar neurons which relay the light-elicited signals from the photoreceptors to ganglion cells in the circuit that connects the photoreceptors to the brain (23)(24)(25)(26). The complex is targeted to the dendritic tips of the ON-bipolar neurons via the association with the mGluR6 (23), a G␣ o -coupled GPCR, that responds to light-induced changes in glutamate release from the photoreceptor terminals (for review, see Ref. 27). Studies with knock-out mice suggest that RGS11 is involved in determining the kinetics of the overall synaptic transmission between photoreceptor and bipolar neurons (24) and that all subunits of the R9AP⅐RGS11⅐G␤ 5 complex are indispensable for its appropriate localization and expression in the bipolar neurons (23).
In this study we have investigated the impact of R9AP on the functional activity of the RGS11⅐G␤ 5 complex. We show that recruitment of RGS11⅐G␤ 5 to the plasma membrane by R9AP results in a substantial increase in the GAP activity of the complex toward lipid-modified but not soluble G␣ o . We reconstituted the mGluR6-G o -RGS11 signaling cascade in Xenopus oocytes and utilized the acceleration of the deactivation kinetics of co-expressed G protein inwardly rectifying potassium (GIRK) channels as a measure of RGS11 GAP function. Activity of RGS11 in this assay system was discerned only when R9AP was co-expressed. These results establish that membrane anchoring by R9AP is a general mechanism for regulating activity of R7 RGS complexes and, in the case of RGS11, is an essential prerequisite for controlling G␣ o signaling through mGluR6.
Purification of G o from Sf9 cells was conducted as described previously with modifications (33). Briefly, Sf9 cells from 1 liter of culture were harvested 48 h after infection with recombinant baculoviruses encoding rat G␣ oA , G␤ 1 , and His-tagged G␥ 2 , resuspended in lysis buffer (20 mM HEPES, pH 8.0, 500 mM NaCl, 2 mM MgCl 2 , 10 M GDP, 20 mM imidazole, 10 mM ␤-mercaptoethanol), and then lysed by sonication. Lysates were centrifuged at 30,000 ϫ g for 30 min, and the resultant pellets were washed with wash buffer (lysis buffer containing 1 mM MgCl 2 ). The pellets were resuspended in wash buffer. C 12 E 10 detergent was added to a final concentration of 1% (w/v), and the mixture was stirred for 1 h before centrifugation at 30,000 ϫ g for 30 min. The supernatants were loaded onto a 1-ml nickel-Sepharose High Performance (GE Healthcare) equilibrated with buffer A (wash buffer containing 0.2% (w/v) C 12 E 10 ). The beads were washed with 25 volumes of buffer A and further washed with 10 volumes of buffer B (wash buffer containing 0.2% (w/v) CHAPS and 1 mM ␤-mercaptoethanol). The G o was eluted with elution buffer (wash buffer containing 500 mM imidazole, 0.7% (w/v) CHAPS, and 1 mM ␤-mercaptoethanol). Fractions containing G o were pooled and concentrated to ϳ10 mg/ml. Finally, buffer was exchanged to storage buffer (20 mM HEPES, pH 8.0, 300 mM NaCl, 1 mM MgCl 2 , 1 M GDP, 0.7% CHAPS, 10% glycerol, 1 mM ␤-mercaptoethanol) by a Zeba Desalt Spin Column (Thermo Scientific).
Membrane Preparations-Rod outer segments treated by urea (uROS) and V8-uROS membranes were prepared as described previously with minor modifications (20). For the preparation of insect cell membranes, Sf9 cells and Sf9 cells infected with amplified recombinant baculoviruses encoding R9AP were sonicated in membrane preparation buffer (10 mM Tris-HCl, pH 7.8, 100 mM NaCl, and 8 mM MgCl 2 ), and crude cellular debris was sedimented at 1200 ϫ g for 5 min. The membranes were then sedimented at 30,000 ϫ g for 30 min and washed 3 times with wash buffer (10 mM Tris-HCl, pH 7.8, 500 mM NaCl, and 8 mM MgCl 2 ). The membranes were stored at Ϫ80°C in membrane preparation buffer containing 10% glycerol. All steps were conducted at 4°C using ice-cold buffers supplemented with protease inhibitors. Total protein concentration was determined by Bradford assay.
Single Turnover GTPase Assay-Single turnover GTPase assays were conducted as described previously (34). Assays were performed at room temperature in GTPase buffer (10 mM Tris-HCl, pH 7.8, 250 mM NaCl, 8 mM MgCl 2 , and 1 mM dithiothreitol). Illuminated uROS or V8-uROS were used as the source of photoexcited rhodopsin required for G protein activation. Phosphate released from hydrolyzed GTP was determined by activated charcoal assay (35). Because high efficiency interactions between RGS9-1⅐G␤ 5 L and G␣ o require the presence of the ␥ subunit of phosphodiesterase, type 6 (PDE␥) (32,36), we conducted the assay in the presence of 1.0 M PDE␥. The relative amount of RGS9-1 and RGS11 used in the assays was deter-mined by Western blotting. Control experiments were performed to ensure that single turnover conditions are maintained and that GDP/GTP exchange does not limit the GTP hydrolysis (see Refs. 34 and 35 for a description of the criteria and supplemental Fig. 1

for results).
Membrane Binding Assay-A membrane binding assay was performed as described previously (20). Briefly, all experiments were conducted in GTPase buffer at 4°C. Recombinant RGS9-1⅐G␤ 5 L or RGS11⅐G␤ 5 S was mixed with membranes in the volume of 100 l, and the mixture was loaded on a 50-l cushion containing 10% sucrose in GTPase buffer. The mixture was sedimented at 120,000 ϫ g for 15 min in a TL-100 ultracentrifuge (Beckman Instruments). Supernatants were collected, whereas the sucrose cushions were discarded. The pellet was washed once by GTPase buffer containing 1 M NaCl to reduce nonspecific protein binding and once by GTPase buffer and then resuspended in 100 l of GTPase buffer. In some experiments, when indicated, the wash step was skipped. Input, resuspended Ppt and Sup fractions were subjected to SDS-PAGE on 12.5% polyacrylamide gels, and the contents of RGS11⅐G␤ 5 S, RGS9-1⅐G␤ 5 L, R9AP, and G␣ o in each sample were determined by Western blot. For the membrane binding assays of G␣ o , 10 mM NaF, 20 M AlCl 3 , and 10 M GDP were added into all buffers to activate G␣ o and release from G␤␥ so that we can test membrane binding activity of free G␣ subunits. The final concentrations in the binding reactions were 50 nM RGS9-1⅐G␤ 5 L complexes, 27 nM RGS11⅐G␤ 5 S, 1.0 M GST, 1.0 M GST-R9AP⌬TM, 0.5 mg/ml Sf9 membranes, 0.5 mg/ml R9AP membranes, and uROS and V8-uROS containing 20 M rhodopsin.
Oocyte Expression Assays-Published methods were used for these studies (37). cRNA was synthesized in vitro from plasmids containing the cDNA and appropriate promoters for cRNA transcription. Plasmids were linearized before cRNA synthesis, and mMESSAGE MACHINE kits (Ambion) were used to generate capped cRNA. cRNA was injected into oocytes at a volume of 50 nl/oocyte using a Drummond microinjector. Oocytes were maintained in a saline buffer (96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , and 5 mM HEPES, pH 7.5) solution supplemented with 5% (v/v) heat-inactivated horse serum, sodium pyruvate (2.5 mM), gentamycin (50 g/ml). A valve system controlled by the data acquisition software pCLAMP 6 (Axon Instruments) was used to control solution changes and to minimize washing and washout times. Two-electrode voltage clamp recordings of the oocytes were performed 36 -72 h after cRNA injection. Membrane potential was clamped at Ϫ80 mV using a AxoClamp 900A amplifier (Molecular Devices) and pCLAMP 6 software. Electrodes were filled with 3 M KCl and had resistances of 0.5-1.5 megaohms. To reveal inward currents through the inwardly rectifying GIRK channels, recordings were performed in oocyte saline buffer with elevated (16 mM) KCl concentrations (other components were 82 mM NaCl, 1 mM MgCl 2 , 1 mM CaCl 2 , and 5 mM HEPES, pH 7.5). After electrophysiological recording, oocytes were sonicated in icecold phosphate-buffered saline containing protease inhibitors. 30 l of phosphate-buffered saline was added per oocyte. The homogenate was centrifuged at 750 ϫ g at 4°C for 10 min, and the resultant supernatant was used as a whole protein extract of oocytes. To compare the expression levels of RGS11 proteins in oocytes, Western blotting of the whole protein extracts was performed using rabbit anti-RGS11 CT antibody.
Curve-fitting and Statistical Analysis-Kinetic analysis and curve-fitting were performed using pCLAMP 6 software. Time constants () were derived from the exponentials fitted to the activation and deactivation phases of the GIRK currents. Cursors were positioned at points on the activation and deactivation curves that corresponded to 20 and 80% of the maximum equilibrium responses, and the exponentials were fitted to the portions of the current trace between these two points. We used Student's t test for comparison of the independent means. Two-tailed p value Ͻ0.01 is defined as significantly different.

RESULTS
R9AP Enhances GAP Activity of RGS11 by a Membrane-delimited Mechanism-Purified uROS have proven to be a valuable model for delineating activity regulation mechanisms of RGS9-1, the closest homologue of RGS11 (34,35). Because urea treatment removes the G protein transducin (G␣ t ) and associated GAPs but leaves endogenous R9AP and GPCR rhodopsin intact (20), we used this preparation as an initial reagent to study the effects of R9AP on the RGS11⅐G␤ 5 complex. Previous studies have shown that anchoring purified RGS9-1⅐G␤ 5 by native R9AP contained within the uROS preparation markedly potentiates its GAP activity toward exogenously added G␣ t (19 -21). Because G␣ t is not the physiological substrate of the RGS11⅐G␤ 5 complex (23), we first demonstrated that the uROS system could be exploited to study the effects of R9AP on the GTP hydrolysis catalyzed by G␣ o , a G protein specifically coexpressed with the RGS11 in the ON-bipolar neurons (23,25,26). The addition of increasing amounts of RGS9-1⅐G␤ 5 to the uROS membranes reconstituted with lipid-modified G␣ o , expressed and purified from Sf9 cells, produced a typical biphasic concentration dependence curve with an initial steep slope of 0.09 Ϯ 0.01 s Ϫ1 M Ϫ1 catalytic activity rate followed by a shallower slope defined by a specific activity of 0.020 Ϯ 0.003 s Ϫ1 M Ϫ1 (Fig. 1A). Previous studies have established that the inflection point in the curve corresponds to the concentration of active R9AP on the membranes (20,21), and therefore, the initial steeper slope is interpreted as the specific activity of the R9AP-bound RGS complex, whereas the second shallower slope reflects the specific activity of excess RGS that remains free after all the R9AP in the reaction mixture has been bound up. Comparison of the two slopes indicates that binding of RGS9-1⅐G␤ 5 to endogenous R9AP on the membranes can potentiate its GAP activity toward G␣ o by ϳ5-fold. Consistent with this, we found that RGS9-1⅐G␤ 5 was efficiently recruited to the uROS membranes and that digestion of R9AP by V8 protease completely abolished association of RGS9-1⅐G␤ 5 with the V8-treated uROS membranes (V8-uROS) (Fig. 1B). Although the magnitude of the stimulatory effect by R9AP is lower than seen in the G␣ t -based system (21), the results clearly indicate that this broken cell preparation can be used to assay the GAP activity of R7 RGS proteins toward G␣ o , making this system amenable for assaying the action of R9AP on the GAP function of RGS11 toward G␣ o .
We found that recombinant RGS11⅐G␤ 5 was also able to strongly bind to native uROS membranes but not to the R9AP Regulates GAP Activity of RGS11 FEBRUARY 12, 2010 • VOLUME 285 • NUMBER 7
Interestingly, when RGS11⅐G␤ 5 was incubated with V8-uROS and the mixture subsequently fractionated by centrifugation, the RGS11 protein was not detected in either the Sup or Ppt fractions ( Fig. 2A). The absence of RGS11 in both V8-treated fractions can be explained if the RGS11⅐G␤ 5 complex can associate directly (i.e. independently of R9AP) with the membranes and sediment with the membrane pellet. The association is weak enough, however, to be disrupted by the stringent condi-tions under which the pellet is washed (see "Experimental Procedures"), resulting in the removal of RGS11⅐G␤ 5 complex from the membrane pellet. Indeed, skipping the washing steps resulted in a robust detection of RGS11 in a pellet fraction (supplemental Fig. 2). RGS9-1⅐G␤ 5 is not targeted to the V8-uROS membranes in this manner (i.e. without R9AP, see Fig. 1B). Consequently, the activity of non-R9AP-bound, RGS11⅐G␤ 5 (0.16 Ϯ 0.01 s Ϫ1 M Ϫ1 , shallow slopes in Fig.  2B) was substantially higher than that of free RGS9-1⅐G␤ 5 (0.02 Ϯ 0.003 s Ϫ1 M Ϫ1 , shallow slope in Fig. 1A), suggesting that the activity of RGS11⅐G␤ 5 can be enhanced by R9AP-independent membrane association that brings it in the vicinity of lipid-modified G␣ o .
To analyze the contribution of such membrane localization to the activity of the RGS11⅐G␤ 5 complex, the above assay was repeated with a soluble G␣ o that was unable to tightly associate with the uROS membranes (Fig. 3A). This preparation of G␣ o was obtained by expression in E. coli known to produce protein devoid of posttranslational lipid modifications (40). The G␣ o purified from E. coli showed intrinsic GTPase activity (k app ϭ 0.021 Ϯ 0.005 s Ϫ1 ) similar to that of the lipid-modified G␣ o from Sf9 cells (k app ϭ 0.027 Ϯ 0.007 s Ϫ1 ) that was effectively accelerated by the addition of soluble RGS proteins (e.g. RGS7, supplemental Fig. 1). However, the RGS11⅐G␤ 5 GAP activity targeted to the soluble nonlipid-modified G␣ o from E. coli was not biphasic. The entire activity versus concentration plot (Fig. 3B) could be fitted with a single shallow slope characterized by a catalytic rate of only 0.043 Ϯ 0.003 s Ϫ1 M Ϫ1 . This rate is closer to the GAP rate displayed by the non-R9AP-bound form of RGS9-1⅐G␤ 5 (0.02 Ϯ 0.003 s Ϫ1 M Ϫ1 ), which does not exhibit intrinsic membrane binding (Fig. 1B).
The role of membrane anchoring by R9AP in activity regulation of RGS11⅐G␤ 5 was further delineated using an R9AP deletion mutant rendered soluble by deleting a transmembrane region (R9AP⌬TM). Incubation with this construct prevented association of RGS11⅐G␤ 5 with the uROS membranes, and the RGS11⅐G␤ 5 complex remained in the supernatant after centrifugation of the reaction mixture (Fig. 4A). Concurrently, R9AP⌬TM also dramatically reduced the GAP activity of RGS11 (Fig. 4B). At the same time, R9AP⌬TM did not have any Two independent experiments were conducted, and all of the obtained data are plotted in one graph. B, recombinant histidine-tagged RGS9-1 expressed in and purified from Sf9 cells co-expressing G␤ 5 protein was incubated with uROS membranes and analyzed for the ability to co-sediment with the membranes. The presence of RGS9-1 and R9AP protein in the total incubation mixture (Input) and the supernatants and membrane pellets produced after centrifugation was determined by Western blotting. The top and bottom panels in B depict Western blots (WB) with an anti-histidine tag antibody (␣His) and with an anti-R9AP antibody (␣R9AP), respectively. In the bottom panel in B, the prominent higher molecular weight bands are full-length R9AP, and an arrowhead indicates the degradation products of R9AP (observed only with V8-uROS).

FIGURE 2. Endogenous R9AP anchors RGS11⅐G␤ 5 on the membranes and potentiates GAP activity toward lipid-modified G␣ o .
A, membrane binding assay was performed with RGS11⅐G␤ 5 as described in the legend for Fig. 1. WB, Western blot. B, single-turnover GTPase assays were performed at a fixed concentration of uROS (E) or V8-uROS (F) with increasing concentrations of RGS11⅐G␤ 5 as described in the legend for Fig. 1. Lipid-modified G␣ o purified from Sf9 cells was used. The rate constants of G␣ GTPase measured in the absence of RGS proteins (0.022 Ϯ 0.001 s Ϫ1 for uROS and 0.019 Ϯ 0.003 s Ϫ1 for V8 uROS) were subtracted from the values measured with RGS proteins, and the resulting k GAP values were plotted on the graph. Three independent experiments with each membrane preparation were conducted, and all of the obtained data are plotted in one graph. effect on the G␣ o GTPase activity in the absence of RGS11⅐G␤ 5 (supplemental Fig. 3). These observations indicate that membrane anchoring of RGS11⅐G␤ 5 enhances its GAP activity by localizing it in close proximity to G␣ o .
To further prove that R9AP mediates the potentiation of RGS11⅐G␤ 5 GAP function, we have performed a reconstitution experiment. For this, membranes prepared from Sf9 insect cells expressing recombinant R9AP were added in trans to the V8-uROS membranes. Data shown in Fig. 5A demonstrate that the addition of the R9AP-expressing Sf9 membranes partitions recombinant RGS11⅐G␤ 5 away from the soluble fraction to the particulate fraction and increases its activity by ϳ2-fold, resulting in nearly complete restoration of RGS11 activity lost upon the V8 treatment of the uROS membranes (ϳ2.5 times; Fig. 5B). At the same time, R9AP-containing Sf9 membranes did not change the basal rates of G␣ o GTPase activity (supplemental Fig. 4). Taken together, these findings indicate that specific anchoring by R9AP potentiates the activity of RGS11⅐G␤ 5 via a membrane-delimited mechanism.
R9AP Augments Regulatory Activity of RGS11 on mGluR6-G␣ o Signaling in Xenopus Oocytes-Recent studies suggest that RGS11⅐G␤ 5 ⅐R9AP complex acts as a GAP for G␣ o in retinal ON-bipolar neurons where it is selectively recruited to mGluR6, a G proteincoupled receptor that is thought to activate the G␣ o when the neurons respond to glutamate released from the photoreceptors (23)(24)(25). Therefore, we have next sought to determine the effect of the R9AP on RGS11 GAP function at a cellular level in the context of this physiologically relevant receptor, mGluR6, and the cognate G protein, G␣ o . We have reconstituted coupling between mGluR6 and G␣ o using a well established Xenopus oocytes expression system and utilized co-expressed GIRK channels to study the kinetics of the physiologically relevant coupling between mGluR6 and G␣ o . The kinetics of the channel deactivation depend crucially on the rate at which the free G␤␥ is re-sequestered by inactive G␣, which in turn depends on GTPase activity of the G␣. Thus, the acceleration of the GIRK channel deactivation rate by co-expressed RGS proteins provides a convenient measure of the GAP function of RGS protein in live cells that has with the high sensitivity and temporal resolution (for review, see Ref. 37). The contribution of endogenous Xenopus G proteins to our kinetic measures was eliminated by co-expressing the catalytic subunit of pertussis toxin (PTX) and by utilizing a pertussis toxin insensitive G␣ o construct, G␣ o C351A. Consistent with an earlier report (41), heterologous expression of mGluR6, PTX catalytic subunit, PTX-insensitive G␣ o C351A, and GIRK1/GIRK4 subunits allowed for the detection of inward currents through the GIRK channels (downward and upward deflections in current traces depicted in Fig. 6A) in response to mGluR6 stimulation by glutamate application and extinction upon glutamate washout (shown in black in Fig. 6, A and B). No glutamateelicited currents were observed when the PTX catalytic subunit was expressed in the absence of the PTX-insensitive G␣ o C351A G protein subunit (data not shown). The injection of oocytes with cRNA for RGS11 and G␤ 5 (2 ng each) did not have any significant effect on the kinetics of the response (blue versus black traces in Fig. 6, A and B) despite robust expression of RGS11⅐G␤ 5 in cells (Fig. 6C). However, coexpression of RGS11⅐G␤ 5 with R9AP dramatically accelerated both the activation and deactivation kinetics of glutamate-evoked GIRK currents (shown in red in Fig. 6, A and B), which are the hallmarks of GIRK channel modulation by the GAP activity of RGS proteins (38,39,42,43). The apparent acceleration of the activation kinetics are in fact an unsurprising and necessary consequence of the accelerated deactivation kinetics; for most reversible reactions the time   Fig. 1C was chosen, and data are expressed as the means Ϯ S.E. from three independent experiments. **, p Ͻ 0.01 (unpaired t test). required to reach equilibrium (i.e. the peak GIRK current amplitude) after the forward reaction is set in motion is shortened when either the forward or the backward rate of the reaction (deactivation rate) is increased (see Ref. 39 for a comprehensive account). Note that the averaged activation and deactivation traces depicted in Fig. 6A have been normalized against the peak level of GIRK current. The normalization was performed so that the altered kinetics depicted by the traces can be compared appropriately. Expression of R9AP alone (i.e. in the absence of RGS11⅐G␤ 5 ) had no effect on GIRK kinetics (data not shown), indicating that the stimulatory effect of R9AP is mediated through RGS11⅐G␤ 5 . Comparison of the protein expression by Western blotting showed a modest increase in RGS11 levels upon co-expression with R9AP (Fig. 6C), consistent with the role of R9AP in protecting RGS11⅐G␤ 5 from rapid proteolytic degradation (23). However, this rather modest increase in RGS11⅐G␤ 5 levels (not exceeding 2-fold) cannot explain an all-or-nothing effect of R9AP addition on the ability of RGS11⅐G␤ 5 to accelerate activation and deactivation kinetics. In summary, these observations indicate that R9AP is necessary for the GAP function of RGS11⅐G␤ 5 protein complex on mGluR6-G␣ o signaling in a cell-based system.

DISCUSSION
The main conclusion from our study is that R9AP positively modulates the GAP activity of RGS11⅐G␤ 5 complex in vitro and is essential for the ability of the RGS11 complex to regulate mGluR6 signaling in cells. In addition, we show that R9APmediated acceleration of the GAP activity is not confined to the RGS9transucin system but is likely to be a more universal mechanism of R9AP action. The stimulatory effect of R9AP requires that it is localized at the membrane with the ␣ subunits of the G proteins. Disruption of this co-localization by either dominant negative soluble R9AP or by the use of non-lipid-modified G␣ o abolished potentiation by R9AP, indicating the critical contribution of membrane association to the action of R9AP. These observations also suggest that changes in lipid modification status of G␣ o (for review, see Ref. 22) would have profound consequences on the acceleration of inactivation by some R7 RGS family members. More specifically, the duration of G␣ o signaling in the mGluR6 cascade of ONbipolar cells is expected to be substantially influenced by dynamics of palmitoylation and/or myristoylation of G␣ o .  . Association with R9AP is essential for RGS11⅐G␤ 5 -mediated regulation of mGluR6-G␣ o signaling in Xenopus oocytes. Control oocytes were injected with cRNAs for mGluR6, G␣ o C351G, and GIRK1 and GIRK4 subunits of heteromultimeric GIRK channels and with cRNA for the catalytic subunit of pertussis toxin to eliminate coupling between mGluR6 and endogenous G i/o G proteins. Some of these control oocytes were subsequently injected with either water or one of the following cRNA combinations in addition to those described above: (i) RGS11 and G␤ 5 or (ii) RGS11, G␤ 5 , and R9AP. A, normalized averaged traces of glutamate (10 M) evoked GIRK currents. The figure shows activation (downward deflection, left) and deactivation (upward deflection, right) phases of the glutamate-evoked GIRK currents in the three oocyte groups. B, a comparison is shown of the activation and deactivation rate constants (1/), where is the time constant derived from the exponential fits of the activation and deactivation phases of the GIRK currents in the different oocyte groups. Data are expressed as the means Ϯ S.E. from 7-9 oocytes from the same donor. The mean steady-state current amplitudes in nA Ϯ S.E. for the different oocyte groups were as follows: control, 941 Ϯ 37; RGS11⅐G␤ 5 , 1024 Ϯ 255; RGS11⅐G␤ 5 ⅐R9AP, 658 Ϯ 76. Similar results were obtained from multiple oocyte donors. *, significantly different from control (p Ͻ 0.05 by Student's t test). WB, Western blot. C, whole protein extracts were prepared from oocytes after electrophysiological recording and expression levels of RGS11 and R9AP were analyzed by Western blotting with specific antibodies or by Coomassie Brilliant Blue (CBB) staining.
Results of this study suggest that the stimulatory effect of R9AP toward the RGS11⅐G␤ 5 complex involves an allosteric enhancement of GAP function acting in addition to the potentiation produced by membrane recruitment (see supplemental Fig. 5 for a schematic illustration). First, we found that RGS11⅐G␤ 5 has an intrinsic propensity to associate directly and independently of R9AP with the uROS membranes, a broken photoreceptor cell preparation. However, recruitment of RGS11⅐G␤ 5 to the same membranes by R9AP produces greater potentiation of its GAP activity that cannot be simply explained by the concentration of the complex at the membrane. Furthermore, in a cellular system reconstituted with the physiologically relevant receptor, mGluR6, RGS11⅐G␤ 5 did not detectably accelerate the GTPase activity of the mGluR6-coupled G protein G␣ o unless R9AP was co-expressed and conferred strong GAP activity to the RGS11⅐G␤ 5 complex. Interestingly, the R9AP-mediated enhancement of RGS11⅐G␤ 5 GAP function at G␣ o coupled to rhodopsin in the uROS system was only ϳ2.5-fold. This observation suggests that the stimulatory effects of R9AP exhibit GPCR selectivity and, cannot be solely explained by membrane recruitment. Finally, comparing the activity of the RGS9-1⅐G␤ 5 complex in the same rhodopsin-based system reveals that despite equal membrane recruitment, R9AP has much greater stimulatory effect toward G␣ t (ϳ70-fold; Ref. 21) as compared with G␣ o (ϳ5-fold; this study). Taken together, these observations indicate that the allosteric effects of R9AP at the R7 RGS proteins might in fact contribute to the G protein and GPCR selectivity of the R7 RGS proteins. In summary, our study establishes R9AP as a subunit of R7 RGS complexes with a general function of GAP activity modulator that acts by allosteric mechanisms in addition to membrane recruitment.