AGS3 Inhibits GDP Dissociation from Gα Subunits of the Gi Family and Rhodopsin-dependent Activation of Transducin*

A number of recently discovered proteins that interact with the α subunits of Gi-like G proteins contain homologous repeated sequences named G protein regulatory (GPR) motifs. Activator of G protein signaling 3 (AGS3), identified as an activator of the yeast pheromone pathway in the absence of the pheromone receptor, has a domain with four such repeats. To elucidate the potential mechanisms of regulation of G protein signaling by proteins containing GPR motifs, we examined the effects of the AGS3 GPR domain on the kinetics of guanine nucleotide exchange and GTP hydrolysis by Giα1 and transducin-α (Gtα). The AGS3 GPR domain markedly inhibited the rates of spontaneous guanosine 5′-O-(3-thiotriphosphate) (GTPγS) binding to Giα and rhodopsin-stimulated GTPγS binding to Gtα. The full-length AGS3 GPR domain, AGS3-(463–650), was ∼30-fold more potent than AGS3-(572–629), containing two AGS3 GPR motifs. The IC50values for the AGS3-(463–650) inhibitory effects on Giα and transducin were 0.12 and 0.15 μm, respectively. Furthermore, AGS3-(463–650) and AGS3-(572–629) effectively blocked the GDP release from Giα and rhodopsin-induced dissociation of GDP from Gtα. The potencies of AGS3-(572–629) and AGS3-(463–650) to suppress the GDP dissociation rates correlated with their ability to inhibit the rates of GTPγS binding. Consistent with the inhibition of nucleotide exchange, the AGS3 GPR domain slowed the rate of steady-state GTP hydrolysis by Giα. The catalytic rate of Gtα GTP hydrolysis, measured under single turnover conditions, remained unchanged with the addition of AGS3-(463–650). Altogether, our results suggest that proteins containing GPR motifs, in addition to their potential role as G protein-coupled receptor-independent activators of Gβγ signaling pathways, act as GDP dissociation inhibitors and negatively regulate the activation of a G protein by a G protein-coupled receptor.

ulate GDP/GTP exchange on the ␣ subunits of G proteins. Following the activational interaction with receptors, G␣GTP and G␤␥ are released to activate their targets, which include adenylyl cyclases, phospholipases, phosphodiesterases, and ion channels (1)(2)(3). A novel class of GTPase-activating proteins (GAPs) for G proteins termed regulators of G protein signaling (RGS) has been identified (4 -6). RGS proteins share a highly conserved RGS domain, which is responsible for the GAP function. Recently, cloning of proteins critical for glial cell development resulted in the identification of the first known Drosophila RGS protein, LOCO (7). The LOCO sequence revealed significant homology to RGS12 and RGS14 within the RGS domain and three additional regions B, C, and D (7). A yeast two-hybrid screen was carried out using G i ␣ as bait in an attempt to confirm the interaction of LOCO with G␣. Interestingly, the D region, rather than the RGS domain of LOCO, was found to bind G i ␣ (7). Sequence analysis of the D region revealed that it contained a segment of homology with four ϳ20amino acid repeats present in the human mosaic protein, LGN. LGN has been previously identified as a G i ␣ 2 -interacting protein using a yeast two-hybrid system (8).
LGN is similar to the activator of G protein signaling 3 (AGS3), which was isolated in a functional screen for receptor-independent activators of heterotrimeric G protein signaling (9). Site-directed mutagenesis and protein interaction studies with AGS3 (9) indicated that the ϳ20-amino acid repeats common to AGS3, LGN, and LOCO were responsible for binding G i ␣. The ϳ20-amino acid repeats were termed the G protein regulatory (GPR) (9) or GoLOCO motif (10). The GPR motif was also identified in Purkinje cell protein-2 (Pcp2) and Rap1GAP, which were identified as G o ␣ binding partners in yeast-two hybrid screens (11,12). These studies suggest that GPR-containing proteins, hereafter termed GPR proteins, are likely to represent a diverse family of proteins that modulate G protein signaling.
At present very little is known about the mechanisms and functions of GPR proteins. The yeast pheromone response pathway is mediated by G␤␥ subunits, and its GPCR-independent activation by AGS3 suggests that it may induce release of G␤␥ from G proteins (9). Pcp2 protein was shown to stimulate GDP release from G o ␣ without affecting the k cat for GTP hydrolysis, thus raising the possibility that GPR proteins may serve as guanine nucleotide exchange factors for G proteins (11). To date, no studies on the regulation of GPCR-mediated G protein activation by GPR proteins have been reported. In this study, we examined the effects of the AGS3 GPR domain (AGS3GPR) on the intrinsic guanine nucleotide exchange of G i ␣ 1 and transducin-␣ (Gt␣) and on the rhodopsin-stimulated nucleotide exchange of G t ␣. Our results suggest that AGS3GPR acts as a GDP dissociation inhibitor (GDI) and may block GPCR-dependent activation of G proteins from the G i family.
Cloning and Expression of AGS3GPR Constructs-A cDNA sequence corresponding to residues AGS3-(572-629) of rat AGS3 and containing GPR motifs III (aa 572-590) and IV (aa 606 -624) (9) was polymerase chain reaction-amplified from the human retinal cDNA lambda gt10 library (provided by J. Nathans, The Johns Hopkins University) using the following primers: ACTTAATCTAGACGACTTCTTCAACATGCTC-ATC and ATCTCTCTCGAGCTAGTCCACCCGCTGCTCGTCCATG. The primers were designed based on the cDNA sequence of a human homologue of AGS3 (GenBank TM accession number 5911952), identified using a Basic Local Alignment Search Tool search at NCBI (Bethesda, MD). In comparison to the rat AGS3-(572-629) sequence, the human sequence had a single homologous substitution, Glu-5733 Asp. The polymerase chain reaction DNA fragment was digested with XbaI and XhoI and subcloned into a pGEX-KG vector (18) digested with the same enzymes. The sequence of an insert was confirmed by automated DNA sequencing at the University of Iowa DNA Core Facility. An additional construct, GST-AGS3-(463-650), was prepared as described. 2 The GST-AGS3-(463-650) construct consists of the last 189 amino acids of rat AGS3 and includes all four GPR motifs (GPR-I (aa 470 -489), GPR-II (aa 524 -542), GPR-III (aa 572-590), and GPR-IV (aa 606 -624)). GST-AGS3-(572-629) and GST-AGS3-(463-650) were expressed in BL21 cells and purified using glutathione-agarose as described (20). The yields of soluble proteins were 15-20 mg/liter of culture.
GTP␥S Binding Assay-G i ␣ 1 and G t ␣ subunits (0.5 M) alone or G t ␣ The filters were then washed three times with 1 ml of the same buffer (ice-cold) and counted in a liquid scintillation counter after dissolution in 3a70B mixture. The apparent rate constant (k app ) values for the binding reactions were calculated by fitting the data to the following equation and uROS membranes (100 nM rhodopsin) with 2 M [␣-32 P]GTP in buffer A for 40 min at 25°C. uROS membranes were removed by centrifugation (30 min, 100,000 ϫ g). Excess of unlabeled GTP (1 mM), uROS membranes (5 nM rhodopsin, only for G t ), and varying concentrations of AGS3-(572-629) or AGS3-(463-650) were added to monitor dissociation of [␣-32 P]GDP from G␣ subunits. Aliquots were withdrawn at the indicated times and passed through Whatman cellulose nitrate filters (0.45 m). The dissociation rate constants (k off ) were calculated by fitting the experimental data to a single exponential decay function: %GDP bound ϭ 100%⅐e Ϫkt .
GTPase Assays-Steady-state GTPase activity measurements were initiated by mixing G i ␣ 1 (1 M) with 10 M [␥-32 P]GTP in 100 l of buffer A. Aliquots (10 l) were withdrawn at the indicated times and transferred to 100 l of 7% perchloric acid. Nucleotides were precipitated using 700 l of 10% charcoal suspension, and 32 P i formation was measured by liquid scintillation counting.
Single turnover GTPase activity measurements were carried out in suspensions of uROS membranes (5 M rhodopsin) reconstituted with G t ␣ (0.5 M) and G t ␤␥ (1 M) essentially as described (21). The reaction was initiated by mixing bleached ROS membranes with 50 nM [␥-32 P]GTP (ϳ5 ϫ 10 4 dpm/pmol) in a total volume of 20 l. The reaction was quenched by the addition of 100 l of 7% perchloric acid. Nucleotides were then precipitated using charcoal, and 32 P i formation was measured by liquid scintillation counting. The GTPase rate constants were calculated by fitting the experimental data to an exponential function: %GTP hydrolyzed ϭ 100⅐(1 Ϫ e Ϫkt ), where k is the rate constant for GTP hydrolysis.
Other Methods-Protein concentrations were determined by the method of Bradford (22) using IgG as a standard or using calculated extinction coefficients at 280 nm. The experimental data were fitted with nonlinear least squares criteria using GraphPad Prizm (v.2) software.

RESULTS
Effects of AGS3GPR on Guanine Nucleotide Binding to G i ␣ 1 -Two AGS3 GPR polypeptides, AGS3-(572-629) and AGS3-(463-650), were utilized to investigate the modulation of guanine nucleotide binding to G i ␣ 1 and G t ␣. AGS3-(572-629) contains the GPR repeats III and IV of AGS3, whereas AGS3-(463-650) includes all four AGS3 GPR repeats. G i ␣ 1 has a lower affinity for GDP than G t ␣ and, consequently, a notable intrinsic nucleotide exchange rate. The GTP␥S binding to G i ␣ 1 , because of intrinsic nucleotide exchange, was characterized by a k app of 0.11 min Ϫ1 (Fig. 1A). Addition of increasing concentrations of AGS3-(572-629) led to a dose-dependent decrease in the GTP␥S binding rates. In the presence of 15 M AGS3-(572-629) the rate of GTP␥S binding to G i ␣ 1 was reduced (k app ϭ 0.018 min Ϫ1 ) (Fig. 1A). The calculated IC 50 value for the inhibitory effect of AGS3-(572-629) was 3.9 M (Fig. 1C). The fulllength GPR domain, AGS3-(463-650) was a significantly more potent inhibitor of GTP␥S binding to G i ␣ 1 . It inhibited the kinetics of GTP␥S binding by G i ␣ 1 with an IC 50 value of 0.12 M (Fig. 1, B and C). appreciably alter the rate of GTP␥S binding to G s ␣, reflecting a lack of interaction between these proteins (Fig. 1D). Two different mechanisms could possibly account for the inhibitory effect of AGS3GPR on GTP␥S binding to G i ␣ 1 . First, AGS3 blocks the dissociation of GDP and, as a result, binding of GTP␥S. Alternatively, AGS3 promotes release of both GDP and GTP (GTP␥S), producing an empty pocket G i ␣. The rate of GDP dissociation of G i ␣ 1 in the absence of AGS3GPR (k off ϭ 0.35 min Ϫ1 ) ( Fig. 2A) was higher than the GTP␥S binding rate in Fig. 1A. Although it is thought that GDP release is a ratelimiting step in the nucleotide exchange by G␣ subunits (24), rates of GDP dissociation considerably exceeding GTP␥S binding rates have also been reported (25). Fig. 2A shows that AGS3-(572-629) significantly slowed the rate of GDP release. The potency of AGS3-(572-629) as a GDI correlated with its inhibition of GTP␥S binding. Not surprisingly, the full AGS3 GPR domain, AGS3-(463-650), was a more effective GDI than was AGS3-(572-629) (Fig. 2B).
Effects of AGS3GPR on Rhodopsin-catalyzed Nucleotide Exchange on G t ␣-G t ␣ binds GDP very tightly and practically does not exchange nucleotides in the absence of R*. The slow GTP␥S binding rate of G t ␣ (k app ϭ 0.001 min Ϫ1 ) in the presence or absence of G t ␤␥ was not affected by addition of AGS3-(572-629) or AGS3-(463-650) (data not shown). The ability of the AGS3GPR to influence R*-induced GTP␥S binding to transducin was tested next. In the presence of ROS membranes containing 5 nM bleached rhodopsin, the rate constant of GTP␥S binding to G t ␣␤␥ was 0.040 min Ϫ1 (Fig. 3A). AGS3-(572-629) significantly suppressed the activation of transducin by R* (Fig. 3A). The IC 50 value for the inhibitory effect of AGS3-(572-629) was 4.6 M (Fig. 3C). As with G i ␣ 1 , AGS3-(463-650) was a more effective inhibitor of transducin activation than was AGS3-(572-629) (Fig. 3B). The IC 50 value for the effect of AGS3-(463-650) was 0.15 M (Fig. 3C). The R*-stimulated GDP dissociation rates from G t ␣␤␥ in the presence of AGS3GPR were measured to determine whether the inhibition of GTP␥S binding to G t ␣ by AGS3GPR was due to its effect on the GDP release. AGS3-(572-629) and AGS3-(463-650) reduced GDP release rates from transducin proportionately to their ability to inhibit the R*-induced GTP␥S binding (Fig. 4).
Effects of AGS3GPR on Steady-state and Single Turnover GTP Hydrolysis-The rate of steady-state GTP hydrolysis is limited by the rate of nucleotide exchange (26). Therefore, a steady-state GTPase assay provides an additional method of confirming the inhibition of G␣ nucleotide exchange by AGS3GPR. The rate of steady-state GTP hydrolysis by G i ␣ 1 was 0.028 mol of P i /mol of G i ␣⅐min) (Fig. 5A). Reflecting the ability to inhibit nucleotide exchange, AGS3-(463-650) effectively suppressed the steady-state rate of GTP hydrolysis by G i ␣ 1 (Fig. 5A).
To test the possibility that AGS3GPR regulates a catalytic step of GTP hydrolysis (k cat ), transducin GTPase activity was measured under single turnover conditions ([GTP]Ͻ [G t ]) (21). The k cat for GTP hydrolysis by transducin reconstituted with uROS was 0.015 s Ϫ1 (Fig. 5B). Addition of 5 M AGS3-(463-650) had no effect on GTP hydrolysis (k cat ϭ 0.014 s Ϫ1 ).

DISCUSSION
Recent findings have identified a novel domain present singly or as multiple repeats in a variety of G protein-interacting proteins (7)(8)(9)(10)(11)(12). The first indication that this domain mediates the interaction with G␣ subunits came from the analysis of the RGS protein LOCO from Drosophila (7). Subsequently this sequence motif containing 19 amino acid residues has been termed GoLOCO (10) or, more generally, a GPR motif (9). GPR proteins identified to date appear to specifically target G␣ subunits from the G i family. GST pull-down experiments showed association of Pcp2 with G o ␣ and G i ␣ but not G s ␣ (11). Similarly, Rap1GAP selectively interacted with G o ␣ and G i ␣ but not with G s ␣ and G q ␣ (12). AGS3 activated the Saccharomyces cerevisiae pheromone response in the G i ␣ genetic background but not in yeast strains expressing Gpa1, G s ␣, or G 16 ␣ (9). The specificity of GPR proteins toward different conformations of G␣ remains unclear. Pcp2 bound equally the GDP and GTP␥S conformations of G o ␣ (11). In contrast, AGS3 displayed preferential interaction with the GDP-bound G␣ subunits (9). Rap1GAP was shown to interact more avidly with G o ␣GDP (12), whereas the N-terminally extended isoform of Rap1GAP, Rap1GAPII, preferentially bound the GTP-complexed G i ␣ (19). Modulation of the Rap1GAP and Rap1GAPII activities by G␣ subunits suggests an interesting link via GPR motifs between G protein signaling and cascades involving small monomeric GTPases (12,19). Currently, the mechanisms of regulation of G protein signaling by GPR proteins are not well understood. By dissociating heterotrimeric complexes G␣␤␥, GPR proteins could serve as GPCR-independent activators of G protein pathways that utilize G␤␥ subunits to transduce signals (9). Effects of GPR proteins on GPCR-dependent G protein signaling have not yet been investigated. To examine potential mechanisms of G protein regulation by GPR proteins, we analyzed the effects of AGS3GPR on the guanine nucleotide exchange of G i ␣ and G t ␣. Two AGS3 constructs, AGS3-(572-629) and AGS3-(463-650), containing two and four GPR motifs, respectively, have been utilized to assess the role of domain multiplicity. The AGS3 GPR constructs markedly inhibited spontaneous GTP␥S binding to G i ␣ 1 because of its intrinsic nucleotide exchange. The full-length AGS3 GPR domain, AGS3-(463-650), was approximately 30-fold more potent than AGS3-(572-629), indicating the possibility that several G␣ subunits can simultaneously bind to a single AGS3 molecule. In addition, a better folding of AGS3-(463-650) cannot be excluded. Further experiments demonstrated that AGS3GPR inhibited the rate of GDP dissociation from G i ␣ 1 . The effectiveness of AGS3-(572-629) and AGS3-(463-650) in suppressing the G i ␣ 1 GDP dissociation rate correlated with their ability to inhibit the rate of GTP␥S binding. The inhibitory effect of AGS3GPR on GTP␥S binding by G i ␣ 1 might be entirely due to the decrease in the GDP dissociation rate, although the higher rate of GDP dissociation precludes a firm conclusion. The GDI activity of AGS3GPR is in contrast with an earlier observation that Pcp2 stimulates GDP release from G o ␣ (11). The basis for the opposite effects of AGS3 and Pcp2 is not clear. It may, conceivably, be related to the structural differences in AGS3 and Pcp2 GPR motifs and/or to Pcp2 sequences outside its GPR domain. AGS3-(572-629) and AGS3-(463-650) had no effects on the basal nucleotide exchange rate of G t ␣. An unstimulated GTP␥S binding rate of G t ␣ is extremely slow (k app Ӎ 0.001 min Ϫ1 ), and therefore it would be practically impossible to discern an inhibitory effect of AGS3GPR. However, any significant stimulation of G t ␣ GDP/ GTP exchange by AGS3 would have been seen.
Transducin activation by R* provided an excellent model system to study the potential mechanisms of regulation of GPCR/G protein coupling by GPR proteins. AGS3-(572-629) and AGS3-(463-650) effectively inhibited activation of transducin by R* in the GTP␥S binding assay. Furthermore, AGS3GPR was capable of blocking the R*-induced GDP release from G t ␣. By analogy to its effect on G i ␣ 1 , AGS3GPR is likely to act directly on G t ␣ as a GDI. Other factors may have contributed to the inhibitory effect of AGS3GPR on transducin activation. AGS3 is capable of dissociating G␣ and G␤␥ (9), and this would also lead to an impairment in the G t -R* coupling. Furthermore, AGS3GPR and R* may compete for binding to G t ␣, thus causing the inhibition. We did not find any evidence that AGS3GPR affects the k cat for GTP hydrolysis by G t ␣. Interestingly, some RGS proteins, such as RGS12 and RGS14, contain GPR motifs (10). These proteins might be capable of very effective dual inhibition of G protein signaling at the level of activation by receptor and at the level of inactivation due to accelerated GTP hydrolysis.
Overall, our results demonstrate the capacity of GPR proteins for negative modulation of GPCR-dependent G protein activation. GPR proteins appear to have an interesting dual potential to regulate G protein signaling. AGS3 may act as a selective activator of G␤␥-regulated effector systems independent of a receptor, while at the same time inhibiting the activation of G proteins by a GPCR.