Stabilization of the GDP-bound Conformation of Giα by a Peptide Derived from the G-protein Regulatory Motif of AGS3*

The G-protein regulatory (GPR) motif in AGS3 was recently identified as a region for protein binding to heterotrimeric G-protein α subunits. To define the properties of this ∼20-amino acid motif, we designed a GPR consensus peptide and determined its influence on the activation state of G-protein and receptor coupling to G-protein. The GPR peptide sequence (28 amino acids) encompassed the consensus sequence defined by the four GPR motifs conserved in the family of AGS3 proteins. The GPR consensus peptide effectively prevented the binding of AGS3 to Giα1,2 in protein interaction assays, inhibited guanosine 5′-O-(3-thiotriphosphate) binding to Giα, and stabilized the GDP-bound conformation of Giα. The GPR peptide had little effect on nucleotide binding to Goα and brain G-protein indicating selective regulation of Giα. Thus, the GPR peptide functions as a guanine nucleotide dissociation inhibitor for Giα. The GPR consensus peptide also blocked receptor coupling to Giαβγ indicating that although the AGS3-GPR peptide stabilized the GDP-bound conformation of Giα, this conformation of GiαGDPwas not recognized by a G-protein coupled receptor. The AGS3-GPR motif presents an opportunity for selective control of Giα- and Gβγ−regulated effector systems, and the GPR motif allows for alternative modes of signal input to G-protein signaling systems.

The G-protein regulatory (GPR) 1 motif or GoLOCO repeat is a ϳ20-amino acid domain found in several proteins that interact with and/or regulate G-proteins (1,2). Such proteins include the activator of G-protein signaling AGS3, the AGS3related protein PINS in Drosophila melanogaster, two members of the RGS family of proteins, and three proteins (LGN, Pcp2, and Rap1GAP) isolated in yeast two-hybrid screens using Gi␣ or Go␣ as bait. Rat AGS3 was isolated in a yeast-based functional screen designed to identify receptorindependent activators of heterotrimeric G-protein signaling (1). The AGS3-related protein PINS is required for asymmetric cell division of neuroblasts in D. melanogaster, where it is found complexed with Gi/Go (3,4), but neither the signal input nor output for this complex is known. Some insight as to how PINS may regulate Gi/Go is provided by studies with AGS3 (1). In the yeast-based system, AGS3 selectively activated Gi␣2 and Gi␣3. The action of AGS3 as a G-protein activator in the yeast-based system was independent of nucleotide exchange as it was not antagonized by overexpression of RGS4, and it was still observed following replacement of Gi␣2 with Gi␣2-G204A, a mutant that is deficient in making the transition to the GTP-bound state (1,5). Both of these manipulations effectively prevent receptor-mediated activation of G-protein signaling in the yeast system and block the action of AGS1, which was isolated in the same screen and apparently behaves as a guanine nucleotide exchange factor for heterotrimeric G-proteins (5,6). These data indicate that the interaction of AGS3 with G-protein influences a unique control mechanism within the activation/deactivation cycle of heterotrimeric G-proteins.
AGS3 exists as a 650-amino acid protein enriched in brain and a 166-amino acid protein (AGS3-SHORT) enriched in heart (1). 2,3 The 650-amino acid protein consists of two functional domains defined by a series of seven amino-terminal tetratrico peptide repeats (TPR) and four carboxyl-terminal GPR motifs. Site-directed mutagenesis, protein interaction studies, and subcellular localization experiments indicated that the GPR motifs of AGS3 were likely responsible for binding G-protein, whereas the TPR domain is a site for binding of regulatory proteins (1, 3, 4). 2,3 AGS3 preferentially binds to G␣ in the presence of GDP (1). AGS3-GPR effectively competed with G␤␥ subunits for binding to Gt␣ and inhibited guanosine 5Ј-O-(3thiotriphosphate) (GTP␥S) binding to Gi␣1. 2 Such an activity likely has significance in a number of aspects of G-proteinmediated signaling events and presents a novel opportunity to control the basal activity of G-protein signaling, as well as influence receptor-mediated activation of G-protein. These observations also raise many interesting questions relative to basic aspects of G-protein structure/function and alternative modes of regulation and functional roles for G-protein signaling systems in the cell. To address these issues, we generated a series of peptides based upon the consensus GPR motif in AGS3 and evaluated their effects on the nucleotide binding properties of Gi␣. A 28-amino acid GPR peptide effectively blocked the interaction of AGS3 with Gi␣ and inhibited GTP␥S binding to Gi␣ by a mechanism that involved stabilization of the GDPbound conformation of Gi␣. The GPR consensus peptide also blocked receptor coupling to Gi␣␤␥ indicating that although the AGS3-GPR peptide stabilized the GDP-bound conformation of Gi␣, this conformation of Gi␣ GDP was not recognized by a G-protein-coupled receptor.
Protein Interaction Assays-The GPR domain of AGS3 (Pro 463 -Ser 650 ) containing the four GPR motifs was generated as a glutathione S-transferase fusion protein by polymerase chain reaction using the full-length cDNA of AGS3 as a template. The AGS3-Pro 463 -Ser 650 segment was also cloned into the pQE-30 vector (Qiagen, Valencia, CA) to generate an amino-terminal His-tagged protein. His-tagged AGS3 was expressed in and purified from bacteria using a nickel affinity matrix (ProBond™ resin; Invitrogen, Carlsbad, CA). The His-tagged AGS3 was eluted from the matrix with imidazole and desalted by centrifugation as with the GST fusion protein (1). The interaction of GST-AGS3-GPR and HIS-tagged AGS3-GPR with G-proteins was assessed by protein interaction experiments using purified G-protein as described previously (1). Gi␣1-3 and Go␣ were purified from Sf9 insect cells infected with recombinant virus as described (8). All purified G-proteins used in these studies were isolated in the GDP-bound form, and G-protein interaction assays contained 10 M GDP.
A separate series of protein interaction experiments were designed to determine whether the Gi␣ complexed with AGS3 contained bound GDP. Gi␣1 (100 nM) was loaded with 3 H-GDP (0.5 M; 2.0 ϫ 10 4 dpm/pmol) by incubation for 20 min at 24°C in binding buffer (50 mM Hepes-HCl, pH 7.5, 1 mM EDTA, 1 mM dithiothreitol, 50 M adenosine triphosphate, and 10 g/ml bovine serum albumin). The 3 H-GDPloaded Gi␣1 was incubated with 300 nM GST or GST-AGS3-GPR in the presence and absence of 10 M GPR peptide and processed as described (1). The washed resin containing bound proteins was transferred to vials for measurement of 3 H-GDP by liquid scintillation spectroscopy.
GTP␥S Binding Assays-GTP␥S binding assays were generally conducted as described (9). G-proteins (100 nM) were preincubated for 20 min at 24°C in the presence and absence of GPR peptides. Binding assays (duplicate determinations) were initiated by addition of 0.5 M GTP␥S (4.0 ϫ 10 4 dpm/pmol), and incubations (total volume ϭ 50 l) were continued for 30 min at 24°C. Both preincubations and GTP␥S binding assays were conducted in binding buffer containing 2 mM MgCl 2 . Reactions were terminated by rapid filtration through nitrocellulose filters with 4 ϫ 4-ml washes of stop buffer (50 mM Tris-HCl, 5 mM MgCl 2 , 1 mM EDTA, pH 7.4, at 4°C). Radioactivity bound to the filters was determined by liquid scintillation counting. Nonspecific binding was defined by 100 M GTP␥S.
GDP Dissociation Assays-Gi␣1 (100 nM) was loaded with 3 H-GDP (0.5 M; 2 ϫ 10 4 dpm/pmol) by incubation for 20 min at 24°C in binding buffer without MgCl 2 . 45-l aliquots of the preincubation mixture (ϳ500,000 dpm) were then added to incubation tubes containing 5 l of vehicle or peptide, and samples were incubated for 30 min at 24°C. Two sets of tubes were set up for each time point to be analyzed. Each set contained duplicate samples for determination of total binding, nonspecific binding, or binding in the presence of peptide. For each time point, one set of tubes served as an internal time control, whereas the other set received added GTP␥S or GDP to initiate dissociation. Data are expressed as % of control, where control represents the level of 3 H-GDP binding at each time point in the set of tubes that did not receive added nucleotide to initiate dissociation. The amount of 3 H-GDP bound fol-lowing the 20-min preincubation (ϳ30,000 dpm) was identical to that observed at the 30-min incubation time point following addition of vehicle or peptide. 3 H-GDP dissociation was initiated by addition of GTP␥S or GDP in a volume of 5 l (final concentration, 100 M). Reactions were terminated at specified time points by rapid filtration through nitrocellulose filters (BA85; Schleicher & Shuell) with 4 ϫ 4-ml washes of stop buffer. Radioactivity bound to the filters was determined by liquid scintillation counting. Nonspecific binding was defined by 100 M GDP.
High Affinity Agonist Binding-Sf9 cell membranes expressing 5-HT 1A receptors were reconstituted with G␣␤␥, and high affinity agonist binding was measured with 3 H-5-HT as described previously (8,10). Membrane aliquots (100 g of membrane protein, 85 nM receptor) were preincubated for 15 min at 25°C with G-proteins (2125 nM G␣␤␥) with or without GPR peptides in a total volume of 17 l (reconstitution buffer, 5 mM NaHEPES, 100 mM NaCl, 5 mM MgCl 2 , 1 mM EDTA, 500 nM GDP, 0.04% CHAPS, pH 7.5). The reconstitution mixtures were then diluted 10-fold with binding buffer (50 mM Tris-HCl, 5 mM MgCl 2 , 0.5 mM EDTA, pH 7.5), and 50 l were added to binding tubes (total volume ϭ 150 l) containing 2 nM 3 H-5-HT. The final concentrations of receptor, G-protein, and peptide in the binding tubes were 2.8 nM, 70.8 nM, and 114 M, respectively. Nonspecific binding was determined in the presence of 100 M 5-HT. Binding reactions were incubated at 25°C for 1.5 h and terminated by filtration over Whatman GF/C FP200 filters using a Brandel cell harvester. The filters were rinsed thrice with 4 ml of ice-cold washing buffer (50 mM Tris-Cl, 5 mM MgCl 2 , 0.5 mM EDTA, 0.01% sodium azide, pH 7.5, at 4°C), placed in 4.5 ml of CytoScint, and counted to constant error in a scintillation counter.

RESULTS AND DISCUSSION
The ϳ20-amino acid GPR motif is repeated four times in AGS3-related proteins, with the exception of the three repeats found in the Drosophila protein PINS (Fig. 1). Alignment of the four GPR repeats from five species revealed a GPR consensus sequence (Fig. 1). The GPR consensus sequence is characterized by the upstream negative charge (Glu-Glu) and hydrophobic cluster (Phe-Phe), Leu/Met 10 , Leu/Ile 11 , Gln 15 , Ser/Ala 16 , Arg 18 , Met/Leu 19 , and the Asp-Asp-Gln-Arg sequence at the carboxyl end of the motif. Helical wheel and Chou-Fasman analysis indicated that this region is capable of existing as an amphipathic helix. Each of the GPR motifs illustrated in Fig. 1 possess a varying number of Proline residues just after and in some cases before the core consensus sequence, which may exert an important influence within the overall organization of the four GPR motifs. As part of an effort to define the structural basis of the interaction of AGS3 with Gi␣ and the functional consequences of this interaction, we asked whether a consensus sequence peptide effectively interacted with Gi␣.
The core GPR consensus sequence was bracketed by additional residues (three amino terminus, five carboxyl terminus) derived from AGS3-GPR-IV, and the carboxyl terminus was amidated (Fig. 1). The 28-amino acid GPR consensus peptide completely blocked the binding of Gi␣1 or Gi␣2 to GST-AGS3-GPR with an IC 50 of ϳ200 nM ( Fig. 2A and B). The GPR consensus peptide also inhibited GTP␥S binding to Gi␣1 and Gi␣2 (IC 50 ϳ200 nM) (Fig. 2, C and D) consistent with the preferential binding of AGS3 to Gi␣ in the presence of GDP (1). The inhibitory effect of the GPR consensus peptide on GTP␥S binding was selective for Gi␣ as it only minimally affected nucleotide binding to Go␣ or brain G-protein (Fig. 2D). The activity of the GPR consensus peptide in both the protein interaction assays and GTP␥S binding assays was lost upon substitution of Phe for the highly conserved Arg 23 (Fig. 2, B, C,  and D). However, substitution of Ala for the invariant Gln 15 did not alter the activity of the GPR peptide (Fig. 2B). 4 Similar results were obtained when these amino acid substitutions were made in the context of GST-AGS3 fusion protein, which contained the terminal 74 amino acids of AGS3 including part of GPR-III and all of GPR-IV ( Fig. 1) (1). 2 We then addressed the mechanism by which the GPR consensus peptide inhibited GTP␥S binding to Gi␣2 and determined the effect of the GPR motif on receptor coupling to G-protein. The inhibition of GTP␥S binding to Gi␣ by the GPR consensus peptide may reflect a reduction in the rate of nucleotide exchange. Indeed, the rate of GDP dissociation was markedly diminished in the presence of the GPR consensus peptide (Fig. 3A). 5 The R23F mutation, which eliminated the effectiveness of the peptide to block interaction of AGS3 with Gi␣ and GTP␥S binding to Gi␣, also did not alter GDP dissociation (Fig.  3A). The inhibition of GDP dissociation by the GPR consensus peptide suggests that the GPR motif is stabilizing the GDPbound conformation of Gi␣. To address this issue we evaluated the interaction of GST-AGS3-GPR with Gi␣2, which had been preloaded with 3 H-GDP. Subsequent analysis of the G-protein complexed with AGS3-GPR on the glutathione affinity matrix indicated that the nucleotide binding site of G-protein bound to AGS3 indeed contained GDP (Fig. 3B). Gi␣ GDP binding to AGS3-GPR was blocked by the GPR consensus peptide (Fig.  3B) consistent with the ability of this peptide to inhibit interaction of GST-AGS3-GPR with Gi␣1/2 (Fig. 2, A and B).
The stabilization of the GDP-bound conformation of Gi␣ by the GPR consensus peptide indicates that the AGS3-GPR motif can influence subunit interactions by interfering with G␤␥ binding to Gi␣. 2 This apparent effect may account for the results obtained in protein interaction assays using GST-AGS3-GPR and brain lysates, where G␤␥ is absent from the AGS3-Gi␣ complex (1). The influence of the GPR motif on subunit interactions would have significant implications for signal processing. First, interaction of the AGS3-GPR motif with G␣␤␥ would release G␤␥ for regulation of downstream signaling events, while stabilizing G␣ GDP (1). Such a mode of signal input may be of utility where there is a need for selective regulation of G␤␥-sensitive effectors. The time frame for termination of such a signaling event (i.e. reassociation of G␤␥ with Gi␣ GDP ) likely differs from that of a more typical signaling event in which there has been an exchange of nucleotide bound to Gi␣, and signal termination involves GTP hydrolysis along with subunit reassociation. A second implication of stabilization of G␣ GDP by a GPR domain is related to receptor G-protein coupling. We addressed this issue experimentally using a membrane assay system where receptor-G-protein coupling is reflected as high affinity binding of agonists. The high affinity binding of agonist observed upon reconstitution of the mem-  A and B, the carboxyl region of AGS3 (Pro 463 -Ser 650 ) containing all four GPR repeats was generated as a His-tagged (A) or GST fusion (B) protein for protein interaction assays as described under "Experimental Procedures." All interactions were done in the presence of 10 M GDP, and the input lanes represent one-tenth of the G-protein used in each interaction assay. A, Gi␣1 (75 nM) was incubated with 300 nM His-tagged AGS3-GPR in the absence and presence of increasing amounts of the GPR peptide, after which bound Gi␣ was isolated on a nickel affinity matrix, and samples were processed for immunoblotting with Gi␣ antisera. The blot in the upper panel of A was stripped and reprobed with AGS3 antisera to provide internal controls for protein loading. B, Gi␣2 (75 nM) was incubated with 300 nM AGS3-GPR or GST in the presence and absence of 100 M GPR consensus peptide, GPR peptide Q15A, or GPR peptide R23F. The immunoblots presented in A and B are representative of three experiments. C and D, GTP␥ 35 S binding to G-proteins (100 nM) was measured in the absence and presence of peptides as described under "Experimental Procedures." Data are expressed as the percent of specific binding (ϳ5 pmol) observed in the absence of peptide and represent the means Ϯ S.E. derived from three experiments. The concentration of peptides in D is 10 M. brane-bound 5-HT 1 receptor with Gi␣␤␥ was inhibited by addition of the GPR peptide (Fig. 4). This action of the GPR peptide was not observed with the R23F peptide and was selective for Gi versus Go (Fig. 4).
The influence of the single amino acid substitution on the bioactivity of the GPR both within the context of a short peptide and a GST fusion protein containing an additional 74 amino acids of AGS3 sequence strongly suggest a relatively discrete and specific surface interaction with Gi␣ ( Fig. 2) (1). 2,6 Helical wheel projections and 3D models indicated that when the GPR consensus peptide is fixed in an ␣ helical conformation, the Phe 8 , Ala 12 , Gln 15 , Met 19 , and Arg 23 residues are on the same face of the helix. On this face of the helix is a hydrophobic sector defined by Phe 8 , Ala 12 , and Met 19 that is bound by polar residues, which may be involved in charge pairing to residues in Gi␣. As was the case for the R23F substitution, disruption of this hydrophobic sector by substitution of Arg for Phe 8 also resulted in a loss of activity for the GST-AGS3-GPR fusion protein in GTP␥S binding and protein interaction assays (1). 2 Thus, either extension (R23F substitution) or shortening (F8R) of the hydrophobic sector on this face of the helix resulted in a loss of bioactivity for the GPR motif. In contrast, strengthening of this hydrophobic sector by substitution of Ala for Gln 15 did not alter the activity of the GPR peptide. These data indicate an important role for a spatially constrained hydrophobic stretch of ϳ16.6 Å that is key for peptide interaction with Gi␣.
The inability of receptor to productively couple to G␣ GDP -GPR is of interest. The G␣ GDP conformation stabilized by the GPR peptide may differ from that stabilized by G␤␥ in such a manner that the receptor cannot recognize G␣. Indeed, the orientations of the amino and carboxyl domains of Gi␣1, which are important interaction sites with receptor, are quite different in the Gi␣ GDP and Gi␣ GDP ␤␥ structures (11)(12)(13). In addition to such differences in the structural orientation of Gi␣ domains interacting with receptor, it is likely that receptor contact points on G␤␥ also play a role in receptor-mediated activation of guanine nucleotide exchange (14 -18). Alternatively, the receptor may indeed interact with the G␣ GDP -GPR complex, but this interaction stabilizes a receptor conformation with low affinity for agonist (19). Ultimately, one may think of the G␣ GDP -GPR complex as a type of dimeric G-protein, and it is not clear what might provide signal input to such a complex.
Although the GPR motif is present in several proteins that interact with G␣ and/or regulate nucleotide binding/hydrolysis (1,2), these proteins have different and often opposing effects on the activation state of G-protein (20,21). 2,5 Pcp2, which contains two GPR motifs based upon this consensus sequence, actually appears to increase the dissociation of GDP from Go␣ (21). Thus, there are either subtle differences in this motif or other residues outside of this motif that play a key role in the specific functional output gendered by interaction of the GPR motif with G␣. Of note is the selective effects of the AGS3-GPR peptide for Gi␣ versus Go␣ in both nucleotide binding assays and the analysis of receptor coupling to G-proteins. Further dissection of the structural basis for this selectivity will provide clues as to the site of interaction of the GPR peptide with Gi␣ and the mechanism by which it stabilizes the GDP-bound conformation. One prominent area of sequence divergence between Go␣ and Gi␣ encompasses switch IV, a region implicated in the formation of Gi1␣ GDP multimers (11).
The role of AGS3 as a GDI is an unexpected concept for heterotrimeric G-proteins, although such proteins serve similar regulator roles for Ras-related G-proteins. Proteins containing the AGS3-GPR motif may promote dissociation of G␣ and G␤␥ in the absence of nucleotide exchange and present an opportunity for selective control of Gi␣-and G␤␥Ϫregulated effector systems. GPR-containing proteins likely play a role in regulating basal activity of G-protein signaling systems in the cell and provide alternative modes of signal input to G-protein signaling systems that may either augment, complement, or antagonize G-protein activation by GPCRs.