A Point Mutation in Gαo and Gαi1Blocks Interaction with Regulator of G Protein Signaling Proteins*

Regulator of G protein-signaling (RGS) proteins accelerate GTP hydrolysis by Gα subunits and are thought to be responsible for rapid deactivation of enzymes and ion channels controlled by G proteins. We wanted to identify and characterize Gi-family α subunits that were insensitive to RGS action. Based on a glycine to serine mutation in the yeast Gα subunit Gpa1sst that prevents deactivation by Sst2 (DiBello, P. R., Garrison, T. R., Apanovitch, D. M., Hoffman, G., Shuey, D. J., Mason, K., Cockett, M. I., and Dohlman, H. G. (1998) J. Biol. Chem. 273, 5780–5784), site-directed mutagenesis of αo and αi1 was done. G184S αo and G183S αi1 show kinetics of GDP release and GTP hydrolysis similar to wild type. In contrast, GTP hydrolysis by the G → S mutant proteins is not stimulated by RGS4 or by a truncated RGS7. Quantitative flow cytometry binding studies show IC50 values of 30 and 96 nm, respectively, for aluminum fluoride-activated wild type αo and αi1 to compete with fluorescein isothiocyanate-αo binding to glutathioneS-transferase-RGS4. The G → S mutant proteins showed a greater than 30–100-fold lower affinity for RGS4. Thus, we have defined the mechanism of a point mutation in αo and αi1 that prevents RGS binding and GTPase activating activity. These mutant subunits should be useful in biochemical or expression studies to evaluate the role of endogenous RGS proteins in Gi function.

Receptor-mediated activation of heterotrimeric guanine nucleotide-binding proteins initiates signals elicited by numerous hormone, neurotransmitter, and sensory stimuli (1). Receptors activate G proteins by stimulating the release of GDP from the ␣ subunit, allowing GTP to bind and to induce dissociation of the G protein ␣ and ␤␥ subunits, which interact with effector proteins to modulate cellular responses (2)(3)(4)(5).
The duration and strength of receptor-generated physiological responses are regulated by the rate at which GTP is hydrolyzed by ␣ subunit (6,7). It has been known for some time that the physiological turn-off of some G protein-mediated signals is faster than would be predicted from the in vitro GTPase activity of isolated G protein subunits (8,9). The solution to this paradox appears to reside in the newly recognized family of regulator of G protein signaling (RGS) 1 proteins, first identified genetically in the yeast Saccharomyces cerevisiae and in the nematode, Caenorhabditis elegans (10 -13). At least 19 RGS protein cDNAs have been identified in mammalian tissues, all sharing a homologous carboxyl-terminal region of ϳ120 amino acid residues termed the RGS domain (13)(14)(15). Biochemical studies with ␣ i and ␣ q family of G proteins demonstrated that RGS4 and G␣-interacting protein (GAIP) act as GTPase accelerating proteins (GAPs) (16,17), which could account for inhibition of G protein-mediated responses (15). GAP activity of Sst2 for Gpa1 has also been recently demonstrated (18). The mechanism by which GTPase activity is enhanced by RGS appears to be the stabilization of the transition state conformation of G␣ for nucleotide hydrolysis (19,20). RGS4 also directly inhibits the interaction of the GTP␥S-bound ␣ q subunit with phospholipase C␤, presumably by binding to the effector region of activated ␣ q (16).
A mutant yeast G␣ subunit, Gpa1 sst , was recently identified in a screen for novel strains showing the "supersensitive to pheromone" (sst) phenotype. It has a single glycine to serine mutation and escapes from negative regulation by the RGS protein, Sst2 (21). Since many RGS proteins affect G i -family G proteins, and the crystal structure of the RGS4⅐␣ i1 complex was recently reported, we wanted to see if the corresponding G 3 S mutation in ␣ i1 and ␣ o would produce insensitivity to RGS. A major objective was to obtain a detailed biochemical and mechanistic analysis of this newly identified class of mutations.
We report that G ␣ i1 and ␣ o subunit G 3 S mutants are insensitive to GTPase activation by two different RGS proteins. Quantitative flow cytometry studies demonstrated that a Ͼ30 -100-fold reduction in affinity of RGS for the ␣ subunit transition state is the mechanism of the insensitivity. In future studies, these mutant ␣ subunits should be useful for evaluating the role of endogenous RGS proteins in the kinetics and function of G i family members. Given the existence of nearly 20 RGS proteins of which at least 5 act on G i -family proteins, it would be difficult to inactivate all of them to determine the physiological role of endogenous RGS proteins in G i signaling. Thus, a G i ␣ subunit insensitive to RGS proteins can be used to assess the combined role of all RGS proteins in G i function in vivo.

EXPERIMENTAL PROCEDURES
DNA Construction and Mutagenesis-The G184S mutation in the ␣ o sequence was introduced by the megaprimer polymerase chain reaction mutagenesis technique using mutagenic antisense primer 5Ј-GGTTTC-TACGATCGAAGTTGTTTTGAC-3Ј as described (22,23). The G183S mutation in ␣ i1 was constructed by overlap-extension polymerase chain reaction using sense and antisense mutagenic primers 5Ј-AGTGA-AAACGACGTCAATTGTGGAAACC-3Ј and 5Ј-GGTTTCCACAATTGA-CGTCGTTTTCACT-3Ј, respectively. The coding region of rat RGS4 was amplified by polymerase chain reaction from an Expressed Sequence Tag obtained from The Institute of Genomic Research and cloned into the pGEX-2T expression vector. A restriction fragment comprising nucleotides 913-1358 of the complete human RGS7 cDNA 2 was cloned into the GST expression vector pgGSTag2. The expressed protein contained the RGS domain (nucleotides 985-1341) but only a portion of the long amino terminus of RGS7.
Purification of His 6 ␣ Subunits and GST-RGS Proteins-All G␣ subunits in this paper were expressed in Escherichia coli and purified as amino-terminal His 6 constructs by a modification of the method of Lee et al. (24). The GST-RGS4 and -RGS7 fusion proteins were purified as described (25). The bacterial supernatant was incubated with glutathione-agarose beads (Amersham Pharmacia Biotech) at 4°C overnight. After washing with phosphate-buffered saline, the GST-RGS4 was eluted with 10 mM glutathione in phosphate-buffered saline and dialyzed against 50 mM Tris and 1 mM EDTA, pH 7.4. The fusion proteins were cleaved by incubation overnight at 4°C with 10 units of thrombin/mg of fusion protein followed by incubation with glutathione-agarose to remove GST and any uncleaved GST-RGS4.
GAP Assays-[␥-32 P]GTP (1 M) was allowed to bind to 50 nM ␣ o for 20 min at room temperature (23-24°C) or 50 nM ␣ i1 for 15 min at 30°C. After lowering the temperature to 4°C, the hydrolysis reaction was started by the addition of MgSO 4 and GTP␥S to final concentrations of 15 mM and 200 M in the presence or absence of 100 nM RGS4 or RGS7. Aliquots (50 l) were diluted in 1 ml of 15% (w/v) charcoal solution (50 mM NaH 2 PO 4 , pH 2.3, 0°C) at the indicated time points. The amount of [␥-32 P]P i released at each time point was fit to an exponential function, Binding of ␣ to GST-RGS4-agarose-Wild type or mutant GDPbound ␣ (1 M) was mixed with GST-RGS4 fusion protein (1 M) bound to glutathione-agarose beads (5 ϫ 10 5 beads/ml) in a final volume of 100 l of HEDML buffer (50 mM Hepes, pH 8.0, 1 mM EDTA, 1 mM dithiothreitol, 10 mM MgSO 4 , 20 ppm deionized Lubrol) with 0.05% bovine serum albumin in the presence or absence of 20 M AlCl 3 and 10 mM NaF (HEDML/AF). After 2 h, the beads were washed three times with 1 ml of ice-cold HEDML/AF buffer with bovine serum albumin, and bound products were separated by SDS-polyacrylamide gel electrophoresis and visualized by Coomassie Blue stain.
Competition Binding of the ␣ Subunit to GST-RGS4 -Labeling of purified His 6 ␣ o with fluorescein isothiocyanate (FITC) was conducted as described (26). Glutathione-agarose beads (Amersham) with sizes between 35 and 78 m were prepared for flow cytometry by filtering through stainless steel Tyler sieves in HEDML buffer. Two nM GST or GST-RGS4 fusion protein was bound to the beads (5 ϫ 10 4 /ml) for 20 min at room temperature in HEDML buffer. Beads were then washed and incubated with 4 nM FITC-␣ o with the indicated amounts of unlabeled ␣ subunit in 100 l of HEDML/AF buffer at room temperature for 2.5 h. The amount of FITC-␣ o bound to the GST-RGS4 was quantitated on a Becton Dickinson FACScan as described (26). Data are presented as a fraction of the control fluorescence after subtracting nonspecific binding (ϳ20% of the total). Results were fit to a one-site competition binding function, and the IC 50 for the competitors were calculated using Prism version 2.01 for Windows 95 (Graphpad Software, San Diego, CA).

RESULTS AND DISCUSSION
Nucleotide Binding to and RGS-stimulated Hydrolysis by ␣ Subunit-The time course of nucleotide binding to ␣ subunits was determined by use of the fluorescent nucleotide derivative, methylanthraniloyl GTP␥S as described (27). Similar rates of binding were seen for both the G 3 S mutants and the wild type proteins. Rate constants were 0.20 Ϯ 0.06 and 0.21 Ϯ 0.08 min Ϫ1 for ␣ o and 0.049 Ϯ 0.002 and 0.12 Ϯ 0.03 min Ϫ1 for ␣ i1 , wild type and mutant, respectively. The k cat values for wild type and G 3 S mutant proteins in the absence of RGS were very similar ( Fig. 1 and Table I). In the presence of 100 nM RGS4, the reaction was completed by the first time point for wild type ␣ o and ␣ i1 (Fig. 1). Even at 4°C, the reaction was too fast to measure with a rate constant greater than 5 min Ϫ1 .
There was no effect of RGS4 on the k cat of either G 3 S mutant (Fig. 1, B and D, and Table I). The RGS domain fragment of RGS7 (100 nM) increased the k cat of wild type ␣ i1 by ϳ9-fold, whereas there was no effect on the ␣ i1 G 3 S (Fig. 1C and Table I).
GAP Activity at Increasing Concentrations of RGS4 -To determine if GAP activity could be restored at higher concentrations of RGS4, GTP hydrolysis at 1 min was measured with 50 nM G␣ subunit and 0 -3 M RGS4. The EC 50 values for RGS4 were 5 Ϯ 1 and 10 Ϯ 4 nM for wild type ␣ o and ␣ i1 , respectively (Fig. 2). There was only a slight increase in GTP hydrolyzed by either G 3 S mutant ␣ subunit, even at the maximum concentration of RGS4 (Fig. 2).
Reduced GAP Activity Is Due to Reduced Affinity-To determine whether the marked reduction in sensitivity of mutant ␣ to RGS4 was due to decreased binding, we tested co-precipitation of ␣ and RGS4. In the absence of AlF 4 Ϫ , GDP-bound ␣ subunit did not bind to GST-RGS4 (Fig. 3). In the presence of AlF4 Ϫ , wild type ␣ o and ␣ i1 showed substantial binding to GST-RGS4 immobilized on glutathione-agarose beads. 3 In con-2 A. Moritz and R. Taussig, unpublished information. 3 The apparent excess of bound G␣ over GST-RGS4 protein was due to incomplete elution of the GST-RGS4 from the beads in this experiment. P]GTP hydrolysis was measured before the reaction (ϳ30% of total) and was subtracted from the data. GTP hydrolysis at the indicated time points was calculated as a fraction of the total GTP hydrolyzed measured at 30 min. Experiments have been replicated at least three times. The data were fitted to a single exponential association function (Graphpad Prism), and the results are listed in Table I. trast, there was no detectable binding of either of the G 3 S mutant ␣ subunits. To more quantitatively characterize the interaction of ␣ subunits with RGS4, we used a recently developed flow cytometry approach (26). Binding of FITC-labeled ␣ o to GST-RGS4 on glutathione-agarose beads was detected by measuring the amount of fluorescence associated with the beads in a FACScan flow cytometer (see "Experimental Procedures"). Unlabeled wild type and mutant ␣ subunits were added to compete for the binding of FITC-␣ o to RGS4. All binding was done in AlF 4 Ϫ -containing buffer with 4 nM FITC-His 6 ␣ o and 2 nM GST-RGS4. Nonspecific binding in the presence of a 50-fold excess of unlabeled ␣ o represented less than 20% of the total binding. Similarly, GTP␥S-bound FITC-␣ o did not show any specific binding (data not shown). Wild type ␣ o and ␣ i1 reduced the binding of FITC-␣ o , with IC 50 s of 30 and 96 nM, respectively (and 95% confidence intervals of 18 -35 and 60 -150 nM; see Fig. 4). As observed for the effects of RGS4 on GTPase activity, we were unable to demonstrate significant interactions of the ␣ i1 G 3 S mutant protein with RGS4 up to 3 M ␣ subunit. Because of the limited purity and amounts of ␣ o G 3 S, the highest concentration used was 300 nM. At this concentration, there was an ϳ20% decrease in bound FITC-␣ o . Thus the affinity for both G 3 S mutant ␣ subunits is at least 30 -100-fold lower than that of wild type.

Generality of G 3 S Mutation in G␣-RGS Interactions-
In this report, we identify a G 3 S point mutation in the conformationally flexible "switch I" region of ␣ o and ␣ i1 that completely eliminates the interaction of these ␣ subunit with RGS proteins. This mutation should be useful in determining the role of endogenous RGS proteins in the physiological functioning of G proteins. This is especially important for G i , since the kinetics of G i -mediated regulation of effectors (e.g. adenylyl cyclase (28) and potassium channels (9)) are faster than expected for the in vitro GTP hydrolysis rates in the absence of RGS proteins. Deactivation of potassium currents was recently shown to be accelerated by overexpressed RGS1, 3, or 4 in oocytes and Chinese hamster ovary cells (29,30). These data show that exogenous RGS proteins can alter G protein function, but the question of whether normal cellular concentrations of RGS proteins alter the kinetics of ion channel function has not been answered experimentally. With the ␣ subunit mutation described here, it is now possible to directly address that question.
This class of G 3 S mutations was first described in yeast by DiBello et al. (21). They identified, in a genetic screen for the supersensitive phenotype, a G 3 S mutation in the yeast ␣ subunit (Gpa1 sst ) that resulted in insensitivity to 1) the functional effects of the yeast RGS protein, Sst2, and 2) the biochemical effects of the G␣-interacting protein. The loss of function due to these G 3 S mutations is selective, since our mutant ␣ subunits retain nearly normal intrinsic GTPase activity and kinetics of GDP release. The corresponding mutation in ␣ q prevented the RGS7-mediated reduction of phospholipase C activation in co-transfection studies (21). These latter data also demonstrate that the G 3 S mutation in ␣ q does not disrupt effector coupling. In preliminary data, 4 a myristolylated mutant G 3 S ␣ i1 inhibited forskolin stimulated type IV adenylyl cyclase activity. Thus, mutating this glycine residue has profound and consistent effects on four different ␣ subunits and their interactions with four different RGS proteins.
Structural Basis of Glycine to Serine Effects-In the crystal 4 M. Nanamori, K-L.. Lan, and R. R. Neubig, unpublished information. Ϫ at room temperature for 2.5 h. Data are presented as a fraction of control with nonspecific binding (ϳ20% of total) subtracted. Experiments have been replicated twice in duplicate. The data for wild type (WT) proteins were fit a to one-site competition function. structure of the ␣ i1 ⅐RGS4 complex (19), the switch I region of ␣ interacts with three of the four different segments of the RGS consensus domain, but there are also contacts with switch II and switch III. Glycine 183 is located in the switch I region of ␣ i1 , forming a turn just before the ␤1 strand (31,33). Interestingly, Natochin and Artemyev (32) recently found that the mutation of Ser-202 in switch II of transducin prevents interaction with retinal specific RGS. Glycine 183 provides a substantial contribution to the buried surface area between ␣ i1 and RGS4 (see Ref. 21 for details). There is direct contact of glycine 183 with the C␣ and C␤ of serine 85 and C␣ of tyrosine 84 of the RGS protein. Introduction of the hydroxymethyl side chain of serine would sterically hinder the formation of a tight complex of ␣ and RGS. Also, threonine 182, which is directly adjacent to glycine 183 mutated in the ␣ i1 G 3 S mutant, exhibits the greatest change in accessibility of any ␣ residue upon forming the ␣ i1 ⅐RGS4 complex (19). Thus the G 3 S mutation may also disrupt local protein conformation and prevent threonine 182 from interacting with the highly conserved residues in RGS.
The glycine at position 183 in ␣ i1 is highly conserved among all ␣ subunits. There are only two exceptions; in ␣7 of Dictyostelium discoidum and in open reading frame B0207.3 of C. elegans it is replaced by a serine (34,35). Interestingly, this is the same residue found in the Gpa1 sst , suggesting that these proteins may be naturally occurring sst variants that are insensitive to modulation by RGS proteins. The contact site in RGS is similarly conserved. Either serine or cysteine is present at the position equivalent to 85 in RGS4 where the G␣ glycine interacts. Druey and Kehrl (36) recently showed that modification of asparagine 88 and leucine 159 in RGS4 disrupted ␣ subunit binding and GAP activity. In the three-dimensional structure, both residues are very close to serine 85 where glycine 183 in G␣ makes contact (19).
Binding of ␣ Subunits to GST-RGS4 -Both in co-precipitation and fluorescence competition studies, the affinity of the AlF 4 --bound mutant ␣ subunits for RGS4 are dramatically reduced. With the flow cytometry method, we obtained quantitative measures of subunit affinities. The IC 50 values of wild type ␣ o and ␣ i1 (30 and 96 nM, respectively) are similar to the K D of 45 nM determined by surface plasmon resonance for transducin binding to retinal-specific RGS (37). Our IC 50 is significantly higher than the K D estimated from on and off rates by plasmon resonance for RGS4 and ␣ i1 (38). The differences between the latter data and ours may be due to methodological differences (kinetic versus equilibrium determination or direct binding versus competition methods). In any case, the 30 -100-fold lower affinity of the G 3 S mutant G ␣ proteins is very clear.
Function of RGS7-Our data also include the first biochemical demonstration of GAP activity by RGS7. At 100 nM, the effect of the RGS7 RGS domain on GTP hydrolysis by ␣ i1 is significantly less than the effect of RGS4. Interestingly, the FITC-␣ o did not show detectable binding to the GST-RGS7 fusion protein (data not shown). These results are probably due to a lower affinity of RGS7 for ␣ o compared with RGS4. This may be due in part to the lack of full-length RGS7 in the expression construct.
In summary, our data show a dramatic disruption in both RGS binding and RGS-mediated GAP activity when the glycine in the switch 1 region of ␣ i1 or ␣ o is mutated to serine. The effect of the G 3 S mutation is both specific in that it only disrupts RGS binding and quite general in the range of ␣ subunit and RGS proteins that it effects. The introduction of this mutation into G protein ␣ subunits can be used in conjunction with expression or transgenic animal studies to evaluate the physiological role of endogenous RGS proteins in the function of a given G protein.