Rapid Activation of Transducin by Mutations Distant from the Nucleotide-binding Site EVIDENCE FOR A MECHANISTIC MODEL OF RECEPTOR-CATALYZED NUCLEOTIDE EXCHANGE BY G PROTEINS*

G proteins act as molecular switches in which information flow depends on whether the bound nucleotide is GDP (“off”) or GTP (“on”). We studied the basal and receptor-catalyzed nucleotide exchange rates of site-di-rected mutants of the a subunit of transducin. We iden-tified three amino acid residues (Thr-325, Val-328, and Phe-332) in which mutation resulted in dramatic increases (up to 165-fold) in basal nucleotide exchange rates in addition to enhanced receptor-catalyzed nucleotide exchange rates. These three residues are located on the inward facing surface of the a 5 helix, which lies between the carboxyl-terminal tail and a loop contact-ing the nucleotide-binding pocket. Mutation of amino acid residues on the outward facing surface of the same a 5 helix caused a decrease in receptor-catalyzed nucleotide exchange. We propose that the a 5 helix comprises a functional microdomain in G proteins that affects basal nucleotide release rates and mediates receptor-cat-alyzed nucleotide exchange at a distance from the nu-cleotide-binding pocket. ) is Biolabs). proteolytic the

G proteins act as molecular switches in which information flow depends on whether the bound nucleotide is GDP ("off") or GTP ("on"). We studied the basal and receptor-catalyzed nucleotide exchange rates of site-directed mutants of the ␣ subunit of transducin. We identified three amino acid residues (Thr-325, Val-328, and Phe-332) in which mutation resulted in dramatic increases (up to 165-fold) in basal nucleotide exchange rates in addition to enhanced receptor-catalyzed nucleotide exchange rates. These three residues are located on the inward facing surface of the ␣5 helix, which lies between the carboxyl-terminal tail and a loop contacting the nucleotide-binding pocket. Mutation of amino acid residues on the outward facing surface of the same ␣5 helix caused a decrease in receptor-catalyzed nucleotide exchange. We propose that the ␣5 helix comprises a functional microdomain in G proteins that affects basal nucleotide release rates and mediates receptor-catalyzed nucleotide exchange at a distance from the nucleotide-binding pocket.
G protein 1 -coupled receptors (GPCRs) activate G proteins by catalyzing the release of GDP bound to the G protein ␣ subunit (G␣) (1)(2)(3). GTP-bound G␣ proteins regulate a variety of cellular effector enzymes or ion channels. The active G␣ is turned off kinetically by intrinsic GTPase activity, which is highly regulated by bound effector and regulator of G protein signaling proteins (4,5). High resolution structural analysis of several nucleotide-bound forms of G␣ (6 -9) and of complexes of G␣ subunits with regulatory proteins (10,11) have provided insights into key conformational changes required for G protein function. In addition, the crystal structure of rhodopsin (12), a prototypical GPCR, provides a basis for understanding several key features of the molecular physiology of rhodopsin (13). However, structural information is lacking for the receptorbound nucleotide-free conformation of G␣, and the molecular details of how GPCRs switch G proteins from their inactive to active states remains unknown. Interestingly, the receptor must act at a distance because extensive evidence from biochemical and mutagenesis work suggests that the cytoplasmic loops of the active receptor (R*) do not directly contact the nucleotide-binding pocket of G␣ subunits (14).
To address the question of how GPCRs catalyze the switching of GDP for GTP by G proteins, we studied the mechanism of nucleotide exchange in transducin (G␣ t ). We chose to study G␣ t because it exhibits an extremely low basal GDP release rate (15) and a very rapid R*-catalyzed nucleotide exchange rate (16) consistent with its role in rod cell visual phototransduction where low biochemical noise levels are required. We tested the hypothesis that the ␣5 helix of G␣ t is one of the structural features that resolves the action at a distance problem and explains how R* induces nucleotide switching by G proteins (17,18). The ␣5 helix of G␣ t (amino acid residues 325-339) connects the carboxyl-terminal tail of G␣ t to the ␤6/␣5 loop. The carboxyl tail of G␣ t , especially amino acid residues 340 -350, is a well characterized binding domain for R* (19 -23), and the ␤6/␣5 loop directly contacts the guanine ring of the bound nucleotide. Consequently, R* might induce nucleotide exchange in G␣ t by using the ␣5 helix to communicate between the carboxyl terminus and the ␤6/␣5 loop.
We report that mutation of several specific residues on the buried surface of ␣5 caused dramatic increases in the basal nucleotide exchange rates in the resulting G␣ t mutants. The evidence suggests that the mechanism of R*-catalyzed nucleotide exchange by G␣ t involves conformational changes in these ␣5 residues. R* may induce changes in ␣5 either directly by binding to ␣5 or indirectly by binding to the adjacent carboxyl terminus. We propose that the ␣5 helix controls the basal nucleotide exchange rate and mechanistically couples R* binding to the guanine nucleotide in G␣ t .

EXPERIMENTAL PROCEDURES
Most of the methods used in this paper have been described previously (24). Brief descriptions are given below.
Site-directed Mutagenesis of G␣ t and Expression in Vitro-Site-directed mutations were created using the QuickChange system (Stratagene). The parent for all G␣ t constructs was pGEM2sT␣, the synthetic bovine G␣ t gene cloned into the pGEM2 plasmid under the control of a SP6 promoter (25). G␣ t and G␣ t mutant proteins were prepared in vitro using the TNT Quick Coupled transcription/translation system (Promega). The translated products were passed over BioSpin 6 gel filtration spin columns (Bio-Rad) twice consecutively to remove excess nucleotides and [ 35 S]methionine. The volume of each sample was then adjusted to 100 l in Buffer A (5 mM Tris-HCl, pH 7.5, 150 mM NaCl, 2 mM MgCl 2 , 1 mM dithiothreitol, 0.01% (w/v) n-dodecyl-␤-D-maltoside). If the sample was to be studied in a R*-catalyzed assay, G␤␥ t was added to a final concentration of 30 nM. Every experiment was performed using freshly translated material. * This research was supported in part by the Allene Reuss Memorial Trust and National Institutes of Health Training Grants GM07739 and GM07982. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
Nucleotide Exchange Rate Assays-Samples (70 l each) of translated G␣ t or mutant G␣ t kept on ice in Buffer A were quickly warmed to room temperature in a water bath. For basal exchange rate assays ( Fig.  1), the experiment was initiated by the addition of GTP␥S to a final concentration of 100 M. Five aliquots (8 l) were withdrawn during a period ranging from 2 to 6 h and digested. For R*-catalyzed assays (Fig.  2), the experiment was initiated by the addition of a mixture of R* and GTP␥S (4 l) to a final concentration of 30 nM R* and 14 M GTP␥S. Immediately before addition to the reaction, the rhodopsin was photolyzed by illumination for 15 s with a fiber optic cable connected to a Dolan Jenner lamp equipped with a Ͼ495-nm long pass filter. The samples were incubated at room temperature under illumination. Aliquots (8 l) were withdrawn and digested at 1, 2, 3, 5, 10, and 20 min. The digestion procedure was adapted from Garcia et al. (20). Aliquots were mixed with 1.5 l of digest buffer (5% Lubrol, 2 mM GDP, 1 mg/ml L-1-tosylamido-2-phenylethyl chloromethyl ketone-treated trypsin) and incubated on ice for 30 min. Digestion was terminated by the addition of 2.5 l of termination solution (10 mg/ml aprotinin, 10 M phenylmethylsulfonyl fluoride), followed by 6 l of 3ϫ SDS sample buffer (New England Biolabs). The proteolytic fragments were resolved by SDSpolyacrylamide gel electrophoresis using precast 15% gels (Bio-Rad). The intensities of the fragments were quantitated by phosphorimaging using a Storm Imager and ImageQuant software (Molecular Dynamics). G␣ t in the GDP-bound inactive conformation yielded an ϳ23-kDa fragment, and G␣ t in the active conformation (GTP␥S-or GDP-AlF 4 Ϫ -bound) yielded an ϳ34-kDa fragment following trypsin digestion. The fraction of G␣ t activated in a given sample was determined from the relative intensity of the two bands. The ϳ23-kDa bands were generally absent from digests of the F185A and F192A mutants of G␣ t (not shown), although they could bind GDP as indicated by activation by AlF 4 Ϫ and the rate of GTP uptake. Instead, the mutations seemed to affect the folding or stability of the proteins. The rates of nucleotide exchange in F185A and F192A could be estimated from the intensities of the ϳ34-kDa bands.

RESULTS AND DISCUSSION
We carried out alanine-scanning mutagenesis of the 14 nonalanine amino acid residues in the ␣5 helix of G␣ t . Proteins were expressed in vitro in rabbit reticulocyte lysate to overcome the well documented problems with the heterologous expression of G␣ t (26,27). Nucleotide exchange rates were determined using a quantitative trypsin digest assay specifically adapted for kinetic measurements. The assay is directly sensitive to the nucleotide-bound status of G␣ t (28) and does not rely on potentially confounding second messenger readouts. The sensitivity and fidelity of this assay has been described in detail (24).
Three of the fourteen mutants (T325A, V328A, and F332A) exhibited tremendous increases in their rates of basal nucleotide exchange (Fig. 1, A and B; Table I). The basal exchange rates of V328A, F332A, and T325A were accelerated by 77-, 151-, and 165-fold, respectively, over the basal exchange rate of wild-type G␣ t measured under the identical conditions. These rates are comparable with or in excess of the basal exchange rate of mutant A322S (63-fold increase) (Fig. 1, A and B; Table  I). Ala-322 is known to reside in the ␤6/␣5 loop and to be in contact with GDP. The A322S mutant is also analogous to mutations in G␣ s and G␣ i that have been reported to greatly increase the basal nucleotide exchange rates (29,30). These results suggest that alteration of the ␣5 helix, which is distant from the nucleotide, can accelerate GDP release to an even greater degree than replacement of amino acids known to be in direct contact with the nucleotide.
The mutants N327A and I339A also exhibited statistically significant gain of function phenotypes (Table I). Together, the five residues that displayed an increase in basal nucleotide exchange rate are found to be clustered along the buried surface of the ␣5 helix (Fig. 3). Thr-325 is located in the first turn of the ␣5 helix, immediately adjacent to the ␤6/␣5 loop. Asn-327 is also located on the first turn, a structure that appears to be stabilized by a network of hydrogen bonds. Substitution of the amino acid side chains in this region might be expected to disrupt some of these hydrogen bonds and alter the structure of the first turn and possibly the ␤6/␣5 loop. Two other residues, Val-328 and Phe-332, point in toward the center of G␣ t and contribute to its hydrophobic core. Val-328 and Phe-332 make extensive contacts with residues from the ␣1 helix, as well as the ␤2 and ␤3 sheets. Loops emanating from these structural elements contribute to the canonical nucleotide-binding surfaces of G␣ t and other G proteins (31). Replacement of Val-328 or Phe-332 with alanine is likely to perturb the packing of the ␤2 and ␤3 sheets and the ␣1 helix in a manner that affects the structure of the nucleotide-binding pocket.
All of the alanine replacement mutants, including V328A, FIG. 1. Basal nucleotide exchange rates of G␣ t and G␣ t mutants determined by quantitative trypsin digests. A, trypsin digest reactions of G␣ t and G␣ t mutants that display accelerated basal nucleotide exchange rates. G␣ t was expressed in vitro in the presence of [ 35 S]methionine and digested with trypsin following treatment with 100 M GDP, 100 M GTP␥S, or 100 M GDP,170 M AlCl 3 ,10 mM NaF. The resulting fragments were resolved by SDS-polyacrylamide gel electrophoresis and visualized by phosphorimaging. Arrows indicate the position of ϳ34and ϳ23-kDa bands. The fraction of G␣ t bound to GTP␥S or GDP was determined from the relative intensity of the ϳ34-kDa band (which corresponds to activated G␣ t ) to the ϳ23-kDa band (which corresponds to inactive G␣ t ). The basal rate of nucleotide exchange was determined by examining aliquots taken at indicated times following addition of GTP␥S (100 M). All the mutants folded properly and could bind GDP, GTP␥S, and AlF 4 Ϫ as indicated by the trypsin digest pattern. B, basal nucleotide exchange time course derived from trypsin digest analysis. The mutants T325A, V328A, and F332A all displayed extremely fast nucleotide exchange kinetics that exceeded the rate of A322S, a mutation of a nucleotide-contacting residue. Data points represent the mean of at least three independent experiments; error bars depict Ϯ2 ϫ S.E. Solid lines are fits to the equation: y ϭ c ϩ 100(1 Ϫ exp(Ϫk app t)).
F332A, and T325A, were capable of binding GDP and GTP␥S and could be activated by GDP/AlF 4 Ϫ (Fig. 1A and data not shown). In addition, the typical band patterns of trypsin digestion were observed with all of the mutants (Fig. 1A and data not shown). These results indicate that the mutations did not disrupt the overall folding of the mutant proteins.
The equivalents of residues Val-328 and Phe-332 are highly conserved not only in G␣ subunits, but also among small monomeric G proteins of the Ras superfamily (32). When the equivalent of Phe-332 was replaced with leucine in Ras, the F156L mutant was found to have transforming properties in vivo and to accelerate greatly the rate of nucleotide exchange in vitro (33). NMR structural studies of Ras F156L revealed alterations in secondary structures ␣1, ␣5, and ␤1-␤3 near position 156. The mutation did not appear to reduce the stability of the protein, perhaps due to induction of new intramolecular contacts not present in the wild-type protein. This observation suggests that G␣ t F332A might exhibit higher stability than wild-type G␣ t in the absence of bound nucleotide.
Because the mutagenesis studies indicate that perturbation of residues in the ␣5 helix in G␣ t causes a dramatic increase in the nucleotide exchange rate, one would predict the existence of intramolecular contacts that maintain the ␣5 helix in a specific rigid conformation to avoid inappropriate release of GDP. In fact, mutation of residues in the protein core that interact with the ␣5 helix also resulted in moderate increases in the basal nucleotide exchange rate (Fig. 3C). In particular, introduction of alanine at positions Gln-168, Phe-185, Lys-188, or Phe-192 led to increases in nucleotide exchange rate (Table I and data  not shown). Gln-168, which is located on ␣F of the helical domain, hydrogen bonds with the main chain carbonyl of Thr-323. Lys-188, which extends from a loop between ␤2 and ␤3, appears to form ionic interactions with two aspartic acid residues on the ␣5 helix, Asp-333 and Asp-337. These interactions likely serve to maintain the proper register of the ␣5 helix with respect to the rest of the protein. Phe-185 and Phe-192, which are located on ␤2 and ␤3, respectively, are part of a phenylala-nine cluster that lies adjacent to Phe-332. These contacts may serve to stabilize the local tertiary structure near the ␣5 helix, to sense alterations in the position of Phe-332, and to communicate structural changes onto the nucleotide-binding pocket. This latter role is suggested by the structural changes observed in ␤2 and ␤3 in NMR studies of the F156L mutant in Ras (33).
It has been proposed previously that interactions between the carboxyl-terminal region of the ␣5 helix and the rest of the protein were involved in controlling the rate of basal (and possibly R*-catalyzed) nucleotide exchange (34 -36). We suggest that these contacts, like those discussed above, stabilize the position of residues in the amino terminus of ␣5 relative to the rest of G␣ t . Weakening these contacts, as in the I339A mutant, moderately accelerated nucleotide exchange (Table I), in agreement with previous results (36). However, the magnitude of increases in GDP release rates resulting from mutations of residues in the carboxyl-terminal region of ␣5 were very small (ϳ4-fold) compared with the effect of mutations closer to the amino terminus of ␣5.
The gain of function phenotypes of the T325A, V328A, and F332A mutants appear to mimic the effect of R* binding on G␣ t in several key respects: (i) the basal rate of nucleotide exchange  a The apparent rate constants for basal nucleotide exchange were derived from fits of each data set to the exponential rise equation, y ϭ c ϩ 100(1 Ϫ exp(Ϫk app t)). Each mutant protein was assayed at least 3 times (wild-type G␣ t protein was assayed 26 times), and an independent fit was made to each data set. The values reported are the mean k app ϫ 10 4 min Ϫ1 Ϯ 2 ϫ S.E.
b The -fold increase in the nucleotide exchange rate measured for the mutant protein relative to that of uncatalyzed wild-type G␣ t was calculated as the k app (mutant)/k app (wild-type).
is tremendously increased, (ii) the primary perturbation is at a distance from the nucleotide-binding site, and (iii) the nucleotide binding properties of the protein are preserved. In principle, perturbation of these residues by R* would be sufficient to catalyze nucleotide exchange.
To explore whether R* actually does induce conformational changes in this region, we measured the R*-catalyzed nucleotide exchange rates in each of the mutants of the ␣5 helix (Fig.  2). The G␣ t mutants that showed the most drastic acceleration of basal nucleotide exchange rates (T325A, V328A, and F332A) could be further stimulated by R* and G␤␥ t (compare Fig. 2A with Fig. 1B). These mutants displayed faster R*-catalyzed rates than that of wild-type G␣ t ( Fig. 2A), as did N327A and I339A (not shown). These data demonstrate that the T325A, V328A, and F332A mutants could interact with R*. Furthermore, the conformations of the mutant proteins appear to be consistent with or possibly similar to the conformation of R*bound G␣ t .
Five mutations caused little or no change in the R*-catalyzed activation rate (Fig. 2B). These mutations, Q326A, K329A, D333A, V335A, and T336A, also did not affect the basal nucleotide exchange rate (Table I). Four of these mutations map to the same surface of ␣5 (Fig. 3D); this surface does not appear to be involved in controlling nucleotide exchange rates in G␣ t .
Four mutants (F330A, V331A, D337A, and I338A) displayed rates of R*-catalyzed activation that were decreased ϳ3-4-fold relative to the wild-type rate (Fig. 2C). Since the mutations did not completely abolish activation by R*, the functions of these residues may be partially redundant. The overall structures of these mutant proteins were not significantly disturbed as indicated by trypsin proteolysis patterns and basal nucleotide exchange rates that were similar to wild-type G␣ t (Table I and not shown). The mutations were unlikely to affect interactions with G␤␥ t , which is known to bind to the opposite face of G␣ t from ␣5 (37). The most direct interpretation of these loss of function results is that Phe-330, Val-331, Asp-337, and Ile-338 make important contacts in the R*-G␣ t complex and that these contacts are disrupted by the substitutions with alanine. In the absence of R*, these residues do not appear to be involved in any contacts that are relevant to nucleotide exchange rates, as FIG. 3. The structure of the ␣5 helix region of G␣ t and the effects of site-directed mutations. A-C were prepared with Molscript (43) and Raster3D (44). A, structure of the GDP-bound heterotrimeric form of a G␣ t /G␣ i chimera (PDB entry 1got) (37). The ␣5 helix (yellow) connects the ␤6/␣5 loop (orange) with an R*-binding domain, the carboxyl-terminal region (amino acids 340 -350) (green). The nucleotide is purple. The coordinates for the 340 -350 region, which are poorly resolved in the crystal structure, were taken from a NMR study of the corresponding peptide in the presence of R* (45) and are not shown in the other panels of this figure. B, the molecule has been rotated 90°about a vertical axis relative to panel A. The ␣5 helix is ϳ50% solvent-exposed and lies in a trough formed by ␣1 and ␤1-␤6. C, close-up view of the ␣5 region. The residues mutated in this study are shown in ball-and-stick representation. Within the ␣5 helix, residues that led to increases in both basal and R*-catalyzed nucleotide exchange rates when substituted with alanine are colored green, residues that led to decreases in R*-catalyzed exchange rate when substituted with alanine are colored red, and residues that were indistinguishable from wild-type when substituted with alanine are yellow. D, helical wheel representation of the ␣5 helix. The coloring is the same as in panel C. The sites of mutations that caused increases in both basal and R*-catalyzed nucleotide exchange map predominantly to a buried surface of the helix. The sites of mutations that caused impairment of R*-catalyzed nucleotide exchange map to a solvent-exposed hydrophobic patch (including Val-331, Ile-338, and Phe-330) adjacent to the buried surface of ␣5. Hydrophilic residues that map to the "bottom" of the helix did not affect either basal or R*-catalyzed nucleotide exchange when replaced with alanine. Together, the data provide evidence that R* catalyzes nucleotide exchange by perturbing the conformation of residues on the buried surface of ␣5 (green line). These residues communicate with the nucleotide-binding pocket. The perturbation may result from direct binding of R* to the exposed hydrophobic surface of ␣5 (red line).
indicated by the crystal structure and the basal nucleotide exchange rates of the corresponding mutants (Fig. 3C, Table I).
In particular, Phe-330 is more than 4 Å from any other residue. Thus, the formation of important contacts involving Phe-330, Val-331, Asp-337, and Ile-338 in the R*-bound conformation necessitates a R*-induced conformational change in ␣5. Consistent results were obtained in site-directed spin-labeling studies with G␣ i , which demonstrated that R*-induced conformational changes occur in a similar set of residues at the amino terminus of ␣5 corresponding to Gln-326, Asn-327, and Phe-330 of G␣ t . 2 Since R* appears to induce conformational changes at Val-331 and Phe-330 and other residues of ␣5, it is plausible that R* also alters the Thr-325/Val-328/Phe-332 region on the adjacent face of the ␣5 helix (Fig. 3, C and D). This is evidence that R* does in fact catalyze nucleotide exchange at a distance by perturbing the Thr-325/Val-328/Phe-332 region and by exploiting the built-in structural connection between these residues and the nucleotide-binding pocket of G␣ t . Subsequent alteration of the ␤6/␣5 loop as previously hypothesized could also occur (18).
The proposed conformational changes of Thr-325, Val-328, and Phe-332 induced by binding of R* need be only subtle. Because mutation of each residue by itself led to a dramatic increase in nucleotide exchange rate, even a minor structural perturbation of more than one residue simultaneously by R* would be expected to produce a potent effect on the nucleotide exchange rate. As a precedent, the 2.6-Å resolution crystal structure of the G␣ i A326S mutant heterotrimer (analogous to A322S in G␣ t ) did not reveal any significant structural alterations although the nucleotide exchange rate was dramatically increased (30).
Some or all of the contacts involving Phe-330, Val-331, Asp-337, and Ile-338 that are inferred to exist in the R*-bound conformation may result from direct binding of R* to ␣5. The amino acid residues Phe-330, Val-331, Asp-337, and Ile-338 map to a conspicuous solvent-exposed, yet predominantly hydrophobic surface of the ␣5 helix (Fig. 3). This surface is contiguous with the carboxyl-terminal region (amino acids 340 -350), a well documented R*-binding domain. Thus, the exposed surface of ␣5 may be part of the R*-binding region on G␣ t , as has been previously suggested (17,18).
The data are also consistent with the possibility that R* indirectly induces the formation of intramolecular contacts involving Phe-330, Val-331, Asp-337, and Ile-338 or a subset thereof. In this scenario, R* would induce changes in the amino terminus of ␣5 by binding to other sites on the protein, such as the carboxyl-terminal region. As a speculative example, R* might induce a counterclockwise rotation of ␣5 such that the exposed hydrophobic surface of ␣5 would become buried, and contacts among Phe-330, Val-331, Asp-337, and Ile-338 would form with the mostly hydrophobic residues in the core of G␣ t . Such a rotation would of course also displace Thr-325, Val-328, and Phe-332 on the inside surface of ␣5. Additional experiments are needed to conclusively determine whether the interactions between R* and ␣5 are direct or indirect.
A number of other proposed mechanisms of R*-catalyzed nucleotide exchange may also be at work. Several studies have suggested that activated receptors might induce opening of the cleft between the Ras-like and helical domains of G␣ subunits (8,38). However, our recent experiments suggest that interdomain interactions are not an energetic barrier to nucleotide exchange in G␣ t (24). Others have suggested that R* could use the G␤␥ subunits as a lever to distort the Switch I/II regions and facilitate nucleotide exchange (14,39). This type of mechanism is suggested by analogy to the structure of complexes of monomeric G proteins and their cognate guanine nucleotide exchange factors, in which the guanine nucleotide exchange factors bind to the Switch I and II regions and induce nucleotide dissociation in part by disrupting the Mg 2ϩ coordination site of the G protein (14, 40,41). However, lack of Mg 2ϩ in G protein heterotrimers indicates that the monomeric G protein exchange processes cannot be an exact model for the heterotrimeric G proteins (31,42). G␣ t is a highly specialized molecular switch that exhibits an extremely low rate of basal GDP release and a very rapid rate of R*-catalyzed GTP uptake compared with other G proteins. However, considering the conserved nature of the ␣5 helix and its orientation to the nucleotide-binding pocket, it is likely that amino acid residues corresponding to Thr-325, Val-328, and Phe-332 in G␣ t serve a primary role in regulating nucleotide exchange rate in other G␣ subunits.