Introduction

Top Abstract Introduction Experimental Procedures Results Discussion References

The ₂AR is a prototypical GPCR that couples to the G protein G_s to activate AC (Gilman, 1987; Kobilka, 1992). According to the two-state model of GPCR activation, GPCRs exist either in an inactive (R) state or an active (R*) state (Lefkowitz et al., 1993; Leff, 1995; Gether and Kobilka, 1998). In the R* state, GPCRs promote GDP dissociation from G, which is the rate-limiting step of the G protein cycle (Gilman, 1987). Agonists stabilize the R* state of GPCRs and increase GDP/GTP exchange at G. Isomerization of GPCR from R to R* can also occur agonist independently and is referred to as constitutive activity. The ₂AR exhibits constitutive activity (Chidiac et al., 1994; Gether et al., 1995). Constitutive activity of the ₂AR gives rise to agonist-independent GDP/GTP exchange at G_s and, consequently, AC activation. Agonist-independent GPCR activity is reduced by inverse agonists that stabilize the R state of GPCR. In this article, I use the term "apparent constitutive GPCR activity" to describe the agonist-independent GPCR activity observed experimentally without implying that a given substance, i.e., a salt, or a G protein possesses a direct effect on R to R* equilibrium.

Although even the purified ₂AR undergoes agonist-independent R to R* isomerization (Gether et al., 1995), the apparent constitutive activity of the ₂AR strongly depends on to which G_s splice variant the ₂AR is coupled (Seifert et al., 1998a,b). G_sL possesses a lower GDP affinity than G_sS, i.e., GDP dissociates from G_sL more readily than from G_sS (Graziano et al., 1989; Seifert et al., 1998b). Thus, the agonist-free ₂AR promotes GDP dissociation from G_sL more efficiently than GDP dissociation from G_sS. Experimentally, this difference between G_sS and G_sL results in strong agonist-independent AC activation by GTP and efficient reduction of this AC activity by the inverse agonist ICI for the ₂AR-G_sL pair, but not for the ₂AR-G_sS pair (Seifert et al., 1998b). In previous studies, it has been difficult to demonstrate differences in the coupling of the ₂AR to G_s splice variants (Graziano et al., 1989; Jones et al., 1990; O' Donnell et al., 1991). To reconcile these differences, I aimed at identifying experimental variables that differentially modulate coupling of the ₂AR to G_sS and G_sL. My research focused on monovalent cations and anions since previous studies had demonstrated that the inclusion of NaCl into reaction mixtures decreases the apparent constitutive activity of several G_i/G_o protein-coupled GPCRs (Costa and Herz, 1989; Gierschik et al., 1989; Hilf and Jakobs, 1992; Tian et al., 1994; Wenzel-Seifert et al., 1998).

Here, I report that monovalent anions have much more pronounced effects on coupling of the ₂AR to G_sL than on coupling of the ₂AR to G_sS. As model system, I used ₂AR-G_s fusion proteins that ensure defined GPCR/G stoichiometry and efficient GPCR/G protein coupling (Seifert et al., 1999b). Additionally, I expressed ₂AR-G_sS and ₂AR-G_sL at similar levels to obtain a defined fusion protein/AC stoichiometry. I propose a model in which anions increase the GDP affinity of G_sL more efficiently than the GDP affinity of G_sS. The increase in GDP affinity of G_sL decreases the efficiency of the agonist-free ₂AR and increases the efficiency of the agonist-occupied ₂AR at promoting GDP dissociation from G_sL.

	Experimental Procedures

Top Abstract Introduction Experimental Procedures Results Discussion References

Materials. The procedures for generation of baculoviruses encoding ₂AR-G_sS, ₂AR-G_sL, and nonfused G_sL were described previously (Seifert et al., 1998a,b). Sodium salts and chloride salts were of the highest purity available and were obtained from Sigma (St. Louis, MO). Sources of all other materials were as described (Seifert et al., 1998a,b).

Sf9 Cell Culture and Membrane Preparation. These procedures followed the published protocol (Seifert et al., 1998a,b).

DHA Binding. The expression of ₂AR-G_sS and ₂AR-G_sL was determined by saturation binding using the ₂AR antagonist DHA (10 nM) (Seifert et al., 1998a,b). DHA competition binding with antagonists, inverse agonists, and agonists was performed as described (Seifert et al., 1998a,b). The effect of GDP on high-affinity binding of the agonist salbutamol at a fixed concentration (1 µM) was studied as described (Seifert et al., 1998b). Binding reactions were carried out in 75 mM Tris/HCl, pH 7.4.

AC Activity. The determination of AC activity in membranes expressing ₂AR-G_sS and ₂AR-G_sL followed the published protocol (Seifert et al., 1998a,b). AC reactions were conducted in 30 mM Tris/HCl, pH 7.4. The relative inhibitory effects of ICI on AC activity were calculated as described (Seifert et al., 1999a; Wenzel-Seifert and Seifert, 2000).

Miscellaneous. Protein was determined using the Bio-Rad DC protein assay kit (Bio-Rad, Hercules, CA). Concentration-response curves shown in Figs. 1, 6, and 7 were analyzed by nonlinear regression, using the Prism III program (GraphPad, San Diego, CA). Data shown in Figs. 4 and 5B were analyzed by linear regression, using the Prism III program. Statistical analyses (i.e., effects of salts versus control, comparison of ₂AR-G_sS versus ₂AR-G_sL, and comparisons of the effects of various anions and cations against each other) were performed using the t test.

View larger version (33K): [in this window] [in a new window]   Fig. 1.   Effects of ISO and ICI on AC activity in Sf9 membranes expressing 2AR-GsL in the absence and presence of NaCl. A and B, GTP dependence of AC activity. Reaction mixtures contained Sf9 membranes (20 µg of protein per tube) expressing 2AR-GsL at 2.3 to 2.6 pmol/mg, GTP at the concentrations indicated on the abscissa, and solvent (basal), ISO (10 µM), or ICI (1 µM). A, experiments performed in the absence of NaCl (control). B, experiments performed in the presence of 150 mM NaCl. C and D, concentration-response curves for ISO and ICI on AC activity in the presence of GTP at a fixed concentration (1 µM). Reaction mixtures contained Sf9 membranes (20 µg of protein per tube) expressing 2AR-GsL at 2.3 to 2.6 pmol/mg and ISO or ICI at the concentrations indicated on the abscissa. C, experiments performed in the absence of NaCl (control). D, experiments performed in the presence of 150 mM NaCl. The dashed horizontal lines are extrapolations of basal AC activities to illustrate the relative contributions of ISO and ICI at the ligand-regulated AC activities (the difference between maximum ISO-stimulated and minimum ICI-inhibited AC activity).  ISO, difference between maximum ISO-stimulated AC activity and basal AC activity.  ICI, difference between basal AC activity and minimum ICI-inhibited AC activity. Data were analyzed by nonlinear regression and were best fit to sigmoidal concentration-response curves (F test). Data shown are the means ± S.D. of three independent experiments performed in duplicate.

	Results

Top Abstract Introduction Experimental Procedures Results Discussion References

Reduction by NaCl of Inverse Agonist Inhibition and Enhancement of Agonist Stimulation of AC in Membranes Expressing ₂AR-G_sL. In the absence of NaCl, GTP strongly increased basal AC activity in membranes expressing ₂AR-G_sL (Fig. 1A). Because of the large stimulatory effect of GTP on basal AC activity, the additional increase in AC activity by ISO was rather modest (~45-70% stimulation) (Fig. 1, A and C). The inverse agonist ICI abolished the increase in AC activity caused by GTP at 100 nM (Fig. 1A). With GTP at concentrations between 1 and 100 µM, the inhibitory effect of ICI on AC activity ( ICI) amounted to ~75% of that of the stimulatory effect of ISO ( ISO). NaCl at 150 mM reduced the maximum stimulatory effect of GTP basal AC activity in membranes expressing ₂AR-G_sL by ~70%, whereas the maximum ISO-stimulated AC activity was not decreased (Fig. 1, A and B). In addition, NaCl substantially reduced the inhibitory effect of ICI on basal AC activity. In the presence of NaCl, the ICI-inhibited AC activity ( ICI) amounted to only ~20 to 25% of that of the ISO-stimulated AC activity ( ISO) (Fig. 1D).

Differential Regulation of AC Activity by Sodium Salts in Membranes Expressing ₂AR-G_sS and ₂AR-G_sL. The effects of NaCl on AC activity in membranes expressing ₂AR-G_sS and ₂AR-G_sL at similar levels were compared. NaCl had virtually no effect on basal AC activity in membranes expressing ₂AR-G_sS and slightly reduced the absolute stimulatory effect of ISO on AC activity (Fig. 2A). In membranes expressing ₂AR-G_sS, the inhibitory effect of ICI on basal AC activity was very small and NaCl abolished the minimal inhibitory effect of ICI in this system. In marked contrast to membranes expressing ₂AR-G_sS, NaCl had a pronounced inhibitory effect on basal AC activity in membranes expressing ₂AR-G_sL and enhanced the absolute stimulatory effects of ISO and diminished the absolute inhibitory effects of ICI.

View larger version (33K): [in this window] [in a new window]   Fig. 2.   Effects of NaCl, NaBr, and NaI on AC activity in Sf9 membranes expressing 2AR-GsS or 2AR-GsL in the absence or presence of ISO or ICI. Reaction mixtures contained Sf9 membranes (20 µg of protein per tube) expressing 2AR-GsS (2.0-2.2 pmol/mg) (A-D) or 2AR-GsL (2.3-2.6 pmol/mg) (E-H), 100 µM GTP, and various salts at the concentrations indicated on the abscissa. Reaction mixtures additionally contained solvent (basal), ISO (10 µM), or ICI (1 µM). Data shown are the means ± S.D. of 3 to 10 independent experiments performed in duplicate. The effects of salts versus control (no added salt) on the individual parameters were assessed using the t test. *p < 0.05. To calculate the relative stimulatory effects of ISO on AC activity (D and H), the basal AC activity for any given salt concentration was set 100%, and the AC activities in the presence of ISO were referred to this basal AC activity.

sL

sS

sL

sS

sS

sL

sS

sL

sL

sS

Salt Regulation of AC Activity in Membranes Expressing ₂AR-G_sL: Comparison of Effects of Anions and Cations. Since the effects of sodium salts on AC regulation by ₂AR-G_sL were much more pronounced than on AC regulation by ₂AR-G_sS (Fig. 2), subsequent studies focused on ₂AR-G_sL. Figure 3 compares the effects of various chloride salts on AC activity. LiCl exhibited pronounced inhibitory effects on basal, ISO-stimulated, and ICI-inhibited AC activity in absolute terms (Fig. 3A). KCl, RbCl, and CsCl exhibited biphasic effects on maximum ISO-stimulated AC activity, i.e., at a concentration of 50 mM, these salts enhanced maximum ISO-stimulated AC activity, whereas at higher concentrations the salts diminished AC activity (Fig. 3, C-E). There was a strong linear correlation between the anion radius and the efficiency of sodium salts at enhancing the relative stimulatory effect of ISO on AC (Fig. 4A). A very similar correlation was obtained with KCl, KBr, and KI (data not shown). In contrast, there was no correlation between the cation radius and the efficiency of chloride salts at enhancing ISO stimulation of AC (Fig. 4B). ChoCl is not included in Fig. 4B because it is not a spherical cation. The spherical cation most closely related to choline is tetramethylammonium, which possesses a radius of 300 pm (McCleskey and Almers, 1985) and thus is almost twice as large as the radius of Cs⁺ (Fig. 4B). Strikingly, however, ChoCl did not differ greatly from CsCl in terms of inverse agonist inhibition and agonist stimulation of AC (Fig. 3, E and F).

View larger version (38K): [in this window] [in a new window]   Fig. 3.   Effects of LiCl, NaCl, KCl, RbCl, CsCl, and ChoCl on AC activity in Sf9 membranes expressing 2AR-GsL in the absence or presence of ISO or ICI. Reaction mixtures contained Sf9 membranes (20 µg of protein per tube) expressing 2AR-GsL (2.3-2.6 pmol/mg) and various salts at the concentrations indicated on the abscissa. Reaction mixtures additionally contained solvent (basal), ISO (10 µM), or ICI (1 µM). Data shown are the means ± S.D. of 3 to 10 independent experiments performed in duplicate. The effect of salts versus control (no added salt) on the individual parameters was assessed using the t test. *p < 0.05.

View larger version (32K): [in this window] [in a new window]   Fig. 4.   Relation between ion radius and the efficiency of salts to enhance ISO-stimulated, and to reduce ICI-inhibited, AC activity in Sf9 membranes expressing 2AR-GsL. The data shown in Figs. 2 and 3 are the basis for the comparison of the effects of salts on ISO-stimulated and ICI-inhibited AC activity. The data shown refer to a salt concentration of 150 mM each. A and B, stimulatory effects of ISO on AC activity; C and D, inhibitory effects of ICI on AC activity. In the experiments illustrated in A and C, Na+ was the cation. In the experiments illustrated in B and D, Cl was the anion. To calculate the relative stimulatory effects of ISO on AC activity, the basal AC activity in the presence of a given salt at 150 mM was set 100%, and the AC activities in the presence of ISO are referred to this basal AC activity. To calculate the relative inhibitory effects of ICI on AC activity, the ratios  ICI/ ISO as illustrated in Fig. 1, C and D, were calculated. Data were analyzed by linear regression. A, slope, 17.0 ± 0.98, p = 0.0033 (significant); B, slope, 0.44 ± 0.18, p = 0.0694 (not significant); C, slope 0.0073 ± 0.0008, p = 0.0109 (significant); D, slope, 0.0005 ± 0.0001, p = 0.0280 (significant). The effects of Br and I versus Cl and of Na+, K+, Rb+, and Cs+ versus Li+, respectively, were compared using the t test. *p < 0.05. Data shown are the means ± S.D. of 3 to 10 independent experiments performed in duplicate.

Effects of Sodium Salts on AC Activation by G_sL. The striking correlation between anion radius and the efficiency of sodium salts to modulate AC regulation by agonists and inverse agonists prompted me to study the effects of sodium salts on AC activation by the hydrolysis-resistant GTP analog GTPS, using Sf9 membranes expressing G_sL without ₂AR. Under these conditions, receptor-independent GDP/GTPS exchange takes place, and G_sL-GTPS then efficiently activates AC (Gilman, 1987). Sodium salts inhibited AC stimulation by G_sL-GTPS in the order of efficiency NaI > NaBr > NaCl (Fig. 5A). As was the case for enhancement of ISO stimulation of AC and diminishment of ICI inhibition of AC by sodium salts, there was a strong linear correlation between anion radius and the efficiency of sodium salts at reducing stimulation of AC by G_sL-GTPS (compare Figs. 4, A and C, and 5B).

View larger version (21K): [in this window] [in a new window]   Fig. 5.   Effects of NaCl, NaBr, and NaI on GTPS-stimulated AC activity in Sf9 membranes expressing GsL without 2AR. Reaction mixtures contained Sf9 membranes (20 µg of protein per tube) expressing GsL (~100 pmol/mg) and various sodium salts. Reaction mixtures additionally contained GTPS (10 µM). A, absolute AC activities in the presence of NaCl, NaBr, or NaI at the concentrations indicated on the abscissa. B, relation between anion radius and the efficiency of salts to reduce GTPS-stimulated AC activity. B, AC activity observed in the absence of salt was set 100% (control), and the inhibitory effects of salts (150 mM each) are referred to this activity. Data shown in B were analyzed by linear regression. Slope, 1.41 ± 0.06, p = 0.0021 (significant). Data shown are the means ± S.D. of three independent experiments performed in duplicate.

Effect of NaCl on GDP Affinity of G_sL. I determined the GDP affinity of G_sL by assessing the potency of GDP at inhibiting high-affinity agonist binding (Seifert et al., 1998b). In the absence of NaCl, the GDP inhibition curve was best fit to a monophasic curve (EC₅₀ = 2.0 µM; 95% confidence interval, 1.4-2.9 µM) (Fig. 6). In the presence of NaCl (150 mM), the GDP inhibition curve was best fit to a biphasic curve; i.e., 66 ± 5% of the G_sL molecules were in a state of high GDP affinity (EC₅₀ = 89 nM; 95% confidence interval, 54-147 nM), and the remaining fraction of the G_sL molecules was in a state of low GDP affinity (EC₅₀ = 5.9 µM; 95% confidence interval, 2.3-15.5 µM). The explanation for the biphasic GDP inhibition curve is that NaCl at a concentration of 150 mM is not maximally effective. As with the incomplete conversion of GDP affinity state of G_sL by NaCl at 150 mM, NaCl at 150 mM did not completely abolish the apparent constitutive activity of the ₂AR (Fig. 1).

View larger version (21K): [in this window] [in a new window]   Fig. 6.   Effect of NaCl on the GDP affinity of 2AR-GsL expressed in Sf9 membranes. Reaction mixtures contained Sf9 membranes (20 µg of protein per tube) expressing 2AR-GsL (5.0-5.5 pmol/mg) 1 nM DHA, 1 µM salbutamol (SAL), and GDP at the concentrations indicated on the abscissa. Reaction mixtures additionally contained solvent (control) or NaCl (150 mM). The DHA binding observed in the absence of GDP was set 0%, and the DHA binding observed in the presence of 1 mM GDP was set 100%. The extent of DHA binding observed with GDP at 10 nM-100 µM was referred to these calibration values. Data were analyzed by nonlinear regression for best fit to monophasic or biphasic inhibition curves (F test). The control data were best fit to a monophasic curve, whereas the data with NaCl (150 mM) were best fit to a biphasic curve. The dashed line represents the function that would have resulted if the data with NaCl (150 mM) had been best fit to a monophasic curve. The effects of GDP at a given concentration in the presence of NaCl versus control (no added salt) were assessed using the t test. *p < 0.05. Data shown are the means ± S.D. of three independent experiments performed in triplicate.

Effects of Chloride Salts on Antagonist, Inverse Agonist, and Agonist Binding to ₂AR-G_sL. In competition experiments, the affinity of ₂AR-G_sL for the antagonist ()-propranolol and the inverse agonist ICI was determined. The K_i value for ()-propranolol in the absence of NaCl was 3.6 nM (95% confidence interval, 2.7-4.8 nM) and in the presence of NaCl it was 5.6 nM (95% confidence interval, 3.8-8.3 nM, not significant). The K_i value for ICI in the absence of NaCl was 1.8 nM (95% confidence interval, 1.3-2.5 nM) and in the presence of NaCl it was 1.5 nM (95% confidence interval, 1.2-1.9 nM, not significant).

View larger version (17K): [in this window] [in a new window]   Fig. 7.   Competition of DHA binding by ISO in Sf9 membranes expressing 2AR-GsL: effects of LiCl, NaCl, and KCl. Reaction mixtures contained Sf9 membranes (20-25 µg of protein) expressing 2AR-GsL (5.0-7.5 pmol/mg), 1 nM DHA, ISO at the concentrations indicated on the abscissa in the absence of salt (no salt), or in the presence of various salts at a concentration of 150 mM each. Reaction mixtures additionally contained solvent (control) or GTPS (10 µM). Data were analyzed by nonlinear regression for best fit to monophasic or biphasic competition curves (F test). Data points are the means ± S.D. of three to six experiments performed in triplicates. The dashed lines in C and D represent the theoretical competition isotherms in the presence of GTPS that would have been obtained if the agonist-binding had been completely GTPS-sensitive.

View this table: [in this window] [in a new window]   TABLE 1 Effects of LiCl, NaCl, and KCl on agonist binding in Sf9 membranes expressing 2AR-GsL The agonist binding data shown in Fig. 7 were analyzed by nonlinear regression for best fit to monophasic or biphasic competition curves (F test). Experiments were conducted either in the absence of salt (no salt) or in the presence of various salts at a concentration of 150 mM each. Data shown are the means of three to six experiments performed in triplicates. Numbers in parentheses represent the 95% confidence intervals. Kh and Kl designate the dissociation constants for the high- and low-affinity agonist-binding to the 2AR, respectively. %Rh indicates the percentage of high-affinity binding sites. The corresponding values obtained in the presence of GTPS (10 µM) are referred to as KhGTPS, KIGTPS, and %RhGTPS, respectively. If data were best fit to monophasic competition curves, data are listed under Kl and KIGTPS, respectively. The effects of salts on Kh, Kl, %Rh, and KIGTPS, respectively, were analyzed using the t test.

	Discussion

Top Abstract Introduction Experimental Procedures Results Discussion References

NaCl Decreases the Apparent Constitutive Activity of ₂AR Coupled to G_sL. In the absence of NaCl, the ₂AR coupled to G_sL shows the hallmarks of high constitutive activity as assessed by the strong stimulatory effect of GTP on basal AC activity and the large inhibitory effect of the inverse agonist ICI on GTP-dependent AC activity (Fig. 1). NaCl diminishes the apparent constitutive activity of the ₂AR as shown by a reduction of basal AC activity and a reduction of the inhibitory effects of ICI. As a result of these effects of NaCl on basal AC activity, the stimulatory effects of the agonist ISO on AC activity are increased. The inhibitory effects of NaCl on the apparent constitutive activity of the ₂AR are not attributable to changes in ionic strength because various salts differed considerably from each with respect to reduction of apparent constitutive GPCR activity (Figs. 2 and 4).

Reduction of the Apparent Constitutive Activity of the ₂AR by Salts Cannot Be Explained by Action of Cations on ₂AR. It has been suggested that Na⁺ stabilizes the inactive (R) state of GPCRs through binding to a highly conserved aspartate residue in the second transmembrane domain of GPCRs (Costa et al., 1990; Kong et al., 1993; Ceresa and Limbird, 1994; Wenzel-Seifert et al., 1998). Thus, the question arose whether the effects of salts on the apparent constitutive activity of the ₂AR are mediated through the putative cation-binding site in the second transmembrane domain. A previous study had shown that alkali salts reduce the agonist affinity of ARs in calf cerebellar membranes in the order of potency Li⁺ > Na⁺ > K⁺ (U'Prichard et al., 1978). I observed the same order of potency of salts when K_lGTPS values were considered (Table 1). In agreement with previous results (U'Prichard et al., 1978), NaCl had no effect on ₂AR antagonist- and inverse agonist binding. LiCl prevented high-affinity agonist binding (Fig. 7B; Table 1), but the effect of LiCl on the apparent constitutive activity of the ₂AR and agonist stimulation of AC was rather modest (Figs. 3 and 4). In contrast to LiCl, NaCl and KCl left high-affinity agonist binding intact (Fig. 7, C and D), but LiCl, NaCl and KCl were similarly efficient at reducing the apparent constitutive activity of the ₂AR coupled to G_sL (Figs. 3 and 4). Moreover, the effects of salts on the apparent constitutive activity of the ₂AR did not correlate with the cation radius but rather with the anion radius (Fig. 4). Collectively, all these data indicate that the effects of salts on the apparent constitutive activity of the ₂AR cannot be explained by modulation of ₂AR function through the putative cation-binding site of the GPCR. Thus, stabilization of the R state by monovalent cations is not a general mechanism by which salts reduce the apparent constitutive activity of GPCRs.

Inhibitory Effect of Salts on the Apparent Constitutive Activity of the ₂AR Coupled to G_sL Can Be Explained by Anion Effects on G Protein. My present results clearly show that the inhibitory effects of salts on the apparent constitutive activity of the ₂AR coupled to G_sL are attributable to anions (Fig. 4), but there is no evidence for an anion-binding site in GPCRs. Thus, the question arose whether the effects of salts on the apparent constitutive activity of the ₂AR coupled to G_sL could be mediated through an increase in GDP affinity of the G protein. The result would be a decreased efficiency of the agonist-free ₂AR and an increased efficiency of the agonist-occupied ₂AR at promoting GDP dissociation and hence AC activation. Several findings support this hypothesis. First, monovalent anions increase the GDP affinity of various purified G proteins, including G_s proteins (Higashijima et al., 1987). Second, G_sS possesses a much higher GDP affinity than G_sL (Graziano et al., 1989; Seifert et al., 1998b), and the effects of salts on ₂AR/G_sS coupling were much less pronounced than the effects of salts on ₂AR/G_sL coupling (Fig. 2). Third, the relation between anion radius and reduction of AC activation by G_sL-GTPS was the same as the relation between anion radius and reduction of the apparent constitutive activity of the ₂AR (Figs. 4, A and C, and 5B). These findings suggest that the same mechanism underlies both processes. Fourth, NaCl increased the GDP affinity of G_sL (Fig. 6).

sL

sL

sL

Differential Regulation of G_s Splice Variants by Anions: Physiological Considerations. ₂AR/G_sS coupling is much less sensitive to regulation by monovalent anions than ₂AR/G_sL coupling (Fig. 2). The molecular basis for the relative resistance of G_sS to regulation by anions presumably is that the intrinsic GDP affinity of this G protein is higher than the GDP affinity of G_sL (Graziano et al., 1989; Seifert et al., 1998b) so that anions can increase the GDP affinity of G_sS only to a little extent, if at all. It should be emphasized that the experiments in this study were conducted with fusion proteins to ensure precisely defined experimental conditions. However, although fusion proteins have experimental advantages relative to conventional coexpression systems, fusion proteins do not occur in vivo. Thus, it will be important to confirm the present results with coexpression systems.

sL

sL

sL

sS

sL

sS

sS

sL

sS

sL

Limitations of Anions as Experimental Tools. Although this study shows that Cl, Br, and I are valuable tools for dissecting functional differences between G_s splice variants, it should be kept in mind that salts exhibit pleiotropic effects. First, chloride salts alter the agonist-binding properties of the ₂AR (Fig. 7; Table 1). Second, in contrast to NaCl and NaBr, NaI was similarly efficient at reducing basal AC activity in membranes expressing ₂AR-G_sS and ₂AR-G_sL (Fig. 2, C and G). Third, among the various chloride salts analyzed, LiCl was particularly efficient at reducing basal and maximum agonist-stimulated AC activity (Fig. 3). These effects of NaI and LiCl could be attributable to direct effects of I and Li⁺, respectively, on AC (Roy et al., 1977).

Conclusion. Monovalent anions differentially modulate coupling of the ₂AR to G_s splice variants. My data are explained by a model in which anions increase the GDP affinity of G_sL more efficiently than the GDP affinity of G_sS. As a result, anions reduce the efficiency of the agonist-free ₂AR and enhance the efficiency of the agonist-occupied ₂AR at promoting GDP dissociation from G_sL and hence activating AC. So far, research has focused on the analysis of the effects of cations on GPCR/G protein coupling (Costa and Herz, 1989; Gierschik et al., 1989; Hilf and Jakobs, 1992; Tian et al., 1994; Wenzel-Seifert et al., 1998). The observation that anions rather than cations determine GPCR/G protein coupling in at least one prototypical system calls for the systematic analysis of anion effects in other GPCR/G protein systems.