A Highly Effective Dominant Negative αs Construct Containing Mutations That Affect Distinct Functions Inhibits Multiple Gs-coupled Receptor Signaling Pathways*

To investigate the subcellular organization of receptor-G protein signaling pathways, a robust dominant negative αs mutant containing substitutions that alter distinct functions was produced and tested for its effects on Gs-coupled receptor activity in HEK-293 cells. Mutations in the α3β5 loop region, which increase receptor affinity, decrease receptor-mediated activation, and impair activation of adenylyl cyclase, were combined with G226A, which increases affinity for βγ, and A366S, which decreases affinity for GDP. This triple αs mutant can inhibit signaling to Gs from the luteinizing hormone receptor by 97% and from the calcitonin receptor by 100%. In addition, this αs mutant blocks all signaling from the calcitonin receptor to Gq. These results lead to two conclusions about receptor-G protein signaling. First, individual receptors have access to multiple types of G proteins in HEK-293 cell membranes. Second, different G protein α subunits can compete with each other for binding to the same receptor. This dominant negative αs construct will be useful for determining interrelationships among distinct receptor-G protein interactions in a wide variety of cells and tissues.

Stimulation of heterotrimeric G proteins by cell surface receptors activates signaling pathways that mediate specific responses to hormones and neurotransmitters. Cells express a wide variety of G protein-coupled receptors as well as numerous G protein ␣, ␤, and ␥ subunits. Many receptors can activate more than one type of G protein, and the G protein subunits can interact with many different types of receptors. The manner in which signaling specificity is maintained in the midst of this vast range of potential interactions is not well understood. This report investigates the interdependence of distinct signaling pathways activated by receptors with broad G protein specificities using a receptor-sequestering dominant negative G protein ␣ subunit.
Many potential mechanisms could establish that distinct receptor-G protein interactions will be independent of each other. Among these, one possibility is that specific receptor-G protein complexes localize to separate membrane compartments (1,2). Differential associations with particular proteins or lipids (3,4) or covalent modifications such as phosphoryla-tion (5) may result in subpopulations of receptors and G proteins that have restricted access to each other. Although G proteins are often expressed at much higher levels than their receptors are (6), there is evidence that different receptors utilize separate pools of G proteins (7,8). An alternative potential mechanism for isolating distinct receptor-G protein interactions is that receptors utilize separate regions for binding different G proteins. If this is the case, then multiple types of receptor-G protein interaction can occur simultaneously without affecting each other. Localization of G protein-binding sites have indicated that separate receptor regions may specify interactions with distinct G proteins (9,10).
Dominant negative G protein ␣ subunits can test potential mechanisms for separating the distinct G protein interactions of broad specificity receptors. For instance, if interactions with different G proteins are localized to separate subcellular compartments and/or receptor regions, signaling from a receptor to one type of G protein ␣ subunit will be blocked by a dominant negative version of that ␣ subunit, whereas the other signaling pathways will be unperturbed. Alternatively, if each receptor has access to multiple types of G protein and these different G proteins can compete with each other for receptor binding, then a dominant negative ␣ subunit will block all of the G protein signaling pathways activated by the receptor.
Several dominant negative G protein ␣ subunits have been developed previously, but they inhibit G protein signaling incompletely and therefore are not optimal for investigating receptor-G protein signaling pathways. One dominant negative ␣ s mutant contains three substitutions that disrupt different ␣ subunit functions, but it is extremely unstable (11), which contributes to its inability to inhibit signaling completely (12,13). Xanthine-binding mutants of ␣ o , ␣ 11 , and ␣ 16 can inhibit signaling of specific G protein-coupled receptor families, but inhibition of receptor-mediated phospholipase C stimulation is incomplete (14,15). ␣ subunit carboxyl-terminal fragments exhibit dominant negative activity (16), but inhibition of G s signaling is only partial (13).
This report describes the development of a highly effective dominant negative ␣ s mutant that contains substitutions that alter distinct ␣ subunit functions, each of which should stabilize the receptor-bound, nucleotide-free state of G s . One set of mutations, located in the ␣3␤5 loop region, specifically increases receptor affinity and decreases receptor-mediated activation without affecting nucleotide handling (17) and also disrupts activation of adenylyl cyclase (18). G226A increases affinity for ␤␥ (19,20), and A366S decreases affinity for GDP (21). Although A366S alone causes ␣ s to be thermolabile, ␣ s (␣3␤5/G226A/A366S), containing all three sets of mutations, is expressed at close to wild-type levels and blocks signaling from the luteinizing hormone receptor to G s by up to 97%. The effects of ␣ s (␣3␤5/G226A/A366S) on signaling by the calcitonin receptor to G s and G q are tested in transiently transfected HEK-293 cells. The results demonstrate that this dominant negative ␣ s mutant can block multiple G protein signaling pathways, which indicates that each receptor has access to multiple types of G protein and that these G proteins can compete with each other for receptor binding.

EXPERIMENTAL PROCEDURES
Construction of ␣ s Mutant Constructs-␣ s mutant constructs in the expression vector pcDNAI/Amp (Invitrogen) were generated from the rat ␣ s cDNA (22) containing the EE epitope (23), which was generated by mutating ␣ s residues DYVPSD (residues 189 -194) to EYMPTE. Mutations were generated by oligonucleotide-directed in vitro mutagenesis using the Bio-Rad Muta-Gene kit except for those in the ␣3␤5 region, which were produced by subcloning mutagenic oligodeoxynucleotide cassettes. Subcloning and mutagenesis procedures were verified by restriction enzyme analysis and DNA sequencing. After an additional 24 h, intracellular cAMP levels in cells labeled with [ 3 H]adenine and inositol phosphate levels in cells labeled with [ 3 H]inositol were determined as described previously (24). cAMP accumulation was measured in the presence of 3-isobutyl-1-methylxanthine and in the presence or absence of agonist as indicated in the figure legends. Inositol phosphate formation was measured in the presence of 5 mM LiCl and in the presence or absence of agonist as indicated in the figure legends. The cells were cultured and assayed at 37°C.
To determine EC 50 values for stimulation of cAMP or inositol phosphate formation, the observed activity was fitted to Equation 1, where X is the concentration of agonist; Y is the observed cAMP or inositol phosphate formation; a is the cAMP or inositol phosphate formation observed in the absence of agonist; b is the maximum observed cAMP or inositol phosphate formation; c is the half-maximal effective concentration (EC 50 ) of the agonist; and d is the slope factor. Transient Expression and Assays for Ligand Binding-HEK-293 cells (6.25 ϫ 10 6 per 150-mm dish) were transfected with 0.2 g/10 6 cells of plasmid encoding the rat luteinizing hormone receptor in pCIS (25) or 0.04 g/10 6 cells of plasmid encoding the rabbit C1a calcitonin receptor in pBKCMV (9) and 2 g/10 6 cells of vector (pcDNAI/Amp) using 62.5 l of LipofectAMINE 2000 Reagent. 24 h after transfection, the cells were replated in 24-well plates. Assays for ligand binding were performed after an additional 24 h.
Receptor numbers per cell were determined by fitting data from saturation binding assays using 125 I-labeled hCG 1 (26,27) or sCT (9) to Equation 2, where X is the concentration of agonist; Y is the specific binding; B max is the maximum number of binding sites, and K d is the equilibrium dissociation constant. In each case, the medium was removed, and the cells were washed once with binding medium (20 mM HEPES-buffered minimal essential medium with Earle's salts without bicarbonate containing 1 mg/ml bovine serum albumin). The medium was then removed and replaced with binding medium containing 0.062-15 nM 125 I-labeled hCG in the presence or absence of 0.44 M unlabeled hCG or 0.030 -15 nM 125 I-labeled sCT in the presence or absence of 1 M unlabeled sCT. The cells were incubated for 1 h at room temperature, washed three times with ice-cold phosphate-buffered saline, and solubilized with 0.5 N NaOH. The samples were collected and counted in a gamma counter.
Membrane Preparations and Immunoblots-HEK-293 cells (6.25 ϫ 10 6 per 150-mm dish) were transfected with 37.5 g of plasmid using DEAE-dextran (28). 48 h after transfection, membranes were prepared as described (24). 25 g of membrane proteins were resolved by SDS-PAGE (10%), transferred to nitrocellulose, and probed with a monoclonal antibody to the EE epitope (23). The antigen-antibody complexes were detected using an anti-mouse horseradish peroxidase-linked antibody according to the ECL Western blotting protocol (Amersham Biosciences).

Combining Substitutions in the ␣3␤5 Loop Region of ␣ s with Mutations That Alter ␤␥ and GDP Binding Results in Highly
Effective Dominant Negative Activity-Replacing five ␣ s residues in the ␣3 helix and the ␣3␤5 loop (29) with the homologous ␣ i2 residues (N271K, K274D, R280K, T284D, and I285T) results in an ␣ s construct, ␣ s (␣3␤5), that exhibits increased affinity for and a decreased ability to be activated by the ␤ 2adrenergic receptor (17). Independently, the mutations also disrupt activation of adenylyl cyclase by ␣ s (18). These properties suggested that ␣ s (␣3␤5) might be able to sequester G scoupled receptors and exhibit dominant negative activity. Indeed, when ␣ s (␣3␤5) was transiently expressed in HEK-293 cells that were co-transfected with plasmid encoding the luteinizing hormone receptor, cAMP accumulation in response to 20 ng/ml hCG was inhibited by 42% (S.E. ϭ 3%, n ϭ 3) (Fig. 1).
With the goal of producing more effective dominant negative activity, additional mutations predicted to stabilize the receptor-G protein complex were introduced into ␣ s in combination with the ␣3␤5 substitutions. The effects of adding each of three mutations, G226A, E268A, and A366S, to ␣ s (␣3␤5) were tested. When combined, these three mutations produce partial dominant negative activity in ␣ s (11). G226A impairs activating conformational changes in switch II required for dissociation of ␣ s from ␤␥ (19,20). E268A disrupts a salt bridge with Arg-231 that may stabilize the activated conformation of ␣ s (30). A366S elevates basal GDP release, causing ␣ s to be constitutively activated and to spend more time in the empty state (21). ␣ s (G226A/E268A/A366S) inhibited cAMP accumulation in response to 20 ng/ml hCG by 69% (S.E. ϭ 1%, n ϭ 3) (Fig. 1).
The dominant negative ␣ s mutants, ␣ s (␣3␤5), ␣ s (G226A/ E268A/A366S), and ␣ s (␣3␤5/G226A/A366S), decreased the effectiveness of signaling from the luteinizing hormone receptor to G s both by increasing the EC 50 value of the cAMP response to hCG and by decreasing the magnitude of cAMP responses to hCG ( Fig. 2 and Table I). Compared with ␣ s (␣3␤5) and ␣ s (G226A/E268A/A366S), the effects of ␣ s (␣3␤5/G226A/A366S) on both of these parameters were greater. These incremental effects of combining the ␣3␤5, G226A, and A366S mutations are consistent with their independent sites and mechanisms of action.
Substitutions in the ␣3␤5 Loop Rather Than in the ␣3 Helix Produce Dominant Negative Activity-To investigate how altering the ␣3␤5 loop region of ␣ s produces dominant negative activity, the effects of smaller numbers of substitutions in this region were tested. Of the five ␣ i2 homolog substitutions, two are located in the ␣3 helix (N271K and K274D) and three (R280K, T284D, and I285T) are in the ␣3␤5 loop. Based on their location, only the loop residues are likely to interact directly with receptors (17). Separately testing the effects of substitutions in these two regions showed that the dominant negative effect is due predominantly to the substitutions in the ␣3␤5 loop, rather than those in the ␣3 helix (Fig. 3). ␣ s (R280K/ T284D/I285T) was only slightly less effective as a dominant negative than ␣ s (␣3␤5) was, whereas ␣ s (N271K/K274D) exhibited the same basal and receptor-stimulated cAMP accumulation as ␣ s did. Therefore, the dominant negative phenotype of ␣ s (␣3␤5) appears to result from the alteration of a receptor contact site on ␣ s .
Individual substitutions of each of the ␣3␤5 loop residues decreased receptor-mediated activation of ␣ s but did not produce dominant negative activity (Fig. 3). ␣ s (R280K) exhibited very little basal or receptor-stimulated activity, whereas the activities of ␣ s (T284D) and ␣ s (I285T) were decreased relative to that of ␣ s . The dominant negative activity produced by simultaneously substituting the three residues may result from additive defects in receptor interaction or may involve conformational changes in the loop due to interactions among the mutated residues.
The ␣3␤5 Substitutions Compensate for the Instability Caused by the A366S Mutation-A previously reported limitation of ␣ s (G226A/E268A/A366S) is its instability (11), due to the A366S mutation, which increases the amount of time ␣ s spends in the thermolabile nucleotide-free state (21). To determine whether the greater dominant negative activity of ␣ s (␣3␤5/ G226A/A366S) compared with ␣ s (G226A/E268A/A366S) is due  in part to a stabilizing effect of the ␣3␤5 substitutions, the expression levels of these ␣ s constructs in membranes of transfected HEK-293 cells were compared (Fig. 4). The expression level of ␣ s (␣3␤5) was similar to that of ␣ s , whereas that of ␣ s (G226A/E268A/A366S) was much lower. The expression level of ␣ s (␣3␤5/G226A/A366S) was reduced somewhat relative to that of ␣ s (␣3␤5) and of ␣ s but was much higher than that of ␣ s (G226A/E268A/A366S). Thus, the ␣3␤5 substitutions appear to counteract the destabilizing effect of A366S. ␣ s (␣3␤5/G226A/A366S) Inhibits Signaling of the Calcitonin Receptor to G s and G q -To investigate how dominant negative ␣ s activity affects the signaling pathways of G s -coupled receptors that also interact with other G protein heterotrimers, signaling of the calcitonin receptor to G s and G q was monitored in the presence and absence of ␣ s (␣3␤5/G226A/A366S). ␣ s (␣3␤5/G226A/A366S) completely blocked ␣ s -mediated cAMP accumulation in response to up to 0.48 nM sCT (Fig. 5A). Even at saturating amounts of sCT, this response was inhibited by 82.3% (S.E. ϭ 0.3%, n ϭ 3). ␣ s (␣3␤5/G226A/A366S) increased the EC 50 for stimulation of cAMP accumulation by sCT by a factor of 10 (Table II). As shown previously (9,31), the calcitonin receptor coupled less efficiently to G q than to G s ( Fig. 5 and Table II). In the presence of ␣ s (␣3␤5/G226A/A366S), ␣ q -dependent inositol phosphate formation in response to all doses of sCT was blocked entirely (Fig. 5B).
The ability of a dominant negative ␣ s mutant to block signaling from the calcitonin receptor to both G s and G q leads to two conclusions regarding calcitonin receptor-G protein signaling. First, each calcitonin receptor has access to both ␣ s and ␣ q in transfected HEK-293 cells, suggesting that these ␣ subunits share a common pool of receptors. Second, ␣ s can compete with ␣ q for binding to the calcitonin receptor. DISCUSSION Combining substitutions in three different regions of ␣ s results in dominant negative ␣ s activity that can inhibit signaling from G s -coupled receptors by close to 100%. The substitutions affect distinct ␣ s interactions as follows: increasing receptor affinity, decreasing receptor-mediated activation, and decreasing activation of adenylyl cyclase (␣3␤5 substitutions), increasing affinity for ␤␥ (G226A), or decreasing affinity for GDP (A366S). Together, these mutations appear to stabilize the nucleotide-free ␣␤␥-receptor complex. The incremental increases in dominant negative activity that result from combining these mutations are consistent with their independent sites and mechanisms of action. Although the nucleotide-free state, which is increased by the A366S mutation, is inherently unstable, ␣ s (␣3␤5/G226A/A366S) is expressed at a level close to that of wild-type ␣ s . This is most likely because the increased receptor affinity caused by the ␣3␤5 substitutions stabilizes the empty state, which has the highest receptor affinity (32). The  ␣3␤5 substitutions, on their own, do not affect the nucleotide handling properties of purified ␣ s (17). The ability of dominant negative ␣ s activity to block signaling of the calcitonin receptor to multiple G protein pathways suggests that, at least in HEK-293 cells, distinct receptor-G protein complexes are not strictly compartmentalized into separate membrane domains. However, some mechanisms for compartmentalizing distinct receptor-G protein signaling pathways might not be detected using a dominant negative ␣ s mutant. For instance, association of receptor subpopulations with distinct ␤␥ combinations could restrict potential ␣ subunit interactions. Inactivation of specific G protein subunits using antisense (33)(34)(35)(36)(37) and ribozyme (38,39) strategies has demonstrated a remarkable specificity of interaction between receptors, ␣␤␥ combinations, and effectors. In particular, in HEK-293 cells, ribozyme-mediated suppression of ␥ 7 specifically reduced expression of ␤ 1 and disrupted activation of G s by ␤-adrenergic but not prostaglandin E 1 receptors (38). Such ␤␥ specificity requirements might be overcome by a dominant negative ␣ s mutant with increased affinity for both receptors and ␤␥. In addition, although the ratio of G s to its receptors in cells is generally ϳ100:1 (6), factors important for compartmentalization might become limiting when a dominant negative ␣ s mutant is overexpressed. If this is so, then coexpressing potential compartmentalization factors with this ␣ s mutant would be predicted to narrow the range of its inhibitory capacities.
The effect of dominant negative ␣ s activity on other G protein pathways may depend on the cell and/or receptor type. For instance, cell-specific factors such as the caveolins, which have been reported to interact preferentially with particular G protein subunits (3), may restrict the accessibility of G proteins to receptors. HEK-293 cells do not have caveoli, although ␤-adrenergic receptors and adenylyl cyclase V/VI localize to low buoyant density membrane domains in these cells (40). In addition, membrane fractionation studies have provided evidence for microdomains in the plasma membranes of neuroblastoma cells (41) and neutrophils (42) that have differences in their G protein content, and polarized epithelial cells differentially sort G protein ␣ subunits (43) and G protein-coupled receptors (44). It will be of interest to use the dominant negative ␣ s mutant described here to sort out the potential role of intracellular compartmentalization in regulating G protein signaling pathways in a wide range of cells and tissues.
The ability of a dominant negative ␣ s mutant to inhibit signaling to G q provides a molecular insight into receptor-G protein interactions in that it demonstrates that different types of G proteins can compete for binding to the calcitonin receptor. Previous studies of calcitonin receptor isoforms containing insertions or deletions identified distinct regions important for specific interactions but did not rule out mutually exclusive binding of different ␣ subunits. The first intracellular loop (45)(46)(47) and an intact seventh transmembrane helix (9) appear to be important for coupling of this receptor to ␣ q but not ␣ s . The ability of different G proteins to compete for receptor binding despite these differences in specificity requirements indicates that either there is overlap in the receptor binding sites for different ␣ subunits or that binding of one type of ␣ subunit to a receptor sterically or allosterically blocks the association of a different one.
Competition between different G proteins for receptor binding raises the possibility that changes in the expression level of a particular G protein may affect other G protein signaling pathways as well. Alterations in G protein ␣ subunit expression levels can take place on several time scales. Short term changes in the expression level of ␣ s due to decreased stability of the activated state have been observed (48). More long term changes in the expression levels of ␣ s (49) and ␣ i2 (50) can occur during and play a role in development. In addition, pseudohypoparathyroidism is associated with decreased levels of functional ␣ s (51).
The dominant negative ␣ s mutant described here will have many applications to the investigation of how receptor-G protein signaling is regulated. In cell types for which there is strong evidence of subcellular compartmentalization, such as neurons and polarized epithelia, it may inhibit receptor subpopulations and be useful for determining which proteins coexist in the same membrane microdomains. In addition, it can be used to investigate the role of receptor-G protein interactions in the targeting of these proteins. For instance, if receptors are involved in the targeting of G protein subunits, receptor sequestration by a dominant negative ␣ subunit may alter the localization patterns of wild-type subunits. The localization patterns of dominant negative ␣ subunits as well as those of wild-type G protein subunits and receptors can be studied using fusions of these proteins to green fluorescent protein (52)(53)(54)(55).