Activator of G-protein Signaling 1 Blocks GIRK Channel Activation by a G-protein-coupled Receptor APPARENT DISRUPTION OF RECEPTOR SIGNALING COMPLEXES*

The Ras-related protein, activator of G-protein signal- ing 1 (AGS1) or Dexras1, interacts with G i /G o (cid:1) and activates heterotrimeric G-protein signaling systems independent of a G-protein-coupled receptor (GPCR). As an initial approach to further define the cellular role of AGS1 in GPCR signaling, we determined the influence of AGS1 on the regulation of G (cid:2)(cid:3) -regulated inwardly recti- fying K (cid:4) channel (GIRK) current ( I ACh ) by M 2 -musca-rinic receptor (M 2 -MR) in Xenopus oocytes. AGS1 ex- pression inhibited receptor-mediated current activation by > 80%. Mutation of a key residue (G31V) within the G 1 domain involved in nucleotide binding for Ras- related proteins eliminated the action of AGS1. The inhibition of I ACh was not overcome by increasing concen- trations of the muscarinic agonist acetylcholine but was progressively lost upon injection of increasing amounts of M 2 -MR cRNA. These data suggest that AGS1 may an- tagonize GPCR signaling by altering the pool of heterotrimeric G-proteins available for receptor coupling and/or disruption of a preformed signaling complex. Such regulation would be of particular importance for those receptors that exist precoupled to heterotrimeric G-protein and for receptors

Heterotrimeric G-proteins transduce a variety of extracellular stimuli into intracellular responses. Such stimuli are primarily sensed by the superfamily of G-protein-coupled receptors. In general, the specific intracellular response is determined by the external stimuli itself and cell type-specific expression of receptors, G-proteins, and effectors. However, additional proteins also operate to influence signaling specificity as well as signal magnitude and duration. Such accessory proteins may act as scaffolding proteins within a signal transduction complex and/or directly influence the basal activation state of G-proteins and effectors independent of an activated GPCR 1 (1)(2)(3)(4)(5)(6)(7)(8). The latter proteins include both the family of regulators of G-protein signaling (RGS), initially defined based upon their ability to accelerate the GTPase activity of selected G i ␣ proteins, and the recently identified activators of G-protein signaling (AGS) 1-3.
AGS proteins (AGS1-3) were identified in a yeast-based functional screen as receptor-independent activators of G-protein signaling (1,2). These proteins do not share any sequence homology, and each entity exerts different effects within the context of the G-protein activation/deactivation cycle (1)(2)(3)(4). AGS1 (AF069506) selectively activated the G i /G o -protein signaling pathway, and it appears to act as a guanine nucleotide exchange factor for G i , somewhat mimicking a GPCR (1,4). AGS1 is a member of the Ras superfamily of small G-proteins providing a potential interface between signaling pathways regulated by the two broad classes of G-proteins. AGS1 is the human counterpart of the Ras-related protein DexrasI (NP_033052), which was identified as a dexamethasone-inducible cDNA in AtT-20 mouse corticotroph cells (9), where it may influence cAMP regulation of hormone secretion (10). AGS1 was also implicated in N-methyl-D-aspartate receptor signaling in neuronal cells, where it is an apparent target of neuronal nitric-oxide synthase (11). Thus, AGS1 is clearly involved in cellular signaling events and binds to as well as activates G i /G o ␣, but it is not known how this apparent G-protein regulator influences the activation of signaling systems by a GPCR. AGS1 may influence the specificity, magnitude, or duration of GPCR signaling events as it may actually interfere with receptor-effector coupling (12) 2 by altering the pool of G-protein available for interacting with receptor.
As an initial approach to define the role of AGS1 in GPCR signaling, we determined the influence of AGS1 on the regulation of G␤␥-regulated inwardly rectifying K ϩ channel (GIRK) current (I ACh ) by M 2 -muscarinic receptors (M 2 -MR) in Xenopus oocytes (13,14). AGS1 expression had little effect on basal levels of current (I hK ) but inhibited the increase in GIRK channel activity elicited by activation of the M 2 -MR. The inhibition of I ACh by AGS1 was progressively lost upon injection of increasing amounts of M 2 -MR cRNA. These data suggest that AGS1 antagonized GPCR signaling by altering the pool of heterotrimeric G-proteins available for receptor coupling and/or by disrupting a preformed signal transduction complex.

EXPERIMENTAL PROCEDURES
cRNA Synthesis-GIRK1, GIRK4, and M 2 -MR in pcDNA3.1 vector (Invitrogen) were kindly provided by Drs. Paolo Kofuji and H. A. Lester (California Institute of Technology). AGS1 and AGS1G31V were prop-agated in the pcDNA3.1-HisA vector (Invitrogen). All plasmids were linearized with appropriate enzymes at sites immediately following translational stop codons and used for capped cRNA synthesis using mMESSAGE MACHINE kits (Ambion Inc., Austin, TX).
Oocyte Preparation and Injection-Xenopus oocytes were surgically extracted and dissociated with defolliculation by collagenase treatment (1 mg/ml collagenase A; Roche Applied Sciences, Indianapolis, IN) (15). Healthy stage V and VI oocytes were selected and maintained at 19°C in ND96 buffer (96 mM NaCl, 2 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 5 mM HEPES, pH 7.5) supplemented with 5% horse serum, 2.5 mM sodium pyruvate, and 50 g/ml gentamycin. Various combinations of cRNAs (final volume, 50 nl in diethyl pyrocarbonate-H 2 O) were injected into stage V and VI oocytes with a Drummond microinjector. Injected oocytes were incubated for 2 days at 19°C with replacement of culture medium twice daily.
Electrophysiology-Electrophysiological measurements were carried out using a standard two-electrode voltage clamp technique (16). Oocytes were clamped at Ϫ80 mV using a Warner oocyte clamp (Warner Instruments, Hamden, CT). Voltage and current recording electrodes had resistances of 0.5-2 megaohms. Constitutive GIRK-associated current responses (I hK ) were elicited by switching the perfusion medium from ND96 to a high potassium solution (2 mM NaCl, 96 mM KCl, 1 mM MgCl 2 , 1 mM CaCl 2 , 5 mM Hepes, pH 7.5). Acetylcholine-induced GIRK current responses were elicited by switching the perfusion solution from high potassium to high potassium solution containing acetylcholine. The difference between the acetylcholine-mediated current and I hK is referred to as I ACh . Switching the perfusion solution back to ND96 returned the current to base-line values (see Fig. 1A).

RESULTS AND DISCUSSION
The ability of AGS1 to interact with heterotrimeric G-proteins raises immediate questions as to its role in signal processing following activation of a GPCR at the cell surface. To address this question, we investigated the role of AGS1 on M 2 -MR coupling to GIRK channels using a Xenopus oocyte expression system. The Xenopus oocyte expression system has been widely utilized to analyze the function and regulation of GIRK channel activities because of the ease of gene expression and functional readouts. In oocytes injected with GIRK1/4 cRNAs and M 2 -MR cRNA, an inward K ϩ current (I hK ) was elicited by exchanging ND96 for a high potassium solution. Acetylcholine elicited an additional inward K ϩ current (I ACh ), and this receptor-mediated event was completely blocked by injection of pertussis toxin. 3 Co-injection of AGS1 with GIRK1/4 and M 2 -MR cRNAs elicited little change in I hk but markedly diminished I ACh (Fig. 1). 4 Mutation of residues in the G 1 domain of AGS1 (G31V) rendered AGS1 inactive (Fig. 1, A and B), as was the case for AGS1 in the yeast assay system and for AGS1 regulation of ERK activity in COS-7 cells (1, 4). The inhibition of I ACh by expression of AGS1 was progressively enhanced by injection of increasing amounts of AGS1 cRNAs (Fig. 2). At the lower expression levels of GIRK channels (0.01 ng/oocyte), AGS1 inhibited I ACh by 76 Ϯ 4.2% (Fig. 2A). In oocytes expressing higher levels of GIRK channels, AGS1 inhibited I ACh by 47 Ϯ 7.7% (Fig. 2B). Expression of AGS1G31V did not alter I ACh at either expression level of GIRK channels (Fig. 2), indicating that the inhibition of I ACh was not because of altered expression of receptor or channels, per se. This thought is further supported by the absence of a decrease in I hk when AGS1 was coexpressed with GIRK channels. Thus, AGS1 likely inhibits I ACh by interfering with the transfer of signal from receptor to G-protein or perhaps from G-protein to the channels. Similar conclusions regarding the action of AGS1 on GPCR signaling were reported by Graham et al. (12) for formyl peptide receptor activation of ERK1/2 kinases in COS-7 cells.
As AGS1 clearly interacts with G i /G o ␣ subunits (1, 4), AGS1 may inhibit I ACh by competing with M 2 -MR for the available pool of G-proteins. This possibility was addressed by examining the dose-response curve for acetylcholine in the presence and absence of AGS1 and by determining the ability of increasing amounts of M 2 -MR to overcome the AGS1-mediated I ACh inhibition. Under standard experimental conditions (cRNA/oocyte: M 2 -MR, 0.1 ng; GIRK1/4, 2 ng for each; AGS1, 5 ng) increasing concentrations of acetylcholine did not overcome the inhibitory effect of AGS1 on I ACh (Fig. 3A). In contrast to the influence of increasing amounts of agonist, the inhibitory action of AGS1 on I ACh was overcome by increasing the levels of expressed receptor. Increased levels of receptor increased the amount of I Ach at fixed amounts of GIRK channels, suggesting that under these experimental conditions receptor was somewhat rate-limiting or that at least G-proteins and effectors were not saturated  (17). At the lowest expression levels of M 2 -MR (0.1 ng of cRNA/ oocyte), AGS1 inhibited I ACh by 80 Ϯ 3.9% I ACh (Fig. 3B). At the higher levels of M 2 -MR expression, AGS1 inhibited I ACh by only 17 Ϯ 5.9% I ACh (Fig. 3B). AGS1G31V did not inhibit I ACh at any levels of M 2 -MR expression (Fig. 3B).
These observations raise several interesting points relative to signal processing through this system. The ability of increased amounts of receptor to overcome the inhibitory effect of AGS1 suggests that, indeed, AGS1 competes with the M 2 -MR for the pool of available heterotrimeric G i /G o to activate GIRK channel. This may reflect an action of AGS1 to disrupt a preconfigured signal transduction complex that is required for agonist activity. Thus increasing concentrations of agonist could not overcome the inhibitory affect of AGS1 action. Such a signal transduction complex may consist of receptor and Gprotein, G-protein and GIRK, or perhaps all three entities (17)(18)(19). By complexing with a defined population of G-proteins, AGS1 may limit the ability of a receptor to act catalytically as it cannot access multiple G-proteins.
These thoughts likely have important implications for G i /G ocoupled receptors that exist in a precoupled state where the receptor is complexed with G-protein and "stabilized" in a conformation exhibiting high affinity for agonist. AGS1 may block the formation of this precoupled complex, and as such, even high concentrations of agonist would not be able to activate downstream effectors. Increasing the amount of receptors in the presence of AGS1 as opposed to agonist itself would allow more of the receptor population to exist in a precoupled state and thus effectively override the inhibitory effect of AGS1 as was indeed the case in the present study.
A flurry of recent reports indicates that multiple proteins interact with and/or regulate the activation state of heterotrimeric G-proteins. Although the role of these proteins in GPCR processing can be varied, they certainly offer unexpected avenues for manipulating the signaling system. Controlling the population of receptors precoupled with G-protein or the generation of signaling complexes may be a key mechanism for regulating the action of hormones and as such provide new pathways for therapeutics that mimic or disrupt specific signaling systems.