Gating properties of girk channels activated by gα_o- and Gα_i-Coupled Muscarinic m2 Receptors in Xenopus Oocytes: The Role of Receptor Precoupling in RGS Modulation

Abstract

‘Regulators of G protein Signalling’ (RGSs) accelerate the activation and deactivation kinetics of G protein-gated inwardly rectifying K⁺ (GIRK) channels. In an apparent paradox, RGSs do not reduce steady-state GIRK current amplitudes as expected from the accelerated rate of deactivation when reconstituted in Xenopus oocytes. We present evidence here that this kinetic anomaly is dependent on the degree of G protein-coupled receptor (GPCR) precoupling, which varies with different Gα_i/o-RGS complexes. The gating properties of GIRK channels (Kir3.1/Kir3.2a) activated by muscarinic m2 receptors at varying levels of G protein expression were examined with or without the co-expression of either RGS4 or RGS7 in Xenopus oocytes. Different levels of specific m2 receptor-Gα coupling were established by uncoupling endogenous pertussis toxin (PTX)-sensitive Gα_i/o subunits with PTX, while expressing varying amounts of a single PTX-insensitive subunit (Gα_i1(C351G), Gα_i2(C352G), Gα_i3(C351G), Gα_oA(C351G), or Gα_oB(C351G)). Co-expression of each of the PTX-insensitive Gα_i/o subunits rescued acetylcholine (ACh)-elicited GIRK currents (I_K,ACh) in a concentration-dependent manner, with Gα_o isoforms being more effective than Gα_i isoforms. Receptor-independent ‘basal’ GIRK currents (I_K,basal) were reduced with increasing expression of PTX-insensitive Gα subunits and were accompanied by a parallel rise in I_K,ACh. These effects together are indicative of increased Gβγ scavenging by the expressed Gα subunit and the subsequent formation of functionally coupled m2 receptor-G protein heterotrimers (Gα_(GDP)βγ). Co-expression of RGS4 accelerated all the PTX-insensitive Gα_i/o-coupled GIRK currents to a similar extent, yet reduced I_K,ACh amplitudes 60-90 % under conditions of low Gα_i/o coupling. Kinetic analysis indicated the RGS4-dependent reduction in steady-state GIRK current was fully explained by the accelerated deactivation rate. Thus kinetic inconsistencies associated with RGS4-accelerated GIRK currents occur at a critical threshold of G protein coupling. In contrast to RGS4, RGS7 selectively accelerated Gα_o-coupled GIRK currents. Co-expression of Gβ5, in addition to enhancing the kinetic effects of RGS7, caused a significant reduction (70-85 %) in steady-state GIRK currents indicating RGS7-Gβ5 complexes disrupt Gα_o coupling. Altogether these results provide further evidence for a GPCR-Gαβγ-GIRK signalling complex that is revealed by the modulatory affects of RGS proteins on GIRK channel gating. Our functional experiments demonstrate that the formation of this signalling complex is markedly dependent on the concentration and composition of G protein-RGS complexes.

G protein-gated inwardly rectifying K⁺ (GIRK) channels mediate slow inhibitory post-synaptic potentials (sIPSPs) in the central and peripheral nervous system and are activated by a variety of G protein-coupled receptors (GPCRs) that selectively interact with pertussis toxin (PTX)-sensitive Gα_i/o proteins (for recent reviews see (Dascal, 1997; Yamada et al. 1998; Mark & Herlitze, 2000). Both the amplitude and the distinguishing slow time course for GIRK-mediated sIPSPs are inherently coupled to the kinetics of the G protein activation-deactivation cycle. The time course for GIRK channel activation is dependent on the receptor catalysed production of free Gβγ dimers that interact directly with the channel and increase open probability (Logothetis et al. 1987; Krapivinsky et al. 1995). Deactivation of GIRK currents is determined by the reaction rate for Gβγ sequestration, a process rate limited by the GTPase activity of associated Gα_i/o proteins, Gα_(GTP) → Gα_(GDP), with a rate constant for GTP hydrolysis of k_GTPase (Breitwieser & Szabo, 1988).

Given the large diversity and ubiquitous expression of heterotrimeric G proteins (Simon et al. 1991), several studies have begun investigating whether functional differences exist among the various Gαβγ subunits that may participate in GPCR → GIRK channel signalling (5 Gα_i/o isoforms, 5 Gβ genes, and 12 Gγ genes) (Lledo et al. 1992; Wickman et al. 1994; Schreibmayer et al. 1996; Sowell et al. 1997; Takano et al. 1997; Valenzuela et al. 1997; Fernandez-Fernandez et al. 1999; Greif et al. 2000; Leaney et al. 2000; Leaney & Tinker, 2000; Lei et al. 2000; Blake et al. 2001; Fernandez-Fernandez et al. 2001). Adding to this molecular complexity is the recently identified ‘Regulators of G protein Signalling’ (RGSs) that accelerate the termination of G protein signalling by increasing k_GTPase through a direct interaction with Gα subunits (for recent reviews see De Vries et al. 2000; Ross & Wilkie, 2000). Several RGS proteins, of which there are now more than 20 mammalian genes identified, are capable of accelerating receptor-dependent GIRK activation and deactivation kinetics in heterologous expression systems (Doupnik et al. 1997; Saitoh et al. 1997; Granneman et al. 1998; Herlitze et al. 1999; Saitoh et al. 1999; Kovoor et al. 2000; Burgon et al. 2001). RGS-accelerated GIRK currents more closely resemble the properties of neuronal and cardiac GIRK currents, indicating native RGS proteins are likely to be important determinants of GIRK-mediated synaptic signalling.

Initial studies of RGS-accelerated GIRK channel kinetics revealed an apparent paradox where the RGS-accelerated activation phase, an expected consequence of accelerated deactivation according to standard kinetic concepts, was not accompanied with a reduction in steady-state current amplitude (Doupnik et al. 1997; Saitoh et al. 1997). These findings suggested that RGS proteins have actions on GPCR → GIRK signalling beyond the well-established GTPase-activating function of the RGS domain on Gα_(GTP) subunits (Zerangue & Jan, 1998). To further explore the effects of RGS proteins on receptor-dependent GIRK channel gating, we examined the actions of two distinct RGS proteins (RGS4 and RGS7) on GIRK channels (Kir3.1/Kir3.2a heteromers) activated by muscarinic m2 receptors coupled to varying concentrations and compositions of Gα_i/oβγ heterotrimers in Xenopus oocytes. Muscarinic m2 receptor-Gαβγ-GIRK channel coupling was established by expressing varying levels of PTX-insensitive Gα_i/o subunits (Wise et al. 1997) while uncoupling oocyte Gα_i/o subunits (xGα_i/o) with PTX. Our findings indicate that GPCR-G protein-RGS-GIRK precoupling is a critical determinant in the anomalous channel gating, and is dependent on the concentration and composition of Gα_i/o-RGS proteins.

Methods

Isolation of Xenopus oocytes

All procedures for the use and handling of Xenopus laevis (Xenopus One, Ann Arbor, MI, USA) were approved by the University of South Florida Institutional Animal Care and Use Committee in accordance with NIH guidelines. Oocytes were isolated from ovarian tissue surgically removed during hypothermia and 0.2% tricaine (MS-222)-induced anaesthesia. After recovery from anaesthesia, frogs were returned to their tanks in the institutional animal housing facility where they were monitored daily. Frogs were killed after a second procedure by exsanguination while under anaesthesia. The time between surgeries was 1-3 weeks.

The oocytes were enzymatically dissociated by a 50 min collagenase A (Boehringer Mannheim) digestion (1.8 %) at room temperature on a rocker platform in Ca²⁺-free oocyte Ringer (OR) solution. The OR solution was composed of 82.5 mmNaCl, 2.5 mm KCl, 1.0 mm CaCl₂, 1.0 mm MgCl₂, 1.0 mm NaHPO₄ and 5.0 mm Hepes, at pH 7.5 (NaOH). Isolated stage V-VI oocytes were then maintained in oocyte culture medium (OCM) at 19 °C in 35 mm dishes on an orbital shaker. OCM was composed of OR solution containing 2.5 mm sodium pyruvate and 5 % heat-inactivated horse serum. OCM was changed 1-2 times daily.

Heterologous expression in Xenopus oocytes

Linearized cDNA-containing vectors (Table 1) were used to transcribe cRNA in vitro using the appropriate RNA polymerase (T7 or T3) as described by the manufacturer (mMessage mMachine, Ambion, Austin, TX, USA). Concentrations and quality of the cRNAs were determined by spectrophotometric absorbance at 260 nm and denatured (formaldehyde) agarose gel electrophoresis. On the first day after enzymatic isolation (Day 1), oocytes were injected with a mixture of cRNAs dissolved in DEPC-treated H₂O at a final injection volume of 50 nl (Nanoliter2000, World Precision Instruments). All oocytes were injected with cRNAs for Kir3.1 (0.5 ng), Kir3.2a (0.5 ng), and the m2 receptor (0.5 ng). The amounts of other cRNAs, including PTX-insensitive Gα subunits and RGSs, were varied as described in the Results section. The PTX-insensitive Gα cDNAs were all constructed by PCR (kindly provided by Stephen Ikeda, NIAAA, Bethesda, MD, USA), having a 5′ Kozak sequence followed by the Gα coding region containing the C-terminal C→G mutation. The 5′ and 3′ untranslated regions of the Gα cDNAs were not included in the constructs, and therefore ribosomal binding and translation initiation of each PTX-insensitive Gα cRNA are expected to be equivalent.

Table 1.

Heterologously expressed proteins in Xenopus oocytes

cDNA	GenBank accession#	Cloning vector
Human muscarinic	X15264	pGEM3(Promega)
m2 receptor
Rat Kir3.1	U01071	pBS-MXT*
Mouse Kir3.2a	U37253	pBS-MXT*
Rat Gαi1(C351G)	M17527	pCI (Promega)
Rat Gαi2(C352G)	M17528	pCI (Promega)
Rat Gαi3(C353G)	M20713	pCI (Promega)
Mouse GβoA(C351G)	M36777	pCI (Promega)
Mouse GβoB(C351G)	M36778	pCI (Promega)
Bovine Gβ1	M13236	pBS-MXT*
Mouse Gβ5	U69145	pcDNA3 (Invitrogen)
Rat RGS4	U27767	pcDNA3.1(Invitrogen)
Bovine RGS7	AF011359	pcDNA3 (Invitrogen)
PTX-S1	M13223	pGEMHE*

^*The pBS-MXT and pGEMHE vectors are modified pBluescript II (KS) (Stratagene) and –3Z (promega) vectors, each containing 5’ and 3’ untranslated regions of the Xenopus β–globin gene that flank the cloned cDNA and enhance RNA stability in Xenopus oocytes (Liman et al. 1992).

Two methods were used to inactivate endogenous oocyte Gα_i/o subunits by PTX-mediated ADP ribosylation. Initially we injected the PTX holotoxin (containing S1, S2, S3, S4 and S5 subunits from Bordetella pertussis, Sigma-Aldrich Chemical) at 1 ng/oocyte on Day 2 of culture (1 day after cRNA injection), and then recorded GIRK currents on Day 4. This method inhibited > 80 % of the m2 receptor-activated GIRK currents coupled to endogenous Gα_i/o subunits that was determined for each batch of oocytes to monitor the efficiency of endogenous G protein uncoupling. In subsequent experiments we included cRNA (1 ng/oocyte) encoding the catalytically active PTX-S1 subunit (kindly provided by Eitan Reuveny, Weizmann Institute, Israel) with the mixture of other cRNAs in the initial Day 1 cRNA injection (Vivaudou et al. 1997). This produced a much more effective > 95 % uncoupling of endogenous Gα_i/o subunits which also was determined for each batch of oocytes tested.

Electrophysiological recordings

Macroscopic GIRK currents were measured using a two-electrode voltage clamp amplifier (GeneClamp 500, Axon Instruments) and standard recording techniques (Stuhmer & Parekh, 1995). Electrodes were fabricated from borosilicate glass tubes (1.5 outside diameter, 0.86 inside diameter, GC150F-10, Warner Instruments) by a programmable microelectrode puller (P-97, Sutter Instruments). Electrodes were filled with 3 m KCl and had tip resistances of 0.8-1.0 MΩ. Membrane currents from voltage clamped oocytes were digitized using a Digidata 1200 acquisition system (Axon Instruments) and a Dell PC computer running pCLAMP 7.0 software (Axon Instruments).

Oocytes were placed in a recording chamber continuously perfused with a minimal Ringer solution composed of 98 mmNaCl, 1 mm MgCl₂ and 5 mm Hepes at pH 7.5 (NaOH). After electrode impalement and clamping the membrane potential to −80 mV, the perfusion solution was changed to a high K⁺ solution composed of 20 mm KCl, 78 mmNaCl, 1 mm MgCl₂, and 5 mm Hepes at pH 7.5 (NaOH). The resulting increase in inward current represents a ‘basal’ K⁺ current (I_K,basal) that is due primarily to receptor-independent GIRK channel activity (Dascal et al. 1993). Rapid application and washout of acetylcholine (ACh, Sigma-Aldrich Chemical) produced the receptor-dependent GIRK current (I_K,ACh), and was performed with a computer controlled perfusion system (SF-77B, Warner Instruments) that rapidly switched the position of two perfusion barrels (barrel A and barrel B) located next to the oocyte. The perfusion barrels contained high K⁺ solution (barrel A) and high K⁺ solution plus ACh (barrel B). For barrel B, a range of ACh concentrations was tested for each oocyte via a manifold that connected multiple reservoirs containing different ACh concentrations. Flow through the perfusion barrels was gravity driven, and the time constant for solution exchange was ≈1 s as determined by the time course change in receptor-independent GIRK current with switching between 20 and 40 mm external K⁺. All recordings were performed at room temperature (21-23 °C).

Electrophysiological data analysis

Time-dependent GIRK current kinetics were analysed using non-linear curve fitting software that fit single exponential functions to derive activation time constants (τ_act) and deactivation time constants (τ_deact) (Clampfit software, Axon Instruments). Dose-response relations were analysed by fitting peak GIRK current amplitudes with the following Hill function:

where the effective concentration producing a 50 % response (EC₅₀) and Hill coefficient value (n_H) were derived from the best fit (Origin 6.0 software, OriginLab Corp., Northampton, MA, USA). In some cases where the GIRK current did not reach a steady-state level at the end of the agonist application period (i.e. low agonist concentrations), the steady-state current amplitude was estimated by the fit of the activation time course to an extrapolated steady-state value.

Statistical comparisons between the various experimental groups were performed by one-way ANOVA where P < 0.05 was considered significant. Experiments were each replicated in oocytes from 2-4 separate batches (dissections) of oocytes.

Radiolabelling, immunoprecipitation and Western blotting of Gα_i/o proteins

To compare the relative expression levels of PTX-insensitive Gα_i/o proteins, [³⁵S]Met/Cys (Pro-mix, Amersham Pharmacia, Piscataway, NJ, USA) was added to OCM (0.5 mCi ml⁻¹) 1-2 h after cRNA injection to initiate radiolabelling of proteins. Fresh OCM containing [³⁵S]Met/Cys was added on a daily basis, and 3 days after cRNA injection ≈20 oocytes from each experimental group were washed with label-free OR solution and homogenized in the following lysis buffer (150 mmNaCl, 50 mm Tris-Cl, 1 mm dithiothreitol (DTT), 1.0 % Trition X-100, and protease inhibitor cocktail (Boehringer Mannheim), pH 7.5) at 50 μl/oocyte. The oocyte lysate was then centrifuged at 10 000 g for 10 min to clear insoluble debris.

Gα_i/o subunits were immunoprecipitated from oocyte lysates using a rabbit polyclonal antibody that recognizes an internal epitope (GAGESGKSTIVKQMK) identical among Gα_i/o isoforms (SA-126, BIOMOL Research Laboratories, Plymouth Meeting, PA, USA). To expose the internal epitope, the supernatant from the equivalent of five oocytes (250 μl) was diluted in SDS denaturing buffer (50 mm NaPO₄, 2 mm EDTA, 1 mm DTT, 0.5 % SDS, pH 8.0) and boiled for 3 min. The denaturated supernatants were then diluted and solubilized further in RIPA buffer (150 mmNaCl, 50 mm NaPO₄, 2 mm EDTA, 1 mm DTT, 1.0 % Triton X-100, 1.0 % deoxycholate, 0.5 % SDS, pH 7.2). Denatured oocyte lysates (≈700 μl each) were then pre-cleared with 10 μl of non-immune rabbit serum (1 h incubation at 4 °C) and 20 μl Protein A/G-agarose beads (Santa Cruz Biotechnology, Inc.). Gα_i/o proteins were immunoprecipitated in an overnight incubation at 4 °C using the SA-126 antibody precoupled to Protein A/G-agarose beads. We initially tested different amounts of oocyte lysate (0.5, 1, 2 and 4 oocytes) with a fixed amount of SA-126 antibody (2 μl) and Protein A/G-agarose (20 μl) to establish non-saturating binding conditions, which was determined to be ≤ 1 oocyte. All immunoprecipitations were then performed using the lysate equivalent of a single oocyte. At the end of the overnight incubation period, beads were washed three times with lysis buffer and the immunoprecipitated proteins eluted by boiling in 40 μl of SDS loading buffer (62.5 mm Tris-Cl pH 6.8, 10 % glycerol, 5 % β-mercaptoethanol, 2 % SDS, 0.05 % bromophenol blue).

Immunoprecipitated proteins were separated by SDS-polyacrylamide (8 %) gel electrophoresis and then transferred to a polyvinylidene difluoride (PVDF) membrane by overnight electrophoretic transfer at 4 °C. ³⁵S-labelled proteins were resolved by autoradiography. Western blot analysis was performed using a polyclonal goat antibody that recognizes a conserved region in Gα_i/o/t/z proteins (sc-12798, Santa Cruz Biotechnology). A ≈41 kDa band corresponding to Gα_i/o proteins was readily detected by an HRP-conjugated donkey anti-goat secondary antibody (sc-2033, Santa Cruz Biotechnology) and enhanced chemiluminescence. Relative radiolabelling and Gα_i/o protein levels were quantified by densitometric analysis (GS-700 Imaging Densitometer, Bio-Rad) of the 41 kDa band following normalization to a sample control.

Results

Conferring specific m2 receptor-Gα-coupling in Xenopus oocytes

To assess the properties of receptor-dependent GIRK channel gating elicited by specific Gα-coupled m2 receptors, oocyte Gα_i/o subunits were inactivated by PTX while expressing one of five different PTX-insensitive Gα_i/o subunits (Gα_i1(C351G), Gα_i2(C352G), Gα_i3(C351G), Gα_oA(C351G), or Gα_oB(C351G)) (Wise et al. 1997). As reported previously, expression of PTX-S1 blocked > 95 % of I_K,ACh and effectively abolished endogenous Gα_i/o protein coupling (Vivaudou et al. 1997). Shown in Fig. 1, co-expression of Gα_i2(C352G) rescued the m2 receptor-coupled GIRK currents and was dependent on the level of Gα_i2(C352G) expression (Fig. 1B). Injection of 5 or 10 ng of Gα_i2(C352G) cRNA per oocyte produced ACh-evoked GIRK currents that were ≈50 and 95 %, respectively, the peak amplitude of GIRK currents recorded from paired oocytes utilizing endogenous Gα_i/o subunits (i.e. no PTX-S1 or Gα_i2(C352G) cRNA). The Gα_i2(C352G)-dependent GIRK currents represent channels activated by m2 receptors coupled specifically to Gα_i2(C352G) subunits and endogenous Gβγ dimers. Gα_i2(C352G)-dependent I_K,ACh was also associated with a parallel 50-60 % decrease in I_K,basal (Fig. 1D). The reduction in I_K,basal by Gα_i2(C352G) is likely due to the sequestration of free Gβγ dimers that mediate receptor-independent basal GIRK channel activity (Lim et al. 1995; Vivaudou et al. 1997; Jeong & Ikeda, 1999).

Figure 1. Specific coupling of Gα_i2(C352G) to m2 receptors and GIRK channel activation in Xenopus oocytes.

A, typical ACh-evoked GIRK currents (I_K,ACh) recorded from different oocytes from three separate experimental groups. Upper traces: I_K,ACh from a ‘control’ oocyte expressing muscarinic m2 receptors and Kir3.1/Kir3.2a channel subunits, utilizing endogenous Gα_i/o proteins for receptor activation. Middle trace: co-expression of PTX-S1 (1 ng cRNA/oocyte) effectively uncouples ACh-evoked GIRK currents utilizing oocyte Gα_i/o proteins. Lower traces: expression of the PTX-insensitive Gα_i2(C352G) subunit (5 ng cRNA) with PTX-S1 rescues m2 receptor-coupled GIRK currents. All GIRK currents were elicited by a 25 s application of different concentrations of ACh as indicated. Bottom traces: superimposed I_K,ACh elicited by 1 μm ACh from the control oocyte (grey trace) and the Gα_i2(C352G)-coupled oocyte (black trace). Peak amplitudes are normalized to illustrate the kinetic differences in the activation and deactivation time courses. B, ACh dose-response relations for GIRK activation via m2 receptors coupled to oocyte Gα_i/o subunits (○, control) and Gα_i2(C352G) at different levels of expression (▪ 1 ng, ▴ 5 ng, and • 10 ng cRNA/oocyte). C, ACh-dose-response curves from B normalized to maximal I_K,ACh. D, comparison of receptor-independent basal GIRK currents (I_K,basal) with varying levels of Gα_i2(C352G) expression. I_K,basal is expressed as the percentage change in the ‘control group’ mean value determined for each batch of oocytes. E, activation time constants (τ_act) and F, deactivation time constants (τ_deact) for GIRK currents coupled to varying levels of Gα_i2(C352G) expression and different concentrations of ACh. ○ control; ▴ 5 ng; and • 10 ng of Gα_i2(C352G) cRNA/oocyte. Data in B-F represent the means ±s.e.m. from at least 3 batches of oocytes with the number of oocytes indicated. * P < 0.05.

ACh dose-response curves derived from oocytes expressing different levels of Gα_i2(C352G) were not significantly different with respect to their EC₅₀ value as shown in Fig. 1C, and this is consistent with previous studies where EC₅₀ values were shown to be relatively unaffected by overexpressed Gα subunits, but most sensitive to levels of GPCR expression (Henry et al. 1995). Since the same amount of m2 receptor cRNA was injected for all experimental groups (0.5 ng cRNA/oocyte), m2 receptor protein levels were assumed to be roughly equivalent across groups and this was supported by the equivalent EC₅₀ values with varying Gα_i2(C352G) expression levels. The ACh dose-response curves for Gα_i2(C352G)-coupled m2 receptors, however, were notably shifted rightward compared to the ACh dose-response relation from m2 receptor coupling to endogenous Gα_i/o subunits (Fig. 1B). The higher EC₅₀ value for Gα_i2(C352G)versus endogenous Gα_i/o-coupled receptors may be attributable to the mixed coupling of oocyte Gα_o and Gα_i subunits, and/or a lower level of m2 receptors precoupled to heterotrimeric G proteins (Shea & Linderman, 1997; Shea et al. 2000) (see Results below and Discussion).

One readily apparent feature of the m2 receptor- Gα_i2(C352G)-coupled GIRK current was the significantly slower activation and deactivation kinetics as compared to GIRK currents coupled to endogenous xGα_i/o subunits (Fig. 1A). The time constants for I_K,ACh activation and deactivation over a range of ACh concentrations are shown in Fig. 1E and F. The slower kinetics for m2 receptor-Gα_i2(C352G)-coupled GIRK currents were not attributable to the level of Gα_i2(C352G) expression since increasing amounts of Gα_i2(C352G) cRNA, which nearly doubled maximal I_K,ACh amplitudes, did not result in significantly different GIRK activation or deactivation time constants. The slower GIRK gating properties may be caused by altered coupling properties of the C→G mutated Gα_i2 subunit, or alternatively the faster properties in the absence of overexpressed Gα subunits may reflect the influence of endogenous RGS proteins (Saitoh et al. 2000).

GIRK channel gating properties via Gα_o- and Gα_i-coupled m2 receptors

The gating properties of GIRK currents elicited by m2 receptors coupled to each of the PTX-insensitive Gα_i/o subunits are shown in Fig. 2. Since there were no noticeable differences in GIRK current kinetics at different levels of Gα_i2(C352G) expression, comparisons between the five different Gα_i/o isoforms were initially made using 5 ng cRNA/oocyte for each Gα isoform. The three Gα_i isoforms (Gα_i1(C351G), Gα_i2(C352G) and Gα_i3(C351G)) each produced ACh-elicited GIRK currents that were indistinguishable from each other with regard to the five parameters analysed (maximal I_K,ACh, I_K,basal, EC₅₀, τ_act and τ_deact) (Fig. 2). Gα_oA(C351G) and Gα_oB(C351G) also produced equivalent GIRK responses but were distinct from Gα_i-coupled GIRK currents in several respects. First, Gα_o-coupled GIRK currents were significantly larger than Gα_i-coupled currents, with maximal I_K,ACh amplitudes being ≈2-fold greater (Fig. 2B). Second, expression of Gα_oA(C351G) and Gα_oB(C351G) more prominently inhibited I_K,basal compared to Gα_i1(C351G), Gα_i2(C352G), or Gα_i3(C351G) (Fig. 2B). Third, the EC₅₀ values for Gα_oA(C351G) and Gα_oB(C351G) were more similar to the EC₅₀ values with endogenous xGα_i/o coupling than with Gα_i1(C351G), Gα_i2(C352G), or Gα_i3(C351G) (Fig. 2C). And finally, the activation of Gα_o-coupled GIRK currents was faster compared to Gα_i-coupled currents and had a more prominent sigmoidal time course (Fig. 2D). The deactivation time course for I_K,ACh was not significantly different among all five Gα_i/o isoforms (Fig. 2D).

Figure 2. Properties of ACh-evoked GIRK currents activated by m2 receptors coupled to five different Gα_i/o subunits.

A, typical ACh-evoked GIRK currents for each PTX-insensitive Gα_i/o subunit examined (Gα_i1(C351G), Gα_i2(C352G), Gα_i3(C351G), Gα_oA(C351G) and Gα_oB(C351G)). Currents were elicited by a 25 s application of 1 μm ACh. B, maximal amplitude of receptor-dependent GIRK currents (I_K,ACh, grey bars) and receptor-independent GIRK currents (I_K,basal, black bars) for each expressed PTX-insensitive Gα subunit (Gα*). Gα* expression was produced by 5 ng cRNA/oocyte. Maximal I_K,ACh responses are to 10 μm ACh. C, EC₅₀ values for control (open bar) and each PTX-insensitive Gα* subunit, derived from ACh dose-response relations for I_K,ACh activation. D, activation (black bars) and deactivation time constants (grey bars) derived from exponential fits of the I_K,ACh time course in response to rapid ACh application and washout. For B-D, data represent the means + s.e.m. from at least 3 batches of oocytes with the number of oocytes indicated (* P < 0.05). In B and D, statistical comparisons were with Gα_i1(C351G), and in C comparisons were with oocyte Gα_i/o coupling (open bar).

The differences between Gα_i and Gα_o-coupled GIRK currents were not attributable to cDNA/vector construction (i.e. untranslated regions), cRNA quantification, or batch-to-batch oocyte variability (see Methods). Several factors could conceivably contribute to these differences and include (1) differences in Gα_i/o biosynthesis or degradation (Li et al. 1996), (2) preferential Gα_o coupling to m2 receptors (Leaney & Tinker, 2000), (3) preferential Gα_o association with oocyte Gβγ subunits (Fernandez-Fernandez et al. 2001), and (4) different Gα_i/o subcellular distributions (Devic et al. 1996).

Biochemical analysis of PTX-insensitive Gα_i/oprotein expression levels

To determine whether PTX-insensitive Gα_i/o subunits are differentially expressed following cRNA injection (5 ng cRNA/oocyte, as in Fig. 2), Gα_i/o proteins were immunoprecipitated from oocytes following metabolic labelling with [³⁵S]Met/Cys. The PTX-insensitive Gα_i/o proteins each contain an equivalent number of Met/Cys sites (19) for potential radiolabelling. As shown in Fig. 3A, a prominent ≈41 kDa band that corresponds to Gα_i/o proteins was radiolabelled in each experimental group. A significant portion represents radiolabelling of endogenous xGα_i/o proteins over the 3 day incubation period as indicated by the oocyte groups that were not injected with PTX-insensitive Gα_i/o cRNA. Quantitative comparisons of the ³⁵S-labelled 41 kDa band via densitometric analysis indicate a small 10-20 % elevation in oocytes injected with PTX-insensitive Gα_i/o cRNAs (Fig. 3B). The observed changes in ³⁵S labelling were also paralleled by densitometric changes in the bands detected by Western blot analysis (Fig. 3B). Since the peak I_K,ACh amplitudes for Gα_o-coupled m2 receptors were ≈2 times higher compared to Gα_i-coupled m2 receptors under equivalent expression conditions (cf. Fig. 2B), the ≈2-fold difference in Gα_oB(C351G)versus Gα_i2(C352G) or Gα_i3(C351G) protein expression could potentially account for the differences in m2 receptor-GIRK channel coupling. Gα_oA(C351G)-elevated protein levels, however, were similar to Gα_i2(C352G) and Gα_i3(C351G) levels (Fig. 3B), and resolving small expression changes was clearly limited by the biochemical approach. Thus despite equivalent cRNA injections (5 ng cRNA/oocyte each), differences in Gα_i/o biosynthesis and/or degradation may result in different steady-state Gα_i/o protein levels that account for the distinct gating properties of Gα_o- versus Gα_i-coupled GIRK currents.

Figure 3. Biochemical analysis of PTX-insensitive Gα_i/o protein levels in Xenopus oocytes.

A, radiolabelling (upper panel) and Western blot analysis (lower panel) of oocytes injected with cRNAs encoding five different PTX-insensitive Gα_i/o subunits (*) as described in Methods and Fig. 2. Oocytes were incubated for 3 days in OCM containing 0.5 mCi ml⁻¹ [³⁵S]Met/Cys. All groups were injected with cRNAs for the m2 receptor and GIRK channel subunits Kir3.1 and Kir3.2a (0.5 ng each/oocyte). PTX-insensitive Gα_i/o cRNAs (5 ng/oocyte) were injected with PTX-S1 cRNA (1 ng/oocyte). Both endogenous and heterologously expressed Gα proteins were immunoprecipitated from the lysate equivalent of one oocyte using a ‘common’ Gα antibody, then separated by SDS-PAGE and transferred to a PVDF membrane for autoradiography and Western blotting. The 41 kDa band corresponding to Gα_i/o proteins is indicated. B, quantitative analysis of Gα_i/o proteins detected by radiolabelling and Western blot analysis. The 41 kDa band from the control group (no PTX-insensitive Gα_i/o cRNA) served as an internal reference and was used to normalize the band intensity among the different experimental groups in each autoradiogram and Western blot. The Western blot results are the means + s.e.m. obtained from 3 independent experiments (separate batches of injected oocytes that were immunoprecipitated and immunostained as described in Methods). The radiolabelling data are the means + s.e.m. obtained from 2 of the Western blot experiments. * P < 0.05.

The GIRK activation time course is dependent on free Gβγ levels

A sigmoidal time course for GIRK activation, as observed with Gα_o expression (Fig. 2), is well documented in cardiomyocytes and neurons and indicative of a multi-step process in channel opening (Inomata et al. 1989; Sodickson & Bean, 1996). Tetrameric GIRK channels can bind up to four Gβγ dimers, and functional studies indicate at least two Gβγ dimers are necessary for channel opening (Nemec et al. 1999; Corey & Clapham, 2001). We questioned whether the sigmoidal GIRK activation time course via Gα_o-coupled receptors was a consequence of a markedly reduced free Gβγ concentration as indicated by the low I_K,basal level. As illustrated in Fig. 4A, conditions favouring high free Gβγ concentrations (high I_K,basal) would yield a single exponential time course for receptor activation, where the binding of a single Gβγ dimer leads to channel opening. At low free Gβγ concentrations (i.e. low I_K,basal), GIRK channels would instead be occupied by fewer Gβγ dimers (less than 3 Gβγ) and require multiple Gβγ binding events during receptor activation (2-4 Gβγ). According to this hypothesis, reducing Gα_o expression levels should increase I_K,basal and promote a single exponential time course for I_K,ACh activation.

Figure 4. Activation kinetics of Gα_o-coupled GIRK currents at different levels of Gα_oB(C352G) expression.

A, kinetic scheme for GIRK channel activation via Gβγ binding to each of the four GIRK channel subunits. Four closed states are depicted with 0 to 3 bound Gβγ dimers. Channel opening (O) occurs with the binding of 4 Gβγ dimers. The level of free Gβγ dimers then determines the channel-bound state that precedes receptor activation. B, left panel: ACh-evoked GIRK currents from oocytes injected with either 1 or 5 ng Gα_oB(C351G) cRNA. Right panel: the activation phase for the 1 μm ACh-evoked GIRK currents are displayed at higher temporal resolution. Red lines represent non-linear fits to the data. A single exponential function (G1) best described the I_K,ACh time course with low Gα_oB(C351G) expression (1 ng RNA/oocyte), having a time constant of 8.1 s. A third-order exponential function (G3) best described the activation time course at high Gα_oB(C351G) expression (5 ng RNA/oocyte), having a time constant of 4.4 s. C, ACh dose-response curves for m2 receptor-activated GIRK currents from control oocytes (○) and oocytes expressing Gα_oB(C351G) at two different levels (▪ 1 ng and ▴ 5 ng cRNA/oocyte). D, effects of Gα_oB(C351G) expression levels on receptor-independent GIRK channel activity (I_K,basal). The open bar is from control oocyte responses, and black bars are from oocytes injected with varying amount of Gα_oB(C351G) cRNA and PTX-S1 cRNA (1 ng/oocyte). Data represent the means + s.e.m. from 4 batches of oocytes with the number of oocytes indicated.

To test this, we compared the activation time course of I_K,ACh at two different levels of Gα_oB(C351G) expression (Fig. 4). As expected, lower expression of Gα_oB(C351G) (1 ng Gα_oB(C351G) cRNA/oocyte) produced smaller I_K,ACh amplitudes compared to high Gα_oB(C351G) expression, and had significantly higher receptor-independent basal GIRK activity (Fig. 4C and D). The activation time course for I_K,ACh was clearly affected by the level of Gα_oB(C351G) expression and was consistent with the gating scheme shown in Fig. 4A. Low Gα_oB(C351G) expression yielded GIRK currents whose activation time course was well described by a single exponential function. In contrast, high Gα_oB(C351G) expression yielded GIRK currents whose activation time course had a significant delay and were best described by a 3rd order exponential function, suggesting the binding of three Gβγ subunits in receptor-dependent GIRK activation (see Fig. 4A). The τ_act was smaller for high Gα_oB(C351G) expression versus low Gα_oB(C351G) expression, and single exponential fits to the rising phase of the sigmoidal time course underestimate the activation rate constant (1/τ_act) under these conditions. Comparisons of τ_deact and EC₅₀ values from ACh dose-response curves indicated the these properties were not significantly affected by these levels of Gα_oB(C351G) expression (data not shown).

RGS4 reduces GIRK currents at low Gα_i/o coupling levels

We next evaluated the effects of RGS4 at saturating expression levels (10 ng RGS4 cRNA/oocyte; Doupnik et al. 1997;

Keren-Raifman et al. 2001) on both Gα_o- and Gα_i-coupled m2 receptors as described in Fig. 2. As shown in Fig. 5, RGS4 accelerated the receptor-dependent GIRK activation and deactivation kinetics for each Gα_i/o isoform, consistent with the non-selective GTPase activating properties of RGS4 on Gα_i/o subunits in vitro (Berman et al. 1996; Watson et al. 1996). Interestingly, however, RGS4 suppressed 60-70 % of the I_K,ACh amplitude for Gα_i-coupled m2 receptors, yet had no effect on the peak I_K,ACh amplitude of Gα_o-coupled receptors. RGS4 did not significantly affect the EC₅₀ for channel activation via Gα_i-coupled receptors, but did cause a small shift the EC₅₀ value for Gα_oA(C351G)-coupled GIRK currents towards a higher ACh concentration, which more closely corresponded to the EC₅₀ values of Gα_i-coupled GIRK currents (Fig. 5F). Kinetic analysis indicated the RGS4-mediated reduction in Gα_i-coupled I_K,ACh (Fig. 5E) was fully explained by the RGS4-accelerated GIRK deactivation rate (see Table 2 and Discussion).

Figure 5. Effects of RGS4 on GIRK currents coupled to specific PTX-insensitive Gα_i/o subunits.

All data are from oocytes injected with 5 ng Gα* cRNA/oocyte, with or without 10 ng RGS4 cRNA/oocyte. Endogenous PTX-sensitive Gα_i/o subunits were inactivated by either PTX injection or PTX-S1 expression as described in Methods. Data are means and s.e.m.; * P < 0.05. A, representative ACh-evoked GIRK currents from oocytes expressing either Gα_i2(C352G) (upper traces) or Gα_oA(C351G) (lower traces) with and without RGS4 coexpression. Scale bars indicate 1 μA and 10 s. B, ACh dose-response relations for m2 receptor-coupled GIRK currents from oocytes expressing Gα_i2(C352G) in the absence (▴) or presence of RGS4 (▿). C, I_K,ACh activation time constants (τ_act) from oocytes expressing PTX-insensitive Gα_i/o subunits alone (black bars) or with co-expressed RGS4 (grey bars). Time constants were derived from a single exponential fit to the rising phase of I_K,ACh elicited by 1 μm ACh. D, deactivation time constants (τ_deact) derived from ACh-elicited GIRK currents recorded from oocytes expressing PTX-insensitive Gα_i/o subunits alone (black bars) or with co-expressed RGS4 (grey bars). Time constants were derived from a single exponential fit to the decay of I_K,ACh after rapid washout of 1 μm ACh. E, effects of RGS4 on maximal I_K,ACh responses elicited by m2 receptors coupled to PTX-insensitive Gα_i/o subunits. I_K,ACh amplitudes (10 μm ACh) from RGS4-expressing oocytes are expressed as a percentage of the mean I_K,ACh amplitude from oocytes expressing the PTX-insensitive Gα_i/o subunit alone (RGS-). F, effects of RGS4 on the EC₅₀ for ACh activation of GIRK currents coupled to m2 receptors and PTX-insensitive Gα_i/o subunits. Black bars are without RGS4 expression (-RGS4), grey bars are with RGS4 co-expression (+RGS4).

Table 2.

‘Expected’ versus ‘Observed’ gating parameters for RGS-accelerated GIRK currents

Expressed proteins	Observed Tdeact (S)	Expected Mean Tact (S)	Observed Tact (S)	Expressed amplituted (% of control)	Observed amplituted (% of control)
Gαi1*	17.6 ± 4.6	11.5	11.2 ± 3.8	—	—
Gαi1*+ RGS4	8.1 ± 5.0	6.5	5.3 ± 2.2	56 ± 28	36 ± 5
Gαi1*+ RGS7	12.5 ± 6.1	9.1	8.8 ± 2.9	79 ± 25	74 ± 6
Gαi1*+ RGS7 + Gβ1	17.5 ± 2.9	11.5	13.0 ± 2.6	—	—
Gαi1*+ RGS7 + Gβ5	13.9 ± 2.1	9.8	11.4 ± 2.5	85 ± 9	88 ± 17
Gαi2*	23.8 ± 7.1	13.9	15.7 ± 7.1	—	—
Gαi2*+ RGS4	9.6 ± 6.3	7.5	6.1 ± 2.1	54 ± 23	33 ± 6
Gαi2*+ RGS7	17.1 ± 6.4	11.3	10.8 ± 4.3	81 ± 18	90 ± 8
Gαi2*+ RGS7 + Gβ1	20.3 ± 3.6	12.6	14.8 ± 2.9	—	—
Gαi2*+ RGS7 + Gβ5	18.6 ± 2.9	11.9	15.7 ± 7.5	95 ± 9	100 ± 18
Gαi3*	15.4 ± 3.4	10.5	14.1 ± 5.2	—	—
Gαi3*+ RGS4	10.0 ± 5.3	7.7	5.7 ± 2.0	73 ± 27	37 ± 3
Gαi3*+ RGS7	13.5 ± 4.4	9.6	10.5 ± 3.6	91 ± 20	61 ± 4
GαoA*	16.7 ± 2.8	11.1	6.8 ± 4.1	—	—
GαoA*+ RGS4	5.7 ± 1.9	4.9	4.2 ± 1.9	44 ± 12	105 ± 11†
GαoA*+ RGS7	3.3 ± 1.7	3.0	3.1 ± 1.8	27 ± 12	79 ± 8†
GαoA*+ RGS7 + Gβ1	6.3 ± 2.2	5.3	2.9 ± 1.5	—	—
GαoA*+ RGS7 + Gβ5	2.2 ± 1.3	2.0	0.9 ± 0.5	39 ± 21	17 ± 2
Gαob*	14.8 ± 4.9	10.3	6.0 ± 3.4	—	—
Gαob*+ RGS4	6.1 ± 2.9	5.2	5.0 ± 1.4	50 ± 19	105 ± 9†
Gαob*+ RGS7	4.0 ± 1.5	3.6	3.1 ± 0.7	35 ± 11	91 ± 13†
Gαob*+ RGS7 + Gβ1	2.7 ± 1.0	2.5	2.4 ± 0.7	—	—
Gαob*+ RGS7 + Gβ5	1.4 ± 0.9	1.3	1.7 ± 1.0	54 ± 32	32 ± 3

^†denotes deviation from expected steady—state GIRK current amplitude.

To determine whether the higher degree of Gα_oversus Gα_i coupling (cf. Fig. 2) could account for the RGS4-dependent effects on steady-state GIRK currents, we examined the effects of RGS4 at a lower level of Gα_oA expression (0.5 ng cRNA/oocyte). At low Gα_oA expression levels, RGS4 significantly reduced ≈90 % of both I_K,ACh (Gα_oA alone, 0.76 ± 0.11 μA, n = 17; Gα_oA + RGS4, 0.07 ± 0.03 μA, n = 9; P < 0.05) and I_K,basal (Gα_oA alone, 2.34 ± 0.18 μA, n = 19; Gα_oA + RGS4, 0.25 ± 0.07 μA, n = 14; P < 0.05). Thus the effect of RGS4 on steady-state GIRK current amplitudes is dependent on the level of Gα_i/o expression and GPCR precoupling. At high levels of m2 receptor-Gα_o coupling (i.e. Fig. 5), RGS4 accelerates GIRK activation and deactivation without significantly reducing steady-state GIRK current amplitudes and resembles the effects of RGS4 with endogenous xGα_i/o coupling (Doupnik et al. 1997).

RGS7 selectively accelerates Gα_o-coupled GIRK currents

RGS7 belongs to a distinct subfamily of RGS proteins that contain a Gγ-like (GGL) domain that specifically binds Gβ5 subunits (Snow et al. 1998; Sondek & Siderovski, 2001). RGS7 has relatively weak effects on GIRK current kinetics when expressed without Gβ5 in Xenopus oocytes (Saitoh et al. 1999), but is significantly enhanced by Gβ5 co-expression (Kovoor et al. 2000; Keren-Raifman et al. 2001). We examined whether RGS7 (10 ng cRNA/oocyte), with or without co-expressed Gβ5 subunits, differentially affected Gα_o- versus Gα_i-coupled GIRK currents. In the absence of coexpressed Gβ5, RGS7 preferentially accelerated Gα_o-coupled GIRK currents having little effect on Gα_i-coupled GIRK kinetics (Fig. 6), which is consistent with other studies indicating RGS7 selectively interacts with Gα_o subunits (Posner et al. 1999; Lan et al. 2000; Rose et al. 2000). Interestingly, RGS7 acceleration of Gα_o-coupled I_K,ACh was greater than that caused by RGS4 (cf. Fig. 5C and D) and was evident for both the I_K,ACh activation and deactivation time course as well as the desensitization rate. As observed with RGS4, RGS7 did not cause a significant reduction in the maximal Gα_o-coupled GIRK current despite accelerating the GIRK channel gating kinetics (Fig. 6E), but did cause a small shift in the ACh dose-response curve (Fig. 6B and F). Yet in contrast to RGS4, RGS7 did not reduce steady-state GIRK currents when Gα_oA expression was markedly reduced (0.5 ng cRNA/oocyte): I_K,basal (Gα_oA alone, 2.34 ± 0.18 μA, n = 19; Gα_oA + RGS7, 2.18 ± 0.20 μA, n = 18) and I_K,ACh (Gα_oA alone, 0.76 ± 0.11 μA, n = 17; Gα_oA + RGS7, 1.25 ± 0.21 μA, n = 17).

Figure 6. Effects of RGS7 on GIRK currents coupled to specific PTX-insensitive Gα_i/o subunits.

All data are from oocytes injected with 5 ng Gα* cRNA/oocyte, with or without 10 ng RGS7 cRNA/oocyte. Endogenous PTX-sensitive Gα_i/o subunits were inactivated by either PTX injection or PTX-S1 expression as described in Methods. Data are means and s.e.m.; * P < 0.05. A, representative ACh-evoked GIRK currents from oocytes expressing either Gα_i2(C352G) (upper traces) or Gα_oA(C351G) (lower traces) with and without RGS7 coexpression. Scale bars indicate 1 μA and 10 s. B, ACh dose-response relations for m2 receptor-coupled GIRK currents from oocytes expressing Gα_i2(C352G) in the absence (▴) or presence of RGS7 (▿). C, I_K,ACh activation time constants (τ_act) from oocytes expressing PTX-insensitive Gα_i/o subunits alone (black bars) or with co-expressed RGS7 (grey bars). Time constants are from GIRK currents evoked by 1 μm ACh. D, deactivation time constants (τ_deact) from ACh-elicited GIRK currents recorded from oocytes expressing PTX-insensitive Gα_i/o subunits alone (black bars) or with co-expressed RGS7 (grey bars). Time constants were derived after rapid washout of 1 μm ACh. E, effects of RGS7 on maximal I_K,ACh responses elicited by m2 receptors coupled to PTX-insensitive Gα_i/o subunits. I_K,ACh amplitudes (10 μm ACh) from RGS4-expressing oocytes as a percentage of the mean I_K,ACh amplitude from oocytes expressing the PTX-insensitive Gα_i/o subunit alone (no RGS7). F, effects of RGS7 on the EC₅₀ for ACh activation of GIRK currents coupled to m2 receptors and PTX-insensitive Gα_i/o subunits. Black bars are without RGS7 expression, grey bars are with RGS7 co-expression.

RGS7-Gβ5 disrupts Gα_o-coupled GIRK currents

Gβ5 binds with high affinity to the GGL domain of RGS7, enhancing the GAP activity and protein stability of RGS7 (Levay et al. 1999; Kovoor et al. 2000; Keren-Raifman et al. 2001). We examined the effects of Gβ5 expression on the RGS7- mediated regulation of both Gα_o- and Gα_i-coupled GIRK currents, with comparisons made using Gβ1 as a negative control which does not bind to RGS7 (Levay et al. 1999; Kovoor et al. 2000). As shown in Fig. 7, co-expression of Gβ5 with RGS7 caused a significant reduction (70-80 %) in Gα_o-coupled GIRK current amplitudes, yet had no effect on Gα_i-coupled currents. The lack of effect of Gβ5 on Gα_i-coupled GIRK currents indicates the level of Gβ5 (or Gβ5γ formation) does not cause a direct inhibition of GIRK currents as observed in transfected HEK-293 cells (Lei et al. 2000). In addition to reducing Gα_o-coupled GIRK currents, Gβ5 enhanced the RGS7-accelerated GIRK kinetics consistent with RGS7-Gβ5 complexes having greater GAP activity compared to RGS7 alone (Kovoor et al. 2000; Keren-Raifman et al. 2001). Thus the formation of RGS7-Gβ5 complexes apparently disrupts Gα_o coupling (either to m2 receptors and/or GIRK channels) and is supported by the effects of Gβ5 on RGS7-Gα_o interactions in vitro (Levay et al. 1999).

Figure 7. Effects of Gβ5 on RGS7-accelerated GIRK currents coupled to specific PTX-insensitive Gα_i/o subunits.

All data are from oocytes injected with 5 ng Gα* cRNA/oocyte and 10 ng RGS7 cRNA/oocyte, with either 5 ng Gβ5 cRNA/oocyte or 5 ng Gβ1 cRNA/oocyte as a negative control. Oocytes also received Kir3.1 and Kir3.2a cRNA (0.5 ng/oocyte each), and m2 receptor cRNA (0.5 ng/oocyte), and endogenous PTX-sensitive Gα_i/o subunits were inactivated by PTX-S1 expression (1.0 ng/oocyte). Data are means + s.e.m.; * P < 0.05. A, representative ACh-evoked GIRK currents from oocytes expressing RGS7 and Gα_i2(C352G) with either Gβ1 or Gβ5 (upper traces), or Gα_oA(C351G) with either Gβ1 or Gβ5 (lower traces). Scale bars indicate 1 μA and 10 s. B, GIRK currents from A, normalized to peak amplitude and superimposed to highlight their temporal features (black traces, RGS7+Gβ1; grey traces, RGS7+Gβ5). C, Gβ5 selectively suppresses RGS7/ Gαo-coupled GIRK currents. I_K,ACh amplitudes are expressed as a percentage of the mean I_K,ACh amplitude from oocytes expressing the negative control Gβ1. D, GIRK activation time constants (τ_act) derived from oocytes expressing different PTX-insensitive Gα_i/o subunits with either RGS7 + Gβ1 (black bars) or RGS7 + Gβ5 (grey bars). Time constants were derived from a single exponential fit to the rising phase of I_K,ACh elicited by 1 μm ACh. E, deactivation time constants (τ_deact) derived from ACh-elicited GIRK currents as in D. Time constants were derived from a single exponential fit to the decay of I_K,ACh after rapid washout of 1 μm ACh.

Discussion

The goal of this study was to determine whether functional differences exist in the GPCR-dependent gating properties of GIRK channels coupled to specific Gα_i/o subunits and RGS proteins. Our findings reveal that m2 receptors coupled to Gα_i and Gα_o subunits are differentially regulated by RGS4 and RGS7. A significant finding of our study is the identification of G protein coupling levels as a critical determinant in mediating the anomalous kinetic effects of RGS proteins on GIRK channel gating (Doupnik et al. 1997; Saitoh et al. 1997; Zerangue & Jan, 1998; Kovoor & Lester, 2002). Incremental increases in Gα_oβγ coupling by increased Gα_o expression revealed a transition in the RGS4-accelerated GIRK currents from ‘expected’ steady-state properties, characterized by a large reduction in maximal GIRK current amplitude, to ‘anomalous’ gating kinetics which displayed accelerated activation and deactivation kinetics with no effect on steady-state GIRK amplitude. We propose that this transition in steady-state GIRK channel kinetics is a functional indicator of the formation of a GPCR-G protein-GIRK channel signalling complex (Huang et al. 1995; Slesinger et al. 1995) that is precipitated at a critical level of Gα_i/o expression and revealed by the modulatory actions of co-expressed RGS proteins.

RGS proteins reveal GPCR-G protein-GIRK coupling levels via steady-state gating properties

RGS proteins were originally identified as ‘negative regulators’ of G protein signalling causing reduced GPCR signalling in yeast, fungi, nematodes and mammals (Dohlman et al. 1995, 1996; Druey et al. 1996; Koelle & Horvitz, 1996; Yu et al. 1996). These findings were readily explained by RGS proteins behaving as GTPase-activating proteins and accelerating the termination of G protein signalling (Berman et al. 1996; Chen et al. 1996; Hunt et al. 1996; Watson et al. 1996). Initial studies of the effects of RGS proteins on GPCR-evoked GIRK currents revealed the anticipated acceleration in current deactivation, which was well explained by the RGS-accelerated GTPase activity of Gα_i/o subunits causing a more rapid sequesteration of channel-activating Gβγ dimers (Doupnik et al. 1997; Saitoh et al. 1997). Unexpectedly, however, the accelerated deactivation was also accompanied with an accelerated activation phase in the absence of reduced current amplitude.

We analysed the temporal and steady-state GIRK current kinetics for the various conditions tested in this study using a simplified two-state gating scheme (Doupnik et al. 1997):

where the forward rate constant for GIRK channel opening (k_open) is dependent on the rate of Gβγ production during GPCR activation (Breitwieser & Szabo, 1988; Yamada et al. 1998), and the closing rate constant (k_close) is dependent on Gβγ clearance, which is rate-limited by the GTP hydrolysis rate of Gα subunits. Accordingly, the time constant for GIRK channel deactivation (τ_deact) is equal to k_close⁻¹, and the time constant for GIRK activation (τ_act) is equal to (k_open +k_close)⁻¹ (Doupnik et al. 1997). Using these relations and an empirically derived k_open value of 0.03 s⁻¹ (held constant to simulate a saturating agonist concentration), the experimentally derived τ_deact values were used to calculate ‘expected’ values of τ_act and steady-state GIRK current amplitudes (k_open/[k_open +k_close]) in the absence and presence of RGS expression according to the kinetic model. The ‘expected’ values were then compared to the experimentally ‘observed’ values (Table 2).

As seen in Table 2, the ‘observed’ effects of RGS4 on τ_act and steady-state amplitude for Gα_i-coupled GIRK currents closely correlate with the ‘expected’ consequences of the RGS4-accelerated deactivation rate, but not so for the Gα_o-coupled GIRK currents. Similarly, a correlation was seen for the effects of Gβ5 on RGS7-accelerated GIRK currents activated by Gα_o-coupled receptors. Thus during reduced levels of G protein coupling, GIRK channels behave according to standard kinetic concepts in response to RGS modulation and are consistent with the collision coupling model of G protein activation (Shea & Linderman, 1997; Shea et al. 2000). In contrast, RGS4 and RGS7 modulate Gα_o-coupled GIRK currents in a manner similar to previous descriptions of the anomalous kinetic effects of RGS proteins on GIRK channel gating which do not correlate with the kinetic model (Table 2) (Doupnik et al. 1997; Saitoh et al. 1997). We propose that these expression conditions promote the formation of a precoupled m2 receptor-Gα_oβγ-GIRK channel complex (Huang et al. 1995; Slesinger et al. 1995), having altered activation properties that are revealed by the modulatory effects of RGS proteins and effectively preserve maximal GIRK current amplitudes (see Fig. 8).

Figure 8. RGS modulation of GPCR → GIRK signal transduction at varying levels of G protein coupling.

A, collision coupling model (low Gα expression) and B precoupled model (high Gα expression). The precoupled model assumes a GPCR-G protein-RGS-GIRK channel complex displaying ‘anomalous’ kinetic behaviour that was characteristic of oocytes expressing Gα_o subunits with either RGS4 or RGS7, together with co-expressed m2 receptors and GIRK channel Kir3.1/3.2a subunits.

GIRK activation kinetics revealed at low levels of basal activity

GIRK channels are composed of four Kir3.0 subunits with each subunit capable of binding a single Gβγ dimer, thus a maximum of four Gβγ dimers can bind to a single GIRK channel (Corey & Clapham, 2001). Several studies indicate GIRK channel activation is a multi-step process and may represent the binding of multiple Gβγ dimers leading to channel opening (Hosoya et al. 1996; Sodickson & Bean, 1996; Nemec et al. 1999). Assuming four bound Gβγ dimers promote the channel open state, receptor-independent basal GIRK activity represents channels transiently occupied by four Gβγ dimers, excluding the potential influence of phosphatidylinositol-4,5-bisphosphate (PIP₂), arachidonic acid, and intracellular Na⁺, Mg⁺² and polyamines (Dascal, 1997; Yamada et al. 1998). I_K,basal activity is directly related to the concentration of free Gβγ in the cell membrane based on the effectiveness of Gβγ scavengers to lower I_K,basal amplitudes (Lim et al. 1995; Vivaudou et al. 1997; Jeong & Ikeda, 1999; Fernandez-Fernandez et al. 2001). According to the GIRK channel gating scheme shown in Fig. 4A, basal GIRK activity reflects channels transiently occupied by three to four Gβγ dimers where random association of a single free Gβγ subunit causes channel opening. Receptor activation of GIRK channels with high free Gβγ concentrations predict a single Gβγ binding event is necessary to activate the channels and I_K,ACh will thus follow a single exponential time course. Conversely, receptor activation with low concentrations of free Gβγ predict multiple binding events are necessary for channel opening, producing a sigmoidal I_K,ACh time course analogous to the sequential movement of multiple gates during the activation of voltage-gated ion channels (Hodgkin & Huxley, 1952). The observed time course for receptor-dependent GIRK activation at the different levels of basal activity produced by high and low levels of Gα_o expression are in agreement with this and similar GIRK gating schemes (Destexhe & Sejnowski, 1995; Hosoya et al. 1996; Ivanova-Nikolova et al. 1998; Yamada et al. 1998). Different levels of GIRK-Gβγ occupancy may also influence the voltage-dependent relaxation phenomena associated with RGS4 expression, which is only observed at low agonist concentrations (Inanobe et al. 2001). Although we deliberately express low levels of m2 receptors to slow GIRK activation rates (Herlitze et al. 1999), it should be emphasized that the GIRK activation and deactivation kinetics measured in the oocyte expression system, even with RGS expression, are consistently slower than native mammalian cells and may be limited by solution exchange rates and/or other intrinsic oocyte factors.

Interestingly, overexpression of a PTX-insensitive and RGS-resistant Gα_o mutant (Gα_{oA(G184S:C351G)}) in rat sympathetic neurons also reveals a prominent sigmoidal GIRK activation time course via α₂-adrenergic receptor activation, and could be dramatically accelerated by coexpression of the N-terminal region of RGS8 (i.e. no RGS domain) (Jeong & Ikeda, 2001). The results of Jeong & Ikeda indicate the RGS8 N-terminal domain facilitates coupling of heterotrimeric Gα_{oA(G184S:C351G)}βγ subunits to α₂-adrenergic receptors, and as a consequence greatly accelerates the GIRK activation time course. Based on our observations of RGS7 reported here, we speculate that N-terminal regions of RGS7 may similarly interact with m2 receptors and facilitate coupling to Gα_o in the absence of Gβ5.

Physiological implications

GIRK channels mediate sIPSPs in the nervous system where the time course and amplitude of inhibitory synaptic events are dependent on the G protein cycle (Luscher et al. 1997). Thus RGS proteins are likely to play an important role in determining the amplitude and temporal properties of GIRK-mediated sIPSPs. Our findings demonstrate GIRK-mediated sIPSPs may be determined by specific GPCR-Gα-RGS signalling complexes that activate postsynaptic GIRK channels, where certain Gα-RGS precoupled complexes accelerate GIRK activation and deactivation kinetics without compromising amplitude. Perturbation of these signalling complexes would reduce inhibitory synaptic transmission and promote cellular excitability. The consequences of these events in the mammalian nervous system remain to be elucidated, yet the demonstrated role of RGS complexes in coordinating locomotion and egg laying behaviour in Caenorhabditis elegans highlight their potential impact on mammalian neural signalling (Chase et al. 2001; Robatzek et al. 2001).