A regulator of G protein signalling (RGS) protein confers agonist-dependent relaxation gating to a G protein-gated K⁺ channel

Abstract

1. The effects of RGS4 on the voltage-dependent relaxation of G protein-gated K⁺ (K_G) channels were examined by heterologous expression in Xenopus oocytes.

2. While the relaxation kinetics was unaffected by the acetylcholine concentration ([ACh]) in the absence of RGS4, it became dependent on [ACh] when RGS4 was co-expressed.

3. Kinetic analyses indicated that RGS4 confers to the K_G channel a voltage-independent inhibitory gating mechanism, which was attenuated by ACh in a concentration-dependent fashion.

4. In vitro biochemical studies showed that RGS4 could bind to the protein complex containing K_G channel subunits.

5. Since the native cardiac K_G channel exhibited similar agonist-dependent relaxation kinetics to that mediated by RGS4, it is suggested that K_G channel gating is a novel physiological target of RGS protein-mediated regulation.

ACh-induced deceleration of heart beat is mediated by the G protein-gated inwardly rectifying K⁺ (K_G) channel (Yamada et al. 1998). The cardiac K_G channel is composed of Kir3.1 and Kir3.4 and its activity can be reconstituted by the expression of both Kir subunits and the m₂ muscarinic receptor in Xenopus oocytes (Kubo et al. 1993; Krapivinsky et al. 1995). It was recently shown that the activation, desensitization and deactivation of the reconstituted ACh-induced K_G current are accelerated by regulators of G protein signalling (RGS) proteins, thus mimicking the properties of native K_G current (Doupnik et al. 1997; Saitoh et al. 1997, 1999; Herlitze et al. 1999). While acceleration of deactivation and short-term desensitization can be accounted for by the enhanced intrinsic GTP hydrolysis of G_α induced by RGS proteins (Doupnik et al. 1997; Saitoh et al. 1997, 1999; Chuang et al. 1998; Herlitze et al. 1999), this cannot explain acceleration of activation without affecting the steady-state current level. It has been shown recently that retinal RGS9 directly inhibits the activity of guanylyl cyclase in addition to its interaction with G_αt (Seno et al. 1998). Thus the RGS proteins might possess functions other than the acceleration of GTPase activity of G_α, which could also be involved in regulation of the K_G channel.

Here, we show that co-expression of RGS4 confers an agonist dependence to the voltage-dependent relaxation of the K_G channel. The RGS4-induced gating behaviour of the reconstituted K_G channel replicated the ACh-induced voltage-dependent relaxation of the native cardiac K_G channel. These findings indicate the possibility that K_G channel gating is a novel physiological target of RGS protein-mediated regulation.

METHODS

Electrophysiology in Xenopus oocytes and native atrial myocytes

Treatment of Xenopus laevis was in accordance with the guidelines for the use of laboratory animals of Osaka Medical School. The frogs were deeply anaesthetized by immersion in water containing 0.35 % tricaine (Sigma Chemical Co.) and oocytes were surgically removed under clean conditions. Following the operation, the frogs were returned to fresh water to allow them to recover from anaesthesia. Adequate time for healling was allowed between each procedure. Following the last collection, the anaesthetized frogs were killed by decapitation.

Rat Kir3.4, rat RGS4, porcine m₂ muscarinic receptor and bovine G_β1 and G_γ2 cDNAs (kindly provided by Drs D. Clapham, C. Doupnik, T. Kubo and T. Haga, respectively) as well as mouse Kir3.1 cDNA subcloned into pGEMHE vector were transcribed in vitro with T7 RNA polymerase and an mRNA capping kit (Stratagene, La Jolla, CA, USA). A mixture of 160 ng RGS4, 80 ng m₂ receptor and 8 ng each of Kir3.1 and Kir3.4 cRNAs was injected into the oocytes which had been defolliculated in 1 mg ml⁻¹ collagenase solution (Wako Pure Chemical, Osaka, Japan). In some oocytes, G_β1 and G_γ2 cRNAs (10 ng each) were co-injected with RGS4, Kir3.1 and Kir3.4 cRNAs. After injection, the oocytes were maintained in a modified Barth's solution for 72–96 h at 18°C and assayed using a commercially available amplifier (CEZ-1250; Nihon Kohden, Tokyo, Japan). Pipettes were filled with 3 M KCl. The bath solution contained (mM): 90 KCl, 3 MgCl₂, 0.15 niflumic acid and 5 Hepes-KOH (pH 7.4).

Single atrial myocytes were enzymatically isolated from hearts removed from adult male Japanese White rabbits deeply anaesthetized with pentobarbital (Yamada & Kurachi, 1995). Using a Langendorff apparatus, the heart was perfused in a retrograde manner through the coronary arteries with 16 mg of collagenase in 100 ml of nominally Ca²⁺-free bathing solution at 37°C for 10 min. Whole-cell patch-clamp analysis was performed as described previously (Yamada & Kurachi, 1995). Pipettes were filled with the following solution (mM): KCl 140, KH₂PO₄ 1, MgCl₂ 1, EGTA-KOH 5, K₂ATP 3, Na₂GTP 0.1 and Hepes-KOH 5 (pH 7.3). The bath solution contained (mM): NaCl 136.5, KCl 5.4, CaCl₂ 1.8, MgCl₂ 0.53, glucose 5.5 and Hepes-NaOH 5 (pH 7.4).

All experiments were performed at room temperature (20-25°C). The ACh-induced K⁺ currents were obtained by subtracting the control currents from those in the presence of various [ACh] at each potential. In the case of the K_G current in the oocytes expressing G_βγ, the Ba²⁺ (1 mm)-sensitive component was analysed. The data were stored on videotape using a PCM data recording system, and subsequently replayed for computer acquisition and analyses.

In vitro binding assay

GST fusion constructs of RGS4 were prepared by PCR tagging of RGS4 cDNA with Bam HI and Eco RI sites at the 5′ and 3′ end, respectively, and were subcloned into pGEX-2T vector (Amersham Pharmacia Biotech, Uppsala, Sweden). GST-RGS4 and GST were expressed in E. coli and purified from the cell lysates through a glutathione-Sepharose column (Amersham Pharmacia Biotech). Kir3.1 and Kir3.4 cDNAs were subcloned into the pcDNA3 expression vector (Invitrogen, San Diego, CA, USA) and transfected to HEK293T cells with LipofectAMINE and PLUS reagents (Life Technologies Inc., Gaithersburg, MD, USA). Cell lysates containing the expressed Kir subunits were collected with buffer containing 20 mm Hepes-NaOH (pH 7.4), 1 mm Na-EDTA, 0.5 mm Na-EGTA, 1 mm DTT, 145 mm NaCl, 5 mm KCl, 1 % (w/v) CHAPS (3-[(3-cholamidopropyl)dimethylammonio]-1-propane sulphonate), 1 mm PMSF (phenylmethylsulphonyl fluoride), 0.2 mg ml⁻¹ benzamidine HCl, and 5 μg ml⁻¹ each of pepstatin, leupeptin and chymostatin. After centrifugation, the clear supernatants were incubated with GST-RGS4 and GST immobilized on the glutathione-Sepharose beads for 3 h at 4°C. The beads were extensively washed and incubated with Laemmli sample buffer for 5 min at 95°C. The bound proteins were analysed by SDS-PAGE and detected with antibodies specific to Kir3.1 (aG1C-1) and to Kir3.4 (aG4N-10) (Morishige et al. 1999).

RESULTS

Effects of RGS4 on the relaxation of K_G channel currents

Xenopus oocytes injected with Kir3.1, Kir3.4 and m₂ receptor cRNAs expressed an inwardly rectifying K⁺ current which responded slowly to ACh (Fig. 1A). On wash-out of ACh, the current disappeared with a time constant of ∼10 s. When RGS4 was co-expressed, the activation and deactivation time courses were accelerated (Fig. 1b; Doupnik et al. 1997; Saitoh et al. 1997, 1999; Herlitze et al. 1999). The steady-state level of ACh-induced current was unaffected by the co-expression of RGS4 (Fig. 1C).

Figure 1. Effect of RGS4 on the relaxation kinetics of the reconstituted K_G channel.

Whole-cell currents were recorded from Xenopus oocytes expressing m₂ receptor and heteromeric Kir3.1/Kir3.4 K_G channels in the absence or presence of RGS4 (A and B, respectively). Test pulses (2 s) to -60 mV were applied every 3 s from a holding potential of 0 mV. Acetylcholine (ACh, 1 μm) was added to the bath as indicated by the horizontal bar above each trace. C, ACh-induced K_G current amplitude was plotted against [ACh]. The mean (±s.e.m.) values of EC₅₀ and I_max of the ACh-induced currents in the presence and absence of RGS4 were 9.1 and 9.9 nM, and 4.3 ± 0.52 and 4.1 ± 0.49 μA, respectively (n= 10). D and E, effect of RGS4 on the gating properties of ACh-induced K_G current. Inward current traces in the absence (D) or presence (E) of RGS4 were recorded with a voltage step to -60 mV from 0 mV and basal currents were subtracted. The arrowheads represent the end and start points of the instantaneous and relaxation components of the currents. Traces in the bottom panel represent normalized and superimposed current evoked by the various [ACh] or directly activated by G_βγ. Traces from the same series of oocytes are compared, and similar results were obtained in oocytes from three different frogs. The dashed lines in A-E indicate the zero current level. F and G, the I_ins/I_max ratio (F) and the time constant (τ) of the relaxation (G) of K_G currents at each [ACh] (open symbols) and with over-expressed G_βγ (filled symbols; n= 10 for each point). Results obtained in the presence and absence of RGS4 are represented by triangles and circles, respectively.

A slowly increasing K⁺ current during hyperpolarizing voltage pulses, known as relaxation, is one of the features of G protein-gated K⁺ (K_G) channels containing Kir3.1. The ACh-induced K⁺ current during a voltage step could be divided into instantaneous and relaxing components. In the control (Fig. 1D), at all concentrations of ACh, the ratio of the instantaneous component to the pulse-end current (I_ins/I_max) was ∼0.71 (Fig. 1F, ○), and the relaxation time constant (τ) was ∼200 ms (Fig. 1G, ○), this notwithstanding that the current amplitude increased as [ACh] was raised (Fig. 1C). The superimposed normalized K_G currents at the bottom of Fig. 1D indicated that I_ins/I_max and τ remained constant at various [ACh] and these values were comparable to those of the K_G current directly activated by G_βγ in the same series of oocytes.

When RGS4 was co-expressed with K_G channels, the relaxation of the ACh-induced currents became dependent on [ACh] (Fig. 1E). At 3 × 10⁻⁹ M ACh, I_ins/I_max was 0.45 and τ was ∼450 ms. As [ACh] was raised, the I_ins/I_max ratio increased and τ became smaller. At 10⁻⁶ M ACh, both values were similar to those with over-expressed G_βγ (Fig. 1F and G, ▵ and ▴). The superimposed normalized currents at the bottom of Fig. 1E clearly show the alteration in the relaxation properties of ACh-induced K⁺ current at various [ACh] in the presence of RGS4. Because I_ins/I_max and τ of the K⁺ current induced by over-expression of G_βγ were the same in the absence and presence of RGS4 (filled symbols in Fig. 1F and G), it would seem that RGS4 may modulate K_G channel gating in an inhibitory manner, which appears to be attenuated by the activation pathway followed by G_βγ.

RGS4 inhibits the gating of K_G channel

In Fig. 2, we examined the quasi-steady-state relative open probability (P_o) of the K_G channel relaxation gate at various potentials in the absence and presence of RGS4. A double pulse voltage-clamp protocol was used. In the absence of RGS4, at all [ACh] examined and with over-expressed G_βγ, the I_ins/I_max ratio of the tail currents recorded at -100 mV was ∼1 following pre-pulse voltage steps to between -100 and -80 mV and it decreased to ∼0.7 with pre-pulses to positive membrane potentials. Thus, the quasi-steady-state relative P_o curves obtained at different [ACh] and with over-expression of G_βγ were very similar (Fig. 2C). On the other hand, when RGS4 was co-expressed, the instantaneous component of the tail currents following the depolarizing pulses was clearly smaller than that in the control (Fig. 2A and B). As [ACh] was increased, the instantaneous component of the tail currents increased and the relative P_o curves were shifted upwards at depolarized potentials in a concentration-dependent fashion (Fig. 2B and D). The relative P_o curve at 10⁻⁶ M ACh even in the presence of RGS4 was similar to that with over-expressed G_βγ (Fig. 2D, ○ and ▾, respectively).

Figure 2. The effect of RGS4 on the quasi-steady-state relative open probability (P_o) curve of the K_G channel relaxation gate.

A and B, comparison of the tail currents of Kir3.1/Kir3.4 K_G channels in the absence (A) or presence (B) of RGS4. Xenopus oocytes were injected with the cRNAs of the m₂ receptor, Kir3.1 and Kir3.4 without (control) or with the cRNA of RGS4. Tail currents at −100 mV were recorded following prepulses for 2 s each from −100 to +60 mV in steps of 20 mV. Basal currents evoked by this protocol without ACh were subtracted. Dashed lines represent the zero current level. C and D, the quasi–steady–state voltage–dependent relative P_o curve of the K_G channel with or without RGS4 at various [ACh]. The I_ins/I_max ratio was calculated from the experimental protocols shown in A and B. Lines are the least–squares fits of the data points to the modified Boltzmann equation:

where d(V) is the voltage-dependent component of the relative P_o curve, V is the membrane potential (in mV), A is a constant value, V_½ is the membrane potential at half the value of d(V) and k is the slope factor. The relative P_o curves of the K_G channel relaxation gate in the presence of 10 nm, 100 nM and 1 μm ACh, and with over-expressed G_βγ could be represented with the following parameters, respectively: A= 0.62, 0.63, 0.61 and 0.68; V_½= -29.6, -29.1, -30.4 and -28.4 mV; k= 16.3, 16.1, 18.4 and 17.2 in the absence of RGS4; A= 0.41, 0.54, 0.65 and 0.70; V_½= -27.7, -26.9, -26.8 and -27.2 mV; k= 21.9, 20.6, 18.4 and 19.1 in the presence of RGS4. n= 6 for each mean value; the s.e.m. value was smaller than each symbol.

The relative P_o curves at various [ACh] and with over-expressed G_βγ were fitted by the modified Boltzmann equation (Fig. 2 legend). RGS4 affected exclusively the A value but neither the V_½ nor k values of the relative P_o curve. Therefore, RGS4 regulates a voltage-independent parameter of the K_G channel relaxation relative P_o curve.

Biochemical study of a link between RGS4 and K_G channel

The possible association between RGS4 and the K_G channel was examined using GST-RGS4 and CHAPS extract from HEK293T cells expressing Kir3.1 and/or Kir3.4 (Fig. 3). Kir3.1 and Kir3.4 immunoreactivity was retained in the complex of GST-RSG4 and glutathione-Sepharose beads when incubated with the lysate from cells co-transfected with Kir3.1 and Kir3.4, but neither signal was observed in the control (GST alone; Fig. 3A and B). Next we prepared lysates from cells separately expressing Kir3.1 or Kir3.4 and mixed them with GST-RGS4 or GST (see Fig. 3C). Although the lysates from cells transfected with the separate Kir subunits contained equal amounts of protein to those co-transfected with both subunits (right panels in Fig. 3A and B), only a small amount of Kir3.1 and little of Kir3.4 were recovered using the GST-RGS4-glutathione-Sepharose beads (Fig. 3A and B). These results suggest that heteromeric Kir3.1/Kir3.4 channel proteins are more effective than homomeric Kir subunits in forming protein complexes containing RGS4.

Figure 3. In vitro association between RGS4 and the Kir3.1/Kir 3.4 K_G channel.

RGS4 fused with GST (GST-RGS4) and GST (1 μg each) were immobilized on glutathione-Sepharose beads and incubated with the lysates from HEK293T cells transfected with various combinations of Kir3.1 and Kir3.4 as indicated above each panel. The proteins recovered on the beads were analysed with anti-Kir3.1 (aG1C-1; A) or anti-Kir3.4 antibodies (aG4N-10; B). Both subunits in the input lysates were detected by this method (right panels). C, protein bands were visualized by staining the gel with Coomassie brilliant blue R-250 (CBB) to show GST-RGS4 and GST. Numbers on the left side of each panel represent the molecular mass of the standard markers in kilodaltons.

[ACh]-dependent alteration of relaxation in native K_G channels

In Fig. 4, we examined the effects of various [ACh] on the relaxation of native K_G channels in rabbit atrial myocytes. The amplitude of ACh-induced K⁺ current increased as [ACh] was raised (Fig. 4A). The ACh-induced K⁺ current increased slowly during hyperpolarizing pulses and upon returning to the holding potential the outward current gradually decreased. As [ACh] was raised, the I_ins/I_max ratio increased (Fig. 4B) and τ became smaller (Fig. 4C). The superimposed normalized currents in Fig. 4A were very similar to those of the reconstituted Kir3.1/Kir3.4 K_G channel with RGS4. This strongly suggests that RGS proteins may participate in the physiological regulation of K_G channel relaxation gating in atrial myocytes.

Figure 4. Relaxation of the ACh-induced K⁺ current in rabbit atrial myocytes.

A, whole-cell currents in response to various [ACh] were obtained by application of hyperpolarizing pulses to -100 mV for 1 s every 7 s from the holding potential of -50 mV. In the bottom panel, superimposed normalized ACh-induced currents are shown. Dashed lines and arrowheads indicate the zero current level and the point separating the instantaneous and relaxation components, respectively. B and C, the I_ins/I_max ratio (B) and the time constant (τ) of relaxation (C; means ±s.e.m., n= 3).

DISCUSSION

This study shows that RGS4 confers ACh dependence on the relaxation gating of Kir3.1/Kir3.4 K_G channels. Since the native cardiac K_G channel shows similar ACh-dependent relaxation kinetics, regulation of K_G channel gating by RGS-mediated signalling may occur physiologically.

The analyses of quasi-steady-state relative P_o curves of K_G channels in the absence and presence of RGS4 clearly indicated that RGS4 decreased a voltage-independent parameter (A). The voltage-independent decrease of A was attenuated by ACh in a concentration-dependent manner. Furthermore, over-expression of G_βγ completely cancelled the inhibitory effect of RGS4 on the relative P_o curve. Therefore, a RGS4-mediated regulatory mechanism may act on the non-activated K_G channels and free G_βγ may hinder the effect of RGS4 on channel gating. The in vitro biochemical experiments indicated that RGS4 and K_G channels composed of Kir3.1 and Kir3.4 were closely associated and also suggested that different Kir3.0 subunits and their combinations may have different affinities for RGS-mediated regulatory mechanisms. However, we could not determine in this study whether the interaction between K_G channel subunits and RGS4 is direct or involves other intermediate proteins. Further studies are needed to clarify the molecular mechanisms for the linkage and interaction between the K_G channel subunits, RGS proteins and G_βγ.

The fast deactivation of the ACh-induced K_G current after wash-out of the agonist and the short-term desensitization have been attributed to the acceleration of the GTPase activity of G_α by RGS proteins (Doupnik et al. 1997; Saitoh et al. 1997, 1999; Chuang et al. 1998; Herlitze et al. 1999). Acceleration of activation was not simply explained by the GTPase-activating protein (GAP) action of RGS proteins (Doupnik et al. 1997). Saitoh et al. (1997) proposed that RGS proteins may function as an on-rate accelerator. They further showed that RGS7 accelerates activation with only a weak effect on deactivation (Saitoh et al. 1999). Herlitze et al. (1999), however, suggested using mutational analysis of RGS2 that the underlying mechanism for acceleration of activation and that of deactivation may share the same pathway. Bünemann & Hosey (1998) showed that RGS proteins, in addition to their GAP action, increase the availability of G_βγ in HEK293T cells. Therefore, it seems likely that RGS proteins possess multiple functional arms to the G protein activation of K_G channel system, which have not yet been fully clarified. The effect of RGS proteins on K_G channel gating reported here has not been recognized at all in previous studies, and may provide a novel insight into the physiological regulation of K_G channel current.