Regulation of a G protein-gated inwardly rectifying K+ channel by a Ca2+-independent protein kinase C


Corresponding author A. Tinker: Centre for Clinical Pharmacology, Department of Medicine, UCL, The Rayne Institute, 5 University Street, London WC1E 6JJ, UK., Email:


  • 1Members of the Kir3.0 family of inwardly rectifying K+ channels are expressed in neuronal, atrial and endocrine tissues and play key roles in generating late inhibitory postsynaptic potentials (IPSPs), slowing heart rate and modulating hormone release. They are activated directly by Gβγ subunits released in response to Gi/o-coupled receptor stimulation. However, it is not clear to what extent this process can be dynamically regulated by other cellular signalling systems. In this study we have explored pathways activated by the Gq/11-coupled M1 and M3 muscarinic receptors and their role in the regulation of Kir3.1+3.2A neuronal-type channels stably expressed in the human embryonic kidney cell line HEK293.
  • 2We describe a novel biphasic pattern of behaviour in which currents are initially stimulated but subsequently profoundly inhibited through activation of M1 and M3 receptors. This contrasts with the simple stimulation seen through activation of M2 and M4 receptors.
  • 3Channel stimulation via M1 but not M3 receptors was sensitive to pertussis toxin whereas channel inhibition through both M1 and M3 receptors was insensitive. In contrast over-expression of the C-terminus of phospholipase Cβ1 or a Gq/11-specific regulator of G protein signalling (RGS2) essentially abolished the inhibitory phase.
  • 4The inhibitory effects of M1 and M3 receptor stimulation were mimicked by phorbol esters and a synthetic analogue of diacylglycerol but not by the inactive phorbol ester 4αphorbol. Inhibition of the current by a synthetic analogue of diacylglycerol effectively occluded any further inhibition (but not activation) via the M3 receptor.
  • 5The receptor-mediated inhibitory phenomena occur with essentially equal magnitude at all intracellular calcium concentrations examined (range, 0-669 nm).
  • 6The expression of endogenous protein kinase C (PKC) isoforms in HEK293 cells was examined by immunoblotting, and their translocation in response to phorbol ester treatment by cellular extraction. The results indicated the expression and translocation of the novel PKC isoforms PKCδ and PKCε.
  • 7We also demonstrate that activation of such a pathway via both receptor-mediated and receptor-independent means profoundly attenuated subsequent channel stimulation by Gi/o-coupled receptors.
  • 8Our data support a role for a Ca2+-independent PKC isoform in dynamic channel regulation, such that channel activity can be profoundly reduced by M1 and M3 muscarinic receptor stimulation.

An inwardly rectifying K+ current activated by acetylcholine was first described in atrial myocytes (Noma & Trautwein, 1978). It was subsequently established that the activation was mediated by M2 muscarinic receptors coupled to pertussis toxin (PTx)-sensitive G proteins (Breitwieser & Szabo, 1985; Kurachi et al. 1986; Pfaffinger et al. 1995). Neurophysiologists have recognized the existence of analogous currents in neurones and neuroendocrine cells activated via a number of Gi/o-coupled transmitter pathways (Luscher et al. 1997; Takano et al. 1997; Knoflach & Kemp, 1998) and these are thought to play an important role in the generation of late IPSPs and in inhibiting hormone release (Luscher et al. 1997; Pennefather et al. 1998; Yamada et al. 1998).

The molecular counterparts of these native currents have now been identified. The channel is a heteromultimer of members of the Kir3.x family. There are five members, Kir3.1-3.5, and also splice variants of Kir3.2 (Dascal et al. 1993; Kubo et al. 1993; Lesage et al. 1994; Krapivinsky et al. 1995; Hedin et al. 1996; Inanobe et al. 1996). Work from a number of laboratories on both cloned and native channels has established that it is the Gβγ dimer that is the direct activator of the channel (Logothetis et al. 1987; Reuveny et al. 1994).

We have previously shown that these channels when expressed in a mammalian cell line are activated by Gi/o- but not Gs-coupled receptors (Leaney et al. 2000; Leaney & Tinker, 2000). It is not clear to what extent this process is dynamically regulated. In this study we have used the family of muscarinic receptors as a means to investigate channel regulation by different classes of heptahelical G protein-coupled receptor. There are five receptor subtypes, M1-M5, which are widely distributed (Caulfield & Birdsall, 1998). It is generally asserted that the M2 and M4 receptors preferentially couple to Gi/o whilst the M1, M3 and M5 receptors couple to Gq/11. In the present study we have utilized the M1 and M3 muscarinic receptors as examples of Gq/11-coupled receptors to investigate the regulation of a cloned counterpart of the neuronal G protein-gated K+ channel (Kir3.1+3.2A) and compared this behaviour with that of receptors predominantly coupling to Gi/o (M2 and M4).


Molecular biology and cell culture

Standard molecular cloning and cell culture techniques were employed throughout. HEK293 cell culture, generation of stable cell lines and transfection procedures were as previously described (Leaney et al. 2000; Leaney & Tinker, 2000). The following cDNAs were used: rat Kir3.1, rat Kir3.2A, human M1, M2, M3 and M4, human RGS2 and rat PLCβct. Transfected cells suitable for patch clamping were identified by epifluorescence from co-transfection of 50 ng of the enhanced variant of the green fluorescent protein (pEGFP-N1; Clontech). In this study a stable cell line expressing both the M3 receptor and Kir3.1 and Kir3.2A was established using dual selection with G418 and Zeocin (Invitrogen) as previously described (Leaney et al. 2000).

Cell extraction and identification of PKC isoforms

HEK293 cells (three 15 cm tissue culture dishes) were extracted in 1.5 ml Laemmli buffer (Laemmli, 1970) and 10 μl of protein was analysed for the presence of PKC isoforms by 9 % SDS-PAGE and Western blotting as previously described (Towbin et al. 1979; Dekker & Parker, 1997). The following antibodies were employed for the Western blot shown in Fig. 8. PKC-α, polyclonal antibody (Kiley & Parker, 1995); PKC-βI, SC-209 (Santa Cruz); PKC-βII, SC-210 (Santa Cruz); PKC-γ, SC-211 (Santa Cruz); PKC-δ, SC-214 (Santa Cruz); PKC-ε, protein A-purified polyclonal antibody (Schaap et al. 1989); PKC-θ, SC-1875 (Santa Cruz); PKC-χ, protein A-purified polyclonal antibody (Ways et al. 1992).

Figure 8.

Expression and translocation of PKC isoforms in HEK293 cells

A, immunoblot illustrating endogenous expression of PKC isoforms in HEK293 cells. Molecular mass (in kDa) is indicated on the left. B, the translocation of PKCδ and PKCε was investigated in the HKIR3.1/3.2 cell line in response to PMA and the inactive analogue 4αphorbol (4α; both 100 nm). The immunoblots, which were quantified by gel-scanning densitometry (see Methods), are shown on the left-hand side of this figure and the corresponding bar charts on the right-hand side. C, cytosol; P, particulate. Data are presented as means ±s.d. (n = 3).

For membrane localization experiments, cells were washed in PBS and homogenized in ice-cold extraction buffer (20 mm Tris-Cl pH 7.4, 2 μg ml−1 aprotinin, 100 μm tosyl-lysine-chloromethyl ketone, 1 μm pepstatin, 50 μg ml−1 PMSF, 1 μg ml−1 diisopropylfluorophosphate (DFP)) by 20 up-and-down strokes in a Dounce homogenizer. The extract was incubated at 4 °C for 15 min and centrifuged at 14 000 r.p.m. at 4 °C. The supernatant was taken (cytosol) and the pellet was re-extracted as above. After clearance the supernatant was discarded and the pellet was homogenized in extraction buffer containing 1 % Triton X-100. The homogenate was incubated at 4 °C for 30 min and centrifuged as above. The supernatant was taken (particulate). One-fifth volume of five times concentrated Laemmli buffer was added to the cytosol and membrane fractions and 10 μl of this was analysed by SDS-PAGE and Western blotting. Immunoreactivity was visualized using the ECL detection system (Amersham) and quantified using Scion Image software (Scion Corporation, MD, USA).


Cells were plated on 13 mm glass coverslips for electrophysiological recordings. Whole-cell membrane currents were recorded using an Axopatch 200B amplifier (Axon Instruments). Patch pipettes were pulled from filamented borosilicate glass (Harvard Apparatus) and had a resistance of 1.5-2.5 MΩ when filled with pipette solution (see below). Prior to filling, the tips of the patch pipettes were coated with a parafilm and mineral oil suspension. Records were filtered at 1 kHz, digitized at 5 kHz and data acquired and analysed using a Digidata 1200B interface (Axon Instruments) and pCLAMP software (version 6.0; Axon Instruments). Cell capacitance was approximately 15 pF and series resistance (< 10 MΩ) was at least 75 % compensated. Recordings of membrane current were commenced after an equilibration period of approximately 5 min. In most of the experiments drugs were applied by a gravity-fed perfusion system. The chamber volume was 250 μl, the flow rate was 7 ml min−1 and the dead time was ≈15 s. Solutions were totally exchanged within 30 s. In some experiments carbachol was applied using a ‘sewer pipe’ system (Rapid Solution Changer RSC-160; Bio-Logic) whereby an array of perfusion capillaries was placed in the bath approximately 40 μm from the recorded cell. This system allows rapid solution switching between capillary tubes and localized application of agonist due to the laminar flow over the studied cell from the pipes. Carbachol was applied to the cell within 200 ms. With either perfusion system only one cell per coverslip was examined due to the essentially irreversible nature of the effect. After the completion of the experiment the coverslip was changed. For current-voltage relationships, currents were measured at the end of each voltage step. Current densities were measured at -60 mV (unless otherwise stated) and all data are presented as means ±s.e.m. (unless otherwise stated) where n indicates the number of cells recorded from. Data were analysed for statistical significance using either Student's t test or one-way repeated-measures ANOVA tests with Bonferroni correction as appropriate; *P ≤ 0.05, **P ≤ 0.01 and ***P ≤ 0.001.

Materials and drugs

Solutions were as follows. Pipette solution (mm): 107 KCl, 1.2 MgCl2, 1 CaCl2, 10 EGTA, 5 Hepes, 2 MgATP and 0.3 Na2GTP (KOH to pH 7.2, ≈140 mm total K+, calculated free Ca2+ 18 nm). In some experiments the concentration of CaCl2 in the pipette solution was increased to 4 or 8 mm resulting in calculated intracellular free Ca2+ concentrations of 111 or 669 nm, respectively, or was omitted altogether resulting in an essentially Ca2+-free pipette solution. Free Ca2+ concentrations were calculated using CaBuf software (G. Droogmans, KU Leuven, Belgium; Nilius et al. 1997). The bath solution comprised (mm): 140 KCl, 2.6 CaCl2, 1.2 MgCl2 and 5 Hepes (pH 7.4). Cell culture materials were from Life Technologies, Inc. and Invitrogen. Molecular biology reagents were obtained from New England Biolabs or Roche Molecular Biochemicals. All chemicals were from Sigma or Calbiochem. Drugs were made up as concentrated stock solutions in ethanol, water or DMSO and kept at 4 or -20 °C.


The experiments described in this paper were performed either on a HEK293 cell line stably expressing the channel subunits Kir3.1 and Kir3.2A (denoted as HKIR3.1/3.2) into which receptors and other signalling components were transiently transfected (Leaney et al. 2000; Leaney & Tinker, 2000) or on a HEK293 cell line that stably expressed the M3 receptor along with both channel subunits (denoted as HKIR3.1/3.2/ M3). Wild-type HEK293 cells, i.e. those not transfected with channel subunits, have a current density of 5.9 ± 0.9 pA pF−1(n = 24) at -60 mV. In this study we have used the non-specific muscarinic receptor agonist carbachol (10 μm). Application of this agent to the HKIR3.1/3.2 line in the absence of expression of muscarinic receptors did not modulate whole-cell currents (basal: 22.7 ± 3.8 pA pF−1, +carbachol: 23.1 ± 3.9 pA pF−1, n = 7).

stimulation of kir3.1+3.2a channels by m2 and M4 muscarinic receptors

We explored the effects of stimulating different muscarinic receptor family members on the G-protein-regulated K+ channel Kir3.1+3.2A. It was first confirmed that stimulation of the Gi/o-coupled M2 and M4 receptors leads to current potentiation upon transient expression in the HKIR3.1/3.2 cell line. Carbachol (10 μm) was applied for 10-30 s resulting in an approximately threefold potentiation of Kir3.1+3.2A currents. Upon washing the current level returned to the baseline level prior to stimulation (Fig. 1A-C for M2 receptor). A similar response was seen on activation of the currents via the M4 receptor (Fig. 1D). These effects were completely abolished by treating cells with PTx (100 ng ml−1, 16 h, Fig. 1D). It has recently been demonstrated that Kir3.1+3.4 channels expressed in CHO cells undergo a novel inhibition in response to prolonged M2 receptor stimulation (Bünemann et al. 2000). We performed analogous experiments by expressing the M2 receptor in the HKIR3.1/3.2 line and stimulating with carbachol (1 μm) for 2 min. We found that M2 receptor stimulation significantly increased current density from 46.6 ± 14.8 to 96.5 ± 18.34 pA pF−1(n = 5, P = 0.02) but we observed little inhibition of channel activity during the application of carbachol (17.5 ± 6.3 %, n = 5). Two minutes later, carbachol was applied again and currents were increased to a similar level (control, 42.4 ± 13.0 pA pF−1; +carbachol, 103.2 ± 22.8 pA pF−1; n = 5, P = 0.03). Thus in HEK293 cells expressing Kir 3.1+3.2A channels we did not observe a significant M2-mediated inhibitory response.

Figure 1.

M2 and M4 muscarinic receptors stimulate Kir3.1+3.2A channels

A, a representative example of the effects of M2 receptor stimulation (10 μm carbachol, CCh) on Kir3.1+3.2A currents. Currents were elicited by holding cells at 0 mV and stepping to potentials between -100 and +50 mV in 10 mV increments for 100 ms. Current traces were recorded before (Control), during (+carbachol) and after receptor stimulation (Wash). B, corresponding current-voltage relationships from data shown in A. C, an example of the effects of 10 μm carbachol (applied as indicated by the bar) upon membrane current in a cell voltage clamped at -60 mV. The dotted line indicates zero current and the dashed line indicates basal current prior to receptor stimulation. Note that after removal of carbachol current returns to basal level. D, bar charts summarizing the effects of stimulating M2 and M4 receptors in control (left-hand panel) and PTx-treated cells (right-hand panel). □, basal current density (Control); ▪, current density due to receptor stimulation (+CCh); inline image, current density after agonist is removed (Wash). Numbers in parentheses indicate the number of cells recorded from. In control cells M2 stimulation increased basal current density from 61.4 ± 8.4 to 214.9 ± 31.3 pA pF−1(n = 16, ** P < 0.01) whilst M4 receptor stimulation increased current density from 82.9 ± 19.2 to 251.9 ± 47.6 pA pF−1(n = 15, ** P < 0.01). In PTx-treated cells (100 ng ml−1, 16 h) carbachol was unable to potentiate Kir3.1+3.2A currents (M2: basal 83.8 ± 15.1 pA pF−1, +carbachol 95.0 ± 12.1 pA pF−1, n = 6, P = 0.12; M4: basal 53.3 ± 11.7 pA pF−1, +carbachol 56.6 ± 13.8 pA pF−1, n = 8, P = 0.44).

biphasic regulation of kir3.1+3.2a channels by m1 and M3 muscarinic receptors

The effects of stimulating the Gq/11-coupled M1 and M3 muscarinic receptors on Kir3.1+3.2A currents were next examined. Brief application (10-30 s) of 10 μm carbachol led to a biphasic response. Currents were initially potentiated by agonist but then decreased to a value substantially lower than that prior to receptor stimulation (Fig. 2a and B). Kir3.1+3.2A channels can be activated directly by over-expressing Gβ1 and Gγ2 (basal, 49.6 ± 4.4 pA pF−1, n = 97; +β1γ2, 341.9 ± 81.0 pA pF−1, n = 24). When we co-expressed Gβ1γ2 with the M3 receptor in HKIR3.1/3.2, there was an elevation of the basal current and the stimulatory effects via the M3 receptor were less pronounced, yet the inhibitory effects were still profound (Fig. 2C and D). The inhibitory effect mediated through both M1 and M3 receptors was long lasting and currents were still inhibited 20 min after carbachol was removed (M3: current density prior to carbachol, 44.8 ± 6.4 pA pF−1; 20 min after carbachol, 18.2 ± 3.6 pA pF−1; n = 8, P = 0.001). Equivalent phenomena were observed in cells that stably expressed the M3 receptor with Kir3.1 and Kir3.2A (HKIR3.1/3.2/M3). A qualitatively similar inhibitory response was observed although the magnitude was greater (78.7 ± 2.7 % inhibition, n = 16).

Figure 2.

M1 and M3 muscarinic receptors have a dual effect on Kir3.1+3.2A channels

A, example of currents (elicited as described in Fig. 1A) recorded from HKIR3.1/3.2 cells co-expressing M1 receptors. Currents were recorded before (Control), during (+carbachol) and 5 min after receptor stimulation (Wash). B, current-voltage relationships from the data shown in A. C, representative examples of the effects of M1 and M3 receptor stimulation on Kir3.1+3.2A current recorded at a holding potential of -60 mV. Carbachol (10 μm) was applied as indicated by the bar, the dotted line indicates zero current and the dashed line indicates basal current prior to receptor stimulation. The left-hand panel shows the effects of M1 receptor stimulation, the middle panel the effects of M3 receptor stimulation and the right-hand panel the effects of co-expression of β1γ2 dimers upon the M3 response. D, bar charts summarizing the effects of stimulating M1, M3 and M31γ2 on Kir3.1+3.2A currents (M1: basal 60.0 ± 13 pA pF−1, +carbachol 123.7 ± 14.0 pA pF−1, wash 24.1 ± 7.6 pA pF−1, equivalent to 62.1 ± 4.8 % inhibition (n = 23); M3: basal 28.1 ± 4.9 pA pF−1, +carbachol 89.2 ± 9.2 pA pF−1, wash 13.2 ± 2.3 pA pF−1, equivalent to 47.6 ± 4.4 % inhibition (n = 31); M31γ2: basal 136.8 ± 28.1 pA pF−1, +carbachol 159.9 ± 24.9 pA pF−1, wash 68.9 ± 20 pA pF−1, equivalent to 53.2 ± 6.6 % inhibition (n = 12)). ** P < 0.01; *** P < 0.001.

We performed experiments where carbachol was applied rapidly and locally to HKIR3.1/3.2/M3 cells using a rapid perfusion system (see Methods). In all experiments application of carbachol resulted in a biphasic response where current density was initially increased and then decreased to a value lower than that prior to receptor stimulation. Even very brief applications (2 s) of carbachol still led to a profound inhibition of current density and there was no significant quantitative difference in the degree of final inhibition observed with any of the time periods (2, 20 and 200 s) of agonist application (Fig. 3).

Figure 3.

The period of application of carbachol does not affect the extent of inhibition of Kir3.1+3.2A currents

A, representative traces illustrating the effect of different periods of application (2, 20 and 200 s) of carbachol (10 μm) on HKIR3.1/3.2/M3 cells voltage clamped at -60 mV. The dotted line indicates basal current prior to receptor stimulation. B, bar charts summarizing data from a number of cells with the periods of carbachol application as indicated in the corresponding data traces shown in A.□, basal current density prior to carbachol application (Control); ▪, peak current density during application (+CCh peak); inline image, current density measured at the end of the drug application (+CCh end); inline image, current density after the removal of stimulus (Wash). Levels of significance are shown with respect to Control current density. * P < 0.05; ** P < 0.01; *** P < 0.001.

All of the experiments described above used a 10 mm EGTA-containing pipette solution in which the calculated free Ca2+ concentration was 18 nm. The high EGTA concentration should potentially minimize fluctuations in intracellular Ca2+ concentration upon receptor stimulation. We investigated the Ca2+ dependence of the M3 response on the background of constant receptor expression in the HKIR3.1/3.2/M3 stable cell line by using pipette solutions with different free Ca2+ concentrations. The concentration of CaCl2 was altered whilst the concentration of EGTA was kept constant (10 mm) to make pipette solutions with calculated intracellular free Ca2+ concentrations of 0, 111 or 669 nm (see Methods). We found that the magnitude of channel inhibition observed in response to M3 receptor stimulation was not significantly enhanced when intracellular Ca2+ was altered to various concentrations between 0 and 669 nm (Fig. 4). Thus the inhibitory effect was essentially Ca2+ independent. In all the experiments described below the pipette solution used had a calculated free Ca2+ concentration of 18 nm.

Figure 4.

The inhibitory effect of M3 receptor stimulation on Kir3.1+3.2A channels is Ca2+ independent

This bar chart summarizes the percentage inhibition of Kir3.1+3.2A current density mediated by M3 receptor stimulation (10 μm CCh) using four different concentrations of calculated intracellular free Ca2+ (0, 18, 111 and 669 nm). These experiments were performed in the stable cell line HKIR3.1/3.2/M3. * P < 0.05.

g proteins mediating the biphasic response to m1 and M3 receptor stimulation

Although it is generally accepted that the M1 and M3 receptors preferentially couple to the Gq family of G proteins they can also activate other signal transduction pathways (Felder, 1995; Nahorski et al. 1997). Thus we examined the sensitivity of the responses to agents known to selectively target certain G protein pathways. G proteins of the Gi/o family were uncoupled using PTx whilst G proteins of the Gq/11 class were targeted by using proteins or domains of protein known to have a GTPase-activating property for this family. Overexpression of such proteins leads potentially to an attenuation of signalling via Gq/11-coupled pathways. We used the C-terminal domain of phospholipase Cβ1 (PLCβ1ct) and a regulator of G protein signalling (RGS2) (Berstein et al. 1992; Wu et al. 1993; Paulssen et al. 1996; Kammermeier & Ikeda, 1999). The HKIR3.1/3.2 cell line was transiently transfected with the M1 or M3 receptor and treated with PTx. PTx abolished the stimulatory effect of M1, but not M3, receptors whilst leaving the inhibitory effect through both receptors intact (Fig. 5A-C). Overexpression of PLCβ1ct in the HKIR3.1/3.2/M3 cell line and co-expression of RGS2 along with either the M1 or the M3 receptor in the HKIR3.1/3.2 cell line significantly reduced the inhibitory effects of carbachol, supporting the involvement of the Gq/11 pathway in mediating the M1 and M3 inhibitory effects (Fig. 5D and E).

Figure 5.

Inhibition of Kir3.1+3.2A channels by M1 and M3 receptors is not mediated through Gi/o proteins

A, current traces recorded from a PTx-treated cell before (Control), during (+carbachol) and after M1 receptor stimulation (Wash). B, current-voltage relationships obtained from the data shown in A. C, bar charts summarizing the effects of PTx upon M1 (left-hand panel) and M3 (right-hand panel) responses. PTx abolishes M1-mediated channel activation (▪) but not inhibition ( inline image), but is ineffective on both channel activation and inhibition due to M3 receptor stimulation. D, the C-terminus of PLCβ1 (PLCβ1ct) attenuated M3-mediated inhibition (M3+PLCβ1ct: basal 13.9 ± 3.2 pA pF−1, +carbachol 55.7 ± 15.3 pA pF−1, wash 11.6 ± 4.2 pA pF−1(n = 9)). Thus M3-mediated percentage inhibition was reduced from 78.7 ± 2.7 %(n = 16) in the absence of PLCβ1ct to 23.5 ± 10.6 %(n = 9, P < 0.001) in its presence. These experiments were performed in the HKIR3.1/3.2/M3 stable cell line. E, bar chart summarizing the effects of co-expression of RGS2 upon M1 and M3 responses (M1+RGS2: basal 22.3 ± 3.8 pA pF−1, +carbachol 74.0 ± 21.3 pA pF−1, wash 15.0 ± 2.5 pA pF−1(n = 5); M3+RGS2: basal 47.7 ± 10.0 pA pF−1, +carbachol 110.3 ± 36.3 pA pF−1, wash 38.9 ± 10.0 pA pF−1(n = 8)). RGS2 reduced M1-mediated percentage inhibition from 62.1 ± 4.8 %(n = 23) to 29.6 ± 8.7 %(n = 5, P < 0.01) and M3-mediated percentage inhibition from 47.6 ± 4.4 %(n = 31) to 20.0 ± 11.7 %(n = 8, P = 0.01). * P < 0.05; ** P < 0.01.

The potential involvement of PKC in mediating the inhibitory response

GTP-bound Gq/11 activates phospholipase Cβ leading to the hydrolysis of phosphatidylinositol-4,5-bisphosphate (PIP2) to form diacylglycerol (DAG) and inositol-1,4,5-trisphosphate (IP3). DAG is then able to activate classical and novel isoforms of PKC. We first investigated the potential involvement of PKC using a number of inhibitors of this enzyme. We found that the non-selective serine/threonine protein kinase inhibitor staurosporine (1 μm) attenuated the M3-mediated channel inhibition by approximately 35 % (Fig. 6a). We then used the more selective compounds bisindolylmaleimide I (GF109203X; 3 μm) and Ro-31-8220 (3 μm). In our experiments both compounds attenuated the M3 inhibitory effects, by approximately 60 and 44 %, respectively (Fig. 6a). Secondly we sought to mimic the inhibitory effects of M1 and M3 receptors by directly activating PKC independently of receptor stimulation using either the phorbol esters phorbol-12-myristate-13-acetate (PMA) and phorbol-12,13-dibutyrate (PDBu) or the synthetic DAG analogue 1,2-dioctanoyl-sn-glycerol (DOG). In these experiments PMA, PDBu or DOG was applied for 3 min to HKIR3.1/3.2 cells and their effects on basal currents measured (Fig. 6B and C). 4αPhorbol (100 nm) was used as a negative control for phorbol esters and was found to have no significant effects on basal current density (Fig. 6C).

Figure 6.

M3-mediated channel inhibition involves a Ca2+-independent PKC isozyme

A, the effects of staurosporine (1 μm), GF109203X (3 μm) and Ro-31-8220 (3 μm) on M3-mediated channel inhibition. PKC inhibitors were applied for at least 5 min prior to M3 receptor stimulation with 10 μm carbachol and the percentage inhibition of basal current density measured. □, percentage inhibition of Kir3.1+3.2A currents by M3 receptor stimulation; ▪, percentage inhibition in the presence of PKC inhibitor. Experiments using the inhibitor GF109203X were performed in HKIR3.1/3.2 cells transiently transfected with M3 receptors whilst the experiments using staurosporine and Ro-31-8220 were done in the stable cell line HKIR3.1/3.2/M3. B, illustration of the effects of DOG (5 μm) on basal Kir3.1+3.2A currents over a 3 min time period. Current traces are shown at 1 min intervals (time points, in minutes, are indicated above each trace) during the 3 min application of DOG and after 5 min wash. C, bar chart summarizing the inhibitory effects of PMA, PDBu (both 100 nm), DOG (5 μm) and 4αphorbol (100 nm) on Kir3.1+3.2A basal current density. For comparison the effects of carbachol on the M3 receptor are also shown. We compared current density before and after application of PDBu, PMA, DOG or 4αphorbol using Student's paired t test. The PKC activators PMA, PDBu and DOG all significantly reduced current density (pre-PDBu: 64.9 ± 15.7 pA pF−1, post-PDBu: 28.3 ± 10.5 pA pF−1, n = 6, P < 0.01; pre-PMA: 32.5 ± 8.6 pA pF−1, post-PMA: 13.7 ± 4.3 pA pF−1, n = 6, P = 0.01; pre-DOG: 35.4 ± 5.2 pA pF−1, post-DOG: 4.7 ± 0.9 pA pF−1, n = 11, P < 0.001). In contrast 4αphorbol had no significant effects upon current density (pre-4αphorbol: 23.72 ± 6.61 pA pF−1, post-4αphorbol: 20.8 ± 3.07 pA pF−1, n = 6, P = 0.7). * P < 0.05;** P < 0.01; *** P < 0.001.

There is currently considerable interest in the direct effects of PIP2 on channel activity. The data presented to date are consistent with a role for the hydrolysis products of PIP2 in channel regulation as well as the depletion of anionic phospholipid itself. We undertook experiments to address the relative contributions of these two processes in the inhibitory response we observed and performed pharmacological occlusion experiments. The HKIR3.1/ 3.2/M3 cell line was treated with DOG (5 μm) and subsequently with carbachol (10 μm) in the presence of DOG, to investigate whether carbachol could induce any further inhibition in addition to that caused by DOG-induced PKC activation. DOG reduced current density by 65.5 ± 3.6 % and subsequent treatment with carbachol caused a further very small additional reduction to 67.9 ± 4.2 % compared to control (n = 6). Carbachol-induced activation of currents still occurred indicating that the receptor was not desensitized by such manipulations (Fig. 7).

Figure 7.

Occlusion of M3 receptor-mediated responses by prior stimulation with DOG

A, current-voltage traces elicited as described in Fig. 1A were recorded in HKIR3.1/3.2/M3 cells. The left-hand panel illustrates a control trace. DOG (5 μm) was then applied until its effects reached a maximum and another trace recorded (centre panel). Carbachol (10 μm) was then applied (in the presence of DOG) for 20 s during which currents were increased (B). After the removal of carbachol another current-voltage trace was recorded (right-hand panel). B, trace recorded at -60 mV from the same cell as illustrated in A. The lower-case letters a-c refer to where data were measured for the bar chart shown in C. C, bar chart illustrating the effects of carbachol on current density when applied after prior inhibition by DOG. Data are normalized to the control current prior to the application of DOG. * P < 0.05.

PKC isoforms endogenously expressed in HEK293 cells

Data thus far indicate that a Ca2+-independent isoform of PKC contributes to mediating the inhibitory response to M1 and M3 receptor stimulation. A number of mammalian PKC isoforms have been identified, which can be divided into three main groups (Mellor & Parker, 1998): the conventional Ca2+- and DAG-dependent (α, βI, βII and γ), the novel Ca2+-independent, DAG-dependent (δ, ε, θ and η) and the atypical Ca2+- and DAG-independent (Π and χ) isoforms. We investigated which isoforms are endogenously expressed in HEK293 cells using Western blotting. Members of each group of isoform were found to be expressed. Strong signals were observed for PKCα, βI, βII and δ whilst PKCγ, ε and χ showed less strong immunoreactivity. PKCη and θ were absent from these cells (Fig. 8a). We examined the translocation of two of these isoforms, PKCδ and PKCε, in response to PMA (100 nm). Both isoforms exhibited translocation from the cytosolic to the particulate fraction in response to PMA but not the inactive analogue 4αphorbol (Fig. 8B).

m1- and M3-mediated channel inhibition prevents subsequent channel activation by Gi/o-coupled receptors

We investigated how inhibition of Kir3.1+Kir3.2A by M1 and M3 receptor stimulation affected subsequent channel activation by Gi/o-coupled receptors. We transfected the Gi/o-coupled A1 adenosine receptor into the HKIR3.1/ 3.2/M3 stable cell line and stimulated this receptor using the non-selective adenosine receptor agonist 5′-N-ethylcarboxyamidoadenosine (NECA, 1 μm) both before and after stimulation of the M3 receptor with 10 μm carbachol. It is clear that NECA-induced currents were reduced following stimulation of the co-expressed M3 receptor (Fig. 9a and B). We ascertained that this was not due to receptor desensitization as repetitive stimulation of the A1 receptor by NECA (1 μm) at the same time points as those indicated in Fig. 9a, in the absence of carbachol stimulation of the M3 receptor, resulted in reproducible levels of current activation (153.1 ± 29.2, 157.6 ± 32.5 and 158.5 ± 33.7 pA pF−1, n = 4). We also performed an additional set of experiments examining the effects of stimulating M1 or M3 receptors upon A1-mediated channel activation in a HEK293 cell line that stably expressed both the Kir3.1+3.2A channel subunits and the Gi/o-coupled A1 adenosine receptor (HKIR3.1/ 3.2/A1; Leaney et al. 2000). We transiently transfected either M1 or M3 receptors into these cells and then looked at the ability of NECA to potentiate currents through the A1 receptor after stimulation of M1 or M3 with carbachol. These data were compared to control A1 receptor responses in cells into which neither the M1 nor the M3 receptors were transfected and in the absence of stimulation with carbachol. Similarly, stimulation of either the M1 or M3 receptors inhibited channel activity such that subsequent stimulation of the currents via the A1 receptor was significantly impaired (Fig. 9C). We also examined whether attenuation of the A1-mediated stimulation could be mimicked by activating PKC in a receptor-independent fashion by using either DOG or PMA. Pre-treatment of HKIR3.1/3.2/A1 cells with 5 μm DOG or 100 nm PMA had similar effects in preventing NECA-induced activation of Kir3.1+3.2A currents (Fig. 9C).

Figure 9.

The effects of M1 or M3 receptor stimulation prevent subsequent channel activation by the Gi/o-coupled A1 receptor

A, individual current traces recorded from a HKIR3.1/3.2/M3 cell transiently transfected with the A1 receptor. The A1 receptor was stimulated with 1 μm NECA prior to and following M3 stimulation with 10 μm carbachol. Time points (in seconds) are indicated above each trace (the experiment was started at t= 0). B, bar chart summarizing the effects of M3 receptor stimulation on A1-induced channel activation recorded in HKIR3.1/3.2/M3 cells transiently transfected with A1 receptors. NECA-induced currents were measured at -60 mV before (95.9 ± 26.2 pA pF−1) and after M3 stimulation (32.8 ± 8.7 pA pF−1, n = 5, P = 0.02). C, bar chart summarizing the effects of M1 and M3 receptor stimulation and DOG and PMA treatment upon NECA-induced Kir3.1+3.2A currents measured in the HKIR3.1/3.2/A1 line. Control NECA-induced currents: 125.9 ± 16.9 pA pF−1, n = 31; NECA-induced currents after M1 stimulation: 38.5 ± 10.0 pA pF−1, n = 13, P < 0.01; NECA-induced currents after M3 stimulation: 17.7 ± 4.8 pA pF−1, n = 12, P < 0.001; NECA-induced currents following DOG: 22.4 ± 6.3 pA pF−1, n = 11, P < 0.001; NECA-induced currents following PMA: 33.15 ± 13.22 pA pF−1, n = 7, P = 0.02. * P < 0.05;** P < 0.01; *** P < 0.001.


The data presented here show that the Gq/11-coupled M1 and M3 muscarinic receptors are able to regulate the cloned counterpart of the neuronal G protein-regulated inwardly rectifying K+ channel in a mammalian expression system. The behaviour is biphasic in nature with a stimulatory and then a profound and long-lasting inhibitory phase. We have also observed qualitatively similar responses mediated via other Gq/11-coupled receptors (α1B and thyrotrophin releasing hormone (TRH) receptors; data not shown). We have also shown that following M3 receptor activation and consequent channel inhibition, the stimulatory response via Gi/o-coupled receptors is significantly attenuated. The inhibitory phase thus represents a profound and prolonged mechanism whereby stimulatory inputs to the channel can be dynamically regulated.

Signalling pathways involved in the response

The key question then is what are the signalling pathways mediating this biphasic response? The transient initial stimulation appears to be mediated through Gi/o for the M1 receptors as the response is PTx sensitive. However the analogous initial response mediated via M3 is not affected by either PTx treatment or over-expression of PLCβ1ct or RGS2. It has been reported that muscarinic receptors can promiscuously interact with multiple G proteins and effectors (Migeon & Nathanson, 1994; Offermans et al. 1994; Nahorski et al. 1997), hence the M3 stimulatory effect may be due to its coupling to PTx-insensitive G proteins such as Gz (Felder, 1995; Rümenapp et al. 2000).

A more intriguing element is the prolonged inhibitory response to M1 or M3 stimulation. Overexpression of PLCβ1ct and RGS2 essentially removes the receptor-mediated inhibitory effect, but does not qualitatively affect the stimulatory effect, implicating Gq/11 and events downstream of the activated G protein in accounting for the response. GTP-bound Gq/11 activates PLCβ resulting in the production of DAG and IP3 from PIP2. The role of PIP2 in regulating the inwardly rectifying family of K+ channels is very topical (Hilgemann & Ball, 1996; Fan & Makielski, 1997; Baukrowitz et al. 1998; Huang et al. 1998; Shyng & Nichols, 1998; Kim & Bang, 1999). A number of laboratories have presented evidence for the physiological depletion of this anionic phospholipid being involved in receptor-mediated inhibition of cloned KATP channels (Xie et al. 1999) and cloned and native Kir3.1+3.4 channels (Kobrinsky et al. 2000; Cho et al. 2000; Meyer et al. 2001).

Our data support a role for the hydrolysis products of PIP2 in Kir3.1+Kir3.2A channel regulation, in particular DAG activation of PKC. The receptor-mediated inhibitory response could be mimicked by DOG and phorbol esters and was Ca2+ independent, suggesting the involvement of a novel PKC isoform. Immunoblotting showed that PKCδ and PKCε are expressed in HEK293 cells and that both of these isoforms translocate in response to PMA. Furthermore, if we activated PKC prior to receptor stimulation we could largely occlude the receptor-mediated inhibition. PKC inhibitors partially, but not completely, block the inhibition. Thus on balance our data support a primary role for PKC, in particular PKCδ and/or PKCε, in mediating the inhibitory response. However, direct PIP2 depletion cannot be excluded from playing a role. It remains to be established whether PKC directly phosphorylates the channel complex or other signalling component(s) to cause its functional effects. The identification of the phosphorylation sites on the channel might enable the relative importance of PKC activation and PIP2 depletion to be elucidated. Our observations are consistent with three previous studies in Xenopus laevis oocytes implicating PKC in Kir3.0 channel regulation through the metabotropic glutamate receptors (Sharon et al. 1997), bombesin receptors (Stevens et al. 1999) and M1 muscarinic receptors (Hill & Peralta, 2000).

Physiological significance

Are there any indications that these phenomena are of physiological importance? To our knowledge there are no instances of channel activation occurring through the Gq/11-coupled M1 and M3 receptors in native tissues and this may be a feature of receptor expression in heterologous systems. For example, after PTx treatment of atrial myocytes, Kobrinsky et al. (2000) only saw inhibition of Kir3.1+3.4 channels after stimulation of the M3 receptor. However, current potentiation was observed following activation of endothelin-A receptors (Yamaguchi et al. 1997).

Inhibition of G protein-gated inwardly rectifying K+ channels has been well documented in a number of native settings. For instance in atrial myocytes an inhibition of current mediated by endothelin-A receptors (Yamaguchi et al. 1997, Meyer et al. 2001), α1-adrenoceptors (Braun et al. 1992; Meyer et al. 2001) and M3 muscarinic receptors (Kobrinsky et al. 2000) has been observed. In central neurones it has been found by Nakajima and colleagues that stimulation of the Gq/11-coupled neurotensin receptors leads to depression of a G protein-gated inwardly rectifying K+ channel (Koyano et al. 1993; Takano et al. 1995; Velimirovic et al. 1995).

Thus it is established that inhibitory phenomena mediated via Gq/11-coupled receptors are physiologically important in Kir3.0 channel regulation. Importantly, are the mechanisms of inhibition different between that observed in heterologous expression systems, by ourselves and others, and that seen in native tissues? There is a growing consensus that PIP2 depletion is a major contributor in atrial myocytes (Kobrinsky et al. 2000; Cho et al. 2000; Meyer et al. 2001) though it is not clear whether it accounts for slow inhibition and/or rapid desensitization. The mechanistic detail for the regulation of the current in central neurones is less clear and there is, to our knowledge, no conclusive data suggesting the importance of anionic phospholipid depletion. The data presented here do implicate downstream hydrolysis products in reproducing equivalent channel regulation.

It is becoming apparent from immunocytochemical and physiological studies that G protein-coupled inwardly rectifying K+ channels are localized on the dendritic tree in spines and shafts in perisynaptic areas (Liao et al. 1996; Ponce et al. 1996). The net result of such inhibitory effects would be to significantly attenuate inhibitory inputs and reinforce excitatory ones. It is possible that the central excitatory effects of muscarinic receptor stimulation could be mediated in part by inhibition of Kir3.0 channel activity, in addition to modulation of currents such as the M-current (Brown & Adams, 1980).

In conclusion, it is apparent from our data that the M1 and M3 muscarinic receptors are able to both activate and subsequently inhibit the current response. We have evidence suggesting that a substantial proportion of this inhibitory effect is mediated via PKCδ and/or PKCε. Kir3.1+3.2A channels inhibited in the above fashion are significantly attenuated in their ability to be stimulated by Gi/o-coupled receptors. The PKC regulation of Kir3.0 channels may form the basis for the dynamic modulation of excitability. This work forms the basis for future studies; namely, whether direct channel phosphorylation occurs and the basis of the interaction of signalling components with the channel.


This work is supported by the Human Frontiers Science Programme, the Royal Society and the Wellcome Trust. J.L.L. is a Royal Society Dorothy Hodgkin Fellow, L.V.D. is supported by the Wellcome Trust and A.T. is a Wellcome Trust Senior Research Fellow in Clinical Science. We thank the following people for providing us with cDNAs: B. Adams (PLCβ1ct), T. Bonner (M1 receptor), L. Y. Jan (M2 receptor), A. Tobin (M3 receptor), E. G. Peralta (M4 receptor) and E. Reuveny (RGS2). We would also like to thank Z. Hafeez for technical help.