Muscarinic Inhibition of Calcium Current and M Current in Gα_q-Deficient Mice

Abstract

Activation of M₁ muscarinic acetylcholine receptors (M₁ mAChR) inhibits M-type potassium currents (I_K(M)) and N-type calcium currents (I_Ca) in mammalian sympathetic ganglia. Previous antisense experiments suggested that, in rat superior cervical ganglion (SCG) neurons, both effects were partly mediated by the G-protein Gα_q (Delmas et al., 1998a; Haley et al., 1998a), but did not eliminate a contribution by other pertussis toxin (PTX)-insensitive G-proteins. We have tested this further using mice deficient in the Gα_q gene.

PTX-insensitive M₁ mAChR inhibition ofI_Ca was strongly reduced in Gα_q −/− mouse SCG neurons and was fully restored by acute overexpression of Gα_q. In contrast, M₁mAChR inhibition of I_K(M) persisted in Gα_q−/− mouse SCG cells. However, unlike rat SCG neurons, muscarinic inhibition of I_K(M) was partly PTX-sensitive. Residual (PTX-insensitive)I_K(M) inhibition was slightly reduced in Gα_q −/− neurons, and the remaining response was then suppressed by anti-Gα_q/11 antibodies.

Bradykinin (BK) also inhibits I_K(M) in rat SCG neurons via a PTX-insensitive G-protein (G_q and/or G₁₁; Jones et al., 1995). In mouse SCG neurons,I_K(M) inhibition by BK was fully PTX-resistant. It was unchanged in Gα_q −/− mice but was abolished by anti-Gα_q/11 antibody.

We conclude that, in mouse SCG neurons (1) M₁ mAChR inhibition of I_Ca is mediated principally by G_q, (2) M₁ mAChR inhibition ofI_K(M) is mediated partly by G_q, more substantially by G₁₁, and partly by a PTX-sensitive G-protein(s), and (3) BK-induced inhibition of I_K(M) is mediated wholly by G₁₁.

The excitability of neurons is controlled by the convergent action of many different ion channels, including voltage-gated K⁺ and Ca²⁺ channels and ligand gated ion channels. N-Type Ca²⁺ channels have complex effects on cell excitability. They permit Ca²⁺ influx at presynaptic terminals, triggering exocytosis, thereby contributing to excitability. However, they also act to limit cell excitability by providing the Ca²⁺ necessary to open Ca²⁺-activated K⁺ channels, resulting in a prolonged afterhyperpolarization and action potential accommodation (Davies et al., 1996; Sah, 1996). The M-type potassium current (I_K(M)) also acts to dampen cell excitability; it increases upon depolarization, hyperpolarizing the cell and keeping it clamped around rest and thus limiting action potential firing (Brown, 1988; Wang and McKinnon, 1995). Both the N-type Ca²⁺ channel andI_K(M) can be inhibited by G-protein-coupled receptors, such as the muscarinic acetylcholine M₁ receptor (M₁ mAChR) (Marrion et al., 1989; Bernheim et al., 1992), and coincident inhibition of both currents results in large increases in cell excitability and action potential discharge (Jones and Adams, 1987;Brown, 1988, 1999).

In rat superior cervical ganglion (SCG), inhibition ofI_Ca andI_K(M) by agonists acting at the M₁ mAChR appears to use very similar transduction mechanisms, and it has been suggested that they may share a common pathway (Hille, 1994). The first step in this pathway is the activation of a G-protein that is insensitive to pertussis toxin (PTX) (Brown et al., 1989; Bernheim et al., 1992). Previous experiments using the injection of antibodies directed against the common C terminus of Gα_q and Gα₁₁ (Caulfield et al., 1994) and the expression of separate antisense constructs specific to Gα_q or Gα₁₁(Delmas et al., 1998a; Haley et al., 1998a) suggested that the G-protein primarily involved is Gα_q. However, because suppression by either method was incomplete, the participation of other G-protein(s) could not be discounted.

Bradykinin (BK) also inhibits I_K(M) in rat SCG neurons through stimulation of B₂ BK receptors (Jones et al., 1995). This effect was strongly attenuated by anti-Gα_q/11 antibodies (Jones et al., 1995), but the specific G-protein involved has not been further identified. It may, in fact, differ from that required for mAChR-induced inhibition because BK-induced inhibition clearly involves the activation of phospholipase C, whereas M₁ mAChR-induced inhibition appears not to (Cruzblanca et al., 1998; Haley et al., 1998b).

In an attempt to further define the role of Gα_qin transducing the effects of M₁ mAChR and B₂ BK receptors, we have therefore investigated the inhibition of I_Ca andI_K(M) by agonists for these receptors in SCG neurons isolated from mice lacking Gα_q(Offermanns et al., 1997a,b).

Parts of this work have been published previously in abstract form (Haley et al., 1998c).

MATERIALS AND METHODS

Gα_q-Deficient mice.Mice deficient in the Gα_q gene (Gα_q −/−) were generated by targeted disruption with a neomycin gene as described previously (Offermanns et al., 1997a). Gα_q-Deficient mice used in the experiments were obtained by mating either Gα_qknock-out males with heterozygous females or heterozygous males and females (to obtain wild-type and knock-out littermates). Mice were kept on a C57BL × 129/Sv background, and genotypes were confirmed by PCR on genomic DNA from tail snips of each mouse.

Cell culture. Sympathetic neurons were isolated from Gα_q knock-out and wild-type mice that were at least 5 weeks old. Wild-type mice were either C57BL/6J (Harlan, Bicester, UK) or from litters also containing Gα_q knock-out littermates; results obtained were the same from both these groups and so have been pooled. We also used ICR mouse (coding for outbred albino mouse; Harlan) to compare properties of SCG neurons derived from different strain of mouse. SCG were removed, and neurons were cultured using standard procedures as described previously for rat (Delmas et al., 1998b).

Microinjection. In a few experiments, Gα_oA+B or Gα_q/11antiserum (OC2 and CQ1 respectively, from G. Milligan) (Caulfield et al., 1994, Delmas et al., 1998a, 1999) or an expression plasmid encoding Gα_q were pressure-injected into the cytoplasm (antisera) or nucleus (plasmid) of SCG neurons 2 d in culture using a microinjector (Eppendorf, Hamburg, Germany). To allow identification of the injected neurons for recording, FITC-dextran (70,000 MW; Molecular Probes, Leiden, The Netherlands) was added in a final concentration of 0.2% (antisera) (Jones et al., 1995) or 0.5% (plasmid) (Haley et al., 1998a). Cells injected with the antisera were recorded at least 2 hr after injection, whereas those injected with the Gα_q-encoding plasmid were recorded 24 hr later.

Electrophysiology. Currents were measured from SCG neurons cultured for 2–3 d, using the amphotericin-B perforated patch technique (Horn and Marty, 1988; Rae et al., 1991). Patch electrodes (2–5 MΩ) were filled by dipping the tip for 40 sec into the appropriate filtered internal solution, and the pipette was then back-filled with the internal solution containing 0.07–0.1 mg/ml amphotericin-B. High-resistance seals (>2 GΩ) were initially achieved and, after amphotericin-B permeabilization, access resistances were generally <25 MΩ for I_K(M)recordings and <15 MΩ for I_Carecordings. SCG neurons were perfused (5–10 ml/min) with an external solution consisting of (in mm): NaCl 120, KCl 3, HEPES 5, NaHCO₃ 23, glucose 11, MgCl₂ 1.2, CaCl₂ 2.5, and tetrodotoxin 0.0005, pH 7.4, maintained at 32°C.

I_K(M) recording. The internal solution comprised of (in mm): potassium acetate 80, KCl 30, HEPES 40, and MgCl₂ 3, adjusted to pH 7.3–7.4 with KOH (280 mOsm/l with K acetate). Cells were voltage-clamped using an Axoclamp-2A (switching frequencies 3–5 kHz, filter 0.1 kHz) from Axon Instruments (Foster City, CA).I_K(M) was measured as a slowly developing inward deactivation relaxation after a 1 sec jump to a command potential of approximately −45 mV from a holding potential of approximately −25 mV (Haley et al., 1998a). Inhibition was measured as the fractional reduction in the amplitude of theI_K(M) deactivation relaxation in response to either cumulatively increasing concentrations of oxotremorine methiodide (Oxo-M) (Research Biochemicals, Natick, MA) or a single application of 1 nm BK (Bachem, Torrance, CA).

I_Ca recording. Perforated patch recordings were conducted primarily as described previously (Delmas et al., 1999). Pipettes had a resistance of 2–3 MΩ when filled with the following solution (in mm): CsCl 30, caesium acetate 110, HEPES 10, and MgCl₂ 1, pH 7.2–7.3 with CsOH (300 mOsm/l). Cells were voltage-clamped using an Axopatch 200A amplifier (Axon Instruments). Series resistance and membrane capacitance were partially compensated (>80%). Current traces were low-pass filtered at 2–5 kHz using a four-pole Bessel filter. Leak and capacitance currents were subtracted digitally using the P/6 subtraction procedure of pClamp6 (Axon Instruments). I_Ca was elicited from a holding potential of −70 mV with a three-voltage pulse protocol consisting of a test pulse to 0 mV applied before and after a conditioning depolarizing step to +90 mV.

For experiments with Bordetella PTX (Speywood, Maidenhead, Berkshire, UK), SCG neurons were incubated with 1 μg/ml PTX in the culture medium for at least 24 hr before recording. Data are expressed as mean ± SEM, and statistical analysis of dose–response curves used two-way ANOVA to compare treatments across all concentrations. If a significant effect of treatment was found overall, further analysis was performed using the two-way ANOVA to determine which treatment groups contributed to this significance. The bradykinin data were analyzed with one-way ANOVA, and if an overall effect of treatment was found, this was followed by Students–Newman–Keuls multiple comparison test. Analysis of I_Ca andI_K(M) current densities and neuron membrane potential was analyzed using Student's t test.p < 0.05 was considered significant.

Immunocytochemistry. Successful injection of Gα_o or Gα_q/11 antiserum was confirmed by staining. After recording, cells were fixed in acetone, washed, incubated with biotinylated Fab2 swine anti-rabbit antibody (the antiserum was raised in rabbit), and then incubated with avidin–biotin complex. The alkaline phosphatase substrate 5-bromo-4-chloro-3-indoxyl phosphate/nitro blue tetrazolium chloride (BCIP/NBT) (Dako, Carpinteria, CA) was applied for 5 min, and the reaction was quenched with water. Neurons containing the injected antiserum were clearly distinguished as containing the dark purple BCIP/NBT product. Detection of expressed Gα_qafter injection of the Gα_q plasmid was confirmed using the above protocol and preceding it with a 1 hr incubation with a 1:2000 dilution of a Gα_qantibody (IQB2) (Milligan et al., 1993).

RESULTS

M₁ mAChR-induced inhibition ofI_Ca and I_K(M) in mouse versus rat SCG

The mAChR agonist Oxo-M inhibited bothI_Ca andI_K(M) (see also Hamilton et al., 1997) in SCG from mouse. However, the dose–response curves were shifted to the right in mouse compared with rat neurons (Fig.1). IC₅₀ values forI_K(M) inhibition were 0.4 and 0.7 μm in rat and wild-type mouse, respectively. For M₁ mAChR inhibition ofI_Ca, IC₅₀ were 800 nm and 1.1 μm in rat and mouse, respectively.

Fig. 1.

Dose–response curves (mean ± SEM plus best-fit curve) for Oxo-M inhibition of I_Caand I_K(M) in rat and mouse SCG.I_Ca, Left, Oxo-M dose–response curves determined after treatment with PTX, thereby isolating the M₁ mAChR-mediated component. nvalues for each concentration point range from 3 to 7.I_K(M), Right, Oxo-M dose–response curve in wild-type mouse SCG (n = 9) is significantly different from that seen in rat SCG (n = 4; p < 0.001). The ratI_K(M) dose–response curve data are taken from Haley et al. (1998a), and the rat I_Cadose–response curve data are from Delmas et al. (1998a).Legend applies to both panels. All cells were recorded using the perforated patch method.

M₁ mAChR inhibition of I_Cais reduced in Gα_q −/− mouse SCG neurons

In rat SCG, inhibition of I_Ca by mAChR agonists results from the activation of two separate pathways: a voltage- and PTX-sensitive M₄ mAChR pathway (M₂ receptors in mice) (Shapiro et al., 1999) that requires Gα_oA and a voltage- and PTX-insensitive M₁ mAChR pathway that mostly involves Gα_q (Delmas et al., 1998a). In mice, total mAChR inhibition of I_Ca was significantly (p < 0.002) reduced when Gα_q was deleted. Thus, 1 and 10 μm Oxo-M resulted in 49 ± 4% (n = 8) and 68 ± 4% (n = 12) inhibition, respectively, in wild-type (Gα_q+/+) mouse neurons and 29 ± 3% (n = 5) and 57 ± 3% (n = 5) inhibition in Gα_q −/− neurons. Inhibition in Gα_q −/− neurons was more voltage-dependent than in wild-type neurons, with facilitation ratio of 1.4 ± 0.2 (n = 10) and 0.94 ± 0.1 (n = 8), respectively (data not shown). Inhibition by 10 μm noradrenaline (NA) (which requires both Gα_oA and Gα_i in rat) (Caulfield et al., 1994; Delmas et al., 1999) was not altered when Gα_q was absent (wild type, 66 ± 3%,n = 8; Gα_q −/−, 62 ± 5%, n = 5).

Incubation with 1 μm PTX inactivates members of the Gα_o/i G-protein family, removes the PTX-sensitive mAChR inhibition, and isolates the M₁ mAChR inhibition, as well as abolishes inhibition by NA (Schofield, 1991; Zhu and Ikeda, 1994). Inhibition by NA was, as expected, abolished in PTX-treated cells from both wild-type and Gα_q −/− mice. After PTX treatment, it became clear that the reduction in M₁ plus M₂ inhibition observed in the Gα_q −/− cells resulted from an almost complete loss of the voltage-insensitive M₁mAChR-mediated inhibition (Fig. 2). To be sure that the loss of M₁ mAChR inhibition ofI_Ca was attributable to the absence of Gα_q, we acutely overexpressed Gα_q in cultured SCG neurons from Gα_q −/− mice. Twenty-four hours after injection of a plasmid encoding for Gα_q, M₁ mAChR inhibition was fully restored (Fig.2C). This reinstated inhibition was specific because NA inhibition was not rescued by Gα_qoverexpression in PTX-treated cells (Fig. 2). These findings accord with previous data from this laboratory that M₁mAChR inhibition of I_Ca in rat SCG is primarily mediated by Gα_q (Delmas et al., 1998a). It should be noted that theI_Ca density (normalized to cell capacitance) was significantly lower in Gα_q−/− neurons (27 ± 2 pA/pF, n = 10) compared with wild-type mouse SCG (38 ± 2 pA/pF, n = 12,p < 0.01). There was no difference betweenI_Ca density in wild-type mouse and rat SCG (rat, 42 ± 2 pA/pF, n = 19).

Fig. 2.

M₁ mAChR inhibition ofI_Ca is markedly reduced in Gα_q−/− mice. All experiments were performed on neurons pretreated with PTX. A, I_Ca waveforms in SCG neurons from wild-type (left) and Gα_q−/− mice (right) in the absence and presence of 10 μm Oxo-M, using the three-step voltage protocol illustrated above each waveform. The conditioning step to +90 mV did not reverse Oxo-M inhibition in either wild-type or Gα_q−/− neurons, confirming the voltage-independent nature of the PTX-insensitive inhibition. Calibration: 5 msec. B, Gα_q is expressed in a Gα_q −/− neuron after injection of a Gα_q expression plasmid (injected cell indicated by arrow). C, In Gα_q −/− neurons, Oxo-M inhibition ofI_Ca is restored by exogenous expression of Gα_q. D, Summary of M₁ mAChR and noradrenergic inhibitions of I_Ca. Note that NA inhibition cannot be restored in Gα_q −/− SCG neurons by expression of Gα_q. n ranges from 6 to 10 for Oxo-M data and from 4 to 7 for NA data.

M₁ mAChR inhibition of I_K(M)is not reduced in SCG from Gα_q −/− mouse

In rat SCG, inhibition of I_K(M)by mAChR agonists is mediated, at least in part, by Gα_q (Haley et al., 1998a), so one would predict a loss of inhibition in neurons lacking Gα_q. We were surprised, therefore, to discover that inhibition ofI_K(M) by Oxo-M was not reduced in Gα_q −/− SCG compared with wild-type neurons (Fig. 3). Indeed, Oxo-M produced significantly more inhibition in Gα_q −/− neurons (p < 0.05), and the dose–response curve lay to the left of the wild-type dose–response (Fig.3B). I_K(M) density (normalized to cell capacitance) was not changed in Gα_q −/− mouse SCG compared with wild type, and neither was the resting membrane potential (wild type, 2.5 ± 0.5 pA/pF, n = 8; −58.0 ± 1.3 mV,n = 9; Gα_q −/−, 3.3 ± 1.1 pA/pF, n = 9; −56.3 ± 1.3 mV,n = 11), and these did not differ from rat SCG (2.4 ± 0.3 pA/pF, n = 10; −59.9 ± 1.3 mV).

Fig. 3.

M₁ mAChR inhibition ofI_K(M) is not reduced in Gα_q−/− mice. A, I_K(M)deactivation relaxation elicited by a −20 mV step for 1 sec from a holding potential of −25 mV. Waveforms (average of 6 traces) are from SCG neurons of wild-type (top) and Gα_q−/− (bottom) mice and are shown in the absence and presence of increasing concentrations of Oxo-M (micromolar).Dotted lines indicate 0 pA. Calibration: 250 pA.B, Mean ± SEM data (plus best-fit curves) for Oxo-M inhibition of I_K(M) in wild-type and Gα_q −/− SCG neurons. Oxo-M dose–response curve in Gα_q −/− neurons (n = 8) was significantly different from wild type (n = 9;p < 0.05). IC₅₀ and Hill slope values for inhibition in Gα_q −/− cells are 0.6 μm and 1.1, respectively. C, Oxo-M dose–response curves in wild-type and Gα_q −/− SCG neurons treated with PTX. IC₅₀ and Hill slope values were 1.3 μm and 0.93, and 1.7 μm and 1.16 in wild-type and Gα_q −/− cells, respectively.

M₁ mAChR inhibition of I_K(M)partly involves PTX-sensitive G-proteins in mouse SCG

One factor contributing to the persistence of mAChR-induced inhibition of I_K(M) in neurons from Gα_q-deficient mice became apparent when we tested the effect of PTX. In contrast to previous observations on rat SCG neurons (Brown et al., 1989; Bernheim et al., 1992; Haley et al., 1998a), PTX treatment significantly reduced inhibition by Oxo-M in neurons from both wild-type mice (p < 0.01) and Gα_q −/− mice (p < 0.0001) (Fig. 4). In fact, the effect of PTX was greater in Gα_q-deficient neurons, and the dose-response curve for the PTX-insensitive component of inhibition showed a significant (p < 0.02) rightward shift in Gα_q −/− neurons. Thus, at 0.3 μm Oxo-M, the mean inhibition ofI_K(M) was reduced from 15 ± 3 to 6 ± 4%. Hence, it appeared that Gα_qcontributed a component of PTX-insensitive inhibition at low agonist concentrations but that, at high agonist concentrations, activation of another PTX-insensitive G-protein predominated.

Fig. 4.

PTX-sensitive G-proteins and Gα₁₁ mediate Oxo-M inhibition ofI_K(M) in Gα_q −/− neurons.A, Representative waveforms (1 sec step from −25 to −45 mV) recorded from Gα_q −/− neurons treated with PTX and injected with either Gα_oA+B antibody or Gα_q/11 antibody. Whereas cumulative concentrations of Oxo-M (micromolar) still produce inhibition in the Gα_oA+Bantibody-injected cell, there was very little inhibition in presence of the Gα_q/11 antibody. Dotted lines indicate 0 pA. Calibration: 250 pA. B, Mean ± SEM data (plus best-fit curves) for Oxo-M inhibition ofI_K(M) in wild-type and Gα_q−/− neurons. PTX treatment significantly reduced Oxo-M inhibition in both wild-type (n = 8; p < 0.01) and Gα_q −/− neurons (n = 10;p < 0.0001). Injection of a Gα_oA+Bantibody had no effect on PTX-treated cells from either wild-type (n = 7) or Gα_q −/− (n = 7) mice and served as an injection control. Injection of a Gα_q/11 antibody reduced the remaining PTX-insensitive inhibition in both the wild-type (n= 4; p < 0.0001 compared with Gα_oA+Bantibody-injected cells) and Gα_q −/− (n = 5; p < 0.0001 compared with Gα_oA+B antibody-injected cells) neurons.Legend applies to bothpanels. Insets demonstrate the presence of the injected antibodies in SCG neurons (injected cells indicated byarrow): left, Gα_oA+Bantibody in wild-type neuron; right, Gα_q/11 antibody in Gα_q −/− neuron.

One possibility is that this residual PTX-insensitive inhibition might have been mediated by G₁₁, because expression of constitutively activated Gα₁₁ is also capable of inhibiting I_K(M) in rat SCG neurons (Haley et al., 1998a). To test this, we injected antibodies directed against the C terminus of Gα_q/11 into the cell cytosol of PTX-treated neurons, using antibodies against Gα_o as a control (Caulfield et al., 1994). Anti-Gα_q/11 strongly reduced PTX-insensitive inhibition in neurons from wild-type mice and virtually annulled the residual inhibition in neurons from Gα_q −/− mice (p < 0.0001). Thus, the residual PTX-insensitive inhibition in neurons from Gα_q −/− mice results from activation of Gα₁₁ because the antibody is specific to Gα_q and Gα₁₁, and Gα_q is absent.

Inhibition of I_K(M) by BK is mediated by Gα₁₁ in Gα_q −/− neurons

For these experiments, a single application of a low concentration (1 nm) of BK was made to each neuron tested because slow recovery from inhibition precluded repeated applications to the same neuron and desensitization precluded applications of incremental concentrations (Jones et al., 1995; Cruzblanca et al., 1998). At 1 nm, BK inhibited I_K(M) in wild-type mouse SCG neurons by 25 ± 4% (n = 8). This was not significantly different from that in rat SCG neurons (29 ± 7%; n = 8; data not shown). Unlike mAChR-induced inhibition, the effect of BK on mouse neurons was not reduced by PTX (Fig. 5). This accords with previous observations on rat neurons (Jones et al., 1995). No inhibition of I_Ca by 1 nm BK could be detected.

Fig. 5.

Gα₁₁ mediates BK inhibition of I_K(M) in Gα_q −/− neurons. A, Representative waveforms (1 sec step from −25 to −45 mV) recorded from Gα_q −/− neurons treated with PTX and injected with Gα_oA+B (left) or Gα_q/11(right) antibody. BK (1 nm) inhibitedI_K(M) in Gα_oA+B-injected neurons but not when the Gα_q/11 antibody was injected. The dotted lines represent 0 pA. Calibration: 250 pA.B, Mean ± SEM data for inhibition by 1 nm BK in wild-type and Gα_q −/− neurons. Neither PTX treatment nor injection of the Gα_oA+Bantibody significantly altered inhibition by BK in either wild-type or Gα_q −/− neurons. Injection of the Gα_q/11antibody reduced BK inhibition in both wild-type and Gα_q−/− cells (p < 0.01 for both compared with PTX plus Gα_oA+B antibody).

No significant reduction of BK-induced inhibition ofI_K(M) was observed in Gα_q −/− neurons. However, injection of an antibody against Gα_q/11 substantially (p < 0.05 vs anti-G_oantibody) reduced I_K(M) inhibition in wild-type mouse neurons (as in rat neurons; Jones et al., 1995) and abolished inhibition in neurons from Gα_q −/− mice (p < 0.01) (Fig. 5B).

DISCUSSION

Using mice lacking Gα_q, we have demonstrated that M₁ mAChR inhibition ofI_Ca requires Gα_q because regulation was essentially lost in Gα_q −/− neurons (Fig. 2). This was not a result of secondary effects on transduction mechanisms arising from an absence of Gα_q during development because inhibition could be fully restored by acute overexpression of Gα_q in Gα_q −/− neurons (Fig. 2); thus, the remaining components of the inhibitory pathway were still present and functional in these mutant cells. In contrast, neither the inhibition produced by NA nor the PTX-sensitive (M₂) component of mAChR inhibition (Shapiro et al., 1999) were reduced after deletion of the Gα_q genes. Hence, these results in mouse neurons are in complete accord with previous conclusions from observations on rat SCG neurons using G-protein antibody injections and antisense depletion (Caulfield et al., 1994; Delmas et al., 1998a) that Gα_q is the primary G-protein involved in M₁ mAChR-induced inhibition ofI_Ca but is not involved in the inhibition produced by activating M_2/4 mAChRs or α₂ adrenergic receptors. It is especially worth noting (particularly in connection with our observations onI_K(M); see below) that Gα₁₁ appears not to be able to substitute for Gα_q in Gα_q −/− mice, although it is able to interact with M₁ mAChRs (Offermanns et al., 1994; Gudermann et al., 1996) and although expression of a constitutively active form of Gα₁₁ can inhibitI_Ca just like Gα_q (unpublished observations; Delmas et al., 1998a).

In contrast [and surprisingly, in view of previous observations in rat SCG neurons using Gα_q antisense (Haley et al., 1998a)], mAChR-induced inhibition ofI_K(M) was not reduced in neurons from Gα_q −/− mice. One contributory reason for this seemed to be that I_K(M)inhibition after mAChR stimulation was partly sensitive to PTX. When this component was eliminated, it appeared that some part of the inhibition (particularly at low agonist concentrations) was mediated by Gα_q because the dose–response curve for Oxo-M-induced M current inhibition was shifted significantly to the right in Gα_q −/− mice. Nevertheless, it was clear that a considerable proportion of the inhibition must have been mediated by another PTX-insensitive G-protein. In Gα_q-deficient neurons, this G-protein was identifiable as Gα₁₁, because the residual PTX-insensitive inhibition was virtually annulled using an antibody directed against the unique but common C terminus to Gα_q and Gα₁₁. Thus, in the Gα_q-deficient mice, we conclude that mAChR-induced M current inhibition is maintained by Gα₁₁ and by an increased contribution of the (unidentified) PTX-sensitive pathway.

Thus, the present experiments reveal two clear differences from the inferences drawn from previous work on mAChR-induced M current inhibition in rat SCG neurons: the additional involvement of a PTX-sensitive G-protein and the greater potential involvement of Gα₁₁. One possible explanation for the apparent involvement of a PTX-sensitive G-protein is that, in mouse ganglion cells, stimulation of M₄ or M₂ receptors might also inhibitI_K(M). However, this seems unlikely, because Hamilton et al. (1997) found that mAChR-induced M current inhibition was completely annulled in mice deficient in the M₁ receptor gene, implying that, as in rat neurons (Marrion et al., 1989; Bernheim et al., 1992), muscarinic inhibition of I_K(M) was mediated entirely by M₁ mAChRs. Although M₁ receptors are usually considered to couple exclusively to PTX-insensitive G-proteins, there have been occasional reports of at least partial PTX-sensitive responses after expression of cloned M₁ receptors (Stein et al., 1988;Ashkenazy et al., 1989). Because this PTX-sensitive pathway seems unique to mouse neurons and may have limited general significance, we have not so far made any serious attempt to identify the species of G-protein involved.

Regarding the involvement of Gα₁₁, in previous experiments on rat neurons, anti-Gα_q antisense produced a rightward shift of the dose–response curve for inhibition of I_K(M) by Oxo-M (Haley et al., 1998a), not dissimilar to that seen in Gα_q−/− mouse neurons. However, although this might also suggest the involvement of another PTX-insensitive G-protein, antisense to Gα₁₁ had no effect on the inhibitory effect of Oxo-M in rat neurons. Hence, it was concluded that the limited effect of Gα_q antisense probably resulted from incomplete protein suppression rather than to the additional effect of another G-protein. Nevertheless, it is necessary to point out that, although Gα₁₁ can sustain mAChR-induced M current inhibition in Gα_q-deficient mice, we have no direct evidence from the present experiments that Gα₁₁ mediates any part of the response of normal (wild-type) mouse neurons to Oxo-M; the reduced inhibition seen in the presence of the Gα_q/11 antibody could equally well have been attributable to antagonism of either endogenous Gα_q or Gα₁₁. The large contribution of Gα₁₁ in Gα_q −/− neurons might then be an adaptive change, perhaps resulting from a redistribution of G-proteins in the plasma membrane, because there is no evidence for compensatory overexpression of Gα₁₁ in the nervous system of these mice (Offermanns et al., 1997b). Hence, the present observations do not necessarily negate our previous conclusion that mAChR-induced M current inhibition in normal rat neurons results primarily from activation of Gα_q. In this context, it should be noted that previous observations have revealed differences between the coupling of muscarinic receptors to ion channels in mouse and rat SCG neurons. Thus, in the mouse, inhibition ofI_Ca is mediated by M₂ receptors (Shapiro et al., 1999) whereas the corresponding response in rat SCG neurons is mediated by M₄ receptors (Bernheim et al., 1992); indeed, although rat neurons possess functional M₂receptors capable of activating PTX-sensitive G-proteins, they appear not to inhibit I_Ca(Fernandez-Fernandez et al., 1999).

The effects of BK on I_K(M) in mouse ganglion cells seem to match those on rat SCG neurons much more closely. Thus, 1 nm BK produced the same amount of inhibition in both, and neither showed any sensitivity to PTX (unlike muscarinic inhibition). Previous experiments on rat neurons using G-protein antibodies (Jones et al., 1995) strongly suggested that inhibition was mediated by either Gα_q or Gα₁₁ (or both) but could not identify which. Because in the present experiments inhibition was unchanged in neurons from Gα_q −/− mice and then annulled by anti-Gα_q/11 antibody, we infer that inhibition in normal SCG neurons is probably mediated exclusively by Gα₁₁ (although with the caveat regarding possible adaptive responses after Gα_q deletion expressed above). Lack of any involvement of Gα_q would accord with the fact that, unlike mAChR stimulation, BK did not inhibitI_Ca in the mouse ganglion cells.

The reason for this apparent selectivity of BK receptors for Gα₁₁ in mouse neurons (and possibly in rat neurons) is unclear. Although it is usually assumed that BK can activate Gα_q, we are unaware of any experiments that unequivocally distinguish effects mediated by Gα_q from those that might equally be mediated by Gα₁₁. On the contrary, Ricupero et al. (1997) have reported that BK-induced stimulation of phospholipase C was enhanced, rather than inhibited, in mouse embryonic stem cells lacking Gα_q, whereas Wilk-Blaszczak et al. (1994b)identified G₁₃ as the G-protein responsible for the PTX-insensitive component of I_Cainhibition in neuroblastoma hybrid cells. [In previous experiments, these authors (Wilk-Blaszczak et al., 1994a) had identified G_q/11 as mediating the phospholipase C-driven activation of I_K(Ca)in these cells but, because this was from anti-G_q/11 antibody infusion, it could have resulted from activation of G_q,G₁₁, or both.]

In conclusion, the present results point to a surprising degree of divergence at the G-protein level in the coupling of M₁ mAChR and B₂ BK receptors to Ca²⁺ and K_M⁺ channels in SCG neurons (Fig. 6). Thus, in the mouse ganglion, M₁ mAChR inhibition ofI_Ca appears to be mediated exclusively (or almost so) by Gα_q, whereas inhibition ofI_K(M) also involves both Gα₁₁ and an additional unidentified PTX-sensitive G-protein. BK does not inhibitI_Ca, and its inhibition ofI_K(M) is probably mediated exclusively by Gα₁₁. This divergence at the level of G-protein coupling may go some way toward explaining the apparent divergence in the subsequent steps to M current inhibition after activation of M₁ mAChRs or B₂ BK receptors, namely, that BK-induced inhibition appears to involve activation of a phospholipase C, whereas M₁ muscarinic inhibition does not (Cruzblanca et al., 1998; Haley et al., 1998b). The difference in G-protein coupling from M₁ mAChRs to Ca²⁺ channels and K_M⁺ channels might equally imply that, contrary to previous inferences (Hille, 1994), the modulation of these two channels may also involve different transduction pathways.

Fig. 6.

Diagram to summarize projected G-protein involvement for M₁ mAChR- and B₂ BK receptor-induced inhibition of I_Ca andI_K(M) channels in rat SCG neurons and in wild-type and Gα_q-deficient mouse SCG neurons.